When Language Breaks Down PDF

This page intentionally left blank
When Language Breaks Down
Doctors, nurses, and other caregivers often know what people with Alzheimer’s
disease or Asperger’s ‘sound like’ – that is, they recognize patterns in people’s
discourse from sounds and silences, to words, sentences, and story structures.
Such discourse patterns may inform their clinical judgements and affect the
decisions they make. However, this knowledge is often tacit, like recognizing
a regional accent without knowing how to describe its features. The absence
of explicit knowledge of discourse patterns may be partly because research
and practice associating neurocognitive function with language has tended to
focus on (often isolated) linguistic ‘deficits’ as signs or symptoms of brain
injury or disorder rather than beginning with comprehensive descriptions of
discourse. In contrast, this is the first book to present models for comprehen-
sively describing discourse specifically in clinical contexts and to illustrate
models with detailed analyses of discourse patterns associated with degen-
erative (Alzheimer’s) and developmental (autism spectrum) disorders. The
authors also suggest how clinical discourse analysis, combined with neu-
ropsychological and imaging data, can add to our understanding of neurocog-
nition. The book is aimed not only at advanced students and researchers in
linguistics, discourse analysis, speech pathology, and clinical psychology but
also at researchers, clinicians, and caregivers for whom explicit knowledge of
discourse patterns might be helpful.
Elissa D. Asp is Associate Professor of English Linguistics in the English

Department and Linguistics Coordinator of the Linguistics Programme
at Saint Mary’s University. She is also Adjunct Professor in the Faculty of
Medicine at Dalhousie University.
Jessica De Villiers is Associate Professor in the Department of English at

the University of British Columbia.
When Language Breaks Down
Analysing Discourse in Clinical Contexts
Elissa D. Asp and Jessica de Villiers

CAMBRIDGE UNIVERSITY PRESS
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore,
São Paulo, Delhi, Dubai, Tokyo
Cambridge University Press

The Edinburgh Building, Cambridge CB2 8RU, UK
Published in the United States of America by Cambridge University Press, New York
www.cambridge.org
Information on this title: www.cambridge.org/9780521889780
© Elissa Asp and Jessica de Villiers 2010
This publication is in copyright. Subject to statutory exception and to the

provision of relevant collective licensing agreements, no reproduction of any part
may take place without the written permission of Cambridge University Press.
First published in print format 2010
ISBN-13 978-0-511-67769-4 eBook (NetLibrary)

ISBN-13 978-0-521-88978-0 Hardback
ISBN-13 978-0-521-71824-0 Paperback
Cambridge University Press has no responsibility for the persistence or accuracy

of urls for external or third-party internet websites referred to in this publication,
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.
Contents
List of figures page vi

List of tables vii
Acknowledgements viii
Transcription conventions x
Introduction 1
1 Introduction to clinical discourse analysis 4
2 Theoretical and clinical contexts 10
3 Talk and speech – conversation analysis and intonation in English 29
4 Grammar 44
5 Phase and contexts of culture and situation 85
6 Study design 93
7 Differential diagnosis and monitoring 114
8 Cognitive models, inferencing and affect 135
9 Modelling information across domains 168
Closing remarks 207
Notes 211
Appendix A: Some basic grammatical terminology and relations 215
Appendix B: Inventory of codes 222
References 225
Author index 249
Subject index 253
v
Figures
2.1 Context, language and discourse relations to

neurocognitive domains and inquiry types page 23
2.2 Interdisciplinary relations in clinical discourse analysis 26
3.1a Tone 1: falling 37
3.1b Tone 1: falling 38
3.2a Tone 2: rising 39
3.2b Tone 2: falling-rising (pointed) 39
3.3 Tone 3: low rising 40
3.4 Tone 4: falling-rising 41
3.5 Tone 5: rising-falling 42
3.6 Tone 2, 2, 1: listing sequence 43
4.1 Options for statements 47
4.2 Options for exclamations 48
4.3 Options for questions 49
4.4 Options for commands (Gregory 1988) 52
4.5 Options for speech functions (adapted from Gregory 1988) 53
4.6 Argument roles – feature synopsis 68
4.7 Circumstantial roles 70
4.8 Referential relations 73
4.9 Conceptual/lexical relations network (from Gregory in
Watt 1990a) 77
4.10 Cohesive analysis 79
4.11 Morphosyntactic resources for marked focus and prominence 84
5.1 Model phasal analysis 91
6.1 Evaluating confabulation, gist, and verbatim recall 108
8.1 Frame for ‘buying a ticket to the movies’ 139
8.2 Decisions and brains (Coricelli et al. 2005) 141
9.1 Feature analysis for The magic of the universe 173
9.2 Visual representation of The magic of the universe logic 180
9.3 Visual representation of propositions in My Mom 184
9.4 Feature analysis for My Mom 186
vi
Tables
3.1 Terms for conversation analysis page 33

4.1 Basic speech function contrasts 45
4.2 Modalization, attitude and evaluation 56
4.3 Conceptual hierarchy with features 58
4.4 Example of predicational and feature analysis 72
4.5 Errors in use of cohesive devices 78
6.1 Narrative performance features 101
6.2 Narrative task scale 102
6.3 Four research questions with pros and cons of study designs 110
6.4 Research design and text type 112
7.1 MLU-w, lexical density and syntactic complexity in the
weather text 117
7.2 Lexical items and repetition counts in the weather text 119
7.3 Pedantic battery 126
7.4 Modalization battery 129
7.5 Modalization features in the discourse of a person with AD 131
7.6 Lexical patterns in the discourse of a person with AD 132
8.1 Gricean maxims in Alzheimer’s disease and Asperger
syndrome 155
9.1 Normative patterns for conversation and relevance construction 200
9.2 Types of information and related systems 203
9.3 Summary of discourse and contextual features 205
vii
Acknowledgements
We would like to thank our friends and families for their encouragement and
support while we worked on this book. To John Foster and Arnaud Goupilliere
we are especially indebted – for making us tea and dinner and cleaning up
afterwards, for making us go out in the sunshine sometimes and for cheer-
fully putting up with us while we worked. We would also like to thank our
mentors. Elissa is especially indebted to Dr Kenneth Rockwood, Professor
of Medicine (Geriatric Medicine and Neurology) at Dalhousie University,
Halifax, Nova Scotia. As well as mentoring her in developing research on
discourse in Alzheimer’s disease and providing access to data and patients,
Dr Rockwood, together with his team, have consistently modelled what inter-
disciplinary research and clinical care can be. Elissa was also privileged to be
invited to participate in the Halifax Symposia on the Treatment of Alzheimer’s
disease from 2003 to 2008 which brought outstanding international scholars in
many disciplines to address particular themes associated with Alzheimer’s and
dementia. Elissa also wants to thank Jennifer Klages, for generously giving
her access to data on vascular cognitive impairment, the Nova Scotia Health
Research Foundation for their support and Saint Mary’s University which has
twice given her sabbatical leave to work on the book and projects related to it,
and her colleagues in the English department who may flinch but vote in favour
of leave to research discourse effects of cholinesterase inhibitors and other
similarly unliterary topics.
Jessica would like to thank Peter Szatmari, whose mentorship and encour-
agement made much of this work possible. He has been a great advisor, collab-
orator and supporter. She would also like to thank the Department of English
at UBC, her research team and language group, the Faculty of Arts at UBC
and the Offord Centre for Child Studies at McMaster University for support-
ing her work in many ways. She is very grateful to the Offord Centre for Child
Studies for giving her access to data and other research supports. She grate-
fully acknowledges financial support from the Social Sciences and Humanities
Research Council and the Canadian Institute of Health Research.
We also both want to acknowledge the mentoring of Michael Gregory. He
taught us much of what we know about language and discourse analysis, and
viii
Acknowledgements ix
although he didn’t believe linguists had any business talking about brains or neu-
rology, we think he would have liked this book. Other linguists who influenced
our work because of their own are Noam Chomsky, William Downes, Jonathan
Fine, Ruqaiya Hasan, Michael Halliday, Rodney Huddleston, Richard Hudson,
Ray Jackendoff and Sydney Lamb. We also would like to thank reviewers and
readers of our manuscript in its various stages for helpful and encouraging
comments, and Andrew Winnard for believing in the project and supporting
us in it. Finally, we thank the individuals and families who participated in our
studies. Our work is for them.
Transcription conventions
SP1 speaker 1
SP2 speaker 2
IV interviewer
CG caregiver
P patient (in caregiver and patient interviews)
CHI child
RES researcher (in semi-structured conversations with researcher)
(1) line 1
(5) line 5
( ) material in parentheses is inaudible or there is doubt of accuracy
(( )) double parentheses indicate clarifying information, e.g.
((laughter))
(.) a pause which is noticeable but too short to measure
(.5) a pause timed in tenths of a second
# noticeable pause, unmeasured
## long pause, unmeasured
: colon indicates an extension of the preceding vowel sound
<text>[>] overlaps following text
<text> [<] overlaps preceding text
… some text intervening
Bold bold is for prominent information (presented as new or
emphasized)
Under underscore is for information that is the focus (unmarked starting
point) of a construction
//text// tone group
//2 text// tone 2
xxx unintelligible
[if] uncertain
^ silent beat
Many of the transcription conventions used are borrowed or adapted from Gail
Jefferson’s system. (See for example Jefferson 1985; 2004.)
x
Introduction
This book is predicated on the relatively uncontentious notions that discourse

patterns – what people do when they talk or write – can provide trained obser
vers with information about cognitive functions and affective states in speakers
and, further, that cognitive functions and affective states may be signs of integ-
rity of neurological function and structure. Neurolinguists, psycholinguists,
aphasiologists, psychiatrists, psychotherapists and speech pathologists all take
some variation on assumptions like this as their point of departure in study-
ing brain–behaviour relationships and treating some neurological and affective
disorders. However, discourse – people’s talk and text – is inherently complex
and apparently unstable and, worse, the neurological substrate and processes
that support even superficially simple things like ‘how words are represented
in the brain’, let alone ‘what happens in brains when people talk’ are matters
of active debate and investigation rather than scientific givens. In the face of
so much uncertainty and complexity, most of the work done on language–
brain relationships has, very sensibly, centred on theoretically discrete and/or
methodologically isolatable phenomena associated with particular semantic,
morphosyntactic or phonological structures or processes. Work on discourse in
clinical environments as another means of investigating neurocognitive (dys-)
function, although often called for, has been less common.
This situation is changing now because of technological developments and,
we think, a sea-change-like shift that is taking place in attitudes to brain–
behaviour relationships. On the technological side, recent developments in neu-
roimaging techniques are providing new tools to investigate neural structure,
chemistry and function, and developments in machine-mediated text analysis
tools, storage and search capacities have made corpus-based discourse studies
much more doable. The change in attitudes to brain–behaviour relationships is
also at least partly technologically mediated insofar as imaging and other tech-
niques enabling in vivo investigation of the effects of cognitive activity suggest
that behaviours can have measurable effects not just on activation patterns but
also on neurochemical and neuroplastic (structural) responses. What is novel in
this is not that behaviour can alter neurochemistry and structure – therapy for
people with brain injury or dysfunction presupposes and evidences this. Rather,
1
2 When Language Breaks Down
it is that the new technologies can make changes observable and measurable,
and so present new possibilities both for understanding brain–behaviour rela-
tionships and, consequently, for developing new therapies to help people with
neurological disorders or injuries. There are other factors that contribute to
this drift – salient among them are recognition of the limits and risks of phar-
macological interventions and an increased, computationally mediated, capac-
ity to conceptualize complex interactions. The first three factors suggest that
people interested in neurological disorders and diseases should have access to
very detailed accounts of the discourse patterns (and other behaviours) of the
populations that they study and treat; the computational capacity to store and
process the data produced by such studies means projects that used simply to
be unworkable can now be fruitfully undertaken.
It is in this environment that we offer this book as a first pass at ‘clinical dis-
course analysis’ or CLDA. It is intended as an introduction to the use of struc-
turally, pragmatically and linguistically based discourse analysis techniques to
investigate relationships between discourse behaviours and patterns and neu-
rocognitive (dys)function in clinically defined groups. Because we work with
teams specializing in the care of people with autism spectrum disorders and
degenerative dementias, most of the examples we use refer to discourse sam-
ples from these groups. However, the techniques that we discuss and model for
discourse analysis were originally developed for description of normal speech
and writing and are applicable to any sort of speech sample, including corpora
representative of the speech associated with other neurological disorders.
Our primary audience is discourse analysts (including linguists and cognitive
scientists) – senior undergraduate or graduate students, faculty and research-
ers interested in investigating relations between discourse and neurocogni-
tive functions. For instance, we see the book as a useful adjunct to courses
in discourse analysis and clinical linguistics. However, it should also be of
value to nurses, speech pathologists, clinical psychologists, neurologists and
psychiatrists interested in the potential of discourse analysis (or working with
discourse analysts) for informing clinical judgements of diagnosis and change
and for addressing their own research questions. Finally, we wanted our book
to be interesting and readable for non-professionals, especially caregivers,
interested in Alzheimer’s disease, autism spectrum disorders or just generally
in discourse and neurocognitive function. So, although some parts of the book
are unavoidably technical, we have worked to make the descriptions of dis-
course patterns in Alzheimer’s and autism speakers accessible for a general
audience. Readers will no doubt let us know whether or not we succeeded.
In the chapters that follow, we first orient readers to clinical discourse analy-
sis (chapter 1) and the theoretical and clinical contexts and disorders our work
engages (chapter 2). Chapters 3–5 present descriptive resources which allow
coding of spoken discourse in terms of conversation analysis and intonation
Introduction 3
(chapter 3), grammatical resources for meaning (chapter 4), and means for
extracting patterns from these and relating the patterns to contexts of culture
and situation presented as articulated aspects of memory (chapter 5). Chapters
6 and 7 address questions in study design associated with various discourse
tasks and model applications for diagnosis (in autism spectrum disorders) and
treatment monitoring (in Alzheimer’s disease). Chapter 8, on cognitive models,
inferencing and affect, and chapter 9 on modelling information across domains,
situate the analytic constructs presented in neurocognitive and clinical perspec-
tives through review of relevant neuropsychological, imaging and lesion stud-
ies and through detailed illustrations of the analyses and inferential processes
involved in clinical discourse analysis. We close with remarks about the future
and potential for clinical discourse analysis.
1 Introduction to clinical discourse analysis
Discourse represents that aspect of mental activity that most clearly reflects
the intimate and over-lapping connections among cognition, language, and
communication.
(Ulatowska et al. 1985)
1.1 What is clinical discourse analysis?

Clinical discourse analysis is the term we use to describe the analysis of
language behaviour observed in clinical contexts. Language behaviour includes
well-defined areas of clinical research addressing syntax, vocabulary, phon-
ology, conversation skills and cohesion. It also includes areas less commonly
described in clinical research such as argument roles, situational features and
functional variation. The focus of clinical discourse analysis is natural lan-
guage behaviour which requires examination of all these aspects of language
use. Even the smallest of texts require analysis that can explore multivariate
features.
Consider the following sentence:
(1) I can remember my Mom.
It is extracted from a brief sample of spoken discourse elaborated below. The
sentence has an interactional function: it makes a statement which is modalized
for capacity (can). It has a predicate that references a cognitive process (remem-
ber) and two argument roles, an experiencer (I) and a percept (my mom). It has
an unmarked clause order: subject occurs first in English statements unless
there is a reason to emphasize some other element of structure. One might also
assume that the speaker interprets the situation as an informal one in that she
chooses to refer to her mother as Mom rather than mother.
The larger text from which this example is taken appears in Text 1.1 below.
Text 1.1 My Mom

Ah the same thing with the – what was it we just discussed? – the stove. I can remem-
ber my Mom, she was as smart as a ticket. She was ninety-six years old when she died
4
Introduction to clinical discourse analysis 5
and she would, you know, she had everything under control. But I knew I knew that
I couldn’t do that I couldn’t go and I don’t know. Now I’m lost again.
The speaker is a 76-year-old woman, Cleo, with moderate Alzheimer’s dis-
ease (AD). She is responding to a question about her ability to use the stove,
and commenting on her mother’s competence in old age as compared with her
own difficulties. Grammatically, her speech is well formed. Most sentences
are complete. They have appropriate subject–verb agreement, and subordinate
and co-ordinate clause structures that are typical for her age group (Mackenzie
2000; Kemper et al. 2001b). She uses idioms she was smart as a ticket, she
had everything under control, and a metaphor I’m lost again. She also uses
cohesive features such as pronouns and other referring expressions appropri-
ately, with reference supplied either in prior or subsequent text. However, there
are conversation and fluency features which reflect planning difficulties and
repairs: she hesitates (the same thing with the) and checks reference (what
was it we just discussed?) but then supplies it herself (the stove). She has false
starts (she would) which she repairs (you know, she had …); repetition (I knew
I knew); and one predication (I couldn’t go) appears incomplete or tangential in
that it does not refer to anything in either prior or subsequent discourse.
Cleo’s abilities and difficulties, including her ability to monitor and repair
her discourse, and her explicit recognition of difficulty (I don’t know, I’m lost
again) are characteristic for her age and stage of Alzheimer’s (Asp et al. 2006a).
Recognizing that this pattern is typical requires not only that all its elements be
described, but also that samples described address both intra- and inter-individual
variations relative to diagnosis, dementia phase, potential treatment effects and
contexts of use. Thus clinical discourse analysis inherently requires both frame-
works that enable comprehensive descriptions of language in use and the develop-
ment of specialized text collections, or ‘corpora’, representative of the language
used by speakers.
As the above brief description suggests, clinical discourse analysis not only
involves description of formal linguistic features such as syntactic structures,
but also characterizes patterns of meaning which may be relevant in under-
standing neural function in speakers. Consider Text 1.2: it is a transcript of
a conversation between a research technician and a six-year-old boy with
autism.
Text 1.2 Lions

(1) CHI: what would scratch you.
(2) CHI: wouldn’t that be terrible?
(3) RES: umhum.
(4) CHI: awful!
(5) CHI: stink.
(6) CHI: and then you hate lions.

(7) CHI: stink hairy.
(8) …
(9) RES: do you like lions?
(10) CHI: na I hate lions.
(11) RES: you hate <lions> [>]?
(12) CHI: <bushy> [>] and hairy stinky.
(13) …
(14) RES: what’s your favourite animal?
(15) CHI: an I hate lions.
(16) RES: yeah.
(17) RES: you hate lions.
(18) RES: but what what animal do you like?
(19) CHI: stinky and then I li ugly:
(20) res: how about bunny rabbits?
(21) RES: do you like bunny rabbits?
(22) CHI: yes:
(23) RES: do you?
(24) CHI: why yes!
The text reflects some normal features of conversational interaction. Ted
initiates a topic (lions). He knows when it is his turn to speak and when to
let others have a turn. He develops his topic, giving characteristics of lions
(bushy, hairy and stinky). And he expresses his attitude to lions (he hates them).
However, when asked to shift topic and say what his favourite animals are, he
doesn’t collaborate in topic development but rather repeats the points that are of
interest to him, that lions are stinky and ugly and he hates them. He also repeats
hairy, stinky and I hate lions. Together, inability to shift focus and repetition,
particularly of single words and phrases, create an identifiable pattern in this
text. For instance, it suggests that Ted has difficulties managing topic and may
have trouble staying with the drift of the conversation. Of course, such features
may occur in the discourse of children who do not have autism. However, if
such patterns appear as normative rather than exceptional in the discourse of
an individual or group with diagnoses of autism spectrum disorders (ASDs),
they may in fact reflect discourse patterns characteristic of ASDs. Identifiable
recurrent patterns are the business of clinical discourse analysts.
Clinical discourse analysis is not as such a theory or discipline. Rather it is
a goal-directed set of practices aimed at describing and explaining language
behaviours as a means of investigating neurocognitive function. This implies a
need for theoretical and descriptive flexibility. At present, there are few studies
which investigate and attempt to fully characterize natural language behaviour
of speakers with particular neurological disorders or diseases, although there
is widespread recognition of the need for such work. We attempt to address
this need by combining conversational analysis with comprehensive linguistic
description of functions and structures as they relate to contextual variables.
As we use it, the term ‘discourse analysis’ refers to the types of description
mentioned and not its more widely used sense associated with, for instance, the
‘discourse of capitalism’ or the ‘discourse of libertarianism’.
1.2 What use is clinical discourse analysis?

Clinical discourse analysis can characterize language behaviour (i.e. discourse)
from which inferences can be drawn about neurocognitive function. Discourse
is a sensitive sign of global and specific function. In clinical contexts, its ana-
lysis can enable the development of tools for diagnosis and evaluation of
treatment response. These may supplement existing measures and provide
information for developing new therapies. Comprehensive descriptions of dis-
course patterns produced by speakers with neurological disorders may also
lead to new understanding of brain–behaviour relationships.
Moreover, clinicians internalize the characteristic behaviours of the treatment
groups they meet, and may use this information in making clinical judgements.
Discourse analyses can make the basis for such clinical judgements explicit,
replicable and generalizable. Such explicit characterizations can lead to the
development of useful models for researchers, healthcare workers and families
and thus help people to understand the behaviours they recurrently notice.
1.3 What use is this book?

Over the last thirty-five years, there has been increasing recognition that lan-
guage behaviour is supported by a wide range of neural capacities, including
attentional and memory systems, and that it is context dependent. While there
is significant interest, a growing literature, and some established descriptive
techniques and norms, there is as yet no agreed upon and validated set of prac-
tices which allow comprehensive analyses of language behaviour in clinical
contexts. In the chapters that follow, we articulate a range of descriptive and
theoretical tools and examples that may be useful for these purposes.
1.4 Sample analysis

Below we present and discuss in non-technical language, an example of dis-
course analysis. We illustrate how discourse techniques can highlight distinct,
observable patterns of language behaviour which may be signs of neurocogni-
tive function.
Text 1.3 The magic of the universe

(1) CHI: what’s my favourite what?
(2) RES: your favourite game on the computer.
(3) CHI: well there’s ex # well there’s uh # eh # there’s the there’s this strange unusual
game.
(4) CHI: uh well # there’s a la a computer called an IBM Aptiva comes with games.
(5) CHI: uh # like my favourite is the # is from I is from a: place where there’s a k.
(6) CHI: it’s the game’s about # it’s a it’s about a light bodied cube # k running get-
ting the opposite colour on another light force called endorfun which is spelled
e n d o r f u n.
(7) RES: umhum?
(8) CHI: and uh uh: light bodied cubes flying everywhere.
(9) CHI: and I have the power.
(10) CHI: I feel the magic of the universe.
(11) CHI: And et cetera et cetera et cetera.
(12) RES: is this a game you play by yourself James?
(13) RES: or with a partner?
(14) CHI: just myself.
(15) RES: hm.
(16) CHI: I am really completely good at it.
The speaker, James, is fifteen years old and has been diagnosed with autism. In
this text he has a conversation with a researcher about his interests. James takes
turns appropriately. In the first paragraph, he uses an echo question for clarifi-
cation of a request for information. Subsequently in lines (3–6) and (8–10) he
responds to and develops the request for information, identifying his favour-
ite game as strange, unusual, the computer that it comes on (4), describing
the game itself (5–6) and the player’s role (9–10). In (14) he responds, again
appropriately, to a question about the number of participants and evaluates
his own ability as a player (16). This is very much a two-way conversation.
James stays on topic throughout and pauses long enough at regular intervals to
allow the interviewer to give feedback (umhum? (7), and hm (15)) and to ask
for details (12, 13). He doesn’t need to be prompted for topic development.
His syntactic structures are varied with some simple and some co-ordinate and
complex structures.
James’ discourse is also characterized by pedantic features and dysfluencies.
Specifically, he repeats information and phrases, supplies technical details, and
has some initial difficulties organizing his talk. He uses redundant attributes
(strange, unusual (3)) and degree modifiers (really, completely (16)). He fully
repeats the phrase light bodied cubes when he refers to it a second time and
repeats et cetera twice. He introduces technical details using one kind of gram-
matical structure, a reduced relative clause using call. He gives the brand name
of the computer (an IBM Aptiva) on which the game is found and provides the
spelling for endorfun. His description of the player’s role (I have the power;
I feel the magic of the universe) comes from the game and has a rehearsed
quality.
In topic initiation there is also repetition: well there’s, repeated three times,
suggests the topic is in fact being reinitiated from the beginning (well is
normally discourse or topic initial (Schiffrin 1987)). In fact, there are marked
difficulties in setting the topic. There are eight pauses, five hesitations, eleven
false starts and three repaired clauses (3, 4, 6) in the first half of the discourse
where James is describing his favourite computer game without actually nam-
ing it. His dysfluency occurs perhaps because he does not know or remember
the name of the computer game. The dysfluency features disappear when James
is talking about his role in the game and his speech becomes more formulaic.
Cumulatively, the amount of repetition, the technical specificity and formu-
laic elements are features typical of pedantic speech in autism. This is thought
to occur across autism spectrum disorders (de Villiers et al. 2007). Even an
informal analysis of a short sample such as James’ The magic of the universe
isolates specific features which contribute to the characterization of ‘pedan-
tic speech’. While the neural substrates of autism spectrum disorders are not
currently known, articulating how dysfluency and pedantic speaking pattern
together may shape research questions about neurocognitive function in autism
spectrum disorders.
The steps by which one moves from observation and description of a dis-
course pattern in an individual or a group with a diagnosed neurological disorder
to hypothesizing possible neurophysiological cause(s) for the pattern are only
a beginning in understanding brain–behaviour relationships. Hypotheses, once
generated, need to be checked if they are to be of any use. Checking requires
designing research projects. For these, a clinic setting really is essential. Even
if discourse data for clinical populations were readily available outside clinic
settings, there are other issues of access. Access to accurate diagnostic infor-
mation for participants, to neuropsychological and neurological expert opinion
and evaluation, to neuroimaging as a potential source of information about
neural structure and/or function are all essential if hypotheses are to be inves-
tigated in ways that have the potential to be useful. And for obvious ethical
reasons, utility is a goal of research in clinical discourse analysis.
That said, knowing how to analyse the data, and being willing to work with
and in interdisciplinary teams is enough to begin with. We hope this book will
be useful to people who might be so inclined and, paraphrasing Orange and
Kertesz (2000: 173), that clinical discourse analyses will become a window
into the cognitive, linguistic and social performances of people with neurologi-
cal disorders.
2 Theoretical and clinical contexts
2.1 Introduction
This chapter is intended to situate clinical discourse analysis in terms of
relevant linguistic and non-linguistic fields and to orient readers to the
developmental and degenerative disorders discussed. Sections 2.2 and 2.3
briefly sketch diagnostic criteria, epidemiological information, current treat-
ment options and potential associations with neurophysiology in each area.
Section 2.4 focuses on the theoretical background and sources for clinical
discourse analysis. These include conversation analysis, ethnographic and
interactional sociolinguistics, functional linguistic discourse analysis, cogni-
tive and philosophical pragmatics, and formal (generative) linguistic models.
Section 2.5 addresses the roles of neurology, neuropsychology, psychiatry
and neuroimaging as essential in developing understanding of relationships
between discourse behaviours and neurological disorders. Finally, section 2.6
addresses the role of normative discourse patterns in evaluating descriptions
of the discourse of clinical groups.
2.2 Autism spectrum disorders

Autism spectrum disorder (ASD) is an umbrella term for a continuum of neurode-
velopmental disorders, the causes of which are unknown. ASD manifests during
infancy and is estimated to affect one in every 165 children (Fombonne et al.
2006). The first account of autism was published by Leo Kanner (1943). Since
that time, an expansion in diagnostic criteria has led to the inclusion of more
diagnostic categories in the autism spectrum. ASD now includes autism, Asperger
syndrome and pervasive developmental disorders not otherwise specified.
ASDs affect more than one domain of functioning and are generally charac-
terized by three core deficits:
1) impairments in socialization and interaction (e.g. lack of shared attention,
lack of peer relationships),
2) impaired language and communication (delay or lack of functional speech,
difficulties with conversation and pragmatics),
10
Theoretical and clinical contexts 11
3) stereotyped behaviours and restricted interests (e.g. a desire for sameness or

routine). These behaviours vary depending on an individual’s age and level
of functioning (American Psychiatric Association 1994).
In trying to understand the underlying causes of the three core deficits of

autism, research has focused on cognition, neurology and genetics. Cognitive
explanations have postulated as underlying causes: difficulty in theory of
mind processing (Baron-Cohen et al. 1985; Baron-Cohen 1995); executive
dysfunction (Russell 1998; Hill 2004) and a unique, locally oriented process-
ing style (Frith and Happé 1994; Happé and Frith 2006). Some investiga-
tions of the neural circuitry associated with these explanations have found
evidence of disturbances in functional connectivity. Specifically, a number of
functional neuroimaging studies have found underconnectivity in autism and
a processing style involving reduced reliance on prefrontal regions (Just et al.
2004; Koshino et al. 2005; 2008). Genetic studies looking at inherited risk for
autism have found evidence that genetic factors play a role in autism. Current
genetic research focuses on risk factors as well as genes that have an effect
on variation in clinical expression (Nicholson and Szatmari 2003; Szatmari
2003).
Autism is a lifelong disorder. Given the high prevalence of ASDs, the need
for services to diagnose and treat autism is overwhelming. With varying needs,
and high costs of early intervention, treatment options are limited and much
research is directed toward improving detection and intervention. Perhaps
the most studied treatment option for ASDs, Applied Behavioural Analysis
(Cooper et al. 2006), engages children in socialization, communication and
behavioural learning with one-on-one teaching for up to forty hours per week.
Other interventions may involve parent education, support programmes, advo-
cacy, self-advocacy and combinations of physical and speech therapies, medi-
cation and social skills training. Often interventions focus on changes in the
environment that can expand opportunities for engagement and social partici-
pation, including the communicative patterns of people who enter into social
networks with people with ASDs. For example, peers and caregivers may be
taught ways to start a conversation with someone with an ASD. They may also
learn ways to use discourse that help facilitate conversational participation.
While some people with less severe forms of ASD can function well and even
excel without treatment, many people with autism are unable to function in
everyday life into adulthood.
Communication difficulties in ASDs are widely varied. Early on, parents
may observe an absence of communicative behaviours such as poor eye con-
tact, lack of shared attention and lack of playing peekaboo. Many children
with autism never develop functional verbal communication, while others
have speech that stands out as being unusual. Frequently observed speech
characteristics of autism include flat or unusual intonation (Fine 1991; Fine

et al. 1991; de Villiers et al. 2007), the use of repetitive or stereotyped language
(Tager-Flusberg 1996) and echolalia (Prizant and Duchan 1981). In Asperger
syndrome, language development may be typical or even advanced, with
complex syntax and wide vocabulary. Stereotyped phrases are often used and
speech may have a scripted or pedantic quality, resembling a written register
(Ghaziuddin and Gerstein 1996; de Villiers et al. 2007). Casual conversation
often poses a distinct challenge for people with Asperger syndrome, especially
in topic management, sustained conversation and contextual relevance. Overall
the communication difficulties found in ASDs may be characterized as making
insufficient use of contextual information in social language use (de Villiers
2005).
While there is no single speaking style that uniquely reflects speech dif-
ficulties in ASDs, the following text, from a 12-year-old speaker with autism
illustrates one kind of difficulty in conversation:
Text 2.1 Free Willy

(1) RES: do you have some movies that you like?
(2) RES: what’s the last movie that you saw?
(3) RES: what was it about?
(4) RES: did you see a movie at Christmas time?
(5) CHI: <yeah: ## yeah> ((faint)).
(6) RES: pardon?
(7) CHI: yeah.
(8) CHI: I did.
(9) RES: what movie did you see?
(10) CHI: Free Willy ((faint)).
(11) RES: pardon?
(12) CHI: Free Willy.
(13) RES: Free Willy.
(14) RES: was it good?
(15) CHI: good.
(16) RES: I’ve never seen that movie.
(17) RES: can you tell me what it’s about Patrick?
(18) CHI: it’s Free Willy.
(19) RES: it’s Free Willy.
(20) CHI: I watched that show.
(22) RES: did you watch it at home?
(23) RES: hm?
(24) CHI: ## yeah.
(25) RES: did you watch it at your house?
(26) CHI: yeah.
(27) RES: tell me about Christmas Patrick.
(28) RES: did you have a good Christmas?
(29) CHI: yeah.

(30) CHI: I have a good Christmas.
(31) RES: what did you get for Christmas?
(32) RES: did you get presents?
(33) CHI: presents.
(34) RES: what did you get?
(35) CHI: I get: toys.
(36) CHI: I should get toys.
(37) RES: get toys?
(38) RES: what kind of toys?
(39) CHI: I: um I # didn’t bring it.
(40) RES: pardon?
(41) CHI: I bring it.
(42) RES: you’ll bring it?
(43) RES: did you bring it to school?
(44) CHI: ## yeah.
(45) RES: yeah?
(46) RES: what was it?
(48) RES: Free Willy hm.
The speaker, Patrick, is a cooperative participant in the conversation, who
nevertheless speaks very quietly and at times appears to require prompting
in response to questions or requests for information (2–4, 23, 28, 32). While
he follows the rules of turn-taking, he rarely offers elaboration. Patrick also
presents difficulties with tense forms (I have a good Christmas (30); I get toys
(35); I bring it (41)). In addition, there is some confusion about whether or not
he has brought a toy (or video), Free Willy, to school (39–44). Patrick’s talk
thus reflects difficulties in morphosyntactic, referential and experiential aspects
of discourse development. He offers too little information and has difficulty
organizing what he does say so that it has obvious relevance to the immediate
context of the researcher’s questions. As we shall show in the following chap-
ters, there are other patterns of ability and challenge within the ASD spectrum.
What they have in common are varying degrees of difficulty in producing the
kinds of contribution that meet the information needs of addressees and/or the
demands of the immediate context.
2.3 Neurodegenerative disorders associated with aging

There are a number of neurodegenerative diseases associated with aging such
as Alzheimer’s disease, vascular cognitive impairment, Lewy Body disease,
Parkinson’s disease, Creutzfeldt-Jakob disease, and the spectrum of fronto-
temporal dementias. With the exception of Parkinson’s disease, they are all
associated with ‘dementia’. Dementia refers to changes in cognitive function
and/or behaviour associated with neurodegeneration rather than identifying a
single disease. While there is some debate as to the combination of features

necessary and sufficient to constitute dementia, there is consensus that such a
diagnosis is to be given when people exhibit chronic changes in cognition and
function which impair their ability to carry out activities of daily living inde-
pendently, and these changes are not caused by psychiatric disorders such as
depression or anxiety, by traumatic brain injury, or other physical causes such
as nutritional deficiency, exposure to toxins, infections or brain tumours.1
Some of these diseases, such as Creutzfeldt-Jakob disease and the fron-
totemporal dementias, are relatively rare. The sporadic type of Creutzfeldt-
Jakob disease associated with aging has an estimated incidence of one in one
million worldwide (Public Health Agency of Canada 1999). The infectious
and familial forms of this disease are rarer still. Estimates of the prevalence of
frontotemporal dementias vary from 2% to 12% of all dementias (Bird et al.
2003; Kertesz 2006; Knapp et al. 2007).2 They constitute a group of neurode-
generative disorders which initially affect frontal cortical regions – judgement,
behaviour and language may be separately or simultaneously impaired. As they
progress, memory and other cognitive functions are also affected. The fronto-
temporal dementias tend to have earlier onset, so that frontotemporal dementia
may account for up to 50% of dementia cases in people less than 60 years old
(Graff-Radford and Woodruff 2007). In rare cases, they are inherited through
an autosomal dominant gene and onset may be earlier. Lewy Body demen-
tia is a neurodegenerative disease which can occur by itself or together with
Parkinson’s and/or Alzheimer’s disease and is estimated to account for between
3 and 26% of all dementias (Zaccai et al. 2005 for review). It is characterized
by the presence of ‘Lewy Bodies’, abnormal proteins called alpha-synuclein
distributed in cortical and subcortical brain regions, as well as by diffuse neu-
ritic plaques (Baba et al. 1998). Dementia with Lewy Bodies most typically
causes fluctuations in cognition, attention and alertness, vivid visual halluci-
nations and impaired motor control (Parkinsonism). As it develops, memory
and other cognitive functions also become impaired. Parkinson’s disease also
involves deposits of the alpha-synuclein protein, initially in subcortical brain
regions (the basal-ganglia). The initial effects of Parkinson’s are usually lim-
ited to impaired motor control and typically develop after age 60, but between
25–40% of people with Parkinson’s later develop dementia (McKeith et al.
1996; Galvin et al. 2001; Nussbaum and Ellis 2003).
The most familiar of the neurodegenerative causes of dementia, because of
their prevalence, are Alzheimer’s disease (AD) and vascular cognitive impair-
ment (VCI). Global statistics for dementia (used inclusively for the neurodegen-
erative diseases associated with aging) suggest that approximately 24 million
people worldwide have dementia, and that 4.6 million people develop dementia
every year (Ferri et al. 2005; Alzheimer’s Association 2007). Projections sug-
gest that the number of people living with dementia will double every twenty
years (Ferri et al. 2005). AD accounts for approximately 60% of dementia

cases. Estimates of VCI suggest that it accounts for about 20% (Hill et al. 1996)
of dementia cases though prevalence may vary by region and gender (Knopman
et al. 2003; Rocca and Knopman 2003; Knapp et al. 2007). Community-based
neuropathological studies suggest that AD and VCI frequently co-occur in indi-
viduals (Massoud et al. 1999; Neuropathological Group 2001; Jellinger and
Attems 2007), suggesting that diagnosis and treatment of ‘mixed dementia’
should be more common (Langa et al. 2004; Jellinger 2007).
AD is normally sporadic (not inherited) and typically affects people after
the age of 65, with about 10% of people over this age being affected. Among
people over 85, estimates of prevalence range from 30 to 50%. In about 5% of
cases, AD is familial, occurring as a consequence of genetic mutation. People
with the autosomal dominant forms of AD may develop symptoms in their 30s
and 40s (Knapp et al. 2007).
Currently, criteria for identifying AD based on the DSM-IV-TR (American
Psychiatric Association 2000) and the recommendations of the National
Institute of Neurological Disorders and Stroke–Alzheimer’s Disease and
Related Disorders (NINCDS-ADRDA) working group include clinical evi-
dence of gradual, progressive development of memory impairment and one or
more impairments in language, motor abilities, ability to recognize and iden-
tify objects, and executive function (McKhann et al. 1984). For a diagnosis of
AD, the DSM-IV further specifies that these deficits significantly impair social
and/or occupational function and are not due to other physical or psychiat-
ric conditions or to toxin exposure. A definite diagnosis of AD can only be
achieved at autopsy. Otherwise diagnoses are ‘probable’ or ‘possible’ and are
clinical judgements which may be supported by neuroimaging or evaluation of
cerebrospinal fluid.3
Some of the structural and functional features of AD neuropathology are
well established. These include the formation of amyloid plaques and neurofi-
brillary tangles first reported by Alois Alzheimer in 1907. Biochemical mark-
ers of these pathological changes in AD may be assessed by cerebrospinal fluid
in vivo (Blennow and Hampel 2003). There is also impaired neurotransmission
associated with loss of cholinergic neurons in the nucleus basilis, reduction of
nicotinic and muscarinic receptors in the cortex, and dysfunction of the major
excitatory neurotransmitter – glutamate (Mesulam 2000; Bleich et al. 2003;
Hynd et al. 2004; Jacob et al. 2007; Kashani et al. 2008). Neuroimaging may
show early localized atrophy in the medial temporal lobe, particularly the hip-
pocampus and entorhinal cortex; reduced temporal-parietal and frontal lobe
function in blood flow and metabolic studies, and altered neurochemical con-
centrations in in vivo imaging of metabolites (Fox et al. 1999; Silverman et al.
2001; Kantarci et al. 2007; Matsuda 2007). Global brain atrophy is evident in
late stage AD (Mesulam 2000).
Although there are a number of biomarkers and established neuropathological

correlates, the actual cause of AD is not known. Studies investigating cause
focus on the role of genetics and molecular mechanisms involved in the depo-
sition of amyloid plaques, the formation of neurofibrillary tangles, cholinergic
dysfunction and neuronal death. Studies of risk factors for the development of
AD point to age, genetics and lifestyle factors such as diet, exercise, social and
intellectual engagement as well as potentially precipitating physical conditions
such as head injury, hypertension, obesity, low oestrogen, elevated cholesterol
and diabetes (Jellinger 2002; Gustafson et al. 2003; Mayeux 2003; Luchsinger
and Mayeux 2004; Blennow et al. 2006). The overall picture with respect to
lifestyle and health conditions is that anything that promotes oxidative stress
and/or increases demands for neuroplastic responses in the brain may increase
the likelihood of developing AD (Mesulam 2000; Mamelak 2007).
2.3.1 Vascular cognitive impairment

Criteria for vascular cognitive impairment (VCI) require that there be evidence
of cerebrovascular disease established by focal neurological signs and symp-
toms and/or by brain imaging; that cerebrovascular disease is judged to be
causally connected to dementia (if present), that the onset is typically abrupt,
and that the disease course is stepwise and fluctuating (Roman et al. 1993).
Prevalence estimates for VCI suggest that it is the second most common cause
for dementia, accounting for approximately 20% of all dementias in Europe
and North America (Hebert et al. 2000; Rocca and Knopman 2003; Knapp et al.
2007). As with AD, there is an inherited form of VCI, cerebral autosomal
dominant arteriopathy with subcortical infarcts and leukoencephalopathy, or
CADASIL (Ruchoux and Maurage 1997). It occurs as a result of a genetic
mutation on a gene (NOTCH 3) which affects blood vessels in the brain and
heart. Prevalence is uncertain, but Razvi et al. (2005) estimated genetically
proven CADASIL to occur in 1.98/100,000 of people in the west of Scotland
and suggested that the mutation prevalence is likely to be higher (4/100,000).
With the exception of CADASIL, VCI is not a disease with a single etiol-
ogy, but rather a range of cognitive states brought about by impaired function
or total obstruction of vasculature supplying blood to the brain (O’Brien et al.
2003). Reduced blood flow reduces oxygen and nutrient supplies, which in
turn cause dysfunction and tissue death. Damage may be localized, or consist of
multiple large or small lesions distributed in cortical and/or subcortical regions.
There is also a variant, small vessel disease, in which arterosclerotic condi-
tions of the small vessels in the brain reduce the blood supply (hypoperfusion)
which may result in a dementia syndrome (Pantoni and Simoni 2003). The
location and extent of the damage will be reflected in the dementia symptoms
of affected people (McPherson and Cummings 1996). However, a common
pattern is for prefrontal systems to be impaired, which may be reflected vari-

ously in behavioural changes, executive dysfunction, attention limitations and
general cognitive slowing (Almkvist et al. 1994). Memory may be relatively
spared at least early in the course of the disease (Desmond 2003; Sachdev and
Looi 2003; Moorhouse and Rockwood 2008). Risk factors for the develop-
ment of VCI include untreated high blood pressure, heart disease, a history of
stroke, atherosclerosis, diabetes, high cholesterol, obesity and lifestyle factors
contributing to these conditions such as diet, exercise, a history of smoking
or substance abuse (Roman 2005; Alagiakrishnan et al. 2006; O’Brien 2006).
As noted above, AD and VCI often occur together (Langa et al. 2004). The
relationship between them is still a matter of research and debated (Skoog and
Gustafson 2003).
At present, there are no cures for either AD or VCI. Treatments address
prevention and symptoms. Reducing risk factors through drug therapy (for
instance, for high blood pressure, diabetes, high cholesterol) and lifestyle
modifications of diet, exercise and toxin exposure may help prevent dementia
or slow the processes of neurodegeneration (Roman 2005; Black 2007). Drug
therapies developed to address cholinergic and glutamatergic system deficits in
AD improve neurotransmission and have consequent benefits in improved cog-
nition and function for some patients (Birks 2006; Burns et al. 2006; McShane
et al. 2006). There is also some evidence that these drugs may have some neu-
roprotective effects. For instance, the cholinesterase inhibitors have been asso-
ciated with reduced atrophy in vivo (Krishnan et al. 2003; Tune et al. 2003) and
work in animal models suggests that memantine may limit neuronal damage
associated with hyper-phosphorylated tau (Li et al. 2004). The cholinesterase
inhibitors are also used to treat VCI patients and appear to have similar effects
(Burns et al. 2006). Research into new drug therapies is aimed at preventing
or reducing the development of plaques and tangles thought to be the principal
causes of neuronal damage in AD (Blennow et al. 2006; Klafki et al. 2006)
and treating vascular disease in both VCI and AD (Gorelick 2003; Inzitari
et al. 2004).
There are some widely recognized patterns of language deficits associated
with AD including early word-finding difficulties, and what has been described
as ‘lexically empty’ and paraphasic speech (Nicholas et al. 1985; Bayles et al.
1987; Dijkstra et al. 2004). There is also evidence of reduced capacity for
processing complex syntactic structures although prototypical syntactic pat-
terns are preserved until AD becomes severe (Kontiola et al. 1990; Lyons
et al. 1994; Bickel et al. 2000; Kemper et al. 2001). People with AD often
modalize their discourse quite heavily and may have difficulty producing or
understanding complex narratives or instructions (Patry and Nespoulous 1990;
Erlich et al. 1997; Nespoulous et al. 1998; Asp et al. 2006a). They also very
characteristically have repetitive speech (Ready et al. 2003; Cullen et al. 2005;
Asp et al. 2006b; Rockwood et al. 2007). As the disease progresses, communi-
cation may be severely impaired and ultimately people with AD become mute,
though the desire to communicate may remain intact (Bayles et al. 1992; Ellis
and Astell 2004).
Because of the variability in VCI, there is no single profile of language or
discourse abilities or deficits. Rather language impairments will vary depend-
ing on the cause, location and type of neural damage. For example, dementia
following a stroke may include any of the aphasic syndromes depending on the
location of the stroke (McPherson and Cummings 1996). A number of studies
investigating VCI caused by diffuse subcortical ischemic damage have found
that patients have reduced letter fluency as compared with category fluency
on word-list generation tasks (e.g. Lafosse et al. 1997; Mendez et al. 1997;
Tierney et al. 2001; Duff et al. 2004; Poore et al. 2006). These tests ask par-
ticipants to name all the words they can think of beginning with a particular
letter (letter fluency) or all the animals they can think of (category fluency). All
these studies compared performance of VCI patients with AD patients in the
interests of finding clinically relevant ways of distinguishing the two dementia
syndromes. The general pattern, not always confirmed statistically, is that AD
patients perform better on letter fluency and VCI patients perform better on
category fluency though both groups may perform worse than control subjects
on both measures. The assumption is that subcortical damage affects prefrontal
function on which letter fluency tasks depend more than temporal lobe func-
tions which are engaged in category fluency. Jones et al. (2006) extended the
analysis to preclinical phases of AD and VCI and found that category fluency
does distinguish the groups, but that compromised prefrontal function in pre-
clinical AD participants matches that of VCI participants so that letter fluency
is similar between groups. Other studies, also comparative of AD and VCI
patients’ language, have reported conflicting results not only for fluency tasks
(e.g. Looi and Sachdev 1999) but also for syntax. Kontiola et al. (1990) suggest
that in AD syntax is more compromised than in VCI, whereas Hier et al. (1985)
and Powell et al. (1988) report the converse. Recent studies directly evaluating
syntax in VCI are lacking as are studies of discourse, though Shindler et al.
(1984) noted increased perseveration and intrusions for VCI patients.
In short the linguistic picture for VCI is very limited. Studies have been
mostly comparative (with AD) and there has been little work on areas such as
syntax and almost none on discourse in VCI. This paucity is perhaps a result
of both the somewhat disappointing findings of early studies that sought to use
linguistic features to distinguish VCI and AD and the heterogeneity of VCI
itself. Those findings that there are mostly point to the effects of impaired
prefrontal function in VCI caused by diffuse subcortical damage – the charac-
terization of discourse for the various groups within this spectrum has yet to be
undertaken. Diagnostic techniques which allow better discrimination between
the subtypes of VCI make the prospect of developing research strategies which
address the linguistic and discourse signs associated with particular vascular
pathologies more possible.
2.4 Theoretical sources

Our approach to discourse analysis is eclectic, drawing from such theoretical
sources as conversation analysis, generative linguistics, functional linguistics
and pragmatics. Here we outline the kinds of phenomena and assumptions of
these different linguistic approaches.
2.4.1 Conversation analysis

Conversation analysis is the analysis of talk. It examines all kinds of spoken
interactions and it focuses on turn-taking, topic and fluency. Turn-taking has
to do with who gets to talk when and how turns at talk are ordered in relation
to each other. For example, questions require responses and the type of ques-
tion asked shapes in some respects the type of response given. Yes/no ques-
tions (e.g. Do you like chai?) require answers ranging between positive (yes)
and negative (no). There are thus both sequence (question before response)
and semantic relations between the pair of utterances. Highly regularized rela-
tions in turn-taking of this type have been described as adjacency pairs (Sacks
et al. 1974; Levinson 1983). Studies of turn-taking address not only sequenc-
ing and semantic relations between turns at talk, but also means of facilitat-
ing or inhibiting speakers’ access to ‘the floor’ through use of features such
as supportive questioning, interruption and speaker selection. Topic refers to
what is discussed and interacts with turn-taking, for instance insofar as speak-
ers’ relative access to the floor affects the extent to which topics of interest to
them are discussed (Danes 1974). Fluency refers to features associated with
information processing and monitoring discourse. Features such as hesitation,
pause frequency and duration, speakers’ repairs to their own or others’ dis-
course and so on, are studied by conversation analysts. Within conversation
analysis, the goal of these analyses is typically to explicate the meaning con-
tributions of such behaviours to talk. There is also a concern in conversation
analysis with identifying turn-taking, topic and fluency conventions, and asso-
ciated speaker expectations for performance. Moreover, conversation analysts
describe the variation in such conventions and expectations associated with
different types of conversational interaction and in different institutional envi-
ronments (e.g. Atkinson and Heritage 1984; Sacks 1992a; 1992b; Heritage and
Maynard 2006).
In clinical contexts, conversational norms associated with turn-taking, topic
and fluency may inform assumptions in terms of which individual speakers’
performance may be assessed. The clinical setting also introduces concerns for
features which conversation analysts do not typically consider. For instance,
in the clinic, concern with fluency may be associated with motor control (e.g.
stuttering), with rate of speech, with integrity of semantic systems, or with
executive function processes such as attention, working memory, or informa-
tion maintenance and monitoring. Similarly, topic management and turn-taking
might be means of investigating ability to (interactively) construct relevance,
in terms of perseveration, tangentiality, attention, confabulation, or social skills
development. These may in turn be linked to neural functions or structures
implicated in different disorders. Thus while the primary business of conver-
sation analysts has been to identify and describe the semiotic function of ele-
ments of conversational interaction, in the clinic these elements may become
the means of investigating relationships between behaviour and cognitive and
neurological functions.
2.4.2 Generative linguistics

Generative linguists have directly presented the fact of language competence
as cognitive and psychological capacities of speakers (Chomsky 1957; 1995;
Jackendoff 2002). The motivation is to provide explanations for the universal
ease with which children acquire languages and the apparently infinite poten-
tial for speakers to say new (never spoken before) things. Formal generativists
typically ask not what do speakers know? but what is it that speakers would
need to know in order to acquire a language? what kinds of mechanisms and
information might support language acquisition? and what formal properties
do languages need to have in order at least rationally to allow for linguistic
creativity? The attention to acquisition and creativity has meant that genera-
tivists have emphasized the role of syntactic (and other) structures in cogni-
tive architectures and so have contributed perspicuous and relatively simple
ways to represent relations in and between syntactic structures and structural
potentials.
2.4.3 Functional linguistics

Functional linguistics is concerned with semiotic potential and language use.
Functional linguists are thus more or less directly concerned with questions
such as what do speakers know that let them communicate relevantly and suc-
cessfully in context. Systemic–functional grammars address the text–context
relationship in part by positing linguistic systems as shaped by their cultural
and situational contexts of use. Within these frameworks, it is possible to
describe particular languages and varieties in terms of experiential, interac-
tional and organizational functions. These functional potentials constitute the
semiotic resources from which speakers can select meanings to produce texts
which are contextually relevant (e.g. Halliday 1976; 1985; Halliday and Hasan
1989). Concern with functional differentiation within the systemic–functional
framework has also meant that there has been explicit attention to areas such as
affect and evaluation in grammars and discourse (e.g. Martin and White 2005).
Because of the concern with texts and discourses, functional linguists have
borrowed and developed models of generic structures, scripts and schemas
from European functional and structuralist traditions, cognitive sciences and
conversation analysis. These cognitive models incorporate semiotically rich
articulations of textual and contextual information into the linguistic model so
that accounts of discourse schemas and genres can be represented as part of the
semiotic resources of a language or variety (e.g. van Dijk 1977; 1981).
2.4.4 Pragmatics
Pragmatics broadly addresses speakers’ ability to infer logical and presup-
positional relationships between sentences based on the content of what is
said, its relevance in context and speaker intentions. Pragmatics took as its
point of departure, questions raised by Austin (1962) about how individual
sentences could function as speech acts and how literal content might differ
from intended or interpreted meaning. For example, Austin drew attention to
the fact that we do not simply make statements, ask questions and give com-
mands, but that a sentence such as I hereby name this ship Floatsome spoken
in the appropriate context by a person with socially designated authority to
name the ship, actually performs the action of naming. Similarly, Let’s call
the kitty Oblong does not merely make a suggestion but actually (assuming the
naming is accepted) may perform the action of naming a new kitten. Austin
made the further point that all our utterances constitute speech acts insofar as
they perform actions, though they need not overtly state the act they perform.
For instance, a statement such as there is a bug on your shirt may be offered
as information. However, it is also possible that it might be spoken with the
intention to warn, as an expression of alarm, or even as a joke. Addressees’
ability to interpret the statement appropriately depends in part on inferring the
speaker’s intention relative to available contextual information. For example,
one can imagine a variety of situations which might motivate a speaker such as
knowledge that bugs in this area are poisonous, an affective state (the speaker
is afraid of bugs), or in the context of humour that the addressee has an image
of a bug on her shirt.
Pragmatics broadly conceived is thus concerned with the problems of
explaining textually and contextually based inferential processes involved in
communication. Aspects of these problems have been addressed with different
emphases and methods in relevance theory, cognitive pragmatics, ethnography
and interactional sociolinguistics as well as directly in some models of language

and discourse processing (e.g. van Dijk 1977; 1981; Lakoff 1987; Gregory
1988; 2009a; 2009b; 1998; Jackendoff 2002; Hudson 2007). Relevance the-
ory pays particular attention to inferencing from text to context (e.g. Grice
1975; Sperber and Wilson 1986; 1995). Interactional sociolinguistics studies
are interested in figuring out how the micro-bits of talk, performance both in
terms of linguistic behaviour and conversational behaviour, instantiate and
construct meaning in context through interaction (e.g. Gumperz and Hymes
1972; Atkinson and Heritage 1984; Clark 1996; Schegloff 2007). Where rel-
evance theorists may be more concerned with abstract relations that character-
ize inferential processes and may therefore work with hypothetical examples,
interactional sociolinguists are concerned with performance and thus work
on instances of talk. For example, interactional sociolinguists are prepared to
investigate the communicative value of superficially meaningless fragments
such as hesitation phenomena and false starts (e.g. Clark and Fox Tree 2002).
Ethnographic studies have also focused on ‘communicative competence’
(Hymes 1971) with particular attention to intercultural differences in conven-
tions associated with many different discourse features (Hymes 1962; 1974).
Studies compare differences not only in linguistic systems but also in conver-
sational norms and generic expectations for narrative, expository and other
types of discourse. Ethnographic work thus informs analyses about potential
and real variation in discourse associated with cultural contexts.
The discourse analysis that we do is eclectic in the sense that it is informed
by all of the above. The grammar that we use is functionally organized (by
experiential, interactional and organizational meaning potential) and generative
(we use a formal syntax). It explicitly incorporates insights from pragmatic,
cognitive and interactional perspectives not only in the grammar but also in
the overall architecture and it assumes that ethnographic locations supply the
content (and may influence the form) of that architecture. That is, we assume
that context, insofar as it is relevant to inferential processes in discourse, is cog-
nitive – matters of speakers’ access to information available in the interaction
and the semantic and episodic memories it activates.4 Moreover, the information
available in the semantic and episodic memory of speakers depends upon the
cultural contexts in which they live and the generic situations in which they par-
ticipate in instances of everyday life. We also assume within this framework that
cognitive processes such as inferencing are, in fact, executive function proc-
esses. Some of these relationships are schematically suggested in Figure 2.1.
Figure 2.1 is not intended as a ‘model’ or paradigm of anything. Rather, its
purpose is just to suggest, reading across rows, some of the areas of correlation
between contextual parameters, linguistic and discourse phenomena, their hypoth-
esized neurocognitive reference domains, and the kinds of study that character-
ize the type of information represented. (Neuropsychological and physiological
Contextual Discourse Neurocognitive reference domains Description domains

parameters and
linguistic
correlates
Context of Language • Language(s) and dialect(s) knowledge • Models/grammars of

culture and • Semantic memory representations of languages and varieties
dialect culturally specific ‘knowledge of the • Ethnographies
variation world’ • Encyclopaedias
Context of Functional • Semantic/episodic memory • Ethnographies of speaking

situation variation by • Representations of typical ways of doing • Conversation analysis
register and and saying things, organizing and • Genre and register
genre affectively responding to information descriptions
(knowledge of genres, registers, scripts,
frames, schemas etc.)
Instantial Discourse • Linguistic processes: selection, retrieval, • Interactional sociolinguistics
situation sequence/order, articulation • Analysis of instances of
• Executive function processes: discourse
information selection/inhibition, • Information and discourse
retrieval, maintenance, monitoring processing models
• Executive control function: judgement, • Pragmatics
planning, reasoning
• Affect: emotional states and evaluative
responses
Figure 2.1 Context, language and discourse relations to neurocognitive

domains and inquiry types
models for ‘semantic memory’ and ‘executive functions’ are not included here
simply because the focus of Figure 2.1 is ‘what information’ needs to be repre-
sented, rather than how or where that information is neurally instantiated.)
Figure 2.1 also says nothing about vertical relations within columns. Some
of the relationships between the elements in context of culture, context of situ-
ation and instantial situation may be thought of abstractly as inheritance rela-
tions of model(s) to instance(s) (Hudson 2007). That is, any particular instance
of discourse will inherit properties from speakers’ knowledge of linguistic and
contextual models which the instance of discourse both references and instan-
tiates. Selection of particular linguistic and discourse features are mediated
by ongoing executive functions associated with information processing and
with executive control functions of planning, reasoning and judgement (Royall
et al. 2002). We assume that executive functions and executive control func-
tions relate instantially available information to speakers’ semantic and epi-
sodic models and that affective states in the speaker in the instance modulate
these processes.
2.5 Other sources

In this section we discuss the roles of neurology, neuropsychology, psychia-
try and neuroimaging in developing understanding of relationships between
discourse behaviours and neurological disorders. Our primary goal here is to

emphasize the need to work with a multi-disciplinary team in clinical discourse
analysis. We do not aim at anything that could be considered a full characteri-
zation of these specializations, but rather just to indicate the types of expertise
necessary for developing and testing hypotheses, and interpreting results.
Neurologists, psychiatrists and (neuro)psychologists are all involved in
diagnosis, assessment, and treatment of people with neurological and/or neu-
ropsychiatric dysfunction. There are significant areas of overlap in their exper-
tise, research and practice. However, each specialization differs to a certain
extent in focus and in the types of care and intervention they typically provide.
Neurologists specialize in understanding the nervous system; that is, the brain,
spinal cord and nerves. This means that they may be consulted not only about
dysfunction but also about insult or trauma to these systems. In relation to
neurophysiology and behaviour, neurology is one of the disciplines which have
the authority and potential to provide explanatory models and attribute cause
relative to neural substrata.
The focus in psychiatry is understanding mental health. This means that psy-
chiatrists’ focus may include emotive, psychological and behavioural disorders
with uncertain neurological etiology such as depression, anxiety, anorexia and
schizophrenia. They are also often involved with many of the disorders (such
as Alzheimer’s) that engage neurologists. Many psychiatrists and neurologists
specialize in periods of the life span (e.g. paediatric or geriatric psychiatry) or
in particular disorders (e.g. vascular disease). They may also be engaged in
epidemiological and preventative research.
Neuropsychologists focus on identifying neural structures and functions
associated with neurological and psychiatric disorders. They develop and use
assessment tools for differential diagnosis, evaluation and rating and are expert
in measurement techniques. They may specialize in particular human proc-
esses such as perception or cognition.
Neurologists, psychiatrists and neuropsychologists may all employ neu-
roimaging as means of investigating neural function and structure in vivo.
Neuroimaging offers additional evidences that may support diagnoses or
neurological hypotheses. They do this by imaging structure using computed
tomography (CT), magnetic resonance imaging (MRI), diffusion tensor imag-
ing (DTI) and by imaging distributions of metabolites using magnetic resonance
spectroscopy (MRS). Function can be assessed through blood flow, oxygen
and glucose uptake using positron emission tomography (PET), single photon
emission computed tomography (SPECT) or functional MRI (fMRI), or by
evaluating electrical activity using electroencephalography (EEG) or magnetic
fields produced by electrical activity using magnetoencephalography (MEG)
(e.g. Toga and Mazziotta 2002). There are also novel technologies for imag-
ing microstructure and functional processes in neurons, cells and vasculature.
Current studies investigating language and discourse abilities that involve an

imaging component most commonly use one of the functional imaging tech-
niques. However, there is potential (given the right team, circumstances and
questions) to bring the other technologies to bear in relating discourse features
to neural structure and function.
An example of a functional investigation is Mason et al.’s (2008) investiga-
tion of the relationship of inferencing in discourse to theory of mind abilities
using fMRI. Building on considerable evidence that people with autism have
a deficit with reading other people’s mental states, they compared two groups
of speakers, 18 with a diagnosis of high functioning autism and 18 matched
controls, in their abilities to draw inferences about situations, intentions and
emotional states in a reading task. They compared left and right hemisphere
regional activations and functional connectivity between groups on task per-
formance and found that the group with autism activated a similar network,
but had less left activation, lower functional connectivity and activations were
not differentiated across tasks. That is, the group with autism employed the
same kinds of cognitive strategies regardless of the specific inferential process.
Mason et al. conclude that the activation patterns during complex information
processing are similar to controls but less efficient in autism.
This study illustrates ways in which various specializations may interact
in coming to understand (or even address) a particular set of brain behaviour
relations in the context of a particular disorder. Minimally, the areas of exper-
tise the study brings together are neurology, pragmatic aspects of discourse
analysis, fMRI imaging expertise (design, execution and analysis), statistics
and diagnostic assessment expertise. These relationships are schematically
indicated in Figure 2.2.
The point of Figure 2.2. is to suggest the collaborative and inherently inter-
disciplinary nature of research which explores discourse in clinical settings.
While particular tasks such as diagnosis or discourse analysis will be carried
out by researchers independently, shaping the questions, establishing the meth-
odology for addressing them, doing the investigation that leads to results and
interpreting the results require consultation and collaboration if the work is to
have potential value for health care.
2.6 Discourse norms

The goal of this section is to clarify the role of normative patterns in grammar
and discourse (behaviour and processing) in interpreting patterns in clinical
settings. Most grammars characterize the linguistic structures and relations
that may (or may not) occur in a language. What constitutes ‘a language’ may
be assumed (based on a speaker’s intuitive investigation of their own knowl-
edge of the particular language) or it may be established through examination
Assumptions
Theory of Mind is supported by a localized and functionally dedicated neural network

(neurology, psychiatry, neuropsychology, neuroimaging)
In autism there are deficits in Theory of Mind

(neurology, psychiatry, neuropsychology)
This may be reflected in differences in inferencing in discourse

(neurology, psychiatry, neuropsychology, discourse analysis)
Questions
Is the neural network supporting Theory of Mind differentially activated between controls
and people with autism on discourse comprehension tasks designed to engage this network?
(neurology, psychiatry, neuropsychology, discourse analysis, neuroimaging)
Ethics
Does the question merit the resources and participation demands?

What if any potential harms may be involved in participation?
(Principal Investigator and supporting institutions)
Method
Participant selection (diagnosis by psychiatrist)

Recruitment, data collection
Task design (neurologist/psychologist/discourse analyst)
fMRI design (technician/physicist; radiologist/neurologist)
Statistical design (psychometrist, biostatistician)
Study coordinator
Investigation
Analysis and interpretation of linguistic data (discourse analyst)
fMRI analysis and interpretation (technician/physicist; radiologist/neurologist)
Statistical analysis (psychometrist, biostatistician)
Results
Disseminate
Figure 2.2 Interdisciplinary relations in clinical discourse analysis

of demographically representative and functionally salient samples of speech

or writing. Increasingly, the speech samples used in developing grammatical
descriptions are based on large electronic corpora such as the British National
Corpus, the parallel American National Corpus, and the COBUILD corpus
of English. Corpora may be specialized for region (e.g. Canadian or Indian
English), mode (e.g. spoken or written), function (e.g. science language, first
or second language acquisition), or some combination of mode and function
(e.g. spoken academic English). Some smaller corpora are already analysed
phonologically, morphologically and/or syntactically.
Even where grammars and corpora are demographically or functionally spe-
cialized, the goal is still to investigate and provide normative representation of
typically occurring speech and/or writing. Grammars and corpora can thus be
used to explicitly establish ‘normal’ speech and discourse patterns. They also
can provide models for conceptualizing problems, recognizing differences, and
may provide data for hypothesis testing. Because grammars and corpora are
normatively based they provide essential reference material for clinical (and
other) types of discourse analysis.
What normatively based grammars and corpora do not do (for obvious rea-
sons) is represent (or indeed systematically exclude) the speech or discourse
of groups with identified, diagnosed, motor, cognitive or affective disorders.
It is also the case that sampling techniques for corpora vary and grammatical
descriptions and postulates may or may not be validated either in corpora or by
other means. Clinical discourse analysis can thus use grammars and represent-
ative corpora as reference material but must develop specialized techniques
and corpora to address the requirements of the medical setting and the specific
patterns relevant for treatment groups. For example, descriptive techniques
may require validation by inter-rater reliability assessment, and theoretical
assumptions about neural representation of grammatical constructs require val-
idation or support from neuropsychological, pathological or imaging sources.
Specialized reference corpora which characterize disorders, including their
stages and level of severity, are necessary for research purposes. Further, it is
necessary to control such reference material by comparison with, for instance,
demographically similar groups without the disorder, and/or with different dis-
orders depending on research goals.
It may also be necessary to develop functionally varied corpora where inves-
tigation is directed to understanding the competence of speakers across a range
of different situation types. For example, in developmental disorders such as
ASDs, verbal communication may be relatively successful in one context (e.g.
school or work) but challenging in others (e.g. phatic communion in casual
social settings). In studies where participants are involved in drug or other ther-
apeutic interventions or where disorders change with time, it is also valuable to
sample periodically (longitudinally) to assess change from a ‘baseline’ state.
Combining cross-sectional and longitudinal sampling with matched controls

is another very rich way to investigate discourse within a study group. Thus,
the choices made with respect to reference corpora for establishing norms or
‘baseline’ patterns will depend on the research goals in the particular study, but
the need for such reference corpora is a given in health research and so also in
clinical discourse analysis with potential translational value.
The clinical discourse analyst thus not only needs to be able to do discourse
analyses and all that that entails, but also needs to know enough about the
disorder(s) she is investigating and to be able to explain the relevance of lin-
guistic and discourse findings to clinicians. This kind of work can happen in
the context of single case studies, and indeed, despite our remarks above about
corpora, the case-study approach can be extremely informative as a point of
departure for research, and for illustrative purposes. But if the goal is ulti-
mately to contribute to the possibilities for helping people in clinical contexts,
then an appropriate corpus and research team is called for.
3 Talk and speech – conversation analysis
and intonation in English
3.1 Introduction
In this and the following two chapters we present descriptive resources which
allow coding of spoken discourse in terms of conversational, grammatical,
pragmatic and contextual features. Features that are coded can be counted. This
allows both ‘pattern finding’, a replicable way of proceeding inductively to
characterizations of discourse (a specialized kind of data mining), and ‘hypoth-
esis checking’ where analyses in terms of the presence, absence, frequency
and/or co-occurrence of discourse features are hypothesized to be associated
with particular disorders, neurocognitive states, or changes in neurocognitive
states.
Our goal is pragmatic: we do not enter into debates or account for the
historical development of particular concepts or constructs, but present crite-
ria for description that allow understanding and reliable coding of discourse
features. Often this means that we are presenting ‘hybrid’ models, synthesized
from works in functional, ethnographic, interactional and formal linguistic
traditions. We refer readers to these works but discuss them only when such
discussion seems essential for understanding.
The order of presentation, from conversational to contextual analysis,
is intended to proceed from the familiar to, perhaps, less familiar ways of
describing discourse. However, as is inevitably the case, some terms appear
that presuppose knowledge of others presented later. For the most part, we
address this by providing examples and/or glosses of terms and references.
We use normative examples rather than examples from clinical contexts in
presenting the grammar. Normative examples are practical insofar as they do
serve to illustrate without the addition of marked features or other distractions.
(See appendix A for some basic grammatical terminology.) Two terms presup-
posed in all of the following discussions are clause and utterance so we offer
definitions of these first.
For English, clauses can be defined syntactically as consisting of a verb,
its arguments and adjuncts as in Cosmo saw Tess yesterday, where the verb
saw takes two arguments, an Experiencer Cosmo and a Percept Tess occurring
29
as subject and complement respectively. Yesterday is an optional element

(syntactically an adjunct) realizing a time circumstance. Clauses which func-
tion independently can also be assigned a speech function. That is, they can
be used alone to ask a question, make a statement or exclamation, or give a
command. This means that the following Examples (1)–(4) are all clauses by
either or both criteria:
(1) Tess listened to the mouse very intently.
This has a speech function (statement), a verb (listened), two arguments (Tess,
to the mouse) and a manner adjunct (very intently).
(2) Where did Cosmo leave his monkey?
This has a speech function (question), verb (leave), and three arguments (Cosmo,
his monkey, where).
(3) What a fabulous day!
This has no verb but has a speech function (exclamation).
(4) Cosmo loves (chasing the monkey).
The bracketed dependent clause has a verb chasing and arguments (the monkey,
tacitly Cosmo), but no speech function.
This description of clauses refers to ideational and interactional features
which are distributed across morphosyntactic and semantic systems. Clauses
are also ‘organized’ as units of information, some of which is ‘given’ and some
of which is presented as ‘news’. Organized information units in discourse
are most commonly identified in terms of their prosodic features as ‘intona-
tion units’ (Chafe 1980; 1994; 2001) or ‘tone groups’ (Halliday 1970; 1994;
Halliday and Greaves 2008). Intonation is discussed in more detail in Section
3.3. For now, a tone group can be thought of as a sound segment that includes
a major pitch movement over information presented as news. Typically, but by
no means always, a simple clause will be spoken as a single tone group.
For many practical purposes, researchers may be content to work with mini-
mum discourse units synonymous with one, two or all aspects of the clause as
just described. If one is interested in, say, measures of information density or
speech function distribution, a clause may be the only unit one needs to refer to.
However, there are phenomena which researchers may want to analyse which
do not meet criteria for a clause: incomplete utterances, minimal responses,
idiosyncratic vocalizations and isolated hesitation fillers are examples. We use
the term utterance as a label to include such heterogeneous bits of speech.
Utterances may be independent clauses which, as just described, are roughly
equivalent to instantiated speech acts (e.g. Searle 1969; 1979). Practically, this
can be operationalized simply as any unit which has a speech function (i.e.
is used to state, exclaim, ask or order). Utterance must also include any spo-
ken (or written) element that signals ideational, interactional or organizational
Talk and speech 31
information about a speaker’s message. Thus, a minimal response umhm

inserted in another’s discourse will be included as an utterance signalling at
least that the speaker of umhm is attending to the discourse. Typically, such
utterances have distinct prosodic organization and one might be tempted to
use Chafe’s (1980; 1994; 2001) intonation unit as the defining ‘minimum dis-
course unit’. But incomplete utterances often don’t meet this criterion – they
are not only partial clauses but partial tone groups – and hesitation fillers such
as uhm and ah may be articulated with a level tone (monotone). If the analyst
is interested in a speaker’s processing abilities, these data matter as utterances.
While they do not meet criteria for speech function or for intonation unit, they
do contribute ideational, interactional or organizational information about a
speaker’s message. So we offer the somewhat inelegant but comprehensive
hybrid definition of an utterance as any unit which
• can be assigned a speech function, and/or
• has a distinct tone group and/or
• is a linguistic signal of ideational, interactional or organizational information
about a speaker’s message.
The result is that, with the exception of the laughter (8) which is not linguis-
tic, each line in the Text 3.1 will be treated in our analyses as an utterance by
one or more criteria.
Text 3.1 Scalding

The discussion is about activities of daily living for a patient (speaker 1) who
has moderate AD. It has been suggested that she might make a mistake with the
hot water taps in the bath. Speakers 2 and 3 are caregivers.
(1) SP1: I still say that that’s not something that’s that a normal person wouldn’t do.
(2) SP2: Well you could get scalded.
(3) SP2: No no
(4) SP2: hell no
(5) SP2: I’m normal
(6) SP2: I would never do that.
(7) SP1: No no
(8) SP1: ((laughter))
(9) SP2: I’m normal /?/
(10) SP3: fingers
The utterances I still say that that’s not something that’s that a normal
person wouldn’t do (1); Well you could get scalded (2), I’m normal (5),
I would never do that (6) and I’m normal (9) all have speech functions (they
are statements) and meet all other criteria. The other utterances cannot read-
ily be assigned speech functions. But (3), (4) and (7) are spoken as single
tone groups and realize information. The emphatic No no (3) and hell no (4)
of the second speaker are spoken as two tone groups so they are treated as
two utterances rather than one. They contradict the suggestion that ‘a nor-
mal person’ could make a mistake (with the hot water tap in the bath) and
thus have ideational and interactional information value. However, given this
speaker’s intervening well you could get scalded, (3) and (4) are not clearly
elliptical statement forms of ‘no that’s not something that a normal per-
son wouldn’t do’. (That is, their ‘speech function’ must be inferred; it is
not formally instantiated.) Similarly, the first speaker’s no no (7) appears
to affirm the second speaker’s negative assertion through reiteration of the
negation, though these ideational and interactional values are inferred rather
than explicitly stated. (Consider that the speaker could have said no, no, of
course, you are normal.) The third speaker’s fingers (10) is a fragmentary
index of ideational value but overlaps with the second speaker’s I’m normal
and no attempt is made to complete or repair the utterance. This may sig-
nal something about this speaker’s interactional (lack of) commitment to the
utterance. She could restate it.
In effect, we treat as an ‘utterance’ any verbalization that engages linguistic
systems. A majority of utterances will also meet classic definitions for speech
act or intonation unit, but others do not. In clinical studies where the goal is to
link behaviour to neurophysiology, with or without pathology, such behaviours
should be included in descriptions where they may have relevance. Our hybrid
coding for utterance allows this.
3.2 Conversation analysis

Conversation analysis examines all kinds of spoken interactions: conversa-
tions, interviews, telephone exchanges, lectures, speeches and oral narratives
are a few examples of the many types of spoken interaction. One focus of
conversation analysis is on the regularities of information exchange. Sacks,
et al.’s (1974) ground-breaking work on turn-taking showed that, among peers,
speakers take fairly regular turns at talk and that they follow a usually tacit set
of rules about ‘what can follow what’. Conversation analysts have also con-
tributed an inventory of elements of talk such as speaker turn, topic, control
of the floor, speaker selection, adjacency pairs and so on; they have identified
different genres of talk in terms of these elements, and they have accounted
for meaningful patterns in our everyday speech which we sometimes think of
as meaningless. For instance, our use of words such as um and uh have been
shown to have a non-random pattern relative to following pause length: uh
precedes a short pause; um occurs before a longer pause (Clark and Fox Tree
2002). In effect, this means that these apparently ‘meaningless’ hesitation fill-
ers function as signifiers for information structure. In Table 3.1, we define and
exemplify terms in conversation analyses and give examples of coding options.
Talk and speech 33
Table 3.1. Terms from conversation analysis
Topic
Ideationally consistent stretches of discourse may be developed over several turns through
cohesive chains. Lexical selections, arguments, predicates and adjuncts are related to each
other such that speakers can make a judgement that what is talked about is ‘the same or a differ-
ent topic’. A discourse may be about a single topic or more than one (see Coates and Cameron
1988; Coates 1996).
Topic initiation occurs normally at the beginning of an interaction. Initiation of new topics
typically occurs at clause boundaries (sometimes called transition relevance places, TRPs).
Topic development occurs when an initiated topic is expanded in subsequent turns which are
cohesive with (i.e. refer back to) the initiated topic. The topic being developed will typically
become background information (rather than new information), occurring for instance in sub-
ject position with pronominal rather than full lexical reference.
Topic shift occurs when the ideational focus of talk changes. New participants, processes
or circumstances will be referred to. The discourse may still have some cohesive ties with
previous talk but will not be ‘about’ the previous topic. The turn introducing a topic shift will
typically have previously mentioned information represented either as news or as an element
of another constituent presented as news so that it can then become topic for the discourse.
Entirely new topics that do not refer back to prior discourse may be introduced at almost any
point. These are regarded as TOPIC INITIATION rather than topic shift.
Text Coding Count

Sp1: I saw John last night. TI: JOHN TOPI 1
Sp2: Oh yeah. And is he well? TD: WELL [JOHN] TOPD 1:1
Sp1: Yeah. He’s fabulous. TD: FABULOUS [JOHN] TOPD 1:2
Sp1: He has a new job, new flat, TD: NEW JOB, FLAT, BABY [JOHN] TOPD 1:3
new baby.
Sp2: Cool.
Sp2: What’s the baby’s name? TS: NAME [BABY] TOPS, T 2
Sp1: Kitty. TD: KITTY [BABY] TOPD 2:1
Notes on coding:
• Interactional elements such as the cool which is an evaluation of the preceding news are
not included in the topic analysis. Evaluations like this one signal a speaker’s reaction to
the topic rather than developing or changing the ideational content.
• Counts for topics, developments, shifts and so on can reference speakers (if one is inter-
ested, for instance, in the level of speaker participation).
Turns – taking a turn at talk. Turns can be of any extent, from minimal (umhum) to extended
monologue. For practical reasons, we treat turn here as turn at talk, rather than as a unit iso-
morphic with clause or utterance (Schegloff 1968; Sacks et al. 1974). Turns can be initiating
(5), or responding (6). Elliptical responses to questions can be verbal or non-verbal (e.g. head
nodding). Both may be counted and coded.
(5) SP1: Well I saw John Coding: Initiating TI
(6) SP2: And was he well? Coding: Responding TR
Count: # of turns and/or # of turns
per speaker
Table 3.1. (cont.)
Hesitations
Sounds (uh, um, er, like) inserted to fill pauses in discourse (see Example (7)).
Alternatively, a sound may be lengthened (Clark and Fox Tree 2002; Jefferson 2004) (see
Example (8)).
(7) Well I uh uh uh saw John Coding: H; Count: 3
(8) A::nd I saw John Coding: HL; Count: 1
Pauses – hesitations in discourse with no vocal filler. Length of pause can be timed using a
stop watch or computer programme. Counts may be specified for total number of pauses, or
pauses judged as short, medium or long by a normative standard. In general, pauses of less
than a quarter of a second are short, those of about half a second are medium and pauses
of more than one second are long (Campione and Veronis 2002). (Campione and Veronis
suggest a categorization of brief (<200 ms), medium (200–1000 ms) and long (>1000 ms)
pauses.) (See also Goldman-Eisler 1968 and Jefferson 2004.)
(9) Well I uh (2) saw John Coding: (2) 2 second pause; Count: 1
Repairs – non-initial correction of utterance (Jefferson 1974; Schegloff et al. 1977). Repairs
may be self-initiated (10) or other initiated (11). In signalled repairs the repair is overtly
signalled as such (12).
(10) I saw Jack, John last night Coding: RSI: repair self initiated; Count: 1
(11) SP1: I saw Jack
SP2: John? Coding: ROI: repair other initiated; Count: 1
SP1: yah John last night Coding: SRSI: signalled repair, self initiated;
(12) I saw Jack, I mean John last night Count: 1
False start: Speaker begins an utterance, stops and restarts (13). The repair can repeat the same start
(14) or start differently (15) (see for example MacWhinney 2000; MacWhinney and Osler 1977).
(13) I’ll tell, I’ll tell you about that later; Coding: false start: FS; Count: 1
(14) I’ll tell, I’ll tell you about that later; Coding: false start repeated: FSR; Count: 1
(15) I’ll tell, we can talk about that later Coding: false start different: FSD; Count: 1
Incomplete utterances lack an obligatory predicate, argument or full argument not supplied
by the context as in Example (16).
(16) We certainly…; We were about… Coding: IU; Count: 1
Overlap: Speech overlaps that of a previous speaker (17). Overlaps can be distinguished as
interruption where the first speaker does not have an opportunity to continue (18) versus
co-construction – overlapping speech where first speaker continues speaking (19). Brackets
can be used in transcripts to indicate overlapping segments of speech (see Sacks et al. 1974;
Coates 1996; MacWhinney 2000).
(17) SP1: I saw John <outside> [<] on the street. Coding: overlap OL; Count: 1
SP2: <Me too> [<]
(18) SP1: I <saw> [>] Coding: interruptive overlap IOL; Count: 1
SP2: <When are> [<] you leaving?
(19) SP1: I saw John <outside> [>] on the Coding: supportive overlap SOL; Count: 1
street
SP2: <cool> [>]
Talk and speech 35
Extended exemplification is taken up in subsequent chapters where we analyse

particular discourses.
3.2.1 Additional features

Other features in conversation researchers may want to code for include:
Latching A person starts speaking immediately after the previous

speaker finishes with no intervening pause (Jefferson 1985; 2004). Frequent
latching can show high levels of engagement and close monitoring of informa-
tion in the discourse on the part of the latching speaker.
Relevant audible non-linguistic behaviour (E.g. laughter, crying,

blowing, sighing, clicks.) Such behaviour may indicate affective, attitudinal
or attentional features of an interaction, a participant or a patient group and so
may be important in transcripts. We code these in brackets. It is useful to set the
representation, for instance ((laughter)), as part of the transcription protocol as
it can then be readily quantified (or removed) from transcripts as needed.
3.3 Intonation patterns in English
3.3.1 Phonological hierarchy

Phonology has to do with speech sounds. The area of phonological study most
relevant for discourse analysis is prosodic phonology. Prosodic phonology,
sometimes referred to as ‘supra-segmental phonology’ is the study of speech
sounds associated with the pitch, length and loudness of sounds. It is contrasted
with segmental phonology which refers to minimum contrastive sound seg-
ments, or phonemes such as the English pairs /p/ and /b/ in pat and bat or the
vowel sounds /æ/ and /ε/ in pat and pet. M. A. K. Halliday (1970; 1994) posited
a hierarchy of phonological units above the phoneme as a constituent hierarchy
as follows:
• a syllable will consist of one vowel phoneme and optional consonants before
it (in the ‘onset’) and after it (in the ‘coda’);
• a foot will consist of one initial stressed syllable and optional unstressed
syllables;
• a tone group will consist of at least one foot. The one required foot will be
the site of a major pitch movement.
Halliday uses the analogy of music such that the syllable is the equivalent of
a beat, the foot a bar, and by extension, the tone group is a phrase. Prosodic
organization structures information in two multifaceted ways:
Tonicity and given/new information In each tone group there will

be a foot containing a syllable which will be more prominent: it will be a lit-
tle louder, longer and will have a bigger pitch movement than others. This
more salient syllable is called the tonic syllable and it signals that the speaker
is presenting the phrase that it occurs in as important information, or news.
Typically, it occurs on the last lexical item in a phrase or clause. Compare
(20) She’s a good student.
(21) She’s good.
If you say these normally and then hum the tune that you’ve used you’ll find
that the biggest pitch movement in (20) occurs on the first syllable of student
while in (21) it occurs on good, the last lexical item in each clause. The loca-
tion of tonic prominence within the tone group is key in interpreting new
information in spoken discourse. For instance, if the tonic syllable is shifted to
she in either example then there is contrastive emphasis on she and the inter-
pretation would be that there is some other person who is not a good student.
It is possible to simply move tonicity within the tone group to (almost) any
element and create patterns of contrastive emphasis in this way. There are also
a number of syntactic resources (outlined in Chapter 4, Section 4.4.2 on mes-
sage organization) which enable changes in the site of tonic prominence, and
therefore changes in what speakers present as new information.
Intonation and speaker stance The direction (falling/rising), height

(high/low) and width (wide/narrow) of the major pitch movement reflect the
speaker’s orientation to what they are saying in terms of
• certainty about polarity (yes/no-ness) where falling tones are broadly cer-
tain and rising tones are uncertain;
• commitment where lower tones express greater commitment on the part of
the speaker to what they are saying;
• the width of pitch range used. This does not correspond to any single
generalization except perhaps that the use of a wider pitch range is broadly,
stereotypically, associated with excited/uncontrolled speech.
(There are cultural differences in the pitch ranges associated with excited or
uncontrolled speech (Lakoff 1975).)
Notation conventions for prosodic phonology include that:
• the tone group is enclosed in double slashes //tone group//

• feet are separated from each other by single slashes /foot/
Talk and speech 37
• tones (numbered 1, 2 etc.) are inserted at the beginning of the tone group,
after the double slashes and before any indication of a silent beat (^)
• the tonic syllable is either underlined or in bold face (tonic)
Thus, //1 where are you / going // is a single tone group spoken on tone 1,
consisting of two feet, the second of which is the site of the tonic go and //2
where are you / going // is the same except that the tonic is on where and the
tone is tone 2.
Beyond these broad patterns, Halliday (1970) posited a system of five primary
tones with more or less specific values as follows. (Also see El-Manoufy 1988;
Brazil 1995; Gussenhoven 2004; Halliday and Greaves 2008.)
Tone 1 falling is the neutral tone for statements (as in Figure 3.1a) and WH
questions (Figure 3.1b). The basic form has a level pretonic, with falling tonic
from mid/mid-high pitch.
600
Pitch (Hz)
75
0.00802561 1.16095
Time (s)
Figure 3.1a Tone 1: falling

Statement: John’s lost his flashlight
500
Pitch (Hz)
75
1.69153 2.74004
Time (s)
Figure 3.1b Tone 1: falling

WH question: What did John lose?
Tone 2 high rising (Figure 3.2a), or falling-rising (pointed) (Figure 3.2b)

with a level pretonic is the neutral tone for yes/no questions:
Talk and speech 39
500
Pitch (Hz)
75
1.03117 2.43627
Time (s)
Figure 3.2a Tone 2: rising

Question: Did John lose his flashlight?
600
Pitch (Hz)
75
0 1.35986
Time (s)
Figure 3.2b Tone 2: falling-rising (pointed)

Polar question: Did John lose his flashlight?
Tone 3 low rising with a mid level pretonic expresses incompleteness or

dependence, as in Figure 3.3.
500
Pitch (Hz)
50
0.693592 2.03203
Time (s)
Figure 3.3 Tone 3: low rising

Statement: John lost his flashlight.
Talk and speech 41
In tone 4 falling-rising (rounded), the pretonic steps down from high to

mid and is used to make statements about which there is some reservation and
to express conditions (as in Figure 3.4).
500
Pitch (Hz)
75
0 1.22381
Time (s)
Figure 3.4 Tone 4: falling-rising

In tone 5 rising-falling (rounded), the pretonic steps up from mid

to mid-high and is used to positively assert that what is said is so (see
Figure 3.5).
500
Pitch (Hz)
75
0.0513036 1.21037
Time (s)
Figure 3.5 Tone 5: rising-falling

Halliday also posited two ‘compound tones’, tone 13 (a falling tone and
then low rising tone) and tone 53 (a rising-falling tone followed by a low ris-
ing tone). These compound tones were hypothesized to occur when a speaker
wants to highlight two points of prominence. The first tone would carry the
main point. Other authors simply regard these as sequences (Halliday 1994).
Sequences such as a series of tone 2, then tone 1 are normal for lists as in
Figure 3.6.
Speech function and tone interact in predictable ways in English such that
there is an unmarked or neutral tone selection for each primary speech function
and the selection of other tones are marked for various kinds of interactional
Talk and speech 43
500
Pitch (Hz)
50
0 3.12834
Time (s)
Figure 3.6 Tone 2, 2, 1: listing sequence

Statement: John lost his flashlight, his hat and his mittens.
and/or affective meaning. For example tone 1, the falling tone, is the unmarked
tone for statements, WH questions and commands, while the rising tone 2 is
the unmarked tone for yes/no questions. Use of a rising tone with a statement
produces something which is often interpreted as a question. However, the
combination of tone 1 and a yes/no question results in something that, at best,
sounds cold and detached, while tone 2 and a command produces something
which may be difficult to interpret as a speech act at all.
4 Grammar
4.1 Introduction
In this chapter, we outline in sequence a grammar for specifying interac-
tional, ideational and organizational features of discourse in English. We use
a functional organization for the grammar because we find this perspicuous
for discourse analysis. The grammar owes much to Michael Gregory (2009a;
2009b). However, in many cases our approach is once again hybrid; we present
a model that draws on social and cognitive perspectives on language, based on
our experience of what works. Section 4.1 presents a grammar of interaction.
In Section 4.2 we outline the grammar of ideation and Section 4.3 presents the
grammar of organization. We presuppose that readers have a basic syntax for
English. As much as possible, technical and model-specific terms are limited.
Appendix A presents basic grammatical terminology and Appendix B lists the
coding options suggested in this chapter.
4.2 A grammar of interaction

The interaction relationship has to do, initially, with two types of activity
speakers can engage in. The first involves the negotiation of role relation-
ships relative to an addressee and the second involves speakers’ expressions
of attitude and evaluations. Role relationships are most easily understood in
institutional settings with highly generic situation types such as the classroom,
court room or clinic, where boundaries may seem relatively fixed: teacher/
student, judge/defendant, doctor/patient, nurse/patient are examples of such
role relationships (Goffman 1959; 1961). Less obviously hierarchic and insti-
tutionally dependent are peer relationships among colleagues, friends, sib-
lings and partners. We refer to the negotiation of role relationships because,
while there may be expectations about behaviours for speakers filling particu-
lar roles, in fact in every instance of interaction there is the possibility that
those expectations may be met, or directly or indirectly challenged or ignored
by participants (Goffman 1974). The defendant who challenges the moral
authority of the judge, or the judge who asks a guilty defendant to specify the
44
Grammar 45
duration of their sentence, the patient who demands a particular treatment or

the doctor or nurse who asks the patient for treatment recommendations are
examples of speakers acting outside their expected roles. More commonly,
speakers make linguistic choices in interactions which, more or less subtly,
signify their orientation to the expected roles for the situation; so the col-
league or friend who is a little more knowledgeable on a given subject may act
or be cast in the role of mentor with respect to that subject. Roles are negoti-
ated through behaviours (Goffman 1974). The use of speech functions and
address terms are perhaps the most obvious ways in which speakers assume
and assign roles in relation to each other. Checks (such as tags) and intonation
may also be involved in negotiating role relationships as well as reflecting
speakers’ attitudes and evaluations. We present criteria, values and possible
coding for each of these in turn in Section 4.2.1 and then discuss attitudes and
evaluations in Section 4.2.2.
4.2.1 Negotiation of roles

Speech functions Particularly significant for the negotiation of
role relationships are the possibilities for selection of SPEECH FUNCTION,
summarized in Table 4.1. In broadly functional terms (e.g. Halliday 1994),
statements and exclamations offer information to an addressee. Questions
seek information from an addressee and commands direct the behaviour of
an addressee.
Table 4.1. Basic speech function contrasts
Statements, Exclamations offer information

Questions seek information
Commands direct behaviour
There are subtypes of each category with different functions. There are also
‘normal’ syntactic structures that express these meaning differences. To keep
life simple (and descriptions replicable), we work with the grammatical fea-
tures (morphosyntax and intonation) of each category. These are spelled out
below and in Section 3.3 on intonation above.
Statements The basic form for a statement is that a subject noun

phrase (NP) or clause precedes a finite (tensed or modal) verb. The verb will
agree with the subject in person and number though, given the amount of syn-
cretism in modern English, this is only observable in the present tense con-
trasts between third person singular and other subjects (Huddleston and Pullam
2002) as in A and B.
A. Present tense verb agreement with third person singular subjects.

(Verbs are in italics. Subjects are in bold face.)
(1) Sam watches TV in bed.
(2) He thinks this is reasonable behaviour.
(3) His mother believes it is bad for him.
(4) What worries her is whether he actually gets enough sleep.
Examples (1)–(4) illustrate ‘normal’ unmarked statement patterns with third
person singular subjects agreeing with third person singular present tense verb
forms.
Coding: Sam watches TV in bed. S
B. Present tense verbs with other subjects.

(5) The children watch TV in bed.
(6) Sam and his dog watch TV in bed.
(7) I watch TV in bed too.
(8) We watch TV in bed.
(9) You watch TV in bed.
(10) Also troubling are the movies that he sees.
(11) I/we/you/Sam might watch TV in bed.
In Examples (5)–(6) subjects are third person plural; in Examples (7)–(8),
subjects are first person singular and plural respectively. In Example (9) the
subject is second person for which there is no number contrast except in dia-
lect varieties that include either you/yous or you/y’all. The verb form doesn’t
change: we follow Huddleston and Pullum (2002: 74–5) in calling this the
plain present tense form. Example (10) illustrates a marked order in which one
can see the importance of agreement in that it is the NP the movies that he sees
following the verb that functions as subject and agrees with the verb. Example
(11) illustrates the fact that modal verbs (will, would, can, could, may, might,
shall, should, must, ought) operate like tensed verbs, but do not change form
regardless of number and person of subjects.
Past tense verbs do not change form in relation to person and number of
the subject. The verb be is an exception to these patterns: it has first person
singular tense (I am …), third person singular tense (he is …), ‘other’ present
tense (we/you/they are …) and singular (I/she was …) and plural (we/you/they
were …) past tense forms. Notice further that negating statements requires the
introduction of a modal or tensed operator verb as in I don’t watch TV in bed
and that negative does presuppose positive. The unmarked tone selection for
statements is falling (tone 1), but see Section 3.3 for other options.
Grammar 47
Statements can be tagged as in Sam watches TV in bed, doesn’t he? A tag

is a clause consisting of a finite operator verb (tensed forms of do/have/be and
modal verbs will, would; can, could; may, might; shall, should; must; ought
and in some dialects dare and need) and a pronominal subject co-referential
with the subject of the clause to which the tag is attached as in Examples
(12)–(18) where the tags are in bold face.
(12) Sam saw an elephant, did he.
(13) The answer is obvious, isn’t it.
(14) We’ve haven’t finished, have we.
(15) Drive carefully, won’t you.
(16) Call me when you get home, will you.
(17) ?Don’t rush me, will you.
(18) Sam saw an elephant, he did.
Coding: Sam saw an elephant, did he. S-TAG
Tags are adjoined to clauses functioning as statements (12)–(14) or com-

mands (15)–(17). Tags can have positive or negative polarity which can be
the reverse of that in the main clause as in (13), (14), (15) and (17) or unre-
versed positive as in (12) and (16). (Negative main clause + negative tag as
in *Sam didn’t see an elephant, didn’t he doesn’t occur.) There is also the
option illustrated in (18), an emphatic tag in which the order of the elements
of the tag are the same as those for the main clause. These options are sum-
marized in Figure 4.1. (Example (18) is included here for comprehensive-
ness, as are (15)–(17).)
Tags work in conjunction with intonation patterns to signify positive or neg-
ative presuppositions and seek confirmation from addressees (e.g. Quirk et al.
untagged Sam watches TV in bed
symmetrical polarity Sam watches TV in bed, does he

statement
Sam watches TV in bed, doesn’t he
reversed
asymmetrical polarity
tagged
Sam doesn’t watch TV in bed, does he

He watched TV in bed, he did
unreversed
He didn’t watch it, he didn’t
Figure 4.1 Options for statements1

1985: 810–13; Sinclair 1990: 433–4). The allowable intonation patterns are as

follows:
+falling ↓ + rising ↑
+rising ↑ +falling ↓
+falling ↓ +falling ↓
The broad patterns are that rising tone expresses uncertainty and falling tone,
certainty, and that positive tags check the polarity of the main clause with neu-
tral presupposition, whereas negative tags presuppose positive polarity. For
example, a positive statement with falling tone asserts positive and expresses
certainty. If it is negatively tagged with falling tone, the positive is presupposed
and certainty is expressed by the tone. The result is that it emphatically asserts
the positive statement as in You saw him / didn’t you / and merely checks on the
presupposed certainty of the proposition you saw him. The expected response
is ‘yes’. The same sentence spoken with rising tone on the tag You saw him /
didn’t you / ↑ differs insofar as the rising tone expresses doubt and so although
the sentence still strongly presupposes positive you saw him, there is less cer-
tainty about the response.
Exclamations Exclamations, summarized in Figure 4.2, may be

either finite (with a tensed verb), or minor (with no verb at all). They often have
an initial WH NP or WH AP construction as in the following. (Again, verbs are
in italics. Subjects are in bold face):
(19) What a brilliant student she is!
(20) (What) a brilliant student!
(21) How wretched that you have to leave!
Coding: What a brilliant student she is! E
Questions Questions have to be subclassified initially into Yes/No

(polar) questions and WH questions (as in Figure 4.3). Polar questions seek
yes/no responses, with graded options (possibly, maybe, probably, certainly,
of course and so on) between the yes/no ‘poles’. The unmarked tone for
major What a brilliant student she is
exclamation
minor What a brilliant student
Figure 4.2 Options for exclamations

Grammar 49
polar Does Sam watch TV in bed
question
WH Where does Sam watch TV
Figure 4.3 Options for questions
yes/no questions is rising (tone 2). WH questions seek information about a

participant or circumstance indicated by an initial WH word (what, where,
when, who, why, which, how). The unmarked tone for WH questions is falling
(tone 1).
The central grammatical difference between questions and statements is that
in a question a finite operator (tensed forms of do, have, be) or modal verb
occurs before the subject element as in (22)–(27) where subjects are in bold
face and operators (or modals) are in italics.
(22) Can you be ready by six?
(23) Is anyone hungry?
(24) Has there been much snow this year?
(25) Would you like some more?
(26) Did the dogs get fed?
(27) Were the children frightened by the film?
Coding: Can you be ready by six? QP
Notice that positive polarity in yes/no questions generally entails no presup-

position about the polarity of the response, whereas negative polarity always
presupposes, not necessarily a positive response, but that the speaker believes
that the positive predication is more likely or ought to be true. The contrast will
seem obvious if one considers likely contexts for examples such as (28)–(30):
(28) Weren’t the children frightened by the film?
(29) Didn’t the dogs get fed?
(30) Wouldn’t you like some more?
Syntactically, WH questions have the same form as polar questions except that
the operator verb is preceded by a WH expression which will be co-referential
with an ‘empty’ argument, or circumstance position elsewhere in the clause.
Co-reference can be indicated by assigning an index (conventionally starting
with i, but in large data samples, better handled with numbers) to the WH
expression and the empty position it refers to (marked with a t for ‘trace’) as
in Examples (31–35).
(31) When1 can you be ready t1?

(32) How much snow2 has there been t2 this year?
(33) What3 would you like t3?
(34) What4 were the children frightened by t4?
(35) Who is hungry?
Coding: When can you be ready? QWH
Notice that (35) is an exception to the general pattern in that when the WH
word refers to the subject position the order of elements is the same as for
statements and there is no obviously ‘empty’ position corresponding to the
WH expression. Notice too that patterns with negation are similar here: What
wouldn’t you like? presupposes that there are many things that you would like
or, alternatively, that you are having difficulty stating a positive preference
for something so the speaker seeks information about negative preferences in
order to limit the set of total options. However, some WH questions actually
entail that X positively is so, in which case it is not possible to negate them.
*How much snow hasn’t there been this year? doesn’t work for this reason.
Commands Commands must also be subclassified into jussive,

optative, and fiat. Unmarked tone selection for all commands is falling (tone 1).
Jussive commands direct the behaviour of the addressee to do something.
They have a base form lexical verb which may optionally be preceded by you
or an indefinite pronoun subject as in (36)–(38).
(36) Call 911.
(37) (You) be quiet.
(38) Somebody call a doctor.
(39) Don’t anybody move.
(40) Do be quiet.
Coding: Call 911. CJ
Examples (39)–(40) illustrate negated and emphatic jussives respectively.

Both constructions require the insertion of base form do. Notice too that (37)–
(39) illustrate the fact that the verb forms here are non-finite (not tensed) base
forms. If the verbs were tensed, we would have you are/were quiet; somebody
calls/called a doctor; doesn’t/didn’t anybody move. These are grammatically
quite acceptable, but they are not commands: the first two are statements and
the third is a question.
Grammar 51
Optative commands propose a course of action to be (putatively) under-

taken by both the speaker and the addressee (41). They are formed with Let
followed by us (contracted to ’s unless it is emphatic) as subject of the follow-
ing verb. Let itself has no subject. The ‘putative’ descriptor refers to the fact
that optative occurs quite frequently in contexts where the speaker is in fact
proposing a course of action which engages the addressee as patient in a proc-
ess carried out by the speaker (42) or where the speaker intends the addressee
to carry out the action themselves (43).
(41) Let’s go out for dinner tonight.
(42) Let’s take your temperature. (imagine a nurse proposing this to a patient)
(43) Let’s eat up all the veggies now. (imagine a parent supervising children’s
dinner)
Coding: Let’s go out for dinner tonight. CO
Optative and jussive commands can usually be tagged, though the options
are more restricted. Optative allows only positive shall we as a tag. Jussive
allows positive or negative will you, won’t you depending on the polarity of
the main clauses.
Fiat commands are those rare utterances in which ‘saying it makes it
so’: the Let there be light construction. Like the optative, they require Let, and
do not allow a subject. The subject of the dependent clause is there or another
third person NP as in royal fiat Let this day be a holiday henceforth. They are
rare because few people have the authority or power to use them except when
joking. Options for Command speech function are summarized in Figure 4.4.
Coding: Let there be light. CF
The ability to use commands may be correlated in some measure with the
power and/or authority of the speaker relative to the addressee and in some cir-
cumstances certain forms may be ‘required’ by the situation (as in the pharma-
cist’s necessary use of jussive in ‘Take two tablets every six hours’). The fiat and
subjunctive almost never occur outside the context of religious texts. Certainly
they cannot be used for anything other than comedic purposes (or by someone
with a severely distorted view of the power of speaking) in ordinary discourse.
But note that there are no isomorphic (one-to-one) correlations here. In some
contexts, intimate and/or equal role relations may be signalled by use of, for
instance, jussive imperatives. It is not possible to simply ‘read off’ role relations
based on speech function selection. One has to look at what else is going on.
untagged (You) be quiet
jussive
symmetrical polarity Be quiet, will you
tagged
Be quiet, won’t you
Don’t be quiet, will you
command
untagged Let’s go
optative
tagged Let’s go, shall we
fiat Let this day be a holiday
Figure 4.4 Options for commands
All the options for speech functions can be summarized as a system as in

Figure 4.5.
Checks and intonation Also important in the negotiation of role

are the use of checks. Tags as described above are one sort of check. Others
include the use of words such as eh, right, huh, with rising intonation adjoined
to statements (e.g. you’re finished, right? or huh? or eh?).
Coding: You’re finished, right? S-CHECK
Rising pitch on a statement can have a similar effect as a tag with rising
tone. The statement positively or negatively asserts something; the rising tone
signals that what is said is not certain and needs confirmation. This sort of
intonation can be suggested in written texts by the use of a question mark
(?) following a statement. (Notice that this is not the same as asking a polar
question. The question did you see them? neither asserts nor presupposes that
an addressee performed the action predicated of them. In contrast, you saw
them? does assert this as an action of the addressee; and you didn’t see them?
presupposes the positive. The use of rising tone (2) adds a ‘needs checking’
feature.)
Coding: You’re finished? S-RT

Grammar 53
untagged Sam watches TV in bed

symmetrical polarity Sam watches TV in bed, does he
statement
reversed Sam watches TV in bed, doesn’t he
tagged asymmetrical polarity
Sam doesn’t watch TV in bed, does he

He watched TV in bed, he did
unreversed
He didn’t watch it, he didn’t
major What a brilliant student she is

exclamation
minor What a brilliant student
speech
function
polar Does Sam watch TV in bed
question
WH Where does Sam watch TV
untagged (You) be quiet
jussive
symmetrical polarity Be quiet, will you
tagged
Be quiet, won’t you
Don’t be quiet, will you
command
untagged Let’s go
optative
tagged Let’s go, shall we
fiat Let this day be a holiday
Figure 4.5 Options for speech functions
Address terms Address terms are often, but not necessarily, real-
ized as Vocative elements and can function as overt markers of the ‘social
distance’ (on a graded scale with poles being NEAR/FAR) between speaker
and addressee as in Darling, Sammy, Samuel, Professor Samuel Johnson, Sir.
Because address terms may overtly mark social distance, they can be used to
index and negotiate role relationships not only in terms of the admissible level
of intimacy, but also in terms of power. The parent–child relationship permits
(indeed requires in many cultures) mutual use of familiar or intimate names.
But young children lack power with respect to their parents and so may not be
free to address them in any way they choose – whereas parents do have such
freedom with respect to children. Similarly, North American professors may
invite students to address them by their first names or by title and name. But
the relationship is not reciprocal – students have little choice about the terms
faculty use to address them, and students who address faculty by first name
without invitation to do so may be regarded as attempting to shift the role rela-
tionship (Brown and Yule 1983; Gregory 1988; van Leeuwen 1996).
Coding: It’s raining Sir. V+SD
All of the above features can be manipulated for ‘politeness’ (Brown and
Levinson 1987), but there are also explicit ‘POLITENESS MARKERS’,
These are the please and thank yous of everyday discourse. Their lexical cat-
egory is somewhat obscure. They were originally verbs (If it please you /
I thank you) but they have become idioms and behave a bit like adverbs. The
degree of politeness required in any particular situation type varies from
culture to culture.
Coding: Two beers, please. PM
4.2.2 Attitude and evaluation

The second major aspect of the interaction relationship has to do with speak-
ers’ ability to express their attitudes and judgements about what it is they are
saying. Important elements here are the use of MODALITY which can express
judgements about the relative probability of what is predicated (as in It is rain-
ing / It might be raining, This should be the answer), about the capacity or
ability to perform some action (You can understand this), and about obligations
(You must/should eat your dinner). Modals can also be used to seek or give
permission (May/can I go, you may/can go). The value of particular modals
varies with the contexts in which they are used. Modal will for instance is typi-
cally interpreted as expressing a prediction in It will rain, an intention in I will
go tomorrow, and an obligation in You will do your homework. However, these
different ‘meanings’ seem to be a function of their construction with other lin-
guistic items rather than of the item will per se. For example, a prediction about
Grammar 55
self (I will) colligated with an action or process which self carries out will quite
reasonably be interpreted as the expression of a speaker’s intention to do some-
thing. But it seems unnecessary to say that will has more than one meaning.
We would rather say that prediction has different values in different contexts.
(For comprehensive discussions of the uses of modal verbs in contemporary
English see Sinclair (1990) and Huddleston and Pullum (2002).) Probability
and capacity values expressed by modal verbs can also be expressed through
adjectives and adverbs with related meanings such as possible/possibly, cer-
tain/certainly, is able, sure/surely, perhaps and so on. Also relevant are indefi-
nite degree expressions variously realized as NP + of + adjective (e.g. sort of,
kind of as in Sam is sort of fabulous); NP AP as in a little bit tired, somewhat
clever.
All of these resources constitute different ways to attenuate the forcefulness
or degree of certainty with which something is expressed. Within the literature
of pragmatics, they are broadly referred to as ‘hedges’ (Lakoff 1973). In clini-
cal work, Nespoulous et al. (1998) and others have referred to discourse where
these features are frequent as ‘modalizing discourse’. Both the pragmatic and
modalizing discourse approaches interpret these features as reflexes of speak-
ers’ psychological or epistemic states. In the heteroglossic approach presented
within appraisal theory (White 1998; Martin and White 2005), these features
are described as opening discourse to other views.
Coding: Sam is sort of MF fabulous.
Speakers can also evaluate the whole of a proposition by adding

ATTITUDINAL elements to a clause. These can be ADVERBS such as unfor-
tunately, happily, sadly and so on, or they can be relative clauses or noun phrases
added to the end of a clause (John broke his leg, which was unfortunate / John
broke his leg, a pity). This sort of marked expression of attitude is less common
than the constant lexical choices of speakers which reflect positive, neutral or
negative evaluations about referents as in A gentleman called / A man called /
Some birk called. When coding attitude, it is useful to code positive attitude
(ATT-P) or negative attitude (ATT-N).
Coding: Unfortunately ATT-N John broke his leg.
Every lexical choice we make indexes our attitudes to what we are talk-
ing about, and so all lexis is ‘evaluative lexis’, though we tend to notice
(and may only want to code for) markedly negative or positive evaluations.
Insofar as speakers orient themselves to others through shared values and
Table 4.2. Modalization, attitude and evaluation
Modalization
Expresses judgements about:
• probability
• permission (seek or give)
• capacity or ability
• obligation
Can be expressed by:

• modal verbs
• adjectives
• adverbs
• indefinite degree expressions (hedges)
Attitudinal elements and Evaluative lexis

Attitudinal elements – evaluate a proposition
• adverbs
• relative clauses or noun phrases added to the end of a clause
Evaluative lexis – reflects speaker’s attitude to referents in a proposition

• positive
• neutral
• negative
judgements, expressions of attitude and evaluation are used in the ‘presenta-

tion of self’ and in the construction of social relationships (Goffman 1959;
1961; 1974; Berger and Luckmann 1966; van Dijk 1977; Martin and White
2005). Table 4.2 lists resources for modalization, evaluation and attitude.
4.3 Conceptual structure and the grammar of ideation

One thing a language does is relate semantic information to phonological
structure. The semantic information may be conventionally signified by a
range of linguistic phenomena including morphemes, words, argument roles,
syntactic structures, discourse structures and patterns and occasionally whole
discourses (Jackendoff 2002; Gregory 2009c). In some formulations, such
as Jackendoff’s (e.g. 1990; 2002), this set of relationships between semantic
information and the linguistic and discourse structures and patterns that signify
it is referred to as ‘conceptual structure’. We use this terminology because the
emphasis on the relational and functional character of the information matches
our analytic experience and is neurally plausible. However, we do not imag-
ine, as Jackendoff seems to (e.g. Jackendoff in preparation), that there is a
dedicated amodal brain region that processes the relationships described in
conceptual structure. Rather, we suspect that ‘conceptual structure’ will turn
Grammar 57
out to be a network of transmodal gateways (of the sort proposed by Mesulam

e.g. 1998) linking information from modal and heteromodal association areas
in widely distributed cortical and subcortical brain regions. This seems to be
the sort of architecture implied by recent imaging and aphasia reviews (e.g.
Bookheimer 2002; Stowe et al. 2005; Ferstl et al. 2008). We take up these
matters in more detail in Chapters 8 and 9 – here we simply want to note that
we do talk about conceptual structure but assume that the neural architecture
is open to revision.
Semantic phenomena may be descriptively represented in taxonomic hierar-
chies at different levels of abstraction and we assume can be usefully thought
of as decomposable into feature bundles, although these do not exhaust seman-
tic meaning potential. Thus it is possible to think of words or morphemes as
designating bundles of meaning features which may be more or less abstract
such that the most abstract are superordinate concepts (animal), less abstract
concepts (dog) are basic level, and most specific (Peekapoo) are subordinate
level concepts (Berlin and Kay 1969; Rosch 1983; Lakoff 1987).
It seems reasonably clear that subordinate, basic level and superordinate
concepts may be represented differently in neural systems. Evidence from
Alzheimer’s and aphasia studies (e.g. Chertkow and Bub 1990; Nespoulous
et al. 1998; Marczinski and Kertesz 2006) suggest this insofar as subordinate
and basic level semantic systems may be degraded, damaged, or be inaccessi-
ble independently of superordinate concepts. Thus a speaker with Alzheimer’s
may be able to say that a picture of a terrier is an animal, but not that it is a
dog or a terrier. There is also evidence from Alzheimer’s and aphasia studies
to support the notion of semantic decomposition into features insofar as speak-
ers may show differential impairment for semantic categories (Warrington and
Shallice 1984; Warrington and McCarthy 1987). For instance, a speaker may
have access to a lexicon for ‘living things’ but not to ‘tools’ (or the converse)
(Marczinski and Kertesz 2006; Devlin et al. 2002; Ilmberger et al. 2002). Such
a speaker might be able to recognize and identify dogs, cats and rabbits and
assign them to the same category, but not be able to do the same thing with ham-
mer, saw and chisel (Warrington and Shallice 1984; Warrington and McCarthy
1987). Thus, while the supporting neural systems are not entirely known or
understood, there is significant evidence to assume both feature decomposition
and taxonomic ordering of semantic features. This is as we might expect given
our capacity to use referring expressions which rely on semantic decomposi-
tion. For example, English personal pronouns refer to superordinate contrasts
such as human/non-human and gender (masculine/feminine/neuter) to estab-
lish the identity of a referent. To use them appropriately, one needs to be able
not only to know the feature set of the pronoun, but to decompose the potential
referents into comparable feature sets and match them. Explanations of funda-
mental processes such as reference are inexplicable unless we can decompose
Table 4.3. Conceptual hierarchy with features
Superordinate
Category Basic Level Category Subordinate Category
Example Thing (animal) Dog Peekapoo

Features • concrete • four footed • (crossbred)
• living • domestic • pet
• mammal • has hair • very small
• barks • longer back than
height
• long woolly fur
• tan/brown
Properties • Concrete: entails • Four footed: shape • Shape, colour, types
associated shape, size, colour feature of hair are specific
with features etc. • Barks: defining • Pet: function
• Living: moves, activity for category
breaths, needs food • Domestic: used by
• Mammal: warm- people for food,
blooded animal work, pets
which lactates and
has one or more
infants at a time
Evaluation In isolation, Examples Examples
features superordinate level • Positive: ‘useful, • Positive: ‘cute’
categories rarely companions’ • Negative: ‘stubborn’
have evaluations • Negative: Fights, can
attached to them. be vicious, unclean,
However, there is ‘scavengers’
positive value attached
to being alive rather
than dead or inert.
words and other signifiers into semantic features. Table 4.3 models differences
in superordinate, basic level and subordinate categories.
In grammatical constructions, the semantic values of words do not account
completely for the meaning of any particular clause. For example, in Cosmo
the cat ate the hazelnut chocolate and chewed the gardenia, cat is a basic
level concept for which we might specify features [DOMESTIC ANIMAL;
PET; MEOWS]; Cosmo, a proper name, indicates some particular individual
cat. However, these meanings do not account for the role of ‘intentional doer
of action’ (Agent) that Cosmo has in both these clauses. The verb eat requires
as one of its participants an Agent. That is, the intentional doer of an action, or
Agent, is a bundle of semantic features associated both with the verb eat and
Grammar 59
with a syntactic position: Agent occurs as subject in unmarked cases. Agent

clearly also occurs with many other verbs and other word classes. Similarly, the
hazelnut chocolate and the gardenia also fulfil a role relative to their respective
verbs eat and chew. Both may be described as Patient, an entity which pre-
exists and undergoes a change of state in a process. And once again, a great
many verbs require the argument role, Patient.
4.3.1 Argument roles

Argument roles are feature bundles expressing superordinate concepts related
to events, relations and states. All clauses involve a predicate and (almost all
involve) one or more arguments. Those arguments which must or can occur
and their default positions with respect to each other are specified by the
predicate(s). The arguments required and the positions in which they occur
vary from language to language but there are probably a core set which recur in
all languages (Fillmore 1968; van Valin and LaPolla 1997). In the description
of particular languages one may want to elaborate argument roles to account
for the variation and semiotics of the language. Argument roles express part
of ideational meaning. An effect of this is that syntactic structures which most
frequently are used to express particular argument roles may acquire their
own meaning potential by association (cf. Jackendoff 2002). (This may be the
source of the common conflation of syntactic positions such as Subject with
semantic roles such as Agent.)
In the following we present a list of argument roles together with their features
and distributions developed for the description of English (Gregory 2009b).
4.3.2 Agent – the ‘doer’ of an action

Agent must be +capacity for intention. This is testable through purpose
adjuncts (in order to, deliberately). For example, JOHN is Agent in John killed
the ducks in order to feed his family. But DISEASE is not Agent in Disease
killed the ducks. (Notice that *Disease killed the ducks deliberately does not
work.) Agent need not be +intention. It only requires the capacity for inten-
tion so that JOHN is still Agent in John killed the ducks accidentally. With
one exception (see below), Agent is realized by a noun phrase [NP] only and
is subject in active clauses. In passive clauses, Agent occurs as an adjunct and
is marked by BYNP as in The ducks were killed by John. In NPs, Agent may
appear as specifier to N′, (John’s race refers to the race that John ran), or it may
appear marked either by BY (the killing of the ducks by John) or OF + posses-
sive NP (the book of Chomsky’s). The exception noted above is that Agent may
be realized as a derived adjectival modifier of N′ (as in the American bombing
of Iraq / parental abuse of children). In the absence of a Theme, only the BYNP
constructions are unambiguously Agentive in NPs. (Thus, John’s murder is

ambiguous, referring either to John as Agent (he committed a murder) or as
Patient (he was murdered by someone) (WH for Agent = WHO).
Coding: John AGENT killed the ducks.
4.3.3 Instrument – an entity used in the performance of an action

An instrument can be concrete or abstract. It is realized as a WITHNP or BYNP
if Agent is also realized as in John killed the ducks with a knife (concrete),
John solved the problem with/by logic (abstract). Instrument may be real-
ized as Subject (in active clauses only) as in The coat-hanger opened the
door. However, if Instrument and Agent are both realized, Agent must be
Subject: John opened the door with the coat-hanger, but not *the coat-hanger
opened the door by John. In NPs, Instrument appears in WITHNP constructions
(John’s opening of the door with a coat-hanger) although BYNP constructions
may be possible for some speakers, or as specifier to N′ (the key’s opening of
the door). (WH = HOW, WITH WHAT)
Coding: The coat-hanger INSTRUMENT opened the door.
4.3.4 Cause – the inanimate cause in events

Cause must be without the capacity for intention and is realized by subject NP
as in The wind eroded the cliffs, or BY/WITHNP in passives (The cliffs were eroded
by/with the wind), or by non-finite ING clauses as subject as in The government
constantly devaluing the dollar hurts everybody. Cause frequently occurs in
causal constructions, e.g. The music prevented them from hearing each other
and alternates with BECAUSE OFNP (They were prevented from hearing each other
because of the music). Note that the alternative, to treat Cause as a metaphoric
Agent, is not really plausible since potentially agentive NPs look metaphoric
in Cause constructions. The wind blew the roof off the house is +Cause and
may be non-metaphoric whereas John blew the roof off the house receives a
metaphoric interpretation in unspecified contexts. In NPs, cause can occur as
specifier to N′ (the wind’s erosion of the cliffs), or as BECAUSE OF/BY/WITHNPs (the
erosion of the cliffs by the wind). (WH = WHAT)
Coding: The wind CAUSE eroded the cliff.

Grammar 61
4.3.5 Experiencer – sentient entity that reacts to,

knows or perceives phenomenon
Experiencer does not allow intention and cannot take a +intention adjunct.
For instance, *John likes Fred deliberately with John as Experiencer is at best
peculiar. The distribution of Experiencer varies with the subcategory of predi-
cate as follows.
(1) Experiencer may be realized by Subject NP, or by BYNP in passives, in
‘experiencer oriented’ reaction predicates as in The guests liked/enjoyed
the show, The show was enjoyed by the guests; in ‘perception’ predi-
cates (The guests saw/heard the show, The show was heard/seen by the
guests), and in ‘cognition’ predicates (John knows/understands/believes
the answer, The answer is known/understood/believed by John). There are
the expected related NP realizations for reaction predicates (the guests’
enjoyment of the show, the enjoyment of the show by the guests); cognition
predicates (John’s knowledge of/belief in the answer, knowledge of/belief
in the answer by John) although the BYNP phrases in the second in each of
these pairs may potentially be construed in relation to the adjacent N (as
in the show performed by the guests, the answer provided by John). NP
realizations of perception predicates are also possible if awkward: ?[the
guest’s seeing/hearing of the show/the sight/hearing of the show by the
guests] before the dress rehearsal was a mistake.
(2) Experiencer may be realized by Direct Object NP, or by Subject NP in pas-
sives, in ‘phenomenon oriented’ reaction events. (His behaviour disgusts/
revolts me / I am disgusted/revolted by his behaviour.) Nominalizations are
peculiar: (my disgust at/with his behaviour, John’s revulsion to/for work)
seem OK but there do not appear to be any PNP realizations (*the dis-
gust of/by me, ??the revulsion of John/John’s). Note that Experiencer here
shares distributional (and semantic) properties with Thematic roles (see
below) and thus might be regarded as a subtype of theme.
(3) Experiencer may be realized by TONP with copula verbs of perception
such as seem/appear/look/sound (It seems/appears to me that John is sick,
John looks sick to me). (WH = WHO)
Coding: The guests EXPERIENCER liked the show.
4.3.6 Stimulus
The Stimulus is the entity or event or state which is reacted to in ‘phenomenon
oriented’ reaction predicates realized by NPs, or by finite or non-finite clauses
as Subject in active clauses as in mushy peas disgust me / that they were

late annoyed her / John’s being early surprised Mary. There are the expected
related passives as in I am disgusted by mushy peas / Mary was surprised
by John’s being early. (The analogue for the finite clause in subject position
is not passive but extraposition as in she was annoyed that they were late).
(WH = WHAT)
Coding: Mushy peas STIMULUS disgust me.
4.3.7 Source – the entity from which motion takes place

In motion processes, Source is typically realized by FROM/OUT OFNPs as in John
came from Toronto / John ran out of the kitchen. In transfer processes Source
may combine with Agent as a Subject NP or a FROMNP Complement as in John
sent the book to Mary / Mary received the book from John. (Note that the sec-
ond example is optionally a modifier of the book.) Within NPs, Source may
be realized as FROMNPs (John’s departure from Toronto), or, with predicate
NPs that allow Source to combine with Agent, as a Possessive NP Determiner
(John’s gift of the book to Mary), and as a FROMNP (the gift of money from the
bank) and (in some dialects) also as a related BYNP construction (?the gift of the
money by the bank) (WH = WHERE/WHO).
Coding: Mary received the book from John SOURCE.
4.3.8 Goal – the entity towards which motion is directed

Goal requires subclassification as:
(1) Goal: location – a spatial destination in motion and transfer processes real-
ized by an NP (John went home/south) or PNP as (John went to Toronto /
John put the roast in the oven), or by a complement particle (preposition)
(John went out / John put the dog out), or by the ‘deictic adverbs’ there/here
(put it there). In NPs, only PNP constructions (with optional there/here sub-
stitution) seem possible as in the placement of students in senior courses /
his departure for home.
Coding: John went to Toronto GOAL-LOC.

Grammar 63
(2) Goal: recipient – the recipient in non-locational transfer processes. In

active clauses it is realized by NP as Subject where it combines with Agent
as in Mary received the book from John, or NP as Indirect Object (John
sent Mary the book), or by a TONP Complement as in (John sent the book
to Mary). The related passives are (i) The book was received by Mary / the
book from John was received by Mary / The book was received by Mary
from John; (ii) Mary was sent the book by John. ‘Verbalization’ predicates
such as tell, write, say, relate, report and so on typically behave as non-loca-
tional transfers insofar as they take the role of Goal: recipient as in John told
Mary the story. The distribution of Goal: recipient in NPs varies predictably
with the head N. Nominal receipt allows either possessive NPs or a BYNP
realization of Goal (Mary’s receipt of the letter / the receipt of the letter by
Mary). Nominalizations corresponding to clauses in which Goal: recipient
appears as an indirect object NP or TONP only allow a TONP realization as in
John’s telling of the story to Mary, sending of the letter to Mary.
Coding: John sent Mary GOAL-REC a letter.
(3) Goal: beneficiary – the entity for whom the process is carried out.

Goal: beneficiary is realized by an Indirect Object NP (Mary built John a
house), or by a FORNP (Mary built a house for John), and (for some varie-
ties) may be realized as Subject NP in a passive (John was built a house by
Mary). Within NPs, Goal: beneficiary may only be realized as a FORNP as in
(John’s building of the house for Mary). (WH = WHERE/WHO/WHOM)
Coding: John built Mary GOAL-BEN a house.
Note: The status of beneficiary and Instrument as argument versus circum-

stantial roles is ambivalent. They both share adjunct-like (circumstantial) prop-
erties insofar as they are not required by any predicate and can combine with
a wide variety of predicates. However, they both also can occur in ‘argument
positions’ and thus syntactically behave as though they are arguments. For
discussion of the issues see for instance Huddleston and Pullum (2002); van
Valin (2001).
4.3.9 Theme
The standard definition describes Theme as the ‘entity in motion’ in motion
and transfer processes, and as the entity located in locational relations. In
active clauses, Theme is realized by a complement NP in transfer processes

(Theme: transferent) as in John gave Mary the book / John gave the book to
Mary. It combines with Agent and is realized by a Subject NP in motion proc-
esses as in John went to Toronto. And it occurs as Subject NP in locational
relations (Theme: located) as in John is in Toronto. In practice Theme is often
used as a term for any argument that does not fit any of the designated roles
above. Thus it may be given a broader gloss to include not only the motion,
transfer and location readings but also the entity affected by, resultant from,
or designated by an event or state. Some grammars refer to a ‘macro-role’ of
‘undergoer’ to characterize this configuration of roles (e.g. Pike and Pike 1982;
van Valin 2001). What unites all these different meanings is that Thematic
arguments are never realized as syntactic adjuncts in English clauses. They are
always realized as either Subjects or complements of the verb. Since Theme
is used for such a variety of arguments it requires further subclassification as
follows.
Theme: patient Theme: patient is the entity that pre-exists the process and
undergoes a change of state in the process. Patient is realized by an NP
Direct Object in active clauses as in John killed Mary / John cooked the
carrots or as Subject in the related passives Mary was killed by John / The
carrots were cooked by John.
Coding: The carrots THEME: PAT were cooked by John.
Theme: resultant – the entity or event resulting from a process. Resultant is

realized by either an NP as Direct Object (John built a house), or by a finite
or non-finite complement clause as in (non-finite) John makes Mary cut her
hair. Note that the much discussed three argument predicates such as per-
suade and teach take both Theme: patient (realized as an NP complement) and
Theme: resultant (realized as a clausal complement [CP] as in John persuaded
them that they should leave / he taught them that they should be polite). See
also the related non-finites as in John persuaded them [PRO to leave] / he
taught them [PRO to be polite].
Coding: John built a house THEME: RES.
Theme: percept Theme: percept is an entity or event which is experienced or

perceived, realized as either an NP or a finite or non-finite clausal complement of
‘perception’ predicates such as see, hear, feel; ‘cognition’ predicates such as know,
Grammar 65
believe, understand; and ‘processor oriented mental reaction’ predicates (like,

enjoy, want, seem, appear ) as in
• perception: Mary saw John / Mary saw that John was running the marathon /
Mary saw John running the marathon;
• cognition: Fred understands the problem / Fred knows that the solution will
be difficult / Fred believes the solution to be difficult;
• reaction: Mary likes John / It seems to me that they are leaving / I want them
to leave.
Coding: Mary saw John THEME: PERC.
Theme: message What is communicated in message transfers (verbalization

predicates) is realized by non-finite complement clauses (The demons told me
PRO to leave), or by finite complement clauses (John said that he would be
here by six and to have dinner ready). Theme: message is distinguished from
Theme: transferent in that the former, but not the latter, may be realized by a
clause.
Coding: John said that he would be here by six THEME: MESS.
Theme: range Theme: range is an entity that designates the nature of a proc-

ess realized by complement NPs (Mary plays tennis/piano/chess / John runs
marathons/the 100 meters / Mary told them a story/a lie).
Coding: Mary plays tennis THEME: RAN.
In addition, by analogue with Theme: located in locational relations, the

Subjects of relational clauses (those that have a form of BE as the main verb)
are treated as subtypes of theme and may be classified according to whether the
‘complement’ identifies, classifies or attributes properties to the Subject which
may be realized by an NP or, in some cases by a finite or non-finite clause.
Theme: identified Obama is the president of the US/The president of the US

is Obama/That they are starving is the problem. The defining characteristic
for Theme: identified is that the predicate NP should be definite and that the
relations be reversible. (Note that ‘reversibility’ implies that predicates in iden-
tificatory relations can be realized by clauses as in The problem is that they are
starving.)
Coding: Obama THEME: IDEN’D is the president of the US.
Theme: classified Clark was a Prime Minister / John is a murderer / People

are bi-peds / To be truly free is an impossibility. Theme: classified requires an
indefinite specifier for the predicate NP and is not reversible outside of poetry.
(*A Prime Minister was Clark is ungrammatical, and although bi-peds are
people is grammatically acceptable its default interpretation is false.)
Coding: People THEME: CLASS’D are bipeds.
Theme: identified and Theme: classified can also occur in appositional con-

structions in NPs as in Obama, the president of the United States … or John,
an opera singer ….
Theme: attribuand Mary is clever/silly/pretty / That they are in trouble is

obvious. Theme: attribuand occurs with a ‘predicate’ AdjP. In NPs, the adjec-
tive usually precedes the noun as in They are in obvious trouble.
Coding: Mary THEME: ATTR is clever.
Theme: possessed Cosmo has a toy monkey / Michelle owns a Vespa.

Theme: possessed occurs with a small subset of lexical verbs indicating pos-
session such as have, own, possess. The possessed element is an NP comple-
ment. The possessor is subject in the unmarked case. In NPs, theme: possessed
can appear as a possessive determiner as in Cosmo’s monkey.
Coding: Cosmo has a toy monkey THEME: POSS’D.
Theme: existent There is a rabbit / I think, therefore I am. Theme: existent

occurs as complement in a clause with an existential there subject or as subject
for the ‘existence’ predicates be and exist.
Coding: There is a rabbit THEME: EX.
Theme: ambient It is snowing. Theme: ambient occurs as predicate with

expletive it as subject.
Grammar 67
Coding: It is raining THEME: AMB.
A summary chart for argument roles is presented in Figure 4.6.
Circumstantial roles Circumstantial roles correspond broadly to adjuncts of

TIME, PLACE and MANNER. Adjuncts are not ‘projected’ by a predicate, but
rather add additional information about the circumstances surrounding an event
or relation. Syntactically, circumstantial roles are always adjuncts regardless of
the constituent that they modify. In most grammars, all circumstantial elements
except sentence adverbs are treated as Adjuncts within the VP.
4.3.10 Time
Time may be realized by NPs (last night, three days ago), PNPs (on Tuesday, at
three), ADVPs (never, always, usually, now, then etc.), and by finite or non-finite
clauses with when/for/while and so on. All but the adverbial class of temporal cir-
cumstantials are arguably modifiers of clauses insofar as tense and aspect selec-
tions for the clause are governed by reference to semantic Time (past, present,
future) and Perspective (whether events are construed as beginning, ongoing,
or completed), and it is precisely this sort of information that the temporal cir-
cumstantials make explicit. The adverbial realizations of Time seem to be V′
adjuncts. McCawley (1988) presents an insightful discussion of the relationships
between Time adjuncts and tense and aspect selections (WH = WHEN).
Coding: They ate last night TIME.
4.3.11 Place
Realizations of Place include PNPs (They ate dinner on a rocky hillside), ADVPs
(somewhere, anywhere, elsewhere, outside), and finite and non-finite clauses
with an appropriate complementizer. Again, Place adjuncts seem to modify
whole clauses rather than just VPs and are thus clause adjuncts rather than VP
adjuncts: Place circumstances can modify conjoined clauses as in [They ate, they
worked and they played] in the kitchen; they are readily realized by separate tone
groups (i.e. don’t necessarily constitute a single tone group within a VP constitu-
ent) and they have no relation as arguments of a predicate (WH = WHERE).
Coding: They ate in the kitchen PLACE.

AGENT – the 'doer' of an action.

e.g. [JohnAGENT] killed the ducks.
INSTRUMENT – an entity used in the performance of an action.

e.g. [The coat-hangerINSTRUMENT] opened the door.
CAUSE – the inanimate cause in events.

e.g. [The windCAUSE] eroded the cliff.
EXPERIENCER – sentient entity that reacts to, or perceives a phenomenon.

e.g. [The guestsEXPERIENCER] liked the show.
Mushy peas disgust [meEXPERIENCER].
It seems [to meEXPERIENCER] that John is sick.
STIMULUS – the entity or event or state which is reacted to in 'phenomenon oriented' reaction predicates.
e.g. [Mushy peasSTIMULUS] disgust me.
SOURCE – the entity from which motion takes place.

e.g. Mary received the book [from JohnSOURCE].
GOAL – the entity towards which motion is directed:
(1) GOAL: LOCATION – a spatial destination in motion and transfer processes.

e.g. John went [to TorontoGOAL:LOC].
(2) GOAL: RECIPIENT – the recipient in non-locational transfer processes.

e.g. John sent [MaryGOAL:REC] a letter.
(3) GOAL: BENEFICIARY – the entity for whom the process is carried out.
e.g. John built [MaryGOAL:BEN] a house.
THEME – the 'entity in motion' in motion and transfer processes, or the entity located in locational relations,
or the only argument in relations.
PATIENT – the entity that pre-exists the process and undergoes a change of state in the process.
e.g. John cooked [the carrotsPAT].
RESULTANT – the entity or event resulting from a process.

e.g. John built [a houseRES].
PERCEPT – an entity or event which is experienced or perceived.

e.g. Mary saw [JohnPERC].
MESSAGE – what is communicated in message transfers (verbalization predicates).

e.g. John said that [he would be here by sixMESS].
RANGE – an entity that designates the nature of a process.

e.g. Mary plays [tennisRAN].
IDENTIFIED – subject NP in identificatory relation (should have a definite specifier and relations are reversible).
e.g. [BushIDEN’D] is the president of the US.
CLASSIFIED – subject NP in classificatory relation (should have an indefinite NP and relation is not reversible).
e.g. [PeopleCLASS’D] are bipeds.
ATTRIBUAND – occurs with a 'predicate' AdjP.

e.g. [MaryATTR] is clever.
POSSESSED – occurs with lexical verbs indicating possession.

e.g. Cosmo has [a toy monkeyPOSS’D].
EXISTENT – occurs as complement in a clause with an ‘existential there’ subject or as subject for the ‘existence’
predicates be and exist.
e.g. There is [a manEX] we know.
AMBIENT – occurs as predicate with expletive it as subject.

e.g. It is [snowingAMB].
Figure 4.6 Argument roles

Grammar 69
4.3.12 Manner
Realizations of Manner include ADVPs (carefully, quickly, slowly), PNP (in a
careful manner, with care/courage/fortitude), the NPs (this way, that way), and
clauses with an appropriate complementizer. Within the general category of
‘manner’ are circumstantial modifiers of clauses (so called ‘sentence’ adverbs
such as unfortunately, frankly, hopefully), VP modifiers (such as quickly, sadly,
slowly) and modifiers internal to V’ (barely, hardly, nearly) (WH = HOW).
Coding: He chewed in a careful manner MANNER.
There are also logical circumstances which occur as syntactic adjuncts to clauses.
4.3.13 Reason
Realized by a complementizer phrase [CP] with because (they stayed home
because it was raining) WH = WHY.
Coding: They stayed at home because it was raining REASON.
4.3.14 Purpose
Realized with (in order) + marked infinitive/that or so that (they stayed home
in order to keep dry / she stay up in order that she might finish her work, she
read so that she would understand) WH = WHY.
Coding: They stayed at home in order to keep dry PURPOSE.
4.3.15 Condition
Typically realized by if clauses (they’ll stay home if it is raining) (WH = ? (only
composites e.g. under what conditions)).
Coding: They’ll stay at home if it is raining CONDITION.
4.3.16 Concession
Typically realized by although clauses (they stayed home although it was nice
out) (WH = ?).
Coding: They stayed home although it wasn’t raining CONCESSION.
These ‘logical’ circumstantials are usually treated as adjuncts within the verb
phrase. They may be represented as either VP Adjuncts or as adjuncts of a
complementizer phrase. Circumstantial roles are summarized in Figure 4.7.
The combination of an articulated set of features and distributions for English
argument roles and circumstances together with the possibility of represent-
ing lexical selections as feature arrays allows for rich coding of both literal and
figurative ideational values in discourse. If one adds to this the semantic values
TIME – may be realized by

• NPs
• pNPs
• ADVPs
• finite or non-finite when/for/while etc. clauses
e.g. They ate [last nightTIME].
PLACE – may be realized by

• pNPs
• ADVPs
• finite and non-finite clauses with an appropriate complementizer
e.g. They ate [in the kitchenPLACE].
MANNER – may be realized by

• ADVPs
• pNP
• the NPs this way and that way
• clauses with an appropriate complementizer
e.g. He chewed [carefullyMANNER].
LOGICAL CIRCUMSTANCES occur as syntactic adjuncts:
REASON
Realized by a complementizer phrase with because.
e.g. They stayed at home [because it was rainingREASON].
PURPOSE
Realized with (in order) + marked infinitive/that or so that.
e.g. They stayed at home [in order to keep dryPURPOSE].
CONDITION
Typically realized by if clauses.
e.g. They’ll stay at home [if it is rainingCONDITION].
CONCESSION
Typically realized by although clauses.
e.g. They stayed home [although it wasn’t rainingCONCESSION].
Figure 4.7 Circumstantial roles

Grammar 71
associated with tense and aspect selections, a fairly complete picture of the idea-
tional information made explicit in a discourse can be coded and analysed. Below
are three examples of clauses coded for argument roles (Examples (44)–(46)).
(44) The male robinAgent/Theme flewAction in to the bushesGoal: location.
(45) MaxAgent/Theme flew Action into a tizzyGoal: location.
(46) MaxAgent/Theme flewAction awaySource Goal: location.
As can be seen, the coded analyses are a kind of shorthand for inheritance
of semantic features associated with predicates and their argument roles and
with features of the actual arguments in particular examples. Table 4.4 presents
a possible predicational and feature analysis for the same examples (44–46).
There are default interpretations which are respectively, literal for example (44)
and metaphorical and idiomatic for Example (45). The third Example, (46),
is potentially ambiguous. It has possible literal, metaphorical and idiomatic
interpretations.
As may be seen from Table 4.4, the default literal interpretation arises
because there is a match between the features for FLY and its argument roles
and the features of the actual arguments filling these roles. The default meta-
phorical interpretation for (45) is based on the mismatch between the features
for the Goal associated with FLY (that it be a location) and the features for
TIZZY (a mental state). To successfully interpret the metaphor, speakers must
allow the features of both the argument role and the argument to be inherited
in the instance. The ambiguity of the third example is based on the possibility
that MAX refers to a non-human participant and the indefiniteness of AWAY
which indicates a spatial orientation from an unspecified source to an unspeci-
fied goal. Literally, Example (46) could mean either (a) or (b):
a. MAX is the name of a bird. The bird flew away.
b. MAX is the name of a person. The person got on a plane which flew away
or Max flew the plane.
Alternatively, the clause could have the metaphoric interpretation in (c):
c. Max is the name of a person. He left (quickly).
The metaphoric possibility is again based on the indefiniteness of AWAY,
which allows Example (46) Max flew away to be interpreted in relation to the
idiom the bird has flown where flight to an unspecified destination is under-
stood as ‘hasty departure’ or ‘escape’. The general points here are that (1) in
interpreting Examples (44)–(46) the features of predicate and argument role
relations and those of lexical and grammatical selections instantiating argu-
ments are inherited; and (2) that coding predicate argument relations functions
as a shorthand for such inheritance relations.
We present detailed analysis of ideational representation in the sample Text
(5.1a) in Chapter 5, where lexical and conceptual information is also mapped.
Table 4.4. Example of predicational and feature analysis
Predicational Event + Theme: transferent FLY +/- Goal: Location

Analysis +/- Agent +/- Source
Features of + entity in motion + Action Spatial location

predicate + doer of action + Motion + Destination (motion
and inherent + intention + travel towards location in space)
arguments through + Source (place motion
the air begins from)
(i) The male robin Flew into the bushes
Features of PAST DEFINITE MALE IN: spatial relation,
arguments BIRD (+WINGS, containership
CAN FLY) TO: spatial relation,
destination
THE: definite
BUSHES: LIVING
THING, PLANT,
can function as
‘container’ LOCATION
since it defines an area
with a periphery and
interior
(ii) Max Flew into a tizzy
MAX: name (for A: indefinite
person or animal) TIZZY: mental
+ CAPACITY FOR state: disturbed, excited,
INTENTION state unwarranted by
circumstances
(iii) Max Flew Away
AWAY: indefinite spatial
orientation: motion from
source to unspecified
destination
4.4 The grammar of message organization

‘Ideational grammar’ has to do with what a text is ‘about’; the ‘grammar of
interaction’ has to do with the role relationships between speaker/writers and
addressees and the expression of attitude to what is predicated. The grammar
of message organization, the ‘textual function’ of language in Halliday and
Hasan’s work (1976; 1989), has to do with the organization of texts as both
cohesive and coherent within their contexts of text production. There are three
general ways in which linguists have addressed these issues: through the study
of presupposition relationships in discourse (cohesion), the study of syntac-
tic and prosodic resources for signalling information as given or new (focus
Grammar 73
and prominence), and the study of the relationships between discourses and
contexts (discourse coherence and intertextuality). We discuss coherence in
Chapters 5 and 8. The discussion of cohesion is based on the foundational
works of Halliday and Hasan (1976; 1989) and the terminology used is largely
theirs.
4.4.1 Cohesion
The study of cohesive features of texts involves isolating those linguistic
features which create relationships of presupposition between sentences,
allowing us to interpret one sentence in relation to preceding and following
sentences, and between sentences and the extra-textual contexts in which
they are produced. This initial distinction gives a binary contrast between
reference internal to a discourse (or endophoric reference) and reference
external to a discourse (or exophoric reference). Within endophoric reference
there are the further contrasts between anaphoric reference (which is refer-
ence ‘backwards’ to information assumed to be already available), cataphoric
reference (reference to information yet to be presented), and homophoric ref-
erence (reference to ‘self’ or to a head element). Within exophoric reference
there is also a possible binary distinction between situational reference, refer-
ence to objects, events and so on present in the immediate perceptual or shared
conceptual environment of speaker/hearers, and intertextual reference, refer-
ence to other discourses presumed to be known by addressees. All of these
relations can be summarized as a system as in Figure 4.8.
Reference
Exophoric
Situational: to environment
Intertextual: to other discourses
Endophoric
Anaphoric: to preceding discourse
Cataphoric: to ensuing discourse
Homophoric: to head
Figure 4.8 Referential relations

Within these broad categories of types of reference, five more or less distinc-
tive subtypes of cohesive relation are recognized. They are reference, substitu-
tion, ellipsis, conjunction and lexical cohesion (Halliday and Hasan 1976).
Reference handles items whose specific function is to index referential rela-
tionships, that is items whose interpretation depends, in part, on the presence
of other elements. There are three general types. The first is called personal
reference. This involves the use of personal pronouns such as I, me, mine, you,
yours and so on and possessive deictics as in my hat, her elephant. Personal
referencing items are definite, referring to known or given information, and
thus are typically used for situational or anaphoric reference. The second type
is demonstrative reference with demonstrative pronouns (this/that, these/those)
and spatial and temporal ‘pro’ words (here/there, now/then). Demonstratives
are also definite and frequently used exophorically as well as being used
endophorically. Halliday and Hasan (1976) include the articles the/a within
demonstrative reference. We find the notion of an indefinite demonstrative (a)
anomalous and prefer to think of the articles as an independent two-item class
indicating only definite/indefinite + singular reference. Comparative reference
includes all items which can be used to establish qualitative or quantitative
comparisons between entities. Thus, the comparative forms of adjectives and
adverbs (e.g. qualitative: easier, cleverer, more sweetly, more cleverly, quanti-
tative: more, fewer, less and so on), deictic adjectives and adverbs expressing
identity (same, equal/ly, identical/ly), similarity (such, similar/ly), and differ-
ence (other, else different/ly) may form cohesive ties insofar as the interpreta-
tion of any comparison requires a reference ‘point of comparison’ as in Sam
is a very clever linguist. However, Mary is cleverer when it comes to syntactic
theory. The domain for the interpretation of cleverer in the second sentence
is the predication A VERY CLEVER LINGUIST: Classified [SAM]. Normal
conventions for coding reference relations are to assign indices of some sort
to the referents and then track them through the discourse as in the coding
below. However, if the reference type needs to be recovered from the coded
data without inspection, more detailed codes may be used (e.g. R1/RP1 for he
in the example).
Coding: Sam1 is a very clever linguist.

He1 is especially clever when it comes to syntactic theory.
Substitution of one element for another is cohesive insofar as substitutes

require textual antecedents. Nominal, verbal and clausal substitutions are rec-
ognized. Nominal substitutions are the items one, ones and same. The verbal
substitute is do. Clausal substitutes are so and not. Examples are John bought
a new hat. Mary bought one too where the item one is a nominal substitute
Grammar 75
for a new hat in the preceding sentence; Does he know? He may do where do
substitutes for the verb know; Does he know the answer? I think so where so
substitutes for the clause he knows the answer.
Coding: Does he know the answer2?

I think so SUB2.
Ellipsis can be thought of as a variant of substitution (substitution by 0 if

you like) and, like substitution, requires an antecedent in a preceding sentence.
It also can be nominal, verbal or clausal. Examples are I haven’t the slightest
idea. Nor have I – where the noun phrase the slightest idea is ellipted; Don’t
move! Why shouldn’t I? – where the verb move is ellipted; and Did you see
anyone? Yes – John where part of the clause (I saw) is ellipted.
Note that substitution and ellipsis are like reference in that they may be
thought of as PRO forms (including 0) for clausal and verbal elements. The
nominal substitutes one, ones and same seem to be treated as substitutes on the
grounds that they involve nominal reference which is not personal or demon-
strative. However, unlike demonstrative, comparative and personal reference,
ellipsis and substitution only work anaphorically within a text. They require
antecedents in preceding sentences for their interpretation.
Coding: Don’t look3!
Why shouldn’t IEllip3?
Conjunction as a type of cohesive relation has to do with logical and tem-

poral relations between sentences which are overtly marked by ‘sentence con-
junctions’ such as however, moreover, nevertheless, consequently, therefore,
meanwhile, afterwards and so on. Such relations are cohesive insofar as they
presuppose other propositions which are temporally or logically related.
Coding: [The view is magnificent.]1 HoweverCONJ1 you shouldn’t look

because the drop will make you dizzy.
The final type of cohesive relation is lexical cohesion. Lexical cohesion

involves the occurrence of lexical items which have some kind of semantic rela-
tion to each other. Halliday and Hasan (1976; 1989) and Hasan (1985) dis-
cuss three types of lexical cohesion. The most obviously cohesive use of lexical
items is simple repetition of an item or items. The occurrence of the lexical
item may be distributed through different word categories and still be treated as
repetition. For instance, our use of the noun cohesion and the adjective cohesive
in the preceding sentences, and indeed throughout this text exemplifies cohesive
repetition. Note that a good deal of intertextual coherence is established by rep-
etition as well. That is, the use of lexical items which have high salience in other
discourses may activate references to those discourses if they are known.
Coding: Don’t look4.
Why shouldn’t I lookLC-REP4?
A second type of lexical cohesion has to do with synonymy where items are
interpreted in relation to each other because they share some semantic compo-
nents. For example, in this text our use of cohesive relation, presupposition and
cohesive tie are close synonyms and form cohesive ties. Halliday and Hasan
(1976) include within the category of lexical cohesion synonymy relations of
superordination and antonymy. Superordination has to do with categorial
structure in taxonomic hierarchies. Here, lexical cohesion functions as a super-
ordinate term that is cohesive with the subtypes introduced (lexical repetition,
synonymy and collocation). The use of antonyms functions cohesively insofar
as word pairs such as dead/alive, white/black, awake/asleep, moving/still or for
that matter synonym/antonym exhibit tendencies to co-occur. If you think of a
conversation, say, in which a speaker reports that they have been feeling sick,
it is likely that they will be asked if they are feeling better or that there will be
some expression of hope that they are soon well.
CODING: Collocation is the third broad type1 of lexical cohesion2 … As a

cohesive L-REP2 category L-CSYN1, this is a sort of rag-bag L-CSYN1.
Collocation is the third type of lexical cohesion recognized by Halliday

and Hasan. As a cohesive category, this is a sort of rag-bag insofar as it is a
general term used to cover anything not dealt with under repetition and syn-
onymy. Moreover, repetition and synonymy involve collocation since items
which are cohesive in these ways are also collocated. The category is useful
though as a gloss for underspecified types of semantic relation. For example,
items such as smoke/fire, ice/cold and so on are often collocated and are cohe-
sive. (The semantic relation in these examples is effect/cause and attributive
respectively.) Similarly, items such as shoot/snap have a high probability of
occurring with collocates such as film/picture (where film and picture might
be expected to appear as result). Figure 4.9 summarizes the options for relations
of this type.
Grammar 77
relational [doctor/plastic surgeon]

hyponymy
inclusion individual [flower/tulip]
co-hyponymy [any relationship between above examples]
converse [doctor/patient]
superordinates
opposition gradable [‘old’/‘new’]
binary
non-gradable [dead/alive]
contrast
non-cline [1981/1982]
non-binary
cline [hot/warm etc.]
near [veer/turn]
similarities synonymy
(+)attitude [win/triumph]
complete [A/A/A]
repetition inflectional [man/men, girl/girls]
scattered
derivational [man/mannish]
unit [flock/geese, herd/elephant]

kind [cattle/cow, mankind/man]
taxonomic collectivity
meronomy [room/chair, house/room]
(concept
co-meronomy [chair/sofa, hand/foot]
oriented)
participant [club/member, team/player]

concept/
relational perspective [top/door]
lexical
relation composition constitution
content [garden/tree]
network (part/whole)
partition
make-up [jacket/lapel, nib/pen]
consistency [knife/steel, table/wood, house/brick]
predicational [process/patient e.g. wash/dishes]

process [event concatenation e.g. want/to go]
extensions
participant [thing/attribute e.g. large/elephant]
nuclear predicational [process/circumstance e.g. sleep/at night]

non-taxonomic enhancements process/manner [e.g. run/quickly, eat/well]
(event oriented) participants [apposition or Qn e.g. my brother/the doctor]
consequentiality [e.g. injure leg/go to hospital/get treatment]
expectancy
compatibility [e.g. go home/eat meal/watch TV/wash self/go to bed]
Figure 4.9 Conceptual/lexical relations network2
In the box below, the collocational relation between kettle and cup of tea
could be characterized as event-oriented compatibility based on purpose.
Coding: Sam put the kettle5 on.

He made a nice cup of tea LC-COL5.
Table 4.5. Errors in use of cohesive devices
Cohesive tie Criterion of error Example

Reference Referent is not in text, referent is Baseball or soccer. | I used
ambiguous or wrong pronoun is used to do that for sportsnight.
Substitution and ellipsis Ellipted or substituted word or phrase Would you like to go to a
cannot be explicitly identified movie? I don’t have one.
Conjunction Semantic relation that conjunction There’s a fourth person.
is signifying is wrong | But there are four people.
Lexical cohesion The second word of the pair What happens if you go on a
occurring in the adjacent sentence passenger train? | You pass
does not contribute to the the test.
cohesiveness, or does not
follow from the previous sentence.
Lastly, items which are collocated but do not entail any overt semantic rela-
tion that would lead you to expect them to co-occur can be reinterpreted or
recoded simply by virtue of the fact that they are collocated. For non-linguist
readers of this book, for instance, the items agent, patient and argument role
may well not have entailed any overt semantic relationship. However, we hope
that having introduced the notion that argument role is a superordinate term in
relation to agent and patient non-linguist readers will begin to expect these to
co-occur.
Investigating the way in which cohesive devices are used by particular speak-
ers may also involve identifying errors or non-conventional use of cohesive
devices, determined on the basis of a clear breach of linguistic rules (McKenna
and Oh 2005). Reference, for example, may not be recoverable. Unrecoverable
references do not form cohesive ties with a reference point of comparison but
can still be coded. We follow McKenna and Oh’s criteria for scoring erroneous
ties, including unrecoverable reference (Table 4.5).
The significance of cohesion in relation to the interpretation of particular
discourses is perhaps best illustrated in relation to different varieties of text. If
for example one considers discourses which vary with respect to the medium
and mode of communication, it is apparent that different modes entail differ-
ent types and different degrees of cohesive relation. Spontaneous face-to-face
dialogues will be different from telephone conversations because in the former,
but not the latter, a physical environment is shared so that situational exophoric
reference is both more likely to be used and has a better chance of succeeding.
Spontaneous spoken monologues might be expected to exhibit more internal
cohesion than conversations, and monologues which are distinctly lacking in
cohesive relations are often understood as signs of fatigue, distraction, mental
illness or other disorder. Within written modes note that, for example, texts
Grammar 79
which are written to be spoken as if not written (e.g. plays) can exhibit all the
types of cohesive relation available for ordinary spoken modes although invari-
ably they will be more tightly structured than spontaneous speech. Similarly,
some texts are ‘written to be read as if thought’. One of the things this mode
‘licences’ are fragments of text which do not have any overt cohesive rela-
tion to the rest of the text (as in ‘stream of consciousness’). Such ‘deviations’
are licensed by the mode because we can interpret them as coherent with the
presentation of a participant’s thoughts. Similarly, the same mode licences
the use of partial or complete exophoric reference because, as presented from
the perception of the ‘knower’, such reference makes sense. Journals kept only
for personal use may present similar features. The general point then is that, on
the one hand, certain modes presuppose or entail certain types of cohesion and,
on the other, that different cohesive relations can in fact be signifiers of modes.
The relevance of this in relation to the interpretation of particular discourses is
that a discourse may be expected to have cohesive features that are appropriate
to its function.
Halliday and Hasan’s (1976) description of cohesion was largely developed
with reference to written texts and was focused on making presupposition
relations with overt signifiers explicit. As Figure 4.10, an extract from a semi-
structured interview between a caregiver and a researcher, illustrates, analysis
based on the possibilities they describe does help to make local presupposi-
tion relationships explicit. The text is densely cohesive, with demonstrative
IV; Can you tell me a little bit about [the way1 her2 memory3 has changed]4?
CG; As far as like I said the pills5.
She R-P2's very forgetful LC-C3,4 in those things R-D5.
[It R-P4's a lot of [little things LC-R5]]4,
um [it R-P4's not so much names or faces6]4,
she R-P2 doesn't tend to forgetLC-R3,4 those R6 things LC-R6.
SheR-P2'll tell me [stories of long ago]7 which I don't know
um [but then again she R-P2'll tell me the R-D7 sameR-D7, 7 storyLC-R7 maybe three days later
and things have changed LC-A7 within the R-D7 storyLC-C7]8
IV; umhum
CG; [So CONJ sheR-P2's forgettingLC-R4, C3 those R-D8 things]4,8 ah.
For [right now it's really just a [memory LC-R3 loss]LC-C4 for herR-P2]4.
Figure 4.10 Cohesive analysis

and personal reference, lexical repetition and collocation used extensively. It

is also evident that the extended turn taken by the caregiver must be read as a
response to the researcher’s question, Can you tell me a little bit about the way
her memory has changed? This means that not only does the adjacency pair,
question–response, contribute to the coherence of the interaction, but also that
there are relationships, not just between ‘words’, but between the interviewer’s
prominent the way her memory has changed and the propositions the caregiver
offers to elaborate the way her memory has changed. We have indicated rela-
tions between presupposing constituents by bracketing and assigning an index
to the constituents.
4.4.2 Focus and prominence

The second major way in which members of the London School have addressed
message organization has to do with the resources available to present parts of
a message as being focal (points of departure, backgrounded information) and
other parts of a message as being prominent (informationally salient [new]
and/or marked). Two major types of phenomenon are often considered in these
contexts. The first is the theory of THEME/RHEME for sentences, which is
a development of the Prague School linguists’ work on Functional Sentence
Perspective (Halliday 1967; Danes 1974; Firbas 1992). The second is the
theory of Given and New (Halliday 1967). We use Gregory’s terms FOCUS/
PROMINENCE for these to avoid confusion with the argument role called
theme (Gregory 1988; Watt 1990).
The basic idea of FOCUS/PROMINENCE is that there is a statistically high
probability that certain types of element will occur as the first element in a
clause and that any departure from such normative probabilities will be in
some degree marked. Thus, statements in which the first element is subject
are said to involve FOCUS only. That is, they are UNMARKED, since statisti-
cally this is highly probable. Similarly, questions beginning with WH words
or an auxiliary verb only involve FOCUS (although in many types of text,
especially written modes, questions themselves are ‘more marked’ than state-
ments). Also, in commands, the unmarked thematic element would correspond
with either a main verb in Neutral (‘bare infinitive’) form or the ‘let’ auxiliary
verb. Which one you get depends on the type of command, and note that the
same proviso holds for commands as for questions, viz. that commands may
overall be more marked than statements and, perhaps, questions. However, the
variety of text being considered really does matter here. The language used by
drill sergeants ordering troops, by directors rehearsing actors for performance,
as well as the language of cooking recipes, and instruction manuals of all types
share a feature which is that there is a predictably high frequency of command
realized by jussive imperatives. That is, command is the unmarked choice of
speech function in these varieties.
Grammar 81
FOCUS elements interact with the creation of PROMINENCE insofar as

prominence is established in part by the location of the major pitch movement in
a tone group. Usually, the major pitch movement coincides with the last lexical
item in a clause and thus the constituent of which that element is a part is promi-
nent. Thus in //1 I like/ coffee// we would say that the subject I is the unmarked
focus of the clause and the complement coffee is prominent because that is
where the major pitch movement occurs. If one moves, say, the complement
to initial position, the effect is to break the sentence up into two tone groups
(and hence two chunks of information) with a major pitch movement occur-
ring both on the moved element and on whatever is the remaining, ‘last lexical
item’. Thus, the statement Coffee, I like is MARKED. The first element is both
FOCUS and PROMINENT (it is site of a major pitch movement) and the verb
like is the locus of the second major pitch movement and so is also prominent.
Somewhat similar effects can be created by other strategies. For instance, it
is possible to simply shift tonicity ‘forward’ to the middle or beginning of a
clause without changing the syntactic structure of the clause. If the initial ele-
ment is selected as tonic, it is both focus and prominence for the message but,
barring other effects, there will not be a second prominent element as there is
in the case of pre- or post-posed elements.
In general, the importance of focus in relation to texts is that the elements
that regularly occur as focus establish the ‘back-grounded’ or ‘given’ informa-
tion for the discourse (and may reflect aspects of the text type). For example,
in a story the protagonist might well appear as FOCUS in successive sentences
about their activities, whereas in a logically structured argument we may find
that what appears as the prominent element in one sentence will be focus in
the next and so on. Contrastively, those elements which are prominent in indi-
vidual sentences can, cumulatively, create prominence for the discourse as a
whole. They may effectively mean that an aspect or aspects of a message are
foregrounded. Focus and prominence are thus semantic functions with a vari-
ety of discoursal, syntactic and phonological realizations that interact with each
other. The morphosyntactic resources enabling marked selections for focus and
prominence are listed below where focus is coded with underlining and promi-
nence with bold face. (The example is originally from Halliday 1967.)
Coding: Peter left his luggage at the station.

(Focus = UNDERLINE; Prominence = BOLD)
4.4.3 Morphosyntactic resources for marked focus and prominence

Topicalization A topic element (TE) may be adjoined to the begin-
ning of a clause to explicitly mark its topic. It usually refers anaphorically to
preceding discourse or exophorically to context.
(47) Speaking of luggage [TE], Peter left his at the station.

(48) According to Peter [TE], he left his luggage at the station.
Preposing An element of clause structure may be preposed [PreP].

The topic and preposed elements are both focus (starting place for the message)
and prominent (they occur as a separate intonation unit). Within the clause, the
last lexical item is prominent.
(49) Peter [PreP], he left his luggage at the station.
(50) His luggage [PreP], he left at the station.
Post-posing Elements of clause structure may be post-posed [PosP] as

in (51). The post-posed element occurs in a separate tone group and is prominent.
(51) Peter left it at the station, his luggage.
Cleft and pseudo cleft clauses allow prominence to occur on a single marked
constituent:
Cleft
(52) It was Peter who left his luggage at the station.
(53) It was at the station Peter left his luggage.
(54) It was his luggage Peter left at the station.
The constituents Peter, at the station and his luggage are made prominent in
these constructions, while the focus of cleft clauses is the speech function
itself. (Compare 52–54 with the corresponding polar questions.)
Pseudo-Cleft
(55) What Peter did was leave the luggage at the station.
(56) What Ellie likes is chocolate.
Clauses with preposed, post-posed, or topicalized elements are similar in
that they each require two tone groups and therefore two points of prominent
(new) information. Clauses with cleft or pseudo-cleft constructions allow
markedness with a single point of prominence.
A related set of systems are those which allow the emphasis or omission of
an argument. Among the most commonly described are passivization, ergativ-
ity and nominalization.
Passivization Passivization reorders the elements of a clause so that

a thematic argument appears in the focus (subject) position and an agent, cause
or experiencer is either deleted (backgrounded) or made prominent in a final
position as in (57), (58), and (59) respectively:
Grammar 83
(57) a. The luggage was left at the station.

b. The luggage was left at the station by Peter.
(58) a. Point Pleasant was destroyed.
b. Point Pleasant was destroyed by Hurricane Juan.
(59) a. The idea was thought ridiculous.
b. The idea was thought ridiculous by most faculty.
Ergative Verbs Verbs of motion and change of state (ergative verbs)

allow inclusion or omission of Agent or Cause without the need for passivization.
When Agent or Cause are present and clauses are active, they are focus as in (60a)
and (61a); when absent, the thematic argument is focus as in (60b) and (61b).
(60) a. Peter rolled the rock.
b. The rock rolled.
(61) a. The water changed the shoreline.
b. The shoreline changed.
Nominalization refers to processes which incorporate predicational informa-
tion into noun phrases. These processes include changing word class from verb
to noun as in destroy > destruction; employ > employment; realize > realiza-
tion. Such nominalizations of verbs can allow backgrounding of processes and/
or participants because they can be presented within the noun phrase as ‘given’
information (through use of definite articles for instance); they can occur in non-
prominent positions, and the arguments required by the verb are not required by
the noun. For example the use of destroy as a verb requires an Agent/Cause and
Patient as in (62a). The nominalization of destroy allows Agent in (62b) and
Patient in (62c) to be back-grounded or deleted altogether (62d).
(62) a. The Americans destroyed Khabul.
b. The destruction of Khabul was sad.
c. The destruction by the Americans was sad.
d. The destruction was sad.
Other nominalizing processes incorporate clauses into NPs which can have
similar effects as in
(63) The destruction which occurred in Khabul was sad.
The importance of morphosyntactic resources for these features in clinical dis-
course analysis is that, while they are all resources which enable speakers to
select focus and prominence and therefore presumably increase the potential
for effective communication, with the exception of ergatives, some of them
also increase information processing loads in both speech production and
speech reception. Processes which re-order clause constituents increase their
syntactic complexity, while those which incorporate predicational information
Topicalization – topic element adjoins to the beginning of a clause to explicitly mark its
topic
Preposing – an element of clause structure may be preposed
Post-Posing – an element of clause structure may be post-posed
Cleft and Pseudo Cleft Clauses – allow prominence to occur on a single marked
constituent
Passivization – reorders the elements of a clause so that a thematic argument appears in

the focus (subject) position and an agent, cause, or experiencer is either back-grounded
through deletion or made prominent in final adjunct position
Ergative Verbs – verbs of motion and change of state allow inclusion or omission of
Agent or Cause without the need for passivization
Nominalization – predicational information is incorporated into noun phrases allowing

for back-grounding of processes and/or participants. Processes for nominalization
include:
• changing word class from verb to noun as in destroy > destruction
• incorporating clauses into NPs
Figure 4.11 Morphosyntactic resources for marked focus and prominence
into noun phrases increase their lexical density (cf. Halliday 1987, 1989). Both
syntactic complexity and lexical density have implications for the evaluation
of language in clinical settings (e.g. van Dijk and Kintsch 1983; Snowdon
et al. 1996; Kemper et al. 2001).
English morphosyntactic resources for marked focus and prominence are
summarized in Figure 4.11.
5 Phase and contexts of culture and situation
5.1 Contexts of culture and situation

All discourse is produced in context and interpretation depends on contexts
of production and interpretation being, in some measure, shared. Early ethno-
graphic work addressed context dependency by positing contexts of culture
and context of situation (Malinowski 1923; 1935; Firth 1957). Context of cul-
ture accounted for sets of culturally specific beliefs, expectations and practices
in terms of which people interpret events around them. Context of situation
referred to patterns of behaviour and talk which appear so regularly in associa-
tion with a particular activity that they are understood as (abstract characteri-
zations of) the function of the situation type. Behaviours which do not reflect
some expected pattern can be interpreted as irrelevant, and behaviours which
appear totally unrelated to the contexts in which they occur may be judged
uninterpretable.
Later work by Halliday, Hasan, Gregory, Martin and others refined and
developed ideas of context. Our view is once again a hybrid, informed by
Halliday’s ethnographic perspective (e.g. 1977; 1978; 1984; 1994), by our
awareness that contexts are significantly matters of what speakers know (e.g.
van Dijk 1977; 2006; van Dijk and Kintsch 1983; Gregory 1988), and by work
in AI, psychology and discourse analysis on top-down cognitive models as to
what ‘contextual knowledge’ might be like. The latter approaches (and ours)
differ from traditional functionalist and ethnographic approaches in explicitly
situating context in neurocognitive domains of semantic and episodic memory
(see also van Dijk 2006). Any and/or all neurocognitive systems involved in
information processing may be engaged when processing physically and/or
temporally present situational information. Which systems are engaged will
depend both on the nature of the information presented and the processors’
capacities for processing it.
Taking this perspective, context of culture which in Gregory’s (1988) formu-
lation referred to known parameters of cultural variation in terms of temporal,
geographical, social and individual provenances of speakers, can be construed
as referring specifically to semantic and episodic memories speakers have
85
that are conditioned by and in these environments (van Dijk 2006). Similarly,
context of situation in this framework refers to configurations of ideational,
interactional and medium information which are encoded episodically and
semantically represented in generic patterns of function, meaning and use.
They are limited by our contexts of culture.
5.2 Frames, schemas, scripts and scenarios

A good deal of the information represented in memory systems is not inherently
linguistic or even discoursal. For example, we have more or less elaborated
knowledge (that varies in different contexts of culture) about what to do if we
feel unwell. Our knowledge can include information that lets us make judge-
ments about whether the feeling is minor, temporary and likely to go away
without help, or sufficiently severe and persistent as to require attention from a
medical professional. Should we decide that help is required, we have knowl-
edge about where to go and who to see, about the types of role relationship we
might expect to have with the healthcare professional and about what they may
do to help us. Such knowledge may be described in frames, schemas, scripts
and scenarios. Frames are representations of (culturally specific) knowledge
and belief associated with events, states or things. In early versions of frame
theory, knowledge was assumed to be propositionally represented and rela-
tively static (e.g. Minsky 1975; van Dijk 2006). Schemata are similar repre-
sentations but include abstract elements in different modalities such as visual
schemata for faces or houses, or discourse schemas for narrative such as those
articulated by Propp (1928), Barthes (1968) and Labov and Waletzky (1967),
as well as conventional and/or stereotypic conceptual organizations (Tannen
and Wallat 1993). Scripts are representations of highly abstract generic struc-
ture potential associated with events – they delineate the inherent elements of
events and their sequences (Schank and Abelson 1977). Scenarios were devel-
oped to refer to characteristic properties of situations that were not necessarily
represented in propositional form (Sandford and Garrod 1981). They cover
some of the same ground as frames and schemas but focus on identifying
multi-modal aspects of situations such as features of the setting.
Thus, in the situation ‘being unwell’, if one makes a decision to seek help
rather than wait, frame knowledge might include information about options
to make an appointment, go to a walk-in clinic, go to an emergency unit at a
hospital, or call an ambulance.
Which option is chosen will depend on other frames/schemas (e.g. what
we know about the services available at the different settings) together with
an assessment of how urgent the need for medical attention is, which requires
integration of proprioceptive experience with (yet more) frames/schemas
about sickness. Choosing ‘go to a walk-in clinic’ suggests an evaluation
Phase and contexts of culture and situation 87
including ‘not immediately life threatening’, ‘not requiring immediate surgery

or resuscitation’, but sufficiently disturbing as to need ‘same day’ assessment
and/or treatment. (There are, of course, other considerations that might moti-
vate such a decision such as access, whether one sees any doctor regularly and
so on.) Going to the walk-in clinic evokes not only frames/schemata, but also
features of the setting (scenarios) which might include: a waiting room which
will have seating and may have other people waiting to be seen; a receptionist
at a desk; a nurse, doctors, consulting and examination rooms; and so on. Even
the quality of the light (often fluorescent) and the type of decor and furniture
may be anticipated. A script for ‘going to the walk-in clinic’ might include
getting there, entering the clinic, talking with the receptionist about the prob-
lem, filling out health insurance and diagnostic forms, establishing a wait time,
being told to go to a consulting/examination room, removing any necessary
clothing and so on.
The relevance of such cognitive models for discourse becomes apparent if
one considers a telephone conversation in which a speaker says ‘I had to go
to the walk-in clinic yesterday.’ This statement immediately evokes not only
frame information about possible severity of the problem motivating the action,
but also associated script and scenario models which condition expectations
about what the speaker might say next. If the speaker reports ‘Yah, I got hit by
a car on my bike and dislocated my shoulder and broke my arm in two places’
an addressee might be disconcerted since, outside war zones, walk-in clinics
are not usually equipped to deal with injuries of this nature. At this point, our
hypothetical addressee will need to suspend expectations and wait (or ask)
about what happened next. An account in which the bone is set and the shoul-
der relocated at the walk-in clinic will require a modification of the frame for
‘walk-in clinic’; an expected one in which the speaker is sent on to an emer-
gency unit is likely to have the addressee wondering whether the speaker didn’t
also hit her head. The general point is that simple mention of an event such as
‘I had to go to the walk-in clinic yesterday’ sets up a whole range of expecta-
tions about what the speaker might say next, based on shared knowledge of
why one goes to walk-in clinics, what walk-in clinics can do, what happens
when people go there and so on.
Frames, schemas, scripts and scenarios are useful insofar as they help to
articulate the mental models which characterize the cultural and generic infor-
mation we use to make sense of what is going on in situations and to construct
relevance and coherence in discourse. There is evidence, which we discuss
in Chapter 8, that people actually do use such mental models in constructing
relevance and making assessments of coherence. Some of these, particularly
highly abstract models for event and discourse schemas appear to have univer-
sal relevance. For example, the schema for stories appears always to involve a
‘complicating event’ (Labov and Waletsky 1967) and a schema for buying and
selling does involve exchange of goods/services for some medium (currency

or other). But the details of what counts as a complicating event or what can
be bought or sold for what medium of exchange vary across cultural contexts
and entail generic models of the kind represented in frames, scripts and sce-
narios. It is also important to acknowledge the extent to which such models
are provisional and open to revision when the instance supplies information
that conflicts with properties of the known model (Hudson 1984; 2007). New
learning is not possible if features of models cannot be altered by experience.
Finally, it is worth emphasizing that discourses usually do not explicitly instan-
tiate top-down models. Rather, the linguistic signals within a discourse provide
indexical signs of such models (Barthes 1968; 1994). Consequently, we follow
van Dijk and Kintsch (1983) in assuming that relevance and coherence are not
properties of texts, but rather achievements of discourse processors in integrat-
ing linguistic input with contextually available information.
5.3 Coherence
Coherence itself is a term for the complex of inferential processes that enable
us to produce discourse that others can make sense of and to interpret what
others say. Coherence thus depends on linguistic information instantiated in
discourse, inferences we draw from this information, presuppositional rela-
tions involved in linguistic cohesion, our situational knowledge (represented in
models such as frames, scripts, scenarios, and schemas), and our attitudes and
evaluations of these. The study of discourse in clinical contexts is a study of
the breakdown, at some level, of discourse coherence. Imaging studies over the
last few years have made the development of a model of discourse coherence
for which the neural substrates are understood begin to seem like an achievable
goal (see e.g. Stowe et al. 2005; Ferstl et al. 2008; Mason and Just 2007 for
suggestive reviews). However, given the wide range of phenomena involved in
producing and interpreting discourse as coherent, and the potential configura-
tions in different disorders that could lead to impairment, isolating particular
features contributing to perceived coherence or incoherence is highly relevant
in clinical settings.
5.4 Registers and dialects

In addition to background knowledge that informs discourse production and
interpretation, speakers’ knowledge of language, including functional varieties
(registers), associated structural knowledge, and dialect varieties, is acquired
in and conditioned by their contexts of situation and culture. So, a speaker of
English will know at least one dialect variety which can be identified in terms
of the temporal, geographical and social provenances in which it was acquired
(such as contemporary, middle class, Toronto English). Of course, they may

also know other English dialects as well as other languages. Depending on the
contexts of culture and situation which speakers have experienced, they will
also know a variety of registers. Pursuing our medical motif, an example is the
register of medical interview which involves an expert medical person, a patient
and some reason (often illness) for the patient’s presence and the interview
occasion. These interviews are normally face-to-face. The medical interview
situation results in linguistic selections in which the experience referred to will
have lexis associated with illness, symptoms, diagnosis, prognosis and treat-
ment and interactional patterns which reflect the medical professional’s role as
expert and the patient’s role as advice/aid seeker. For instance, the professional
asks questions about the patient, may make statements about the patient, and
may give commands to the patient (take these three times a day and call me if
you experience any dizziness). The patient may ask questions and make state-
ments but these are unlikely to be about the healthcare professional’s personal
life. Other interactional systems reflecting the social role of the expert and their
relationship to the patient may result in differential use, for instance of address
terms (Dr. X/Sarah) and politeness markers (lift up please). Organizationally,
the face-to-face interview will result in spoken discourse, dependent on instan-
tial context, and therefore is likely to be highly exophoric (Does it hurt here or
here? Do you feel that now?). If there has been previous interaction there may
be shared and therefore underspecified reference.
5.5 Phase
Phase is the construct used here to link context and discourse and track relevant
information (e.g. Gregory 2002). A phasal analysis maps the changing contexts
of situation at different levels of delicacy based on coding and interpretation
of shifts in linguistic patterns. One does this by analysing the grammatical and
discourse features presented, and identifying patterns of regularity in one or
more functions. Transitions between phases, when they occur, are marked by
presence of both anaphoric and cataphoric reference, signalling how discourse
to come is to be interpreted relative to what has already been said. We illustrate
phasal patterns in Text 5.1, which is a slightly more extended fragment of the
caregiver–interviewer Text 4.1 in Chapter 4.
Text 5.1 How her memory has changed

(1) IV: Can you tell me a little bit about the way her memory has changed?
(2) CG: As far as like I said the pills. She’s very forgetful in those things. It’s a lot of little
things # um it’s not so much names or faces # she doesn’t tend to forget those things.
She’ll tell me stories of long ago which I don’t know um but then again she’ll tell me
the same story maybe three days later and things have changed within the story.
(3) IV: umhum

(4) CG: So she’s forgetting those things ah. For right now it’s really just a memory loss
for her.
(5) IV: What did her daughter pick up that
(6) CG: Um just repeating
(7) IV: Just repeating?
(8) CG: Repeating constant repeating. Ahm # I’m trying to think of what else ahm #
I don’t know you’d have to speak to her what changes she might of found in her.
But I find the repeating is yeah.
(9) IV: She doesn’t get sort of balled up in the kitchen or anything like that?
(10) CG: No: no: If she’s concentrating on doing the dishes # she’ll do the dishes. She’s
not jumping from one thing to the next. No I haven’t seen her do that. Like ah no.
(11) IV: And she does do the dishes? That is something.
(12) CG: Yeah yeah yeap.
(13) IV: How about the paper?
(14) CG: Does she read the paper? Ah they don’t generally get the paper but when they
do yeah I think she picks it up. Um I’ve noticed too now when she she likes to read.
I did take notice this one one time she had a book mark in it where she had stopped
reading and I came back the next day the next morning I came back and I noticed
that the bookmark had moved but it went backwards instead of forwards. Like she
started more to the beginning of the book and started reading there.
Consistent selections from experiential, interactional and organization
f unctions will correspond with a single primary phase and its associated generic
situation. For example, this bit of discourse is part of a home-visit interview,
the goal of which is to establish a phenomenologically realistic baseline evalu-
ation for the patient’s memory and ability to participate in activities of daily liv-
ing (ADLs) (Rockwood et al. 2002). This situation is reflected in the medium
(face-to-face semi-structured talk) and in turn and topic patterns instantiated
by speech function and predicational selections. The researcher asks questions
which direct topic selections – here about memory (1)–(8), household chores
(9)–(12) and reading (13)–(14). For each topic, the caregiver responds, stat-
ing details that elaborate on particular abilities and deficits, and then links the
observations back to the researcher’s more general questions with a summary
statement. There is thus an overall consistency of roles as represented in the
speech function and turn-taking patterns of speakers. The discourse is also
cohesive as elaborated above and there is a global coherence of ideation insofar
as what is talked about is the patient’s cognitive well-being as reflected in her
memory and ability to stay focused. So in terms of the generic function of the
situation we might treat this as a single primary phase.
Secondary phases within it are marked by the shifts in ideation from mem-
ory, to household chores, to reading. Within each of these secondary phases,
we see the caregiver taking extended turns at talk, moving into a monologic
mode in order to fulfil the role of ‘giver of information’, the speaker respon-
sible for topic development. The organization of information within these
1. IV; Can you tell me a little bit about the way her memory has changed?
IV 2. CG; As far as like I said the pills. She's very forgetful in those things.
3. CG; It's a lot of little things.
4. CG; um it's not so much names or faces,
MEMORY
5. CG; she doesn't tend to forget those things.
CHANGES
6. CG; She'll tell me stories of long ago which I don't know um
7. CG; but then again she'll tell me the same story maybe three days later
CG
8. CG; and things have changed within the story.
9. IV; Um hum
10. CG; So she's forgetting those things ah.

11. CG; For right now it's really just a memory loss for her.
OTHERS 12. IV; What did her daughter pick up that

IV
NOTICE
13. CG; Um just repeating
14. IV; Just repeating?

CG
15. CG; Repeating constant repeating.
16. CG; Ahm,, I'm trying to think of what else ahm ,,,
HOME 17. CG; I don't know
VISIT 18. CG; you'd have to speak to her what changes she might of found in her.
19. CG; But I find the repeating is yeah.
20. IV; She doesn't get sort of balled up in the kitchen or anything like that?
IV
21. CG; No: no:
DAILY 22. CG; If she's concentrating on doing the dishes, she'll do the dishes.
ACTIVITIES 23. CG; She's not jumping from one thing to the next.
CG 24. CG; No I haven't seen her do that.
25. CG; Like ah no.
26. IV; And she does do the dishes?

IV 27. IV; That is something.
EVALUATE
CG
28. CG; Yeah yeah yeap.
29. IV; How about the paper?
IV Tr 30. CG; Does she read the paper?

31. CG; Ah they don't generally get the paper
READING 32. CG; but when they do yeah I think she picks it up.
33. CG; Um I've noticed too now when she she likes to read.
34. CG; I did take notice this one one time she had a book mark in it where
she had stopped reading
CG 35. CG; and I came back the next day
36. CG; the next morning I came back
37. CG; and I noticed that the bookmark had moved
38. CG; but it went backwards instead of forwards.
39. CG; Like she started more to the beginning of the book and started
reading there.
Figure 5.1 Model phasal analysis

mini-monologues also differs in that, after initial responses which are often
elliptical, clauses are full (not elliptical) and there is internal cohesion as sub-
topics are introduced and developed. As noted, the caregiver also links specific
topics back to the interviewer’s questions, in effect telling the interviewer how
to interpret the details offered and at the same time signalling the end of each
phase. The phasal pattern of this text is laid out in Figure 5.1.
Phase is a heuristic for showing patterning of information in discourse by
which we can interpret people’s ongoing understanding of the context of situ-
ation relative to the topic at hand and their knowledge of the addressee. It also
helps to make explicit current information, on the basis of which people infer
what they should say. Such choices and inferences are typically made in terms
of the speaker’s cultural and situational knowledge in semantic and episodic
memory systems relative to the information available to them in the current
situation and whatever goals they may consciously or unconsciously be pursu-
ing in the interaction. In this sense, phasal analyses differ from van Dijk and
Kintsch’s (1983) microstructural analysis of a text base primarily insofar as
they are developed from an inherently multi-functional view of language so that
interactional and organizational features form part of the base and can inform
inferential processes (as can ideational information). Moreover, because what
phasal analysis does is aid the identification of patterns in functional linguistic
selections, it has the potential to show patterns that are not otherwise obvious
but may nevertheless influence the sorts of inferences speakers make.
Neurocognitive disorders can interrupt or make unavailable any of the
resources necessary to process texts: ability to encode, monitor, plan, select
or maintain information online, access semantic or episodic memory, infer
from current and contextual information what might come next, integrate
currently available information with existing models, or figure out the needs
of the addressee. All or any of these can be compromised, as of course can
more obviously linguistic systems, and affect discourse production and
comprehension. Discourse analysis lets us design studies which can help us
investigate such effects. Clinical discourse analysis seeks also to contribute
to understanding them.
6 Study design
6.1 Introduction
The abilities to talk, to carry on a conversation, and to tell stories are central
to us; they are constitutive acts through which we create, embody and perform
our selves. Conversation is ontogenetically prior to narrative, developing in
tandem with language: two and three year olds can carry on simple conversa-
tions: story telling develops around four years of age, co-incident with the
emergence of episodic memory and theory of mind. Because of their centrality
in our social and cognitive lives, narrative and conversation tasks are increas-
ingly used in study designs to investigate linguistic, discoursal and cognitive
patterns. Narrative and conversation tasks provide naturalistic, ecologically
valid data which can be used to identify both positive and negative features in
discourse.
In this chapter, we explore narrative and conversation tasks for three com-
monly investigated areas: linguistic structure, narrative and memory. We out-
line issues in study design based on the models presented in previous chapters.
Linguistic structure is commonly investigated in contexts of language develop-
ment and in speech disorders and is central in the evaluation of speech perform-
ance in educational and clinical settings. We discuss morphology and syntax as
two areas of linguistic structure in interaction with information processing load
as represented in the conversation of speakers with AD and an ASD in Section
6.2. Narrative occurs universally and is important in our cognitive representa-
tion of events (van Dijk and Kintsch 1983; van Dijk 2006). Narrative tasks
are used not only to investigate linguistic structure and discourse, including
narrative skills, but also to evaluate cognitive abilities such as comprehension
and recall. In Section 6.3 we outline classic structural definitions of narra-
tive elements and describe the types of narrative tasks used for research in
clinical settings. We then characterize the spontaneous narrative abilities of
a speaker with an ASD. In Section 6.4 we examine performance of speakers
with Alzheimer’s disease and mild vascular dementia on controlled narrative
recall tasks. In Section 6.5 we discuss the design of research questions and the
corpora required to address them.
93
6.2 Linguistic structure: morphology and syntax
6.2.1 Morphosyntax
For clinical populations, there are two kinds of structural questions researchers
ask about morphosyntax. The first is ‘what kinds of errors and/or omissions
occur in the speech of a particular group or in an individual?’ The second
is ‘what kinds of complexity are there in speakers’ language?’. Omissions
and errors in morphological form are assessed by frequency and type. The
misuse or omission of determiners, pronouns, number, case, gender, tense,
verb auxiliaries, aspect and prepositions are standardly investigated mor-
phological features. Morphological features are used both to track language
development in children and to identify problems for people whose speech
is affected by neural trauma or disease. In cases of trauma or disease, mor-
phological deficits may be associated with damage or dysfunction in par-
ticular regions. However, which areas are implicated depends on the nature
of the morphological deficit so assumptions about regional specificity can-
not be made with confidence without reference to the specific deficit and
diagnosis.
Inability to use morphological features appropriately can have obvious
syntactic consequences such as lack of agreement between subject and predi-
cate (e.g. *They sings), omission of all morphological inflections and func-
tion words (e.g. *John – store – shop for John went to the store to shop) or
pragmatic consequences such as problems of cohesive reference. Argument
structure problems are also possible, either through omission of arguments or
prepositions required by a head (sometimes called ‘incomplete predication’),
inclusion of arguments which are not projected by the relevant head (these are
sometimes called ‘faulty predication’), or inability to use a head element either
at all or appropriately:
(1) Argument omitted: *They put the milk
(2) Preposition omitted: *They put milk __the fridge
(3) Wrong argument: *she ate the bread to Sam
(4) Predicate omitted: *she___ bread
(5) Wrong predicate: *she was bread
6.2.2 Argument structure

Argument structure problems can be a function of not knowing what argu-
ments are required by a head, essentially not knowing (how to use) a word.
Or they can be caused by specific neurocognitive dysfunctions affecting,
singly or in combination, phonological, syntactic and semantic systems. For
instance, while the nature and causes of these differences are not yet clear
Study design 95
(and indeed are hotly debated), a number of studies suggest that agrammatic
and fluent aphasics differ in their predicate–argument structure abilities from
normal speakers and from each other. Agrammatic aphasics appear to have dif-
ficulty with non-canonical dependency relations but are comparatively able to
process argument roles in canonic positions (see Drai and Grodzinsky 2006a;
2006b for meta-analysis of data for agrammatic aphasia). Fluent aphasics on
the other hand are apparently more impaired in the interpretation of argu-
ment roles (Shapiro et al. 1993). (For the range of views in this area, see, for
instance: Bird et al. 2000; Shapiro and Caramazza 2000; Collina et al. 2001;
Webster et al. 2001; Druks 2002 for review; Black and Chiat 2003; Thompson
2003; Lee and Thompson 2004; Nankano and Blumstein 2004.)
Alternatively, argument structure problems may occur as a consequence of
information processing load. We use ‘information processing load’ to refer
to situations in which there is evidence that a speaker is struggling to access,
maintain and/or monitor information online. We evaluate information load in
terms of signs of processing difficulty presented by the speaker: for instance, a
speaker who shows frequent hesitation, long pauses, false starts and incomplete
utterances will be presumed to be experiencing some difficulty in discourse
production. In such contexts, argument structure problems are also likely to
occur as speakers fail to monitor, or cannot access lexical selections, or lack
working memory capacity to complete their utterances. Most speakers will be
familiar with how multi-tasking and fatigue can affect their own and/or oth-
ers’ discourse so that cognitively undemanding tasks can become challenging,
discourse less coherent, and well-known lexis unavailable.
Particular neurocognitive disorders may influence the experience of infor-
mation load, resulting in patterns associated with the dysfunction. In AD for
instance, information processing load is partly associated with verbal working
memory deficits (e.g. Kemper et al. 2001b). This may result in incomplete
sentences, where the speaker begins a sentence and stops part way through
because they cannot retrieve lexis, have forgotten what they meant to say, are
interrupted, or are distracted by new information as in Example (6).
(6) Incomplete sentence:
IV: = would would that be a hope for you uh, that you would be able to uhm, go and do
things like you used to be able to do, around the farm
P: yeah, *there’s maybe a few things that,
IV: mhmm
CG: and be able to go and do things without, me sort of being there to help him, and
show him what to do and what not to do
In *They put the milk __, *she ate the bread to Sam, the speaker produces
sentences as though they were complete, but they lack an argument or include
an inappropriate one. In Example (6) above, the patient who has AD just stops
part way through the utterance and pauses. This may be an effect of work-
ing memory deficit and/or other systems since the structure that he begins is
complex, requiring a dependent clause complement for the that which he does
not supply. Other speakers wait and when he doesn’t complete his turn they
take the floor. If the patient started again and corrected or completed the utter-
ance, it would be treated as a false start as in Example (7):
(7) False start:
It was um, it sounded like, i, it was a new, re- recent one.
False starts appear to signal planning problems, but the repairs indicate that
the speaker is monitoring and so does successfully complete the utterance. Thus,
information load processing problems can arise from multiple causes but do
tend to be overtly marked by features such as hesitation, false start and incom-
plete utterances. The latter can of course affect argument structure. Information
processing load also increases with increases in grammatical complexity (e.g.
Stowe et al. 1995; 1998; Caplan et al. 2001; Michael et al. 2001).
The simplest measures of morphosyntactic complexity and diversity
assess presence or absence of simple or complex units and their frequencies
expressed as proportions of some relevant constituent or text. Thus, one can
look at the proportion of morphologically complex words in a text sample
according to length. Length can be specified by number of words, clauses,
predications, utterances or turns. It is also valuable to record time of speak-
ing. Measures of syntactic complexity are evaluated in the same domains,
and assess the number of words per clause, the number of predications per
clause or sentence, or the number of simple versus complex and/or compound
clauses per sentence. Syntactic complexity evaluations may also address types
of clause constructions and so refer to embedding of clauses functioning
within the structure of other clauses or phrases, their distribution to the left
or right (where left embedding is more complex than right), and dependency
relations in passive, relative, unaccusative and WH constructions which entail
processing arguments or adjuncts in non-canonical positions relative to their
traces. Thus, although the clauses in Examples (8a) and (8b) are predication-
ally similar, (8a) is simpler than (8b) because (8a) consists of two independent
clauses coordinated with the alternative conjunction but whereas in (8b) the
predications in the first independent clause of (8a) are relativized within the
subject noun phrase of the main clause [[Sam [RELATIVE CLAUSE]] eats heaps of
chocolate every day].
(8) a. Max says Sam is supposed to be on a diet but he eats heaps of chocolate
every day
b. Sami, whoi Max says ti is supposed to be on a diet, eats heaps of
chocolate every day
Study design 97
The relative clause requires that Sam and who be co-referential with the
r elative gap ti and that Sam be kept in working memory until the verb of the main
clause (eats) occurs. Both the working memory demands and the co-reference
demands make (8b) more complex and so more challenging to process. While
(8b) is demonstrably more complex than (8a) and increased complexity gener-
ally has the effect of slowing processing speed (Kemper and Herman 2006),
it is worth noting that ‘not all complexity is equal’. Cognitively healthy older
speakers are more likely to produce (8a) than (8b) (Kemper et al. 2001a), but
may be as accurate, if slower, than younger speakers in processing (Saxton
et al. 2001; Grossman 2002). (But see Kemper and Herman (2006) where older
speakers made more efforts because of effects of increased memory load and
increased syntactic complexity.) An AD speaker might have serious difficulties
with (8b) presumably because of the working memory demands, but would
process the active and passive in Examples (9a) and (9b) similarly (Almor et al.
1999). However, for an agrammatic aphasic speaker, (9b) might be challenging
because of its non-canonical presentation (e.g. Shapiro et al. 1993; Thompson
2003), though whether the grammatical structure itself or the need to maintain
grammatical information online for processing is the real source of the prob-
lem is currently unresolved (see Stowe et al. 2005 for discussion).
(9) a. The guests devoured all the food in the first hour
b. All the foodi was devoured ti by the guests in the first hour
Other factors which may increase clause complexity include textual
re-ordering for focus and prominence with resources such as preposing, post-
posing, cleft formation and topicalization which were illustrated in Chapter 4.
These structures increase syntactic complexity; however, in normal speak-
ers they also appear to increase, rather than reduce, processing efficiency for
prominent elements (e.g. Sturt et al. 2004; Sandford et al. 2005).
Lexical density and richness may also affect information processing. Lexical
density in a discourse is evaluated by averaging the number of lexical (as
opposed to form) words per clause in samples of a specified length (Halliday
1987; 1989; Bucks et al. 2000). Lexical richness is evaluated similarly but in
terms of the number of lexical items relative to other words. A Type–Token
Ratio (TTR) is the most commonly used formula for evaluating lexical rich-
ness. TTR is the total vocabulary words V as a ratio of the total number of
words in N. Thus, TTR = V/N, where a higher value for V reflects a richer
(more varied) vocabulary. A limitation of TTR is that it is sensitive to length.
Brunet’s Index is a length insensitive version of TTR. Brunet’s Index W is
calculated as W = NV–0.165. The value for W is the index, lower values are richer
(Brunet 1978; Bucks et al. 2000). Finally, there is the notion of idea density
which evaluates the amount of information in a discourse in terms of the aver-
age number of propositions per ten words in a specified sample (Kintsch and
Keenan 1973; Small et al. 2000; Kemper et al. 2001) and processing capac-
ity (Kemper and Sumner 2001). Syntactic complexity and lexical density are
modelled in Chapter 7.
Morphosyntactic features can be investigated either through discourse tasks,
or through sentence completion tasks which require participants to supply
correct forms or to correct errors. This we illustrate with conversational text
between an interviewer and a young man with an Autism Spectrum Disorder.
The first topic is bowling, the second is a film.
Text 6.1 Bowling

(1) CHI: I’m on a bowling team.
(2) RES: oh I don’t know much about bowling.
(3) RES: I haven’t bowled a lot.
(4) CHI: ((laughs))
(5) CHI: this is about this is about five-pin we do.
(6) CHI: I used # I used to even bowl ten-pin too.
(7) RES: uhhuh?
(8) CHI: I was pretty good at ten-pin.
(9) CHI: I’m also pretty good at five-pin too as well.
(10) RES: good.
(11) RES: how many are on a team in bowling?
(12) CHI: um I’d say about five or five or six I would say.
(13) RES: umhum?
(14) CHI: five or six.
(15) RES: yeah.
(16) CHI: well sometimes if there’s a smaller team there’s four is an exception too as
well.
The young man is very responsive. He chuckles when the interviewer says
she knows little about bowling. He tries to inform her about the topic, answer-
ing her questions with careful detail, and taking turns appropriately. When
he is uncertain, he uses hedges appropriately as in um I’d say about five or
five or six I would say (12). His use of I’d say and about indicate uncertainty
about the number of players, as does his presentation of alternative numbers
(or five or six). However, the way he organizes his text here and later shows
some difficulties in information processing, which he tries to compensate for
by adding redundant information. Here, he repeats the modalization I would
say in final position. Other information processing features are reflected in
common dysfluencies such as false starts (I used, I used (6)), and hesitation
markers (um).
There are also problems with two of his morphosyntactic constructions. In
lines (5) and (16) he completes a predication successfully but then adds predi-
cative content that creates an error as in well sometimes if there’s a smaller
team there’s four is an exception too as well (16). Here there are three possible
Study design 99
interpretations. In one reading, is an exception too as well appears without an

external argument (there is no subject for the verb is). Alternatively, four can be
interpreted as the subject of is in which case we could say the existential clause
is incomplete (has no predicative complement), or that four is an exception too
as well is the predicative complement of the existential clause in which case
we might suppose that a relative clause was intended. There is also redundant
repetition, most obviously in the two occurrences of too as well (9, 16), five or
six (12, 14), and I would say (12). Normally, intonation would disambiguate
a structure like this but this speaker structures information using fewer tone
groups: there’s four is an exception too as well is spoken as a single tone group
(de Villiers et al. 2006).
Text 6.2 Transvestites

(1) CHI: <it’s> [<] about these three men that dress up as females.
(2) CHI: they’re like um.
(3) CHI: what do they call them?
(4) CHI: they call # transvestites.
(5) CHI: that’s what <they’re called> [>].
(6) RES: <uhhuh> [<]?
(7) CHI: sexual transvestite.
In Text 6.2, there is a missing argument (they call # transvestites (4)). Sexual
transvestite (7) lacks number agreement with its apparent antecedent (they in
they are called transvestites). In isolation one might conclude that the speaker
has difficulty with linguistic structures. However, examination of a 1,000
word sample of his conversation suggests that problems of argument structure
and agreement are infrequent. There are seven including the four described
here: one sentence has added predicative content; there is one missing rela-
tive pronoun and one wrong preposition. However, problems associated with
information processing and perhaps lexical retrieval are pervasive. This can be
shown by giving the ratios of the number of words used to construct complete
predications as a proportion of the total number of words. The rest of the words
are processing features such as false starts, repetitions, ums and uhs. Such fea-
tures occur in all spoken discourse relative to the difficulty the speaker experi-
ences in text production (Clark and Wasow 1998). However, the frequency of
features here, 254/1000, approximately 25% of the speech, is unusual. Here, in
examples (10–11), we see the speaker struggling:
(10) there’s it also delivers to to places in Saint Catherine’s as well and also
the um and also places like in um like s like international places like
Germany and England and stuff like that.
(11) and then and then we then we work then we work um something like uh
nine-thirty to nine-thirty to twelve.
In (12), the speaker actually refers to his own word finding (fluency) and
cognitive processing difficulties.
(12) um.
guess to s uh sort sort out um sort out uh mail and stuff.
put them into little um # categories and stuff.
and it’s um # and uh um #.
well # it’s hard to think now.
…
god I forgot all these machine names what they’ve called them
…
I forget what they call them
…
I found that I find those names sort of a bit greek to me sometimes
… uh I’m trying to think now
… I’m trying to think of it.
The features of repetition, false starts and hesitation fillers clearly reflect
processing difficulties which may involve planning, monitoring, retrieval
and/or flexibility. Those morphosyntactic problems that do appear may be an
effect of such processing difficulties and simultaneously reflect the speaker’s
attempts to compensate for them. His efforts to provide accurate and detailed
information are characteristic of ASDs and the challenges he faces are in some
senses constructed by the tension between his attempt to give detail and his
ability to retrieve appropriate lexical items. It is possible, given this, that some
apparent grammatical ‘errors’ are in fact other kinds of constructions serving
different discourse functions. For instance, the lack of agreement cited for Text
6.2 (transvestites…transvestite) may be serving a textual function of ‘recap’,
comparable to the other repetitions such as too as well.
It may of course also be the case that this speaker’s processing difficulties
are simply signs of fatigue or distraction in the interview situation. Collecting
more data from a range of different situations at different times would allow
one to evaluate the extent to which such features are constant in his discourse,
or occur only in specific situations. If it were established that these features are
constant in his discourse, that would be indicative of information processing
load difficulties, the specifics of which could then be investigated.
6.3 Narrative
Narrative as distinct from other generic text types such as dialogue or report
has a number of conventional features. Classic descriptions (e.g. Propp 1928;
Labov and Waletsky 1967; Barthes 1968) identify:
• title (now I’m going to tell about the time …);
Study design 101
Table 6.1. Narrative performance features
Formal elements of Other commonly Other possible

narrative structure considered aspects considerations
Title Themes Lexical density

Setting/orientation Evaluation and causation Narrative voice
Participant(s) Character Syntactic complexity
Events Cohesion and coherence Fore- and backgrounding
Initiating event Reported speech Prosodic organization
Complicating event Episodes
Resolution Gist/recall
Coda Morphosyntactic features
• setting/orientation (time, place, participants);

• events (including an initiating event and at least one complicating event – the
complicating event is surprising in some way and requires a protagonist to act);
• resolution (an action or event which alters the situation created by the com-
plicating event);
• coda (a comment on narrative significance and/or termination).
Most elements may be omitted but the complicating event sequence is defining
of narrative: it is what makes a story a story rather than, say, a simple report.
Narratives are commonly evaluated both in terms of the presence, absence or
accuracy of story elements, and also in terms of themes, evaluation, charac-
terization of attitudes, cohesion and coherence. Narratives may also be used
to assess morphosyntactic features, lexical density, recall, gist formation, fore-
and backgrounding and prosodic organization. Table 6.1 summarizes features
used to evaluate narrative performance.
Narratives may be spontaneous or they may be elicited through a narra-
tive task. A common task is picture description, where pictures without texts
are used to elicit narrative discourse. Picture description as study design is
informative about language organization and comprehension and has the virtue
of controlling topics of speech. However, picture description does not tell you
about the effects of memory loss in the discourse of patients with Alzheimer’s,
nor about interactive conversational skills of people with autism. Moreover, as
a consequence of controlled topic selection, the information picture descrip-
tion can provide about topic management and development is limited in any
group (Duong et al. 2005). Other common tasks include narrative recall tasks
and cued narratives (e.g. what’s the most exciting/frightening thing that ever
happened to you?). Conversational tasks, interviews, semi-structured inter-
views and/or spontaneous conversations may be examined for spontaneous
narratives. (Spontaneous conversations can be acquired by, for instance, giving
Table 6.2. Narrative task scale
Controlled Prompted Spontaneous
No demand on Picture
memory description,
cartoons
puppets, films Cued narrative
(semantic, episodic,
prospective)
Maximum demand Read aloud, Non-elicited narrative occurring
on memory read to self spontaneously in conversation
participants something to do and recording the conversation that occurs while

they are doing it, or by asking about something.)
One could think of narrative tasks on a graded scale with controlled elicita-
tion using story stimuli and spontaneously occurring narratives on opposite
ends. Picture description requires speakers to construct a narrative sequence
from supplied material. It controls for topic, participants and event sequence
but other elements are determined by the speaker. There is no demand on
memory. Films similarly require speakers to encode and recall event sequences
from supplied visual and auditory representation. They control for topic, par-
ticipants and event sequence. Other elements are determined by speakers and
demand on memory is quite high. Cued narrative controls for topic but not
event sequences unless the cue is for semantic events (socially known, public/
historical events). That is, cues can be focused to place demands on episodic,
semantic or prospective memory (see Chapter 2 for definitions). Read aloud or
read-to-self recall tasks may assess comprehension, verbatim and/or gist recall
of a story. They are the most controlled narrative tasks. Table 6.2 summarizes
narrative tasks on two dimensions: memory demands and speaker production
demands.
Evaluation of narrative can include not only accuracy relative to story ele-
ments and language features as described above, but also narrative patterns
associated with particular disorders. For instance, ASD speakers may sponta-
neously offer narratives about subjects of particular interest to them.
Text 6.3: Creature from the Blue Lagoon

This is a story retell of a film, the original title of which was ‘Creature from
the black lagoon’. The speaker has retitled it ‘The Blue Lagoon’, although he
refers to the ‘black lagoon’ as the origin of the monster.
(1) CHI: I made it.
(2) RES: what is it?
Study design 103
(3) CHI: ever seen The Blue Lagoon?

(4) RES: um: I don’t remember if I have.
(5) RES: I don’t think so.
(6) CHI: uh the monster that comes from the black lagoon.
(7) CHI: and he has webbed fingers and webbed toes and webbed things.
(8) CHI: and this is supposed to be his hand.
(9) RES: what’s the story about?
(10) CHI: um: these explorers go out into the rainforest.
(11) CHI: and they find fossils of # of the the # the monster from the black lagoon.
(12) RES: umhum?
(13) CHI: they find clay mo they find skeletons and stuff f from the.
(14) CHI: and they find one that’s living still that was caught in ice.
(15) RES: and then what happens?
(16) CHI: and they uh killed all of them.
(17) CHI: they all get killed.
(18) CHI: and then there’s number two.
(19) CHI: and then they had to go catch er the thing.
(20) CHI: and bring it back to America.
(21) CHI: and then he tries to kidnap a beautiful woman.
(22) RES: does it succeed?
(23) CHI: no.
(24) CHI: he gets shot.
(25) CHI: and he falls into a a lake.
(26) RES: uh.
(27) CHI: and then uh # in number three they go down there.
(28) CHI: and uh they accidentally torched him.
(29) CHI: and uh he and then they wrapped him up like a mummy.
(30) CHI: and uh done surgery and stuff.
(31) CHI: and uh he transformed into a kinda like a man.
(32) RES: hm.
(33) CHI: but still he was an amphibian.
(34) RES: umhum.
(35) CHI: they put one of those things that people get put in their throats so then they
can breath.
(36) RES: oh right.
(37) RES: umhum.
(38) RES: they usually talk funny after that don’t they?
(39) CHI: they don’t he doesn’t talk at all.
(40) CHI: ((laughs))
(41) RES: no?
(42) CHI: and then they put clothes on him.
(43) RES: umhum?
(44) CHI: and then wa runs away.
(45) CHI: and breathes in the swampy area.
(46) RES: do you like movies like that?
(47) CHI: umhum.
(48) RES: with monsters and things?
(49) CHI: that’s right.
In Text 6.3, Creature from the Blue Lagoon a young man with an ASD is
asked about a film and responds with real enthusiasm, giving plot summaries
for the original (Creature from the Black Lagoon) and two sequels. He initiates
the topic as a response to the researcher’s question about a model he has made
of the creature’s hand. When asked what the story is about, he supplies all three
plot summaries, with prompting only for resolving events. For the first one, he
gives the setting and participants, and the event sequence including the compli-
cating event (they find one that’s living still that was caught in ice) and, when
asked, supplies the resolving event (they uh killed all of them). He volunteers a
title in and then there’s number two (18) and gives the initiating event (go catch
thing and bring it back to America), complicating event (thing tries to kidnap
a beautiful woman) and, with a prompt (does it succeed?), he offers the resolu-
tion (no he gets shot and he falls into a lake). Similarly, in the third story he
offers a title (and then in number three they go down there), supplies initiating
(go there, accidentally torch him) and complicating events – those surrounding
the transformation of the creature through surgery into something kind of like
a man, but still … an amphibian (31–33), who when clothed runs away back
to the swamp. In the third story, the resolution is not stated. The only element
of evaluation of the film, apart from his very evident interest, occurs when the
researcher asks:
(13) RES: do you like movies like that?
CHI: umhum.
RES: with monsters and things?
CHI: that’s right.
What we see here then is a well-formed set of narratives, with conventional

narrative elements, spontaneously produced with minimal prompting. The
absence of evaluative comments and resolving events is perhaps part of a dis-
order specific pattern.
For instance, there is a lack of attention to the movie’s emotional elements
in the young man’s narratives. He never mentions that the creature loves the
woman in the film. He also does not say that the creature dies at the end, an
implied outcome that is a key dramatic element of the movie. The creature
is unable to survive in the water because, following his operation, his lungs
rapidly evolved and he lost his gills. Dramatic music and contextual clues (pro-
ceeding directly to the ocean after his escape as he has done repeatedly before;
standing longingly by the water’s edge; a slow but steady march into the water)
inform the audience of the unfortunate creature’s fate. The speaker does not
appear to make use of these clues of context to infer the story’s outcome or
the mental state of the creature at the end of the film. Failure to use contex-
tual clues to interpret and predict behaviour is a characteristic feature of ASDs
(Baron-Cohen 1995; Frith 2003).
Study design 105
6.4 Narrative tasks designed to investigate memory

In contrast with spontaneously produced narratives occurring in conversation
or interview, some narrative tasks may be designed to make specific demands
on memory. Participants are read or asked to read short narratives and to
remember them as they will be asked to repeat them verbatim. Repetitions can
occur immediately or after a delay (the latter task is typically referred to as a
delayed recall task). The sample texts below are examples of delayed recall
of a single narrative. The recall task in the study was the WMS III, logical
memory (Weschsler 1997). The stories have been modified here to protect the
source narrative. The first (Text 6.4) is produced by a participant with a diagno-
sis of subcortical ischemic vascular disease (SIVD), and mild vascular demen-
tia (VaD). Texts 6.5 and 6.6 are produced by speakers with clinical diagnoses
of mild AD. The speaker of the fourth text (Text 6.7) is an older adult with
no cognitive impairment. These texts are from a small study which evaluated
gist, verbatim recall, confabulation and modalization on narrative recall tasks
for participants with diagnoses of mild AD, mild vascular dementia (VaD),
and older adults with no cognitive impairment (NCI) (Klages 2006; Asp et al.
unpublished1). Verbatim recall and confabulation distinguished the participants
with mild VaD from participants with mild AD. AD participants confabulated
more and had poorer verbatim recall than VaD participants on two out of three
repeats. VaD and AD participants did not differ significantly from each other
on any neuropsychological or cognitive measure.
Text 6.4 Delayed recall (SIVD)

(1) P: // the what//
(2) IV: the second story //and start at the beginning//
(3) P: // the second story// what was it about//
(4) IV: The story was about a road report
(5) P: //oh yes// this chap was uh getting to ready uh to go out // around breakfast time//
and he listened to the uh the uh news the road report// there had been a big accident
and whatever whatever// so he changed his plans// and he put away the keys // and
he went into the yard // and he uh worked // and worked in the garden//
In Text 6.4 the speaker provides a very accurate gist based on the narra-
tive event structure, including orientation (this chap was uh getting to ready
uh to go out around breakfast time), initiating (and he listened to the uh the
uh news the road report), complicating (there had been a big accident and
whatever whatever) and resolving events (so he changed his plans and he put
away the keys and he went into the yard // and he uh worked // and worked in
the garden). However, there is limited detail supplied about participants and
circumstances. For instance, the speaker uses indefinite and general references
as in the chap instead of a proper name, around breakfast time instead of the
actual time and whatever whatever instead of the accident details. The absence
of lexical specificity is characteristic of the delayed recall performance of the
SIVD participants. Nevertheless, 29% of the lexical items were repeated ver-
batim from the source narrative and his excellent gist contrasts markedly with
that produced by the speaker of Text 6.5.
Text 6.5 Delayed recall (AD)

(1) P: // what was the second story//.
(2) IV: the story was about a road report
(3) P: // oh yes// can’t remember it dear//
(4) IV: any little detail?
(5) P: //aah //had there been an accident//
(6) IV: umhum
(7) P: // yeah// a crash//
(8) IV: ye
(9) P: // (I) don’t know anything else// (accident) (15)//1 can’t think//
(10) IV: ok
(11) P: //oh you poor thing//do you have many like me//
(12) IV: oh, you’re doing fine
(13) P: ((laughs))
(14) IV: you’re doing your best. [Instructions]
(15) P: oh, I see. that’s good
The only event that the speaker recalls is the complicating event had been an
accident and this is a question. Accident and crash are also the only words
repeated verbatim. The speaker is very aware of her recall difficulties, saying
that she can’t remember, and doesn’t know anything else. She also comments
on her cognitive state (I can’t think) and appears to believe her performance is
cause for commiseration with the interviewer (oh you poor dear. Do you have
many like me?). So while her recall is very limited, her awareness of deficit and
perception of it as burdensome is acute.
The third text (Text 6.6) is also produced by a speaker with mild AD. He
is similar to the second speaker in that he comments directly on his perform-
ance (I can’t word it word for word) and his cognitive ability (there’s no way
the old brain’ll). He differs in that his delayed recall is a significantly more
detailed gist (similar to the one produced in Text 6.4), but much of the detail
is confabulated.
Text 6.6 Delayed recall (AD)

(1) P: // the second story was the fella who got up at dawn to go on a trip// fishing// and
the road report came on// and gave him a bad accident report// an that there’d been
a huge pile up, and roads closed and traffic all rerouted// and that was the advisory
to people not to go out// and he did// and it was a very disastrous day// that’s what
I got out of it//(laughter)//1 but that’s about the gist of the whole story//
Study design 107
(2) IV: (laughter) yeah

(3) P: // I can’t word it word for word//
(4) IV: ok
(5) P: // there’s no way the brain’ll ()//
(6) IV: Not a problem. Just what you can remember.
Story elements are judged to be confabulated when processes, participants
or circumstances (or entire predications) which are intended to represent story
elements are reported but are not present, nor synonymous, with the control
narrative. This speaker confabulates entire predications representative of story
elements. Specifically, the setting (the second story was the fella who got up at
dawn to go on a trip fishing) and an added element of the complicating event
(that was the advisory to people not to go out) as well as the resolving event
(and he did) are all confabulated. The speaker also offers a narrative coda (and
that was a very disastrous day for him) which, though potentially coherent
with his version, has no relationship to the control narrative. In the control
narrative there is no mention of a fishing trip or an advisory for people not to
go out, although this is a reasonable inference to draw from a major accident
report. The setting is nine, not dawn, and the participant stayed at home. The
only two elements that are accurately represented from the control narrative
are the initiating (getting ready to go out) and the complicating event (the bad
accident report). 17% of lexical items are repeated verbatim from the source
narrative.
There are three types of confabulation recognized (Kern et al. 1992; Dalla
Barba et al. 1999). The first is error of recall as in mistakenly asserting that
one has heard something from an individual when in fact it was on a news
broadcast. The second type is sometimes referred to as ‘fantastic’. Typically,
it involves statements about personal past which are entirely imaginary (and
often grandiose). The third type of confabulation involves replacing one set
of experiences or reports with others (Berlyne 1972; Kopelman 1987). The
examples in Text 6.6 appear to be largely of this type. It is also worth noting
that the speaker assesses his performance as personal that’s what I got out of
it and appears unaware of the confabulated elements but that’s about the gist
of the whole story.
The amount of confabulation can be straightforwardly modelled in narrative
recall tasks as counts of words referring to processes, participants or circum-
stances which are intended to represent story elements but are not present in
the control narrative, nor synonymous with the control narrative. All words are
counted when the entire predication is confabulatory: Uncle Thomas got up at
dawn to go fishing as a representation of a story in which no such event or par-
ticipant occurs counts as 9 words. He heard it on the television in a story where
the participant has heard something on the radio counts as 1 (confabulation of
source television for radio).
Verbatim recall Narrative recall Match to control (percentage

(read or listen to a story) of words repeated verbatim)
Gist Narrative recall 1) Percentage of story elements

(read or listen to a story, accurately repeated
watch a film etc.) 2) Predicational analysis of
percentage of participants, events
and circumstances accurately
repeated
Confabulation Narrative recall (as above) Match to control (word counts

and/or gist elements)
Report of activities Check with caregiver
Figure 6.1 Evaluating confabulation, gist and verbatim recall.
The last delayed recall sample (Text 6.7) is produced by an older adult with
no cognitive impairment. The text differs markedly from all others not only in
the accuracy of gist recall but in the amount of participant and circumstance
information incorporated verbatim from the control narrative.
Text 6.7 Delayed recall (No cognitive impairment)

(1) P: //It was nine o’clock on a Saturday morning// and Gino Costa was, getting ready
to go to market// and listening to the radio// and the program was interrupted by a
road report //to say there had been a multi-vehicle accident on the main highway //
the highway was closed to outbound traffic, and secondary routes were at a crawl
// so Gino decided he wouldn’t go to the market // he put away his keys // and he
went out to work in the garden//
(2) IV: um terrific
This speaker reproduces a verbatim account, including all story elements
and high lexical specificity (52% of lexical items from the source narrative). In
the first two repeats, the speaker even reproduces the intonation patterns occur-
ring in the read-aloud original story. We might call repetition of intonation
‘prosodic verbatim’. None of the other speakers represented does this.
What we have done here is illustrate different kinds of narratives, taking
examples from opposite poles of the Narrative Tasks Scale. We looked at a
speaker who has an ASD who produced well-formed narratives spontaneously,
highlighting the role of clinical discourse analysis in looking at abilities as well
as deficits. We then showed how performance can vary on a single narrative
delayed recall task. The speaker with no cognitive impairment differs from
the speakers with AD/VaD both in accurate production of story elements and
verbatim lexical recall. This is unsurprising given the diagnoses. AD and VaD
Study design 109
participants differ from each other in that the VaD speaker produces a very
good gist, while gist is poor in the AD participants, verbatim lexical recall
minimal and confabulation high. Figure 6.1 summarizes the features investi-
gated, the sort of tasks that may be used to investigate them, and possible ways
of evaluating the results.
6.5 Designing research questions

Research questions can be narrowly targeted or open ended and clinical dis-
course analysis may play a role in either case. When addressing a clinical
group whose discourse patterns are not well established or described in the
existing literature, a first step would be to develop full characterizations of
a representative sample of texts produced by speakers diagnosed with the
disorder in question. This step may seem extravagant insofar as full charac-
terizations take time. However, the initial full characterization of sample texts
functions rather like taking a medical history, and can inform the process of
developing research questions and study designs that focus on particular diag-
nostic groups. In Table 6.3 we give four sample research questions and possi-
ble study designs, and indicate some positive and negative features associated
with each design.
These study designs illustrate that there are different ways to address spe-
cific research questions and that depending on the nature of the pathology and
the research questions being asked, study designs will be more or less struc-
tured. Semi-structured tasks are less natural, but can be thought of in relation to
context of situation and functional varieties, determining what is relevant given
the nature of the disorder. The choice of register and context of situation should
be informed by the nature of the pathology.
Decisions also must be made about whether or not a control group is needed
and if so what kind(s) of controls are most suitable. A variety of statisti-
cal methods (e.g. regression, analysis of variance, chi-square) may be used.
Bridging the gap between qualitative approaches to discourse analysis and
research which uses statistical methods may present an initial hurdle for some
humanities researchers. Statistical methods books in the health sciences may
be consulted, but a team-based approach or working with statistical consultants
can also reduce the divide.
6.6 Corpora design

Any discourse study requires the development of some sort of corpus. In its
simplest sense, a corpus is a collection of language samples. Corpus-based
studies of discourse skills present comparable sampling issues of power as
occur in epidemiological studies. Thus, an issue that clinical discourse analysts
Table 6.3. Four research questions with pros and cons of study designs
Approaches to
Research questions study design Pros Cons
1. Is pedantic Unstructured or – quite naturalistic, – large number of

speaking semi-structured spontaneous participants needed
characteristic conversation – ecological validity – hard to measure
of all with inter-rater – can use measures degree replicably
subgroups reliability on rating (e.g. lexical density) – hard to compare
within the pedantic speech in for other questions patterns with other
autism different subgroups – more detail disorders
spectrum? and/or identify – costly for coding
specific linguistic and can’t be used by
characteristics of clinicians as a rating
pedantic speech and tool
code in different
subgroups
Control group is – measures may have
optional since the known neurocognitive
question addresses bases
variation within the
autism spectrum
2. Do people Elicit favourite – may be informative – topic not controlled
with ASDs topics in interview for developing – heterogeneous data
express (semi-structured or interventions for – requires
attitude and open question/with individuals sophisticated coding
evaluation family input) and analysis
comparatively or
less than Elicit narratives – picture description – less informative
a control using picture book is a well-established about recurring
group? stimulus. Look for elicitation method individual behaviour
presence or absence – controlled for topic – less naturalistic
of evaluation of – can look at particulars
narrative in discourse of evaluation without
+ control group distraction
3. Do low lexical 1000 word samples – small corpus – data requires
density and from variety of – naturalistic sophisticated coding
syntactic discourse types – some results may be and analysis
complexity Code for patterns scalable
scores correlate + control group
with high/low
modalization
values in AD?
4. Is confabulation Narrative recall – controlled
associated with – simple scoring
disease phase
in AD?
or
Study design 111
Table 6.3. (cont.)
Approaches to
Research questions study design Pros Cons
Episodic memory – may target personal – needs a greater

task (e.g. ‘what state/experience more number of
did you do last directly participants
weekend’) – may be informative – heterogeneity of
for developing data
interventions for – sophisticated
individuals coding and
analysis
will need to be concerned with is sampling size, both for the number of partici-
pants and the amount of discourse produced by each participant, relative to the
research questions being asked.
The number of participants needed varies, ranging from a single participant
case study or small group case study to large groups. Limitations imposed
by time demands of data transcription and analyses can be issues if the aim
is full characterization. Coding and analysis, even using good programmes,
take time. In the case of smaller case and group studies, it is simplest if those
small groups are clinically both very well defined and very homogeneous since
it then becomes possible to see whether participants with similar diagnoses
and demographic profiles exhibit the behaviour. (This of course, though, begs
questions as to whether other participants with other profiles behave similarly.
It also presents the rare case from the clinician’s perspective, where heteroge-
neity is the norm.)
The amount of text needed can also vary depending on research questions. A
standard thousand-word sample may be adequate to provide information about
grammatical structures for speakers, but is unlikely to provide information
about specific highly salient features of discourse which may be characteristic
but comparatively rare. For example, it is widely recognized that people with
autism are known to be literal, including having difficulty in both producing
and understanding metaphor (Frith 2003). But the frequency with which such
cases arise is not sufficient to ensure an occurrence of difficulty with metaphor
or other figurative language in any given task. The rare case is only likely to be
observable in corpora when there are large amounts of data. Thus, a thousand-
word sample is suitable for observing some language features. Longer sam-
ples may be necessary for discourse features and spontaneous performance. If
specific genres are addressed (e.g. narrative), there is a possibility for shorter
sampling. Other important considerations in corpora design are whether to
Table 6.4. Research design and text type
Type of analysis Text type Considerations
Cohesion 100 sentences from a Requires continuous prose – tasks such as

analysis speaker (McKenna and prompted narrative, story retell and picture
Oh 2005) description could be used
Grammatical 1,000 word sample (Bucks Prompted narratives or spontaneous
structures et al. 2000) conversation can be used for continuous
and/or fluency sample
15 independent clauses The advantage of this type of sampling
from 2 different sections is that a speaker’s discourse can be
of an individual’s text, or characterized in different registers.
30 clauses where texts are More clauses will be necessary for some
collected from multiple questions.
settings (Rochester and
Martin 1979)
Type token ratios 100 words from at least Widely used as measures of lexical density,
3 different parts of a although requires attention to contexts of
speaker’s discourse, or use for representative sampling, and has
350 word samples (Hess been shown to be less stable and sensitive
et al. 1986), or 250 words than other measures. (See e.g. Watkins et al.
segmented into 10 samples 1995; Owen and Leonard 2002; Chipere
(Bradac 1988) et al. 2004 for discussion and alternatives.)
Metaphor, Task specific samples – Allows investigation of particular areas of
homograph (Happé 1997) discourse comprehension and production,
or other including rare phenomena
comprehension – Difficulty can be an issue relative to the
disorder being investigated
– Samples from multiple language varieties
may be required where evaluation
(scoring) depends on pronunciation
Discourse Longer samples in multiple Registerial variety required for fuller
features and settings (Fine 2006) characterization and generalizability of
spontaneous findings
speech Specified parameters for Prompted, structured, semi-structured and
full characterization spontaneous spoken and written discourse in
a range of generic situation types:
– with family
– with peers
– with schoolmates
– in jobs
– in the clinic
– in typical situations for individuals
– narratives
– conversations
Cross-sectional studies are useful for
developing a characterization and for
specific questions. Looking at change
over time requires longitudinal studies.
Study design 113
look at different time periods (to investigate stability and change) and whether
to collect data from multiple contexts for full characterization and variation.
It is usually possible to design corpora for specific questions. In Table 6.4
we suggest some sampling considerations in relation to areas of inquiry and
text types. There are corpus linguistic techniques and technologies, includ-
ing corpus collection software (for audio and video), standard concordancing
and search tools, parsers, taggers, tools for data extraction and phonetic and
prosodic analyses and speech-to-text software for transcription. These make
the development of disorder specific corpora more than feasible. In addition to
specialized corpora developed to investigate particular questions or disorders,
it is also invaluable to have access to corpora which are broadly representative
of the speech community at large. Such corpora can inform investigators about
normative patterns and be resources for checking hunches both about what is
typical and what may be rare for the speech community.
7 Differential diagnosis and monitoring
7.1 Introduction
We noted in Chapter 1 that experts have often internalized patterns of discourse
behaviour characteristic of neurological disorders and their phases which
inform their clinical judgements. However, these internalized patterns may not
be explicitly recognized and consequently appear as ‘intuitive’ responses. Just
as people can often recognize regional or social dialect variation without being
able to say precisely what it is they recognize, so the patterns informing clini-
cians’ impressions may not be explicit. Descriptions of discourse can make
experts’ tacit knowledge about discourse patterns associated with diagnostic
groups explicit, in the same way that descriptions of dialect variation can aid
dialect recognition. Such descriptions can also be used to model characteris-
tic patterns for healthcare workers and families who may not have experts’
breadth of experience.
Beyond explicitness, one role of clinical discourse analysis is to add tools
to existing diagnostic resources where diagnosis is still a clinical decision.
Another is to provide characterizations for diagnostic categories which are
under-investigated. These may help with diagnostic clarification and planning
for treatment. A third role for description of discourse is to track change over
time both intra-individually and for group applications. There is potential for
monitoring developmental and degenerative processes and tracking responses
to treatments and interventions. There is also a fourth role which is to improve
understanding of relationships between everyday discourse behaviours and
neurocognitive function.
In relation to these roles, in this chapter we model study designs which address
discourse correlates of diagnoses and monitoring. Except where primary signs
and symptoms of neurocognitive dysfunction are linguistic, descriptions of dis-
course patterns in these contexts are neither abundant nor widely available. There
is thus an a priori value in the descriptions as descriptions. They may also be
used to develop rating scales for assessing categorical and continuous variations
and, in conjunction with other assessments, to improve understanding of the
neurophysiology of disorders and the effects of treatment and intervention.
114
Differential diagnosis and monitoring 115
7.2 Pedantic speech in ASDs

Pedantic speech is a recognized characteristic of ASDs (Ghaziuddin and
Gerstein 1996; de Villiers 2006). It has been described as an area of difficulty
in ASDs because of its pragmatic and functional effect on social communi-
cation. Descriptions highlighting attributes of pedantic speech include that it
may seem mechanical, formal or more technical and detailed than the context
demands (Baltaxe 1977), that it resembles written speech (Asperger 1944), and
may be more literal and include redundancies which are not necessarily cohe-
sive (de Villiers et al. 2007). Such characterizations are based on behavioural
observation but do not articulate the specific discourse features that are signs of
pedantic patterning. One unanswered question is the extent to which pedantic
speaking characterizes different subgroups in the spectrum. Having a compre-
hensive characterization of pedantic speech may help clarify this question and
can lead to development of clinical diagnostic practices and research tools.
Here we operationalize the patterns for pedantic speech using a case study
approach. We look at lexical and syntactic patterning and message organization
in a conversation between a young man with Asperger syndrome, Jay, and an
interviewer. We then present a novel battery for analysis of pedantic speech.
The conversation consists of three primary phases. The opening is informal
chat. Then later in the conversation, Jay uses pedantic speech to talk about a
meal (phase 2) and then about the weather (phase 3), a favourite topic. Text 7.1
is an excerpt from the weather phase:
Text 7.1 Weather

(1) CHI: um: what month was this?
(2) RES: that was in # July.
(3) CHI: that might have been the uh time of the severe thunderstorm outbreak.
(4) RES: I think so.
(5) CHI: with the # six tornadoes.
(6) RES: umhum.
(7) RES: I think it was.
(8) CHI: cause they were saying that was the worst severe thunderstorm outbreak to
ever occur in cottage country.
(9) RES: uhhuh?
(10) RES: <I think there> [>].
(11) CHI: <there were> [<] six tornadoes.
(12) RES: umhum?
(13) RES: I think there was just too much hot weather at once eh?
(14) CHI: yeah.
(15) RES: and that caused it.
(16) CHI: yeah.
(17) CHI: when you get temperatures of ninety-five the # the air sometimes rises itself.
(18) CHI: and then # causes a storm without a cold front.
(19) RES: umhum?

(20) CHI: so you can get severe storms either from # hot air rising up so high that it cools
off or # a hot air mass actually colliding with a different cold air mass.
(21) RES: umhum?
(22) CHI: and I’m sure it must o been a cold front if the storms were that bad # with
tornadoes.
For the three primary phases in the text, lexical density and syntactic com-
plexity are compared. The analyses are based on a transcript which has been
cleaned for false starts and hesitations, but which is otherwise a verbatim
record of an audio-recorded interview. (Note that cleaning transcripts in this
way is not essential, or even desirable insofar as leaving such features in means
that scores reflect actual performance (Clark and Fox Tree 2002; Sampson
2003). We cleaned transcripts here in order to model standard practice.) Two
standard measures of syntactic complexity are used: mean length of utterance
(MLU-w) defined as the average number of words per independent clause
(Thordardottir and Weismer 1998; van Dijk and van Geert 2005), and clause
complexity defined here as the ratio of dependent clauses (all types) to inde-
pendent clauses. The third measure, lexical density, (LD), is calculated here as
the number of open class lexical items (as opposed to closed class ‘form words’
such as ‘a’ and ‘up’) occurring per independent clause (LD = [lex/Cl]). Lexical
density measures give a rough guide to how much information is packed into a
clause or text. In general, the greater the lexical density, the more information
per clause.1 Lexically dense clauses demand more information processing and
consequently are more common to writing than speech. Similarly, more com-
plex clause patterning is characteristic of formal written (rather than spoken)
English because, again, it may require greater planning on the part of a speaker
and places higher demands on listeners for processing (van Dijk and Kintsch
1983; Sampson 2001; 2003; and cf. Halliday 1987; 1989).
The overall pattern of these three measures, indicated in Table 7.1, suggests
that the speaker’s conversation has grammatical and organizational properties
which make it more similar to written or ‘expert’ prose than to ordinary adult
casual conversation. For example, Sampson (2003) gives length and complex-
ity means for an 80,000 word speech sample and a 63,500 word published
writing sample from the British National Corpus (BNC) as a way of measuring
comparative ‘wordiness’. He calculates length somewhat differently (averag-
ing the number of words over each immediate constituent in a construction that
dominates more than one word), but the results appear comparable to other
measures. Mean length for the speech sample is 4.62 and for the published writ-
ing sample it is 9.45. These means are roughly parallel to the contrast between
the first conversational phase of Jay’s discourse in which the participants are
discussing recent personal events (shopping, a visit to the doctor and so on)
where his average independent clause length is five words, and subsequent
Table 7.1. MLU-w, lexical density and syntactic complexity in the weather text
Words per Clause complexity

independent (N clause/ Lexical density
Words Lexical words Sentences Clauses Independent clause N independent (lexical words
(N) (N) (N) (N) (N) (MLU-w) clause) per IC)
Phase 1 125 42 18 27 25 5 1.08 1.68

Casual (0.34)
Phase 2 106 63 8 13 11 10.47 1.18 5.73
Recipe (0.59)
Phase 3 1000 442 92 182 100 10 1.82 4.42
Weather (0.44)
Continuous 1231 547 118 222 136 9.05 1.63 4.02
(0.44)
phases where he moves into a pedantic speaking style and his average clause
length doubles to ten. Similarly, in the same corpus Sampson gives clause com-
plexity means (evaluated in terms of depth – the extent of clause embedding
for words in subordinate clauses) for published (adult) writing in the same
corpus as 1.857 and for speech as 1.365. The pattern seen in this young man’s
three phases parallels this, although the syntactic complexity measure cannot
be directly compared. Jay’s weather phase is much more complex (1.82) than
his casual conversation phase (1.08).
Lexical density scores follow the same pattern. In the conversational phase,
Jay’s lexical density score (1.68) is quite close to the sample presented in
Halliday (1987: 329) as characteristic of adult spontaneous speech (1.8). And
this contrasts markedly with the second and third phases. In the second, when
asked about a meal by the interviewer, he actually supplies the whole recipe for
the meal he plans to prepare. Here, the lexical density increases by more than
two thirds (5.73), again following written language patterns. In the third phase,
when he shifts to a favourite topic, there is a slight decrease in lexical density
(4.42), but this is still very much in the ‘written’ range (Halliday 1987), and is
offset by increased syntactic complexity.
The pattern exhibited highlights two points: first, the sorts of differences
suggested by descriptors such as ‘wordy’ and ‘formal’ appear to be associated
with specific properties in the speaker’s discourse. Longer sentences and com-
paratively complex syntax can translate into ‘wordiness’, and greater lexical
density does sound more ‘formal’ and like written prose because (published)
writing generally is lexically more dense than speech. Second, the analysis of
phases shows that the speaker is not limited to the use of pedantic speech. He is
quite capable of producing the simpler, less lexically dense syntactic construc-
tions more typical of adult conversations, but moves into ‘pedantic speech’
when possible.
7.2.1 Lexical repetition and collocation

Patterns of lexical repetition and collocation bind text and help to create cohe-
sion and coherence. However, frequent repeated lexis or marked textual or
collocational patterns can create redundancy. Jay’s discourse is cohesive and
coherent but may appear verbose. A prominent feature of the third phase of this
discourse is the lexical repetition of items relating to weather. In Table 7.2, the
highlighted lexical words referring to weather occur and are often repeated.
For all repeated items, the number of occurrences is included.
Collocations of verbs from the register of weather report are also promi-
nent, contributing to an impression of ‘expert discourse’. The verbs themselves
occur in a range of registers, with potentially different argument structures and
associated meanings. For instance, in they’re calling for more beer, calling
Table 7.2. Lexical items and repetition counts in the weather text
18 occurrences 13 occurrences 8 occurrences 7 occurrences 6 occurrences

thunder lightning severe tornado clouds (clouds 3
(thunder 7 (severe 7 (tornadoes 5 cloud 2 cloudy 1)
thunderstorm 7 severest 1) tornado 2)
thunderstorms 3
thunders 1)
5 occurrences 4 occurrences 3 occurrences 2 occurrences 1 occurrence
wind (wind 4 hot high fall strikes cold front downpour flashing, cracks,
winds 1) rain (rains mass conditions growing,
storm (storms 4 rainstorm rain bolt (bolts 2 temperatures gusts, loud,
storm’s 1) rainy) bolt 1) hurricane sounds, gear,
power actually shower threaten activity, humid,
(shower 3 (threatening 1 forecasting,
showers 1) threatened 1) breaks,
brew (brewed 1 calm, bang,
brew 1) continuous, line,
rise (rising 1 strokes, chain,
rises 1) ground
cool (cools 1
cool 1)
knock (knocked 1
knocking 1)
streak (streak 1
streaks 1)
takes three arguments and is synonymous with ask. However, given the collo-
cations here, they are stereotypically associated with weather reports where to
call for x takes only two arguments and is synonymous with predict. Collocates
of ‘weather’ verbs are in italics in the list below.
• they’re still calling for showers today and tonight maybe.
• and if conditions are right that could trigger off a thunderstorm later.
• we’re getting back into fall temperatures.
• I’m surprised though today with that big cloud mass that formed it didn’t
fully develop into a thunderstorm.
• <the> storm didn’t fully develop.
• and I watched it as it brewed.
• <and> then it blew over.
• well today they were forecasting cloudy and rainy.
• cause they were saying that was the worst severe thunderstorm outbreak to
ever occur in cottage country.
• that must o been scary when the lightning knocked the power out.
• so it wasn’t like a continuous storm that never ceased.
• so you can get severe storms either from # hot air rising up so high that
it cools off or # a hot air mass actually colliding with a different cold air
mass.
• when you get temperatures of ninety five the the air sometimes rises itself
and then causes a storm without a cold front.
The three most commonly occurring lexical bases are thunder, lightning and
severe. Below are listed three sets of sentences in which they occur. Many of
the sentences are found in more than one list since they contain the lexical item
from more than one set (e.g. ‘thunder’ and ‘severe’). In most cases, the sen-
tences listed are not adjacent to each other, but the lexical features of repeated
items and their collocates combine to connect the discourse, making it very
cohesive and, at the same time, giving the speech a formal, and in some cases
repetitive quality.
Collocations of the three most frequent lexical items

Thunder
today it threatened to uh thunderstorm.
there was # a little bit of distant thunder I think I heard in the
distance.
I could tell from the clouds that it was a thunderstorm and not just
a shower.
and if conditions are right that could trigger off a thunderstorm
later.
I’m surprised though today with that big cloud mass that formed it
didn’t fully develop into a thunderstorm.
I’ll tell you that uh # I think the hot weather and the severest thunder-
storm activity is over now for this year.
I don’t think we’re gonna get any more thunderstorms of the severe
nature that we got this summer.
that might have been the uh time of the severe thunderstorm outbreak
with the six tornadoes.
cause they were saying that was the worst severe thunderstorm
outbreak to ever occur in cottage country.
there was severe lightning and thunder?
and you hear this big bang of thunder after the lights come on
again.
it’s amazing that a line of thunderstorms can come through in a half
hour.
was it right after was the thunder right after the lightning?
so how many miles was this one away from the lightning the
thunder?
I woke up when I uh heard wind and a tiny bit of thunder and heavy
rains.
no thunder in it?
hearing the distant thunders growing tends to relax me somehow.
in the thunderstorms I’ve seen usually we have gusts around maybe
eighty kilometres an hour.
Lightning
was there any severe lightning in the storm you were at at your
cottage?
<there> was severe lightning <and> thunder?
what’s chain lightning?
some people call them streaks of lightning.
<was> this during lightning strikes?
that must o been scary when the lightning knocked the power out.
the lightning wasn’t that close I guess.
when your lights went off uh how far away was the lightning?
was it right after was the thunder right after the lightning?
<so> how many miles was this one away from the lightning the
<thunder>?
several areas of Niagara were without power after several lightning
strikes.
there were several lightning strikes.
and there was lightning flashing every two seconds.
Severe
was there any severe lightning in the storm you were at at your
cottage?
<there> was severe lightning <and> thunder?
during a severe morning electrical storm.
I’ll tell you that uh # I think the hot weather and the severest thunder-
storm activity is over now for this year.
I don’t think we’re gonna get any more thunderstorms of the severe
nature that we got this summer.
that might have been the uh time of the severe thunderstorm outbreak
with the six tornadoes.
cause they were saying that was the worst severe thunderstorm
outbreak to ever occur in cottage country.
so you can get severe storms either from # hot air rising up so high
that it cools off or # a hot air mass actually colliding with a different
cold air mass.
There is register-specific lexis and a high frequency of co-occurrence of

particular lexical items such as severe lightning, severe thunderstorm, severe
storm(s). Selections from the grammar in other areas such as quantification are
also markedly consistent. Jay uses several three times (e.g. several lightning
strikes; several areas of Niagara) and repeatedly questions the interviewer
about the details of a storm she mentioned as in was there any severe light-
ning…?; how far away was the lightning?; how many miles was this one away
from the lightning…?. An examination of the internal structure of noun phrases
(NPs) provides further information about the interaction of high frequency
items. In they were saying that was the worst severe thunderstorm outbreak,
severe thunderstorm is treated as a compound (with severe categorizing storm
type in weather report register), since worst and severe would not otherwise
occur together in this kind of construction.
7.2.2 Message organization

It has been suggested that in ASDs there are problems recognizing relevance
in context (Happé 1993; Frith 2003). This can also be reflected in how rel-
evance is marked in discourse (de Villiers and Szatmari 2004). Marked use of
relevant and irrelevant information can be a contributing factor to a pedantic
quality in speech, since the inclusion of known information can make speech
appear detailed or verbose (Baltaxe 1977; de Villiers and Szatmari 2004). So,
in pedantic speech it is likely that examination of linguistic backgrounding and
foregrounding of information will be revealing since the linguistic and discour-
sal resources for organizing information as new and relevant or as known are
aspects of how relevance is constructed in talk and text. In addition to the kinds
of lexical pattern already discussed, this involves consideration of reference
type, syntactic constructions and prosodic features that background or fore-
ground relevant information. Here we present a battery for message organiza-
tion. It consists of resources for the description of a whole system of patterns
associated with information structuring.
Reference Participants are represented by noun phrases. Typically,

when a speaker introduces information about participants that they want to
present as new, indefinite articles and full noun phrases are used. Where infor-
mation is already known, anaphoric pronominal reference and definite articles
would be expected. Known information may be established through mention
in the specific context of situation/discourse, or known information may be
presupposed in a given context of situation and context of culture. Relevance
construction is in part dependent on knowing and negotiating what may be
textually and contextually presupposed as known and what must be introduced
as new.
Below are the frequencies of reference type in Jay’s text. (Pronominal refer-
ence here includes only pronominal reference to non-human participants and
events. Personal pronoun reference is excluded because these referents are
always definite and the context of situation here involves only two speakers
so inclusion is uninformative. In other situations, personal pronoun referents
could be highly salient and one would want to include them.) Included are only
pronominal it and that. Indefinite NPs are those which occur with indefinite
articles as in a severe thunderstorm, or if the head noun is plural, with no arti-
cle as in severe thunderstorms. Definite NPs include those with proper nouns
(Niagara) as head, and those with definite articles, quantifiers or possessive
articles as in the lightning, six tornadoes, your lights respectively.
Indefinite reference: 39% (59/151)
Definite reference: 42% (64/151)
Pronominal reference: 22% (33/151)
One observation that can be made is that Jay’s use of pronominal reference
is relatively infrequent (22%) as compared with his use of definite and indefi-
nite noun phrases (42% and 39% respectively). Limited pronominal reference
might suggest a text in which many participants are introduced as new, and per-
haps not referred to again. However, in this text Jay introduces 59 participants
as new using indefinite NPs, and refers to them 97 times. Two thirds of his
referencing, though, is in full, definite NPs. Here, the use of full noun phrases
instead of ellipsis or pronominal anaphora adds a pedantic redundancy. For
instance, in the complex sentence in Example (1), the noun phrase the severe
thunderstorm outbreak is fully specified when it is repeated in the dependent
clause, although it is given information:
(1) that might have been the uh time of the severe thunderstorm outbreak
with the six tornadoes cause they were saying that was the worst severe
thunderstorm outbreak to ever occur in cottage country.
In the second mention it would be quite possible to either ellipt severe thunder-
storm outbreak or substitute one for it, as in the worst or the worst one.
Syntax and prosodic features Consideration of syntax and pro-

sodic features is also involved in a full characterization of message organi-
zation. Relevant syntactic constructions include the systematic options for
clause rank alternative sequences and embedding listed and exemplified
below.
Cleft constructions Cleft constructions can mark information that is to be

presented as news. In it’s amazing that a line of thunderstorms can come
through in a half hour, Jay highlights the attribution amazing by placing it in
the complement position of the it is ______ that… construction.
Sentence topic constructions Sentence topic constructions serve to orient an

addressee to a topic shift as in speaking of the weather, they’re calling for
rain tomorrow where the topic element is highlighted. Jay does not use such
constructions. He introduces the topic of the conversation with a comment on
the weather: today it threatened to uh thunderstorm.
Ellipsis As noted, ellipsis refers to the deletion of elements. It is normally

used when information is redundant as a type of zero anaphora. Jay typically
does not delete where deletion is possible. There are certainly cases of ellipsis,
as in the Example (2) where the subject (marked with t1) is ellipted in the third
clause:
(2) when you get temperatures of ninety five the the airi sometimes rises
itself. and then ti causes a storm without a cold front.
However, more typically, Jay repeats the subject in coordinate clauses where
ellipsis would be possible, as in Example (3).
(3) I’ve read a bit.
and uh I’ve learned a bit from tv.
Here it would be possible to ellipt the second mention of subject (I) and the
operator (have) as in I’ve read a bit and learned a bit from tv.
Relative and other subordinate clause types Relative clauses can be used
to background information within NPs. Other subordinate clause types such
as content clauses can serve to move information out of given positions such
as subject into positions after the complement (which typically receives
tonic stress). In Jay’s I’m surprised though today with that big cloud mass
that formed that it didn’t fully develop into a thunderstorm there is a relative
clause (in bold face) and the second that clause (underlined above) provides
the content for the adjective surprise. High frequency of these constructions
increases the overall syntactic complexity of the discourse. Additionally, the
proportional use of logical and temporal conjunctions for subordination (e.g.
because, if, although, consequently) versus coordination adds to relative syn-
tactic complexity of message organization. Another speaker might have pro-
duced the propositional information in Jay’s sentence above as a sequence of
simple independent paratactically linked clauses as in A big cloud mass formed
today but it didn’t fully develop into a thunderstorm. That surprised me. The
complexity of Jay’s style is partly created by his habitual selection of hypotac-
tic rather than paratactic relations between clauses. This is in part perhaps an
effect of his interest in the weather and what causes events like storms and
tornadoes, but also results from his foregrounding of his reactions to meteoro-
logical events as in this instance.
Preposing Preposing moves an element to initial position so that it is fore-

grounded and lets the final element also take tonic prominence. Preposing is
used quite regularly in Jay’s discourse. In Example (4) temporal information
(when you get temperatures of ninety five) and conditional information (if con-
ditions are right) are foregrounded:
(4) [when you get temperatures of ninety five]i the the air sometimes rises
itself ti.
and [if conditions are right]i that could trigger off a thunder storm later ti.
Passive and ergative Passive voice as in the tea was brewed by Sam is rare in
weather report registers presumably because weather is culturally construed as
happening or being caused by inanimate forces and thus there is limited pos-
sibility for agency deletion or emphasis. Ergative use, where an affected entity
appears as subject in an active clause, is more frequent as in and I watched as it
brewed where the storm, the referent of it is construed as ‘brewing’.
Prosody The syntactic constructions illustrated change the potential

distribution of tonic prominence such that information can be marked as new
and salient, or known.
Example (5) marks the high frequency lexical item actually as salient:
(5) CHI: <uh> [<] ninety mile an hour winds <can be a tornado> [>].
res: <is that right> [<]?
res: is <that right> [>]?
CHI: <yeah> [<].
res: oh.
CHI: anywhere from ninety to one hundred and twenty.
res: I see.
CHI: actually it’s anywhere from ninety to two hundred and twenty.
Here, the probability adverb actually was given emphatic contrastive stress,
which its preposed initial position permits. In the last element two hundred and
twenty, two is the site of contrastive stress.
Unsurprisingly, in pedantic speech we find features typical of speech in gen-
eral such as the use of cleft, topic and pre- and post-posing to organize focus
and prominence, but also features which are either more characteristic of written
modes such as the use of embedded clauses, or, in the case of frequent full lexical
repetition of known information, perhaps unique to pedantic speaking in ASDs.
A message organization battery can be combined with other features that
are characteristic of pedantic speech to create a ‘pedantic battery’ as in
Table 7.3. Calculations of the proportion of pedantic speech relative to total
speech in c onversation might add to this characterization.
Table 7.3. Pedantic battery
favourite topic
• pedantic speech is more likely when favourite topic is the topic
• speaker is more informative than might be expected in context
lexical repetition of given information
collocational patterns associated with the (often technical) lexis of the registers of the favourite
topic
syntactic complexity – as for written norms
lexical density – as for written norms
syntactic aspects of message organization
• reference (through lexical repetition more than pronoun reference)
• theme marked constructions
• sentence topic constructions
• limited ellipsis
• embedding: relative and other clause types
• preposing and postposing
• passive and ergative constructions
While the lexical repetition is a marked and characteristic element of pedantic

speaking, other kinds of lexical and syntactic patterning contribute to the pedan-
tic quality. As has been shown, these features cannot be adequately addressed
without consideration of registers and register variation. With pedantic speech,
lexis associated with specialized registers is a partial sign of special interests.
7.3 Monitoring and change over time

If one or more discourse behaviours are clearly associated with particular diag-
noses, stages, or levels of severity, it may be possible to use the behaviours as
clinical signs of stability or change. Especially in contexts where new treat-
ments and/or interventions have been or are being developed and biomark-
ers are absent or not widely available, establishing behavioural correlates of
responsiveness to treatment can be valuable. In such situations clinical discourse
analysis can function as a way of discovering and tracking how treatment affects
discourse patterns. In this section, we give an example of the use of clinical dis-
course analysis to discover behaviours associated with drug treatment response
for patients with Alzheimer’s disease, and then present a battery for examining
interactional aspects of discourse as a means of monitoring treatment.
7.3.1 Monitoring change over time

A number of researchers have observed that pragmatic, interactional and
highly automatized aspects of language may be preserved in the discourse
of speakers with significant impairments in left hemispheric regions classi-

cally associated with linguistic competence (Cardebat et al. 1993; Nespoulous
et al.1998) and in speakers with AD (Ulatowska and Bond-Chapman 1991;
Duong et al. 2003). Nespoulous et al. (1998) characterize discourse with pre-
served highly automatized lexis, preserved syntactic structure and capacity to
express attitude, evaluation and probability as ‘modalizing discourse’. They
contrast this with ‘referential’ discourse which has fully lexicalized proposi-
tional content. In research on AD, Duong et al. (2003) found that participants
with AD produce more modalizing discourse relative to referential discourse
than a group of older adults with no cognitive impairment. Duong et al. (2003)
suggest modalization reflects participants’ efforts to maintain their commu-
nicative role while faced with referential difficulties and their awareness of
processing deficits.
7.3.2 Tags in Alzheimer’s disease

In work describing the discourse of patients participating in a clinical trial for
donepezil, a cholinesterase inhibitor widely used to treat patients with AD, Asp
et al. (2006a) identified a particular use of tags as characteristic of patients’
speech. The first example of this tag use, which we called the checking ‘self-
referential tag’, is one in which people with AD use tags with statements that
refer to their own daily routines and activities. Thus if asked ‘what time do
you get up in the morning?’ a patient might respond with a tagged statement
such as I get up at six, don’t I? or they might check a response about their age
(I’m 83, aren’t I?). The tag checks the certainty of the proposition, typically
with caregivers who participate in interviews. What is unusual about check-
ing self-referential tags is that they check information normally available from
the speaker’s own episodic memory. A second kind of self-referential tag is
used to monitor information flow (I told you that already, didn’t I?). These are
common to speakers without AD. We paid attention to them because we were
interested in what speakers with AD were doing with self-referential tags.
Examination of interviews with one hundred patients and caregivers showed
that tag frequencies differed hugely between patients and caregivers, that they
used them differently, and that the type of tags patients used correlated with
their response to treatment. In brief, more patients (71%) than caregivers (21%)
used self-referential tags and used them more frequently: patients produced
93% of all self-referential tags in the corpus. Moreover, more patients used a
greater percentage of episodic checking tags than caregivers: 80% of patients’
tags are checking tags and most of these (72%) are singular – the sort used to
check episodic facts about oneself as in I don’t have a dog, do I? More than
half of caregivers’ tags were monitoring (54%) and their episodic checking
tags were mostly plural (77%). Qualitative analysis of this difference suggested
that caregivers tended to use plural episodic checking tags to engage patients
in the interaction (We went out for dinner last night, didn’t we?) rather than as
expressions of uncertainty about episodic information. Among patients, the
31 who used monitoring tags with or without checking tags had significantly
better scores on standard neuropsychological measures of cognitive function
after twelve months of treatment than patients either only using checking tags
(40) or those not using tags at all (29) (Asp et al. 2006a).
We hypothesized that ability to monitor information flow reflects preserved
and/or up-regulated prefrontal function which may occur endogenously or as
treatment response in AD and that, while the frequent use of self-referential
tags generally is a sign of episodic and information processing deficits, aware-
ness of these deficits as reflected in tag use by patients with AD is more posi-
tive than lack of such awareness. We are currently in the process of checking
another corpus from a recently completed double-blind placebo-controlled
trial with 120 patients to find out whether this is indeed the case. This work, as
well as the research mentioned above, suggests that a ‘modalization battery’
might be helpful as a general index of processing difficulty and of monitoring
abilities. If tailored for a specific group, such as AD patients being treated with
cholinesterase inhibitors, it may have the further benefit of providing another
way of evaluating treatment response for individual patients.
A model for such a modalization battery is offered in Table 7.4, and its use
illustrated in a single text. The modalization battery consists of resources for
the description of a system of patterns associated with modalizaton as sug-
gested by Nespoulous et al. (1998) and as suggested by the sociocognitive
model. It thus includes interactional, ideational and organizational selections
which may be relevant.
If the battery were being developed specifically to evaluate monitoring capac-
ity relative to episodic and semantic memory deficits in AD, other features such
as the frequency of self-initiated repairs, incomplete utterances and syntactic
complexity might be included in the initial investigation on the grounds that
these may function as positive or negative signs of discourse monitoring abili-
ties. Below we analyse a single text of an interview with a speaker with AD to
illustrate the modalization battery.
Text 7.2 Memory problems

(1) IV: Mmhm.
(2) So can I ask you to tell me a little bit about the time when your health changed.
(3) When you started noticing you were having memory problems.
(4) P: Well I I guess it would only be {(laughs)} recently and not very long ago.
(5) IV: Mmhm
(6) P: Um I suppose # in the past year.
(7) IV: Mmhm.
Table 7.4. Modalization battery
(1) Speech functions

(i) proportion of the number of statements, exclamations, questions (WH/polar), commands
(jussive/optative) relative to the total number of independent clauses
(ii) number of questions used to check self-referential episodic information or monitor
discourse as a proportion of independent clauses
(iii) number of minimal responses as a proportion of the number of speaker’s turns
(iv) number of minor clauses relative to total number of independent clauses
(2) Tagged statements in terms of the tag types, intonation contours, number of tags in
proportion to the number of statements by speakers, and the number of these which are self-
referential (monitoring or checking)
(3) Checks and intonation contours: code for checks (e.g. you know) and proportion of
statements with rising tone
(4) Modal verbs, adjectival and adverbial modulations: code for likelihood, capacity,
hypothesis, obligation, habitual action and assess their frequency in relation to the number of
statements and minor clauses. Include alternatives, when these limit the certainty with which
a statement is offered (e.g. I went shopping or something).
(5) Cognition predicates: (e.g. think, believe, suppose, know) assess their frequency in relation
to the number of statements and minor clauses. Note the proportion of cognition predicates
which are negative (i.e. deny knowledge or express doubt about episodic or semantic
information).
(6) Automatization of lexis
(i) proportion of high-order taxonomic reference relative to basic or subordinate level
categorization in N/V/A/Av selections
(ii) indefinite reference relative to definite reference
(iii) lexical density and/or richness measures
(8) What sort of things started happening that made you aware?
(9) P: Um # well um of course I would think that I had said or done something
(10) and someone would tell me I hadn’t and that sort of thing you know.
(11) IV: Mmhm
(12) P: Nothing nothing very drastic but ah just,
(13) CG: The little everyday things wasn’t it?
(14) P: Little little things.
(15) IV; : Mmhm
(16) and what sort of took you off to the doctor the first time to check it out?
(17) P: Ah well just what other people were saying to me
(18) that I was getting forgetful and so on.
(19) IV: Mmhm.
(20) P: But ah of course ah just the ordinary doctor that doesn’t # have much of a remedy
for that. ((laughs))
(21) But ah I I went to my doctor anyway
(22) IV: And and did your doctor refer you to the clinic?
(23) P: Ah # # did he?
(24) CG : Yes
(25) P: Yeah
(26) CG: Yes Dr Smith did Dear.

(27) P: Yes Dr Smith yeah.
(28) CG: He wanted the Geriatric assessment test done.
(29) IV: Umhum.
(30) And how was that for you?
(31) Was that was that test ah
(32) P: I I I ah I don’t know.
(33) IV: Guess you went through it
(34) and maybe didn’t
(35) P: Didn’t even know it ((laughs))
Text 7.2 is a fragment of a home-visit interview with a patient, Doris, and
her caregiver. The patient has moderate, untreated AD. As can be seen from the
interview, Doris participates actively. She takes almost half (12/25) of all turns,
attempting responses to every question addressed to her. She is obviously atten-
tive, even successfully (if ironically) completing the interviewer’s final turn.
Her lexical density score in this fragment, calculated here as the proportion of
non-repeated lexical words per independent clause, is 2.6 (24/9). This is in the
normal range for spontaneous spoken adult discourse. However, there is some
evidence represented in pauses, hesitations, false starts and incomplete utter-
ances (21 altogether) that she has some processing difficulties. Examination
of modalization features and lexical patterning (summarized in Tables 7.5 and
7.6) suggest the source of some of these problems may be episodic memory –
that is, since she is being asked about when her memory problems started, what
she personally noticed, what sent her to the doctor and so on, the questions
make demands on her recall of her own recent past experience.
As Table 7.5 indicates, half of Doris’ responses to questions are statements.
Otherwise, she uses elliptical polar responses, minor clauses and one polar
question to respond to questions. One of her statements is checked with a mon-
itoring you know, but otherwise they are untagged and have falling tone (so tags
and tone are not represented in Table 7.5). Her only question (about referral) is
addressed to her caregiver (did he (23)), indicating that she doesn’t know who
referred her to the memory clinic. Three of her other responses are hypotheti-
cal: she uses a modal would, (It would only be recently (4)), and lexical verbs,
(I guess (4) and I suppose ( 6)). In full independent clauses, the proportion of
lexical verbs referring to mental cognition processes is 71% (5/7). If minor
clauses are included , the proportion of mental cognition predicates is lower at
41% (5/12). In addition to guess and suppose, there is the habitual past cogni-
tion (I would think (9)) and two negative mental cognition predicates (I don’t
know) in response to questions seeking episodic information about her visit to
a specialist.2 Elliptical responses following her episodic polar question echo
her caregiver’s responses and thus add no new information. The sense that
she speaks well but not very informatively is thus partly a consequence of the
hypothetical modalization, and in some cases frank denial, of knowledge about
Table 7.5. Modalization features in the discourse of a person with AD
Element Category Feature(s) Number Value
Speaker turns 12/25 48% of all turns

Independent (include elliptical 9/12 75% of patient’s
clauses responses to polar turns
questions)
Minor clauses 3/12 25%
Speech Statements Full 6/9 66% of patient’s
functions independent
clauses
Elliptical (positive) 2/9 22%
Polar question Episodic check 1/9 11%
Modalizations Statement check (monitoring – 1/9 11% of patient’s
you know) statements and
minor clauses
Modal verbs Hypothesis 1/9 11%
(would)
Mental cognition Hypothesis (guess, 5/9 55%
suppose)
Not know (2)
Think
her past: leaving out the habitual would and the monitoring check, it is the case
that 58% of her clauses are modalized in ways that limit episodic certainty.
The lexical patterns displayed in Table 7.6 augment this impression. For
example, of the nouns she uses, only two (remedy and Dr Smith) could be clas-
sified as subordinate level and Dr Smith is an echoic repetition of her caregiver
so would be excluded in a count. Of the others, half are superordinate and half
are basic level and one of these is repeated. None of her other lexical selections
are subordinate, except perhaps the adjective drastic. Excluding personal pro-
noun reference, Doris uses nine noun phrases (NP). Five (55%) of her NPs are
indefinite. Moreover, three of her definite NPs (excluding the echoic Dr Smith)
have very general reference: one has superordinate lexis (that sort of thing);
one gives a time span (in the past year) but is attenuated by a mental cognition
verb indicating hypothesis, (I suppose) as well as by the preposition in which
makes this a rather underspecified elaboration of recently; and one refers to
the ordinary doctor, rather than the ‘family doctor’ or ‘general practitioner’
or the contrastive ‘geriatrician’ or even ‘specialist’. Only my doctor refers to a
specific, known participant.
The other word categories show a somewhat similar pattern. All verbs (5)
referring to actions or states are superordinate, and Doris’ basic level verbs (8)
refer exclusively to mental cognition and verbalization processes such as think
Table 7.6. Lexical patterns in the discourse of a person with AD
Taxonomic relations
Reference
(n) = number
of occurrences Nouns Verbs Adjectives Adverbs
Personal I (9) Superordinate sort thing do get ago

pronouns me (2) people (become)
you (1) have go
he (1)
Definite – the past year Basic level year guess long past recently
reference – that sort of doctor (2) suppose little
thing think (2 – echo)
– the ordinary tell forgetful
doctor say (2) ordinary
– my doctor know (2)
Dr Smith
(echo rep)
Indefinite – something Subordinate remedy ?drastic?
reference – someone Dr Smith
– nothing (2) (echo rep)
– things (echo)
– other people
– a remedy
and say. There are no subordinate level verbs or adverbs and only one poten-
tial candidate (drastic) as a subordinate level adjective. Pooling these patterns,
we see that only 8% of Doris’ lexical selections instantiate subordinate level
concepts. The rest refer to superordinate (36%) and very high frequency basic
level (56%) concepts.
Prevalence of indefinite rather than definite reference, superordinate and
high-frequency basic level concepts rather than subordinate level concepts,
together with discourse which is heavily modalized for uncertainty about epi-
sodically salient information all contribute to the overall pattern of speech
which is rather uninformative.
The pattern of co-operative but ‘empty’ speech in AD has been attributed
variously to the breakdown of lexical–semantic systems, conceptual systems
or to problems with executive function processes which limit access to these
systems (e.g. Hier et al. 1985; Nicholas et al. 1985; Carlomagno et al. 2005).
Each of these difficulties is progressive in AD so establishing change patterns
of modalization and lexical selection over time, and in the context of treatment,
may be effective means of monitoring change.
7.4 Steps to usability

Above we have presented descriptive batteries for characterizing pedantic
speech in ASDs and modalizing discourse in AD. These may be used in other
contexts. For instance, the initial observations about modalizing discourse
were made with regard to speakers with aphasia. We suggested that these bat-
teries could provide descriptions which in themselves have an a priori value as
descriptions in under-investigated areas. We also are interested in the potential
of using the information from such characterizations for monitoring change
and as a supplement to diagnoses. Another useful application for discourse
characterizations is in the development of rating scales.
A rating scale is an instrument used to diagnose disorders, evaluate severity
or subgroup, or track change over time. A rating scale tells you how to inter-
pret an observed set of behaviours by assigning a value to them. Linguistically
explicit scales that measure aspects of discourse offer a useful complement to
other scales and neurocognitive assessment tools.
Observations allow qualitative and quantitative evaluation of a group (or an
individual relative to a group or a baseline state) based on some described, coded
set of phenomena. From the initial observations, linguistic descriptions can be
used to develop scales. Hypotheses can then be generated to test whether a feature
or bundle of features will be present, absent or occur with a specified frequency
likely to be associated with the disorder, disease phase, treatment response, or
other research question being investigated. Depending on what a scale evalu-
ates, rating scales may be used by clinicians, researchers, families, educators and
other caregivers in clinical contexts or a variety of life situations.
Rating scales based on discourse analysis directly and explicitly address dis-
course features. Insofar as discourse is a readily observable everyday behav-
iour and reflects neurocognitive states, rating scales that refer to discourse may
be quite useful. This is well established as practice in assessing child language
development, and of course in aphasia studies. The design of scales can be
targeted to specific users and take into consideration their potential linguistic
experience, time demands and likely contexts of use. For instance, a general
practitioner may need a tool that can help inform a decision to refer. Clinical
discourse analysis can also be useful in designing tools that are specific to
particular disorders and that are sensitive to register: in assessing AD, perform-
ance in the clinic is important because this is where medical personnel see and
evaluate patients. A scale for evaluating change from a baseline in relation to
treatment would also be useful for assessing AD. Thus descriptions that inform
rating scales need to be based on data collected from appropriate contexts of
situation and their associated registers. The approval process for rating scales
varies in different communities, but the timeline is never short.
We end the chapter with an example of a rating scale developed for use in
ASDs. Conversation is an area of particular difficulty in ASDs. Thus a tool
for measuring conversational difficulties specific to ASDs might be useful in
assessing the degree of impairment, and in evaluating change over time or in
response to treatment. In research aimed at understanding social communica-
tion difficulties in high-functioning autism and Asperger syndrome, de Villiers
et al. (2007) developed a preliminary scale for rating conversational impair-
ment in ASDs. Based on a descriptive linguistic approach, this scale identified
five areas of conversational difficulty in ASDs:
1) atypical intonation
Speech has a monotone quality and/or atypical stress selection
2) semantic drift
Includes abrupt switching of topics and ‘disengagement from verbal
context’
3) terseness
Minimally responsive, including short, delayed responses or necessary
prompting
4) pedantic speech
Stereotypic or rehearsed sounding speech with more factual or technical
detail than is required for the situation
5) perseveration
Excessive persistence on a particular chosen topic
These five constructs were derived from a wider set of characteristic dis-
course features which were then collapsed to create a more useful version of
the scale. As it makes explicit the nature of certain social communication diffi-
culties in ASDs, this scale may be usefully applied to measure variation within
an ASD population. A full description of the development of the rating scale
can be found in de Villiers et al. 2007.
8 Cognitive models, inferencing and affect
8.1 Introduction
This chapter outlines approaches to top-down cognitive modelling and
inferencing, and addresses functionally grounded work on affect. We describe
each area and illustrate its potential for addressing questions in clinical dis-
course analysis. We also review recent work from neuroimaging and lesion
studies to suggest some of the relevant neural systems. As usual, we draw on
various disciplinary perspectives and theoretical models. Our practical moti-
vation here is to use what works, and has potential for coding corpora in the
various linguistic contexts and situations encountered doing clinical discourse
analysis.
Cognitive models in general characterize information bundles of various
kinds. Perhaps the most familiar are those used to represent words or word-like
concepts. Models for words may be more or less detailed depending on the
tolerance for elaboration within a particular framework, but morphosyntactic
class, inflection and distribution features are typically indicated. How a word is
pronounced – its phonological form and regular phonetic variants – will be spelt
out in phonological and phonetic representations. Semantic features are often
specified only at superordinate levels as in THING/EVENT or merely indexed
through the use of the ‘CAPS-for-concept’ convention. Thus, the model for the
lexeme ‘cat’ will include the information that it is a common count noun, with
the inflectional and distributional features of this class – it can occur as head of
a noun phrase and it inflects for plural number /s/. It is pronounced /kæt/. The
entry might also include semantic features such as THING, ANIMATE and so
on, or just CAT as a shorthand for some presupposed set of features.
Conventionally the semantic features associated with word-like concepts
have been presented as modality-neutral conceptual representations in linguis-
tics, either in some sort of algebraic formulation or, less formally, in lists of
lexically represented features. Depending on the model, there may be point-
ers to schematic, modal, functional and encyclopedic information included as
relevant. One rationale for such a presentation is that a modality-neutral concep-
tual system would allow communication between otherwise incommensurable
135
modes such as language, vision and taste (e.g. Jackendoff 2002; in preparation).
However, work in cognitive neurology (Mesulam e.g. 1998), cognitive neu-
ropsychology (Warrington and Shallice 1984; Warrington and McCarthy 1987)
and neuroimaging (Thompson-Schill 2003; Martin 2007) raises questions about
the plausibility of modality neutral concepts and the existence of a modality
neutral conceptual system (Thompson-Schill 2003). We do not address these
debates here but offer brief synopses of some of the issues as they are relevant
for clinical discourse analysis in chapter 9. Here we merely make the point that
agreement even on something as seemingly basic as the neural architecture that
supports word-like concepts is absent, so all comments about neural substrate
need to be read with caution. That said, we favour distributed models which
treat modality/functional specificity as central to neural organization and link
information across modes through transmodal gateways (e.g. Mesulam 1998).
We discuss this sort of model further in the final chapter.
While there is limited consensus about the neural instantiation of concepts,
there is robust evidence that we use information represented in cognitive mod-
els of concepts. Cognitive models for concepts are ways of characterizing what
we know about real and abstract objects and events, relations and attributes.
Top-down cognitive models are similar except that they highlight aspects of
generic situation potential. Following van Dijk and Kintsch (1983), we think
of top-down models as contributing to the construction of ‘situational models’
which, as we use the terms, are speakers’ mental representations of particular
discourses. We assume that conceptual and top-down models have relation-
ships to executive functions on (at least) two dimensions. First, they affect our
capacity to maintain and monitor information ‘online’ by making recurrent
features redundant and informing selection and inhibition processes (e.g. van
Dijk and Kintsch 1983; Shallice and Burgess 1996; Tinaz et al. 2006). Second,
they inform executive control processes (EFC) associated with inferencing,
judgement, decision making and action planning (Norman and Shallice 1986;
Baddeley and Della Sala 1996; Goel et al. 1997; Royall et al. 2002; Paxton
et al. 2008).
Our emotional states and affective responses continuously modulate and are
modulated by experience in situations (Frijda 1986; Mesulam 1998). Insofar
as the conceptual and top-down cognitive models we bring to bear in the
interpretation of experience contribute to shaping what that experience is, posi-
tive, neutral and negative evaluations attached to those models (either directly
through experience or indirectly as matters of cultural transmission) will influ-
ence not only emotional states and affective responses, but also directly or
indirectly the judgements, decisions and plans we make (Coricelli, et al. 2007;
Rushworth et al. 2007 for reviews). Thus, affect and emotion have roles to play
in ‘top-down’ cognitive processing associated with executive control processes
(Bechara et al. 2000; Royall et al. 2002; Coricelli et al. 2007). We include a
Cognitive models, inferencing and affect 137
section on affect in this chapter because of this association. However, from

linguistic, if not psychological, psychiatric and AI perspectives, affect and
emotion in discourse have been little studied. Thus, our goals here are to index
their importance and point to some work where affect and emotion have been
the focus of study.
8.2 Top-down models

This section outlines top-down models and their sources and discusses their
use in clinical discourse analysis. Specifically, we discuss scripts, frames,
scenarios and schemas as developed in psychological and Artificial Intelligence
(AI) literature as well as in discourse analysis and linguistics. In Chapter 5, we
introduced these elements as representations of event and situational knowl-
edge relative to contexts of situation and culture.
Scripts refer to conventional action sequences associated with generic
situations (Schank and Abelson 1977). That is, when we do something or plan
to do something such as going to a movie, or going to a restaurant, we have
expectations about the essential actions that characterize the generic structure
potential of the activity and their order (Hasan 1978).1 Galambos (1986) has
distinguished actions which may be more central to a script rather in the way
that some features may be more central to a concept, defining its prototype,
while other features may be distinctive without being defining. So the script
for going to a movie will include going to the cinema, buying tickets, possibly
buying popcorn, finding a seat, watching the movie. Here, finding a seat is
neither central nor distinctive. We do this at public lectures and entertainments,
on buses and so on. Interestingly, although buying popcorn is optional, it is
relatively distinctive insofar as one does not, for instance, buy popcorn to eat
at plays, concerts or at the opera. Watching a movie is central and distinctive.
Order constraints require that finding a seat cannot precede buying tickets nor
follow watching the movie. People usually know both the elements and their
order.
There is an extensive body of literature associating script knowledge and
performance with prefrontal function and medial temporal lobe function (e.g.
Luria 1965; Norman and Shallice 1986; Stuss and Benson 1986; Grafman
1989; Godbout and Doyon 1995; Shallice and Burgess 1996; Zanini et al.
2002; Godbout et al. 2004; Zanini 2008). Specifically, patients with lesions in
the prefrontal cortex (Sirigu et al. 1995; 1996; Zalla et al. 2000; 2003) and with
neurodegenerative diseases focally affecting prefrontal systems have difficulty
recognizing event sequences and ordering events appropriately (Cosentino
et al. 2006). People whose primary impairment is in semantic memory have
difficulty with event components but not necessarily with sequences (Cosentino
et al. 2006) and degenerative disorders such as Alzheimer’s and Parkinson’s
disease may impair knowledge of components and their sequences (Godbout

and Doyon 1995; Allain et al. 2008). In autism, it has been shown that there are
problems generating scripts with all the central elements, although participants
typically have such knowledge (Volden and Johnston 1999). There is debate as
to whether script knowledge is represented in toto (e.g. Grafman 1989; 2002;
Zalla et al. 2003) or separately as knowledge of sequence and component ele-
ments. Evidence supporting the latter view comes from studies indicating that
sequence and component elements may be discretely impaired by focal lesion
damage (Sirigu et al. 1995; Sirigu et al. 1996; Cosentino et al. 2006), and from
imaging and lesion studies that suggest that script knowledge is distributed
and its activation engages a functional network that includes the dorsolateral
prefrontal cortex, temporal and parietal regions and basal ganglia (Godbout
and Doyon 1995; Tinaz et al. 2008).2 There is consensus that script knowledge
is closely linked to executive function processes supported by the prefrontal
cortex and that these processes are especially associated with abilities to plan
and carry out activities in daily life (Shallice 1982; Rusted and Shepherd 2002;
Royall et al. 2005; Allain et al. 2008; Zanini 2008).
Frames refer to information associated with situation specific concepts
(Minsky 1975; van Dijk 1977). For example, an event frame would include an
event structure specification familiar as predicate and participant role relation-
ships and circumstances, and other information such as social evaluations and
logical conditions of the event and its participants. Frames are thus a device
for characterizing acquired attitudinal, encyclopedic and lexical knowledge.
So, within the script of going to the movies, we could postulate a frame for
buying a ticket. This will involve a buyer and a seller, a ticket and money
(the predicate–argument relationship). It will also include information that the
ticket is likely made of paper and is required for admission to the cinema and
an expectation of cost. This can be represented in an argument structure type
format with typical default features and options included for arguments and
circumstances. Depending on the amount of detail one wants for the descrip-
tion, it may also include explicit evaluations and conditions as propositions. An
example for buying a ticket might look like Figure 8.1.
Frames have been used to examine how people orient to and within events.
In AI and linguistics, the role of frames is the representation of conventional
knowledge structures (Fillmore 1976; 1982; Minsky 1975; 1977). In discourse
analysis, Goffman (1959; 1961; 1974; 1986), Gensler (1977), van Dijk (1977),
Tannen (1993a; 1993b) and others have used framing to explore contextualiza-
tion processes. Goffman looked at how people (re)position themselves with
respect to situation types and each other and at the discourse behaviours that
serve to ‘reframe’ situations. Van Dijk (1977) investigates the role of frame
knowledge in inferring unstated information relevant to the interpretation of
discourse. Medical discourse is one area where these approaches have been
Buy ticket
Condition: necessary for admission
Source +/– Agent [person/machine]1
Agent and Goal recipient [movie goer]2
Theme transferent [TICKET]1�2
Theme transferent [MONEY: approximate cost]2�1
Figure 8.1 Frame for ‘buying a ticket to the movies’ (The superscripts are
indices for the direction of transfer in the BUY process, where the ticket goes
from the seller/machine to the movie goer and money goes from the movie
goer to the seller/machine.)
used extensively (e.g. Tannen and Wallat 1993; Coupland et al. 1994). Gensler
(1977) used frames as abstracts for interpreting anaphora. Framing is also used
to study risk-taking behaviour and decision making more generally (e.g. De
Martino et al. 2006; Kahneman and Frederick 2007). Insofar as frames involve
relations between concepts and propositions relative to generic situation types,
they are inherently more complex than concepts and therefore present more
challenges for neural characterizations. Van Dijk and Kintsch (1983) assume
that frames are represented in semantic memory. So long as we allow that
semantic memory for frames is likely to be distributed according to the modal
and functional values of concepts entailed by particular frames, this seems
reasonable.
Frames (or something similar) are used in presuppositional and inferential
processing and decision making. Some imaging studies have examined the way
particular frames may affect these processes and findings suggest that the ante-
rior cingulate cortex is more activated in making decisions that involve analytic
judgements (Coricelli et al. 2007) in contrast with decisions which differen-
tially activate the amygdala. The latter are presumed to involve an affective
style (De Martino et al. 2006; Kahneman and Frederick 2006). (The anterior
cingulate cortex is normally involved in ‘difficult task’ processing since its
generic roles are linked to behaviour monitoring, error correction and response
selection (Duncan and Owen 2000; Krawczyk 2002; Rushworth et al. 2007).)
The orbitofrontal cortex also appears to be involved in decision mak-
ing and response evaluations. It receives projections from sensory areas and
from limbic regions associated with emotion and memory and it feeds back
through monitoring and attentional systems (Mesulam 1998; Royall et al.
2002; Rushworth et al. 2007). It has been suggested that the role of the medial
orbitofrontal cortex in decision making relative to frames is that it integrates
emotional and cognitive information, so that greater activation in this region
may signify more ‘rational’ than emotive decision making relative to a frame
(De Martino 2006; Coricelli et al. 2007; Kahneman and Frederick 2007). For
example, Camille et al. (2004) report that patients with lesions to the orbito
frontal cortex did not experience the emotion of regret when they lost in posi-
tively framed gambling tasks and did not learn from their losses, though they
did feel happy or disappointed depending on whether they won or lost. Control
participants did feel regret and learnt from the experience of regret so that
they modified their behaviour to resist positive frames associated with high
gain and risk and selected instead certain modest gain options. Coricelli et al.
(2005) showed in an fMRI study that heightened regret increased activity in
the medial orbitofrontal cortex, the dorsal anterior cingulate and the anterior
hippocampus. Coricelli et al. (2007) hypothesize that the engagement of the
hippocampus in this loop reflects the activation of consciously accessible
information, incorporated through the orbitofrontal cortex activity into the
decision-making process – allowing the high risk frame to be evaluated and
rejected because it leads to the experience of regret. Activation of the anterior
cingulate when people attend to emotional information has been observed in
other studies (Rushworth et al. 2007 for review). Lane et al. (1997) suggest
that such activity is coherent with increased monitoring and attention needed
for response selection. Figure 8.2 shows some of the brain regions associated
with this sort of decision-making process.
There is overlap in what frames and scenarios describe insofar as they
both refer to features of generic situations such as participant roles. However,
scenarios were originally posited to address situational knowledge in non-
propositional form, whereas frames presupposed relatively fixed propositional
knowledge (e.g. Kintsch 1974. But see also van Dijk and Kintsch 1983 where
non-propositional knowledge is not ruled out; and Kintsch 1988 where it is
assumed though not addressed because of the challenges it presented). Scenarios
and frames also differ from each other in that scenarios can be used to describe
elements of ‘setting’ in the generic situation, whereas frames characterize con-
cepts relevant to such generic situations. Thus, a frame can be a model that
refers to concrete entities (theatre ticket), processes (buy ticket), abstractions
such as GAIN/LOSS, or emotions such as DEPRESSION/HAPPINESS, while
scenarios will refer to generic situation types such as AT THE RESTAURANT,
AT THE MOVIES, IN THE CLASSROOM. The scenario AT THE MOVIES
will include a film and possibly trailers (and increasingly advertisements).
(1) Scenario: AT THE MOVIES
• ticket booth at the entrance
• it will be dark
• it will be public
• there is seating with aisles separating the rows
• people may eat popcorn
Anterior cingulate cortex

associated with monitoring
response selection and error correction
Orbital
frontal
cortex
integrates Hippocampus
cognitive, accesses prior experiences of regret
sensory and induced by high-risk gamble
emotional
information
Figure 8.2 Decisions and brains
• a film will be shown

• other
Evidence that people may use scenarios in processing discourse came initially
from a study by Sanford and Garrod (1981) that compared reading times in
generic versus non-generic situations: people processed more quickly when
the scenario was a generic one suggesting that they used knowledge of the sce-
nario to interpret instantiated information (Sanford and Garrod 1981). A nicely
illustrative recent set of examples are two studies investigating the effects of
the presence or absence of a title on processing a series of paragraphs (St. George
et al. 1994; 1999). In both studies, participants read pairs of paragraphs which
were identical except that one in each pair of paragraphs had a title which made
texts coherent. The paragraphs without titles were quite difficult to interpret.
For example, one of the paragraphs describes horseback riding including the
facts that it’s expensive even if you have the gear, that it can be dangerous and
instruction is important, and describes the process of mounting. Without the
title, ‘horseback riding’, from which one can infer the scenario of a first riding
lesson, it is incomprehensible, though each of the sentences is well formed.
The first study (St. George et al. 1994) investigated the electrophysiologi-
cal response of participants under both conditions and found that the N400
wave (typically associated with semantic processing) was consistently bigger
for words in the untitled rather than in the titled paragraphs. Similarly, in the
second study, fMRI showed a frontotemporal network of bilateral activation
which was much more extensive for the untitled versus the titled paragraphs.
These and similar studies suggest that the time course for processing with and
without top-down models corresponds to the greater processing demands of
the unfamiliar condition.
Brown and Yule (1983) describe schemas as models that are operationalized
in the interpretation of discourse. They refer not only to narrative and other
discourse schemas, but also to concepts that account for stereotypic processing
effects in terms of interests, preoccupations and other culturally conditioned
expectations of participants. Examples of the operation of schemata include
different inferences about the nature of an event type based on interests (where
music students and a weight-lifting team interpreted the same text as refer-
ring to a musical evening and a card game respectively, in Anderson et al.
(1977: 372)) and different orientations to events which help to explicate the
different discourse acts of participants (Tannen and Wallat 1993).
Like scripts and frames, schemas and scenarios facilitate contextualization
and integration processes. The variability and potential complexity of these
models suggests that they are likely to have widely distributed representations
depending on their contents and the situations they refer to. For example, in
talking about the scenario ‘at the movies’, we used propositions such as ‘it will
be dark’ and ‘people may eat popcorn’. This information though, might well
be neurally represented in regions associated with darkness and the sound,
taste and smell of popcorn. So what the scenario refers to may be a set of
episodically encoded multi-modal experiences. We may also know these as
propositions but that is not necessary. In this respect, scenarios contrast with
frames which are assumed to refer to propositional information associated
with semantic memory. Similarly, schemas viewed as stereotypic expecta-
tions affecting interpretations of situations may be difficult to distinguish from
frames since they can be propositionally represented. However, when the term
schema is used to refer to highly abstract structures such as those occurring
in narrative or those representing the basic shapes of common objects this is
clearly not the case.
The four types of top-down model we have described do show areas of

overlap as we have suggested and, indeed, some authors treat them simply
as variant labels for situational knowledge (e.g. Fillmore 1982; Bednarek
2005). However, in their original formulations each construct was articulated
to address particular aspects of situational knowledge. Research in discourse
processes and neurocognitive function over the past thirty years suggests that
the nuances early researchers sought to articulate are more than discursive epi-
phenomena. For example, that we use non-propositional information (scenar-
ios) to process discourses coherently seems to be reflected in recent imaging
studies. Ferstl et al. (2005) and Ferstl and von Cramon (2007) have shown that
activations associated with the representation of spatial information needed
for coherent interpretations of stories are very close to activations associated
with space perception and object location in general and these activations are
distinct from activation patterns associated with processing temporal and emo-
tional information in discourse.
In developmental and neurodegenerative disorders, difficulties in discourse
processing can be reflected in problems with top-down models. For instance,
Alzheimer’s impairs semantic and episodic memory. People with AD may thus
have difficulty accessing knowledge of frames, scripts and scenarios. Moreover,
cholinergic denervation of the prefrontal cortex and/or other neurotransmitter
deficits can limit capacity for the sorts of complex integrative processes that
are required in order to integrate top-down models with instantially presented
information. These difficulties may be reflected in what people say and how
they respond to situations as well as in actual discourse abilities requiring spe-
cific generic knowledge. For example in Text 8.1 below a patient with moder-
ate AD responds to questions about an incident in which he got lost while out
on a walk close to his home.
Text 8.1 Getting lost

(1) IV: Can you tell me what happened that day.
(2) P: Well # uh it was uh just it it looked different to me,
(3) P: Like I mean # some of them streets # I uh see I been in that place for quite a few
years # but I still # I just couldn’t quite get them streets figured out # so I just had
to keep on walking # slow walking # and just try to figure out where # where I had
to go # you see. But uh it took a good part of the day to get it figured out.
(4) IV: So what did you do to find your way back # did you ask someone or call?
(5) P: I did talk to the odd one # like you know # but they were uh kind of uh new too #
like the ones I were # was talking to and uh so they didn’t really give me any help
as far as the directions were that I wanted to go.
(6) IV: Umhum.
(7) P: I wasn’t a bit worried about it.
(8) P: I just thought I’ll keep on going till I find exactly where to go.
(9) IV: Do you know your address off by heart?
(10) P: Uh it’s at uh # well # the one I’m staying at that’s at her her uh her place here.
(11) IV: Can you tell me what the address is?
(12) P: Well that’s uh # I I should have been able to.
(13) P: I have it written down # too. But
(14) CG: Yeah he has it in his pocket
(15) P: Yeah ( )
The situation (getting lost) might be thought of as generic. Children, espe-

cially in cities, are usually encouraged to memorize their home addresses and
phone numbers and are taught quite early about what to do in case they get lost.
Adults do this learning automatically for new addresses and phone numbers.
That is, people learn early on one or more scripts for resolving ‘being lost’.
Disorientation in space and time is a clinical sign of AD, and the situation in
which people with AD get lost close to home is common (Hirono et al. 1998b).
It is also common for people with AD not to follow generic scripts to resolve
the problem. Here the issues actually appear in what the patient says about his
experience. The area near his home did not look familiar to him – he was not
able to access the relevant episodic experiences to orient himself. His memory
is sufficiently unreliable that he carries his home address on a piece of paper in
his pocket. However, he did not use this, or find someone who could help him
to use it, to get home. He did ask ‘the odd one’ about directions, but he appar-
ently did not persist until he got help. (He may in any case have been unable
to follow offered directions.) The result was that he went for his morning walk
but did not arrive home until late evening. The initial problem then of disori-
entation is here compounded by problems in episodic and semantic memory,
and perhaps ineffective use of script knowledge. This situation may also be
exacerbated by social conditioning which inhibits elderly people from seeking
appropriate help, perhaps because this would require admitting the extent of
their dysfunction.
Ineffective use of script knowledge may be a way of characterizing the typi-
cally poor responses to direct questions about instrumental activities of daily
living in AD, especially those that require some problem solving or calculation
(Royall et al. 2005). Cosentino et al. (2006) report that people with AD dif-
fer from people with behavioural dysexecutive syndrome (a neurodegenerative
disorder in which prefrontal cortical regions are differentially impaired) in that
while the latter have problems recognizing inappropriate sequencing of script
elements (e.g. John got dressed and then had a shower), people with AD are
more likely to have difficulty with recognizing conceptually impossible script
sequences (e.g. John turned the water off and then had a shower) that require
information integration and reasoning abilities which rely on both limbic and
prefrontal systems. To the extent that semantic systems are impaired in AD,
knowledge of scripts, schemas, frames and scenarios may also be impaired.
However, semantic impairments are typically progressive and people lose
subordinate and basic level concepts before they lose superordinate categories.
In discourse terms, this is reflected in patterns such as preserved use of indefi-
nite reference terms, stereotyped idioms and common syntactic structures. It
also may be reflected in preserved interactional and social skills that require
high-order (over-learned and perhaps multiply represented) scripts, schemas,
frames and scenarios but do not make significant demands on online process-
ing and reasoning abilities. Thus, people with AD may know the routines for
talking about the weather (though recent episodic details may be missing or
confabulated) and can repeat the story that they have told many times before,
but may be unable to tell a new story because of the demands it makes for inte-
grating new information into organized event sequences and deciding which
events are the most salient, even when that information is supplied as in picture
description tasks (Bschor et al. 2001).
8.3 Inferencing in discourse

Our ability to make sense of discourse partially depends on our ability to infer
relationships between what is said, what is intended and contexts of situation
and culture. The input for inferential reasoning includes not only linguisti-
cally encoded information, but also information available in the context and
retrieved from memory.
The latter kinds of information may include generic knowledge of scripts,
frames, scenarios and schemas, as well as instantial information from modali-
ties such as vision, hearing, touch, smell, taste, proprioception and affect. The
ability to use all these different kinds of information to produce and interpret
coherent texts requires minimally:
(a) that the information is available,
(b) that information from different sources can be integrated and
(c) that there are sorting procedures for selecting relevant information and
inhibiting irrelevant information.
Availability of information presupposes integrity of linguistic, sensory and
memory systems. Integration of information is mediated by activity in hetero-
modal and prefrontal cortical regions and includes the ability to maintain
and monitor information online. Maintenance and monitoring, together with
selection and inhibition are executive function processes necessary to sup-
port inferencing (Shallice1982; 1988; Baddeley 1998; Royall et al. 2002).
Executive processes may also involve explicit (conscious) manipulation of
information – what we speak of as thinking, reasoning, planning and judging
(Royall et al. 2002; 2005). Once again, the neurological substrate for inferenc-
ing processes are matters of ongoing research and debate. However, there are a
few areas where there is, if not consensus, at least an observable trend toward
convergence. We sketch some of these and then discuss the ways in which
inferencing has been addressed in the discourse of people with ASDs.
First, it would appear that there is no localized ‘inference processor’ – rather,
as the minimally necessary elements for inferencing above suggest, even the
simplest inference may require the recruitment of large-scale networks. Second,
there is convergence on the fact that the prefrontal cortex plays an essential
role in inferencing processes – and that specific regions may be preferentially
recruited for particular tasks. We discussed earlier in this chapter, for instance,
the putative role of the medial orbitofrontal cortex in integrating information
from emotion and memory systems in the evaluation of risk and reward. In
addition to its role in decision making and risk assessment, the orbitofrontal
circuit is associated with inhibition of socially inappropriate behaviour and
perhaps monitoring of events (regardless of expected outcomes). Lesions in the
orbitofrontal cortex result in dysinhibition, and loss of insight and judgement
(Royall et al. 2002).
Similarly, the anterior cingulate, dorsal and ventral prefrontal cortices each
enter into complex circuits which support inferential processes. The anterior
cingulate cortex receives information from the amygdala, orbitofrontal and
motor cortices, and hippocampus, and appears to be specialized for monitoring,
error correction and initiation. People with lesions in this region have difficul-
ties initiating activities, reduced emotional responses and may have attentional
problems (Royall et al. 2002; Rushworth et al. 2007). The dorsolateral pre-
frontal cortex is linked in a complex network to the basal ganglia and parietal
regions, and is centrally involved in verbal and spatial working memory, plan-
ning, goal selection, hypothesis generation, sequencing and set shifting, self-
awareness and self-monitoring. Lesions are associated with deficits in these
areas (Cabeza and Nyberg 2000; Royall et al. 2002; Alvarez and Emory 2006).
The left ventrolateral prefrontal cortex normally has dense bi-directional links
to the temporal lobes and hippocampus and is associated with semantic selec-
tion, working memory and episodic encoding and retrieval (Nobre et al. 1999;
Cabeza and Nyberg 2000; Petrides et al. 2002) as well as more generally with
linguistic processes (see Bookheimer 2002 for review).
Finally, the functional roles of the more medial aspects of both dorsal and
ventral prefrontal cortex are the subjects of very active research programmes
in relation to their roles in inferencing and discourse comprehension (see Ferstl
and von Cramon 2002 for review). The dorsomedial prefrontal cortex has been
associated with reasoning and evaluation, and may be engaged in internally
driven evaluation processes. Such processes are potentially self-reflexive and
can include affective content which has led some researchers to associate
activations here with emotion processing and theory of mind (e.g. Fletcher
et al. 1995; Frith and Frith 2003). However, comparable activations have been
observed in inferencing tasks unrelated to theory of mind suggesting that this
region is more generally involved in making inferences associated with coher-

ence (e.g. Ferstl and von Cramon 2001; 2002; Ferstl et al. 2005; Seiborger
et al. 2007; Ferstl et al. 2008 for discussion). In contrast, Ferstl et al. (2005)
found that responses to emotional information about a protagonist’s feelings
in simple narratives activated the ventromedial prefrontal cortex bilaterally
along the supraorbital sulcus and the left amygdaloid complex. Activation of
the ventromedial prefrontal cortex does not appear to be replicated by other
inferencing tasks and has been associated with emotion processing, decision
making and empathy and so may be linked to the ability to infer mental states
in others (Bechara et al. 2000; Luan et al. 2002; Ferstl et al. 2005; 2008). One
non-frontal region should also be mentioned in this context: there is robust
evidence, primarily from fMRI studies, suggesting that the anterior temporal
lobes bilaterally are involved in discourse comprehension. Ferstl et al. (2008)
suggest that this activity is evidence of integration of linguistic and episodic
information into a gist of text meaning. Frith and Frith (2003) suggest instead
that this activity reflects script retrieval. Given the widely varied text types
and tasks in which this activation appears, the latter suggestion seems the less
likely of the two.
These circuits are not ‘discourse processors’ or ‘inference generators’ per
se. Rather they each contribute components to processing discourse or other
data as input and to generating inferences necessary for coherence. Within
inferencing, which systems are most engaged may depend on the type (e.g.
inductive/deductive), phase (e.g. evaluation/conclusion) and complexity of
the inference process examined, and may also be affected by the ideational or
affective content and media related to the inference.
For example, Reverberi et al. (2007) have identified a left lateralized fronto-
parietal network that is activated in generating simple deductive conclusions.
They make the point that variability in activations across studies may be associ-
ated with type and stage of reasoning. They suggest that propositional reasoning
may differ from relational reasoning. Such a dissociation would explain the left
frontal lobe (BA10) and right basal ganglia activations observed in Fangmeier
et al. (2006) whose study investigated relational reasoning. Similarly, patterns
of activation during evaluation of premises may differ from concluding activa-
tions, and more complex arguments may result in regionally distinct as well as
more widespread activations (Noveck et al. 2004; Monti et al. 2007).
Kuperberg et al. (2006) identify a more widespread temporal/inferior pari-
etal/prefrontal network associated with constructing causal inferences based
on intermediately related (as opposed to obviously or unrelated) three-word
prompts. They interpret their findings consistently with the above suggestions
in that they see the temporal engagement as involving stored semantic infor-
mation, inferior prefrontal regions are activated for retrieval and selection, and
posterior dorsolateral prefrontal regions may be involved with maintenance and
manipulation of new information for integration and consolidation. They also

suggest that activations in superior medial prefrontal regions may be involved
in directed search for sequence relations relevant to inferencing and that acti-
vation in the right temporal and inferior prefrontal regions may be associated
with monitoring for incoherence.
Chow et al. (2008) examine fMRI correlates of inferencing in a reading task.
They contrasted reading for understanding with reading with explicit instruc-
tions to predict an outcome for sentences such as ‘as the plane approached
the cliffs the passengers began to scream’. The target prediction is the caus-
ally related item ‘crash’. In contrast with a baseline task (reading nonsense
words), the reading task activated a network in the left hemisphere including
the posterior superior temporal sulcus, anterior temporal lobe, the temporal
parietal junction and inferior frontal gyrus. The predictive reading task resulted
in additional activity prefrontally in the left anterior prefrontal cortex (BA
9/10), left ventral inferior frontal gyrus (BA 47) and left dorsolateral inferior
frontal gyrus (BA 44/45). They suggest that these regions each have specific
roles – respectively evaluation, semantic retrieval and integration of strategic
inferences. Using dynamic causal modelling (Friston et al. 2003), Chow et al.
(2008) also observed that inferential predictions consistently increased interac-
tions between the left dorsolateral inferior frontal gyrus and the left superior
temporal sulcus and interpret this as evidence of top-down control of activation
and retrieval of lexico-semantic information. Thus, they hypothesize a network
for intentional predictive inferencing in which ventral, dorsal and anterior pre-
frontal cortical regions guide and integrate information selection from lexico-
semantic systems in the temporal lobe and evaluate it for coherence. Since
the anterior prefrontal activations are seen in the predictive reading task but
not in the contrast task (read for understanding), they suggest that the anterior
prefrontal activations are specifically associated with the intentional nature of
the prediction task.
Den Ouden et al. (2005) also suggest that the anterior prefrontal cortex (BA10)
(together with the right parietal cortex and precuneus) is specifically involved
with prediction or ‘prospective memory’ and furthermore point to differences
in activation patterns depending on what thinkers are thinking about – that is,
in their fMRI study activations varied depending on whether participants were
thinking about their own intentions versus events caused by the physical envi-
ronment. ‘Intentional causality’, where participants were thinking about their
own intended actions, activated a wide network that included the precuneus/
posterior cingulate cortex, posterior medial dorsolateral prefrontal cortex, the
superior temporal sulcus bilaterally and the temporal poles bilaterally. They
associate each of these regions with component aspects of theory of mind, but
point to increased activation in an anterior portion of the dorsomedial prefron-
tal cortex as differentially engaged in intentional causality. Ferstl et al. (2005)
and Ferstl and von Cramon (2007) also describe content specific activations
during text comprehension. However, as noted above they make the point that
the ‘theory of mind network’ activates the same regions as their general pur-
pose discourse coherence network (see also Ferstl et al. 2008).
Studies such as these consistently suggest complex interactions in functional
networks in inference processing – with the type, complexity and context of
the inference all affecting the systems involved. They also show some acti-
vation in the right hemisphere which has been proposed as being especially
important in inference processing and coherence (Marini et al. 2005), as well
as being linked to social and emotion processing. However, it is not yet clear
what roles the right hemisphere plays in inferencing. One possibility is that
prefrontal activity in the right hemisphere may reflect, as Tulving (1983) sug-
gested, an episodic ‘retrieval mode’ that allows the integration of modality
specific information associated with memories and their relation to the remem-
berer. A second view is that the right hemisphere functions to boost process-
ing capacity, when task demands exceed those of the left hemisphere (e.g.
Schacter et al. 1996; Mason and Just 2007). This fits well not only with imag-
ing in healthy populations where right hemisphere engagement appears more
often when processing demands are high (e.g. Wagner et al. 2001), but also in
aging and demented populations where right hemisphere activity appears to
be compensatory (e.g. Dolcos et al. 2002; Grady et al. 2003). Specific roles
have also been suggested for particular regions. For example, Henson et al.
(1999), Shallice et al. (1994) and others (Rugg et al. 1996; Schacter et al.
1996) have suggested that right-sided prefrontal activity is involved in ‘post
episodic retrieval processes’ such as monitoring. Kuperberg et al.’s (2007) ten-
tative proposal of an ‘incoherence detector’ is somewhat similar. It should be
noted that these are not mutually exclusive possibilities: compensatory activity
in the right hemisphere might coincide with an episodic retrieval mode and nei-
ther of these exclude monitoring or other processes. It is also worth noting that
despite all the interest in identifying specific roles for the right hemisphere in
inferencing, theory of mind and discourse processing, recent reviews suggest
that language processing is (at least in the right-handed populations studied)
predominantly left lateralized. Exceptions appear to be for consistent bilateral
activations of the anterior temporal lobes and right lateralized activation of the
ventromedial prefrontal cortex for emotional processing (Ferstl et al. 2008).
Other activations observed in the right hemisphere, associated for instance
with metaphor processing are, so far, inconsistent and might be accounted
for in terms of increases in processing load or episodic memory search as a
response to processing load demands (Wagner et al. 2001; Stowe et al. 2005;
Ferstl et al. 2008).
Just as the neurological picture for inferencing abilities is not complete, so the
discoursal signs of impaired inferencing abilities are still under investigation.
In autism, abilities to construct coherence and infer intentions are actively

being investigated. Investigations are framed within models such that autism
is said to be characterized by inability to imagine the cognitive and emotional
states of others (the theory of mind deficit) which is thought to lead to diffi-
culties in inferring the intentions of others (Baron-Cohen et al. 1985; Happé
1993; Baron-Cohen 1995). Differences in processing style and ability to inte-
grate information from different sources have also been probed as character-
istic of ASDs and have been associated with difficulties in the production and
interpretation of discourse. The characteristic processing style is described in
central coherence theory (e.g. Frith and Happé 1994; Jolliffe and Baron-Cohen
1999a). Studies have addressed these questions globally from model theoretic
perspectives or they have addressed specific aspects of inferential processing
and coherence construction.
For instance, Happé (1993) found that there are problems with relevance
in social communication: participants could answer questions, but responses
were not always relevant to context. She also investigated abilities to under-
stand similes, metaphors and irony and found problems in inferencing. In a
small study, Dennis et al. (2001) investigated a range of tasks involving infer-
encing in discourse and found that children with autism had more difficulty
with implication than with presupposition. They hypothesize that children
with autism make inferences from texts to fixed generic knowledge rather
than using instantial information. Happé (1997) also investigated problems
of inference from instantial situations using a homograph task where children
were asked to select appropriate values for potentially ambiguous words, as
in there was a big tear on her cheek or there was a big tear on her dress
(Happé 1997). The study found that children with autism were more likely to
make errors because they did not integrate the sentence context in producing
their interpretations. These findings were supported in a later study by Jolliffe
and Baron-Cohen (1999a) who found that people with Asperger syndrome
and high functioning autism performed less well than controls on three tasks
requiring use of relevant contextual information, designed to evaluate ability
to interpret text as an integrated whole. On homographs and auditorily pre-
sented ambiguous sentences, people with an autism spectrum disorder were
slower and less accurate than control participants. They also took longer and
made incorrect inferences in an interpretation task that required ‘bridging’
inferences (the inferring of a connection between the current clause and a
preceding clause or passage).
More recently, Norbury and Bishop (2002) investigated inferencing and
pragmatic abilities in typically developing children, children with high func-
tioning autism, and children with specific language impairment and pragmatic
language impairment without autism. In questions about text-connecting and
gapping inferences in a story, children with autism had more difficulty with
inferencing than others. The ‘text-connecting’ inferences required children to

infer a relationship between two sentences (e.g. that ‘juice’ and ‘the drink’,
mentioned in different sentences, referred to the same object). In gap-filling
inferences, children were required to use general knowledge and textually sup-
plied information to infer facts not stated in the text, such as the location of
the seaside based on mention of a swimming costume, sandcastles and water
in a story. Again, the children tended to use general knowledge or personal
experience rather than the information presented in the story to answer the
inferential questions. Norbury and Bishop (2002) suggest that these differ-
ences reflect difficulty in incorporating the story information in order to make
inferences. These studies may appear to contrast with others such as Saldaña
and Frith (2007), who found that sixteen adolescents with autism did not dif-
fer from matched control participants in ability to make bridging inferences
from world knowledge. However, the inferences in the latter study differ in that
they require participants to use generalized knowledge of the world (e.g. that
rocks can be big, and that if you drop them on people, people might get hurt),
as compared with textually supplied information required for inferences in
the Norbury and Bishop (2002) study. (For example, in Norbury and Bishop’s
study, the story mentioned a clock chiming when some participants were at a
pier. Participants were later asked where the clock was. A correct inference
would use the story – ‘the clock was on the pier’. An incorrect inference would
use general knowledge as in the example response from Norbury and Bishop
the clock was in the participants’ bedroom.)
Several studies have looked at humour and found difficulties in ASDs, includ-
ing interpreting ironic jokes (Martin and McDonald 2004), and explaining the
motivations behind story characters’ utterances in ways that are appropriate to
context (Happé 1994; Jolliffe and Baron-Cohen 1999b). Ozonoff and Miller
(1996) found that on tasks measuring humour appreciation, inference and indi-
rect requests in narratives, seventeen high-functioning adults with autism pro-
duced significantly more errors in all tasks than a control group. Error analysis
of responses to humorous stories showed that the people with autism tended to
choose literal or non sequitur funny responses, instead of simply funny or other
responses. Ozonoff and Miller suggest: (a) that people with autism do under-
stand that jokes are supposed to be funny, despite the fact that they have sig-
nificantly more difficulty selecting the ‘correct, straight-forwardly funny’ punch
lines than controls; (b) that the completion of the joke required participants to
reinterpret the preceding story in ways that demand a cognitive flexibility which
people with autism may have difficulty with; and finally (c) that the tendency
for autism participants to choose straightforward story endings or non sequitur
funny story endings may reflect problems in set shifting – that is, the straight-
forward story ending is quite coherent and within the set of the story, while the
non sequitur funny ending does not require reinterpretation of the story.
One part of the emergent picture from these and similar studies is evidence
that people with autism and Asperger syndrome process information differ-
ently, with tendencies to rely on general world knowledge or otherwise given
information (from personal experience or from information that is ‘old’ in the
context) in responding to questions and drawing inferences, instead of using
new information supplied in the instance. Another fairly robust finding in
ASDs is that people often have difficulty with the Gricean maxim of quan-
tity, for example supplying either too little or too much information (Bishop
and Adams 1989; Surian et al. 1996; de Villiers et al. 2007). Loukusa et al.
(2007) associated the type of information supplied with development stage.
In a study investigating the abilities of children with high functioning autism
and Asperger syndrome to respond relevantly, they found both younger and
older children violated the Gricean maxim of quantity by responding to ques-
tions with additional information that created a semantic drift. In the younger
children, extra information was about their own experiences, while in older
children extra information was more general. Loukusa et al. (2007) suggest the
speakers with high functioning autism and Asperger syndrome may not stop
processing after giving a relevant answer.
The converse of this pattern, where too little information is supplied, is illus-
trated in Text 8.2 below (repeated from Chapter 2) in which, Patrick, a twelve
year old boy with autism, responds to questions about movies and Christmas.
Text 8.2 Free Willy

(1) RES: do you have some movies that you like?
(2) RES: what’s the last movie that you saw?
(3) RES: what was it about?
(4) RES: did you see a movie at Christmas time?
(5) CHI: yeah: ## yeah ((faint)).
(6) RES: pardon?
(7) CHI: yeah.
(8) CHI: I did.
(9) RES: what movie did you see?
(10) CHI: Free Willy ((faint)).
(11) RES: pardon?
(13) RES: Free Willy.
(14) RES: was it good?
(15) CHI: good.
(16) RES: I’ve never seen that movie.
(17) RES: can you tell me what it’s about Patrick?
(18) CHI: it’s Free Willy.
(19) RES : it’s Free Willy.
(20) CHI: I watched that show.
(22) RES: did you watch it at home?

(23) RES: hm?
(24) CHI: ## yeah.
(25) RES: did you watch it at your house?
(26) CHI: yeah.
(27) RES: tell me about Christmas Patrick.
(28) RES: did you have a good Christmas?
(29) CHI: yeah.
(30) CHI: I have a good Christmas.
(33) CHI: presents.
(35) CHI: I get : toys.
(37) RES: get toys?
(39) CHI: I : um I # didn’t bring it.
(40) RES: pardon?
(44) CHI: ## yeah.
(45) RES: yeah?
(48) RES: Free Willy hm.
Patrick provides frequent polar responses without elaboration (5, 7–8, 24,
26). He also responds to questions with partial repetitions of given informa-
tion, repeating parts of questions (tautologies (15, 30, 33)), or previously
presented information. So when asked, for instance, what Free Willy is
about he simply repeats the film title (it’s Free Willy), which in Gricean
terms is under-informative. This response and I watched that show (20)
might be delayed responses to the earlier questions about whether he had
seen a movie (2, 4) and what it was (9). The repetition of Free Willy (21)
may be intended to clarify the reference of that show, or to emphasize the
response, or again, it may be a delayed response to the listed earlier ques-
tions which are phrased in a variety of ways. It may also be an attempt to
please his interlocutor; Patrick has already received positive reinforcement
for saying Free Willy several times. Alternatively, it may be a sort of idea-
tional perseveration (Bayles et al. 1987). It is also possible that Patrick does
not understand what is wanted as a response, and finally, if he does, it may
be that he has difficulty organizing his knowledge of the film in ways that
allow him to answer the question with sufficient information – that is as a
script with an ordered sequence of events.
The exchange from lines (31–47) also deserves comment.

(33) CHI: presents.
(35) CHI: I get : toys.
(37) RES: get toys?
(39) CHI: I : um I # didn’t bring it.
(40) RES: pardon?
(44) CHI: ## yeah.
(45) RES: yeah?
Patrick’s response to the question about what kinds of toys he got for
Christmas I : um I # didn’t bring it (39) appears irrelevant in part because
of the pronoun it which doesn’t have a clear antecedent, but also because the
response does not appear to address the question what kind of toys (did you
get) (38). It is possible that the question what kind of toys, which calls for cat-
egorization, is too abstract for Patrick. However, assuming that since he does
reply he is attempting to be co-operative, it may also be reasonable to infer
that his I didn’t bring it refers to the generic situation type of ‘show and tell’ in
which children bring new toys and experiences to share in school. His response
then might be thought of as drawing on his generic knowledge of toy sharing
and in this sense is indirectly relevant – it also highlights the patterns noted in
Loukusa et al. (2007) and others that children with autism may use personal
or general knowledge to respond to questions rather than responding to the
instantial demands of the situation.
Much of the work on inferential processing in discourse has been directed
toward considerations of relevance, as developed, for example, in Grice’s
model (1975) of conversational maxims and extended in Sperber and Wilson
(1986; 1995). Theories of relevance or the Gricean co-operative principle are
predicated on models of ‘normal’ behaviour with the assumption that features
such as truthfulness are controlled by the speaker. The context of pathology
poses a challenge for some of these assumptions. People who confabulate, for
instance, may be quite unaware of the falseness of their assertions, and some
disorders, such as autism, can make the demands for relevance quite difficult.
What such a framework does help with is in pointing out areas of difference
imposed by particular neurological impairments.
Table 8.1. Gricean maxims in Alzheimer’s disease and Asperger syndrome
Relevance (model presup-

poses neurotypical function) Alzheimer’s disease Asperger syndrome
Quality (truthfulness: don’t – Confabulation – Literalness

say things you know to be – Paranoid suspicions
untrue or for which you
lack evidence)
Quantity: too little – Underspecified, deictic – Poor or linear topic
information reference development, terseness
too much information –Repetition, verbosity – Pedantic, specificity and
detail of NPs
Relevance: be relevant – Tangential (can’t track – Difficulty contextualizing
(cohesive and coherent), conversation, so incoherent)
topic sharing assumes – Perseverative ideation: return
mutual interest or at least to same topic despite current
socially conditioned one
behavioural norms that
allow performance of
mutual interests
Manner: be clear, avoid – Poor planning, information- – Difficulty foregrounding
obscurity, avoid ambiguity, processing problems lead and backgrounding
be orderly to fragmented incomplete, information
unplanned ‘disordered’ – Sometimes poor discourse
discourse which may affect planning, though discourse
cohesion and/or coherence can be very orderly
In Section 8.2 we outlined cognitive models involved in ‘top-down’

processing of information. These models highlight generic features that peo-
ple use to make inferences which allow them to interpret what is being said
and decide how to respond to it. In this section we have looked at research on
inferencing in neuroimaging and in the discourse of people with ASDs and
considered some of the ways in which people do not incorporate informa-
tion from models in ways that let them arrive at an appropriate inference or
use it in an appropriate response. We end the section with a presentation of
Grice’s conversational maxims and their applicability to Alzheimer’s disease
and Asperger syndrome. As can be seen from Table 8.1, Asperger syndrome
does not present problems for Grice’s first maxim, Quality or truthfulness,
though other conversational maxims may not be followed. In Alzheimer’s
disease, although observation of turn-taking patterns and responsiveness
suggest that people with AD genuinely try to co-operate in conversation
(Ramanathan-Abbott 1994; Temple et al. 1999), they regularly fail to follow
Gricean maxims.
8.4 Affect
In our discussions of top-down processing and inferencing abilities, the
primary focus has been on experiential aspects of discourse. In this section we
shift focus to affect in discourse. By affect, we refer to the ability to modal-
ize what we say in terms of deontic (obligation) and epistemic (likelihood)
values, to express attitudes and evaluations about what is said, and to express
emotions. Modalization, attitude and evaluation have been described in terms
of interpersonal systems of mediation (Halliday 1994) or interactional sys-
tems (Gregory 2009c; 2009d). These resources have been further explicated
in appraisal theory (White 1998; Martin 2000; Martin and White 2005). In
Chapter 4, we outlined the linguistic elements for expressing modalizations,
attitudes and evaluations in detail. Here we illustrate their instantiations in two
sample texts and describe how emotion is represented in one of the texts.
In the appraisal framework, speakers are assumed to systematically and
continuously orient themselves to others and to what they are saying through
expressions of modalization and positive, neutral or negative evaluation and
attitude. Such expressions can be unconscious or deliberate. Discourse may
also be described as being more or less univocal or ‘monoglossic’ (White
2000). In monoglossic discourse, speakers may signal positive or negative
evaluations and attitudes and express emotions, but they do not modalize in
ways that acknowledge other possible states of affairs or interpretations other
than their own. Discourse which is monoglossic contrasts with heteroglos-
sic discourse which does make such gestures to others. The following text is
a conversation in which an eleven-year-old speaker with autism responds to
questions about a potentially emotional and exciting event, Christmas, with a
marked absence of expressions of affect and no evident attempts to orient to
his addressee except insofar as he does respond to questions. Later in the same
conversation the topic shifts from Christmas dinner to favourite foods and then
to after-school activities.
Text 8.3 Christmas

(1) RES: it’s almost Christmas time Joseph.
(2) RES: it’s December.
(3) RES: are you getting excited?
(4) CHI: yeah.
(5) RES: yeah?
(6) RES: what would you like to get for Christmas?
(7) CHI: a Real Talkin Bubba.
(8) RES: what’s a Bubba?
(9) CHI: a um oh uh uh a real talking joke playing best friend bear.
(10) RES: oh I see.
(11) RES: do they have them in stores?
(12) CHI: uh uh uh uh: the: uh : they have them on tv.

(13) RES: oh is that where you’ve seen them?
(14) CHI: yeah.
(15) RES: umhum.
(16) RES: and what did your mom ask for something for Christmas?
(17) CHI: um um he a.
(18) CHI: no no.
(19) CHI: she didn’t.
(20) RES: no?
(21) RES: are you going to get her a present?
(22) CHI: uhhuh.
(23) RES: do you know what you’re going to get her?
(24) CHI: a Real Talkin Bubba.
(25) RES: oh : .
(26) RES: you’re going to get your mom one too?
(27) CHI: yeah.
(28) RES: so that’ll mean you’ll have two at your house.
(29) CHI: yeah.
(30) RES: wow!
(31) RES: that’ll be nice.
(32) RES: so what do you do on Christmas day Joseph?
(33) CHI: I play games.
(34) RES: umhum.
(35) RES: and do you have some friends or family in?
(36) CHI: no.
(37) CHI: I don’t.
(38) RES: no.
(39) RES: do you go somewhere for Christmas?
(40) CHI: I don’t go anywhere.
(41) RES: no?
(42) RES: hm.
(43) RES: do you have a turkey on Christmas Day?
(44) CHI: uh : .
(45) CHI: I do eat turkey.
(46) CHI: um I do eat turkey.
(47) RES: do you like turkey?
(48) CHI: umhum.
(49) RES: I do too.
(50) RES: I like turkey and dressing and potatoes.
(51) CHI: yeah.
(52) RES: what’s your favourite food Joseph?
(53) CHI: ah ah my favourite food is rice.
(54) RES: is which?
(55) CHI: my favourite is my favourite food is turkey.
(56) RES: turkey?
(57) RES: mm.
(58) RES: what else do you like what else do you like eating?
(59) CHI: nothing.
(60) CHI: nothing.

(61) RES: mm.
(62) RES: what do you do when you go home from school Joseph?
(63) CHI: ah : I play games.
(64) RES: what kind of games?
(65) CHI: every game.
(66) RES: mm.
(67) RES: can you tell me one of the games that you have at your house?
(68) CHI: I don know what is the games.
(69) RES: hm.
(70) RES: do you have board games or card games?
(71) CHI: uh uh I play board games.
(72) RES: mm.
(73) RES: what board game do you play?
(74) CHI: I pay I pla # I play a card game.
(75) RES: umhum?
(76) RES: and do you watch television?
(77) CHI: umhum.
(78) RES: what’s your favourite show on television.
(79) CHI: a card a Real Talkin Bubba.
(80) RES: oh I see.
(81) RES: ((laughter)).
(82) RES: are they on television a lot?
(83) CHI: umhum.
The interviewer unsurprisingly asks all the questions. In doing so, she sup-
plies prompts that might elicit some expression of emotion or attitude. For
example, she asks Joseph are you getting excited (3) and expresses positive
attitude (wow that’ll be nice (30–31)) about having two talkin’ bubbas at his
house. (A talkin’ bubba is his desired Christmas present and also apparently
what he plans to give to his mother.) Joseph replies only with a yeah to the first
prompt, without marked intonation that might reflect real excitement, and does
not respond to or acknowledge the second evaluation. A speaker who shares the
positive evaluation of the interviewer might be expected to at least acknowl-
edge it (yeah), and often speakers will reiterate or intensify its value (e.g. ‘yeah
it’ll be great’) to show interpersonal alignment. Joseph does not overtly appear
to align himself in relation to the interviewer, except where he acknowledges
her positive evaluation of a turkey dinner (51) and in that he may also try to
provide responses he thinks the interviewer wants. For example, his first vote
for favourite foods is rice, but then he switches to turkey, which is given infor-
mation that the interviewer has already positively evaluated (50–55):
(50) RES: I like turkey and dressing and potatoes.
(51) CHI: yeah.
(52) RES: what’s your favourite food Joseph?
(53) CHI: ah ah my favourite food is rice.
(54) RES: is which?

(55) CHI: my favourite is my favourite food is turkey.
While Joseph readily provides answers to the questions he is asked, he does
not modalize his discourse in terms of possibilities or alternatives. He appears,
for instance, to only want one thing for Christmas (a Real Talkin’ Bubba). He
never says things such as ‘I might play games’ or ‘I think we’ll have turkey’,
and provides unequivocal responses to all questions. His discourse is in this
respect monoglossic (White 2000).
Similarly, topics are raised which might be expected to elicit some sort of
affective response. Joseph is asked what he and his family do on Christmas Day
(32–46), whether he likes turkey (47) and what his favourite food is (52–60). To
all of these he responds with statements without any expression of evaluation,
attitude or emotion. He neither evaluates positively for things he might well
enjoy such as playing games at home, nor does he express disappointment or
boredom about not having visitors or going somewhere at Christmas. The only
place where he appears to offer positive appraisal is in his description of what a
Real Talkin’ Bubba is. In responding to this question he incorporates positively
weighted lexical selections (best friend), and material from the advertisement
(tells jokes) into the name of the toy. Real Talkin’ Bubba becomes a real talk-
ing, joke-telling, best friend bear. The addition of best friend appears to be his
own, but may have been part of a television ad.
Joseph does not elaborate his responses. His responses to both questions
about what he does on Christmas Day and what he does after school are identi-
cal (I play games). However, he has real difficulties in classifying the games he
plays (every game), identifying the games (I don know what is the games), and
offers no descriptions. He also has some fluency difficulties (12, 74). It may be
that the apparent monoglossic style and lack of appraisal features are related to
problems with expressive language.
In contrast with the absence of affect we see in Joseph’s discourse, people
with AD often have conspicuous affect, including emotional responses.3 As
noted in Chapter 6, people with AD often also modalize extensively, reflecting
their characteristic epistemic uncertainty. The patient seen in Text 8.4, Julia, is
somewhat different in that she has very mild, and successfully treated, AD.
Text 8.4 Travel plans

(1) P: I’m # booked now to go to Toronto on the first of May for fifteen days
(2) IV: nice
(3) P: yeah so I’m looking forward to that
(4) IV: so you’re planning that trip
(5) P: um hum yes I’m planning that (8) and that’s for my oldest son who works there
at the hospital and my daughter-in-law there is a radiologist. I think I’ve told you
that before. So Sarah # that’s her name # is busy and I had a great visit with her.
Did I tell you about that IV? I was there for Christmas and into New Year yeah I
guess I told you.
(6) IV: this year?
(7) P: yeah oh maybe I didn’t. I don’t know. When did you last come?
(8) IV: I was here before Christmas
(9) P: yes I probably told you I was going and I did go and we had a really pleasant
visit
(10) IV: isn’t that great
(11) P: yes and ….
…
(12) P: oh yes and I’m sure there’ll be other young people and ah let’s see Emily’s
fourteen and Alice is eleven so they’re at a nice age to travel they’ll really enjoy
that get a lot out of it # I’m glad for Alice who needs a change and ( ) too he works
very hard he works for the television company in Toronto and ah ## now where
were we IV?
(13) IV: ( )
(14) P: I’m telling you about things that I’m going to do and then you get a picture and
what I have been doing
(15) IV: so how about are you planning even longer term like after um in the summer
doing something or
(16) P: I haven’t made many plans I don’t think that we’ve made any from the time I
get back # ah # you know ah # I’ll come back from Toronto of course mid May the
fifteenth I guess and ah well I like being here a lot of the summer because I love
gardening and ah you know that’s the time # so when I come back in May I’ll enjoy
you know ah I’ll go with either Carol or Ellen and get some plants and well I usu-
ally plant either the last week of May or early June and if it’s if there’s still a frost
warning as there usually is in early June Ellen helps me and I’ll still have some of
this pine I’ve got over my gardens for the winter that’s for my bulbs
(17) IV: yes
(18) P: she and I had put those in and of course I love ( ) towards the back there and that
will be really nice there’s something so special about the first flowers springtime
(19) IV: our crocuses are up
(20) P: oh are they?
(21) IV: we have a big bed of crocuses and so I picked some yesterday to take to the
hospital for mom and dad
(22) P: oh how nice
Julia monitors information in her discourse and does modalize episodic
details but her discourse is very informative and comparatively lexically dense
with a high proportion of subordinate category references. She orients continu-
ously to the interviewer’s needs for information, expresses positive attitude
and evaluations both about her own and others’ activities and aligns herself
with the interviewer’s positive evaluations and expressions of pleasure. For
example, Julia asks questions of the interviewer to monitor information flow
(Did I tell you about that IV? (5) Now where were we IV? (12)) and only once
asks an episodic checking question (When did you last come? (7)). She also
monitors information flow through modalized statements about the discourse
as in I think I’ve told you that before (5), yeah I guess I told you (5), yes I prob-
ably told you I was going (9) and twice expresses uncertainty as in oh maybe
I didn’t. I don’t know. Once she appears to monitor an interior (unspoken)
question about a participant’s name in So Sarah # that’s her name # is busy (5).
Elsewhere in the discourse, she actively clarifies the reference of phrases that
might be unclear for the interviewer by supplying relevant information in rela-
tive clauses or NP adjuncts as in my oldest son who works there at the hospital
(5) or Alex, my husband, was a … (ellipted from Text 8.4). She also explicitly
comments on her own discourse strategy for the benefit of the interviewer in
I’m telling you about things that I’m going to do and then you get a picture and
what I have been doing (14). In these and similar examples she shows herself
to be actively monitoring information for the benefit of her addressee. It is also
the case that she supplies details about participants and events so that the inter-
viewer can develop a really complete picture of what and who she is talking
about, and share her experience and enthusiasm.
Julia responds to all prompts from the interviewer and aligns herself consist-
ently with the interviewer’s evaluations. So for example, when the interviewer
offers positive evaluation nice (2) of the planned trip, Julia acknowledges with
a yeah, and then elaborates with another positive evaluation (I’m looking for-
ward to that (3)). She contextualizes her anticipation of enjoyment with refer-
ence to an earlier visit in which she positively evaluates (I had a great visit
with her (5) and we had a really pleasant visit (9)). She responds to the inter-
viewer’s evaluation (isn’t that great (10)) with a yes and then a long narrative
about other ‘great’ activities she’s been involved in (omitted from the text).
When asked about longer term plans, she modalizes and hesitates a bit initially,
but then elaborates a specific plan to stay home in the summer because she
likes gardening:
I haven’t made many plans I don’t think that we’ve made any from the time I get
back # ah # you know ah # I’ll come back from Toronto of course mid May the
fifteenth I guess and ah.
This passage contains not only explicitly positive statements about her own
experiences (… I like being here a lot of the summer because I love garden-
ing …(16)) and evaluations of the phenomena which stimulate these experi
ences (… that will be really nice there’s something so special about the first
flowers springtime …(18)) but also again shows her actively engaging and
aligning with the interviewer’s offer of information (our crocuses are up (19))
with another question with marked intonation (oh are they? (20)) signalling
(perhaps delighted) surprise and then positively evaluating the interviewer’s
reported action of taking flowers to her parents with an oh how nice.
In his work on the language of affect, Downes (2000) points out that emotion
shares with language a system of meanings that are categorized. He outlines
resources for experiencing and expressing emotional affect including seman-

tic and lexical resources and indexical or iconic resources. We can see Julia
making use of these resources. For example, it’s expected that people will be
excited before a trip. Not surprisingly then, when Julia projects ahead in con-
versation to her upcoming trip, she expresses affect (I’m looking forward to
it). In this she exploits semantic and lexical resources for affect in order to
show her response to the projected situation. Similar uses of semantic and lexi-
cal resources appear in Julia’s categorizations of her feelings: she uses both
oppositional verbs such as enjoy, like, love and gradable adjectives (glad, nice,
pleasant and special), the latter two with intensifiers (really and so).
Downes also suggests that any linguistic feature or feature combination can
indexically or iconically signal affect in real or imagined situations where emo-
tion is evoked. For example, the overall impression that Julia conveys of being
happy, thoughtful and engaged might be thought of as iconically represented in
her discourse not only through the frequency and repetition of positive evalua-
tive lexis, but also in the amount of information which she volunteers in order
to convey her plans and attitudes to her addressee. This elaboration suggests
her emotional state and attitude. Finally, as mentioned above, Julia’s response
to the interviewer’s early crocuses (as signs of spring) with prosodic features
of increased loudness and rising tone index her enthusiasm.
The affective features of discourse are describable in terms of lexical, syn-
tactic, prosodic and conceptual features which modalize, evaluate or express
attitudes. Cognitive models such as scripts and frames describe generic expec-
tations relative to situations. These models may embed evaluations which
influence people’s judgements in and about instantial situations. (Think of the
script for preparing a tax return!) Clinical investigations, especially of fram-
ing effects, have shown that patients’ judgements and emotional states may be
influenced by positive and negative evaluations attached to these models (e.g
Everingham et al. 2006; Ehrenreich et al. 2007; Berg et al. 2008). Indeed, the
development and use of cognitive therapies to treat affective disorders such
as depression and anxiety are predicated on links between cognitive models
(especially, but not exclusively, of self) and their relation to emotional states
(see Beck 2005 for review).
Here we have shown reduced presentation of affect in one text (8.3) by a
speaker with autism, and a display of positive affect in text 8.4 by a speaker
with Alzheimer’s. In presenting these analyses, we deliberately hedged our
descriptions of possible emotional states in the speakers. What discourse anal-
ysis allows is merely a description of presented patterns. Inferencing beyond
patterns is of course what we all do in our everyday lives. However, in clini-
cal contexts, particularly where speakers’ abilities to express affect may be
compromised by cognitive impairments, such attribution needs to be guided
by neuropsychological and/or psychiatric assessments of emotional states and
by patient and caregiver input. The absence, for instance, of emotional prosody
is not necessarily evidence of detachment or an absence of emotional involve-
ment – it may reflect an inability to use prosodic resources.
Apart from mentioning the debated theory of mind network and the func-
tion of the ventromedial prefrontal cortex in emotional awareness, we have
not attempted to articulate neural substrates for affect and emotion. The few
areas where there is consensus are well described in the literature. There are
some brain regions such as the amygdala, hippocampus and medial prefrontal
cortex which are particularly associated with emotional processing (Mesulam
1998; Phan et al. 2002; Wager et al. 2003; Sergerie et al. 2008). There is also
some evidence, based in part on impressions from neural trauma, that the right
hemisphere may be more involved in emotion and mood. However, the meta-
analyses of imaging studies (just cited) do not support greater right laterali-
zation for emotion processing and, in a review of neuropsychiatric disorders
associated with brain injury, Cummings (1997) makes the point that while
right brain injury is often associated with neuropsychiatric symptoms, most
can occur as a consequence of injury to either hemisphere and ‘only mania and
certain types of personality change are uniquely associated with right brain
damage’ (p. 33). Cummings suggests that the right hemisphere is important in
mood and other disorders because of its role in supporting interaction abilities.
There are many observations which associate capacity for expressing and inter-
preting emotional prosody, facial expression and gesture with damage to the
right hemisphere, as well as with pragmatic difficulties in discourse (Mesulam
2000). The relationships of emotion and affect to neurotransmitters such as
seratonin, dopamine and noradrenaline have also received extensive attention,
perhaps because they are targets for pharmacological interventions (Delgado
and Moreno 2006 for review).
There are a great many more areas where research is underway includ-
ing investigation of relationships between discourse patterns and the neural
instantiation of affect and emotional states. However, the work of correlat-
ing discourse patterns with neurally, or even neuropsychologically identified
or emotional affective states is only beginning. At present, there are some
emergent patterns. Just as it seems that ideational components of concepts
and models appear to have distributed representations associated with their
modal and/or functional values, so emotionally associated words, and indeed
discourses appear to differentially activate neural networks linked to emotion
processing. For example, Beauregard et al. (1997) investigated blood flow
in relation to passive viewing of concrete (bear, elephant, rabbit), abstract
(ego, purity, rumour) and ‘emotionally laden’ (sex, murder, sadness) words.
They found that the emotional words differentially activated orbitofrontal
and medial frontal cortices, compared with abstract or concrete words, or
a baseline image. Similarly, Ferstl et al. (2005) observed activation in the
ventromedial prefrontal cortex associated with participants reading stories

with emotional information about the protagonist’s feelings. And Goldin
et al. (2005) observed similar activations in anterior medial prefrontal cortex
as part of a network associated with watching sad or amusing film clips. This
region has been associated in many studies with different aspects of emo-
tion processing – including imagining emotional events in the distant future
(D’Argembeau et al. 2008) and viewing emotionally laden negative pictures
(Ochsner et al. 2002). In a meta-analysis of eighty studies, Seitz et al. (2006)
associate the region with empathy, but articulate empathy as a collection of
processes which are said to be represented in anatomically distinct subre-
gions of the medial prefrontal cortex. The salient point for present purposes
is that this region does seem to be involved in networks which process infor-
mation from a wide variety of emotional stimuli – pictures, films, thoughts
about the future, stories and words.
There is also a growing body of imaging evidence which distinguishes acti-
vation patterns for different emotions. For example, in the study mentioned
above by Goldin et al. (2005), amusing films differentially activated sub-
cortical regions associated with positive emotional states (love, happiness)
and memory whereas the sad films differentially activated parts of the visual
object processing path as well as the amygdala. Two other areas for considera-
tion are emotional intensity (or arousal) and emotional suppression. Dolcos
et al. (2004) studied arousal for positive and negative emotions (valence) in
a picture viewing and rating fMRI task. They found that the dorsolateral pre-
frontal cortex (generally associated with monitoring) was activated bilater-
ally, but the left hemisphere was more active for positive pictures and the
right for negative pictures. Further, the dorsomedial prefrontal cortex (associ-
ated with self-awareness) was more active for more intense emotions, while
the ventromedial area (associated with (self-)reflection) was more active for
positive emotions. Finally, in an fMRI study designed to address the effects
of reappraisal, Oschsner et al. (2002) found that appraisal activated a system
that included the amygdala and orbitofrontal cortex. Reappraisal deactivated
these regions but increased activations in the dorsal and ventral left prefron-
tal cortex and the dorsomedial prefrontal cortex as part of a more extended
network.
The studies described are not intended to create a complete picture of
emotion processing in relation to discourse, nor to outline the neural basis
for emotion processing – rather, they are intended to make some more mod-
est points that seem to be emerging from current work on emotion. First, the
prefrontal cortex seems to function, as Mesulam would put it (1998) like a
hub, not for emotions per se but for processing emotional responses and the
information that prompts them. Second, specific subregions of the prefrontal
cortex appear to be involved with particular parts of this processing. So, as
with inferencing, the dorsolateral prefrontal cortex monitors, the dorsomedial

region engages for reflection and so on. Third, the prefrontal areas form parts
of larger networks that engage cortical and subcortical regions associated with
memory, emotion and attention. And just as there is no ‘inference generator’
or ‘coherence generator’, so there is no single ‘emotion processor’, or genera-
tor for that matter. Instead, the pattern is of a large-scale functional network,
the components of which are differentially engaged according to the content,
valence and value of the emotional experience. Finally, the overlap between
networks engaged in studies of emotion processing and studies of inferencing
and discourse processing suggest at the very least that there is a lot of interest-
ing work to be done.
8.5 Samples analysis

We end Chapter 8 by considering some of the situational models we have
explored in reference to a text produced by Will, a 7-year-old speaker with
autism. Here he is talking about construction sites in a context where there are
toy trucks.
Text 8.5 Construction sites

(1) RES: And then what would it do?
(2) CHI: It would move along.
(3) RES: Where would it move along to?
(4) CHI: The construction site.
(5) RES: The construction site.
(6) RES: And what would it do there?
(7) CHI: Pour cement.
(8) RES: Pour cement?
(9) CHI: Yes.
(10) CHI: And then it hardens.
(11) RES: And when it hardens what happens?
(12) CHI: Then it dries.
(13) RES: It dries.
(14) RES: Uhhuh.
(15) RES: And what does it make?
(16) CHI: Um # and it turns that way instead of coming this way.
(17) CHI: That’s because they are many machines in the construction site.
(18) RES: there are many which?
(19) CHI: Them there are many machines in the construction site.
(20) CHI: Be careful.
(21) CHI: There might be danger.
(22) CHI: Okay.
(23) CHI: Uh okay bye bye!
(24) CHI: That was great.
Will has a script for what happens in a construction site (a cement truck moves
along, it pours cement, cement hardens). When prompted, he offers these
actions (2, 7), and in one case he adds an action, in proper sequence, without a
prompt (and then it hardens (10)). He provides additional generic information
about the truck’s actions (it turns that way instead of coming this way (16)),
though this response is not directly relevant to the question posed by the inter-
viewer. Thus the script provided for pouring cement includes:
• a cement trucks moves along to a construction site (2, 4)
• pours cement (7)
• cement hardens and dries (10, 12)
• cement truck navigates around other machines in the site (16–17).
While Will has some success in generating this script through co-construction,
he also appears inflexible or limited in his script knowledge. When asked what
does it make he responds by continuing to talk about the truck’s activities (Um
# and it turns that way instead of coming this way (16)), even elaborating with
an explanation for these actions (that’s because they are many machines in the
construction site (17)). He stays with the script he knows in response to what
may for him be a difficult or confusing question. There may be any number of
responses to ‘what does it make’ in this context, but it is possible he doesn’t
know what it will make. Alternatively, this may not be what he is interested in.
8.5.1 Scenario and frame

Will’s scenario for AT THE CONSTRUCTION SITE includes a specification
of the setting (the construction site (4)) which he repeats fully (17) and generic
elements of that setting: many machines in the construction site, cement,
danger. He also performs a generic warning one might hear at a construction
site: Be careful. There might be danger (20–21). Interestingly, as matters of
frame and scenario models, the participants mentioned are all mechanical. The
cement truck moves along, but no driver or other animate agent is included
in the event frame. However, the performance of the generic warning and his
subsequent okay bye bye suggest that while he has not directly mentioned any
people as participants on the site, he does know that there would be drivers
and/or other workers there and that they might interact in these ways. So, we
might infer from his imaginative digression that his situational model includes
people.
Though these generic elements are presented in such a way that they cre-
ate a fairly vivid picture, they are not always linked together in predictable
ways. Will’s perspective shift to being in the site with the cautionary warning
Be careful. There might be danger. okay bye bye (20–23) comes without a
signal that would contextualize it. Similarly, his evaluation of his imaginative
digression That was great! (24) apparently refers to his internal experience
rather than being addressed to the interviewer. In these digressions the absence
of contextualization cues may mean he either doesn’t recognize the need to
give the contextual details or the need to stay with the interaction. His conver-
sational schema might actually not include either of these requirements.
Will has resources for expressing affect. For example, he uses increased
intensity to indicate his enthusiasm for an event (then it dries! (12)), and in
saying bye bye. He also positively evaluates in that was great. But he appears
to use these resources without real reference to his addressee. We don’t really
know why he expresses enthusiasm for drying cement or what was great.
Again, it may be that his conversational schema isn’t weighted to use affect to
align with addressees.
9 Modelling information across domains
9.1 Introduction
As we outlined in Chapter 5, the term context may be used to refer to a variety
of apparently rather different constructs including the physical and social envi-
ronments in which communication takes place, common patterns of interaction
that might occur in those settings, as well as meanings made relevant by ongo-
ing events and discourse. Our use of the term context includes these senses,
but interprets them specifically as information available to speakers. This
information is acquired, some of it actively ‘learned’, in interactions which
individuals participate in and is represented in episodic, semantic, linguistic
and other repertoires. Speakers use this acquired and learned information to
interpret what is happening, to shape responses to new, incoming, informa-
tion and to communicate. We use context of culture, context of situation and
phase to describe information relevant for interpretation of particular texts or
discourses. Respectively, these reflect broad patterns of culture associated with
language and dialect variation, situation types which constitute patterns of cul-
ture, and the specific discourses which instantiate culture and situation. We are
interested in instances of discourse because new information is acquired ‘live’,
moment by moment in situations and because examining instances allows us to
observe discourse behaviour which may be clinically relevant and from which
generalizations may be abstracted and related to other evidences of neurocog-
nitive integrity and function.
In the first part of this chapter we present detailed phasal analyses of dis-
course from conversations of people with autism spectrum disorders, and inter-
views with caregivers and people with Alzheimer’s disease and relate these to
contextual models. For the first two texts we also provide feature analyses that
characterize some of the conceptual information and morphosyntactic infor-
mation that informs the phasal analysis. These detailed analyses are intended
for people likely to be directly involved in analysing clinical discourse. We
also generalize and relate these analyses to the top-down models discussed
in Chapter 8. We discuss some of the ways in which cultural and situational
168
Modelling information across domains 169
p atterns relate to neurocognitive function. Finally we consider future directions

for cross-cultural applications.
9.2 Phasal analysis A: The magic of the universe

In the first phasal analysis we return to The Magic of the Universe, introduced
in Chapter 1. In the Magic text, a 15-year-old boy, James, with Asperger syn-
drome has a conversation with a research technician about his favourite video-
game, Endorfun. Recall that primary phases are distinguished from each other
by major shifts in one or more of the ideational, interactional or organizational
functional relationships. Insofar as such shifts change the function of the situa-
tion, they change the situation type. This text has 2 primary phases, a dialogic
phase (lines 1–2 and 11–15), in which James is asked about his favourite game,
and a monologic phase on the topic of the game.
Text 9.1 The magic of the universe

(1) SP1: what’s my favourite what?
(2) RES: your favourite game on the computer.
(3) SP1: well there’s ex # well there’s uh # eh # there’s the there’s this strange unusual
game.
(4) SP1: uh well # there’s a la a computer called an IBM Aptiva comes with games.
(5) SP1: uh # like my favourite is the # is from I is from a: place where there’s a k.
(6) SP1: it’s the game’s about # it’s a it’s about a light bodied cube # k running get-
ting the opposite colour on another light force called endorfun which is spelled
e n d o r f u n.
(7) RES: umhum?
(8) SP1: and uh uh: light bodied cubes flying everywhere.
(9) SP1: and I have the power:
(10) SP1: I feel the magic of the universe.
(11) SP1: And et cetera et cetera et cetera.
(12) RES: is this a game you play by yourself James?
(13) RES: or with a partner?
(14) SP1: just myself.
(15) RES: hm.
(16) SP1: I am really completely good at it.
Each primary phase has two sub-phases. In Phase 1a, there is a clarifying
(echo) question (What’s my favourite what? (1)) and James is asked to identify
a favourite computer game ({what’s} your favourite game on the computer
(2)). In Phase 1b, he is asked about the circumstances in which he plays the
game (is this a game you play by yourself or with a partner? (12–13)) and he
replies that he plays it alone (14). The second primary phase is distinguished
from the first interactionally by the monologic pattern of turn-taking (all
continuing turns by James), with a positive minimal response by the researcher

(umhum?). Ideationally, phase 2 has a description of details and features of the
game in its first sub-phase (Phase 2a: 3–8), and the perspective is third person.
Phase 2a is also characterized organizationally by hesitations (4 occurrences
of uh) and false starts (a la, is the, is from, i, it’s, the game’s about, it’s a, k,
running). In contrast, Phase 2b (9–11) has no hesitation fillers or false starts
but is characterized by a perspective switch to first person (I have the power
(9)) where James describes what he experiences when he plays the game, and
by generic situation specific lexis associated with the game (I feel the magic of
the universe (10)) and lexical repetition of et cetera (11).
In addition, there is a transition (16) which blends elements from both
the first and the second phase (I am really completely good at it). Here the
speaker’s intonation indicates information structure for repeated elements as in
really completely good where really and completely modify good rather than
really being an adjunct of completely. Intonation also disambiguates strange
unusual game in that strange unusual are coordinate (redundant) rather than
contrastive attributes of game.
Phase helps to identify selections that realize generic schemas and scripts
specific to the contexts of culture and situation of the participants. The situ-
ational and cultural parameters for each text are summarized below. However,
phase also presupposes feature analyses since our interpretations of even the
most mundane bits of discourse depend on our being able to relate semantic
features to morphosyntactic and phonological features. In processing discourse
online, we use knowledge of top-down models and features to construct (from
the bottom up) relevant meanings that we associate with semantic information
about situations. In order to illustrate these relationships we present a feature
analysis for this text in Figure 9.1 and discuss the relations between contexts of
culture and situation, top-down models, features and phase.
In Text 9.1, and indeed in all these texts, the temporal and geographical
provenances of the context of culture are 1990s Canadian. Social and individ-
ual provenances vary. In this text, the researcher is an adult woman with terti-
ary eduction, whereas the research participant is a teenage boy with Asperger
syndrome. Attributes such as age, gender, social class, education and ability
as well as role relationships within a context of situation affect power and
social distance relations between participants. Here, at least by virtue of her
age (adult) and education, the researcher is the more powerful speaker relative
to the research participant and, given their contextual roles, the default value
for social distance is ‘far’.1 However, speakers may do things to shift default
values, either deliberately because of their communicative goals or they may in
some cases be unaware of the default parameters for social distance and power.
The context of situation involves a characteristic type of research activity with
this group which is for researchers to engage in and record conversational
interactions with participants. So the mode of communication is face-to-face

talk and the topic, ‘favourite game’, has potential since it refers to an experi-
ence that James knows well. James’ understanding of the researcher’s goal is
not known.
Interactionally, the role relationships between researchers and study partici-
pants are relevant in terms of social power and distance. We noted that within
the context of culture, the researcher’s social provenance gives her greater
power relative to James and the default social distance is far. The generic func-
tion of the researcher as interviewer interacts with and compounds features of
provenance insofar as it is expected that the researcher will ask questions to
guide the talk and the research participant will try to respond to them. However,
since another explicit goal of the interaction is to gather information about
the conversational skills of participants in unstructured settings, the researcher
may try to minimize her position and role. She lets James develop his topic
and offers a minimal response as supportive feedback. She also asks a question
designed to allow him to keep talking about the game. That is, in conversation
analysis terms, she lets him dominate the talk.
The scenario of medical (psychiatric) research interviews is underspecified,
but generic features include:
• locations such as schools and clinics, as well as home visits
• there will be some sort of recording strategy used (audio, audio-video,
other)
• there may be prompts, task materials, games or toys
• there is usually some attempt to ensure the comfort of the research partici-
pant in the interview setting by providing seating, tables and so on.
This part of this interview takes place on a break at school and is audio-
recorded.
In Text 9.1, the two primary phases might be thought of in terms of the
framing relevant to the overall generic situation. That is, the talk takes place
in the context of a research interview so the researcher asks a question likely
to elicit conversation and James responds with an extended monologue. James
clearly has a frame for playing computer games. Games have themes, names,
participants, action, one or more players and different skill levels. He offers
information about all of these including his evaluation of his own abilities. The
sub-phases within this second phase reflect James’ knowledge of the computer
game and his script for participating in the game.
9.2.1 Schemas
As suggested above, the researcher’s understanding of her role in the generic
situation presumably motivates her questions and the topic. James’ responses
show no overt awareness of the researcher’s goals, but he does effectively

respond to the immediate demands of the interaction. The primary phase
boundaries correspond with these schemas. Schematic knowledge may also be
involved in his negotiation of the perspective shift in the secondary phases of
his monologue. While his presentation of his role in the game is relevant for the
description, the absence of a transition between the game and his role might be
surprising to an addressee. It may be that he does not have a schema for signal-
ling perspective shift before performing it.
9.2.2 Script
The phasal analysis also reflects script knowledge at work in this text. The
researcher clearly has a script for ‘getting the research participant to talk’ and
James has good frame and script knowledge for responding to the particular
question. He gives a detailed, organized and accurate description of the game
showing that he knows it, and knows how to describe it.
The phasal analysis helps to identify patterns of functional selections that
correspond with top-down models such as the scripts and schemas which are
generically associated with the context of situation. However, as noted, these
analyses presuppose feature analyses. The feature analysis for this text appears
in Figure 9.1.
The feature array shows that the broad interactional patterns distinguish-
ing the primary phases are instantiated not just by turn-taking sequences
(where dialogue involves regular exchange of speakers, in contrast with
monologue where one speaker ‘holds the floor’), but also by morphosyntac-
tic features such as the order of elements distinguishing statements (subject
before tensed verb), from questions (subject following tensed verb), and
question types from each other (+/– WH = WH/polar question). Dialogic
turn-taking patterns include options for question/response pairs that we see
in the first primary phase. The speakers (and we as observers) ‘know what’s
going on’ here and how to respond because morphosyntactic features cor-
respond with properties of models for speech functions that characterize the
interactional dimension of the situation. Similarly, the ideational domain
reflects a particular context of culture: computer games are temporally,
socially and geographically situated. The density of information and order-
liness of presentation in James’ description of the game can be shown if
one extracts from the feature array the salient information about the game.
Figure 9.2 is an attempt to give a visual representation of the density and
links in his description.
In the line above each text box in Figure 9.1, the text is given with all false
starts, hesitations and so on, included and coded. There is also some mark-up
for syntactic structure and an indication of focus (underlined) and prominence
1.SPKR 1: [what j ['s i [my [favourite] what j ] [t i [t j ]]?
what 's my favourite what

i j j
[WHj [V [DET [A] WH ] [t [t i ]]]]
PRESENT PERSONAL GRADE

3RD PERSON 1ST PERSON ATTRIBUTE
SINGULAR SINGULAR
POSSESSIVE
THING/EVENT POSITIVE THING/EVENT

PREFERENCE
SUPERLATIVE
IDENTIFIED IDENTIFIER
RESPONDING
QUESTION
2. RES: [your [favourite] [game [on [the computer]]]]
your favourite game on the computer

[DET [A] [N [P [DET N]]]]
PERSONAL GRADE COMMON LOCATION DEFINITE COMMON
2ND PERSON ATTRIBUTE COUNT COUNT
POSSESSIVE SINGULAR SINGULAR
POSITIVE ACTIVITY THING

PREFERENCE ENTERTAINMENT MACHINE
SUPERLATIVE +AGENT
+RULES
+WINNERS/LOSERS LOCATOR
LOCATED
RESPONDING
MINOR (ELLIPTICAL)
Figure 9.1 Feature analysis for The magic of the universe

3. SPKR 1: [there's]FS [the]FS [there ['s [this [strange] [unusual] game]]]
there 's this strange unusual game
[[N] [V [DET [A] [A] N]]]
PRO PRESENT DEMONSTRATIVE GRADE GRADE COMMON

EXISTENTIAL 3RD PERSON SINGULAR ATTRIBUTE ATTRIBUTE COUNT
SING SING
NEAR NOT ORDINARY NOT ORDINARY ACTIVITY

NEUTRAL NEUTRAL ENTERTAINMENT
+AGENT(S)
+RULES
+WINNERS/LOSERS
EXISTENT
RESPONDING
STATEMENT
4.SPKR 1: [there ['s [a]FS [la]FS [a computer 2 [t 2 called [an IBM Aptiva 2 [t 2 comes with games]]]]]]
there 's a computer called an IBM Aptiva comes with games

2 2 2
[[N] [V [DET N [t [V [DET N N [t [V [P N]]]]]]]]
PRO PRESENT INDEFINITE COMMON EN INDEFINITE PROPER PROPER PRESENT COMMON
EXISTENTIAL 3RD PERSON SING COUNT SING 3RD COUNT
SING SING SING SING
THING ACTION: NAME NAME ACTION: ACCOMPANI- ACTIVITY
MACHINE RESULTATIVE COMPANY BRAND MOTION MENT ENTERTAINMENT
+AGENT(S)
+RULES
+WINNERS/LOSERS
EXISTENT
TH: PATIENT EVENT TH:RESULTANT
TH: TRANSFERENT EVENT CIRC: MANNER
CONTINUING
STATEMENT
Figure 9.1 (continued)

5. SPKR 1: [uh]H # [like] [[my favourite t 2] [is]FS [the]FS # [is [from [a: place [where there's a ]]]]] [k]FS INC
2
my favourite t is from a place where there 's a
2 3 3
[[DET [A] t ] [V [P [DET N [WH PRO V DET 0]]]]]
1ST GRADE PRES INDEFINITE COMMON EXISTENTIAL PRES INDEFINITE

PERSON ATTRIBUTE 3RD PERSON SING COUNT 3RD PERSON SING
POSSESSIVE SING SING SING
SING
PREFERENCE [GAME] LOC LOCATION LOCATION
POSITIVE
SUPERLATIVE
LOCATED SOURCE
LOCATED LOCATOR
EXISTENT: 0
CONTINUING
STATEMENT (INCOMPLETE)
6. SPKR 1: [it's]FS [the game's about]FS # [it's a]FS [[it] ['s [about [a light bodied cube # k [running ][getting [the opposite colour [on another light force [called [endorfun [which is spelled e n d o r f u n]]]]]]]]]]]
it 's about a light bodied cube running getting the opposite colour on another light force called endorfun which is spelled endorfun
2
[[N ] 4 4 4 5 6 6 12 12
[V [P [DET A N [[t ] t V] [[t ] t V [DET [A] N] [P [DET N N [[t ] t V N [[WH ] V V N]]]]]]]]]
PRO PRES INDEF GRADE COMMON REL TENSE ING REL TENSE ING DEF ?GRADE? COMMON INDEF COMMON COMMON REL TENSE EN PROPER PRO PRES EN COMMON
IMPERS 3RD SING ATRRIBUTE COUNT ASPECT CONTINUOUS ASPECT CONTINUOUS ATTRIBUTE COUNT COUNT PRO PASSIVE PASSIVE RELATIVE 3RD PERS PASSIVE COUNT
3RD SING PERS SING ASPECT ASPECT SING SING SING NOUNS
NEUTER SING PASSIVE
NOM
ACTIVITY SMALL THING ACTION ACTION POLAR PROPERTY ALTER- PROPERTY THING ACTION NAME ACTION THING
ETC. WEIGHT SHAPE MOTION TRANSFER CONTRAST NATIVE ? ABSTRACT RESULT BRAND VERBAL LETTERS
BODY MADE GEO- ON FOOT +POWER ATIVE RESULT-
OF LIGHT METRIC FAST ATIVE
ATTRIBUAND ATTRIBUTE
AGENT EVENT AGENT EVENT TH: TRANSFERENT GOAL: LOCATION
TH: EVENT TH: RESULT
PATIENT
TH: EVENT TH:
PATIENT RESULT
CONTINUING
STATEMENT

7. RES: mmhm?
mmhm
RESPONDING
8. SPKR 1: [and [uh]H [uh:]H [[light bodied cubes] [ t [flying [everywhere]]]]]

and light bodied cubes flying everywhere
4
[C [[A N] [t [V [PRO]]]]]
ADD GRADE COMMON TENSE ING INDEFINITE
ATTRIBUTE COUNT ASPECT CONTINUOUS LOCATION
PLURAL ASPECT
SMALL WEIGHT THING ACTION INCLUSIVE
BODY MADE OF SHAPE MOTION
LIGHT GEOMETRIC IN AIR/SPACE
AGENT/TH: TRANSFERENT EVENT CIRC: LOCATION
CONTINUING
STATEMENT
MINOR (ELLIPTICAL)
9. SPKR 1: [and [[I] [have [the power]]]]
and I have the power

7 8
[C [[N] [V [DET N ]]]]
ADD PRO PRESENT DEFINITE COMMON
PERS COUNT
1ST PERSON SINGULAR
SING
NOMINATIVE
DEF POSSESSIVE ABSTRACT

SPEAKER RELATION FORCE
POSITIVE
POSSESSOR STATE POSSESSED

CONTINUING
STATEMENT

10. SPKR 1: [[I] [feel [the magic of the universe]]]
I feel the magic of the universe
[[N]7 [V [DET N9 P DET N]]]]

PRO PRESENT DEF COMMON DEF COMMON
PERS NON-COUNT COUNT
1ST PERS SINGULAR
SING
NOMINATIVE
DEF MENTAL ABSTRACT LOCATION

SPEAKER REACTION PROCESS INCLUSIVE
POSITIVE
EXPERIENCER STATE PERCEPT
CONTINUING
STATEMENT
11 SPKR 1: [and [et cetera] [et cetera] [et cetera]]
and et cetera et cetera et cetera
[C [C PRO] [C PRO] [C PRO]]
ADD ADD ADD ADD
THE SAME THE SAME THE SAME

EVENTS EVENTS EVENTS
STATES STATES STATES
REPEATED REPEATED REPEATED
CONTINUING
NO SPEECH FUNCTION

12. RES: [[isi [ [this] [ t i [a game [[you] play [by yourself [James]? [or[ with a partner?]]]]]]]]]
is this a game you play by yourself James or with a partner
i 4 i 4 7 7 7 10
[[V [[N] [t [DET N [[N ] [V [P N ] [N ] [C [P [DET N ]]]]]]]]]]
PRES PRO INDEFINITE COMMON PRO PRES PRO PROPER INDEF COMMON
3RD DEMONSTRATIVE SING COUNT PERSONAL PLAIN REFLEXIVE SING COUNT
SING SING SING 2ND PERS 2ND PERS SING
NOM SINGULAR
NEAR ACTIVITY SINGULAR ACTION NAME ALT ACCOMPANIMENT HUMAN

ETC. ADDRESSEE ENTERTAIN PERSONAL CO-PARTICIPANT
FIRST EQUAL
MALE
CLASSIFIED CLASSIFIER
AGENT EVENT CIRC: MANNER CIRC: MANNER
RESPONDING
QUESTION
POLAR (ALTERNATIVE)
13. SPKR 1: [just [[myself]]
just myself
[AV [N7]]
DEGREE (SINGLE) PRONOUN

REFLEXIVE
1ST PERS
SINGULAR
SPEAKER
MANNER
RESPONDING
STATEMENT (MINOR)

14. SPKR 1: [[I] [am [[really] [completely] good [at it]]]]
I am really completely good at it

7
[[N ] [V [[AV] [AV] A [P N4]]]]
PRO PRES DEGREE DEGREE GRADE PRO

1ST PERS 1ST PERSON ATTRIBUTE 3RD PERS
SING SINGULAR SING
NOMINATIVE ACCUSATIVE
NEUTER
SPEAKER REAL INCLUSIVE VALUE THING

POSITIVE OTHER
ATTRIBUAND ATTRIBUTE
CONTINUING
STATEMENT
Key for abbreviations and codes

A = adjective, ACC = accusative case, ADD = additive relation, ALT = alternative relation, Av = adverb, C = conjunction, CIRC = circumstance, DEF = definite, DET = determiner,
EN = past participle, FS = false start, H = hesitation, INC = incomplete, INDEF = indefinite, ING = present participle form, LOC = locational relation, NOM = nominative case,
P = Preposition, PERS = Personal (after PRO), PERS = person after 1st/2nd/3rd, PRO = pronoun, RES = Researcher, SING = singular, SPKR = speaker 1, t = trace, V = verb,
WH = interrogative/relative pronoun, # = pause, [:] = lengthening of the immediately preceding sound, 0 = potential but unrealized argument roles/predicates, superscript indices
are for coreference.

NAMING GAME REGISTER
game
light bodied cubes flying
game everywhere
it’s about a light bodied cube

GAME getting the opposite colour on another
computer
light force called endorfun
IBM
endorfun
Aptiva
EVALUATION
I have the power
my favourite I feel the magic of

the universe
strange unusual
I am really completely good at it
Ideational Evaluations Role in game

IDENTIFIED game ATTRIBUTES – strange, unusual, POSSESSOR –I have the power
IDENTIFIED (NAME) my favourite REACTION TO GAME – I feel the
endorfun magic of the universe
CLASSIFIER –computer EVALUATION OF
SOURCE –IBM Aptiva PARTICIPATION IN GAME – I am
PARTICIPANTS IN THE GAME – light really completely good at it
bodied cubes, another light force called
endorfun
WHAT THE PARTICIPANTS DO –
flying everywhere
GOAL OF GAME –it's about a light
bodied cube getting the opposite colour
on another light force called endorfun
Figure 9.2 Visual representation of The magic of the universe logic

(bold face). In the first line of each text box the text is repeated without flu-
ency features, but with focus and prominence marking. The second line gives a
syntactic parse, the third line gives subcategory information about lexical items
while the fourth glosses semantic features (in lexical ‘shorthand’) for the same
items. The next lines provide argument role analyses. The final line in each
text box provides information about speaker turn, speech function, and clause
status (where this is not major and full).
Trying to imagine the demands of this discourse in functional and neurocog-
nitive terms offers some additional insight with respect to James’ performance.
His response requires that he decide what he wants to say. Leaving out the deci-
sion itself, this presumably involves not simply semantic and episodic search
and retrieval, engaging (at least) prefrontal (minimally left inferior frontal
gyrus) and left temporal lobes and hippocampus (Petrides et al. 1995; Fletcher
et al. 2000), but also that, once retrieved, he maintains the information online
(also hypothetically, left inferior frontal gyrus (Stowe et al. 2005) and perhaps
a dorsomedial-superior temporal sulcus loop (Vigneau et al. 2006)), while he
organizes it into a coherently related series of sentences. This organization is
likely to engage several additional regions in the prefrontal cortex as well as
anterior, medial and posterior temporo-parietal areas associated with language
and discourse processing. (See Vigneau et al. 2006 for a meta-analysis and
review of left hemisphere language processes.) Some of these activations, such
as the anterior temporal poles, would typically be bilateral, minimally adding
processing power given the complexity of the recall required for the response.
(See e.g. Ferstl et al. 2008 and Mason and Just 2007 for discourse networks
and Frith and Frith 2003 and Gallagher and Frith 2003 for theory of mind net-
works. See Ferstl and von Cramon 2002 for comparison.)
These networks, or something like them, are needed not only to gather the
ideational information and organize its linguistic representation but additionally,
in order for the discourse to be judged ‘successful’, James also needs to imagine
what his addressee already knows and what she will need to be told in order to
interpret his response and then he has to organize what he says with those needs
in mind. That is, there are layers of processing requiring imagination, planning,
maintenance and manipulation of message that must be co-ordinated with the
linguistic information to produce something that works in its context.
Thought about in these terms, James’ initially dysfluent speech must first be
interpreted in relation to the complexity of the ideational representation that he
attempts – the detail and logical order of the first phase of his response are cog-
nitively quite demanding. That he manages as well as he does suggests that he
has significant cognitive resources available, and perhaps that his dysfluencies
are a result of online processing limitations relative to cognitive load. We do
not know if this is so of course. The dysfluencies could reflect a restricted area
such as retrieval or planning. But these questions can be investigated.
Perhaps of more obvious interest though is his failure to provide a transition

for his shift to first person performance of his role in and experience of the
game. This sort of unexpected shift is not uncommon in the discourse of people
with autism spectrum disorders (de Villiers et al. 2007) and may reflect, as we
suggested, problems with schemas for transitions. Again, this is quite open to
investigation. One possibility is that speakers know the schemas but find the
demands of taking the addressee into account in monitoring and manipulation
of information beyond their online processing capacities (so establishing what
those capacities are is of interest). Alternatively, monitoring and manipulating
behaviour in relation to the needs of an addressee might be specifically dif-
ficult for speakers with ASDs with the result that they either do not acquire
schemas for transition or that they do not use them. Current models for dis-
course comprehension and theory of mind suggest that there is a central role
for subregions of the medial prefrontal cortex in discourse monitoring and in
emotion processing and self-awareness. A number of recent studies suggest
that these same regions are functionally and structurally different in people
with autism spectrum disorders (e.g. Fletcher et al. 1995; Gilbert et al. 2008;
McAlonan et al. 2005; Thakkar et al. 2008). Finding out whether the absence
of transitions in ASD discourse is an effect of such functional and structural
differences would help to explicate relations between neurocognitive function
and behaviour in ASDs.
9.3 Phasal analysis B: My Mom

Text 9.2 is from a medical research interview involving two caregivers, a
researcher and a patient with AD, Cleo, who we introduced in Chapter 1. They
have been discussing the possibility of Cleo coming to harm without people
to help her.
Text 9.2 My Mom

(1) CG1: It’s a concern to us when you go for a shower or to me when you go for a
shower and turn the hot water on and jump in without cooling it off and the same
thing you go for a drink sometime and you turn the hot water on instead of cold
water well that is a concern.
(2) P: I still say that that’s not something that’s that a normal person wouldn’t do.
(3) CG1: Well you could get scalded.
(4) CG: No no hell no
(5) CG: I’m normal
(6) CG: I would never do that.
(7) P: No no ((laughter))
(8) CG: I’m normal <xxx> [>].
(9) CG2: <xxx fingers> [<].
(10) P: My Mom my Mom used to have a: people used to say she doesn’t even put her
you know [if] she doesn’t even bother to put the this here and that there
(11) P: I don’t know

(12) P: she is always taking care of herself
(13) P: she’s always you know
(14) P: and then she’d turn right around and disconnect what I had been talking about
(15) P: and I mean she’s everybody used to think you know I I’ll be lucky if I’m doing
that or I’ll be lucky
(16) P: if I you’re not lucky girl you’ve had it.
(17) P: And ah # I don’t know #
(18) P: what have we got to now?
(19) CG1: Scalding scalding in the hot water
(20) P: No no
The passage of dialogue here is part of a primary phase (1–9) in which one of
Cleo’s caregivers articulates the possibility of harm (scalding). Cleo suggests
that this could happen to a ‘normal person’ (2). The caregiver responds by
claiming normality for himself and denying that he could make such a mistake
(3–8). Cleo appears to agree and after brief incomplete exchanges, she shifts
topic and begins the second primary phase (10–18) as a monologue. It may be
that the topic here is related to scalding if she is trying to link and contrast her-
self with her mother’s self-sufficiency (she was always taking care of herself)
and competence. One possible interpretation of her statements I mean she’s
everybody used to think you know I I’ll be lucky if I’m doing that or I’ll be
lucky is that people used to think that they would be lucky if they were like her
mother. However, Cleo makes other statements that appear to contradict this
inference. For instance, one could infer that Cleo is suggesting her mother was
disorganized and incoherent from the statements people used to say she doesn’t
even put her you know [if] she doesn’t even bother to put the this here and that
there and and then she’d turn right around and disconnect what I had been
talking about. If this is so, then the if you’re not lucky girl you’ve had it could
refer to the possibility that if Cleo is not lucky she will be disorganized and
incoherent like her mother. Figure 9.3 graphically presents these contrasts.
We cannot really resolve the contradictions that appear here because the
frequency of incomplete utterances, false starts, indefinite reference and ref-
erence without antecedent, together with limited lexicalization and marked
modalization are features which make it difficult to identify topic or even a
coherent relationship between the predications that are actually complete. It is
also uncertain how her monologue is related to the preceding topic and this is
in fact apparent both in her final question (what have we got to now? (18)) and
in her caregiver’s reintroduction of the topic of scalding.
If we were to describe this exchange in terms of contextual features and top
down models, we might note that the social provenance differs from Text 9.1 in
that all the speakers here are adults, and so are peers in that respect. However,
since one of the participants is being treated for AD, her competence to inde-
pendently and safely carry out every day tasks is being discussed. The social
she is always taking care of herself
and the she’d turn right around

and disconnect what I had been
talking about
she doesn’t even bother to put the this MY MOM I don’t know
here and that there
And ah, I don’t know
she’s always you know

what have we got to now?
and I mean she’s everybody used

to think you know I I’ll be lucky if
I’m doing that or I’ll be lucky
if I YOU’RE NOT LUCKY GIRL

YOU’VE HAD IT.
Figure 9.3 Visual representation of propositions in My Mom
distance between speakers is also affected by individual provenance insofar

as the caregivers are Cleo’s family members and so they may be described
as being socially ‘near’ with respect to each other, while all of them are dis-
tant from the researcher. (We omit details of family relations although these
of course will affect power and social distance and, in more ethnographically
detailed descriptions, would help to inform the analyses.)
Within the context of situation, the function of the interview as a whole is to
gather information that will aid in the assessment of the patient’s response to
treatment with a cholinesterase inhibitor and to set goals for treatment evaluation
(Rockwood et al. 2002). Input from the patient and caregivers are thus highly val-
ued, and so although the usual minimal scenario supports for research interview
are in place (here, a tape recorder and a set of objectives which the researcher has
explained to the participants), the interview occurs as a home visit and the inter-
viewer does not closely follow a script. The interactions are thus fairly sponta
neous. We see this insofar as the researcher does not intervene in the monologue
and allows the caregivers to respond to Cleo’s request for discourse monitoring
support (what have we got to now). Nevertheless, the ideational direction is guided
by the presence and interventions of the researcher as she has asked the family
about any concerns they may have with respect to Cleo’s ability to carry out every
day activities. The first primary phase (scalding) overtly reflects this.
This phase also suggests the participants have conflicting schemas and
frames. Caregiver’s explicit expression of concern about harm reflects an expe-
riential frame:
• Cleo’s ability to manipulate every day objects is impaired,
• so she could get hurt,
• the cause of her inability is Alzheimer’s disease.
Cleo’s statement (2) that ‘normal’ people might also be at risk of scalding
makes explicit a schematic contrast between ‘normal people’ and people with
AD that is at work in the discourse. Her attempt to include ‘normal people’ as
at risk may be a rejection of the schema. Her caregiver’s subsequent insistence
on his own ability and normality (hell no # no # no # I’m normal. I would never
do that…I’m normal (4–8)), leave Cleo with little room (or dignity). She is by
implication both not competent and not normal. It is in this context that she
shifts the topic to her mother and attempts a comparison.
The second primary phase (10–16), her monologue about her mother, inco-
herent though it is, can be read as motivated by a desire to resist the schema
that has been operationalized. She does not manage the new topic well, but the
strategy of shifting topic when you do not like what is being said and you have
not succeeded in arguing for another view does reflect schematic knowledge
of options within turn-taking models for conversation (Garcia and Joanette
1997).
The details of her difficulty in coherently developing the new topic can
be made explicit through a feature analysis. We present one for the second
primary phase in Figure 9.4. As with Figure 9.1, the line above each text box
in Figure 9.4 gives the text with fluency features included and coded, some
mark-up for syntactic structure and an indication of focus (underlined) and
prominence (bold face). In the first line of each text box the text is repeated
without fluency features, but with focus and prominence marking. The second
line gives a syntactic parse, the third line gives subcategory information about
lexical items while the fourth glosses semantic features for the same items.
The next lines provide argument role analyses. The final line in each text box
provides information about speaker turn, speech function and clause status
(where this is not major and full). The feature analysis for Text 9.2 shows that
Cleo has low levels of basic and subordinate level features as a proportion
of her talk (Nicholas et al. 1985; Smith et al. 2001). There are, for exam-
ple, thirty NP positions in her ninety-nine word monologue referring to seven
different object concepts shown in Figure 9.4. Only three of these concepts
are lexically represented. There are the two basic level items, mom and girl,
and the superordinate people, each of which is repeated once. The only other
object references occur pronominally and include the speaker (pronominal I),
the addressee (pronominal you), and unspecified transferred objects (this and
10. [My Mom]FS/TOP [my Mom used to have]FS [a:]HES [people used to say [she doesn't even put her]FS [you know]CHECK [if]FS [she doesn't even bother to
put theFS this here and that there]]]
people used to say she doesn’t even bother to put this here and that there
[V [V 1 [ V AV V 1 V N PROLC + N PROLC ]]]]}
{[N]2 [[N] [t
INDEF PAST INF PERS PRO PRES DEG BASE INF PRO PRO
COMMON ASPECT 3RD PER NEG DEM DEM
PLURAL SINGULAR SING
FEMININE
NOMINATIVE
THING2 ACTION NOT AFFECT t1 ACTION NEAR NEAR FAR FAR
HUMAN HABIT VERBAL MENTAL TRANSFER
REACTION
AGENT EVENT MESSAGE
1 STATE PERCEPT
EXPERIENCER
1 EVENT THEME GOAL:LOC + THEME GOAL:LOC
AGENT
RESPONDING
STATEMENT
11. [[I] [don't know]]]
I don’t know
3 V]]
[[N]
PRO [V BASE
PERSONAL PRESENT
1ST SING NEG
NOMINATIVE
DEF
NOT MENTAL COGNITION
SPEAKER
EXPERIENCER STATE PERCEPT: 0
CONTINUING
STATEMENT
Figure 9.4 Feature analysis for My Mom

12. [[she] [is always [taking care of herself]]]
She is always taking care of herself
{[N] [V AV [V [N [P N]]]]}
PRO1 PRESENT ASPECT ING COMMON PRO1

PERSONAL 3RD SING CONTINUOUS ASPECT NON-COUNT REFLEXIVE
3RD SING CONTINUOUS ASPECT 3RD SING
FEMININE FEMININE
NOMINATIVE
DEF FREQUENCY ACTION ACTION/STATE ?DEF

OTHER OTHER?
AGENT CIRC: TIME EVENT EVENT TH: PATIENT
CONTINUING
STATEMENT
13.[ [she][’s always]]IU [[you] [know]]CHECK

She ’s always you know
[[N1] [V [AV]]] [[N] [V]]
PRO PRES ASPECT PRO PRES
PERSONAL 3RD SING PERSONAL PLAIN
3RD SING 2ND
FEMININE NOMINATIVE
NOMINATIVE
DEF OTHER FREQUENCY MENTAL COGNITION
NO ARGUMENT ROLE CIRC:TIME EXPERIENCER STATE PERCEPT : 0
CONTINUING
NO SPEECH FUNCTION

14. and {then [[[she] ['d turn right around] and [disconnect [what I had been talking about]] ]}
and then she ’d turn right around and t3 disconnect what I had been talking about
1
C {C [[[[N] [V V AV P]] C [[t 3] [V [N4 N3 V V V P t4]]]]]}
PRO MODAL BASE DEG BASE WH PRO PAST EN ING
PER PAST PER PERF CONT
3RD SING ASPECT 1ST SING ASPECT ASPECT
FEMININE NOM
NOM
ADDITIVE SEQUENCE DEF HABIT ACTION FULL CIRCLE ADD DEF ACTION MESSAGE DEF ACTION
OTHER OTHER SPEAKER
AGENT EVENT GOAL: AGENT EVENT THEME: SOURCE
LOCATION TRANSFERENT 0
AGENT EVENT RANGE RECIP 0
CONTINUING
STATEMENT
15. and [I mean] [[she's]FS [everybody [used to think [you know]CHECK [I]FS [[I ['ll be [lucky [if I'm doing that]]]] or [I'll be lucky]]]]] (??IU)
and I mean everybody used to think I5/3 ’ll be lucky if I5/3 ’m doing that or I ’ll be lucky
3 5 ? 5/3
[[N V] [[N [V V [[N [V V A] [C [N V V N ]]] [C [N V V A]]]]]
PRO PRESENT PRO PAST INF PRO MOD BASE GRADE COND PRO PRES ING PRO PRO MOD BASE GRADE
PER PLAIN INDEF PER PER 1ST CONT DEM PER AL
1ST SING 1ST SING 1ST SING SING ASPECT SING 1ST S
NOM NOM NOM CONT NOM
ASPECT
ADD DEF MENTAL INCLUSIVE HABIT MENTAL DEF? FUT POSITIVE COND DEF? ACTION ? ALTER- DEF? FUT POSITIVE
SPEAKER COGNITION HUMAN COGNITION SPEAKER? CHANCE SPEAKER? NATIVE SPEAKER CHANCE
AGENT EVENT MESSAGE (of an elliptical?)
EXPER EVENT PERCEPT

ATTRIBUAND STATE CIRC: CONDITION OR ATTRIBUAND STATE
AGENT EVENT RANGE

CONTINUING
STATEMENT

16. {if [ [you] [’re not lucky]] [girl] [[you] [’ve had it]]}
if you ’re not lucky girl you ’ ve had it
3?
{C [[N] [V NEG [A]]] [N3?] [[N3?] [V V N]]]}
PRO PRES GRADE COMMON PRO PRES EN PRO
PERSONAL PLAIN COUNT PERSONAL PERSONAL
2ND SING SING 2ND SING PERF 3RD SING
NOM NOM ASPECT
ACCUSATIVE
NEUTER
COND ADDRESSEE? NOT CHANCE HUMAN ADDRESSEE? FINISH
SPEAKER? POSITIVE FEMALE SPEAKER?
YOUNG
ADDRESSEE
CIRC: CONDITION POSSESSOR STATE POSSESSED
ATTRIBUAND NOT STATE
CONTINUING
STATEMENT
17. and uhH, (pause) [[I] [don’t know]]
I don’t know
3
[[N] [V V]]]
[N3 [V V]]
PRO PRES BASE

PERSONAL NEG
1ST PERSON
SING
NOM
DEF NOT MENTAL COGNITION
SPEAKER
EXPERIENCER STATE PERCEPT: 0
CONTINUING
STATEMENT

18. [what 6 [have i [we [t i [got [to t 6]] now]]]]?
what have we got to now

[WH6 [VI [N [t i [V [P t 6]] PRO-TIME]]]]
PRES PRO EN
PLAIN PER
PERF 1ST P
ASPECT NOM
COMPLETE SPEAKER ACTION DISCOURSE PRESENT
+ ADDRESSEE TOPIC
AGENT/THEME EVENT GOAL6 CIRC: TIME
CONTINUING
STATEMENT (INFORMATION MONITORING)
Key for abbreviations and codes
A = adjective, ACC = accusative case, ADD = additive relation, ALT = alternative relation, Av = adverb, C = conjunction, CIRC = circumstance, DEF = definite, DEG = degree,
DET = determiner, EN = past participle, EXPER = experiencer, FS = false start, H = hesitation, INC = incomplete, INDEF = indefinite, ING = present participle form,
LOC = locational relation, NEG = negative element, NOM = nominative case, P = Preposition, PERS = Personal (after PRO), PERS = person after 1st/2nd/3rd,
PRO = pronoun, SING = singular, t = trace, TOP = topic element, V = verb, WH = interrogative/relative pronoun, # = pause, [:] = lengthening of the immediately preceding sound,
0 = potential but unrealized argument roles/predicates, superscript indices are for co-reference.

that). There is additionally, an expletive (empty) it, relative and interrogative

uses of what, and personal pronouns referring to the participants although
often ambiguously as can be seen from the reference chains in the feature
analysis. For instance, three of the four uses of first person singular I in (15)
could refer either to the indefinite everybody of the main clause or to Cleo and,
similarly, the second person you in (16) could refer to Cleo, to an unspecified
addressee, or to a generic, narrative ‘you’. Girl, used here in second person
address, shares the same ambiguity. These unresolved ambiguities together
with the lack of antecedents for definite referring expressions contribute to the
incoherence of Cleo’s monologue.
The effects of under-lexicalization, referential incompleteness and ambigu-
ity are compounded in Cleo’s discourse by other features. For example, there
are eight false starts, two hesitations, and at least one incomplete predication
(13) in her brief monologue. (The second instance of I’ll be lucky in (15) is
grammatically complete but lucky can take a complement and in context it
seems to be lacking one so it is coded as (IU) – incomplete utterance.) Her
discourse is also modalized with three instances of you know, a monitoring
question in (18) (what have we got to now), two negative mental cognition
predications and an I mean. Most of the fluency features occur at the beginning
of her monologue suggesting she has an initial problem with planning what she
has to say. However, the ambiguities and incomplete utterances, modalization,
under-lexicalization, and references without antecedent are persistent, which
suggests that she has ongoing ideational and executive difficulties. Generously
construed, her If you’re not lucky girl you’ve had it may be an attempt to com-
ment on her own (or someone else’s) situation as we noted earlier. However,
the absence of clear referents together with the fact that she repeats lucky (I’ll
be lucky if…I’ll be lucky) and speaks markedly loudly raises the possibility that
Cleo is motivated here simply by the words just spoken – that she has moved
into stereotypic utterance or is just echoing her own talk.
Cleo’s limited lexicalization together with the other features described result
in a monologue that is genuinely incoherent. We know she is talking about her
mother and that she begins by trying to make some sort of comparison, but
beyond that it is really not possible to sort out what she is saying. Indeed, by
the end of the monologue it seems that she herself has lost the thread of her
discourse – she is not just at a loss for words, she has lost her idea. Because
of this, she fails even in the underspecified goal of changing the topic since
the response to her where have we got to now is a reinitiation of the topic (and
schema) of scalding and her competence.
Cleo’s difficulties at word and clause level reflect not only problems of lexi-
cal access (the this here and that there) and local planning (My Mom my Mom
used to have a: people used to say she doesn’t even put her you know [if] she
doesn’t even bother to put the this here and that there (10)), but also executive
difficulties in selecting, inhibiting and monitoring her discourse for coherence

so that the problems we see in detail at word and clause level redound through
the discourse. Cleo has a schema for getting out of a sticky conversational spot,
and her syntax seems adequate, but she cannot here marshal the executive and
conceptual resources needed to communicate successfully.
Again, thinking about these difficulties in relation to typical functional
impairments in AD can shed light on the possible causes of the difficulties
Cleo has. For example, glucose metabolism (a measure of functional integrity)
is reduced in early AD in posterior cingulate, temporo-parietal and prefron-
tal association cortices (Herholz et al. 2002). The prefrontal deficits include
bilateral ventrolateral and ventromedial regions (Herholz et al. 2002) associ-
ated with language and emotion processing. Moreover, women are more likely
to have reduced prefrontal metabolism (Herholz et al. 2002).2 It seems quite
likely then that the difficulties we see in Cleo’s monologue are not just effects
of a memory deficit or restricted lexical access, but actually correspond to dys-
function in prefrontal cortex associated with executive functions such as selec-
tion and planning and sub-components of verbal working memory systems
(Baddeley et al. 2001; Baddeley 2003; Baudie et al. 2006). Unravelling the
relations between cause and effect could help us better understand and monitor
change in AD.
9.4 Phasal analysis C: Cow parts

Text 9.3, Cow parts, is a semi-structured conversation between a research tech-
nician and a fifteen-year-old boy, Warren, with Asperger syndrome. Thus it
shares social provenance with Text A, the Magic of the Universe, in that the
researcher’s position is +power and the default social distance is far because
she is an educated adult woman talking to a teenage boy with an ASD. The
generic functions of the situation are also similar. This is a medical research
interview in which the researcher’s primary goal is to engage Warren in casual
conversation so his conversational skills may be observed. The interaction
takes place in a setting familiar to Warren and is audio recorded. This text has
three phases, one about edible cow parts, one about summer vacation, and a
short transitional phase (18–21) which blends these experiences.
Text 9.3 Cow parts

(1) SP1: <and> [<] uh I don’t think we’re gonna eat the liver.
(2) RES: ((coughs))
(3) SP1: I don I don’t think we’re gonna eat the kidneys either # I don’t think.
(4) SP1: but we could eat the stomach.
(5) SP1: put them in water.
(6) SP1: and then put different meats and stuff in.
(7) SP1: vegetables.

(8) SP1: cook it up.
(9) SP1: they used to we might be able to y eat the brains.
(10) RES: did you take a holiday this summer?
(11) SP1: this summer?
(12) RES: umhum.
(13) SP1: nope.
(14) RES: no?
(15) RES: two summers ago you went to Europe didn’t you?
(16) SP1: yeah.
(17) RES: that was a nice long holiday.
(18) SP1: once um I saw this show that <this> [>] these two girls went to England.
(19) RES: <umhum> [<]?
(20) SP1: and then they chose out something.
(21) SP1: and it turns out to be cow’s brain.
(22) RES: mm: .
(23) SP1: have you ever tried those before?
(24) RES: no.
(25) RES: I haven’t.
(26) SP1: well I want to.
(27) SP1: I might want to try a brain soup.
(28) SP1: ((laughs)).
In Phase 1 (1–9, 22–28), Warren is describing parts of a cow that might
be eaten. He displays reasonable frame knowledge of internal organs, but has
less information about what is actually edible. His speech is both coherent
and interactive. He maintains and develops topic (1–9) and asks the researcher
about her own experiences with his topic have you ever tried those before? (23).
His questioning of the researcher is consistent with a schematic model for a
casual conversation between peers. In (28) he laughs which suggests that he
may be comfortable in the interactive situation.
In Phase 2 (10–17), the researcher takes five of eight turns, and Warren’s
turns consist of two polar responses without elaboration and an echo question
this summer? (11) seeking clarification. In this phase, the researcher introduces
a new topic which Warren does not take up. First, she asks him if he took a
holiday the previous summer (10). Her attempted transition to holiday fails
perhaps because Warren did not take a holiday. She then moves on to the topic
of a particular summer vacation she is aware of (two summers ago you went
to Europe didn’t you? (15)), and begins to develop this topic (that was a nice
long holiday (17)). In short, we see her functioning in her role as researcher,
attempting to shift the topic from edible cow parts to the more socially neu-
tral ‘summer holidays’. Warren’s response (lines 18–21) is interesting in that
it blends elements from Phases 1 and 2, (went to England and cow’s brain),
creating a transition back to the topic of edible cow parts. He reinitiates Phase
1 with a shift in speech function (Have you ever tried those before?). This
may suggest that Warren does not recognize the background generic scheme
of the research interview, nor the provenance differences but is responding to
the researcher as a genuine peer. Additionally, he may not be aware (or con-
cerned) that his topic might be distasteful for her. Nevertheless, in combining
travel with brain soup to get his own topic back on the floor, Warren shows
conversational skill.
Warren’s renewal of Phase 1 may be a discourse sign of ‘special interests’,
a characteristic pattern of the discourse of people with ASDs. Such shifts to
previous topics are often perseverative, particularly after others’ unsuccessful
attempts at topic changes. For researchers, tracking such shifts may be a way
of identifying and monitoring ideational perseveration in discourse. Its rela-
tionship to, for instance, anterior cingulate function could then be investigated
(Thakkar et al. 2008).
9.5 Phasal analysis D: Squirrels

The final phasal analysis is of a conversation in which a woman being treated
for AD, Doris, has been asked about a visit to friends in Montreal. This is
another medical research interview which occurred as a home visit: it thus has
many of the provenance and situational features we saw in Text 9.2. Here, all
participants are adult women and so are equal at least in these respects. Doris
and her caregiver are family and so are socially near. Their social relationship
to the researcher is far (because of her professional role in the situation) but
this is the fourth home visit so they are familiar with each other. However,
Doris is elderly and is being treated for AD, so it is possible that she will be
treated and/or will act as a less powerful participant. The role of the researcher
is similar to that of the researcher in Text 9.2. Treatment goals were set on the
first visit and she is now helping Doris and her caregiver assess whether Doris
is better, worse or stable on those goals. Text 9.4 arose as part of a response
to questions about Doris’s mood. The caregiver has said that Doris was really
happy about a visit the family made to some friends in Montreal. Here she is
recalling an earlier visit to the same friends.
Text 9.4 Squirrels

(1) P: I know you see I know what nice people they are because I
(2) CG: they sure are
(3) P: stayed with them for a I guess it was a week in Montreal
(4) CG: a whole week yes
(5) P: but not # may not have been a whole week but I almost believe it was # Mary
and I
(6) CG: yes yes that’s right
(7) P: yes
(8) CG: yes # and you knew them anyway from xxx.
(9) P: I knew them anyways I went and stayed there with them
(10) CG: yes
(11) P: and I can see myself swinging on the swing.
(12) CG: is that right
(13) P: ((laughter))
(14) CG: in their backyard
(15) P: in their backyard
(16) CG: yes
(17) P: and uh the the thing that makes me remember it I guess is because uh # a a
squirrel came out along the fence while I was swinging and I never saw that type
of squirrel before. They’re very large to what our squirrels are and they’re much
darker in colour
(18) CG: yes # they have very large black squirrels there
(19) P: mmm
(20) IV: umhum
(21) CG: and if you haven’t seen them before it does seem unusual
(22) P: yes it did to me ((laughing))
(23) CG: and very large grey squirrels as well dear # I didn’t tell you # we didn’t see any
black ones this time but we saw a couple of grey ones.
(24) P: yes # I don’t remember seeing any grey ones but I saw the black ones
(25) CG: the black ones
(26) P: yes
(27) CG: and they go in there and they eat up Myrtle’s garden
(28) P: umhum # yes that’s what she said when I went in complaining to her about the
squirrel ((laughing))
(29) CG: yes
(30) P: and uh she said well dear they won’t hurt you but ((laughter)) they’ll hurt my
garden
(31) CG: yes
(32) P: chase them away she said they’ll tear hurt my garden
(33) IV: How long ago was this visit # approximately?
(34) P: oh I don’t know three or four years
(35) CG: oh I # longer than that
(36) P: longer than that
(37) CG: oh yes # gosh
(38) P: ((giggle))
(39) CG: a good ten
(40) P: ((laughter)) yes I guess it was
(41) CG: a good ten or more
(42) IV: ok
(43) P: I I went uh # I went there to the what was it? I forget ((chuckles))
(44) CG: Rotary Club meetings
(45) P: yes #
(46) CG: or was it the world # Rotary Club
(47) P: Rotary Club meetings yeah um hum
(48) CG: there in Montreal
(49) P: in Montreal # and uh # we went and and stayed uh at uh # uh # Prof and Mrs Smith’s
in oh ah they they lived in # live in Granby # just outside of Montreal. Everyday we
went in to Montreal to the meetings and uh it was wonderful uh # uh Gwen uh the girl
that I went with # and I stayed there and went in and we always went early and got uh
uh seat way up ((laughing))
(50) CG: on the top of the Olympic Stadium I think it was
(51) P: Olympic Stadium yes # it was Olympic Stadium uh we had a wonderful time #
there were people there from everywhere # Rotary Club # I guess ((laughing)) oh
we had a wonderful time # but uh that’s where I first saw those black squirrels
(52) CG: umhum
(53) P: I thought all squirrels were little brown ones
(54) CG: I know # the tiny little things we have here # yes
(55) P: they’re big ((laughing)).
Text 9.4 has two primary phases, both dialogic, but differing ideationally
according to the details being recalled. In Phase 1, (1–10, 33–51), the topic is
the visit, when it occurred and the reason for it. Phase 2 (11–32, 51–55) elabo-
rates on a particular memory of a backyard and squirrels. Overall, Doris is
able to offer good accounts of these events though her caregiver supplies some
details and actively supports her recall. Ideationally, both phases are charac-
terized by mental processes associated with memory and proper names, con-
sistent with the generic situation of trying to recall details of events and their
participants. The discourse has informal features such as laughing and terms
of endearment (dear (23, 30)) though, noticeably, all instances of laughing are
Doris’ (13, 22, 28, 30, 38, 40, 49, 51, 55).
In Phase 1, Doris reports her experience of the visit – its general location and
duration and the people she stayed with. Proper names for people, times and
places appear in syntactically prominent positions reflecting this topic. There is
also positive evaluation, as it was a visit she enjoyed: the people were nice (1),
and the visit was wonderful (49, 51). These features all reflect good schematic
knowledge for reporting a trip and visit.
In Phase 2, Doris reports with vivid clarity a personal experience of swinging
(I can see myself swinging on the swing) and discovering a large black squirrel.
A discussion of squirrels, their attributes and behaviours follows, including
reported events (and reported speech) concerning squirrels in Myrtle’s garden.
This discussion is interrupted when the interviewer asks about the time of the
visit, resuming Phase 1's topic, but is renewed in (51), when Doris again recalls
her first encounter with ‘those black squirrels’: (P: we had a wonderful time,
but uh that’s where I first saw those black squirrels).
In the reconstruction of these experiences, we see a pattern of echoing lexical
words and phrases between the caregiver and Doris which continues throughout
the discourse (CG: you knew them anyway. P: I knew them anyways. (8–9) //
CG: in their backyard. P: in their backyard (14–15) // P: I thought all squirrels
were little brown ones. CG: I know # the tiny little things we have here # yes
(53–54)). Interactionally, the participants positively align themselves to each
other when they do this, particularly when there is no possible doubt about
what is repeated as in (53–54). The pattern of repetition also appears recipro-
cal as does their habit of giving positive feedback. This reciprocity allows the
caregiver to unobtrusively supply details and development for Doris, and to
confirm details Doris offers on her own. For example in lines (3–4) the caregiver
confirms the duration of Doris’ visit by repeating it:
(3) P: I guess it was a week in Montreal.
(4) CG: a whole week yes.
Indeed, the caregiver confirms almost every detail Doris offers, and when she
does not confirm she supplies information as in (43–49), where she supplies
both the name of the event and the place, which Doris confirms and repeats.
(43) P: I I went uh # I went there to the what was it? I forget ((chuckles))
(44) CG: Rotary Club meetings
(45) P: yes #
(46) CG: or was it the world # Rotary Club
(47) P: Rotary Club meetings yeah um hum
(48) CG: there in Montreal
(49) P: in Montreal # and uh # we went and and stayed uh at uh # uh # Prof and Mrs Smith’s
in oh ah they they lived in # live in Granby # just outside of Montreal.
Nevertheless, Doris’ account is good. She recalls proper names of people
(e.g. Annie, Prof and Mrs Smith), names the place she stayed in (Granby) and
provides substantial information in Phase 2 about the squirrel she saw, the yard
and her friend’s reaction to squirrels, including reported speech (she said well
dear they won’t hurt you but they’ll hurt my garden (30)). It is only with the
renewal of Phase 1 in line (33) by the interviewer (How long ago was this visit
# approximately?) that Doris shows difficulty with recalling details of time or
names:
(33) IV: How long ago was this visit # approximately?
(34) P: oh I don’t know three or four years
(35) CG: oh I # longer than that
(36) P: longer than that
(37) CG: oh yes # gosh
(38) P: ((giggle))
(39) CG: a good ten
(40) P: ((laughter)) yes I guess it was
(41) CG: a good ten or more
(42) IV: ok
Here she has misjudged the time elapsed but accepts and confirms her
caregiver’s correction and goes on to offer the reason for her visit, and then
reiterate what was for her the episodically salient part of the experience, big
black squirrels.
Doris manages not only to describe her visit, but also to say why she went,
and why she recalls it so vividly. This is a kind of discourse that people with
mild/moderate AD often can manage quite well, relying as it does on some very
general discourse schemas and recall of personal experiences from a relatively
distant past (more than ten years ago). Her report makes few demands on her
ability to recall more recent experiences, nor to reference or relate information
from an encyclopedic inventory. Moreover, her caregiver provides continuous
support, offering names and times, confirming her statements and positively
aligning with her value judgements. The result is a coherently co-constructed
account of why she was so pleased about the more recent family visit. This
co-construction, supportive as it is, does limit our ability to evaluate Doris’
performance since she does not have to rely exclusively on her own abilities for
lexicalization. Note however, that when she is asked to estimate the time of the
visit she is dramatically wrong. This sort of temporal conflation is very com-
mon in AD, although whether it is linked to the clinically recognized symp-
tom of temporal disorientation and its associated neural substrate is another
question to be answered (Hirono et al. 1998a).
9.6 Summary discussion

We have presented phasal analyses of four texts involving two high function-
ing youths with ASDs and two women in different stages of AD. The phasal
analysis lets us relate local variation in functional selections in instances of
discourse to individual and generic information patterns. These individual and
generic information patterns may then be abstracted as characteristic of the
speaker and his or her negotiation of the situation. The features in Figures 9.1
and 9.4 provide more explicit representations of the information that is used to
do phasal analysis. When a discourse analyst says, for instance, ‘the interac-
tion relationship has shifted’, that is based on a deconstruction of some set of
features signifying turn, speech function, addressee relationships and so on.
Similarly, a shift in ideational relationship will be realized by some change in
the predications, lexical selections, tense and/or aspect choices which the fea-
ture analysis makes explicit. Obviously, we do not expect doctors, nurses and
families to conduct research on discourse phase. What we have suggested is
that phasal analysis is a tool that will be of use to researchers and /or discourse
analysts for descriptions that relate functional selections made within gram-
matical structures to generic models of situation and contexts of culture.
Contexts of culture and situation provide parameters for describing and
interpreting contextual variations relevant to discourse phase. At times, spell-
ing out the generic properties of models results in information that appears
banal – as ‘normal process’ generic features often are. But precisely because
they are the unnoticed and unexamined fabric of our everyday experience, they
inform and shape it and our ability to recognize difference.3 Another way of
looking at cultural and situational norms is through the lenses of conversation
analysis and relevance theory as discussed in previous chapters. Conversation
analysis highlights conventions around interaction and ideation by focusing on
exchange structures and topic. Relevance theory draws attention to (schematic)
default assumptions for the goals of communication and some of the processes
that contribute to the construction of relevance in discourse. We can use the
information from the phasal analyses to relate discourses to such norms and
to provide explicit bases for judgements about normativity and for new gener-
alizations. For example, as outlined in Chapter 3, the model for conversation
between peers assumes there will be a relatively equal distribution of talk and
topic. Who ‘counts’ as a peer depends on both contexts of culture and situation
of participants. Where the context of culture and situation are shared among
speakers, recognition and performance of role relationships relative to talk and
topic may be automatic and unproblematic. Speakers will not only take turns
at talk and share responsibility for topic development, they will shape their
discourse in terms of what they believe the other speaker(s) know and/or need
to know, and in terms of the goals of the interaction.
Conversations and interviews which occur in medical and research contexts
share many of the properties of conversation but also differ. In the present
chapter, Texts 9.2 and 9.4 are extracts from semi-structured interviews involv-
ing patients with Alzheimer’s disease. Part of the role of the researcher here
is to elicit specific information about a patient’s cognitive status and ability
to perform activities of daily living (ADLs), but also to allow participants to
expand freely on these topics and to introduce their own concerns. The role of
caregivers and patients is to supply this information. Topic initiation, sharing
and relevance construction are shaped by these goals. Texts 9.1 and 9.3 are
extracts from research interactions designed to investigate the informal conver-
sational skills of speakers with autism spectrum disorders. While the research
participants in each of the texts conform to expectations in many ways, in
all of the texts, one or more elements of interaction, topic and/or relevance
construction do not conform to expected norms. These normative patterns and
differences for research participants are summarized in Table 9.1.
People with particular disorders may habitually employ different discourse
features to negotiate generic situations, creating new patterns within familiar
contexts of situation. We have illustrated some of these types of differences
for individual speakers with AD and ASDs. For example, people with mild
or moderate AD tend to be good interactionally, usually respecting turn and
trying to participate, but they may also show a level of dependence which is
atypical for adults in an interview situation when they are being asked about
Table 9.1. Normative patterns for conversation and relevance construction
Text 9.1: Magic Text 9.2: Text 9.3: Text 9.4:

of the universe My Mom Cow parts Squirrels
Role
Be responsive    
Contribute    
Relevant  X  
Develop topic  X  (?)
Develop mutual topics na X X 
Topic
Start    
Develop coherently  X  (?)
Complete  X  
Quantity
Provide enough information  X  (?)
Don’t provide too much X na  (?)
information
Manner
Signal shifts in perspective X X  na
Give sufficiently lexicalized (?) X  (?)
reference for coherence
Quality
Truthful    
themselves. The interactional dependence is most likely created by ideational

and organizational difficulties – limited (access to) episodic information,
difficulties with lexical retrieval and reduced executive function resources can
force one to rely on caregivers or other conversational participants for personal
information, discourse monitoring information, and even the words needed to
articulate an experience or desire. This shows up in discourse in features such
as those we have discussed – high proportions of modalization, discourse and
episodic monitoring questions, comparatively low lexical density, ideational
perseveration, difficulties with fluency and coherence, and in some patients
confabulation and other more severe manifestations of cognitive dysfunction.
Notably too, when we see successful discourse of speakers with AD, it may
be the case that caregivers or other speakers are supporting their conversa-
tional participation in precisely these areas. The extent to which such patterns
co-vary with disease phase and treatment is another matter of ongoing research,
as is the matter of their relationships to neural structure and function. In the
next section we sketch, very provisionally, the sort of architecture and neural
components implied by the schemata we have used.
9.7 Discourse and neurocognitive function

Neurological function and dysfunction affect the information available to
speakers, their abilities to relate information from different domains and
consequently the ways in which they use language to negotiate contexts.
Increasingly, these different domains are described in terms of large-scale
functional neural networks. The emphasis on such networks has been partly
created by developments in neuroimaging – as investigations into particular
processes, structures or disorders reveal simultaneous, or at least co-ordinated,
activations in different brain regions (Guye et al. 2008 for review). Moreover,
diffusion tensor imaging is enabling the mapping of long distance fibre tracts
which support such distributed networks (e.g. Catani et al. 2005; Catani and
Mesulam 2008; Catani and Thiebaut de Schotten 2008). However, the idea of
large-scale neural networks has been around for much longer than the technolo-
gies supporting visualization of activation patterns and structural connections.
In the ongoing interactions between biological sciences and developments in
AI, the use of large-scale neural networks as ways of thinking about (and/or
formally modelling) complex dynamic systems has contributed to models of
language (e.g. Hudson 1984; 2007) and neural organization (e.g. Mesulam
1998).
Mesulam (1998) proposed a model in which large-scale functional networks
are linked by transmodal communication centres. These centres or hubs ena-
ble complex multi-modal representations which we think of as ‘cognition’,
‘memory’ and ‘emotion’, which we refer to when we talk, and which consti-
tute the content and texture of consciousness. Among the large-scale networks
are those dedicated to memory, emotion, executive functions and processing
capacities (working memory), and language. Sensory modes are also served by
networks as are attention and awareness, and motion processes. Memory pro
cesses are supported through a network centred in the hippocampal–entorhinal
complex. There are links to unimodal and heteromodal association areas as
well as to prefrontal cortex which are differentially activated depending on
what is being encoded or recalled. Top-down processes of reasoning, plan-
ning, inferencing and so on are mediated by centres in prefrontal cortex as
are working memory, selection, inhibition and monitoring processes. There
are bidrectional links to limbic and other systems including those centred
in the amygdaloid complex associated with emotion processing. Systems
associated with attention and awareness which modulate all other pro
cesses are distributed through the hypothalamus, amygdala and basal fore-
brain. In this framework, language was conceptualized as supported through
networks with centres in the temporal lobe (Wernicke’s area) and (typically)
left inferior frontal gyrus in prefrontal cortex (Broca’s area). Revisions to
the language component of the model have been proposed in the light of
recent imaging findings supporting a more extended language network with
three pathways in fibre tracts linking prefrontal, temporal and parietal lobes
(e.g. Cantini and Mesulam 2008). Exactly how recent findings will affect
our understanding of the role of regions such as ‘Broca’s area’ is not clear,
but interpretations of the imaging data suggest that subregions contribute
different functions to language processing, potentially including a phono-
logical working memory component and one or more subregions associated
with semantic processes such as retrieval and categorization. It is also not
clear yet whether any specific syntactic processing occurs in this region –
a number of studies suggest that activation patterns are equally consist-
ent with working memory demands (e.g. Stowe et al. 2005; Wartenburger
et al. 2004; Fiebach et al. 2005; Vigneau et al. 2006). Similarly, the exact
roles of the parietal area and extended temporal involvement are still being
worked out. However, it seems clear that subregions of the temporal lobe are
extensively involved not only in semantic but also syntactic and phonological
processing, and that parietal areas are involved in processing complex sen-
tences and inferential relations in discourse (Vigneau et al. 2006).
What is most salient for our present purpose, however, is the architecture
in a model such as that proposed by Mesulam. There are no ‘stores’ or ‘word-
hoards’ for lexemes, concepts, frames or schemas, but directories linking differ-
ent kinds of information in unimodal, heteromodal and transmodal association
areas. Thus, beginning a long and auspicious line stretching, at least, from de
Saussure and Hjelmslev at the beginning of the twentieth century to Lamb
(1966; 1998) and Jackendoff (e.g. 2002), in current work a word is regarded
as a set of relations between meanings and auditory, visual or tactile signs.
Similarly concepts do not ‘exist’ in particular neural regions but are distributed
across unimodal and heteromodal areas according to how they are acquired or
learned and their subsequent associations formed through transmodal interac-
tions.4 Accessing a concept or a word is thus a matter of linking its meaning,
value and representation through transmodal gateways such as those proposed
for Wernicke’s area (Mesulam 1998).
We would like to summarize some of the systems and networks we have
discussed so we offer Table 9.2. However we do so very tentatively. The
complexity of the systems involved in discourse processing means that only
the most obvious neural systems can be included. Also, the transitional state
of knowledge with respect to the details of the neural networks associated
with language and discourse processing argue for extreme caution. Until the
publication of Catani et al. (2005), linguists and neurologists simply did not
know about the linguistic role of the secondary fibre tracts in the arcuate
Table 9.2. Types of information and related systems
Normal potential Neurocognitive systems Neural network centres Discourse effects
Context of culture Languages and dialects Semantic memory Medial temporal lobes (MTL) Lecalization
(contexts of user) (Temporal, geographical, Linguistic systems (temporal-parietal lobes) Dialects
social and individual (semantic, morphosyntactic, and
variations) phonological feature patterns)
Context of situation Functional varieties Semantically and MTL and parahippocampal Functional variation in linguistic
(contexts of use) (ideation, interaction, episodically encoded formation, Amygdala, tempo- feature patterns:
medium combination scripts, schemas, parietal lobes Event sequences; discourse structures;
options) scenarios, frames contextualization features; affect and
interactional features
Phase Discourse instances Semantic recall R/L VLPFC, Temporal-parietal

lobes and HC Predication
Episodic encoding (episodic recall more likely Lexical selection
than encoding to engage the
Episodic recall
right hemisphere)
Maintaining D/VLPFC information processing effects

associated with message organization,
Monitoring D/VLPFC, VMPFC, interaction and affect
ACC, OFC
Planning
Attention Thalamus modulates everything
Inferencing Extended language network:

Integration D/VMPFC Coherence
anterior temporal poles, Semantic complexity
posterior temporal-parietal regions Spatial imaging
ACC = anterior cingulate cortex; D = dorso; DL = dorsolateral; HC = hippocampal complex; L = left; M = medial; MTL = medial temporal lobe; OFC = orbito-frontal cortex;
R = right; VL = ventrolateral; PFC = prefrontal cortex
fasciculus. Moreover, up until about the same time, most linguists and neu-
rologists would have confidently assured anyone who was interested that
syntax is processed in Broca’s area – we do not know this with such certainty
anymore. So Table 9.2 must be read as merely a partial summary of our dis-
cussion, intended to be suggestive rather than definitive – and likely to need
revision as new technologies and means to interpret their results change what
we think we know.
Only the final row in Table 9.2 is intended to refer to anything that might
be considered a process and the items listed in the column ‘Neural network
centre’ are not intended to suggest traditional localizations. We do not mean
for instance that ‘semantic memory is in the medial temporal lobe’ or that
‘inferencing happens in dorsal/ventromedial prefrontal cortex’. Rather, we are
simply labelling the neural areas that we have discussed as possible network
centres for the processes to the left of them. The items in the column labelled
‘Discourse effects’ are, similarly, not ‘outputs’ of the named neural centres.
They are discourse phenomena that might be affected by activity centred in
these regions.
Much more confidently we offer the summary of discourse and contextual
features given in Table 9.3. This summarizes the major areas of investigation
in clinical discourse analysis that we have discussed and suggests some inter-
actions. Context of culture and Context of situation identify parameters
for what speakers know. This variation is semiotic: it is relevant to speakers’
semantic and episodic memory, and includes the languages and varieties they
know. Language refers specifically to linguistic structures, features and func-
tions as resources for instantiating semiotic potential. Discourse processes is
self-explanatory. Some of these processes are cognitive, such as inferencing,
metaphor and topic shift. Others are or may be signs of cognitive function,
such as confabulation, hesitation phenomena and modalization. Still others,
such as turn-taking and evaluation, are discourse behaviours which may reflect
cognitive and affective processes and functions. Finally, the Research areas
column indicates some topics we have discussed related to discourse processes,
language and context.
9.8 Intra- and intercultural relevance

We have found the approaches we have presented helpful in the struggle to
understand how discourse may be interpreted in relation to people’s neurocog-
nitive functioning. In closing we would like to suggest that the model, which
was originally developed for ethnographically grounded discourse analysis,
may also be useful as an aid for achieving intra- and intercultural relevance.
There are two ways to go about this. One is to attempt to design tools which are
sufficiently context independent so as to overcome differences in, for instance,
Table 9.3. Summary of discourse and contextual features
Context of culture Context of situation Language Discourse processes Research areas
Provenances Medium
Individual participant Spoken (spontaneous, Phonetic and phonological features Hearing/articulation Processing abilities
demographics as prepared, read aloud) Prosodic features (e.g. intonation Parsing/producing Relevance construction
relevant in terms Written (spontaneous, tonicity) Discourse intonation Syntactic complexity
of age, geographical edited) Syntactic markedness options Inferencing (reference Lexical density and richness
provenance, and social Cohesive options presupposition, implicatures) Executive functions
provenance (class, Discourse schemas Information processing indices
gender, education, (e.g. hesitation, pause, false start,
other) repairs, incomplete utterances)
Interaction
Roles Speech functions Negotiation of speaker roles and Social skills
+/– power Modalization options alignment (selection of speech Conversational schema
+/– social distance Lexis acts, modality, turn, overlap, Theory of mind
Affect (evaluation, evaluative lexis)
attitude)
Emotion
Ideation
Concepts Lexis Lexical selection, Predication, Relevance
Frames Argument and circumstantial roles Metaphor, Metonymy, Irony, Conceptual structure organization
Schemas Morphosyntactic structures Humour, Confabulation, Gist/recall, Semantic and episodic memory
Scripts Topic (initiation, Executive functions
Scenarios development, shift)
levels of literacy, types of experience relationship available, differences in

writing systems and so on. Another approach is to develop culturally specific
neuropsychological tests. Both approaches presuppose articulated descriptions
of the relevant languages, discourse patterns and contexts of situations within
particular contexts of culture. These approaches are not mutually exclusive,
and both are likely to provide somewhat different kinds of information about
clinical populations. For instance, from an epidemiological perspective having
neuropsychological inventories and rating scales with universal applicability
and relevance allows commutability of information across cultural bounda-
ries. This is important in understanding epidemiological patterns and disease
processes. However, neuropsychological inventories and rating scales which
are intended to operate interculterally may be limited in the amount of detail
they provide precisely because they are universal. (As the level of abstraction
increases, the amount of specific information decreases.) The second approach,
of designing culturally specific neuropsychological inventories and scales
avoids this but presents problems of commutability of information. An alterna-
tive, third way is to develop both kinds of tool so that there is commutability at
the level of generalization and sensitivity to culturally specific variations.
Clinical discourse analysis is a novel approach to enduring questions
about relationships between brain and behaviour. The challenges involved
in mapping information from different kinds of study such as neuroimaging,
neuropsychology, cognitive neurology, medicine and discourse analysis are
many. However, the opportunities for achieving new understanding of disor-
ders and diseases in terms of how they alter everyday behaviour are also many
and, if pursued, may help some people live the best lives they are capable of.
Closing remarks
We wrote this book because we thought a description of the work we do could be

helpful for people who want to investigate and understand discourse in clinical
contexts. In modelling comprehensive discourse analyses and showing how such
analyses may be systematically related to aspects of neurocognition, we hope to
have illustrated the usefulness of taking a unifying approach to investigating nat-
ural language behaviour, and particularly extended discourse, in relation to neu-
rocognition. Our practice of hybridizing techniques from functional and formal
linguistic models, from conversational analysis, ethnomethodology, situational
linguistics and pragmatics, as well as from structuralist and semiotic discourse
models, artificial intelligence and neuropsychology and bundling them together
in order to account for all the different aspects of language that potentially con-
tribute to discourse patterns is driven by the need for comprehensive accounts
that are beyond the scope of individual frameworks or discourse models (that we
know). We have suggested ways in which the resultant hybrid methodology for
clinical discourse analysis can be combined with neuropsychological and neu-
roimaging techniques and adapted to different situations and cultures. While we
have been most concerned to address the need for comprehensive analyses, we
have also shown that the amount of detail included in analyses can be adjusted
according to the scope and purposes of particular investigations.
We have further suggested that clinical discourse analysis has at least the
following potential applications. Most broadly, clinical discourse analyses can
be used to identify linguistic and discourse patterns associated with affect and
neurocognitive function generally and so may be useful in researching a wide
range of brain–behaviour relationships. In the context of affective or neurologi-
cal disorder, clinical discourse analyses can be used
• to make explicit the linguistic and discourse patterns that inform clinicians’
intuitive diagnostic judgements;
• to refine and develop diagnostic criteria for particular disorders;
• to develop tools for monitoring and assessing endogenous or therapy-induced
changes in affect and neurocognitive function;
• to inform the development of cognitive and/or behavioural therapies.
207
These broadly stated applications suggest that there is extraordinary potential

for clinical discourse analyses to aid and inform research and clinical prac-
tice in affective and neurocognitive disorders. However, we have also been
at pains to point to limitations. Clinical discourse analysis provides another
set of tools for informing clinical judgements and research – but the value of
analyses will be determined by its relation to data from other sources such as
neuropsychological evaluation and neuroimaging. Clinical discourse analy-
sis is thus by definition work that is most relevantly pursued collaboratively
with multi-disciplinary teams in hospital or research labs. Discourse analysts
are not typically also neurologists or psychologists and discourse analyses
may inform, but are unlikely to replace, other diagnostic practices.
We have not tried in this book to do more than point to the existence and
utility of technologies for text collection, analysis and comparison. There are
many such tools and they improve annually so that it is now quite practical for
instance to work with parsing programmes in big projects instead of manually
analysing syntactic data. However, from our perspective there is real value
in training discourse analysts to be able to do analyses manually in the first
instance so they know intimately the structures, patterns and meanings emerg-
ing from analyses independently of computer mediation. Similarly, when
beginning to work with the discourse of any clinical group, taking time initially
to perform detailed manual analyses can be invaluable. In both cases, analysts
will be better prepared to design and interpret computationally mediated analy-
ses or searches because they will have the experiential base for understanding
the discourse patterns they encounter. What we have tried to do is provide
enough information and models for (future) discourse analysts to develop that
experiential base, and to show relevance and potential of such activity for clini-
cians and other caregivers.
Finally, given that we have pointed regularly to the importance of neuropsy-
chological evaluation and neuroimaging in research, diagnosis and monitoring,
one may wonder whether we are not proposing elaborate training and research
that is or will soon be made redundant by these practices and technologies. To
this we respond with three observations.
First, we have tried to show that detailed descriptions even of very aberrant
discourse of individuals can be informative about their cognitive and/or affec-
tive states and so may have independent clinical and research value. The sort
of information generated from such analyses may complement, for instance,
information about blood flow or glucose metabolism from imaging studies or
executive function reflected in ability to count backwards by sevens or draw
clocks, but it is clearly qualitatively different and potentially more informative
about how, for instance, regional metabolism and executive functions may be
reflected in everyday talk and behaviour.
Closing remarks 209
Second, the sort of comprehensive detailed descriptions of discourse

p atterns for particular neurological disorders has, for the most part, not yet
been undertaken (or published) so the extent to which such analysis may be
of value in areas other than those we examine is simply not known.1 This is
an odd state of affairs given the centrality of discourse to human sociocogni-
tive behaviour. One cause has perhaps been the overwhelming complexity of
the task, combined with what appeared to be limited rewards for the work
involved. However, as we suggested in the introduction, the new technologies
for developing and analysing corpora are in fact making such projects much
less daunting and potentially more rewarding. Also, another possible cause for
the lack of detailed published discourse descriptions of neurological and affec-
tive disorders is a tendency to search for deficits in discourse behaviours rather
than to begin with comprehensive characterizations of patterns. The search for
markers and deficits is quite understandable but, when it doesn’t yield results,
it is possible to conclude that discourse analysis is not useful. Against this,
we have argued that comprehensive analysis and identification of patterns as
illustrated in the preceding chapters are more likely to be rewarding, at least as
a point of departure for investigations.
Third, there are extremely active searches for reliable neuroimaging corre-
lates of Alzheimer’s, autism and a host of other neurological disorders and dis-
eases, and some well-established neuropsychological diagnostic tests. Some
of these searches have already produced results and no doubt others will – so
aren’t clinical discourse analyses likely to become almost immediately redun-
dant? To this we point to the social realities surrounding the diagnosis and
management of neurological disorders. Alzheimer’s may serve as an example.
Most people who get Alzheimer’s will never visit a memory clinic, see a neu-
rologist or have even a CT scan. At best, they will be seen and treated by a fam-
ily physician or other primary caregiver who may not have (or request) access
to neuroimaging or expert neuropsychological advice. This is the case even in
the wealthiest countries where extensive resources are dedicated to developing
the technologies and there is wide access to health care. Given that the number
of people with dementia is expected to triple or quadruple by 2050, it seems
unlikely that access will change dramatically, even if imaging becomes less
costly and more people are trained as neurologists and neuropsychologists.
Thus, while we are excited about the opportunities provided by new technolo-
gies, we see these as research and diagnostic tools that are likely to remain
primarily in teaching hospitals and research institutes. Primary care provid-
ers will continue to need tools that effectively allow on-the-spot evaluation
for diagnosis and monitoring. We hope clinical discourse analysts will help to
develop such tools to support and improve on clinical practice in at least some
of the ways we have discussed in this book.
Finally, as we pointed out in Chapter 9, imaging technologies are unsettling

even some of the most well-founded assumptions about the neural instantia-
tion of language and offering new information and new sources of informa-
tion about what happens in our heads when we talk. But interpreting what
the new information means is another matter and requires neuropsychologi-
cal and other expertise. We believe clinical discourse analysis of otherwise
well-characterized neurological populations can contribute to the interpretive
processes and to our understanding of language and discourse in neurocogni-
tive terms. The project is ambitious and the challenges enormous – but the
resources are available and the needs addressed are real so we hope people will
accept the challenge.
Notes
Notes on Chapter 2
1
The debate revolves around the salience given to executive function relative to
memory disorder. Diagnostic criteria which treat AD as prototypical of demen-
tia typically require memory impairment and impairment in another area such as
executive function (e.g. DSM-IV). However, in dementia arising from causes other
than AD, memory impairment may not be a prominent feature. For discussion, see
for instance Bowler and Hachinski (2003: 7–8), Mesulam (2000) and Royall et al.
(2005).
2
Researchers have differed in their diagnostic and inclusion criteria in epidemiologi-
cal studies of the dementias. The variance in prevalence estimates with the frontotem-
poral and other dementias is partly an artifact of such differences. Knapp et al. (2007)
review the general difficulties.
3
Dubois et al. (2007) have proposed a revision of diagnostic/research criteria to
allow earlier diagnosis of AD and the elimination of the ‘possible’ NINDS-ARDRA
category. They propose a combination of episodic memory impairment plus neu-
roimaging, spinal fluid evaluation, or proven AD mutation in the family. Their goal is
to allow identification of AD before patients are actually demented.
4
Semantic memory refers to stored information that is decontextualized. For example,
speakers will normally know the names and some attributes of common plants and
animals. However, it is unusual for speakers to remember the contexts in which they
acquired such information – they simply ‘know’ that cats are called cats, that they
meow, and so on. In contrast, episodic memory refers to contextualized memory –
rememberers not only ‘know’ about something, but they remember some aspect or
aspects of the context they learnt it in – episodic memories are at least bimodal.
Episodic memory is often associated with personal or autobiographical memory.
However, personal and autobiographical memories may over time become decon-
textualized parts of a semantic repertoire about one’s life. See Tulving (e.g. 1972 for
an early articulation, 2002) for a recent account. While some parts of the networks
supporting semantic and episodic memory systems are matters of ongoing research,
lesion studies, imaging and studies of neurodegenerative diseases all suggest that the
hippocampal complex is central to episodic memory and that the left temporal cortex
is central to semantic memory. (See Cabeza and Nyberg 2000 for an imaging review.)
Neurodegenerative diseases differentially affect memory systems. For example, hip-
pocampal involvement in AD impairs recent episodic memory while damage to the
temporal lobe is a hallmark for semantic memory. (See e.g. Perry and Hodges 1996
for a review.)
211
212 Notes to pages 47–192
Notes on Chapter 4
The systems for speech functions are adapted from Asp 2001; Gregory 1988; 2009c.
1
The conceptual/lexical relations are from Gregory, published in Watt (1990).

2
Notes on Chapter 6
1
This material was presented in poster format as Asp, E., Fisk, J., Klages, J., Kydd,
D., Song, X, and Rockwood, K. Language performance may distinguish mild AD
and SIVD. VAS_COG 2005, Florence, Italy. Details are available from E. Asp.
Notes on Chapter 7
1
Alternatively, or additionally, one can measure lexical richness using Brunet’s Index
or Honoré’s Statistic (Brunet 1978; Honoré 1979). Brunet’s index quantifies the
amount of lexical variation in a text calculated over the text length. Honoré’s Statistic
gives a value for the number of words used only once (Holmes 1992; 1994). These
measures are not sensitive to text length. Lexical richness measures give an idea of
how varied that information is and how much a speaker is inclined to select alterna-
tive lexical items and thus they add valuable perspective to the standard measures
of lexical pattern such as type token ratio or lexical density, particularly in clinical
contexts (Holmes and Singh 1996; Singh and Bookless 1997).
2
The would in this instance is not included in modalization because it is used in its
aspectual sense to refer to habitual aspect rather than hypothesis.
Notes on Chapter 8
1
Some authors include elements of setting and participants in scripts. However, since
these can be handled by other constructs, we find it useful to limit script knowledge to
ordered action sequences that characterize a situation. This is consonant with Schank
and Abelson’s (1977: 41) description of scripts as ‘a predetermined stereotyped
sequence of actions that defines a well known situation’.
2
Chwilla and Kolk (2005) examine ERP responses to three-word lexical items around
which conventional scripts might be constructed (e.g. director, bribe, dismissed) and
find that activating script knowledge presents an N400 effect similar to the one that
occurs for semantic relations. They conclude that script information is accessed and
integrated immediately, that it constitutes a central aspect of word meaning, and that
the patterns of activation (insofar as they are deducible) spread from Wernicke’s area
to more anterior sites in the left hemisphere and then to anterior temporal and orbital
sites in the right hemisphere.
3
Emotional liability is included as a clinical sign of AD (DSM-IV). It refers to a
tendency to display uncontrolled and sometimes extreme or incongruous emotions.
Notes on Chapter 9
1
Hypothetically, the gender difference might have contrapuntal effects – increasing
social distance but decreasing the perceived social power of the researcher.
2
Women also have more white matter disease which is linked to cognitive decline (van
Dijk et al. 2008) and the development of dementia.
Notes to pages 199–202 213
3
In the present discourse analyses, the experiential fields the speakers are absorbed in
are culturally specific, in temporal, geographic and social provenances. The research
activities that include school, clinic or home visits, lengthy interviews and other data
collection designed to investigate specific disorders and treatments are also likely
to be culture specific and generate their own generic situations. Even the disorders
themselves may be culture specific in several kinds of ways. One is that prevalence
of particular disorders may vary, affecting everything from conceptual schemas and
frames related to the disorder to public health policy and social supports. Another is
that depending on factors such as socioeconomic status and attitudes, people may
or may not seek treatment. This may be not only because the kinds of treatment and
their cost to individuals vary (and so may be beyond the means of some sociocultural
groups) but also because frames and schemas vary. If memory loss is regarded as
normal in old age, for instance, people with AD may not be defined as ill until the
symptoms are severe. And if drug treatments are only publicly funded for moderate
or severe AD, then treatment of people with mild AD may be limited to people whose
families can afford it.
4
Such a model may shed light on the current debate about category specific deficits for
object concepts. The discussion hinges around evidence that damage to neural tissue
in different brain regions differentially affects (access to) concept categories such as
animals and tools. For example, perceiving animals (and faces) appears to be associ-
ated with bilateral activation of the lateral fusiform gyrus in contrast with perception
of tools which bilaterally activates the medial fusiform gyrus (Chao et al. 1999; Grill-
Spector 2003). Current ‘competing’ explanatory hypotheses for these differences
include ‘sensory-functional’, ‘domain specific’ and ‘conceptual structure’ approaches
(reviewed in Tyler and Moss 2001; Caramazza and Mahon 2003; Thompson-Schill
2003).
The sensory-functional hypothesis prioritizes modality and use as organizing prin-
ciples for categories. The overarching hypothesis is that concepts are represented as
features at least partially distributed along the neural pathways associated with their
acquisition. So, tool concepts might be partially represented in neural regions close
to those activated by their use whereas animal concepts might be primarily repre-
sented in neural regions close to those associated with, for instance, their shapes (e.g.
Warrington and Shallice 1984; Warrington and McCarthy 1987; Thompson-Schill
2003). The domain-specific hypothesis prioritizes taxonomic domains: animals,
fruits and vegetables, conspecifics and tools are postulated as potentially phyloge-
netically motivated, conceptual domains (Caramazza and Shelton 1998; Caramaza
and Mahon 2003). Under either of these hypotheses, damage in a region that affected
one category need not affect concepts in another category because the categories are
stored in different areas of the brain.
The conceptual structure approach also posits distributed features for concepts,
but prioritizes their internal feature structure as the salient factor in category specific
deficits. In this framework, domains such as animals and tools have different inter-
nal conceptual organizations. Animals have many shared features but few distinctive
ones. Tools have a few highly salient distinctive features (they are designed usually
for one purpose), but few shared features. The model predicts that the superordinate
category of animals will be less vulnerable to damage because it will have multi-
ple connections with other concepts within the same domain. However, basic level
concepts such as CAT and COW will be more vulnerable because they have few
214 Notes to page 209
distinctive features and those they have are not highly correlated with each other.
Tool concepts may be resistant to damage (as compared with basic level animal
concepts) because although they have fewer features in general, those features they
do have are distinctive and correlated with each other. Under this view, conceptual
structure is an emergent property of the connections between features, and category
deficits are interpreted as damage to feature structure levels (Moss et al. 2007).
Mesulam’s model is most obviously consistent with the sensory-functional
hypothesis of Warrington and Shallice. However, it would not appear to rule out
the other possibilities of domain specific and modality neutral concept organiza-
tion. Rather, as suggested by Thompson-Schill (2003), domain specific organization
(and apparent deficit) could occur as an emergent property of a distributed sensory-
functional organization of semantic categories. That is, the two apparently clear
categories which may be differentially impaired, tools and animals, are respectively
associated with function and vision. Deficits for these categories might thus be evi-
dence of deficit for functional or modal attributes rather than for the categories as
such. Martin’s (2007) review of functional neuroimaging studies of object concepts
similarly suggests that modal-functional and domain-specific views of conceptual
organization are not necessarily mutually exclusive: studies collectively suggest that
distributed representation of features for object concepts according to modality or use
may co-exist with dedicated domains for categorical learning.
Notes on Closing remarks

1
An exception to this general state of affairs is work on schizophrenic discourse. See
for example the early work of Rochester and Martin 1979.
Appendix A: Some basic grammatical
terminology and relations
Grammatical terminology
clauses and utterances
In Chapter 3 we defined the terms Clause and Utterance (pp. 29–32). Here
we include brief definitions for convenient reference. The reader is referred to
Chapter 3 (pp. 29–32) for broader definitions and examples.
English Clauses can be defined syntactically as consisting of a verb, its
arguments and adjuncts as in Cosmo bit Piper yesterday. Independent clauses
can also be used alone to ask a question, make a statement or exclamation,
or give a command. A simple Clause is also typically spoken as a single tone
group.
Meaningful speech phenomena which do not meet the criteria for clause
such as incomplete utterances, minimal responses, idiosyncratic vocalizations,
and isolated hesitation fillers are labelled as utterances. Utterance refers to
any unit which can be assigned a speech function, and/or has a distinct tone
group, and/or is a linguistic signal of ideational, interactional, or organizational
information about a speaker’s message.
morphemes
Morphemes are smallest contrastive units in the grammar. Rabbit, dog, -s, in
dogs, -ed in called, -ity in fatality, un- in unhappy, brush in toothbrush are
all morphemes. None of these items can be further analysed. For instance,
rabbit refers to a small furry long-eared animal that hops, but no part of rab-
bit is associated with one of these meanings. Morphemes may or may not be
words: Whereas rabbit is a word, -s in dogs signifies plural but does not occur
as a word.
Words may be simple (e.g. rabbit/dog), compound (toothbrush) or com-
plex (fatality/unhappy). What differentiates a word from a morpheme is
simply the fact that we have a word for those morphemes or compound/com-
plex combinations of morphemes which can occur independently. In each
case, the unit signifies one or more meanings with which it is conventionally
associated.
215
216 Appendix A
Word categories
The major categories are NOUN, VERB, ADJECTIVE and ADVERB.
nouns
Nouns are Proper (e.g. John, Canada), or Common (cat, human, kindness). If
common they can be countable (i.e. be singular or plural as in boy/boys, goose/
geese etc.), or non-countable (butter, milk, happiness etc.). Included in this
category are PRONOUNS which can substitute for noun phrases (e.g. Personal
Pronouns as in I/me/mine, we/us/ours, you/you/yours, he/him/his, she/her/hers,
it/it/its, one/one/ones, they/them/theirs). There are also Reflexive Pronouns
(myself, yourself, ourselves etc.), Relative Pronouns (who, whom, whose, which,
which, what, that), Indefinite Pronouns (compounds formed with some-, any-,
every- or no- and -one, -body or -thing as in somebody, anything, everyone etc.)
and two ‘special’ pronouns: existential there occuring as Subject in existential
clauses (as in There is a woman at the door). Notice that existential there must
be unstressed), and Ambient or ‘expletive’ it which doesn’t refer to anything
(as in the it in it’s sunny).
Characteristic modifiers of nouns:

• DETERMINERS (e.g. the/a/an; this/that/these/those; some/any/every/no
etc. as in the dog, a man, some milk, these puppies and so on).
• ADJECTIVAL PHRASES (e.g. very tall women, totally unacceptable
behaviour where very tall and totally unacceptable are adjective phrases
modifying the nouns women and behaviour respectively).
• PREPOSITIONAL PHRASES (e.g. a book about physics, a manual of
style, a teacher of linguistics where about physics, of style and of linguistics
are prepositional phrases modifying the nouns book, manual and teacher
respectively).
• RELATIVE CLAUSES (e.g. the man who answered the letter, the dog that
bit me where who answered the letter and that bit me are relative clauses
operating as modifiers of the nouns man and dog).
It is normal for determiner and adjectival phrase modifiers to occur before the
noun that they modify and for prepositional phrases and relative clauses to
follow the noun they modify.
Nouns operate as the head element of noun phrases. Noun Phrases can occur
as Subject, Complement, or Adjunct in a clause. Noun Phrases can also appear
case marked by prepositions as in the above examples. That is Prepositional
Phrases are Noun Phrases to which a preposition has been added. Most modern
grammars describe prepositional phrases as consisting of a Preposition and a
Noun Phrase complement. In English it is normal for the head noun of a noun
Appendix A 217
phrase occurring as Subject of a clause to agree with the verb in person and
number if the clause is indicative declarative or indicative interrogative. The
bracketed phrases in the following examples are Noun Phrases operating as
elements of clauses. The underlined words are nouns or pronouns.
(i) [The dog] eats [rabbits].
(ii) [A very serious discussion about ethics] was going on in [the pub].
(iii) [John] is [a person of considerable integrity].
(iv) [They] say [he] will phone again [tomorrow].
(v) [I] told [them] [you] would be [home].
verbs
Verbs can be inflected for tense (past or present as in walk/walked, eat/ate,
go/went), number (singular or plural am/are, was/were, though the singular/
plural distinction in past tense is only relevant for the verb BE), and for present
tense verbs person (third person, present tense singular; he eats, she walks, it
talks; and general present I/we/you/they eat/walk/talk).
The ‘modal’ verbs are an exception. They are can/could, may/might, will/
would, shall/should, must, ought, (and dare, need, and BE in special circum-
stances). Many linguists assume that modals are tensed pairs, others do not. In
any case, whether they are treated as tensed or not, their forms have no clear
relationship to semantic time (see for example I would not care to see the crime
go unpunished, I could leave tomorrow, Mary will have finished by now, ?John
will have left yesterday). Notably, they have no other inflected forms (i.e. there
is no *canning, *mighted, *wills and so on). Linguists treat the modals as finite
forms whether or not they recognize them as tensed.
Tensed verbs are often called finite verb forms and they contrast with non-
tensed or non-finite forms. The non-finite inflection forms are:
1) the BASE FORM (be/walk/have/eat/talk)
2) the MARKED INFINITIVE FORM (to be/to walk/to eat)
3) the -ING FORM as in (walking/eating/going)
4) the -EN FORM (eaten/walked/gone).
Verbs are central to the formation of clauses for two reasons. One is that it is
often the choice of verb that determines both the number and type of other ele-
ments of the clause. For example, eat requires both an NP subject and an NP
complement and, if the clause is active, the subject will be Agent (the doer of
the action) and the complement will be Theme: patient (the entity that under-
goes a change of state). Thus the boy ate the apple is acceptable, but not *of
the boy eats at the apple, nor *the table eats the idea and so on. The other
reason for the centrality of verbs in clause formation is that only certain types
of verb forms can occur in certain types of clauses. For instance, interrogatives
218 Appendix A
and declaratives require that the first verb word be a finite form (modal or
tensed verb). (John is at home/ John will be at home/ Will John be at home/ Is
John at home are all acceptable but *John been at home/ *Being John at home
and so on are not.) Similarly, imperative clauses require that the first verb word
is either the modal let or a base form (Be quiet/ Eat your dinner/ Do study for
you exam/ Let’s go).
adjectives
Adjectives are uninflected forms, although many can form comparatives and
superlatives with the suffixes -er and -est repectively as in clever, cleverer,
cleverest, kind, kinder, kindest, nice, nicer, nicest. Adjectives which cannot
form a comparative or superlative with -er or -est can usually have more or
most as a modifier with the same effect (e.g. *astuter, astutest are bad, but
more astute, most astute are acceptable). Comparatives and superlatives of
adjectives (whether derived through suffixation of -er/-est or modification
with more/most) function as determiners; they select or identify individuals
in terms of a base attribute. These forms are consequently treated as derived
rather than inflected. Adjectives realize ‘attributes’ and they occur either as
modifiers of nouns (clever student, kinder era, nicest dog) or as the predicate
in attributive relational clauses with a form of BE as the main verb (This stu-
dent is clever/ That era was kinder/ Susan’s dog is nicest).
adverbs
Adverbs are also uninflected forms. Many are derived from adjectives with
the addition of an -ly suffix (as in cleverly, nicely, kindly, astutely and so on).
Like adjectives, comparatives and superlatives of adverbs are formed by the
addition of more or most as a modifier (as in more cleverly, most astutely),
and they can be intensified by DEGREE ADVERBS such as very, so, too,
almost, quite and so on. Otherwise adverbs accept neither modification nor
complementation. They occur as ADJUNCTS only, modifying adjectives
(e.g. cautiously clever, totally silly), other adverbs in the case of the degree
subclass (very cleverly, somewhat slowly, quite suddenly), verbs (com-
pletely exhausted, entirely gone, utterly destroyed), verb phrases (left the
room quietly/carefully/silently), or whole sentences (frankly, I don’t believe
it/ Mary left for Montreal quite suddenly/ unfortunately, John is sick). This
description of adverbs excludes from the category comparatives and superla-
tives formed with -er and -est. Forms such as quicker and fastest in Sam ran
quicker (than Sue), Sam runs fastest (of all) are regarded here as determin-
ers derived from adjectives, not as adverbs. Note that prototypical adverbs
(suddenly, happily, quickly, botanically etc.) do not allow suffixation of
-er/-est.
Appendix A 219
In addition to the major categories of words above, there are also three impor-
tant minor categories; PREPOSITIONS, DEICTICS and CONJUNCTIONS.
prepositions
Prepositions are also uninflected forms. They form a closed class (as opposed
to open set categories like nouns and verbs) insofar as the number of them is
finite (i.e. new prepositions are very infrequently added to English). The class
consists of items such as to, by, with, for, at, in, out, on, over, up, down, under,
beneath, beside, above, below and so on. Prepositions occur as modifiers of
Noun Phrases (to John, on the table, at home, in the afternoon) in which case
the phrases are often called Prepositional Phrases. They can also occur alone
as, for instance, the sole realization of a Goal: location (John went out) or a
locative predicate (Mary is in).
deictics
Deictics include the definite and indefinite articles (the/a), the demonstratives
(this/that/these/those), the possessive articles (my, our, your, our, his, her, their,
its, one’s) and lexical possessives (John’s, the president’s, Canada’s), quantifi-
ers (one, two, three…, first, second …) and a few other elements. (The compar-
atives and superlatives of adjectives are sometimes included among the deictic
class.) All of these occur as specifiers of Noun Phrases and will precede any
adjectival modifiers of the nouns as in the dog/ the clever dog / the first clever
dog and so on, but not, for example, *clever first the dog. Some of them, such
as the demonstratives, can substitute for a whole noun phrase as in give me the
book / give me that.
conjunctions
Conjunctions are of three general types:
1) co-ordinating conjunctions such as and/or/but which can be used to
co-ordinate words, phrases, and clauses,
2) subordinating conjunctions such as because/if/although/in order/so that
which are used to subordinate one phrase or clause in relation to another,
3) sentence conjunctions such as however, moreover, nevertheless and so on
which indicate logical relationships between independent sentences.
Some basic types of grammatical relation and terminology

constituency and dependency
There are two fundamental types of grammatical relation, constituency and
dependency.
220 Appendix A
Constituency characterizes the structure of phrases (including clauses and

sentences) in terms of a vertical dimension such that, for instance, sentences
consist of clauses, clauses consist of phrases, phrases consist of words and/or
other phrases, and words consist of morphemes.
Dependency characterizes phrases on a horizontal dimension in terms of
relationships of ‘companionship’. In dependency terms, a noun phrase must
have a ‘head’ element (a noun) and may optionally have a specifier (such as an
article) as well as other pre-head and post-head modifiers. Modifiers that occur
before a head noun are typically called modifiers, those after it may be called
complements or adjuncts, depending on their relationship to the head noun.
(Some linguists call post-head noun modifiers ‘qualifiers’.) Similarly, a Verb
Phrase must have a verb and may have complements and so on.
Head as a technical term refers to the element of any construction that must
be present in order for the construction to occur (Noun in Noun phrases, Verb
in Verb Phrases, Adjective in Adjective Phrases, Adverb in Adverb Phrases and
Preposition in Prepositional Phrases).
subject
A subject is an element of clause structure. It is typically a Noun Phrase (NP)
(He/John/The tall man frightened them), but may also be a clause (What they
saw frightened them). If it is a Noun Phrase and the sentence is finite (has
a verb that is inflected for tense, number and person) the noun will ‘agree’
with the verb in person and number, and will be nominative case (He is sick/
They are sick, but not *He are sick, Them is sick). In Declaratives, the subject
will precede all verbs (Mary is answering the question/ Mary might have been
being watched). In Interrogatives, the subject will follow the first verb, usually
an auxiliary in Modern English (Is Mary answering the question?/ What did
Mary hear?). (Interrogatives such as Who is sick? are exceptions to this order.)
In Imperatives (of the jussive type) the subject will be absent or it will be you/
somebody (Go home!/ you go home/ somebody call an ambulance). A vocative
element may be present (John, go home). Some linguists treat such elements
as subjects.
complement
A complement is an element of phrase (including clause) structure.
Complements are realized by all categories of phrases except adverb phrases.
They follow the main verb in clauses (unless they have been moved). There
can be more than one complement in a clause. For example, verbs like put
and give require two complements as well as a subject (Sam put [the milk]
[in the fridge] / Sam gave [Sue] [a ring]), and a verb like sell can have three
complements (Sam sold [Fred] [the ring] [for two dollars]). Complements
Appendix A 221
are arguments of a predicate: if an element following a main verb (or other

predicate) is either required or at least expected given the meaning of the verb
it is a complement. (Note: in traditional grammars only NP complements are
recognized. Sometimes they are referred to as indirect and direct objects as in
John gave Mary the book.)
predicate
Predicate refers to any element that enters into an argument structure that is
not itself an argument. (From the point of view of the syntax of the clause, this
means that anything that is not a subject or a complement must be a predicate.)
Typically, the elements that count as predicates are the lexical head of any
construction. Verbs are prototypical predicates insofar as they take both com-
plements and subjects and, as we have seen, the particular verb determines the
number and types of arguments that are required to make acceptable clauses
(e.g. put requires an Agent, a Theme: transferent, and a Goal: location which,
all other things being equal, are realized as Subject, Objective complement and
Locative complement respectively). However, nouns and adjectives also func-
tion regularly as predicates as in the Iraqi invasion of Kuwait where invasion
(a noun) has a complement (of Kuwait), and John is sick where the adjective
sick is the predicate.
adjunct
An adjunct is an element of phrase (including sentence) structure. Adjuncts
are circumstantial elements usually having to do with the time, place, or man-
ner of an event. They usually follow all complements (but they can be moved
around). They are often realized by prepositional phrases (on Wednesday, at
the movies, with alacrity etc.). Adverbs are always adjuncts. We also recognize
‘logical’ adjuncts (phrases beginning with because, in order to, if, although
etc.). Adjuncts are ‘optional’ elements in the sense that phrases and sentences
may be grammatically acceptable without adjuncts. Adjuncts predicate some-
thing about the consitutuent they are adjoined to. For example, the sentence
Mary is astute is perfectly acceptable. When the adjunct fortunately is added to
it we have another predication about the sentence viz. It is fortunate that Mary
is astute or [[Mary is astute] fortunately].
Appendix B: Inventory of codes
Speech functions
Exclamations
E after clause
Statement
S after clause
Statement tagged
S-TAG after clause
Question polar
QP after clause
Question wh
QWH after clause
Command jussive
CJ after clause
Command optative
CO after clause
Command fiat
CF after clause
Checks and intonation

Checks
S-CHECK after clause
Rising Pitch
S-RP after clause
Address terms
Vocative + social distance
V+SD after vocative
Politeness marker
PMR after politeness marker
Attitudes and evaluation

Modality
M after modal
Modality feature
MF after modality feature
Attitude
ATT-N or ATT-P after attitude
222
Appendix B 223
Appendix B (cont.)
Argument roles
Agent
AGENT after agent
Instrument
INSTRUMENT after instrument
Cause
CAUSE after cause
Experiencer
EXPERIENCER after experiencer
Stimulus
STIMULUS after stimulus
Source
SOURCE after source
Goal: Location
GOAL-LOC after Goal: Location
Goal: Recipient
GOAL-REC after Goal: Recipient
Goal: Beneficiary
GOAL-BEN after Goal: Beneficiary
Theme: Patient
THEME-PAT after Theme: Patient
Theme: Resultant
THEME-RES after Theme: Resultant
Theme: Percept
THEME-PERC after Theme: Percept
Theme: Message
THEME-MESS after Theme: Message
Theme: Range
THEME-RAN after Theme: Range
Theme: Identified
THEME-IDEN’D after Theme: Identified
Theme: Classified
THEME-CLASS’D after Theme: Classified
Theme: Attribuand
THEME-ATTR after Theme: Attribuand
Theme: Possessed
THEME-POSS’D after Theme: Possessed
Theme: Existent
THEME-EX after Theme: Existent
Theme: Ambient
THEME-AMB after Theme: Ambient
Circumstantial roles
Time
TIME after time
Place
224 Appendix B
Appendix B (cont.)
PLACE after place

Manner
MANNER after manner
Reason
REASON after reason
Purpose
PURPOSE after purpose
Condition
CONDITION after condition
Concession
CONCESSION after concession
Message organization
Reference
R after reference, co-indexed with first instance
Substitution
SUB after substitution, co-indexed with substituted item
Ellipsis
Ellip after ellipsis, co-indexed with ellipted item
Conjunction
CONJ after conjunction
Lexical Cohesion
LC after lexical item, co-indexed with presupposed item
Lexical Cohesion – Repetition
LC-REP after lexical item, co-indexed with first instance
Lexical Cohesion – Collocation
LC-COL after lexical item, co-indexed with collocate
Focus and Prominence:

Focus = Underline; Prominence = Bold.
References
Alagiakrishnan, K., McCracken, P. and Feldman, H. 2006. Treating vascular risk factors
and maintaining vascular health: is this the way towards successful cognitive ageing
and preventing cognitive decline? Postgraduate Medical Journal, 82 (964), 101–5.
Allain, P., Le Gall, D., Foucher, C. et al. 2008. Script representation in patients with
Alzheimer’s disease. Cortex, 44 (3), 294–304.
Almkvist, O. 1994. Neuropsychological deficits in vascular dementia in relation to
Alzheimer’s disease: reviewing evidence for functional similarity or divergence.
Dementia, 55, 203–9.
Almor, A., Kempler, D., MacDonald, M. C., Andersen, E. S. and Tyler, L. K. 1999. Why
do Alzheimer patients have difficulty with pronouns? Working memory, seman-
tics, and reference in comprehension and production in Alzheimer’s disease. Brain
and Language, 67(3), 202–27.
Alvarez, J. A. and Emory, E. 2006. Executive function and the frontal lobes: a meta-
analytic review. Neuropsychology Review, 16(1), 17–42.
Alzheimer, A. 1907. On a peculiar disease of the cerebral cortex. Allgemeine Zeitschrift
fur Psychiatrie und Psychish-Gerichtlich. Medicin, 64, 146–8.
Alzheimer’s Association 2007. Alzheimer’s Disease Facts and figures. Report. New
York: Alzheimer’s Association New York City Chapter.
American Psychiatric Association 1994. Diagnostic and Statistical Manual of Mental
Disorders, fourth edition. Washington, DC: American Psychiatric Publishing.
2000. Diagnostic and Statistical Manual of Mental Disorders, fourth edition. Text
Revision. Washington, DC: American Psychiatric Publishing.
Anderson, R. C., Spiro, R. J., and Montague, W. E. 1977. Schooling and the Acquisition
of Knowledge. Hillsdale, NJ: Lawrence Erlbaum.
Asp, E. 2001. How to do different things with words: some observations on speech
acts in relation to a socio-cognitive grammar for English. In de Villiers, J. and
Stainton, R. (eds.), Communication in Linguistics: Vol. 1. Toronto: Editions du
Gref, pp. 1–13.
Asp, E., Song, X. and Rockwood, K. 2006a. Self-referential tags in the discourse of
people with Alzheimer’s disease. Brain and Language, 97, 41–52.
Asp, E., Cloutier, F., Fay, S., Cook, C., Robertson, M. L., Fisk, J., Dei, D. W. and
Rockwood, K. 2006b. Verbal repetition in patients with Alzheimer’s disease who
receive donepezil. International Journal of Geriatric Psychiatry, 21, 426–31.
Asperger, H. 1944. Autistic psychopathology in childhood. In Frith, U. (ed.), Autism
and Asperger Syndrome. Cambridge: Cambridge University Press, pp. 37–92.
225
226 References
Atkinson, J. M. and Heritage, J. (eds.) 1984. Structures of Social Action: Studies in
Conversation Analysis. Cambridge: Cambridge University Press.
Austin, J. L. 1962. How to Do Things with Words: the William James Lectures delivered
at Harvard University in 1955, (ed. Urmson, J. O.). Oxford: Clarendon.
Baba, M., Nakajo, S., Tu. P. H., Tomita, T., Nakaya, K., Lee, V. M., Trojanowski, J. Q.
and Iwatsubo, T. 1998. Aggregation of alpha-synuclein in Lewy bodies of spo-
radic Parkinson’s disease and dementia with Lewy bodies. American Journal of
Pathology, 152, 879–84.
Baddeley, A. D. 1998. Working memory. Comptes rendus de l’académie des sciences,
321(2–3), 167–73.
2003. Working memory and language: an overview. Journal of Communication
Disorders, 36, 189–208.
Baddeley, A. D. 2000. The episodic buffer: a new component of working memory?
Trends in Cognitive Sciences, 4(11), 417–22.
Baddeley, A. D. and Della Sala, S. 1996. Working memory and executive control.
Philosophical Transactions of the Royal Society of London B: Biological Sciences,
351, 1397–404.
Baddeley A. D., Baddeley, H. A., Bucks, R. S. and Wilcock, G. K. 2001. Attentional con-
trol in Alzheimer’s disease. Brain, 124, 1492–508.
Baltaxe, C. 1977. Pragmatic deficits in the language of autistic adolescents. Journal of
Pediatric Psychology, 2, 176–180.
Baron-Cohen, S. 1995. Mindblindness: An Essay on Autism and Theory of Mind.
Massachusetts: MIT Press.
Baron-Cohen, S., Leslie, A. and Frith, U. 1985. Does the autistic child have a ‘theory of
mind’? Cognition, 21, 37–46.
Barthes, R. 1968. Elements of Semiology. New York: Hill and Wang.
1994. The Semiotic Challenge. Berkeley and Los Angeles: University of California
Press.
Baudie, S., Dalla Barba, G., Thibaudet, M. C., Sagghe, A., Remy, P. and Traykov, L.
2006. Executive function deficits in early Alzheimer’s disease and their relations
with episodic memory. Archives of Clinical Neuropsychology, 21, 15–21.
Bayles, K. A., Kaszniak, A. W. and Tomoeda, C. 1987. Communication and Cognition
in Normal Aging and Dementia. Boston: Little, Brown and Company.
Bayles, K. A., Tomoeda, C. K. and Trosset, M. W. 1992. Relation of linguistic commu-
nication abilities of Alzheimer’s patients to stage of disease. Brain and Language,
42, 454–72.
Beauregard, M., Chertkow, H., Bub, D., Murtha, S., Dixon, R. and Evans, A. 1997. The
neural substrate for concrete, abstract and emotional word lexica: a positron emis-
sion tomography study. Journal of Cognitive Neuroscience, 9(4), 441–61.
Bechara, A., Damasio, H. and Damasio, A. R. 2000. Emotion, decision making and the
orbitofrontal cortex. Cerebral Cortex, 10, 295–307.
Beck, A. T. 2005. The current state of cognitive therapy: a 40-year retrospective.
Archives of General Psychiatry, 62(9), 953–9.
Bednarek, M. A. 2005. Frames revisited – the coherence-inducing function of frames.
Journal of Pragmatics, 37, 685–705.
Berg, C., Raminani, S., Greer, J., Harwood, M. and Safren, S. 2008. Participants’ per-
spectives on cognitive-behavioural therapy for adherence and depression in HIV.
References 227
Psychotherapy Research: Journal for the Society of Psychotherapy Research,

18(3), 271–80.
Berger, P. and Luckmann, T. 1966. The Social Construction of Reality: A Treatise in the
Sociology of Knowledge. Garden City, NY: Anchor Books.
Berlin, B. and Kay, P. 1969. Basic Color Terms. Berkeley and Los Angeles: University
of California Press.
Berlyne, N. 1972. Confabulation. The British Journal of Psychiatry, 120, 31–9.
Bickel, C., Pantel, J., Eysenbach, K. and Schröder, J. 2000. Syntactic comprehension
deficits in Alzheimer’s disease. Brain and Language, 71(3), 432–48.
Bird, H., Howard, D. and Franklin, S. 2000. Why is a verb like an inanimate object?
Grammatical category and semantic category deficits. Brain and Language, 72,
246–309.
Bird, T., Knopman, D., Van Swieten, J. et al. 2003. Epidemiology and genetics of fron-
totemporal dementia/Pick’s disease. Annals of Neurology, 54(S5), S29–31.
Birks, J. 2006. Cholinesterase inhibitors for Alzheimer’s disease. Cochrane Database
Systematic Review, 25(1), CD005593.
Bishop, D. V. and Adams, C. 1989. Conversational characteristics of children with
semantic-pragmatic disorder: II. What features lead to a judgement of inappro-
priacy? British Journal of Disorders of Communication, 24(3), 241–63.
Black, S. E. 2007. Therapeutic issues in vascular dementia: studies, designs and
approaches. Review. Canadian Journal of Neurological Sciences, 34(1),
S125–30.
Black, M. and Chiat, S. 2003. Noun–verb dissociations: a multi-faceted phenomenon.
Journal of Neurolinguistics, 16 (2–3), 231–50.
Bleich, S., Romer, K., Wiltfang, J. and Kornhuber, J. 2003. Glutamate and the gluta-
mate receptor system: a target for drug action. International Journal of Geriatric
Psychiatry, 18, S33–40.
Blennow, K. and Hampel, H. 2003. CSF markers for incipient Alzheimer’s disease.
Lancet Neurology, 2(10), 605–13.
Blennow, K., de Leon, M. J. and Zetterberg, H. 2006. Alzheimer’s disease. The Lancet,
368, 387–403.
Bookheimer, S. 2002. Functional MRI of language: new approaches to understanding
the cortical organization of semantic processing. Annual Review of Neuroscience,
25, 151–88.
Bowler, J. V. and Hachinski, V. 2003. Vascular Cognitive Impairment: Preventable
Dementia. Oxford: Oxford University Press.
Bradac, J. J. 1988. Language variables: conceptual and methodological problems
of instantiations. In Tardy, C. H. (ed.), A Handbook for the Study of Human
Communication: Methods and Instruments for Observing, Measuring, and Assessing
Communication Processes. Westport, CT: Ablex Publishing, pp. 301–22.
Brazil, D. 1995. A Grammar of Speech. Oxford: Oxford University Press.
Brown, G. and Yule, G. 1983. Discourse Analysis. Cambridge: Cambridge University
Press.
Brown, P. and Levinson, S. C. 1987. Politeness. Cambridge: Cambridge University
Press.
Brunet, E. 1978. Le vocabulaire de Jean Giraudoux: Structure et Evolution. Genève:
Skatline.
228 References
Bschor, T., Kuhl, K. P. and Reischies, F. M. 2001. Spontaneous speech of patients with
dementia of the Alzheimer type and mild cognitive impairment. International
Psychogeriatrics, 13(3), 289–98.
Bucks, R., Singh, S., Cuerden, J. M., and Wilcock, G. 2000. Analysis of spontaneous,
conversational speech in dementia of Alzheimer type: evaluation of an objective
technique for analyzing lexical performance. Aphasiology, 14 (1), 71–91.
Burns, A., O’Brien, J., Bap Dementia Consensus Group, et al. 2006. Clinical practice
with anti-dementia drugs: a consensus statement from the British Association for
Psychopharmacology. Journal of Psycholpharmacology, 20(6), 732–55.
Cabeza, R. and Nyberg, L. 2000. Imaging Cognition II: an empirical review of 275 PET
and fMRI studies. Journal of Cognitive Neuroscience, 12, 1–47.
Camille, N., Coricelli, G., Sallet, J., Pradat-Diehl, P., Duhamel, J. R. and Sirigu, A.
2004. The involvement of the orbitofrontal cortex in the experience of regret.
Science, 304, 1167–70.
Campione, E. and Veronis, J., 2002. A large-scale multilingual study of silent pause
duration. ESCA-workshop on speech prosody. April 2002, Aix-en-Provence,
pp. 199–202.
Caplan, D., Vijayan, S., Kuperberg, G. et al. 2001. Vascular responses to syntactic
processing: event-related fMRI study of relative clauses. Human Brain Mapping,
15, 26–38.
Caramazza, A. and Mahon, B. Z. 2003. The organization of conceptual knowledge: the
evidence from category-specific semantic deficits. Trends in Cognitive Sciences,
7(8), 354–61.
Caramazza, A and Shelton, J. R. 1998. Domain-specific knowledge systems in the
brain: the animate–inanimate distinction. Journal of Cognitive Neuroscience,
10, 1–34.
Cardebat, D., D’Monet, J. F. and Doyon, B. 1993. Narrative discourse in dementia.
In Brownell, H. H. and Joanette, Y. (eds.), Narrative Discourse in Neurologically
Impaired and Normal Aging Adults. San Diego: Singular, pp. 317–32.
Carlomangno, S., Santoro, A., Menditti, A., Pandolfi, M. and Marini, A. 2005. Referential
communication in Alzheimer’s type dementia. Cortex, 41, 520–34.
Catani, M. and Mesulam, M. 2008. The arcuate fasciculus and the disconnection theme
in language and aphasia: history and current state. Cortex, 44, 953–61.
Catani, M. and Thiebaut de Schotten, M. 2008. A diffusion tensor imaging tractography
atlas for virtual in vivo dissections. Cortex, 44, 1105–32.
Catani, M., Jones, D. K. and Ffytche, D. H. 2005. Perisylvian language networks of the
human brain. Annals of Neurology, 57, 8–16.
Chafe, W. 1980. The deployment of consciousness in the production of a narrative. In
W. Chafe (ed.), The Pear Stories: Cognitive, Cultural, and Linguistic Aspects of
Narrative Production. Norwood, NJ: Ablex, pp. 9–50.
1994. Discourse, Consciousness, and Time: The Flow and Displacement of Conscious
Experience in Speaking and Writing. Chicago: University of Chicago Press.
2001. The analysis of discourse flow. In Schiffrin, D., Tannen, D. and Hamilton,
H. E. (eds.), The Handbook of Discourse Analysis. Malden, MN: Blackwell,
pp. 673–88.
Chao, L. L., Haxby, J. V. and Martin, A. 1999. Attribute-based neural substrates in
temporal cortex for perceiving and knowing about objects, Nature Neuroscience,
2, 913–19.
References 229
Chertkow, H. and Bub, D. 1990. Semantic memory loss in dementia of Alzheimer’s

type: what do various measures measure? Brain, 113, 397–417.
Chipere, N., Malvern, D. and Richards, B. 2004. Using a corpus of children’s writ-
ing to test a solution to the sample size problem affecting type-token rations. In
Aston, G., Bernardini, S. and Stewart, D. (eds.), Corpora and Language Learners.
Amsterdam: John Benjamins, pp. 137–50.
Chomsky, N. 1957. Syntactic Structures. The Hague: Mouton.
1995. The Minimalist Program. Cambridge, MA: MIT Press.
Chow, H. M., Kaup, B., Raabe, M. and Greenlee, M. W. 2008. Evidence of fronto-
temporal interactions for strategic inference processes during language compre-
hension. NeuroImage, 40, 940–54.
Chwilla, D. J. and Kolk, H. H. J. 2005. Accessing world knowledge: evidence from
N400 and reaction time priming. Cognitive Brain Research, 25, 589–606.
Clark, H. H. 1996. Using Language. Cambridge: Cambridge University Press.
Clark, H. H. and Fox Tree, J. E. 2002. Using uh and um in spontaneous speech.
Cognition, 84, 73–111.
Clark, H. H. and Wasow, T. 1998. Repeating words in spontaneous speech. Cognitive
Psychology, 37, 201–42.
Coates, J. 1996. Women Talk. Oxford: Blackwell.
Coates, J. and Cameron, D. (eds.) 1988. Women in Their Speech Communities. London
and New York: Longman.
Collina, S., Marangolog, P. and Tabossi, P. 2001. The role of argument structure in the
production of nouns and verbs. Neuropsychologia, 39, 1125–37.
Cooper, J. O., Heron, T. E. and Heward, W. L. 2006. Applied Behaviour Analysis,
2nd edition. New Jersey: Prentice Hall.
Coricelli, G., Critchley, H., Joffily, M., O’Doherty, J., Sirigu, A. and Dolan, R. 2005.
Regret and its avoidance: a neuroimaging study of choice behaviour. Nature
Neuroscience, 8(9), 1255–62.
Coricelli, G., Dolan, R. J. and Sirigu, A. 2007. Brain, emotion and decision making: the
paradigmatic example of regret. Trends in Cognitive Sciences, 11(6), 258–65.
Cosentino, S., Scarmeas, N., Albert, S. and Stern, Y. 2006. Verbal fluency predicts mortality
in Alzheimer disease. Cognitive and Behavioral Neurology, 19(3), 123–9.
Coupland, J., Robinson, J. and Coupland N. 1994. Frame negotiation in doctor–elderly
patient consultations. Discourse and Society, 5(1), 89–124.
Cullen B., Coen R. F., Lynch, et al. 2005. Repetitive behaviour in Alzheimer’s dis-
ease: description, correlates, and functions. International Journal of Geriatric
Psychiatry, 20, 686–93.
Cummings, J. L. 1997. Neuropsychiatric manifestations of right hemisphere lesions.
Brain and Language, 57(1), 22–37.
Dalla Barba, G., Nedjam, Z. and Dubois, B. 1999. Confabulation, executive func-
tions, and source memory in Alzheimer’s disease. Cognitive Neuropsychology, 16
(3/4/5), 385–98.
Danes, F. 1974. Functional sentence perspective and the organization of the text. In
Danes, F. (ed.), Papers on Functional Sentence Perspective. The Hague: Mouton,
pp. 106–28.
D’Argembeau, A., Xue, G., Lu, Z.-L., Van der Linden, M. and Becharab, A. 2008.
Neural correlates of envisioning emotional events in the near and far future.
NeuroImage, 40, 398–407.
230 References
Delgado, P. L. and Moreno, F. A. 2006. Neurochemistry of mood disorders. In Stein,
D. J., Kupfer, D. J. and Schatzberg, A. F. The American Psychiatric Publishing
Textbook of Mood Disorders. Arlington, VA: American Psychiatric Publishing,
Inc., pp. 101–16.
De Martino, B., Kumaran, D., Seymour, B. and Dolan, R. J. 2006. Frames, biases, and
rational decision-making in the human brain. Science, 313(5787), 684–7.
Dennis, M., Lazenby, A. L. and Lockyer, L. 2001. Inferential language in high-func-
tioning children with autism. Journal of Autism and Developmental Disorders,
31(1), 47–54.
Den Ouden, H. E. M., Frith, U., Frith, C. and Blakemore, S. J. 2005. Thinking about
intentions. NeuroImage, 28, 787–96.
Desmond, D. 2003. Vascular dementia. Clinical Neuroscience Research, 3(6), 437–48.
de Villiers, J. 2005. Discourse Analysis in Autism Spectrum Disorder, Linguistics and
the Human Sciences, 1(2): 245–60.
2006. Syntactic and semantic patterns of pedantic speech in Asperger’s syndrome. In
Hwang, S., Sullivan, B. and Lommel, A., eds., LACUS Forum XXXII: Networks.
Houston: LACUS.
de Villiers, J. and Stainton, R. (eds.) 2009. Communication in Linguistics Vol. II:
Michael Gregory’s Proposals for a Communication Linguistics. GREF.
de Villiers, J. and Szatmari, P. 2004. Message Organization in Autism Spectrum
Disorder. In Fulton, G. D., Sullivan W. J. and Lommel, A. R. (eds.), LACUS Forum
XXX: Language, Thought and Reality. Houston: LACUS, pp. 207–14.
de Villiers, J., Fine, J., Ginsberg, G., Vaccarella, L. and Szatmari, P. 2007. A scale for
rating conversational impairment in autism spectrum disorder. Journal of Autism
and Developmental Disorders, 37(7), 1375–80.
Devlin, J. T., Moore, C. J. and Mummery, C. J. 2002. Anatomic constraints on cognitive
theories of category specificity. NeuroImage, 15(3), 675–85.
Dijkstra, K., Bourgeois, M. S., Allen, R. S. and Burgio, L. D. 2004. Conversational
coherence: discourse analysis of older adults with and without dementia. Journal
of Neurolinguistics, 17(4), 263–83.
Dolcos, F., Rice, H. J. and Cabeza, R. 2002. Hemispheric asymmetry and aging: right
hemisphere decline or asymmetry reduction. Neuroscience and Biobehavioral
Reviews, 26, 819–25.
Dolcos, F., LaBar, K. S. and Cabeza, R. 2004. Dissociable effects of arousal and valence
on prefrontal activity indexing emotional evaluation and subsequent memory: an
event-related fMRI study. NeuroImage, 23, 64–74.
Downes, W. 2000. The language of felt experience: emotion, evaluative and intuitive.
Language and Literature, 9(2), 99–121.
Drai, D. and Grodzinsky, Y. 2006a. A new empirical angle on the variability debate:
quantitative neurosyntactic analysis of a large data set from Broca’s aphasia. Brain
and Language, 96, 117–28.
2006b. The variability debate: more statistics, more linguistics. Brain and Language,
96 (2), 157–70.
Druks, J. 2002. Verbs and nouns – a review of the literature. Journal of Neurolinguistics,
15, 289–315.
Dubois, B., Feldman, H. H., Jacova, C. et al. 2007. Research criteria for the diagno-
sis of Alzheimer’s disease: revising the NINCDS-ADRDA criteria. The Lancet:
Neurology, 6(8), 734–46.
References 231
Duff Canning, S. J., Leach, L., Stuss, D. T. Ngo, L. and Black, S. E. 2004. Diagnostic
utility of abbreviated fluency measures in Alzheimer’s disease and vascular demen-
tia. Neurology, 62, 556–62.
Duncan J, and Owen A. M. 2000. Common regions of the human frontal lobe recruited
by diverse cognitive demands. Trends in Neurosciences, 23(10), 475–83.
Duong, A., Tardif, A. and Ska, B. 2003. Discourse about discourse: what is it and how
does it progress in Alzheimer’s disease? Brain and Cognition, 53, 177–80.
Duong, A., Giroux, F., Tardif, A. and Ska, B. 2005. The heterogeneity of picture-
supported narratives in Alzheimer’s disease. Brain and Language, 93, 173–84.
Ehrenreich, B., Hilden, J. and Malterud, K. 2007. Patients’ written life stories: a gateway
for understanding. Scandinavian Journal of Primary Health Care, 25(1), 33–7.
Ellis, M. P. and Astell, A. J. 2004. The urge to communicate in severe dementia. Brain
El-Manoufy, A. 1988. Intonation and meaning in spontaneous discourse. In Benson,
J., Cummings, M. and Greaves, W. (eds.), Linguistics in a Systemic Perspective.
Philadephia: John Benjamins, pp. 301–29.
Erlich, J. S., Obler, L. K. and Clark, L. 1997. Ideational and semantic contributions to
narrative production in adults with dementia of the Alzheimers type. Journal of
Communication Disorders, 30, 79–99.
Everingham, C. R., Heading, G. and Connor, L. 2006. Couples’ experiences of postnatal
depression: a framing analysis of cultural identity, gender and communication.
Social Science and Medicine, 62(7), 1745–56.
Fangmeier, T., Knauff, M., Ruff, C. C. and Sloutsky, V. 2006. fMRI evidence for a
three-stage model of deductive reasoning. Journal of Cognitive Neuroscience, 18,
320–34.
Ferstl, E. C. and von Cramon, D. Y. 2001. The role of coherence and cohesion in text
comprehension: an event-related fMRI study. Cognitive Brain Research, 11(3),
325–40.
2002. What does the fronto-median cortex contribute to language processing: coher-
ence or theory of mind? NeuroImage, 17, 1599–612.
2007. Time, space and emotion: fMRI reveals content specific activation during text
comprehension. Neuroscience Letters, 427, 159–64.
Ferri, C. P., Prince, M., Brayne, C. et al. for Alzheimer’s Disease International. 2005.
Global prevalence of dementia: a Delphi consensus study. Lancet, 366, 2112–17.
Ferstl, E. C., Rinck, M. and von Cramon, D. Y. 2005. Emotional and temporal aspects
of situation model processing during text comprehension: an event-related fMRI
study. Journal of Cognitive Neuroscience, 17, 724–39.
Ferstl, E. C., Neumann, J., Bogler, C. and von Cramon, D. Y. 2008. The extended lan-
guage network: a meta-analysis of neuroimaging studies on text comprehension.
Human Brain Mapping, 29, 581–93.
Fiebach, C. J., Schlesewsky, M., Lohmann, G., von Cramon, D. Y. and Friederici, A. D.
2005. Revisiting the localization of syntax: syntactic integration vs. syntactic
working memory. Human Brain Mapping, 24, 79–91.
Fillmore, C. J. 1968. The case for case. In Bach, E. and Harms, E. (eds.), Universals in
Linguistic Theory, New York: Holt, Rinehart, and Winston, pp. 1–88.
1976. Frame semantics and the nature of language. In Harnad, S. R., Steklis, H. D.
and Lancaster, J. (eds.), Origins and Evolution of Language and Speech. New
York: Annals of the New York Academy of Sciences, Vol. 280, pp. 20–32.
232 References
1982. Frame semantics. In The Linguistic Society of Korea (ed.), Linguistics in the
Morning Calm. Seoul: Hanshin Publishing Co, pp. 111–37.
Fine, J. 1991. The static and dynamic choices of responding: toward the process of build-
ing social reality by the developmentally disordered. In Ventola, E. (ed.) Functional
and Systemic Linguistics. The Hague: Mouton DeGruyter. pp. 213–34.
2006. Language in Psychiatry: A Handbook of Clinical Practice. London: Equinox
Publishing.
Fine, J., Bartolucci, G., Ginsberg, G. and Szatmari, P. 1991. The use of intonation to
communicate in pervasive developmental disorders. Journal of Child Psychiatry,
32(5), 771–82.
Firbas, J. 1992. Functional Sentence Perspective in Written and Spoken Communication.
Cambridge: Cambridge University Press.
Firth, J. R. 1957. Papers in Linguistics 1934–1951. London: Oxford University Press.
Fletcher, P. C., Happé, F., Frith, U. et al. 1995. Other minds in the brain: a func-
tional imaging study of ‘theory of mind’ in story comprehension. Cognition, 57,
109–28.
Fletcher, P. C., Shallice, T. and Dolan, R. J. 2000. ‘Sculpting the response space’ – An
account of left prefrontal activation at encoding. NeuroImage, 12, 404–41.
Fombonne, E, Zakarian, R., Bennett, A. Meng, L. and McLean-Heywood, D. 2006.
Pervasive developmental disorders in Montreal, Quebec, Canada: prevalence and
links with immunizations. Pediatrics, 118(1), 139–50.
Fox, N. C., Scahill, R. I., Crum, W. R. and Rossor, M. N. 1999. Correlation between
rates of brain atrophy and cognitive decline in AD. Neurology, 52(8), 1687–9.
Frijda, N. H. 1986. The Emotions. Cambridge: Cambridge University Press.
Friston, K. J., Harrison, L. and Penny, W. 2003. Dynamic causal modeling. NeuroImage,
19, 1273–302.
Frith, U. 2003. Autism: Explaining the Enigma, 2nd edition. Oxford: Blackwell.
Frith, U., and Frith, C. D. 2003. Development and neurophysiology of mentalising.
Philosophical Transactions of the Royal Society of London (Series B), 358, 459–73.
Frith, U. and Happé, F. 1994. Autism: beyond ‘theory of mind’. Cognition, 50,
115–32.
Galambos, J. A. 1986. Knowledge structures for common activities. In Galambos,
J. A., Abelson, R. P. and Black, J. B. (eds.), Knowledge Structures. Hillsdale,
NJ: Erlbaum, pp. 21–46.
Gallagher, H. L. and Frith, C. D. 2003. Functional imaging of theory of mind. Trends in
Cognitive Sciences, 7(2), 77–83.
Galvin, J. E., Lee, V. M. and Trojanowski, J. Q. 2001. Synucleinopathies: clinical and
pathological implications. Archives of Neurology, 58,186–90.
Garcia, L. and Joanette, Y. 1997a. Conversational topic-shifting analysis in dementia.
In Bloom, L., Obler, L. K., De Santi, S. and Erlich, J. (eds.), Discourse Analysis
and Application: Studies in Adult Clinical Populations. Hillsdale, NJ: Earlbaum,
pp. 169–92.
Gensler, O. 1977. Non-syntactic anaphora and frame semantics. Proceedings of the
Third Annual Meeting of the Berkeley Linguistics Society. Institute of Human
Learning, University of California, Berkeley, pp. 321–34.
Ghaziuddin, M. and Gerstein, L. 1996. Pedantic speaking style differentiates Asperger
syndrome from high-functioning autism. Journal of Autism and Developmental
Disorders, 26(5), 585–96.
References 233
Gilbert, S. J., Bird, G., Brindley, R. Frith, C. D. and Burgess, P. W. 2008. Atypical
recruitment of medial prefrontal cortex in autism spectrum disorders: an fMRI
study of two executive function tasks. Neuropsychologia, 46(9), 2281–91.
Godbout, L. and Doyon, J. 1995. Mental representation of knowledge following frontal-
lobe or postrolandic lesions. Neuropsychologia, 33, 1671–96.
Godbout, L., Cloutier, P., Bouchard, C., Braun, C. M. J. and Gagnon, S. 2004. Script gen-
eration following frontal and parietal lesions. Journal of Clinical and Experimental
Neuropsychology, 26(7), 857–73.
Goel, V., Grafman, J., Tajik, D., Gana, S. and Danto, D. (1997). A study of the perform-
ance of patients with frontal lobe lesions in a financial planning task. Brain, 120,
1805–22.
Goffman, E. 1959. The Presentation of Self in Everyday Life. New York: Doubleday.
1961. Asylums: Essays on the Social Situation of Mental Patients and other Inmates.
New York: Doubleday Anchor.
1974. Frame Analysis: An Essay on the Organization of Experience. New York: Harper
and Row.
1986. Stigma: Notes on the Management of Spoiled Identity. New York:
Touchstone.
Goldin, P. R., Hutcherson, C. A. C., Ochsner, K. N., Glover, G. H., Gabrieli, J. D. E.
and Gross, J. J. 2005. The neural bases of amusement and sadness: a comparison
of block contrast and subject-specific emotion intensity regression approaches.
Goldman-Eisler, F. 1968. Psycholinguistics: Experiments in Spontaneous Speech. New
York: Academic Press.
Grady, C., MacIntosh, A. R., Beig, S., Keightley, M., Burian, H. and Black, S. 2003.
Evidence from functional imaging of a compensatory prefrontal network in
Alzheimers disease. The Journal of Neuroscience, 23(3), 986–93.
Graff-Radford, N. R. and Woodruff, B. K. 2007. Frontotemporal dementia. Seminars in
Neurology, 27(1), 48–57.
Grafman, J. 1989. Plans, actions, and mental sets: managerial knowledge units in the
frontal lobes. In Perecman, E. (ed.), Integrating Theory and Practice in Clinical
Neuropsychology. Hillsdale, NJ: Erlbaum, pp. 93–138.
2002. The structured event complex and the human prefrontal cortex. In Stuss, D. T.
and Knight, R. T. (eds.), The Frontal Lobes. New York: Oxford University Press,
pp. 292–310.
Gregory, M. 1988. Generic situation and register: a functional view of communication.
In Benson, J., Cummings, M. and Greaves, W. (eds.), Linguistics in a Systemic
Perspective. Philadelphia: John Benjamins, pp. 301–29.
2002. Phasal analysis within communication linguistics: two contrastive discourses.
In Fries, P., Cummings, M., Lockwood, D. and Spruiell, W. C. (eds.), Relations
and Functions within and around Language. London: Continuum.
2009a. Remarks on a theory of grammar for a socio-cognitive linguistics. In de
Villiers, J. and Stainton, R. (eds.), pp. 269–81.
2009b. Arguments, roles, relations, prepositions and case: proposals within a socio-
cognitive grammar of English. In J. de Villers and R. Stainton (eds.), pp. 282–311.
2009c. Notes on communication linguistics. In de Villiers, J. and Stainton, R. (eds.),
pp. 143–266.
2009d. English patterns. In de Villiers, J. and Stainton, R. (eds.), pp. 1–142.
234 References
Grice, H. P. 1975. Logic and conversation. In Cole, P. and Morgan, J. L. (eds.), Syntax
and Semantics. Vol. III: Speech Acts. New York: Academic Press, pp. 41–58.
Grill-Spector, K. 2003. The neural basis of object perception. Current Opinion in
Neurobiology, 13(2), 159–66.
Grossman, M., Cooke, A., DeVita, C. et al. 2002. Age-related changes in working mem-
ory during sentence comprehension: an fMRI study. NeuroImage, 15(2), 302–17.
Gumperz, J. J. and Hymes, D. (eds.) 1972. Directions in Sociolinguistics: The
Ethnography of Communication. New York: Rinehart and Winston.
Gussenhoven, C. 2004. The Phonology of Tone and Intonation. Cambridge: Cambridge
University Press.
Gustafson, D., Rothenberg, E., Blennow, K., Steen, B. and Skoog, I. 2003. An 18-year
follow-up of overweight and risk of Alzheimer disease. Archives of Internal
Medicine, 163, 1524–8.
Guye, M., Bartolomei, F. and Ranjeva, J. P. 2008. Imaging structural and functional
connectivity: towards a unified description of brain organization? Current Opinion
in Neurology, 21(4), 393–403.
Halliday, M. A. K. 1967. Intonation and Grammar in British English. The Hague:
Mouton.
1970. A Course in Spoken English: Intonation. Oxford: Oxford University Press.
1976. Halliday: System and Function in Language: Selected Papers, (ed. by
G. Kress). London: Oxford University Press.
1977. Text as Semantic Choice in Social Contexts. In van Dijk, T., Janos, A. and Petofi,
S. (eds.), Grammars and Descriptions. New York: Walter de Gruyter, pp. 176–225.
1978. Language as Social Semiotic: The Social Interpretation of Language and
Meaning. London: Edward Arnold.
1984. The Semiotics of Culture and Language. London: Pinter.
1985. Dimensions of discourse analysis: grammar. In van Dijk, T. A. (ed.), Handbook
of Discourse Analysis, Vol. II: Dimensions of Discourse. London: Academic Press,
pp. 29–56.
1987. Spoken and written modes of meaning. In Horowitz, R. and Samuels, S. J.
(eds.), Comprehending Oral and Written Language. New York: Academic Press,
pp. 55–82.
1989. Spoken and Written Language. Oxford: Oxford University Press.
1994. Introduction to Functional Grammar (2nd edition). London: Arnold.
Halliday, M. A. K. and Greaves, W. S. 2008. Intonation and the Grammar of English.
London: Equinox.
Halliday, M. A. K. and Hasan, R. 1976. Cohesion in English. London: Longman.
1989. Language, Text and Context, 2nd edition. London: Longman.
Happé, F. 1993. Communicative competence and theory of mind in autism: a test of
relevance theory, Cognition, 48, 101–19.
1994. An advanced test of theory of mind: understanding of story characters’
thoughts and feelings by able autistic, mentally handicapped, and normal children
and adults, Journal of Autism and Developmental Disorders, 24(2), 129–54.
Happé, F. 1997. Central coherence and theory of mind in autism: reading homographs
in context, British Journal of Developmental Psychology, 15, 1–12.
Happé, F. and Frith, U. 2006. The weak coherence account: detail-focused cogni-
tive style in autism spectrum disorders, Journal of Autism and Developmental
Disorders, 36(1), 5–25.
References 235
Hasan, R. 1985. Linguistics, Language, and Verbal Art. Oxford: Oxford University
Press.
Hébert, R., Lindsay, J., Verreault, R., Rockwood, K., Hill, G. and Dubois, M. F. 2000.
Vascular dementia: incidence and risk factors in the Canadian Study of Health and
Aging. Stroke, 31(7), 1487–93.
Henson, R. N. A., Shallice, T, and Dolan, R. J. 1999. Right prefrontal cortex and epi-
sodic memory retrieval: a functional MRI test of the monitoring hypothesis. Brain,
122, 1367–81.
Herholz, K., Salmon, E., Perani, D. et al. 2002. Discrimination between Alzheimer
dementia and controls by automated analysis of multicenter FDG PET. NeuroImage,
17, 302–16.
Heritage, J. and Maynard, D. W. (eds.) 2006. Communication in Medical Care: Interaction
Between Primary Care Physicians and Patients. Cambridge: Cambridge University
Press.
Hess, C. W., Sefton, K. M. and Landry, R. G. 1986. Sample size and type-token ratios
for oral language of preschool children. Journal of Speech and Hearing Research,
29(1), 129–34.
Hier, D. B., Hagenlocker, K. and Shindler, A. 1985. Language disintegration in demen-
tia: effects of etiology and severity, Brain and Language, 25(1), 117–33.
Hier, D. B., Stein, R. and Caplan, L. R. 1985. Cognitive and behavioral deficits after
right hemisphere stroke. Current Concepts in Cerebrovascular Disease, 20, 1–5.
Hill, E. L. 2004. Executive dysfunction in autism. Trends in Cognitive Sciences, 8,
26–32.
Hill, G., Forbes, W., Berthelot, J.-M., Lindsay, J. and McDowell, I. 1996. Dementia
among seniors. Health Reports, 8(2), 7–10.
Hirono, N., Mori, E., Ishii, K. et al. 1998a. Hypofunction in the posterior cingulate
gyrus correlates with disorientation for time and place in Alzheimers disease.
Journal of Neurology, Neurosugery and Psychiatry, 64, 552–4.
1998b. Regional metabolism: associations with dyscalculia in Alzheimer’s disease.
Journal of Neurology, Neurosurgery and Psychiatry, 65, 913–6.
Holmes, D. I. 1992. A stylometric analysis of Mormon scripture and related texts.
Journal of the Royal Statistical Society (A), 155, 91–120.
1994. Authorship attribution. Computers and the Humanities, 28, 87–106.
Holmes, D. I. and Singh, S. 1996. A stylometric analysis of conversational speech of
aphasic patients. Literary and Linguistic Computing, 11, 45–60.
Honoré, A. 1979. Some simple measures of richness of vocabulary. Association of
Literary and Linguistic Computing Bulletin, 7, 172–7.
Huddleston, R. D. and Pullum, G. K. 2002. The Cambridge Grammar of the English
Language. Cambridge: Cambridge University Press.
Hudson, R. 1984. Word Grammar. Oxford: Blackwell.
2007. Language Networks. The New Word Grammar. Oxford: Oxford University
Press.
Hymes, D. 1962. The ethnography of speaking. In Gladwin, T. and Strutevant, W. C.
(eds.), Anthropology and Human Behaviour. Washington, DC: Anthropological
Society of Washington, pp. 15–53.
1971. On Communicative Competence. Philadelphia: University of Pennsylvania Press.
1974. Foundations in Sociolinguistics: An Ethnographic Approach. Philadelphia:
University of Pennsylvania Press.
236 References
Hynd, M. R., Scott, H. L. and Dodd, P. R. 2004. Glutamate-mediated excitotoxicity
and neurodegeneration in Alzheimer’s disease. Neurochemistry International, 45,
583–95.
Ilmberger, J., Rau, S., Noachtar, S., Arnold, S. and Winkler, P. 2002. Naming tools
and animals: asymmetries observed during direct electrical cortical stimulation.
Neuropsychologia, 40(7), 695–700.
Inzitari, D., Lamassa, M., Pantoni, L. and Basile, A. M. 2004. Therapy of vascular
dementias, Archives of Gerontology and Geriatrics Supplement, 9, 229–34.
Jackendoff, R. 1990. Semantic Structures. Cambridge: MIT Press.
2002. Foundations of Language. Oxford: Oxford University Press.
(In preparation). Conceptual Semantics. To appear in Semantics: An International
Handbook of Natural Language Meaning, de Gruyter.
Jacob, C. P., Koutsilieri, E., Bartl, J. et al. 2007. Alterations in expression of gluta-
matergic transporters and receptors in sporadic Alzheimers disease. Journal of
Alzheimer’s Disease, 11(1), 97–116.
Jefferson, G. 1974. Error correction as an interactional resource. Language in Society,
2, 181–99.
1985. An exercise in the transcription and analysis of laughter. In van Dijk, T. A.
Handbook of Discourse Analysis, Vol. III. London: Academic Press, pp. 25–34.
2004. Glossary of transcript symbols with an introduction. In Lerner, G. H. (ed.),
Conversation Analysis: Studies from the First Generation. Amsterdam: John
Benjamins, pp. 13–31.
Jellinger, K. A. 2002. Vascular-ischemic dementia: an update. Journal of Neural
Transmission Supplementum, 62, 1–23.
2007. The enigma of mixed dementia. Alzheimers and Dementia, 3(1), 40–53.
Jellinger, K. A. and Attems, J. 2007. Neuropathological evaluation of mixed dementia.
Journal of the Neurological Sciences, 257(1–2), 80–7.
Jolliffe, T. and Baron-Cohen, S. 1999a. A test of central coherence theory: linguistic
processing in high-functioning adults with autism or Asperger syndrome: is local
coherence impaired? Cognition, 71, 149–85.
1999b. The strange stories test: a replication with high-functioning adults with autism
or Asperger syndrome. Journal of Autism and Developmental Disorders, 29(5),
395–419.
Jones, S., Jonsson, E. and Bäckman, L. 2006. Differential verbal fluency deficits in the
preclinical stages of Alzheimer’s disease and vascular dementia. Cortex, 42(3),
347–55.
Just, M. A., Cherkassky V. L., Keller, T. A. and Minshew, N. J. 2004. Cortical activa-
tion and synchronization during sentence comprehension in high-functioning
autism: evidence of underconnectivity. Brain, 127, 1811–21.
Kahneman, D. and Frederick, S. 2007. Frames and brains: elicitation and control of
response tendencies. Trends in Cognitive Sciences, 11(2), 45–6.
Kanner, Leo. 1943. Autistic disturbances of affective contact. Nervous Child, 2, 217–50.
Kantarci, K., Weigand, S. D., Petersen, R. C. et al. 2007. Longitudinal 1H MRS changes
in mild cognitive impairment and Alzheimer’s disease. Neurobiology of Aging, 28,
1330–9.
Kashani A., Lepicard, E., Poirel, O. et al. 2008. Loss of VGLUT1 and VGLUT2 in
the prefrontal cortex is correlated with cognitive decline in Alzheimer disease.
Neurobiology of Aging, 29 (11), 1619–30.
References 237
Kemper, S., Herman, R. E. 2006. Age differences in memory-load interference effects
in syntactic processing. Journal of Gerontology Series B: Psychological Sciences
and Social Sciences Online, 61, 327–32.
Kemper, S. and Sumner, A. 2001. The structure of verbal abilities in young and older
adults. Psychology and Aging, 16, 312–22.
Kemper, S., Greiner, L. H., Marquis, J., Prenovost, K. and Mitzner, T. L. 2001a.
Language decline across the life span: findings from the Nun Study. Psychology
and Aging, 16(2), 227–39.
Kemper, S., Marquis, J. and Thompson, M. 2001b. Longitudinal change in language
production: effects of aging and dementia on grammatical complexity and propo-
sitional content. Psychology and Aging, 16, 600–14.
Kern, R. S, Van Gorp, W. G., Cummings, J. L., Brown, W. S. and Osato, S. S. 1992.
Confabulation in Alzheimer’s disease. Brain and Cognition, 19, 172–82.
Kertesz, A. 2006. Progress in clinical neurosciences: frontotemporal dementia-Pick’s
disease. Canadian Journal of Neurological Science, 33(2), 141–8.
Kintsch, W. 1974. The Representation of Meaning in Memory. Hillsdale, NJ: Lawrence
Erlbaum Associates.
1988. The role of knowledge in discourse comprehension: a construction–integration
model. Psychological Review, 95, 163–82.
Kintsch, W., and Keenan, J. M. 1973. Reading rate and retention as a function of the
number of the propositions in the base structure of sentences. Cognitive Psychology,
5, 257–74.
Klafki, H.-W., Staufenbiel, M., Kornhuber, J. and Wiltfang, J. 2006. Therapeutic
approaches to Alzheimer’s disease. Brain, 129, 2840–55.
Klages, Jennifer. 2006. Magnetic resonance spectroscopy, magnetic resonance imag-
ing and neuropsychology in persons with vascular cognitive impairment and
mild Alzheimer’s disease: will new technology help with old problems? PhD
Dissertation, Department of Psychology, Dalhousie University.
Knapp, M., Prince, M., Albanese, E. et al. 2007. Dementia UK. A report to the
Alzheimer’s Society on the prevalence and economic cost of dementia in the
UK produced by King’s College London and London School of Economics.
London: Alzheimer’s Society.
Knopman, D. S., Parisi, J. E., Boeve, B. F. et al. 2003. Vascular dementia in a popula-
tion-based autopsy study. Archives of Neurology, 60(4), 569–75.
Kontiola, P., Laaksonen, R., Sulkava, R. and Erkinjuntti, T. 1990. Pattern of language
impairment is different in Alzheimer’s disease and multi-infarct dementia. Brain
Kopelman, M. D. 1987. Amnesia: organic and psychogenic. The British Journal of
Koshino, H., Carpenter, P. A., Minshew, N. J., Cherkassky, V. L., Keller, T. A. and Just,
M. A. 2005. Functional connectivity in an fMRI working memory task in high-
functioning autism. NeuroImage, 24, 810–21.
Koshino, H., Kana, R. K., Keller, T. A., Cherkassky, V. L., Minshew, N. J. and Just, M. A.
2008. FMRI investigation of working memory for faces in autism: visual coding
and underconnectivity with frontal areas. Cerebral Cortex, 18(3), 289–300.
Krawczyk, D. C. 2002. Contributions of the prefrontal cortex to the neural basis
of human decision making. Neuroscience and Biobehavioral Reviews, 26,
631–64.
238 References
Krishnan, K. R. R., Charles, H. C., Murali Doraiswamy, P. et al. 2003. Randomized,

placebo-controlled trial of the effects of donepezil on neuronal markers and hip-
pocampal volumes in Alzheimer’s disease. American Journal of Psychiatry, 160,
2003–11.
Kuperberg G. R., Kreher D. A., Sitnikova T., Caplan, D. N. and Holcomb, P. J. 2007.
The role of animacy and thematic relationships in processing active English
sentences: evidence from event-related potentials. Brain and Language, 100,
223–37.
Kuperberg, G. R., Lakshmanan, B. M., Caplan, D. N. and Holcomba, P. J. 2006.
Making sense of discourse: an fMRI study of causal inferencing across sentences.
Labov, W. and Waletzky, J. 1967. Narrative analysis. In Helm, J. (ed.), Essays on Verbal
and Visual Arts. Seattle: University of Washington Press, pp. 12–44.
Lafosse, J. M., Reed, B. R., Mungas, D., Sterling, S. B., Wahbeh, H. and Jagust, W. J.
1997. Fluency and memory differences between ischemic vascular dementia and
Alzheimer’s disease. Neuropsychology, 11, 514–22.
Lakoff, G. 1973. Hedges: a study in meaning criteria and the logic of fuzzy concepts.
Journal of Philosophical Logic, 2(4), 458–508.
1987. Women, Fire, and Dangerous Things: What Categories Reveal About the Mind.
Chicago: University of Chicago Press.
Lakoff, R. 1975. Language and Woman’s Place. New York: Harper and Row.
Lamb, S. M. 1966. Outline of Stratificational Grammar. Washington: Georgetown
University Press.
1998. Pathways of the Brain: The Neurocognitive Basis of Language. Amsterdam: John
Benjamins.
Lane, R. D., Fink, G. R., Chau, P. M. and Dolan, R. J. 1997. Neural activation dur-
ing selective attention to selective emotional responses. Neuroreport, 8(18),
3969–72.
Langa, K. M., Foster, N. L. and Larson, E. B. 2004. Mixed dementia: Emerging con-
cepts and therapeutic implications. Review. Journal of the American Medical
Association, 292(23), 2901–8.
Lee, M. and Thompson, C. K. 2004. Agrammatic aphasic production and comprehen-
sion of unaccusative verbs in sentence contexts. Journal of Neurolinguistics, 17(4),
315–30.
Levinson, S. C. 1983. Pragmatics. Cambridge: Cambridge University Press.
Li, L., Sengupta, A., Haque, N., Grundke-Iqbal, I. and Iqbal, K. 2004. Memantine
inhibits and reverses the Alzheimer type abnormal hyperphosphorylation of tau
and associated neurodegeneration. FEBS Letters, 566, 261–9.
Looi, J. C. L. and Sachdev, P. S. 1999. Differentiation of vascular dementia from AD on
neuropsychological tests. Neurology, 53, 670–8.
Loukusa, S., Leinonen, E., Jussila, K. et al. 2007. Answering contextually demand-
ing questions: Pragmatic errors produced by children with Asperger syndrome or
high-functioning autism. Journal of Communication Disorders, 40(5), 357–81.
Luan, P. K., Wager, T., Taylor, S. F. and Liberzon, I. 2002. Functional neuroanatomy
of emotion: a meta-analysis of emotion activation studies in PET and fMRI.
Neuroimage, 16, 331–48.
Luchsinger, J. A. and Mayeux, R. 2004. Dietary factors and Alzheimer’s disease. Lancet
Neurology, 3, 579–87.
References 239
Luria, A. R. 1965. Two kinds of motor perseveration in massive injuries of the frontal
lobes. Brain, 88, 1–10.
Lyons, K., Kemper, S., LaBarge, E., Ferraro, F. R., Balota, D. and Storandt, M. 1994.
Oral language and Alzheimer’s disease: a reduction in syntactic complexity. Aging
and Cognition, 2, 271–94.
McAlonan, G. M., Cheung, V., Cheung, C. et al. 2005. Mapping the brain in autism: a
voxel-based MRI study of volumetric differences and intercorrelations in autism.
Brain, 128(2), 268–76.
McCawley, J. D. 1988. The Syntactic Phenomena of English, Vol. I–II. Chicago: University
of Chicago Press.
McKeith, I. G., Galasko, D., Kosaka, K. et al. 1996. Consensus guidelines for the clini-
cal and pathologic diagnosis of dementia with Lewy bodies (DLB): report of the
consortium on DLB international workshop. Neurology, 47, 1113–24.
McKenna, P. and Oh, T. 2005. Schizophrenic Speech: Making Sense of Bathroots and
Ponds that Fall in Doorways. Cambridge: Cambridge University Press.
Mackenzie, C. 2000. Adult spoken discourse: the influences of age and educa-
tion. International Journal of Language and Communication Disorders, 35(2),
269–85.
McKhann, G., Drachman, D. A., Folstein, M., Katzman, R., Price, D. L. and Stadlan, E. M.
1984. Clinical diagnosis of Alzheimer’s disease – report of the NINCDS –ADRDA
work group under the auspices of Department of Health and Human Services Task
Force on Alzheimer’s disease. Neurology, 34, 939–44.
McPherson, S. and Cummings, J. 1996. Neuropsychological aspects of vascular demen-
tia. Brain and Cognition, 31, 269–82.
McShane, R., Areosa, S. A. and Minakaran, N. 2006. Memantine for dementia. Cochrane
Database Systematic Reviews, 19(2), CD003154.
MacWhinney, B. 2000. The CHILDES Project: Tools for Analyzing Talk. 3rd edition.
Mahwah, NJ: Lawrence Erlbaum Associates.
MacWhinney, B. and Osler, H. 1977. Verbal planning functions in children’s speech.
Child Development, 48, 978–85.
Malinowski, B. 1923. The problem of meaning in primitive languages. In Ogden, C. K.
and Richards, I. A. (eds.), The Meaning of Meaning. New York: Harcourt Brace,
pp. 296–336.
1935. An ethnographic theory of language. In Coral Gardens and their Magic,
Vol. II. London: Allen and Unwin, pp. 4–78.
Mamelak, M. 2007. Alzheimer’s disease, oxidative stress and gammahydroxybutyrate.
Neurobiology of Aging, 28(9),1340–60.
Marczinski, C. A. and Kertesz, A. 2006. Category and letter fluency in semantic demen-
tia, primary progressive aphasia, and Alzheimer’s disease. Brain and Language,
97, 258–65.
Marini, A., Carlomagno, S., Caltagirone, C. and Nocentini, U. 2005. The role played by
the right hemisphere in the organization of complex textual structures. Brain and
Language, 93(1), 46–54.
Martin, A. 2007. The representation of object concepts in the brain. Annual Review of
Psychology, 58, 25–45.
Martin, I. and McDonald, S. 2004. An exploration of causes of non-literal lan-
guage problems in individuals with Asperger syndrome. Journal of Autism and
Developmental Disorders, 34, 311–28.
240 References
Martin, J. R. 2000. Beyond exchange: appraisal systems in English. In Hunston, S. and

Thompson, G. (eds.), Evaluation in Text: Authorial Stance and the Construction of
Discourse. Oxford: Oxford University Press, pp. 142–75.
Martin, J. R. and White, P. R. R. 2005. Appraisal: The Language of Attitude and
Intersubjective Stance. London and New York: Palgrave.
Mason, R., and Just, M. A. 2007. Lexical ambiguity in sentence comprehension. Brain
Research, 1146, 115–27.
Mason, R. A., Williams, D. L., Kana, R. K., Minshew, N. and Just, M. A. 2008. Theory
of mind disruption and recruitment of the right hemisphere during narrative com-
prehension in autism. Neuropsychologia, 46(1), 269–80.
Massoud, F., Devi, G., Stern, Y. et al. 1999. A clinicopathological comparison of
community-based and clinic-based cohorts of patients with dementia. Archives of
Neurology, 56, 1368–73.
Matsuda, H. 2007. Role of neuroimaging in Alzheimer’s disease, with emphasis on
brain perfusion SPECT. Journal of Nuclear Medicine, 48(8), 1289–1300.
Mayeux, R. 2003. Epidemiology of neurodegeneration. Annual Review of Neuroscience,
26, 81–104.
Mendez, M. F., Cherrier, M. M. and Perryman, K. M. 1997. Differences between
Alzheimer’s disease and vascular dementia on information processing measures.
Brain and Cognition, 34, 301–10.
Mesulam, M. M. 1998. From sensation to cognition. Brain, 121(6), 1013–52.
2000. Aging, Alzheimer’s disease, and dementia: clinical and neurobiological per-
spectives. In Principles of Behavioural and Cognitive Neurology, 2nd edition.
Oxford: Oxford University Press, pp. 439–522.
Michael, E. B., Keller, T. A., Carpenter, P. A. and Just, M. A. 2001. FMRI investigation
of sentence comprehension by eye and by ear: modality fingerprints on cognitive
processes. Human Brain Mapping, 13, 239–52.
Minsky, M. 1975. A framework for representing knowledge. In Winston, P. (ed.), The
Psychology of Computer Vision. New York: McGraw-Hill, pp. 211–77.
1977. Frame-system theory. In Johnson-Laird, P. N. and Wason, P. C. (eds.),
Thinking: Readings in Cognitive Science. Cambridge: Cambridge University
Press, pp. 355–76.
Monti, M. M., Osherson, D. N., Martinez, M. J. and Parson, L. M. 2007. Functional neu-
roanatomy of deductive inference: a language-independent distributed network.
Moorhouse, P. and Rockwood, K. 2008. Vascular cognitive impairment: current con-
cepts and clinical developments. The Lancet: Neurology, 7(3), 246–55.
Moss, H. E., Tyler, L. K. and Taylor, K. I. 2007. Conceptual structure. In Gaskell, G.
(ed.), Oxford Handbook of Psycholinguistics, New York: Oxford University Press,
pp. 217–33.
Nankano, H., and Blumstein, S. E. 2004. Deficits in thematic integration processes in
Broca’s and Wernicke’s aphasia. Brain and Language, 88, 96–107.
Nespoulous, J-L., Code, C., Virbel, J. and Lecours, A. R. 1998. Hypothesis on the dis-
sociation between referential and modalizing verbal behavior in aphasia. Applied
Psycholinguistics, 19, 311–31.
Neuropathology Group. 2001. Medical research council cognitive function and aging
study. Pathological correlates of late-onset dementia in a multicentre, community-
based population in England and Wales. Lancet, 357, 169–75.
References 241
Nicholas, M., Obler, L. K., Albert, M. L. and Helm-Estabrooks, N. 1985. Empty
speech in Alzheimer’s disease and fluent aphasia. Journal of Speech and Hearing
Research, 28(3), 405–10.
Nicholson, R. and P. Szatmari 2003. Genetic and neurodevelopmental influences in
autistic disorder. Canadian Journal of Psychiatry, 48(8), 27–38.
Nobre, A. C., Coull, J. T., Frith, C. D. and Mesulam, M. M. 1999. Orbitofrontal cor-
tex is activated during breaches of expectation in tasks of visual attention. Nature
Neuroscience, 2, 11–12.
Norbury, C. F. and Bishop, D. 2002. Inferential processing and story recall in children
with communication problems: a comparison of specific language impairment,
pragmatic language impairment and high-functioning autism. International
Journal of Language and Communication Disorders, 37(3), 227–51.
Norman, D. A. and Shallice, T. 1986. Attention to action: willed and automatic control
of behaviour. Centre for Human Information Processing (Technical Report #99).
Reprinted in revised form in Davidson, R. J., Schwartz, G. E. and Shapiro, D. (eds.)
1986, Consciousness and Self-Regulation: Vol. IV, New York: Plenum, pp. 1–18.
Noveck, I. A., Goel, V. and Smith, K. W. 2004. The neural basis of conditional reasoning
with arbitrary content. Cortex, 40, 613–22.
Nussbaum, R. L. and Ellis, C. E. 2003. Alzheimer’s disease and Parkinson’s disease.
New England Journal of Medicine, 348, 1356–64.
O’Brien, J. T. 2006. Vascular cognitive impairment. Review. American Journal of
Geriatric Psychiatry, 14(9), 724–33.
O’Brien, J. T., Erkinjuntti, T. and Reisberg, B. et al. 2003. Vascular cognitive impair-
ment. Review. The Lancet: Neurology, 2, 89–98.
Ochsner, K. N., Bunge, S. A., Gross, J. J. and Gabrieli, J. D. 2002. Rethinking feel-
ings: an FMRI study of the cognitive regulation of emotion. Journal of Cognitive
Neuroscience, 14, 1215–29.
Orange, J. B. and Kertesz, A. 2000. Discourse Analyses and Dementia, Brain and
Language, 71, 172–4.
Owen, A. J. and Leonard, L. B. 2002. Lexical diversity in the spontaneous speech of
children with specific language impairment: application of D. Journal of Speech,
Language, and Hearing Research, 45(5), 927–37.
Ozonoff, S. and Miller, J. 1996. An exploration of right-hemisphere contributions to the
pragmatic impairments of autism. Brain and Language, 52, 411–34.
Pantoni, L. and Simoni, M. 2003. Pathophysiology of cerebral small vessels in vascular
cognitive impairment. International Journal of Psychogeriatrics, Supplement 1,
59–65.
Patry, R. and Nespoulous, J.-L. 1990. Discourse analysis in linguistics. In Joanette, Y.
and Brownell, H. H. (eds.), Discourse Ability and Brain Damage: Theoretical and
Empirical Perspectives. New York: Springer-Verlag, pp. 3–29.
Paxton, J. L., Barch, D. M., Racine, C. A. and Braver, T. S. 2008. Cognitive control, goal
maintenance and prefrontal function in healthy aging. Cerebral Cortex, 18(5),
1010–28.
Perry, R. J. and Hodges, J. R. 1996. Spectrum of memory dysfunction in degenerative
disease. Current Opinion in Neurology, 9(4), 281–5.
Petrides, M., Alivisatos, B. and Evans, A. C. 1995. Functional activation of the human
ventrolateral prefrontal cortex during mnemonic retrieval of verbal information.
Proceedings of the National Academy of Science USA, 92, 5803–7.
242 References
Petrides, M., Alivisatos, B. and Frey, S. 2002. Differential activation of the human
orbital, mid-vetrolateral, and mid-dorsolateral prefrontal cortex during the process-
ing of visual stimuli. Proceedings of the National Academy of Sciences, USA, 99,
5649–54.
Phan, K. L., Wager, T., Taylor, S. F. and Liberzon, I. 2002. Functional neuroanatomy of
emotion: a meta-analysis of activation studies in PET and fMRI. NeuroImage, 16,
331–48.
Pike, K. L. and Pike, E. G. 1982. Grammatical Analysis. Dallas: Summer Institute of
Linguistics.
Poore, Q. E., Rapport, L. J., Fuerst, D. R. and Keenan, P. 2006. Word list generation
performance in Alzheimer’s disease and vascular dementia. Neuropsychology,
Development and Cognition. Section B, Aging, Neuropsychology and Cognition,
13(1), 86–94.
Powell, A. L., Cummings, J. L., Hill, M. A. and Benson, D. F. 1988. Speech and language
alterations in multi-infarct dementia. Neurology, 38(5), 717–9.
Prizant, B. and Duchan, J. 1981. The functions of immediate echolalia in autistic chil-
dren. Journal of Speech and Hearing Disorders, 46, 241–9.
Propp, V. 1928 [trans. 1958]. The Morphology of the Folktale. Philadelphia: American
Folktale Society.
Public Health Agency of Canada. 1999. Surveillance for Creutzfeldt-Jacob disease in
Canada. Canada Communicable Disease Report – Vol. XXV.
Creutzfeldt-Jakob Disease Surveillance System (CJD-SS) www.phac-aspc.gc.ca/
hcai-iamss/cjd-mcj/cjdss-ssmcj/stats_E.html, accessed 2007.
Quirk, R., Greenbaum, S., Leech, G. and Svartik, J. 1985. A Comprehensive Grammar
of the English Language. London: Longman.
Ramanathan-Abbott, V. 1994. Interactional differences in Alzheimer’s discourse: an
examination of Alzheimer’s speech across two audiences. Language and society,
23, 31–58.
Razvi, S. S., Davidson, R., Bone, I. and Muir, K. W. 2005. The prevalence of cerebral
autosomal dominant arteriopathy with subcortical infarcts and leucoencephalopa-
thy (CADASIL) in the west of Scotland. Journal of Neurology, Neurosurgery and
Psychiatry, 76(5), 739–41.
Ready, R. E., Ott, B. R. and Grace, J. 2003. Amnestic behaviour in dementia: symp-
toms to assist in early detection and diagnosis. Journal of the American Geriatrics
Society, 51, 32–7.
Riverberi, C., Cherubini, P., Rapisarda, A. et al. 2007. Neural basis of generation
of conclusions in elementary deduction. Journal of NeuroImage, 38(4),
752–62.
Rocca, W. A. and Knopman, D. S. 2003. Prevalence and incidence patterns of vascular
dementia. In Bowler, J. V. and Hatchinski, V. (eds.), Vascular Cognitive Impairment:
Preventable Dementia. New York: Oxford University Press, pp. 21–32.
Rochester, S. R. and Martin, J. R. 1979. Crazy talk: A Study of the Discourse of
Schizophrenic Speakers. New York: Plenum.
Rockwood, K., Graham, J. E. and Fay, S. 2002. Goal setting and attainment in Alzheimer’s
disease patients treated with donepezil. Journal of Neurology, Neurosurgery and
Rockwood, K., Moorhouse, P. K., Song, X. et al. 2007. Disease progression in vas-
cular cognitive impairment: cognitive, functional and behavioural outcomes in
References 243
the Consortium to Investigate Vascular Impairment of Cognition (CIVIC) cohort

study. Journal of the Neurological Sciences, 252(2), 106–12.
Roman, G. C. 2005. Vascular dementia prevention: a risk factor analysis. Review.
Cerebrovascular Diseases, 20 Supplement 2, 91–100. Epub 2 Dec 2005. Review.
Roman, G. C., Tatemichi, T. K., Erkinjuntti, T. et al. 1993. Vascular dementia – diag-
nostic criteria for research studies. Report of the NINDS-AIREN International
Workshop. Neurology, 43, 250–60.
Rosch, E. 1983. Prototype classification and logical classification: the two systems.
In Scholnick, E. (ed.), New Trends in Cognitive Representation: Challenges to
Piaget’s Theory. Hillsdale, NJ: Lawrence Erlbaum Associates pp. 73–86.
Royall, D. R., Palmer, R., Chiodo, L. K. and Polk, M. J. 2005. Executive control medi-
ates memory’s association with change in instrumental activities of daily living: the
Freedom House Study. Journal of the American Geriatric Society, 53(1), 11–17.
Royall, D. R., Lauterbach, E. C., Cummings, J. L. et al. 2002. Executive control func-
tion: a review of its promise and challenges for clinical research. A report from the
Committee on Research of the American Neuropsychiatric Association. Review.
Journal of Neuropsychiatry and Clinical Neurosciences, 14(4), 377–405.
Ruchoux, M. M. and Maurage, C. A. 1997. CADASIL: cerebral autosomal domi-
nant arteriopathy with subcortical infarcts and leukoencephalopathy. Journal of
Neuropathology and Experimental Neurology, 56(9), 947–64.
Rugg, M. D., Fletcher, P. C., Frith, C. D., Frackowiak, R. S. J. and Dolan, R. J. 1996.
Differential activation of the prefrontal cortex in successful and unsuccessful
memory retrieval. Brain, 119(6), 2073–83.
Rushworth, M. F., Behrens, T. E., Rudebeck, P. H. and Walton, M. E. 2007. Contrasting
roles for cingulate and orbitofrontal cortex in decisions and social behaviour.
Trends in Cognitive Sciences, 11(4),168–76.
Russell, J. (ed.) 1998. Autism as an Executive Disorder. Oxford: Oxford University
Press.
Rusted, J. and Shepherd, L. 2002. Action-based memory in Alzheimer’s disease: a lon-
gitudinal look at tea-making. Case Studies in Neuropsychology, Neuropsychiatry
and Behavioural Neurology, 8(1–2),111–26.
Sachdev, P. S. and Looi, J. C. L. 2003. Neuropsychological separation of vascular
dementia and Alzheimer’s dementia. In Bowler, J. V. and Hachinski, V. (eds.),
The Vascular Dementias: A Review and Re-appraisal. Oxford: Oxford University
Press, pp. 153–75.
Sacks, H. 1992a. Lectures on Conversation: Vol. I. Cambridge, MA: Blackwell.
1992b. Lectures on Conversation: Vol. II. Cambridge, MA: Blackwell.
Sacks, H., Schegloff, E. A. and Jefferson, G. 1974. A simplest systematics for the organ-
ization of turn-taking for conversation. Language, 50, 696–735.
Saldaña, D. and Frith, U. 2007. Do readers with autism make bridging inferences from
world knowledge? Journal of Experimental Child Psychology, 96(4), 310–19.
Sampson, G. 2001. Emprical Linguistics. London: Continuum.
2003. The structure of children’s writing: moving from spoken to adult written
norms. In Granger, S. and Petch-Tyson, S. (eds.), Extending the Scope of Corpus-
based Research: New Applications, New Challenges. Amsterdam: Rodopi,
pp. 177–93.
Sanford, A. J. S., Sanford, A. J., Filik, R. and Molle, J. 2005. Depth of lexical–semantic
processing and sentential load. Journal of Memory and Language, 53(3), 378–96.
244 References
Sanford, A. J. and Garrod, S. C. 1981. Understanding Written Language: Explorations

of Comprehension Beyond the Sentence. Chichester: John Wiley & Sons.
Saxton, J. A., Ratcliff, G., Dodge, H., Pandav. R., Baddeley, A. and Ganguli, M. 2001.
Speed and capacity of language processing test: normative data from an older
American community-dwelling sample. Applied Neuropsychology, 8, 193–203.
Schacter, D. L., Alpert, N. M., Savage, C. R., Rauch, S. L. and Albert, M. S. 1996.
Conscious recollection and the human hippocampal formation: evidence from
positron emission tomography. Proceedings of the National Academy of Sciences,
USA, 93, 321–25.
Schank, R. C. and Abelson, R. P. 1977. Scripts, plans, and knowledge. In Johnson-
Laird, P. N. and Wason, P. C. (eds.), Thinking: Readings in Cognitive Science.
Cambridge: Cambridge University Press, pp. 421–32.
Schegloff, E. A. 1968. Sequencing in conversational openings. American Anthropologist,
70,1075–95.
2007. Sequence Organization in Interaction: A Primer in Conversation Analysis: Vol.
I. Cambridge: Cambridge University Press.
Schegloff, E. A., Jefferson, G. and Sachs, H. 1977. The preference for self-correction in
the organization of repair in conversation. Language, 53, 361–82.
Schiffrin, D. 1987. Discourse Markers. Studies in Interactional Sociolinguistics 5.
Searle, J. 1969. Speech Acts: An Essay in the Philosophy of Language. Cambridge:
Cambridge University Press.
1979. Expression and Meaning. Cambridge: Cambridge University Press.
Seiborger, F. T., Ferstl, E. C. and von Cramon, D. Y. 2007. Making sense of nonsense: an
fMRI stude of task induced inference processes during discourse comprehension.
Brain Research, 1166, 77–91.
Seitz, R. J., Nickel, J. and Azari, N. P. 2006. Functional modularity of the medial pre-
frontal cortex: involvement in human empathy. Neuropsychology, 20(6), 743–51.
Shallice, T. 1982. Specific impairments of planning. Philosophical Transactions of the
Royal Society of London (Series B), B298, 199–209.
Shallice, T. and Burgess, P. W. 1996. The domain of supervisory processes and tempo-
ral organisation of behaviour. Philosophical Transactions of the Royal Society of
London (Series B), 351,1405–12.
Shallice, T., Fletcher, P., Frith, C. D., Grasby, P., Frackowiak, R. S. J. and Dolan, R. J.
1994. Brain regions associated with acquisition and retrieval of verbal episodic
memory. Nature, 368(6472), 633–5.
Shallice, T., Venable, N. and Rumiati, R. I. 1988. From Neuropsychology to Mental
Structure. Cambridge: Cambridge University Press.
Sergerie, K., Chochol, C. and Armony, J. L. 2008. The role of the amygdala in emo-
tional processing: a quantitative meta-analysis of functional neuroimaging studies.
Neuroscience and Biobehavioural Reviews, 32(4), 811–30.
Shapiro, K. and Caramazza, A. 2000. Sometimes a noun is just a noun: comments on
Bird, Howard, and Franklin. Brain and Language, 76, 202–12.
Shapiro, L. P., Gordon, B., Hack, N. and Killacky, J. 1993. Verb-argument structure
processing in complex sentences in Broca’s and Wernicke’s aphasia. Brain and
Language, 45(3), 423–47.
Shindler, A. G., Caplan, L. R. and Hier, D. B. 1984. Intrusions and perseverations. Brain
References 245
Silverman, D. H., Small, G. W., Chang, C. Y. et al. 2001. Positron emission tomogra-
phy in evaluation of dementia: regional brain metabolism and long-term outcome.
Journal of the American Medical Association, 286(17), 2120–7.
Sinclair, J. (ed.) 1990. Collins Cobuild English Grammar. London: Harper Collins.
Singh, S. and Bookless, T. 1997. Analyzing spontaneous speech in dysphasic adults.
International Journal of Applied Linguistics, 7, 165–82.
Sirigu, A., Zalla, T., Pillon, B., Grafman, J., Agid, Y. and Dubois, B. 1995. Selective
impairments in managerial knowledge following pre-frontal cortex damage.
Cortex, 31, 301–16.
Sirigu, A., Zalla, T., Pillon, B., Grafman, J., Agid, Y. and Dubois, B. 1996. Encoding
of sequence and boundaries of scripts following prefrontal lesions. Cortex, 32,
297–310.
Skoog, I. and Gustafson, D. 2003. Hypertension, hypertension-clustering factors and
Alzheimer’s disease. Neurological Research, 25, 675–80.
Small, J. A., Kemper, S. and Lyons, K. 2000. Sentence repetition and processing
resources in Alzheimer’s disease. Brain and Language, 75(2), 232–58.
Smith, E. Rhee, J., Dennis, K. and Grossman, M. 2001. Inductive reasoning in
Alzheimer’s disease. Brain and Cognition, 47, 494–503.
Snowdon, D. A., Kemper, S. J., Mortimer, J. A., Greiner, L. H., Wekstein, D. R. and
Markesbery, W. R. 1996. Linguistic ability in early life and cognitive function
and Alzheimer’s disease in late life: findings from the Nun Study. Journal of the
American Medical Association, 275, 528–32.
Sperber, D. and Wilson, D. 1986. Relevance: Communication and Cognition.
Oxford: Blackwell.
1995. Relevance: Communication and Cognition. 2nd edition. Oxford: Blackwell.
St. George, M., Mannes, S. and Hoffman, J. E. 1994. Global semantic expectancy and
language comprehension. Journal of Cognitive Neuroscience, 6, 70–83.
St. George, M., Kutas, M., Martinez, A. and Sereno, M. I. 1999. Semantic integration in
reading: engagement of the right hemisphere during discourse processing. Brain,
122, 1317–25.
Stowe, L. A., Wijers, A. A., Willemsen, A. T. M. et al. 1995. Effects of sentence com-
plexity on rCBF: A PET study. Poster presented at the Meeting of the Cognitive
Neuroscience Society, San Francisco.
Stowe, L. A., Haverkort, M. and Zwarts, F. 2005. Rethinking the neurological basis of
language. Lingua, 115(7), 997–1042.
Stowe, L. A., Broere, C. A. J., Paans, A. M. et al. 1998. Localizing components of a com-
plex task: sentence processing and working memory. Neuroreport, 9, 2995–9.
Stowe, L. A., Broere, C. A., Paans, A. M. et al. 1998. Localizing components of a
complex task: sentence processing and working memory. Neuroreport, 9(13),
2995–9.
Sturt, P., Sanford, A. J., Stewart, A. J. and Dawydiak, E. 2004. Linguistic focus and
good-enough representations: an application of the change-detection paradigm.
Psychonomic Bulletin & Review, 11, 882–8.
Stuss, D. T. and Benson, D. F. 1986. The Frontal Lobes. New York: Raven.
Surian, L., Baron-Cohen, S. and van der Lely, H. 1996. Are children with autism deaf
to Gricean maxims? Cognitive Neuropsychiatry, 1, 55–71.
Szatmari, P. 2003. The causes of autism spectrum disorders (ASD). British Journal of
246 References
Tager-Flusberg, H. 1996. Brief report: Current theory and research on language and
communication in autism. Journal of Autism and Developmental Disorders, 26(2),
169–72.
Tannen, D. (ed.) 1993a. Framing in discourse. New York: Oxford University Press.
(ed) 1993b. Gender and Conversational Interaction. London: Oxford University
Press.
Tannen, D. and C. Wallat. 1993. Interactive frames and knowledge schemas in inter-
action: Examples from a medical examination/interview. In Tannen, D. (ed.),
Framing in Discourse. New York: Oxford University Press, pp. 56–76.
Temple, V., Sabat, S. and Kroger, R. 1999. Intact use of politeness in the discourse of
Alzheimer’s sufferers. Language and Communication, 19(2), 163–80.
Thakkar, K. N., Polli, F. E., Joseph, R. M. et al. 2008. Response monitoring, repeti-
tive behaviour and anterior cingulate abnormalities in autism spectrum disorders.
Brain, 131, 2464–78.
Thompson, C. K. 2003. Unaccusative verb production in agrammatic aphasia: the
argument structure complexity hypothesis. Journal of Neurolinguistics, 16(2,3),
151–67.
Thompson-Schill, S. L. 2003. Neuroimaging studies of semantic memory: Inferring
‘how’ from ‘where’. Neuropsychologia Special Issue: Functional Neuroimaging
of Memory, 41, 280–92.
Thordardottir, E. T. and Ellis Weismer, S. 1998. Mean length of utterance and other
language sample measures in early Icelandic. First Language, 18, 1–32.
Tierney, M. C., Black, S. E., Szalai, J. P. et al. 2001. Recognition memory and verbal flu-
ency differentiate probable Alzheimer disease from subcortical vascular dementia.
Archives of Neurology 58, 1654–9.
Tinaz, S., Schendan, H. E., Schon, K. and Stern C. E. 2006. Evidence for the importance
of basal ganglia output nuclei in semantic event sequencing: an fMRI study. Brain
Research, 1067, 239–49.
Tinaz, S., Schendan, H. E. and Stern, C. E. 2008. Fronto-striatal deficit in Parkinson’s dis-
ease during semantic event sequencing. Neurobiology of Aging, 29(3), 397–407.
Toga, A. W. and Mazziotta, J. C. 2002. Brain Mapping: The Methods. London: Academic
Press.
Tulving, E. 1972. Episodic and semantic memory. In Tulving, E. and Donaldson, W.
(eds.), Organization of memory. New York: Academic Press, pp. 381–403.
1983. Elements of Episodic Memory. New York: Oxford University Press.
2002. Episodic memory: from mind to brain. Annual Review of Psychology, 53,
1–25.
Tune, L., Tiseo, P. J., Ieni, J. et al. 2003. Donepezil HCl (E2020) maintains functional
brain activity in patients with Alzheimer disease. American Journal of Geriatric
Tyler, L. K. and Moss, H. E. 2001. Towards a distributed account of conceptual knowl-
edge. Trends in Cognitive Sciences, 5(6), 244–52.
Ulatowska, H. and Bond-Chapman, S. 1991. Discourse changes in dementia. In
Lubinski, R. (ed.), Dementia and Communication. Philidelphia: Mosby Publishing,
pp. 115–32.
Ulatowska, H., Cannito, M. P., Hayashi, M. M. and Fleming, S. G. 1985. Language abil-
ities in the elderly. In Ulatowska, H. K. (ed.), The Aging Brain: Communication in
the Elderly. Austin, TX: PRO-ED, pp. 125–39.
References 247
van Dijk, T. A. 1977. Text and Context. Explorations in the Semantics and Pragmatics
of Discourse. London: Longman.
1977. Knowledge frames and speech act comprehension. Journal of Pragmatics, 1,
211–32.
1981. Studies in the Pragmatics of Discourse. The Hague/Berlin: Mouton.
2006. Discourse, context and cognition. Discourse Studies, 8(1), 159–77.
van Dijk, T. A. and Kintsch, W. 1983. Strategies of Discourse Comprehension. New
York: Academic Press.
van Dijk, M, and van Geert, P. 2005. Disentangling behavior in early child develop-
ment: interpretability of early child language and the problem of filler syllables
and growing utterance length. Infant Behavior and Development, 28 (2), 99–117.
van Dijk, E. J., Prins, N. D., Vrooman, H. A., Hofman, A., Koudstaal, P. J. and Breteler,
M. M. 2008. Progression of cerebral small vessel disease in relation to risk factors
and cognitive consequences: Rotterdam Scan study. Stroke, 39(10), 2712–9.
van Leeuwen, T. J. 1996. The representation of social actors. In Caldas-Coulthard, C. R.
and Coulthard, M. (eds.), Texts and Practices – Readings in Critical Discourse
Analysis. Routledge: London, pp. 32–70.
van Valin, R. 2001. An Introduction to Syntax. Cambridge: Cambridge University Press.
van Valin R. and LaPolla, R. J. 1997. Syntax: Structure, Meaning and Function.
Vigneau, M., Beaucousin, V., Herve, P. Y. et al. 2006. Meta-analyzing left hemisphere
language areas: Phonology, semantics, and sentence processing. NeuroImage,
30(4), 1414–32.
Volden, J. and Johnston, J. 1999. Cognitive scripts in autistic children and adolescents.
Journal of Autism and Developmental Disorders, 29(3), 203–11.
Wager, T. D., Phan, K. L., Liberzon, I. and Taylor, S. F. 2003. Valence, gender, and later-
alization of functional brain anatomy in emotion: a meta-analysis of findings from
neuroimaging. NeuroImage, 19, 513–31.
Wagner, A. D., Pare-Blagoev, E. J., Clark, J. and Poldrack, R. A., 2001. Recovering
meaning: left prefrontal cortex guides controlled semantic retrieval. Neuron, 31,
329–38.
Warrington, E. K. and McCarthy, R. A. 1987. Categories of knowledge: further frac-
tionations and an attempted integration. Brain, 110,1273–96.
Warrington E. K. and Shallice T. 1984. Category specific semantic impairments. Brain,
107, 829–54.
Wartenburger, I., Heekeren, H. R., Burchert, F., Heinemann, S., De Bleser, R. and
Villringer, A. 2004. Neural correlates of syntactic transformations. Human Brain
Mapping, 22, 72–81.
Watkins, R. V., Kelly, D. J., Harbers, H. M. and Hollis, W. 1995. Measuring children’s
lexical diversity: differentiating typical and impaired language learners. Journal of
Speech and Hearing Research, 38(6),1349–55.
Watt, D. L. E. 1990b. The Phonology and Semology of Intonation in English. PhD dis-
sertation, York University, Toronto.
Webster, J., Franklin, S., and Howard, D. 2001. An investigation of the interaction
between thematic and phrasal structure in nonfluent agrammatic subjects. Brain
Wechsler, D. (1997). Wechsler Adult Intelligence Scale, 3rd edition. San Antonio,
TX: The Psychological Corporation.
248 References
White, P. R. R. 1998. Telling Media Tales: the News Story As Rhetoric. Unpublished
PhD dissertation, University of Sydney, Sydney.
2000. The sixth modal: dialogism and the language of intersubjective Positioning.
First Workshop of the Systemic Functional Research Community on Interpersonal
and Ideational Grammar, Leuven, 16–18 Number 2000.
Zacccai, J., McCracken, C. and Brayne, C. 2005. A systematic review of prevalence and
incidence studies of dementia with Lewy bodies. Age and Ageing, 34(6), 561–6.
Zalla, T., Pradat-Diehl, P. and Sirigu, A. 2003. Perception of action boundaries in
patients with frontal lobe damage. Neuropsychologia, 41, 1619–27.
Zalla, T., Sirigu, A., Pillon, B., Dubois, B., Agid, Y. and Grafman, J. 2000. How patients
with Parkinson’s disease retrieve and manage cognitive event knowledge. Cortex,
36, 163–79.
Zanini, S. 2008. Generalised script sequencing deficits following frontal lobe lesions.
Cortex, 44(2), 140–9.
Zanini, S., Rumiati, R. I. and Shallice, T. 2002. Action sequencing deficit following
frontal lobe lesions. Neurocase, 8, 88–99.
Author index
Abbott 155 Blumstein 95

Abelson 86, 137, 212 n.1 Bond-Chapman 127
Adams 152 Bookheimer 57, 146
Alagriakrishnan 17 Bookless 212 n.1
Allain 138 Bowler 211 n.1
Almkvist 17 Bradac 112
Almor 97 Brazil 37
Alverez 146 Brown G. 54, 142
Alzheimer 15 Brown P. 54
Alzheimer’s Association 14 Brunet 97, 212 n.1
American Psychiatric Association 11, 15 Bschor 145
Anderson 142 Bub 57
Asp 5, 17, 18, 53, 105, 127, 128, 212 n.1 Bucks 97, 112
Asperger 115 Burgess 136, 137
Astell 18 Burns 17
Atkinson 19, 22
Austin 21 Cabeza 146, 211 n.4
Cameron 33
Babba 14 Camille 140
Baddeley 136, 145, 192 Campione 33
Baltaxe 115, 122 Caplan 96
Baron-Cohen 11, 104, 150, 151 Caramazza 95, 213 n.4
Barthes 86, 88, 100 Cardebat 127
Bayles 17, 18, 153 Carlomagno 132
Beauregard 163 Catani 201, 202
Bechara 136, 147 Chafe 30, 31
Beck 162 Chao 213
Bednarek 143 Chertkow 57
Benson 137 Chiat 95
Berg 162 Chipere 112
Berger 56 Chomsky 20
Berlin 57 Chow 148
Berlyne 107 Chwilla 212 n.2
Bickel 17 Clark 22, 32, 33, 99, 116
Bird H. 95 Coates 33
Bird T. 14 Collina 95
Birks 17 Cooper 11
Bishop 150, 151, 152 Coricelli 136, 139, 140
Black, M. 95 Corsun 137
Black, S.E. 17 Cosetino 137, 138, 144
Bleich 15 Coupland 139
Blennow 15, 16, 17 Cullen 17
249
250 Author index
Cummings 16, 18, 163 Garrod 86, 141

Gensler 138, 139
Dalla Barba 107 Gerstein 12, 115
Danes 19, 80 Ghaziuddin 12, 115
D’Argembeau 164 Gilbert 182
De Martino 139, 140 Godbout 137, 138
de Villiers 9, 12, 99, 115, 122, 134, 152, 182 Goel 136
Delgado 163 Goffman 44, 45, 56, 138
Della Sala 136 Goldin 164
Den Ouden 148 Goldman-Eisler 33
Denni 150 Gorelick 17
Desmond 17 Grady 149
Devlin 57 Graffman 137, 138
Dijkstra 17 Graff-Radford 14
Dolcos 149, 164 Greaves 30
Downes 161 Gregory 22, 44, 54, 56, 59, 68, 77, 80, 85, 89,
Doyon 137, 138 156, 212 n.1
Drai 95 Grice 22
Druks 95 Grill-Spector 213 n.4
Dubois 211 n.3 Grodzinsky 95
Duchan 12 Grossman 97
Duff 18 Gumperz 22
Duncan 139 Gussenhoven 37
Duong 101, 127 Gustafson 16, 17
Guye 201
Ehrenreich 162
Ellis 14, 18 Hachinski 211 n.1
El-Manoufy 37 Halliday 21, 30, 35, 37, 42, 45, 72, 73, 74, 75,
Emory 146 76, 79, 80, 81, 84, 85, 97, 116, 118, 156
Erlich 17 Hampel 15
Everingham 162 Happé 11, 112, 122, 150, 151
Hasan 21, 72, 73, 74, 75, 76, 79, 137
Fangmeier 147 Herholz 192
Ferri 15 Heritage 19, 22
Ferstl 57, 88, 143, 146, 147, 148, 149, 163, Herman 97
181 Hess 112
Fiebach 202 Hier 18, 132
Fillmore 59, 138, 143 Hill 11, 15
Fine 12, 99–112 Hirono 198
Firbas 80 Hodges 211 n.4
Firth 85 Holmes 212 n.1
Fletcher 146, 181, 182 Honoré 212 n.1
Fombonne 10 Huddleston 45, 46, 55, 63
Fox 15 Hudson 22, 23, 88, 201
Fox Tree 22, 32, 33, 116 Hymes 22
Frederick 139, 140 Hynd 15
Frijda 136
Friston 148 Ilmberger 57
Frith C. D. 146, 147, 181 Inzitari 17
Frith, U. 11, 104, 111, 122, 146, 147, 150,
151, 181 Jackendoff 20, 22, 56, 59, 136, 202
Jacob 15
Galambos 137 Jefferson 33, 35
Gallagher 181 Jellinger 16
Galvin 14 Joanette 185
Garcia 185 Johnson 138
Author index 251
Jolliffe 150, 151 MacWhinney 34

Jones 18 Mahon 213 n.4
Just 11, 88, 149, 181 Malinowski 85
Mamelak 16
Kahneman 139, 140 Marczinski 57
Kanner 10 Marini 149
Kantarci 15 Martin, A. 136, 214 n.4
Kashani 15 Martin, I. 151
Kay 57 Martin, J.R. 214 n.1, 21, 55, 56, 112, 156
Keenan 98 Mason 25, 88, 149, 181
Kemper 5, 17, 84, 95, 97, 98 Matsuda 15
Kern 107 Maurage 16
Kertesz 9, 14, 57 Mayeux 16
Kintsch 84, 85, 88, 92, 93, 97, 116, Maynard 19
136, 139, 140 Mazziotta 24
Klafki 17 Mendez 18
Klages 105 Mesulam 15, 16, 57, 136, 139, 163, 164,
Knapp 14, 15, 16, 211 n.2 201, 202, 211 n.1, 214 n.1
Knopman 15, 16 Michael 96
Kolk 212 n.2 Miller 151
Kontiola 17, 18 Minsky 86, 138
Kopelman 107 Monti 147
Koshino 11 Moorhouse 17
Krawczyk 139 Moreno 163
Krishnan 17 Moss 213 n.4
Kuperberg 147, 149
Nankano 95
Labov 86, 87, 100 Nespoulous 17, 55, 57, 127, 128
Lafosse 18 Neuropathological Group 15
Lakoff G. 22, 55, 57 Nicholas 17, 132, 185
Lakoff R. 36 Nicholson 11
Lamb 202 Norbre 146
Lane 140 Norbury 150, 151
Langa 17 Norman 136, 137
LaPolla 59 Noveck 147
Lee 95 Nussbaum 14
Leonard 112 Nyberg 146, 211 n.4
Levinson 19, 54
Li 17 O’Brien 16, 17
Looi 17, 18 Ochsner 164
Loukusa 152, 154 Oh 78, 112
Luan Phan 147 Orange 9
Luchsinger 16 Osler 34
Luckman 56 Owen A. J. 112
Luria 137 Owen A. M. 139
Lyons 17 Ozonoff 151
McAlonan 182 Pantoni 16–17

McCarthy 57, 213 n.4 Patry 17
McCawley 67 Paxton 136
McDonald 151 Perry 211 n.4
McKeith 14 Petrides 146, 181
McKenna 15, 78, 112 Phan 163
Mackenzie 5 Pike E. 64
McPherson 16, 18 Pike K. 64
McShane 17 Poore 18
252 Author index
Powell 18 Sturt 97
Prizant 12 Stuss 137
Propp 86, 100 Sumner 98
Public Health Agency of Canada 14 Surian 152
Pullum 45, 46, 55, 63 Szatmari 11, 122
Quirk 47 Tager-Flusberg 12
Tannen 86, 138, 139, 142
Razvi 16 Temple 155
Ready 18 Thakkar 182, 194
Reverberi 147 Thiebaut 201
Rocca 15, 16 Thompson 95, 97
Rochester 112, 214 n.1 Thompson-Schill 136, 213 n.4, 214 n.4
Rockwood 17, 18, 90, 184 Thordardottir 116
Roman 16, 17 Tierney 18
Rosch 57 Tinaz 136, 138
Royall 23, 136, 145, 146, 211 n.1 Toga 24
Ruchoux 16 Tulving 149, 211 n.4
Rugg 149 Tune 17
Rushworth 136, 138, 139, 140, 144, 146 Tyler 213 n.4
Russel 11
Rusted 138 Ulatowska 4, 127
Sachdev 17, 18 van Dijk, E. 212 n.2

Sacks 19, 32, 33 van Dijk, M. 116
Saldana 150 van Dijk, T. 21, 22, 56, 84, 85, 86, 88, 92, 93,
Sampson 116, 118 116, 136, 138, 139, 140, 212 n.2
Sandford, A. J. 86, 141 van Geert 116
Sandford, A.J.S. 97 van Leeuwen 54
Saxton 97 van Valin 59, 63, 64
Schacter 149 Veronis 33
Schank 86, 137, 212 n.1 Vigneau 181, 202
Schegloff 22, 33 Volden 138
Schiffrin 9 von Cramon 143, 146, 147, 148, 181
Searle 30
Seiborger 147 Wager 163
Seitz 164 Wagner 149
Sergerie 163 Waletzky 86, 87, 100
Shallice 57, 136, 137, 138, 145, 149, 213 n.4 Wallet 86, 139, 142
Shapiro, K. 95 Warrington 57, 136, 213 n. 4
Shapiro, L.P. 95, 97 Wartenburger 202
Shelton 213 n.4 Wasow 99
Shepherd 138 Watkins 112
Shindler 18 Watt 77, 80
Silverman 15 Webster 95
Simoni 16 Weschler 105
Sinclair 48, 55 White 21, 55, 56, 156, 159
Singh 212 n.1 Wilson 22, 154
Sirugu 137, 138 Woodruff 14
Skoog 17
Small 98 Yule 54, 142
Smith 185
Snowdon 84 Zaccai 14
Sperber 22, 154 Zalla 137, 138
St. George 141, 142 Zanini 137, 138
Stowe 57, 88, 96, 97, 149, 181, 202
Subject index
affect treatment 17, 209

affective processing style 139 treatment effects in 5, 126–32
appraisal framework 156 treatment monitoring 127, 192
cognitive models 162 ambient 66
definition 156 argument roles (see also circumstantial roles)
iconic signalling of affect 162 definitions 59–67
lesions studies 140 summary 68
model analyses argument structure
AD 159–61 aphasia 94–5
autism 156–9 information processing load 95
modulate discourse processes 23, 136 problems with 94–5
neural networks 146, 147, 163–5, 201 Asperger’s syndrome
prosody 162, 163, 167 casual conversation 12
semantic and lexical resources 21, 161–2 contextual relevance 12
Agent 59–60 discourse examples 169–82, 192–4
Alzheimer’s disease language development 12
cause of 16 pedantic speech 12, 115
dementia phase 5 topic management 12, 169–82, 193
diagnostic criteria 15 attitude (see also affect)
discourse characteristics 5, 17–18, 126–32, ASDs 6, 167
199 lexical selection 55, 196
confabulation 105–7, 200 linguistic resources 55–6
discourse monitoring 191, 200 Attribuand 66
episodic (self-referential) checking tags autism
127–8 contextual relevance 13
episodic recall 196–8, 200 discourse examples 5–6, 7–9, 12–13,
fluency 191, 200 156–9, 165–7
hesitations 191 expression of attitude, examples 6, 167
coherence 17, 183, 185–91, 194–8, 200 expression of evaluation, examples 8, 159
ideational perseveration 17, 200 fMRI study of inferencing 25
incomplete utterances 183, 191 morphosyntactic difficulties, examples 13
lexicalizalization 17, 183, 185–91, 198, 200 speech characteristics 11
discourse examples 4–5, 105–7, 106, topic management examples 13
126–32, 182–92, 194–8 autism spectrum disorders (ASDs)
modalization 17, 126–7, 183, 191, 200 causes 11
morphosyntax 13 communication difficulties 11–12
reference 183, 185–91 defined 10–11
repetition 17, 191, 196–7 diagnostic criteria 10–11
topic development 17, 183, 196–8 discourse characteristics 6, 200
turn-taking 199 dysfluency 8, 9, 98–100, 181
neurophysiology 15, 192 evaluation 104
prevalence 14, 15 humour 151
risk factors for 16 inferencing in discourse 149–54, 181
253
254 Subject index
autism spectrum disorders (ASDs) (cont.) cognitive models (see also top-down models,

pedantic speech 8, 115–26, 181 affect)
see also pedantic speech distributed models 136
perseveration 194 for words 135
quantity of information 152–4, 172 in functional linguistics 21
special interests 194 modality neutral conceptual representations
transitions 169–82, 193–4 135–6
use of contextual information 152 transmodal association areas 136
neuroimaging 11 cohesion
treatment 11 and discourse varieties 78–9
example analysis 79–80
brain–behaviour relationships types of relation
attitudes to 1–2 conjunction 75
study of 1 definition 73
understanding of 9, 114, 207 ellipsis 75
Brunet’s Index 97 lexical cohesion 75
collocation 76
Cause 60 repetition 75–6
checks synonymy 76
definition 52 reference 74
examples 5, 191 AD 183, 185–91
Circumstantial roles substitution 74–5
definitions 67–70 cohesive irregularities (evaluation) 78
summary 70 coherence
Classified 66 clinical relevance 88
clause definition 88
definition 29–30 examples
examples 30, 31 AD 183, 185–91, 198
clinical discourse analysis Asperger’s syndrome 192–4
clinical research areas 4, 203, 204 neural substrates 88–9, 145–9, 147, 181, 192
comprehensive discourse descriptions commands
5, 6, 7, 109, 207, 209 morphosyntactic features 50–1
defined 7, 22 sub-classification 50
detailed examples summary 51
AD 182–92, 194–8 use and role relationships 51
Asperger’s syndrome 169–82, 192–4 conceptual structure 56–8 see also argument
diagnosis 2, 5, 7, 114, 207, 209 roles, cognitive models
discourse patterns 2, 5, 6, 9, 207, 209 and argument roles 58–9
future prospects 9, 208 decomposition and taxonomic structure
goals 6–7, 9, 92, 207 57–8
informing clinical judgements 7, 114, 198, example analysis 70–1
207 confabulation
interdisciplinary expertise 9, 25, 208, 210 discourse definition 107
intra- and intercultural relevance 204–6 example on delayed recall (AD) 105–7
limitations 208 modelling 107, 108
neurocognitive (dys)function 1, 2, 5, 7, neuropsychological definition 107
201–4; AD 192 contexts of culture and situation see also top-
Asperger’s syndrome 181–2, 185–91, 192 down cognitive models and phase
non-technical examples 4–5, 7–9 cultural relevance 204–6
theoretical sources 2, 10, 19–23, 207 discourse analysis 85–9, 168, 198–200
therapies 7, 114, 207, 209 example analyses
treatment response 7, 114, 126–32, 192, home visit (AD) 183, 194
207, 209 psychiatric research interview (ASD)
coding 170, 192
non-linguistic behaviour in transcripts 35 ethnographic definition 85
relevance in clinical discourse analysis 29 functional and dialectal variation 88, 204
Subject index 255
inheritance relations 23 Asperger’s syndrome 9.31; (AD)

instantial situation 23 192, 200
linguistic systems 20 inferencing 145–55
neurocognitive definition 22, 85–6, 168 top-down models 136–145
semantic and episodic memory 22, 85, 204 Existent 66
conversation analysis 6, 10, 21, 199 Experiencer 61
definition 20, 32–5
technical terms 35 false starts
conversation drift 6, 152 definition 34
conversation skills 170–1, 192, 199 examples (normative) 5, 9
conversation tasks AD 183, 191
morphosyntax and information processing Asperger’s syndrome 170
load (in ASD) 98–100 information processing load 96
study design 93 feature analysis
corpora examples
computationally mediated analysis 208 AD 185–91
definition 109 Asperger’s 172
demographically and functionally relevance 170, 198
representative 25–7, 113 fluency features
manual analysis 208 clinical relevance 19–20, 96
sampling issues 109–13 definition 19
specialized for clinical discourse analysis 2, examples
5–6, 27–8, 113 AD 5, 130, 191
Autism 9
dementia see also Alzheimer’s disease, Focus and Prominence 79–80
neurocognitive disorders, vascular clinical relevance 83–4
cognitive impairments definitions 80–1
definition 13–14 Given/New 80
diagnostic criteria 14 importance 81
mixed dementia 15, 17–18 lexical density 83
prevalence and incidence 14 morphosyntactic resources 81–3, 84
discourse (see also clinical discourse analysis) syntactic complexity 84–78
as language behaviour 1, 7 Theme/Rheme 80
complexity 1, 4 functional linguistics 10, 20–1
discourse monitoring 127, 191, 192
discourse norms 24, 25–8, 198–200 generative linguistics 10, 20, 22
discourse patterns 1, 2, 10, 199–200 genres 21
discourse processes 204 generic structures 21
discourse tasks 3, 93 gist recall
example
echolalia 12, 191 AD 106
ethnographic linguistics 10, 21, 22 VaD 105–6
evaluation modelling 108
examples grammatical resources 46–8, 204
AD 196 functionally organized 22
ASDs 8 grammar of ideation 56–71, 72
hedges 55 see also argument roles, circumstantial
in appraisal theory 55 roles, conceptual structure, feature
linguistic resources for 21, 54–5 analysis
modalizing discourse 55 grammar of interaction 44–56, 72
neural systems 146 see also affect, appraisal theory, feature
exclamation analysis
morphosyntactic features 46–8 address terms 53–4, 196
executive functions alignment (AD) 197
executive control processes 23 attitude and evaluation 54–6
information processing 23, 130, 201 checks and intonation 52
256 Subject index
grammar of interaction (cont.) contexts 23

politeness markers 54 instantial situations 23
role relationship 44–5, 199 Instrument 60
AD home visit 183–4, 194 interactional sociolinguistics 10, 21, 22
Asperger’s syndrome 193, 194 interruption see overlap
psychiatric research interview 170–1, intonation 35–43
192, 193 see also prosodic phonology and the
speech functions 45–51 grammar of interaction
tags (see tags) clause 30
grammar of message organization 72–84 notation 36–7
see also feature analysis primary tones 37–42
coherence and intertextuality 73 speech function 42–3
see also coherence speaker stance 36
Focus and Prominence 72 see Focus and tags (see tags)
Prominence utterance 31
presupposition and cohesion 72
see also cohesion language–brain relationships 1, 201–4
Gricean co-operative principle 154 large scale functional neural networks
Gricean maxims 201–2
in ASDs and AD 155, 200 affect 163–5
discourse 181–2
hesitation inferencing 144–9
definition 34 language 201–4
examples 5, 9, 34, 170, 191 top-down models 136–45
latching 35
idea density 97 lexical access (AD) 191, 198
Identified 65 lexical density
idioms 5, 71 measures 97
incomplete utterances examples
definition 34 AD 130, 185–91
examples 95 ASD 116, 118
AD 183, 191 lexical richness measures 97
inferencing (see also coherence, pragmatics, literal interpretation 71
relevance theory, top-down models) Logical circumstantial roles 69–70
and right hemisphere activations 149
causal inferences 147–8 Manner 69
deductive conclusions 147 memory see also contexts, inferencing,
discourse comprehension 145–9 large-scale functional networks,
in discourse 144–5 top-down models
ASDs 149–54 episodic memory definition 211 n.4
intentions 148 episodic models 23
neural systems and large scale networks episodic recall 198, 200
144–9, 181–2, 201 neural systems 201
predictions 148 self-referential tags 127–8
theory of mind 146, 147, 148 semantic memory definition 211 n.4
Asperger’s syndrome 179 semantic models 23, 198
information processing load working memory 201
definition 95 deficits in AD 95–6, 192
fluency features 95–6 Message 65
AD 191 metaphor
Asperger’s syndrome 181 examples in AD 5
example (ASD) 98–100 feature analysis 70–1
idea density 97 modalizing discourse
lexical density and richness 97 in AD 126–7, 183, 200
morphosyntactic complexity 97 episodic memory deficits 130–2, 200
working memory deficits in AD 95–6, 192 lexical selection and indefinite reference
inheritance relations in 131–2
Subject index 257
AD 183, 185–91 normative 33

modalization battery 128 pedantic speech see also autism, autism
model analysis (AD) 128–32 spectrum disorders, Asperger
monitoring syndrome
change over time 2, 114, 126–32, 207 examples of 8–9, 115–26, 170
treatment effects (in AD) 126–32, 184 features of 9, 115, 116–25
morphosyntactic complexity formulaic elements 5, 9
example (ASD) 116–18 lexical density 116, 118
information processing load 97 lexical repetition 118, 170
measures 96–7 message organization 122–5
register specific selection and collocation
narrative 118–22, 126, 170
delayed recall task (AD/VaD) 105–9 relevance construction 122
episodic memory 93 repetition 9, 170
evaluation of narrative performance 101 syntactic complexity 116–18
schema 100–1 technical specificity 9
study design 93, 100–2 pedantic battery 122–5
spontaneous narrative (ASD) 102–4 Percept
task types 101–2 definition 64–5
theory of mind 93 examples of 5
neurodegenerative disorders 13–19 phasal analysis
Creutzfeldt-Jakob disease 13, 14 clinical relevance 168, 181–2, 198–200, 199
fronto-temporal dementias 13, 14 definition 89, 169–82
Lewy Body disease 13, 14 examples
Parkinson’s disease 13, 14 AD 183–5, 196
(see also Alzheimer’s disease, dementia, Asperger’s 169–82, 192–3
vascular cognitive impairments) feature analysis examples
neuroimaging AD 185–91
activation patterns 1, 11, 25, 201 Asperger syndrome 169–82
clinical discourse analysis 1, 3, 10–24, 25, multifunctional 92, 172, 198
201–4, 209 pattern identification 92, 168, 172,
developments in 1, 202 198–200
diffusion tensor imaging 201, 202 phonology (see also intonation)
fMRI example – inferencing in phonological hierarchy 35
discourse 25 primary tones 37–42
frame use 139–40 prosodic phonology 35–43
inferencing 145–9 segmental phonology 35
types 24 tonicity and information structure 36
neurological function and structure 201–4 tonic syllable 36
see also inferencing, Alzheimer’s disease, Place 67
ASD’s, neuroimaging, large-scale planning (see also executive function)
functional neural networks, examples of difficulties, 5, 191, 192
transmodal association areas Possessed 66
neurology pragmatics 10, 21–2
clinical discourse analysis 10 goals 21
neuropsychology speech acts 21
clinical discourse analysis 3, 10 predication
incomplete examples 5
overlap AD 183, 191
definition 34 tangential 5
examples 34 AD 183
psychiatry
Patient 64 clinical discourse analysis 10–24
pause
definition 33 questions
examples of morphosyntactic features 49–50
ASD 9 presupposition 49, 50
258 Subject index
Range 65 affect 136

rating scales as mental models of contexts of culture and
definition 133 situation 87–8
example (ASD) 134 coherence 88
use 133 comparison and overlap of models 142
relevance definition 136
contextual 12, 21, 88, 152 executive function 136, 201
relevance theory 21, 22, 154, 199, 200 examples 86–7
repairs (see also executive function) AD 143–4
definition 33 autism 165–7
examples 5 relevance 88
ASD 9 frames
normative 33–4 anaphora 139
repetition contextualization processes 138
Alzheimer’s disease 17, 191, 200 decision making 139
ASD’s 6, 8, 12, 118, 122–5, 170 definition 86, 138
fluency feature 5 examples
perseveration 194 Asperger’s syndrome 171, 193
Resultant 64 autism 166
inferencing in discourse 138–9, 145–9
selection 192, 201–4 neural characterizations 139
semantic decomposition 57–8 neural activation patterns 139–40
Source 62 representations of conventional knowledge
subject–verb agreement 5, 46 138, 185
subordinate clause structures risk-taking 139–140
AD 5 roles and reframing situations 138
ASDs 8 scenarios
speech functions definition 86, 140–1
definitions 45–51 examples
summary 53 autism 166
statement home visit 184
examples 46 psychiatric research interview 171, 192
morphosyntactic features 45–6 processing speed evidence 141
tags see tags imaging studies 141–2
unmarked tone 46–8 schemas
Stimulus 61–2 contextualization cues
study design 93 Asperger syndrome 169–82
context of situation and register 109 autism 166–7
examples 109 definition 86, 142
measures for morphosyntactic complexity distributed multi-modal representations
96–7 142, 143
morphosyntactic features 94 examples
role of comprehensive descriptions 109 AD 185, 192, 196, 197
see also conversation tasks, narrative tasks, Asperger’s syndrome 171–2, 192–4
corpora in functional linguistics 21
scripts
tags activities of daily living 138, 144–5
Alzheimer’s disease 127 Alzheimer’s disease and script use
information monitoring tags (AD) 127 143–5
intonation, and presupposition 47–8 autism 138, 166
morphosyntactic features 47 definition 86, 137
self-referential tags (AD) 127–8 executive function 136–49
technological developments 1–2, 208, 209 examples
Theme (argument role) 63 Asperger’s syndrome 172
Time 67 normal 184
top-down cognitive models functional linguistics 21
Subject index 259
functional network for 138 transmodal association areas 136, 201–2

medial temporal lobe function 137 language 201–4
prefrontal function 137, 144–9 turn-taking
topic clinical relevance 20, 199
definition 19, 33 definition 19, 33
examples 33–4 examples
AD 183, 196–8 AD 130
Asperger’s 171, 193 ASDs 6
topic development autism 8
definition 33 normative 33
examples Asperger’s syndrome 169, 193
AD 183, 196–8 Type-token ratio 97
ASD 6, 8, 193
normative 33–4, 193 utterance
topic initiation clinical relevance 31, 32
definition 33 definition 30–2
examples examples 31–2
ASD 6, 8
normative 33–4, 193 vascular cognitive impairment (VCI)
topic management see also Alzheimer’s diagnostic criteria 16
disease, Asperger’s syndrome, example delayed recall task 105–6
autism language and discourse deficits 18–19
clinical relevance 20, 199 neurophysiology 16
examples prevalence 14–15, 16
AD 183, 185–91 risk factors 17
ASDs 6, 8, 13, 169–82, 193, 194 symptoms 16–17
topic shift treatment 17
definition 33 verbatim recall
examples 6, 33–4 examples
AD 183, 185 AD, VaD 105–9
Asperger’s syndrome 169–82, 193, 194 control 108
normative 184, 193 modelling 108

When Language Breaks Down PDF

Uploaded by

Copyright:

Available Formats

When Language Breaks Down PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

When Language Breaks Down PDF

Uploaded by

Copyright:

Available Formats

This page intentionally left blank

When Language Breaks Down

Elissa D. Asp is Associate Professor of English Linguistics in the English

Jessica De Villiers is Associate Professor in the Department of English at

Elissa D. Asp and Jessica de Villiers

Cambridge University Press

This publication is in copyright. Subject to statutory exception and to the

ISBN-13 978-0-511-67769-4 eBook (NetLibrary)

ISBN-13 978-0-521-71824-0 Paperback

Cambridge University Press has no responsibility for the persistence or accuracy

List of figures page vi

2.1 Context, language and discourse relations to

3.1 Terms for conversation analysis page 33

This book is predicated on the relatively uncontentious notions that discourse

1.1 What is clinical discourse analysis?

Text 1.1 My Mom

Text 1.2 Lions

(6) CHI: and then you hate lions.

1.2 What use is clinical discourse analysis?

1.3 What use is this book?

1.4 Sample analysis

Text 1.3 The magic of the universe

2.2 Autism spectrum disorders

3) stereotyped behaviours and restricted interests (e.g. a desire for sameness or

In trying to understand the underlying causes of the three core deficits of

characteristics of autism include flat or unusual intonation (Fine 1991; Fine

Text 2.1 Free Willy

(29) CHI: yeah.

2.3 Neurodegenerative disorders associated with aging

single disease. While there is some debate as to the combination of features

years (Ferri et al. 2005). AD accounts for approximately 60% of dementia

Although there are a number of biomarkers and established neuropathological

2.3.1 Vascular cognitive impairment

pattern is for prefrontal systems to be impaired, which may be reflected vari-

2.4 Theoretical sources

2.4.1 Conversation analysis

2.4.2 Generative linguistics

2.4.3 Functional linguistics

and interactional sociolinguistics as well as directly in some models of language

Contextual Discourse Neurocognitive reference domains Description domains

Context of Language • Language(s) and dialect(s) knowledge • Models/grammars of

Context of Functional • Semantic/episodic memory • Ethnographies of speaking

Figure 2.1 Context, language and discourse relations to neurocognitive

2.5 Other sources

discourse behaviours and neurological disorders. Our primary goal here is to

Current studies investigating language and discourse abilities that involve an

2.6 Discourse norms

Theory of Mind is supported by a localized and functionally dedicated neural network

In autism there are deficits in Theory of Mind

This may be reflected in differences in inferencing in discourse

Does the question merit the resources and participation demands?

Participant selection (diagnosis by psychiatrist)

Figure 2.2 Interdisciplinary relations in clinical discourse analysis

of demographically representative and functionally salient samples of speech

Combining cross-sectional and longitudinal sampling with matched controls

as subject and complement respectively. Yesterday is an optional element

information about a speaker’s message. Thus, a minimal response umhm

Text 3.1 Scalding

3.2 Conversation analysis

List of figures page vi

2.1 Context, language and discourse relations to

3.1 Terms for conversation analysis page 33