SplitPDFFile 124 To 195

5
The epistemic primacy of belief in the Trinity
108 An investigation of its practices has suggested that the church

can endure as a coherent community only if it continually tests other
beliefs for truth or falsity by seeing whether they are consistent with its
trinitarian identification of God. Should this distinctive epistemic
procedure be incoherent or otherwise unreasonable, then what follows
is presumably not the continuing embrace of irrational practices, but
their abandon- ment, and so a basic change in communal identity. We
therefore need to determine whether the church has the epistemic right
to regard its own most central beliefs as the primary criteria of truth,
and if so, what confers this right. This requires a closer look at the
contents of the church’s trinitarian nexus of belief.
Jesus’ universal primacy
Jesus undertakes his journey from Bethlehem to Golgotha, so the

church believes, for the life of the world (cf. Jn. 6:51). The gift of life
which Jesus undertakes to give the world is not, it seems, an event or
state of affairs which comes to pass apart from or in addition to his
acceptance and enactment of the mission from the Father which leads
to the cross. Rather the world’s deliverance from death, its redemption
and reconciliation to God, coincide with the particular and
unrepeatable sequence of actions and events by which the church
identifies Jesus. Jesus’ death does not simply symbolize or promise the
world’s deliverance, but actually puts death to death; Jesus’
resurrection, ascension, and his gift of the Spirit do not simply
symbolize or promise new life from the dead, they bring it about and
impart it.1 This means that what happens on the way
1. The language here is in part that of the Orthodox Divine Liturgy, but the thought is
not a uniquely Eastern one. Thus Augustine: “The immortal one took on mortality in
order that
The epistemic primacy of belief in the
Trinity
10
from Bethlehem to Golgotha and the Emmaus road has universal
scope. If Jesus’ action and passion are genuinely for the life of the
world, then their power and significance extends to all creation. And
the difference Jesus makes to the rest of reality is not superficial or
transitory, but utterly basic; it is the difference between death and life,
between non-being and being, for every particular thing.
It thus seems that a garden-variety belief in redemption through
Jesus entails belief in Jesus’ unrestricted primacy with respect to all
created reality. If in virtue of his life, death, and resurrection Jesus is
“for the life of the world,” he must have what the New Testament calls
“primacy in everything” (Col. 1:18). That creatures are and what they are
must depend on Jesus himself – and so on what comes to pass between
Bethlehem and the Emmaus road – if he is to be their redeemer. Only
one upon whom creatures depend for their existence, qualities, and
relations can give them forgiveness and new life – can make them be
what they were not and not be what they were. And all creatures must
be dependent upon Jesus if any creature is to be redeemed by him.
Any creature outside the scope of this dependence might be capable
not only of resisting Jesus’ bestowal of life in his own case, but of
counteracting it in others. Jesus can be the redeemer of the world, it
seems, only if he is the one in whom “all things hold together” (Col.
1:17), and through whom “all things came into being” (Jn. 1:3; cf. I Cor.
8:6).2
All this may seem more than a little paradoxical, since the one on
whom this argument supposes that all creation depends is himself a
creature, who as such came to be at a particular moment, before
which he was not, and who lived in a particular and therefore limited
stretch of space and time. This appears to generate wildly incoherent
claims. If all creation depends on Jesus in the manner I have
described, then what came to be before Jesus was born depended for
its existence on Jesus. But Jesus did not yet exist. So what existed
before him depended on what did not exist. But surely nothing can
depend for its existence on what does not exist, so what existed before
Jesus cannot have depended on him. The New Testament claim that in
the human being Jesus “all things hold together”
he might die for us, and by his death put to death our death.” Sermo 23a, 3, CCL, vol. xlI,
p. 322.
2. All this, it should be stressed, is said in the New Testament of the particular human
being Jesus. The one through whom “all things came into being” in Jn. 1:3 is the Logos,
but precisely the Logos who became flesh (Jn. 1:14); the one in whom “all things hold
together” in Col. 1:17 is the same as “the firstborn from the dead” (1:18), who “has made
peace through the blood of his cross” (1:20); the one Lord “through whom are all things”
Trinity and
1
in I Cor. 8:6 is the same as the human being who died for those who now eat in the
temples of Corinthian idols (see 8:11). See also note 8 below.
Trinity
11
and through him “all things came into being” seems thoroughly self-
contradictory.
Col. 1, which has helped to organize these last reflections, handles
this problem by recourse to the Christian community’s trinitarian
identification of God, and in particular by characterizing Jesus as “the
icon (ϵt’nω´ v) of the unseen God” (1:15). In this context the Father is
clearly the “unseen God” of whom Jesus is the “icon” or image. To
characterize Jesus as the Father’s image suggests that his human
visibility depicts the Father in the world, in virtue of a resemblance or
likeness to the one whose image he is. While unseen in himself, the
Father enables the world to see him in another, or to see him by seeing
this other. As Paul puts the point, the face of Jesus Christ is the very
image of God (cf. II Cor. 4:6,4). And Jesus’ human face is not a partial
or transitory, and therefore perfectible or replaceable, image of the
Father; rather, in the expressive phrase of Heb., Jesus “bears the exact
imprint of God’s very being” (1:3). Without preju- dice to his full
humanity, so these texts propose, this human being com- pletely
shares the attributes of the God whom he calls “the Father”; if he is the
Father’s perfect image, then “in him all the fullness of God was pleased
to dwell” (Col. 1:19; cf. 2:9 for Θϵo´ r roç).3
The New Testament’s characterization of Jesus as the image of the
unseen God conceptualizes, we could say, the incarnational logical
structure of the narratives which identify him. These narratives
attribute the actions and passions of this human being to God (such as
dying on the cross), and the actions of God to this human being (such
as forgiving sins). Statements which conform to this logical pattern
imply that the one who dies on the cross and the one who forgives the
paralytic’s sins – and in whom all things hold together – are the same,
in the strict sense: they must, if such statements are true, be
numerically identical. By having this implication, these statements give
rise to the traditional Christian doctrine of incarnation, insofar as this
doctrine is concerned (perhaps chiefly concerned) to assert explicitly
that “our Lord Jesus Christ” is “one and the same (ϵ‘´ vα nαt` rov αv’ro`
v) . . . the same one perfect in divinity and the same one perfect in
humanity.”4
3. For an analysis of the extensive modern exegetical debate on this passage from Col.,
see Pierre Benoit, “L’hymne christologique de Col. 1, 15–20,” Exégèse et théologie, vol. IV
(Paris: Cerf, 1982), pp. 159–203. Alois Grillmeier charts the role of the passage in the
christological debates of the ancient church in Jesus Christus im Glauben der Kirche, vol. I,
2nd edn (Freiburg: Herder, 1982), pp. 102–21. For a theological reflection see Hans Urs von
Balthasar, Theodramatik II/2 (Einsiedeln: Johannes Verlag, 1978), pp. 229–38 (ET, pp. 250–
9).
Trinity and
1
4. DS 301 (the definition of Chalcedon). Chalcedon’s repeated insistence on the
logical point that the one who is perfect (that is, complete) ϵ’v α’ vӨpωπo´ r rt is
numerically the
Trinity
11
In the prologue to John’s Gospel the church has generally found the
chief paradigm for this teaching. Jesus is there described as the Word
of God become our flesh. That is: the Word “expresses the total reality
of the Father” (to recall a phrase of Thomas Aquinas), and the human
being Jesus is himself this Word, incarnate; therefore Jesus himself
expresses the total reality of – is the perfect image of – the Father.5 Col.
1, we can see, reverses the subject–predicate relation of the Johannine
paradigm. Whereas John takes the Logos as subject and ascribes
“flesh” to him, Col. takes the crucified human being Jesus as subject
and ascribes “the fullness of God” to him. These two paradigms
jointly display the logical pattern which governs the developed
doctrine of incarnation, and is clearly at work in the claim that in the
creature Jesus, all things hold together: whatever is true of the human
being Jesus of Nazareth is true of God (John) and whatever is true of
God is true of the human being Jesus of Nazareth (Col.).6
For present purposes, the crucial outcome of these reflections on
Jesus as the icon of the Father is that he shares fully in the Father’s
creative power. Only if this human being fully possesses the Father’s
divine capacity to give being in every respect is it true to say that in
him, “all things hold together.” But because the capacity to create is
shared by the Father and Jesus (and also of course by the Spirit, about
whom more in a moment), the actual work of creation, of making all
things and making everything hold together, will also be shared by
them. The work, like the capacity, will however be shared in a certain
order. It will originate with the Father, who is the source, himself
unoriginate, of the Son and the
same – ro` v αv’ ro` v – as the one who is complete ϵ’ v Өϵo´ r rt , gives the rule, as it were,
for its subsequent use of the notions of person, hypostasis, and nature (see DS 302). A
detailed discussion of Chalcedon’s doctrine is beyond the scope of this book, but we may
observe that the definition’s main concern is not to make an arcane metaphysical point,
but to make explicit a pattern for uttering true sentences about Jesus Christ. It thereby
addresses a matter of basic Christian concern: worship of Jesus presumes that sentences
uttered according to this pattern are true, otherwise such worship would be idolatry.
5. See chapter 1, note 2. This suggests that in describing Jesus as the perfect image of the
Father we have located for at least one of the divine persons a feature of the sort for which
the liturgy suggests we look: a characteristic not only unique to him, but non-contingent,
and therefore constitutive (at least in part) of his identity. On this see chapter 9, pp. 269–
71.
6. The medievals developed in considerable detail the use of logical devices to articulate
Chalcedon’s doctrine (see Marshall, Christology in Conflict, chapter 5), but this procedure
stems from the sometimes self-conscious preoccupation of ancient Christian theology
with what can and cannot be said christologically. See, for example, Cyril of
Alexandria’s 4th anathema against Nestorius: “Whoever allocates the terms contained in
the gospels and apostolic writings and applied to Christ . . . to two persons or hypostases
and attaches some to the man considered separately from the Word of God, some as
divine to the Word of God the Father alone, shall be anathema.” Lionel R. Wickham, ed.
and trans., Cyril of Alexandria: Select Letters (Oxford: Clarendon Press, 1983), p. 30; for
Trinity and
1
applications of this rule see Cyril’s Quod unus sit Christus, Patrologiae Cursus Completus, series
graeca (=PG), ed. J. P. Migne (Paris: 1857–66), vol. lxxv, 1289B–1293a; 1327B–1329D.
Trinity
11
Spirit, and it will be received and accepted by the Son and by the
Spirit. The Father and the Son will each therefore have their own
distinct role in the one work of creation, as of redemption, though the
roles will be inseparable.7
The work of the Trinity in making all things hold together in the
one who is the Father’s image has two aspects with particular
epistemic significance. Col. 1 expresses these by saying “all things have
been created (a) through him (6t’ αv’rov˜) and (b) for him (ϵt’ç αv’ro`
v)” (1:16).
Saying (a) that the Father creates through Jesus suggests that as the
Father eternally wills to create a temporal world, his own Word in the
flesh accepts this intention and shares in its enactment – or, if it helps
put down the specter of incoherence which led us to go incarnational
as well as trinitarian in thinking about creation, his own Word who is
to become flesh (the traditional Verbum incarnandum).8 He is himself,
with the Father, the agent of creation, upon whom all creation entirely
depends. More than that: to say that the Father creates through Jesus
suggests that when the Father wills to create the world, he sees and
knows it – indeed sees and knows all possible worlds – in and through
his Word in the flesh. The Word in the flesh is, to use the traditional
term, the “exemplar” for the Father of all things, real and possible.
That the Father sees all things in the enfleshed Word does not mean
that Jesus himself is all things, but rather that all things have their
reality and particular character in virtue of their relation and ordering
to him.
The complementary phrase from Col. 1 brings this out: (b) “all
things have been created for him.” As he wills to create the world, the
Father
7. On this see the discussion in chapter 9, pp. 251–8.

8. It seems, however, to make no substantive difference whether one speaks of the Word
incarnate (Verbum incarnatum) or the Word “to be incarnate” (Verbum incarnandum), since
from God’s point of view – which is the one that counts in the present case – the Word is
always incarnate. Like divine acts generally, the Father’s sending of the Son into human
flesh, the Son’s acceptance of this mission, and the Spirit’s creation of the humanity of
Mary’s first child by uniting that humanity to the eternal Son – everything which makes
up the incarnation from God’s side – is always actual in and for God. The created terms of
this act of incarnation – the humanity of Jesus and its union with the Word – are temporal;
they come to be at a particular time. One therefore need not resort to talk of the Verbum
incarnandum for fear of paradoxical or incoherent consequences which might follow
from speaking of creation through the Verbum incarnatum; for the creating God the two
come to the same thing. On this see, e.g., Thomas Aquinas on the missions of the divine
persons: “Divine action (operatio) can be considered in two ways: either from the side of
the agent, with respect to whom it is eternal, or from the side of the effect of the action,
with respect to which it can be temporal. But God’s action is not a medium between
Trinity and
1
himself and his effect, rather his action is in him and is his entire substance; therefore his
action is by its very essence eternal, but the effect is temporal.” Scriptum super Sententiis I,
14, 1, 1, ad 3 (ed. R.
P. Mandonnet and M. F. Moos, 4 vols. [Paris: Lethielleux, 1929–47]) (hereafter In Sent.).
Trinity
11
orders each thing, and all things together, around his own Word in the
flesh; the Father wills a world which fits, in its totality, with Jesus
Christ. The fitness is twofold; it embraces both the origin and the
destiny, the beginning and the end, of all things. That is: the Father
wills a world in which all particular things and their various properties
and relations “hold together” in Jesus, and all particular things reach
the goal the Father wills only on account of Jesus’ cross and
resurrection – in the event, despite their own refusal of the goal and
consequent captivity to evil, so that their attainment of the final goal is
not only creation, but reconciliation and redemption. And the Son in
turn glorifies the Father by accepting and enacting the particular role
the Father appoints in the creation and redemption of the world the
Father wills.9
Only by the outpouring of the Spirit, however, will all things
actually reach the goal that the Father and the Son, each in his own
particular way, establishes for them. That goal, briefly put, is for all
things to share as fully as each is capable in the infinite beauty,
goodness, and truth of the divine being and life, which the Father
rejoices eternally to share in its totality with his Son and his Spirit,
and which they together rejoice to receive from him. The Father
creates a world which expresses, and thereby resembles, him by
conforming all things to his incarnate Word – to the one who alone
fully expresses his own total reality. The Father creates a world which
not only resembles him, but radically desires him and succeeds in
attaining him, because he makes that world by his Spirit. The Spirit’s
distinctive role in the triune God’s act of making “all things hold
together” is perhaps clearest with regard to the destiny of creatures. It
belongs chiefly to the Spirit to give all things their proper share in the
divine life. Let loose on “all flesh” (Acts 2:17) at Pentecost by the risen
and exalted Christ (cf. Acts 2:33), the Spirit enlivens and moves
9. On this see, e.g., Barth’s treatment of creation as a work of the triune God in Kirchliche
Dogmatik III/1 (Zurich: Evangelischer Verlag, 1932–67 [for the complete work]): “In view of
this one, his Son, who would become a human being and the bearer of human sin, God
loved humanity from eternity, before he created it, and with humanity the whole world –
in and in spite of its complete lowliness, non-divinity, and indeed anti-divinity. And he
created it precisely because he loved it in his own Son, who stood eternally before his
eyes as the one rejected and slain on account of its sin” (pp. 53–4; [ET, pp. 50–1]). The
suggestion that the Father envisions the world (indeed any world) through and for the
particular human being Jesus of Nazareth raises, to be sure, a host of difficult questions
about the relation between creation, sin, and redemption which cannot be treated in detail
here. Col., at any rate, is quite clear about the basic claim which raises the problems: the
one through whom and for whom “all things have been created,” and in whom “all things
hold together” (1:16, 17), is the crucified Jesus, viz., the very same one (αv’ro´ ç) “through
whom [God] was pleased to reconcile all things to himself, making peace through the
Trinity and
1
blood of his cross” (1:20).
Trinity
11
every creature in the way suitable to it so that, joined to the Father’s
crucified and risen ϵt’nω´ v, all may enjoy the good which the Father
intends for them. The Spirit is thus the agent who immediately brings
it about that all things receive the life of God by holding together in
Christ, and is in that sense the principal agent who moves them to
their final goal.10 So while willing that there be a world at all which
holds together through union with his Word in the flesh belongs
chiefly to the Father, and willing to be the flesh to which all things are
ordered belongs chiefly to the Word, the realization of this will in
creatures belongs chiefly to the Spirit.11
But the Spirit’s distinctive role in creation, like that of the Son, per-
tains to the origin as well as the destiny of all things. Pentecost is the
definitive enactment of the Spirit’s mission to join all things to the
incarnate Son, but that mission is already anticipated and prepared,
so the church’s trinitarian exegesis has regularly proposed, from the
beginning of creation: it is the same Holy Spirit who moves over, and
then gives form to, the formless chaos of Gen. 1:2, who is breathed into
the first human beings to give them life (Gen. 2:7) – and who is then
withdrawn from them on account of sin (cf. Gen. 3:19).12
The Spirit, this suggests, is a secondary exemplar, as well as the
chief agent, of the movement of creatures into the life of God. The love
of the Father and the Son for one another seems bound up with the
person of the Spirit in a distinctive way (though precisely how this is
so remains a matter of theological dispute).13 The Father loves the Son
Jesus, in eternity and in time, precisely by giving him the gift of the
Spirit – the one gift equal to and so worthy of both the giver and the
receiver – to repose in and on him; the Son loves the Father precisely
by gratefully receiving and rejoicing in this gift, and (in time) by
sharing the gift with the world. Father and Son are thus eternally
united or joined with one another in
10. As Aquinas argues: “In [created] things, the motion which is from God seems to
be attributed properly to the Holy Spirit.” Summa Contra Gentiles IV, 20 (no. 3571).
11.The Word’s willing to be flesh is shared with the Father and the Spirit, not, of course,
the being flesh itself. The former can therefore be appropriated to him, while the latter, as
unique or proper to him, cannot.
12. For a sketch of the classic exegesis, see Boris Bobrinskoy, Le Mystère de la Trinité: Cours de
théologie orthodoxe (Paris: Cerf, 1986), pp. 21–70.
13.Viz., as to whether the mutual love of the Father and the Son in the Spirit requires, or
even permits, the affirmation that the Spirit eternally proceeds ex Patre Filioque. For
some reflection on how traditional East–West disputes about the eternal procession of the
Holy Spirit might embody a disagreement which makes no difference there is any need
to resolve, see Bruce D. Marshall, “Action and Person: Do Palamas and Aquinas Agree
about the Spirit?” St. Vladimir’s Theological Quarterly 39/4 (1995), pp. 379–408. For a different
argument to a similar conclusion from the Orthodox side, see Serge Bulgakov, Le
Trinity and
1
Paraclet, trans. Constantin Andronikof (Paris: Cerf, 1946), pp. 87–143.
Trinity
12
and through the Holy Spirit as a person distinct from both, though the
love which unites them, as it springs only from fullness and not from
lack, does not have the distinctive note of desire. That creatures
succeed in attaining the God who perfectly expresses himself in Jesus
Christ results from the gift of this Spirit; that they, made from nothing
and so lacking all, cannot help desiring union with this God, whether
or not he ever wills to pour out his Spirit so that they may attain him,
results from their creation in the image of this same Spirit. Anything
which is, however remotely, like the Spirit to whom it eternally belongs
to unite the Father and the Son in love, will naturally seek its own
share in that love.14
Epistemic right as christological coherence
What then of deciding about truth? We undertook this exploration of

the church’s trinitarian identification of God in the hope of answering
the question by what right the church takes the beliefs which make up
that identification as primary when it decides about truth. What does
the content of these beliefs suggest about their epistemic status?
Identification of Jesus as epistemic trump

As he is identified by the church’s canonical narrative, the particular
person Jesus of Nazareth holds all things together, with regard to both
14. These last remarks are suggested by medieval Western views of the Spirit as not only
agent but exemplar of the love of creatures for the Father in the Son. Thus, e.g.,
Bonaventure, commenting on John 17:22: “In prayer the Lord asks that his disciples be
united, not by nature, but by love (dilectionis), in conformity to the highest unity. Now
the members of Christ are united by mutual love (amorem). Therefore in the divine there
is an exemplar of this,” namely “the third person, who proceeds in the manner of mutual
love (caritatis).” In I Sent. 10, 1, 3, a (Opera Selecta I, p. 160a). With regard to all creation,
and linked to the exemplarity of the Son: “All creatures come forth from God by thought
and will. But in the divine we have to suppose, prior to the production of creatures, the
eternal emanation of the Word, in whom the Father laid out all things that were to be
done. For the same reason, therefore, it was necessary that a person emanate in whom he
willed and gave all things” (In I Sent. 10, 1, 1, d; Opera Selecta I, p. 156a). While not always
thought of as having Bonaventure’s interest in exemplarism, Aquinas makes a similar
argument (though without the suggestion that knowledge and will in God not only
explicate, but demonstrate, the trinitarian processions). “Assuming, according to our faith,
the procession of the divine persons in a unity of essence (which no argument can be
found to prove), the coming forth of the persons, which is perfect, will be the pattern
(rationem) and cause of the coming forth of creatures . . . Thus the coming forth of the
creature, insofar as it stems from the generosity of the divine will, may be traced back to
one principle, which is as it were the pattern (quasi ratio) of this entire generous conferral
. . . and this is the Holy Spirit,” that is, “a person . . . in the divine who comes forth in the
mode of love” (In I Sent. 10, 1, c; see In I Sent., prologus: “Just as a waterway is diverted
from a river, so is the temporal procession of creatures from the eternal procession of
persons.”). On these texts see Gilles Emery, La Trinité créatrice (Paris: J. Vrin, 1995).
Trinity and
1
their creation and their redemption. Only that can be whose existence
and attributes fit with his; even that in creation which comes to oppose
him must be wholly capable of being redeemed by him. This fitness
between Jesus and other creatures is an asymmetrical rather than a
mutual relationship. The being of other creatures is not simply
compatible with his own, but as a whole causally dependent upon
him; he is not simply the exemplar but the agent of creation – together,
of course, with the Father and the Spirit, each in his own way.
This suggests that when it comes to the epistemic relation between
beliefs – when it comes to deciding what is true – the identification of
Jesus which relies on the church’s canonical narrative must have
primacy. When we ascend from the content of the church’s central
beliefs to the question of their epistemic status, these beliefs seem to
require that we accept the following conditional: if identifying
descriptions of Jesus in the church’s canonical narrative are held true,
then the sentences by which we identify and describe other things
must, if we are to hold them true, at least be compatible with (that is,
not contradict) the sentences by which we identify and describe Jesus.
This consistency relationship too is asymmetrical; holding true the
beliefs by which the church identifies Jesus requires deciding about the
truth of other beliefs by seeing whether they are at least consistent
with the narrative identification of Jesus, and not deciding about the
truth of that narrative identification by seeing whether the beliefs
which make it up are consistent with others. The narratives which
identify Jesus are epistemic trump; if it comes to conflict between these
narratives and any other sentences proposed for belief, the narratives
win.
That the meaning of the narratives which identify Jesus implies
their epistemic primacy may perhaps most clearly be seen by
considering the obvious alternative.15 It might be held that when
conflict with other beliefs arises, then beliefs expressed by holding
these narratives true lose the epistemic confrontation.16 This epistemic
decentralization of the narratives which identify Jesus might be
maintained as a general principle, or it might be proposed ad hoc; the
range of sentences with respect to
15. The implication is, more precisely, that if the narratives are held true, they must be
held to be epistemically primary, and if they are true, they must be epistemically primary.
16. There might of course be cases where the consistency of this or that belief with the
canonical narrative was undecidable by us, at least in our present state of ability to work
out the logical relations among our beliefs. This would have no bearing on the
epistemic primacy of central Christian beliefs, since the rule which defines this primacy
is that conflicts must always be decided in favor of the Christian beliefs, not that we can
Trinity
12
always tell when there are conflicts.
Trinity and
1
which the narratives were epistemically subordinate might be wide or
narrow. Either way, when beliefs constitutive of the narrative
identification of Jesus are held false, it becomes impossible to believe
that in Jesus so identified, all things hold together. To regard as false
chief elements in the New Testament’s identification of Jesus is to hold
that there is no one who meets or satisfies that description.17 And if
there is no one who answers to the New Testament’s identifying
description of Jesus, then all things cannot hold together in him, since
nothing can hold together in that which is not. More precisely: to hold
false sentences indispensable to the New Testament’s identification of
Jesus because holding them true would create conflict with some other
beliefs is to suppose that states of affairs obtain which are incompatible
in some way with Jesus’ existence and the attributes unique to him.
The status of “that which holds all things together” might or might not
be ascribed to such states of affairs; in either case, the canonically
identified person Jesus would not be the one who holds all things
together. If he is, then the nexus of belief by which the church
identifies him cannot lose epistemic conflicts with other beliefs.18
So if all things hold together in Jesus, crucified and risen, then at the
epistemic level it seems that all true beliefs must hold together in – be
logically consistent with – the narratives which identify him, and the
triune God with him. He must have “primacy in everything,”
including decisions about truth.
Unrestricted epistemic primacy

As this line of argument already suggests, the epistemic primacy of the
church’s narrative identification of Jesus, and with him of the triune
God, must be unrestricted; it must range across all possible beliefs.
There can be no type or area of belief which is exempt from the
requirement that it be at least consistent with the body of beliefs which
identify the crucified and risen Jesus. That is: no matter what the
contents of our various
17. This holds good regardless of how one specifies what these chief elements are. The
present claim is not, therefore, that everything in the New Testament, or in the Gospels,
is equally important to the identification of Jesus. That the narratives of the passion and
resurrection are among the principal elements, as here supposed, is relatively non-
controversial.
18. The same point can be put in terms of Tarski–Davidson definitions of truth for
sentences. If all things in fact hold together in Jesus Christ, then the right branch of
any T-sentence whose left branch could possibly be true must state truth conditions for
the sentence on its left which are compatible with the truth conditions (also as stated
by
T-sentences) of the narratives which identify Jesus and the triune God.
Trinity
12
beliefs, no matter how remote the meaning of many sentences we hold
true may be from that of the church’s canonical narrative, those
sentences must still be tested for truth against the church’s most
central beliefs. If we hold true the canonical identification of Jesus,
then we cannot decide about the truth of our philosophy, our politics,
or even our science without seeing whether those beliefs are consistent
with, and to that extent subject to correction by, the beliefs which
identify Jesus.
This is, once again, required by the content of the beliefs
themselves. Suppose we hold beliefs, or at least accept the possibility
of beliefs, which do not have to be tested for their consistency with the
narrative identification of Jesus. We then build into our epistemic
structure the possibility that there are objects, events, or states of
affairs, whether past, present, or future, which in one way or another
do not fit with Jesus Christ crucified and risen, which are incompatible
with his existence and attributes, and so with creation and redemption
through him. He will therefore not “have the primacy” with respect to
those things; they will not “hold together” in him; rather, on account of
them, he will lack the features ascribed to him in the canonical
narrative. Thus if all things, not just some things, hold together in
Jesus, then it seems that all beliefs, not just some beliefs, must fall
within the epistemic range of the beliefs which identify him.
Believing the gospel (that is, the narratives which identify Jesus and
the triune God), therefore, necessarily commits believers to a
comprehensive view of the world centered epistemically on the
gospel narrative itself. On such a view there will be no regions of belief
and practice which can isolate themselves from the epistemic reach of
the gospel. But conversely having such a comprehensive view also
means that Christians will not be able to isolate the gospel from the rest
of their beliefs, to be held true for whatever restricted purposes, pious
or otherwise, Christians may want to use it. On the contrary, believing
the gospel at all means that Christians must always venture forth,
prepared to engage as best they can all life and reality – to interpret
and assess whatever alien or novel beliefs they may encounter – in
light of the narratives which identify Jesus.19
Taking the narrative identification of Jesus and the Trinity as
primary in decisions about truth, and that without restriction, is not at
all the same thing as taking the beliefs which make up that
identification as the sole criteria of truth. Nor does the unrestricted
epistemic primacy of this
Trinity and
1
19. On this see chapter 6.

Trinity
12
nexus of belief require this sort of implausible epistemic exclusivity.
Many, indeed the great majority, of our decisions about what is true
cannot be made by appeal to the church’s central beliefs alone, but
require that we advert to other relevant beliefs. Consistency with the
narratives which identify Jesus is the condition sine qua non which all
other beliefs must meet in order to be true, but this condition will
rarely be sufficient to decide about the truth of philosophical, political,
natural scientific, or other beliefs, or to settle disputes between
competing claims in those areas. Thus the unrestricted epistemic
primacy of the church’s central beliefs can be stated only negatively: no
sentences which are inconsistent with these beliefs can be true
(including any other beliefs – such as logical laws – to which we may
advert epistemically), but consistency with the church’s central beliefs
does not normally guarantee, all by itself, that other beliefs are true.20
It is important to observe that one can, at least in principle, fail to
maintain the unrestricted epistemic primacy of the gospel narrative
without regarding any elements of the narrative as false. Holding
aspects of the narrative false is, of course, the clearest way of denying
its epistemic primacy. But the fact that one holds the narrative true
need not indicate anything more than that one has a consistent system
of belief. In any consistent belief structure even those members which
have the lowest epistemic status will win conflicts with beliefs which
are inconsistent with them. By itself, a consistent belief system says
nothing about the epistemic priorities of those who hold the beliefs.
Epistemic primacy is by contrast a normative relationship, such that
for any beliefs A and B, A is epistemically primary with respect to B if
and only if, should inconsistency arise between A and B, A is held true,
and B rejected or modified. When taken as A, the narratives which
identify Jesus and the triune God will be epistemically primary
without restriction just in case B could be any other possible belief. To
ascribe unrestricted epistemic primacy to the gospel narrative is thus
not simply to hold, as a matter of fact, no beliefs which are inconsistent
with it, but to be prepared to reject any possible belief which is
inconsistent with it.
Of course questions about one’s epistemic priorities normally come
up only when conflict arises among beliefs one is pretty deeply
committed to
20. This formulation guarantees that unrestricted epistemic primacy belongs exclusively to
the narrative identification of Jesus; saying that all other beliefs must be consistent with
these ensures at a stroke both that these are unrestrictedly primary and that the rest are
Trinity and
1
not.
Trinity
12
holding true. Faced with conflict, one might attempt to maintain a con-
sistent system of belief which included the chief elements in the
church’s narrative identification of Jesus, but at the same time reversed
the epistemic priorities implied by the narrative. Resolving epistemic
conflict in this fashion is the aim of the dependence thesis in theology.
Some reflection later in this chapter on the belief that Jesus is risen
will suggest how difficult this turns out to be in practice.
Epistemic primacy and highest truth

If we hold true the narratives which identify Jesus and the Trinity, we
are committed by the content of those narratives to regarding them as
epistemically primary across the board. This implies that a
community which has the beliefs expressed by holding these
narratives true must regard that body of beliefs as what could be called
the highest truth – not only the highest available truth, but the highest
truth there can be. To say that central Christian beliefs have
unrestricted epistemic primacy means that any possible belief which
contradicts them must be false. Ascribing genuinely unrestricted
primacy to these particular beliefs, moreover, pre- empts the
application of the category; no other beliefs will be able to enjoy this
logical status. From this it follows that no true belief can contradict the
narratives which identify Jesus and the Trinity. If these narratives are
believed, therefore, those who hold them true cannot con- sistently
suppose that they could possibly turn out to be false. And this is just
what it means to say, at least when it comes to deciding about truth,
that a set of beliefs is the highest truth.
That the narratives which identify Jesus and the Trinity have to
function, if held true at all, as the highest truth helps give a clearer
picture of the idea that epistemic justification is finally christological
and (thereby) trinitarian coherence. On this view, the totality of beliefs
which human beings might hold can be seen as forming an open field,
ordered around a christological and trinitarian center. This center is
also, as it were, a peak or summit from which the whole field can be
surveyed, though of course its distant parts will be seen less clearly
from the center than those close by. Every part of the field is
contiguous with every other part, since nothing can belong to the field
– can count as a sentence or belief – unless we can assign to it a
meaning or content, which links it in manifold ways to the rest of the
field. There are thus no regions inaccessible from the rest of the field,
no boundaries which the intrepid and interested explorer cannot cross.
Trinity and
1
The distance between sentences and beliefs in this logical
Trinity
13
space follows upon their difference or similarity of meaning and
content; those sentences are closer to one another whose meanings are
more alike, more intimately connected.
In this open field of possible sentences or beliefs, no belief which
fails of consistency with the christological center can be counted as
true. Consistency is of course the most minimal kind of coherence
among beliefs, but for just that reason it enables us to define the
unlimited epistemic reach of the narratives which identify Jesus in a
precise and infor- mative way. Since no two inconsistent sentences can
both be true, regardless of what the sentences mean, consistency with
these narratives can be required of all sentences as a test of truth, and
we have readily available interpretive and inferential procedures for
figuring out whether the test has been met in most cases. Ascertaining
whether this minimal sort of coherence with the canonical narrative
obtains will leave the truth of a great many beliefs undecided, but it
will already configure the field of belief in quite definite ways, by
eliminating lots of beliefs which lack consistency with those which are
centrally Christian.
Coherence of course comes in many varieties. If the minimal kind of
coherence among beliefs is consistency, the maximum is identity. A set
of sentences are consistent when its members are all possibly true; the
sentences in the set are identical, for present purposes, when they all
mean the same thing. So we normally assume that “Grass is green” and
“Schnee ist weiß,” while they mean quite different things, are
consistent with one another (they can both be true), and yield
compatible beliefs; we normally assume that “Grass is green” and
“Gras ist grün” mean the same thing, and that those who hold these
sentences true have the same belief. A sentence proposed for belief
which meant the same thing as one of the sentences which make up
the church’s narrative about Jesus and the Trinity would thus have the
maximal sort of coherence with those beliefs to which unrestricted
epistemic primacy belongs.
If consistency with the beliefs which are epistemically primary is
necessary for any other belief to be regarded as true, it would seem
that identity (of meaning) with the primary beliefs would have to be
regarded as guaranteeing truth. Almost as strong a form of coherence
as identity of meaning is necessary implication; logically necessary
inference from those beliefs which are epistemically primary would
likewise seem to guarantee truth to whatever sentences were inferred.
A weaker form of coherence, but still stronger than consistency, is
Trinity and
1
what the medievals called convenientia. One belief coheres with another
ex convenientia when
Trinity
13
the one makes it particularly fitting or appropriate, but not necessary,
to believe the other. So belief in the incarnation, for example, fits in an
especially beautiful way with belief in the infinite and selfless
goodness of God, but (insofar as the incarnation is also believed to be a
contingent act of divine freedom) it is not a necessary implication of
God’s goodness; the truth of belief in God’s goodness does not by
itself, therefore, guarantee the truth of belief in God’s incarnation.21
Instead of continuing this general taxonomy of christological coher-
ence, this and subsequent chapters will attend to particular cases.
These general remarks indicate, however, that when it comes to the
varieties of christological coherence, scope and decisiveness vary
inversely. The more narrow the range of beliefs across which a
particular type of coherence with the church’s central narrative can
function as a relevant test of truth, the more coherence tends to
guarantee truth. Conversely, the wider the range across which the test
applies, the more it tends simply to permit truth rather than support or
guarantee it. Taken by itself, consistency with the christological and
trinitarian center does not guarantee truth, but it does, crucially,
guarantee falsity to those beliefs which lack it, and that on the widest
possible scale. It alone applies as a relevant test to all possible beliefs;
its scope is unlimited, and so it alone can serve to define the epistemic
primacy of the church’s central narrative.
The Father’s epistemic role

By what right, then, does the Christian community finally take its
narrative identification of Jesus as epistemically primary across the
board, and so as the highest truth? The answer proposed by the
content of the narratives, read in their full trinitarian depth, is that
Jesus is the icon of the Father. This human being perfectly expresses
the Father in the world – not all by himself, in isolation from
everything else, but by being the one in whom all other things hold
together.
When the Father envisions this and any world which he might actu-
ally will to create, and when he wills that this world in fact be and be
redeemed, he orders it in its totality around his Word in the flesh.22 It
is
21. For the scholastics themselves, it was normally states of affairs which were conveniens,
rather than beliefs about them (the incarnation of God, rather than the belief that God is
incarnate). For examples regarding the case at hand, see Aquinas, Summa theologiae III, 1, 1;
Bonaventure, In III Sent. 1, 2, 1 (who speaks of congruitas rather than convenientia).
22. As Eph. 1 also suggests: the mystery of the Father’s will now made known in Christ (v. 9)
is his resolve (v. 9b) “before the foundation of the world” (v. 4) to gather up all things in
Trinity and
1
heaven and on earth under Christ as their one head (v. 10). See the comments of Heinrich
Schlier, Der Brief an die Epheser, 6th edn (Düsseldorf: Patmos Verlag, 1968), pp. 63–6.
Trinity
13
entirely natural for the Father to order the world in this way. As the
Father’s own Word become our flesh, who alone makes the human
journey from Bethlehem to Golgotha and the Emmaus road, Jesus is
the Father’s very image, the one who perfectly expresses the total
reality of the Father. What the crucified and risen Jesus expresses
naturally includes the Father’s mind and will, which he shares not
simply by qual- itative but by numerical identity.23 Indeed, as the
Father’s Word and image incarnate, Jesus Christ is the Father’s way of
uttering or expressing the very Trinity of divine persons, as well as
every creature the Trinity creates.24 As a result the whole ordering of
creation will have to fit with – hold together in relation to – what
happens in the crib at Bethlehem, on the cross of Golgotha, and with
the disciples on the Emmaus road. Any ordering of creation as a whole
which did not fit with the features by which the church identifies this
particular person would on that account fail to fit with the Father’s
own mind and will (since it would be incompatible with what
perfectly expresses that mind and will); on any such ordering the
Father would, per impossibile, be at war with himself.
The upshot of these trinitarian considerations is that we have
reached the end of the epistemic road. The Father’s knowledge is
definitive. This means more than that his knowledge cannot be
mistaken. The Father is, together with his Word and Spirit who come
forth from him, the source of all things in their entirety, and so of their
order and relation. As such the Father (and therefore the triune God)
cannot be thought of as waiting, the way we must, for things to exist in
order to know them. His very knowledge of them must be productive
of their existence and attributes (presuming, of course, that he knows
that they exist; their existence itself must, on the same assumption that
the triune God is the source of all things, be wholly dependent on his
will). The Father is the final measure of all things; nothing can be other
than as he orders it, other than it is in his mind and will.25
23. The enfleshed Word expresses the will of the Father in two different senses: (i) he
fully shares that capacity to act which belongs to the Father’s divinity, and is
communicated to him by eternal generation, and (ii) he freely accepts the Father’s
eternal but contingent decision that he in fact be enfleshed, viz., that he be Jesus of
Nazareth.
24. As Thomas Aquinas suggests: “The whole Trinity is spoken in the Word, and every
creature as well” (Summa theologiae I, 34, 1, ad 3). See Anselm, De Processione Spiritus Sancti
11: “One who sees the Son sees the Holy Spirit, just as he sees the Father.” F. S. Schmitt,
ed., S. Anselmi Opera Omnia (Edinburgh: Thomas Nelson, 1946–61), vol. II, p. 208, 23–4.
25. Or: nothing can be incompatible with the Father’s mind and will, since all things are
just because he (and with him the Son and Spirit) knows and wills them. This obviously
brings to mind conceptual problems about evil which I cannot pursue in detail here. It
may simply be observed that the argument here about epistemic primacy coheres with two
Trinity and
1
Thus if we believe things to be other than they are, and are ordered,
in the Father’s mind and will (other than the Father believes them to
be, we would say, if the Father had beliefs), our belief cannot be true.
In the Father’s mind and will, all things are ordered to Jesus, crucified
and risen. Consequently to hold any belief inconsistent with the
narrative identification of Jesus is to believe things to be other than
they are in the Father’s mind and will. And any such belief must be
false. In sum: if the Father expresses his own total reality by ordering
all things around the crucified and risen Jesus – if Jesus is the Father’s
icon – then we will have the epistemic right to hold only those beliefs
which are ordered around (at minimum, are consistent with) the
beliefs which identify Jesus. Beyond the Father’s knowledge no further
or more basic justification for the unrestricted primacy of the church’s
central beliefs could reasonably be sought, or given, since none can be
conceived.
That all things hold together in Jesus Christ, crucified and risen, is
the work not only of Jesus and the Father, but of the Spirit in whom
they are united in love, and whom they pour out upon all flesh. As
exemplar the Spirit orients all things to the Father’s icon as their chief
desire, and as agent the Spirit chiefly brings about the realization of
this desire. This suggests that the epistemic habits which define
Christian identity are skills only the Spirit can teach us. Ordering all
of our beliefs around the gospel of Christ requires a massive reversal of
our settled epistemic habits and inclinations, of our usual ways of
deciding what is true. Only the Holy Spirit is up to the epistemic
effort involved. The Spirit alone can teach us to recognize in the
narratively identified Jesus the Father’s own icon, and to interpret and
assess all of our beliefs accordingly. To the way the Spirit carries out
his distinctive epistemic role we will return in detail in a later chapter.
But we already have the outline of an answer to our question about
epistemic justification. The Christian community decides what is true
by learning from the Spirit to read the narratives which identify Jesus
in
traditional solutions to these conceptual problems. (1) To say that nothing can be
inconsistent with the Father’s knowledge and will is not to say that the Father wills evil,
nor that he fails to know it (since in knowing it he would produce it) but that evil is not,
strictly speaking, part of creation; it is the privation of goodness and therefore being, and
so not what the Father wills, which eo ipso comes to be. (2) There is a sense in which, given
this privative metaphysical status, evil “fits” with creation in that nothing which suffers it
will fail to be redeemed; the ordering here claimed of the origin and destiny of all things
around Jesus would only prove incoherent if there was evil inherently beyond the scope of
redemption, evil not destined, however mysteriously, to be made good, and in that way to
Trinity
13
disappear.
Trinity and
1
their full trinitarian depth, and so to recognize the crucified and risen
Jesus as the perfect icon of the Father; the church finds in the New
Testament depiction of this distinctive relationship between Jesus and
the Father the ultimate warrant for taking the trinitarian narratives as
epistemically primary across the board. When it comes to epistemic
right, this is the force of saying, “Jesus Christ is the truth.”
Precedents
Here there is no room to explore some of the deep precedents in the
tradition for this epistemic outlook. We may simply indicate that
theologians agree on it who are often thought to have utterly divergent
epistemic commitments. Along lines we have sought to make explicit
Thomas Aquinas, for example, argues that “The chief matter in the
teaching of the Christian faith is the salvation accomplished by the
cross of Christ,” which is fool- ishness to the world “since it includes
something which seems impossible according to human wisdom,
namely that God dies (Deus moriatur), and that the omnipotent becomes
subject to the power of the violent.”26 This requires that Christians
and Christian theology keep their epistemic priorities straight – as
Aquinas concretely and colorfully puts it, “Whatever is not in agreement
with Christ is to be spewed out . . . because he is God.”27 But Luther’s
approach to matters epistemic, while motivated by concerns quite
unlike Thomas’s and expressed in a very different way, is, as I have
argued elsewhere, basically the same in substance.28
The epistemic picture developed here is also drawn with
remarkable clarity and economy by Anselm, in a text from the De
concordia. Anselm here speaks simply of consistency with scripture
rather than specifically with the narratives identifying Jesus, and the
concepts I have used are not entirely Anselm’s own. But little of the
foregoing argument is not stated or implied in his remark; indeed one
key feature of Anselm’s picture – the epistemic relation between the
church’s central beliefs and what seems supported by “evident reason”
– we have yet to treat in detail. His words can stand as an apt summary
of the argument so far.
26. In I Cor. 1, 3 (nos. 45, 47). S. Thomae Aquinatis super Epistolas S. Pauli Lectura, vol. I, 8th
edn, ed. R. Cai, O. P. (Turin: Marietti, 1953).
27. In Col. 2, 2 (nos. 95–6). S. Thomae Aquinatis super Epistolas S. Pauli Lectura, vol. II. In a
more familiar idiom: “It does not belong to sacred doctrine to prove the principles of the
other sciences, but only to judge them: for whatever is found in other sciences which
contradicts the truth of this science is totally to be rejected as false, according to II Cor.
10:[5]” (Summa theologiae I, 1, 6, ad 2).
28. See my “Faith and Reason Reconsidered: Aquinas and Luther on Deciding What
Trinity
13
is True,” The Thomist 63/1 (1999), pp. 1–48.
Trinity and
1
We proclaim nothing of use to spiritual salvation which holy
scripture, made fruitful by a miracle of the Holy Spirit, does not
either put forward directly or contain within itself. For if at times we
say something by reason which we cannot either point out explicitly
in what scripture says or prove on that basis, we can know by
scripture whether it is to be accepted or rejected, in the following
way. If it is gathered by evident reason, and scripture at no point
contradicts it (since scripture, just as it is opposed to no truth, also
supports no falsity), then what is said by reason is taken under the
authority of scripture, by the very fact that scripture does not negate
it. But if scripture opposes our view beyond any doubt, then
although our reasoning may seem unavoidable to us, we cannot
suppose that it is supported by any truth. Thus holy scripture
contains within it authority over every truth which reason gathers,
since scripture either openly affirms it or at least does not negate
it.29
Epistemic priorities, truth commitments, and plausible

interpretation
The unrestricted epistemic primacy of those beliefs which are most

central for the Christian community has so far come to light as a
necessary implication of their content. To hold these beliefs true at all,
so an analysis of their content suggests, requires treating them as
epistemic trump. Recalling the philosophical considerations about the
truth-dependence of meaning introduced earlier reinforces this
conclusion. The key issue is how we may plausibly connect the truth
value, the meaning, and the epistemic status we assign to sentences.
The last chapter argued that we cannot on the whole interpret sen-
tences plausibly (assign meanings to them) except by maximizing the
ascription of truth to sentences which speakers hold true, and so by
opti- mizing agreement between speakers. This is one way (mainly
Davidson’s) of working out in a full-scale, non-question-begging
theory of interpretation Frege’s contention that the meaning of a
sentence is given by its truth conditions, and the meaning of the
constituents of a sentence (in a
29. De concordia III, 6, F. S. Schmitt, ed., S. Anselmi Opera Omnia vol. II, pp. 271, 26–272, 7.
See Cur Deus homo I, 18: “It is certain that if I say anything which without doubt
contradicts holy scripture, it is false, and I would not want to hold it if I knew” (Opera,
vol. II, p. 82, 8–10). Commenting on these passages, Michel Corbin asks rhetorically,
“Can we say that
. . . [Anselm] has taken so seriously the christological titles Logos (Jn. 1:1) and sophia (I Cor.
1:22) that only that to which faith gives its assent is rational?” (cf. “justified”). “Louange et
Trinity
14
grâce, nuit et jour,” Michel Corbin and Henri Rochais, eds., L’œuvre de S. Anselm de
Cantorbery, vol. v (Paris: Cerf, 1988), p. 25.
Trinity and
1
natural language, the words) is the contribution they make to the
meaning of sentences in which they might be used.30 We need now to
develop this thought so as to factor in the issue of epistemic primacy.
How is the epistemic status of a sentence related to belief and meaning
– to the truth value we assign the sentence and the interpretation we
give it?
The argument – to state its conclusion at the outset – will display
the great strength of the links between the epistemic status of a
sentence, its plausible interpretation, and the truth value we set on it.
A change in any one of these variables affects the others in quite
specific ways. These connections have the look of logical bonds rather
than merely psychological ones; attempts to break them (whether or
not deliberate) lead in predict- able ways to incoherence or
implausibility. Since we want to test the claim that when central
Christian beliefs are held true they must function with unrestricted
epistemic primacy, our chief concern for the moment is the link
between truth value (that is, the assignment of positive truth value, or
belief ) and epistemic status.
At first glance, the link between truth value and epistemic status
looks obvious: roughly put, the more persistently we try to uphold the
truth of a belief, the wider the range of beliefs over which it has
epistemic primacy, and thus the higher (or more deeply rooted) the
epistemic status it has for us. We can grasp the significance of this
bond by looking at what happens to interpretation when theologians
committed to the dependence thesis try to break it.
Interpreting Jesus’ resurrection

The dependence thesis is the claim that we may hold central Christian
beliefs true without regarding them as epistemically primary; on this,
so the argument goes, rests the possibility of being an intellectually
responsible Christian in the modern world. Some reflection on belief
in Jesus’ resurrection will provide a useful demonstration of the
difficulties inherent in this epistemic procedure.31 Though
dependence theorists differ on which beliefs are essential to
Christianity, many would agree with the
30. Davidson’s holistic formulation takes the matter a step further: “Frege said that only
in the context of a sentence does a word have meaning; in the same vein he might have
added that only in the context of the language does a sentence (and therefore a word)
have meaning” (“Truth and Meaning,” Inquiries into Truth and Interpretation [Oxford:
Clarendon Press 1985], p. 22).
31. For a cognate argument, suggested by both Luther and Aquinas, regarding belief in the
Trinity
14
incarnation, see Marshall, “Faith and Reason Reconsidered,” pp. 33–46.
Trinity and
1
claim made in chapter 2 that belief in Jesus’ resurrection is central to
communal and individual Christian identity (though perhaps for
reasons different from the ones there offered). For present purposes
the decisive question is whether it is possible to give a plausible
interpretation of “Jesus is risen” when, in dependence-theoretical
fashion, one both (a) wants to hold it true, even in the face of epistemic
conflict, and (b) treats it as epistemically subordinate to and dependent
upon other beliefs, with which conflict may arise.
The New Testament’s talk of Jesus’ resurrection is of course
complex. But the meaning of “Jesus is risen,” whatever else it may
involve, seems at least to include the predicates placed on the lips of
the glorified Jesus himself in Revelation: “I was dead, and behold I
am alive forever and ever” (Rev. 1:18). This seems to capture the
basic action of the Gospel stories of Jesus’ resurrection, which narrate
(a) that Jesus once was dead, and (b) that now he lives forever, such
that people can now encounter him in person. “Is risen” appears, in
other words, to be basically a two part predicate, composed of “was
dead” and “lives forever.” “Dead” and “lives” have their usual
meanings; the astonishment lies in conjoining them in this sequence –
in saying that there is a temporal point at which “lives” may rightly be
applied to this person which follows the point at which “is dead” may
rightly be applied to him. This, to be sure, gives rise to some novel
connotations of “lives,” such that it can, for example, now be joined to
“forever.” Using the standard logical form of attribu- tion, the New
Testament applies this two part predicate to “Jesus.” “Was dead”
and “lives forever” are (in that sequence) both said to be true of the
person Jesus, and this same person is the subject of both of these
predicates (a point on which the Gospel narratives, with their
references to the risen Jesus bearing the wounds of his crucifixion,
lay particular stress). To believe that Jesus is risen, therefore, means at
least to believe that
(1) Jesus was dead,
(2) Jesus now lives,
and that
(3) “Jesus” has the same referent in both (1) and (2).
Theologians committed to the dependence thesis, striving to make
“Jesus is risen” come out true while avoiding conflict with what they
regard as the primary standards of truth for that sentence, have tended
to come up with two different interpretations of it. Schleiermacher can
rep-
Trinity
14
resent the first proposal, and Rudolf Bultmann the second (and much
more common) one.
In his lengthy discussion of the resurrection and ascension
narratives at the end of Das Leben Jesu, Schleiermacher insists that
Jesus must have appeared to and conversed with his disciples after his
crucifixion and burial. Otherwise the historical reliability of all the
Gospel reports of the appearance of the redeemer and of the
redemptive impression he made upon his followers would be void.32
This means that he is committed to the truth of the resurrection (or
more precisely, the appearance) narratives, but also to (3) – that is, to
accepting the logical form which attributes “is risen” as a genuine
predicate to Jesus, and not to anyone or anything else. The question is
what to make of the predicate itself, that is, of (1) and (2).
Schleiermacher insists with particular clarity on the interpretation
of (2). Jesus’ “second life” with his disciples after his cross and burial
is a wholly natural one, a return to the same sort of human life he
enjoyed before the cross.33 This follows naturally from his epistemic
priorities. In order to harmonize Christian claims with the best
deliverances of scientific and historical reason, we should not hold
any beliefs which invoke the supernatural or miraculous.34 The
resurrection narratives are true; therefore they must be interpreted in a
way which attributes to Jesus a life after the cross entirely continuous
with his life before it. What then of (1)? In a remarkable passage,
Schleiermacher proposes that it makes no difference to our
interpretation of the New Testament’s talk of Jesus’ resurrection
whether or not we take the text to mean that Jesus actually died. The
one sure sign of death, he argues, is the onset of bodily decom-
position, and we do not know whether Jesus’ body began to
decompose; he might simply have been catatonic, and thus only have
appeared to be dead. Our ignorance need not concern us, since what
matters about Jesus’
32. See Friedrich Schleiermacher, Das Leben Jesu, ed. K. A. Rütenik, Friedrich
Schleiermacher’s sämmtliche Werke, pt. I, vol. vI (Berlin: G. Reimer, 1864), p. 471. ET: The Life
of Jesus, trans. S. MacLean Gilmour (Philadelphia: Fortress Press, 1975), p. 442.
33. “Through the resurrection Christ returned to a genuinely human life”; the risen
Jesus “says explicitly that he is not a being existing outside the ordinary course of
nature, but a fully human body.” Therefore “one should think of his condition as the
restoration of his life entirely in the former manner,” and interpret any contrary
indications in the texts (such as the risen Jesus passing through closed doors) accordingly
(Das Leben Jesu, pp. 498; 473, ET, pp. 469, 444).
34. Except insofar as the very existence of the redeemer is a miracle, in that it could not
have been caused by any antecedent natural or historical states of affairs. See Das Leben Jesu,
p. 474 (ET, p. 445); Der christliche Glaube, 2nd edn (ed. Martin Redeker, Berlin: de Gruyter,
1960), § 93,3.
Trinity and
1
death is not whether it actually occurred, but his attitude toward it, the
continuance of his fully unimpeded God-consciousness even in his suf-
fering.35
For Schleiermacher this is not simply an historical hypothesis to the
effect that Jesus might not actually have died on the cross, but also a
proposal about what the Gospels actually mean when they say that
Jesus “died.” He does not hold, as one might expect, that the Gospels
clearly affirm Jesus’ death on the cross, but we cannot be sure whether
what they say is true. He seems, rather, to be sure that what they say is
true, but not to be sure what they mean. He finds the problem acute
because he has already fixed on an interpretation of the resurrection
narratives which requires the elimination of miraculous or
supernatural elements from Jesus’ life after the tomb. The task is now
to find a plausible interpretation of the accounts of the crucifixion
which fits with this reading of the resurrection narratives. He hits on
the solution of saying that when ascribed to Jesus in the New
Testament, “dead” can be taken to mean “no longer alive,” but it can
equally well be taken to mean “asleep,” that is, “not no longer alive.”
Of course this proposal does not interpret the troublesome term,
but rather fails to interpret it. To propose that a stretch of discourse
really supports contradictory interpretations ascribes to the discourse a
level of vagueness or confusion which puts it at the margins of
intelligibility, where we are not sure we are dealing with language at
all. As an interpretive recourse this suggestion can be employed only
sparingly, and as a last resort. Failure to arrive at a consistent
interpretation signals with far greater likelihood that one’s
interpretive effort has somewhere gone astray.
In fact Schleiermacher never suggests that the problem is unusual
vagueness or incomprehensibility in the texts. Having said that both
interpretations are permissible he instead struggles to strike some sort
of balance between the two, oscillating between a plausible
interpretation
35. See Das Leben Jesu, pp. 443–5 (ET, pp. 415–17). Schleiermacher sometimes seems to lean
toward denying that Jesus died on the cross: one “has to grant that in Christ’s situation
there was not yet the least beginning of decomposition, and thus not death.” But there is
finally no need to decide the issue: “In this matter we can take a position of complete
equanimity, and give an unbiased account of the details without any particular interest in
whether it comes out one way or the other” (p. 444; ET, pp. 416–17). Though he probably
strives to find more consistency than the texts actually have to offer (they are, after all,
student lecture notes), Emanuel Hirsch’s reading may do better justice than the texts
themselves to Schleiermacher’s aims. See his Geschichte der neuern evangelischen Theologie, vol.
v (Gütersloh: Bertelsman, 1954), pp. 37–8.
Trinity
14
which does not fit with his epistemic priorities and an implausible one
which does.
To take “dead” here to mean “no longer alive” is to take it in the
usual sense, and so to take the New Testament’s affirmation that Jesus
is risen along the lines suggested in (1)–(3) above. But it is hard to see
how this interpretation could square with Schleiermacher’s epistemic
commitments. In that case to hold true “Jesus is risen” allows for an
event in time which at best fits very poorly with the expectations of
natural and historical science as Schleiermacher understands them.
So this interpretation is quite plausible, but given his epistemic
commitments should lead Schleiermacher to hold “Jesus is risen” false.
To take “dead” to mean “asleep” or “catatonic” is, by contrast, much
more congenial to Schleiermacher’s epistemic commitments, since it
allows him to understand Jesus’ life after this “death” to be entirely
natural and continuous with that before it, lacking any undesirable
miraculous or supernatural element. But this is a wholly unsatisfactory
interpretation of “dead”: were we to take speakers this way when we
find them holding true “x is dead” we would have to find them
massively false; we should therefore not interpret the troublesome
term this way.36
When it comes to “Jesus is risen,” therefore, Schleiermacher seems
unable to link up his epistemic commitments, the truth value he
assigns to the sentence, and a plausible interpretation of it. If he holds
to his epistemic priorities and remains committed to regarding the
sentence as true, the result, while it cleaves to the manifest logical form
of the sentence, is an astonishingly forced interpretation of the
predicate. A workable theory of interpretation requires us to adjust
the meanings we assign to a speaker’s utterances in order to find her
right most of the time. This sometimes requires us to assign meanings
to the speaker’s words which make particular sentences the speaker
holds true come out false, as the price of assigning meanings to those
same words which allow us to find her right most of the time. If this
Davidsonian suggestion is basically correct, then given
Schleiermacher’s epistemic priorities, the interpre- tively efficient
adjustment would simply be to find “Jesus is risen” false,
36. Schleiermacher’s willingness to forgo interpretive plausibility here for the sake of his
epistemic commitments is, as Hans Frei observes, startling. “Far more remarkable [than
Schleiermacher’s skepticism about physical miracles] is the fact that, no matter what he
may have chosen to believe about the facts of the case, it never occurred to him that there
is something unfitting, indeed ludicrous, about rendering the story of Jesus in a way that
makes such a thundering anticlimax possible.” The Eclipse of Biblical Narrative (New Haven:
Trinity and
1
Yale University Press, 1974), p. 313.
Trinity
14
rather than hold it true at the cost of systematic interpretive
implausibility. The apparent alternative would be to adjust one’s
epistemic priorities
– to seek a way of construing one’s relevant beliefs about nature and
history, and the inferences which must, may, and may not be made
from them, which is at least consistent with “Jesus is risen,” plausibly
interpreted.
Together with almost all modern readers of the resurrection narra-
tives who are committed to the dependence thesis, Bultmann disagrees
with Schleiermacher about where the main interpretive difficulty lies.
While Schleiermacher never quite proposes any definite interpretation
of (1), Bultmann takes it for granted that the New Testament asserts
that Jesus died on the cross, and he interprets “died” in the usual
sense. The problem becomes what to make of (2) – of the sense in
which Jesus “lives” after his genuine death.
Were we to interpret the resurrection stories according to their
mythological surface meaning, Bultmann assumes, we would surely
have to regard them as false.37 But this does not mean they actually
are false; Bultmann is utterly committed to the truth of the New
Testament proclamation of Jesus’ cross and resurrection. Instead it
means that we have to interpret the narratives differently, so that they
are consistent with, and thus can come out true by, the same standards
which would otherwise make them come out false – we have to
“demythologize” them. Like Schleiermacher, therefore, he proposes his
account of “Jesus is risen” as an interpretation of what the New
Testament actually says, of the sentences to be found there, and not
as an alternative to sentences which he thinks he has to hold false.
In some famous passages, Bultmann argues that to believe in the
resurrection of Jesus just is to believe in the saving significance of his
death, to grasp the cross as God’s definitive promise of salvation or
authentic existence.38 Jesus’ resurrection is not, by contrast, to be
thought of as a spatio-temporal event which happens to a dead
person.39 As Bultmann
37. In the well-known throw-away line: “One cannot use the electric light and the radio . .
. and at the same time believe in the New Testament world of spirits and miracles.”
“Neues Testament und Mythologie,” Kerygma und Mythos, vol. I, ed. Hans Werner Bartsch
(Hamburg: Evangelischer Verlag, 1948), p. 18; on the resurrection specifically, see p. 21 (ET:
“New Testament and Mythology,” Kerygma and Myth: A Theological Debate [New York:
Harper, 1961], pp. 5, 8).
38. “Faith in the resurrection is nothing other than faith in the cross as the saving event, in the
cross as the cross of Christ” (“Neues Testament und Mythologie,” p. 50; the emphasis is
Bultmann’s [ET, p. 41]).
39. In the New Testament, “next to the historical event of the cross stands the resurrection,
Trinity and
1
which is no historical event.” So “it is not as though the cross could be seen as simply the
Trinity
15
sees it, the problem with believing in the occurrence of such an event is
not simply that it runs counter to our most basic assumptions about
what can happen in nature and history. We should not think of
locating the resurrection in the realm of “objective” happenings to
begin with; unlike the cross it is not datable, even in principle.40 To
take Jesus’ resurrection as a spatio-temporal event would require
numbering “Jesus is risen” among our historical and natural-scientific
beliefs, where we are entitled to hold only those sentences true for
which proof can be offered. In the case of the belief that Jesus is risen
no such proof is available. If we want to maintain these epistemic
priorities and still hold “Jesus is risen” true, we will have to find an
interpretation according to which “Jesus is risen” does not assert that
“rises” – that is, “lives” – is a spatio-temporal event which happens to
the dead Jesus.41
Bultmann develops this interpretation by elaborating the
suggestion that believing in Jesus’ resurrection is the same thing as
believing in the saving significance of his cross. To the historical event
of the cross is “added” the present proclamation of the significance of
that event as
death and disappearance of Jesus, upon which the resurrection followed, reversing the
death” (“Neues Testament und Mythologie,” pp. 44, 48; ET, pp. 34, 38).
40. See Rudolf Bultmann, “Zum Problem der Entmythologisierung,” Kerygma und
Mythos, vol. II, ed. Hans Werner Bartsch (Hamburg: Evangelischer Verlag, 1952), p. 206
(partial ET: “Bultmann Replies to his Critics,” Kerygma and Myth; here, p. 209).
41. “Certainly,” Bultmann emphasizes, “faith’s relation to its object cannot be proven. That
faith cannot be proven is, however, precisely its strength, as Wilhelm Herrmann taught.
To assert that faith could be proven would be to assert that God can be known and
ascertained outside of faith, and consequently to put God at the level of the world, which
is always at hand and subject to objectification. Here the demand for proof is indeed
appropriate” (“Zum Problem der Entmythologisierung,” pp. 199–200; see p. 207 [ET, p.
201]).
Here Bultmann evidently follows the strategy, widespread among dependence
theorists, of treating “history” and “faith” as wholly discrete (though, so the argument
usually goes, not conflicting) epistemic realms, correlated with quite different sorts of
things (“the objective world,” “God”). How, though, do we decide which sentences to put
into which box? If sentences having to do with “history” must be susceptible of proof, we
will have to assign countless sentences about the past, and not simply those having to do
with God or divine action, to that realm, outside “the world subject to objectification,”
which is for Bultmann the homeland of faith. “Napoleon had a red handkerchief in his
pocket at Waterloo” is no more susceptible of historical proof than “Jesus is risen”; short
of the eschaton we will very likely lack evidence which would conclusively establish
either one as true. If states of affairs with regard to which the only beliefs we can have are
unprovable must be moved outside the objective world, then what Napoleon had in his
pocket when Marshal Ney charged Wellington’s line is no less outside the sphere of
objective reality than Jesus’ resurrection – and so, presumably, is no less the object of
saving faith. One might reply that “Jesus is risen” belongs to “faith,” while “Napoleon
had a red handkerchief in his pocket at Waterloo” belongs to “history,” because the
former has to do with God and salvation, while the latter does not. But this seems wholly
arbitrary; why we need to assign sentences about God and sentences about the past to two
discrete epistemic realms in the first place is just the question at issue. If not being subject
Trinity and
1
to proof fails to move the contents of Napoleon’s pockets outside the objective world,
then there is apparently no cause to suppose that it moves Jesus’ resurrection there either.
Trinity
15
God’s ever new offer of salvation. To proclaim Jesus’ resurrection, it
seems, just is to proclaim his cross as saving, even as to believe in the
resurrection just is to believe in the cross, preached. Jesus’
resurrection thus appears to be identical with the event of its
proclamation, and faith in the resurrection with faith in that
proclamation.42 There is life after the cross, to be sure, but what lives
seems not to be the crucified Jesus. What “lives” are the words which
proclaim his cross as the event of salvation, and the faith of those who
find their lives transformed by this proclamation. Jesus dies a human
being, but what rises are words and faith; while the cross is clearly an
event which happens to Jesus, the resurrection is an event which
happens not to him, but to us.
Or so it often seems. Like Schleiermacher, though, Bultmann
wavers between an implausible interpretation of “Jesus is risen” which
squares with his epistemic priorities and a more plausible and
standard one which appears inconsistent with those priorities, and
with what he says about the resurrection when he adverts to them.43
Here too interpretive irresolution results from an effort to hold the
gospel proclamation of the resurrection true in the face of apparently
conflicting beliefs, without regarding it as epistemically primary over
against those beliefs (that is, without having to give up beliefs with
which the gospel proclamation apparently conflicts). This allows him
either his epistemic priorities or a plausible interpretation of the
resurrection texts, but not both. As with Schleiermacher, we need not
settle which interpretation he finally decides on; indeed there may be no
way to settle it.44
42. “It is the word [of reconciliation] which is ‘added’ to the cross and, by demanding
faith, makes the cross comprehensible as the saving event” (“Neues Testament und
Mythologie,” p. 51; ET, p. 42).
43. Side by side with the remarks just cited, Bultmann insists that it is “Christ the crucified
and risen [who] encounters us in the word of proclamation”; this suggests that Jesus’
resurrection is the content (and not simply the event) of the proclamation, and indeed
that in hearing the proclamation which has this content we encounter the risen – living –
Jesus himself, the very one who died on the cross (“Neues Testament und Mythologie,” p.
50, my emphasis; ET, p. 41).
44. Not unreasonably taking his cues from Bultmann’s explicit epistemic priorities
(though he does not call them that), Karl Barth reads Bultmann as settling for a
thoroughly forced reading of the New Testament resurrection texts, and poses some
trenchant questions about the possibility, necessity, and desirability of such an
interpretation (see Kirchliche Dogmatik III/2, pp. 531–7; 541–5 [ET, pp. 442–7, 451–4], also
Rudolf Bultmann: ein Versuch, ihn zu verstehen [Theologische Studien 34, 2nd edn, Zurich:
Evangelischer Verlag, 1953]). Arguing inter alia against readings of Bultmann influenced
by Barth, James F. Kay plays up the passages (like those cited in the previous note) which
suggest Bultmann’s commitment to a more plausible interpretation (see Christus
Praesens: A Reconsideration of Rudolf Bultmann’s Christology [Grand Rapids: Wm. B.
Eerdmans, 1994]). Both ways of reading Bultmann probably seek more consistency than
Trinity and
1
the texts actually exhibit (which is not to say they are both equally persuasive); the reason
each side can
Trinity
15
The problem with an approach like Bultmann’s lies not in the inter-
pretation of (1) “was dead,” nor even primarily of (2) “now lives,” but
in the way he handles (3) – the logical form by which the New
Testament attributes both (1) and (2) to “Jesus.” Whereas
Schleiermacher took the logical form of “Jesus is risen” in a plausible
way but leaned toward a highly unlikely interpretation of the
predicate, Bultmann makes some- what better sense of the predicate at
the cost of losing track of the logical form of the sentence. One way to
get at the difficulty which arises here is to ask whether Bultmann’s
interpretation of “Jesus is risen” succeeds in fixing the referent of
“Jesus” in a consistent way.
He clearly has no trouble agreeing to “Jesus died on the cross,”
which implies that “Jesus” refers to “the person who died on the
cross” (this slightly vague formulation could easily be indexed in order
to guarantee uniqueness of reference). He also wants to hold true
“Jesus is risen”; let us say this means that he in fact wants to hold true
“Jesus now lives.” Since “Jesus” refers to “the person who died on the
cross,” assent to these last two sentences requires assent to “The
person who died on the cross is risen” and “The person who died on
the cross now lives”; save in oblique contexts, singular terms refer to
the same thing just in case we can substi- tute one for the other salva
veritate.
But now, it seems, Bultmann clearly wants to demur. To hold
true “The person who died on the cross is risen” and “The person who
died on the cross now lives” is to say that resurrection is an event
which happens to a dead person. This our epistemic priorities require
us to deny. But Bultmann, committed to avoiding any interpretive
outcome on which “Jesus is risen” ends up false, seems to opt for
saying that “Jesus is risen” and (let us suppose) “Jesus now lives” are
indeed true, but “Jesus” in these sentences does not refer to the
person who died on the cross; it refers rather to the proclamation of
his cross as saving, or to faith in that proclamation. So (1) “Jesus was
dead” and (2) “Jesus now lives” both come out true, but at the cost of
denying (3) outright: “Jesus” does not refer to the same thing in both
sentences. This in turn calls for an extended or meta- phorical
interpretation of the predicate in (2), one which captures the sense
in which “lives” may rightly be applied to the experience of faith in the
proclamation of Jesus’ death or (still more extended) to the words of
the proclamation themselves, as distinguished from the sense in which
it
Trinity and
1
always find ammunition against the other is that the conflict between Bultmann’s
epistemic priorities and his truth commitments bars him from coming up with a
consistent interpretation in the first place.
Trinity
15
applies to persons. Thus by varying the referent of “Jesus” (by denying
[3]), Bultmann ends up taking “Jesus is risen” to mean something like
“Because Jesus died on the cross, we can now experience authentic
existence.”
Fixing the reference of singular terms plays a basic role in the inter-
pretation of any language, since without it we will have scarcely any
idea of what the speakers of the language are talking about. The
complexities of reference need not concern us here, save to observe
that here the Davidsonian (and Quinian) principle of charity has
important work to do: if we fix the reference of a singular term in such
a way that we have to count as false most of what our speakers assert
about it, the great likelihood is that we have the reference wrong.
Bultmann wants to count the “speakers” of the New Testament right
when they assert “is risen” of “Jesus,” but at the dimly acknowledged
cost – given his epistemic priorities – of counting them wrong in the
rest of the assertions they make about him. Obviously words and faith
did not die on the cross, but a person did; if “Jesus” refers to words
and faith, as it often seems to do in Bultmann’s interpretation of “Jesus
is risen,” then “Jesus died on the cross” will be false, as will “Jesus was
born around 4 bc,” “Jesus was a Jew,” and so forth (or perhaps these
and similar sentences will have to be counted not so much as false but
as nonsensical – an equally good sign that interpretation has taken a
wrong turn, since the point of interpretation is presumably to make
sense of speakers’ utterances). Better to avoid such implausible
interpretive desperation, take “Jesus” to refer consis- tently to “the
person who died on the cross” (as Bultmann himself obviously wants
to do, most of the time), and if our epistemic priorities forbid us to
assert “now lives” of “the one who died on the cross,” regard the New
Testament speakers as simply mistaken in their belief that Jesus is
risen. The alternative, once again, is to adjust our epistemic priorities –
to reas- sess our assumption that tenable beliefs about nature and
history forbid us to believe that the same person who died on the cross
now lives.
The point of these examples is not to single out Schleiermacher
and Bultmann as exceptionally misguided interpreters of the New
Testament’s talk of Jesus’ resurrection. On the contrary: within the
epistemic and interpretive boundaries they broadly share, each offers
an especially subtle and complex account of “Jesus is risen.” The
different interpretive difficulties in which each tends to get entangled
when guided by his epistemic priorities illustrate instead the inherent
Trinity and
1
problems of the project they undertake.
To be committed to the truth of a sentence is to be committed to the
Trinity
15
falsity of those sentences which, given the best interpretation we
can provide, are inconsistent with it. The intensity of our commitment
to different sentences does of course vary, and with that the readiness
with which we are prepared to reject other sentences should they come
into conflict with those to which we are committed. When sentences to
which we are committed seem to come into conflict we can try to show
that the conflict is apparent rather than real, or we can try to adjust the
interpretation of some of the sentences to make them consistent with
the others – thereby showing that we regard the adjusted sentences as
epistemically subordinate to those whose interpretation remains fixed
(and with that our more intense commitment to the truth of the
sentences whose meaning we decline to adjust). But there are limits
on the extent of the interpretive adjustments we can plausibly make.
Those limits are reached when making a given adjustment in meaning
would require us to attribute noticeably more falsity to sentences held
true by speakers of the interpreted language than would result from
simply holding false the sentences we are trying to adjust.
If these principles are roughly right, then our epistemic priorities
will on the whole have to match our truth commitments. The modern
theological dependence thesis, we can now observe, is basically an
effort to arrange a divorce between the two: to hold at least some
centrally Christian beliefs true in virtually all conceivable
circumstances, while not being committed to rejecting virtually any
conceivable belief which cannot be reconciled with those central
Christian claims. For these truth commitments, the epistemic priorities
lie elsewhere. In this effort to sunder the highest epistemic priorities
from the deepest truth commitments something has to give, and the
only thing left to give is plausible interpretation. This is not an all-or-
nothing proposition, but admits of degree. The stronger our
commitment to a given belief, the more likely we will interpret it badly
if we fail to accord it an equivalent degree of epistemic primacy. Ifwe
are not willing to give up central Christian beliefs, and with them
Christian identity, the only way to guarantee that we will interpret
them plausibly is to treat them as epistemically primary without
restriction. And correlatively: if we want to interpret central Christian
beliefs plausibly, the only waytoensure that wewill not have to give
them up in the process is to treat them as epistemically primary across
the board.
What to correlate?
Trinity and
1
Ascribing unrestricted epistemic primacy to the most central Christian
beliefs might seem like a deliberately anti-correlationist position. But it
Trinity
16
would be misleading to put the objection to the dependence thesis here
developed in this way. An anti-correlationist view would presumably be
one which was concerned to deny something that correlationists want
to affirm. But the foregoing considerations regarding how we may
viably adjust the assignment of meaning, truth value, and epistemic
status suggest that epistemic correlation in theology is not so much
mistaken as empty. In order to locate the entities which standard
accounts would have us correlate, we must already have excluded the
possibility of the sort of epistemic correlation we are invited to
undertake. Should the theological notion of epistemic correlation thus
fail to state anything which could coherently be affirmed, there is no
point in denying it; it needs not refutation, but elimination.45
The enterprise of epistemic correlation in theology seems to be that
of distinguishing Christian belief from other regions of human
discourse, and then comparing Christian talk with these other types of
belief in order to see which beliefs in the various regions come out
true. Correlationists mark off Christian belief from other areas of
discourse in various ways, sometimes by distinguishing the presumed
sources of beliefs (such as “revelation” and “reason,” or as we saw
with Schleiermacher in chapter 3, two different sorts of consciousness),
at other times by adverting to what the beliefs are about (such as, in
Bultmann’s case, an objective world of spatio-temporal events and the
world of human subjectivity).46
In order to identify beliefs as “religious,” “historical,” “scientific,”
or the like – in order to assign them their proper region – we have to
know what the sentences mean which state the contents of the beliefs.
If the relations between meaning, truth, and belief here outlined
basically hold, we will on the whole not be able to figure out the
meanings of sentences, and a fortiori the meanings of the words of
which sentences are composed, except by assigning truth values to
them which optimize
45. We are here concerned only with epistemic “correlation” aimed at judgments about
the truth of Christian beliefs. This often goes together with an enterprise from which
it is nonetheless distinct, namely a correlation which aims to establish the existential
meaningfulness or relevance of Christian beliefs. See my Christology in Conflict for an
argument against the latter sort of correlation, at least in its christological applications.
46. On “reason” and “revelation” as terms of epistemic correlation, see David Tracy’s
“Foreword” to Jean-Luc Marion, God Without Being, trans. Thomas A. Carlson (Chicago:
University of Chicago Press, 1991). Tracy there rightly observes that non-correlational
theologies are not for the most part those which unaccountably refuse to correlate terms
like “reason” and “revelation,” but which regard the very idea of epistemic correlation
as “at best a category mistake” (p. x). The present remarks are an attempt to locate the
mistake.
Trinity and
1
agreement about which sentences are true with speakers of the
language in which the beliefs are stated. In order to assign beliefs to
various regions, therefore, we must already on the whole have decided
about their truth. And this leaves us nothing to correlate: no large
regions of beliefs whose meaning we know, but whose truth value we
need to find out, and might hope to discover by relating the regions to
one another in some way. If we do not know the meanings of sentences
believed, we have no bodies of belief to correlate, and if we have fixed
sentence meanings we must have in large part already assigned truth
values to sentences believed; there is no epistemic work left for
correlation to do.
There will, of course, always be some leftovers the meaning of
which we suppose we have fixed, but about the truth value of which
we are unsure. When this happens we will not always have any clear
way to decide whether to assign one meaning to a sentence and agree
with our speakers in holding it true, or to assign a different meaning
and hold it false. In the nature of the case this interpretive margin for
error will have to be the exception rather than the rule. To the extent
that we can settle such cases at all, they will presumably be decided by
the most relevant beliefs whose truth value we know – which is to say,
if we are thinking in terms of “regions” of belief, those to whose region
the leftovers belong, rather those which make up some other area of
discourse.
Appeals to epistemic correlation in modern theology have normally
been made in defense of epistemic priorities suggested by the depen-
dence thesis; the point has been to show that beliefs which belong to
the Christian “region” are at least consistent with, and perhaps
warranted by, the contents of other areas of belief. But the problem
with such appeals has nothing specifically to do with the epistemic
status of beliefs. The very idea of correlation turns out to be empty. It
would make no more sense to try to correlate “reason” and
“revelation” if revelation were epistemically primary, or if the
correlation were “mutually critical,” than it does when reason or
Wissenschaft is epistemically primary. Insofar as the dependence thesis
is at bottom the effort to divorce truth commitments from epistemic
priorities in theology (or, one could say, to divorce central Christian
beliefs from their properly primary epistemic function), one can be a
dependence theorist without correlation, that is, without appealing to
areas of belief that can be compared in the hope of favorable epis-
temic result. Yet it is not hard to see why the dependence thesis has so
Trinity
16
often taken an essentially correlationist form. Correlation holds out the
promise that theologians can do justice both to the contents of
Christian
Trinity and
1
beliefs and to the epistemic priorities of modernity while keeping each
in its place. Like the dependence thesis generally, correlation makes a
promise which in the end it lacks the logical resources to keep.
Philosophical considerations thus support the conclusion we drew
from an analysis of the content of the church’s trinitarian identification
of God: if central Christian beliefs are to be held true at all, they must
have unrestricted epistemic primacy. A Christian view of things will
be one which orders the whole open field of possible sentences or
beliefs so as to achieve at least consistency with a christological and
trinitarian center; in this way it will match its epistemic priorities to its
truth commitments, and so hope to come up with a plausible
interpretation of the whole field. These philosophical considerations
also support the conclusion of our empirical analysis of Christian
practice. Since we can only pick out central Christian beliefs in the first
place by reference to their function across the whole field – this
community’s publicly observable commitment to retaining these
beliefs in all possible circumstances, which displays its conviction
that they are epistemic trump – there can be no beliefs external to a
Christian view of the world, nothing outside it with which it might
conceivably be correlated. It is, of course, possible to order the whole
field differently than the Christian community does, around different
epistemic priorities. But that is another matter, and the subject of the
next chapter.
6
Epistemic priorities and alien claims
Fideism and epistemic priorities 141
The suggestion that the church’s trinitarian identification of God

enjoys unrestricted epistemic primacy may prompt a certain unease.
To suppose that the Christian community is rationally justified – let
alone correct – in rejecting as false any statement which is inconsistent
with its own most central beliefs seems to have the whiff of fideism
about it. Surely no one wants to be a fideist, so we need to see whether
this oppro- brious term fits the epistemic outlook so far proposed.
The charge of “fideism” labels the following contention. Far from
answering the question of the church’s right to hold its central beliefs,
the argument so far has not yet succeeded in addressing it. To take
the church’s central beliefs as epistemic trump is to help oneself to a
whole range of trinitarian and christological beliefs, simply
assuming that these beliefs are true rather than producing any
argument for them. But surely the question which most needs to be
answered is why we should think these Christian convictions are true
in the first place; the crucial question is what grounds, if any, there are
for holding the central beliefs themselves true. By failing even to
address, let alone answer, that question, we have taken a wholly
fideistic position, the chief aim of which, so an objector might
conclude, is to insulate Christian belief from criticism. This objection
exploits our deep conviction – surely correct and important, as far as it
goes – that in order to hold a belief in an epistemically responsible
way (to hold it rationally, rather than fideistically) we must be able to
offer reasons for the belief. We do this in all sorts of ways, but in giving
reasons we naturally appeal to beliefs which we and (we hope) our
interlocutors already hold true. Reasons may in turn be
Trinity and
1
requested and given for the beliefs to which we appeal, and so forth.
Sometimes the charge of fideism seems simply to be that it is irrational
to hold beliefs for which no reasons or “grounds” are given.
But surely this cannot be right. The giving of reasons has to come to
an end.1 Beliefs succeed in terminating justificatory arguments just in
case we do not have to give reasons for them. Without a terminus to
the regress of reasons we will lack a way to decide whether the reasons
we have given are good ones, that is, whether they actually support the
belief which prompted our search for reasons in the first place. This
sort of decision has to rely, it appears, on a nexus of belief which can
serve as a reason, but does not itself require the giving of reasons.
Eventually the practice of justifying beliefs will have to appeal to
beliefs which we and (again we hope) our interlocutors hold true, but
for which no reasons are given – beliefs which, so far as we can tell,
settle the matter with which we began. So holding beliefs rationally
requires not only the ability to give reasons, but the ability to
distinguish between those occasions when we need to give reasons,
and those when we do not.
If we had to give a reason for everything, then whatever beliefs,
Christian or otherwise, were invoked to justify other beliefs would be
held irrationally, since beliefs terminate arguments when no reasons
need to be offered for them. Since in that case it would be
unreasonable to hold these beliefs, it would also be unreasonable to
hold the beliefs which they are supposed to justify; reasonable beliefs
presumably cannot be justified by unreasonable ones. So if it were
fideistic to hold beliefs without giving reasons for them, then all of our
beliefs would be irrational, both those which we hope to justify by
giving reasons, and those which are supposed to do the justifying.
Were this the case, Christian beliefs obviously could not be faulted by
comparison with others for their lack of rationality. Understood in this
way, the charge of fideism seems self- refuting.
To say that the Christian community’s central beliefs are
epistemically primary without restriction does not, of course, mean
that these beliefs must terminate all justificatory arguments. Naturally
they will conclude only a small portion of such arguments, namely
those where the matter in dispute – what God is like, for example, or
what human beings may
1. This way of putting the point is Wittgenstein’s; see On Certainty, ed. G. E. M.

Anscombe and G. H. von Wright (Oxford: Basil Blackwell, 1969), e.g., §§ 110, 192. But the
thought goes back at least as far as Aristotle, e.g., Posterior Analytics I, 3 (72b, 5–24);
Epistemic priorities and alien
1
Metaphysics IV, 4; 7 (1006a, 4–12; 1012a, 21–2).
Trinity and
1
hope for – is of a kind which makes appeal to these beliefs relevant to
settling the dispute (cases, in other words, where the meaning of the
disputed beliefs places them in logical space close to the narratives
which identify Jesus and the triune God). Taking these beliefs as
epistemically primary across the board does mean, however, that in
general justificatory arguments cannot be accepted as yielding true
statements when those arguments terminate in beliefs incompatible
with the church’s canonical narrative. If that narrative has epistemic
primacy we cannot for the most part suppose, in other words, that we
may regard statements as true whose supporting arguments depend
on appeal to convictions incompatible with those narratives. Thus if
the narratives which identify Jesus are epistemically primary, they are
the final and decisive, though of course not the sole, criteria for
deciding what count as good reasons for holding sentences true.2
So far, then, it seems epistemically permissible to hold central
Christian beliefs true without giving reasons for them. Indeed if the
church’s central beliefs do have epistemic primacy, there will in the
nature of the case be no way to show that these beliefs are true without
appealing to these same beliefs as decisive tests of truth, no non-
circular (and therefore epistemically trivial) way of offering reasons for
them. That is part of what it means to say they are epistemically
primary.
In response the charge of fideism might be more narrowly confined.
It is not that all beliefs require reasons in order to be held rationally,
but rather that these particular beliefs, the centrally Christian ones,
need justifying grounds, and cannot themselves suitably terminate
justificatory arguments.
Were foundationalism tenable this version of the objection might
have considerable force. In that case, beliefs would have to meet
certain stringent requirements in order to terminate justificatory
arguments in a rational way. Such beliefs would have to be self-
evident, empirically evident, givens of immediate self-consciousness,
or something of the kind. Indeed much of the appeal of
foundationalism lay in the sense that without recourse to beliefs
having certain marks which place them
2. The qualifications “in general” and “for the most part” are important here. In order to
figure out whether a given claim is genuinely incompatible with central Christian beliefs
it is necessary to establish strictly inferential connections between the claim in question
and the negations of Christian beliefs. But many sorts of justificatory arguments yield
their conclusions in less stringent ways, and in such cases a claim supported by reasons
themselves genuinely opposed to the church’s central beliefs might nonetheless be
1
compatible with these beliefs.
Trinity and
1
beyond doubt for any rational person, there would be no adequate way
to terminate justificatory arguments. This attraction has not been lost
on theologians, who have tried to meet the requirements it imposes by
packing their epistemic foundations with theological content, or
showing that Christian beliefs were at least plausible, and perhaps
more robustly warranted, by appeal to the supposed foundations.
As we have already seen, however, epistemic foundationalism
seems untenable. Even if there are beliefs of the needed types (which is
often disputed), they cannot plausibly be invoked for foundationalist
purposes. Apart from foundationalist assumptions, it is not clear that
central Christian convictions lack any marks (like empirical evidence or
interior immediacy) which beliefs need to have in order to be
epistemically primary, and so a fortiori to terminate justificatory
arguments. Here too, the fideistic charge seems to require a distinction
between rational and irrational ways of holding beliefs which turns
out to be impossible to sustain.
Theologians have, however, often drawn a rather different lesson
from the widely acknowledged demise of foundationalism. In order to
avoid fideism in a post-foundationalist world, so the argument some-
times goes, we have to grant that all of our established beliefs, and
with them all of our criteria of truth, are at hazard whenever we try to
make decisions about what is true. The very idea that there could be
perma- nently fixed epistemic priorities, let alone priorities which
applied without restriction across the field of possible beliefs, is for this
reason inherently fideistic – that is, irrational.
In this strong sense the further reformulated charge of fideism
seems no more plausible than the earlier variants. It proposes a
standard for rational (that is, non-fideistic) conversation which seems
impossible to meet, in theology or anywhere else. It seems impossible
that we could doubt all of our beliefs at once, or even be prepared to
doubt them all. Perhaps the most familiar arguments against this
possibility are those proposed by Wittgenstein: doubt is logically
possible only against a background of beliefs held true, since doubt
(or preparedness to doubt) requires reasons for doubting (or being
prepared to), and giving reasons requires appeal to beliefs held true
(that is, not doubted).3 If foundationalism is untenable, and with it
the necessity (and perhaps the possibility) of terminating justificatory
arguments by appeal to putatively self-
3. See again Wittgenstein, On Certainty.

1
evident or self-justifying beliefs, then it becomes especially important
to recognize that rational conversation and argument do not require,
but rather preclude, holding all of our beliefs (including our criteria of
truth) open to doubt at the same time.
Theologians sometimes compound the problem of inflated expecta-
tions about what may reasonably be doubted by basing the charge of
fideism on a curious sort of self-loathing foundationalism. Such an
outlook takes for granted the dualism of conceptual scheme and
experi- enced content, and so is structurally foundationalist, but
despairs of ever escaping the scheme so as to find the unblemished
content – a kind of foundationalism which has given up on locating the
foundations. On this view we will never be able to discover which
conceptual scheme, if any, tells us the truth about the world and our
experience of it, and for this reason we must regard all of our beliefs as
up for grabs. But rather than trying to doubt all of our beliefs, the more
plausible course would be, as chapter 4 argued, to jettison the dualism
of scheme and content which tempts us to suppose that we could make
sense of such doubt in the first place.4
So far it seems difficult to state the worry about “fideism” in a
way which mounts a plausible objection to the unrestricted epistemic
primacy of central Christian beliefs. Yet there is something to the
worry, even if standard attempts to formulate it as an argument are
unsuccess- ful. The collapse of foundationalism surely does not mean
that we may believe whatever we like, nor does it mean that we may
choose our epistemic priorities at will. Holding beliefs in an
epistemically responsible way does not require that we be able to give
reasons for every one of our beliefs. But our epistemic responsibilities
do seem to require that we be open to changing any one of our beliefs if
given sufficient reasons to do so. From the possibility of doubting any
particular belief we cannot argue to the possibility of doubting all our
beliefs, but the impossibility of doubting all our beliefs does not by
itself entitle us to cling to any particular belief.
An account of the justification of beliefs, Christian or otherwise,
therefore ought to include an explanation of the kinds of reasoning (as
distinguished from the social or psychological motives) which might
4. For a theological example of the scheme–content distinction appropriated wholesale,
see Gordon D. Kaufman, In the Face of Mystery (Cambridge, Mass.: Harvard University
Press, 1993); on the putative need to play off language as a human contrivance against the
ascription of truth to sentences, and so to regard the truth of our sentences as a question to
which we can never hope to give an answer, see pp. 49, 347.
Trinity and
1
lead us to change our established beliefs, including our epistemic
priorities. Explaining what might give us reason to change our
epistemic priorities will very likely involve saying what gives us
reason to keep them, so we could equally well say that we need an
account of how we may justify our epistemic commitments – not
simply our beliefs, but the priorities which guide our decisions about
the truth of beliefs.
Since we have already offered an account of how the Christian
community may justify its distinctive epistemic priorities, it may
seem odd that this issue should now recur; the logic of the argument
so far seems to suggest that this question has already received, for
good or ill, the most complete answer we can give to it. Were that so,
however, we would be left with a kind of standoff between the charge
of fideism and the epistemic primacy of the Trinity. The objector to
fideism has raised what seems like a reasonable question: why
suppose that the beliefs which make up the church’s identification of
God are true in the first place? Though the objection has not succeeded
in showing that the epistemic primacy of the Trinity is irrational, the
objector still wants reasons for believing in the Trinity.
It might be replied that this demand is simply confused, and
requires not to be met, but to be eliminated. The very idea of epistemic
primacy is that for beliefs which have this property without restriction,
there can be no further beliefs to which one might appeal in order to
decide about their truth – that is, no beliefs of just that sort which the
objector seems to think he needs in order to believe in the Trinity. But
while right as far as it goes, this leaves us without a satisfying response
to the objector worried about fideism. If we accept the objector’s
gambit we fall into the confusion of treating beliefs which (if held
true) have to be epistemically primary as though they were
epistemically subordinate; if we refuse the gambit we also decline to
offer any reasons for the priorities by which we decide about truth
which might have some epistemic purchase with those who do not
already share them.
What we need is a way of giving reasons for beliefs without
creating epistemic subordination and dependence. This would meet
the anti- fideist’s legitimate concern that we ought to be able to say
something about the justification of our epistemic priorities
themselves, without letting the objector’s demand for reasons tempt us
to lose track of those priorities. Identifying reasons of a kind which
involves no epistemic dependence would permit a genuinely
1
justificatory conversation about epistemic priorities between adherents
of rival belief systems, individu-
Trinity and
1
ated by different priorities. Armed with such reasons, the unrestricted
epistemic primacy of central Christian beliefs would be entirely
compatible with at least some of what has traditionally been called
“apologetics.”
Inclusion and assimilation
The success of the Christian community’s belief system at coping with

alien and novel beliefs – its inclusive and assimilative power – might
count as a reason for having this community’s epistemic priorities
which nonetheless brings about no epistemic subordination of the
beliefs for which reasons are given.5
By “alien” beliefs I mean those within the open field of possible
belief which are not central to the belief system of the Christian
community.6 An alien belief is therefore one which a community may
hold false without loss to its own identity. At any given time the
church’s belief system as a whole will likely contain countless alien
beliefs. The capacity of the church’s view of the world to permit the
acceptance of alien beliefs when its holders meet with good reasons to
hold them true is its inclusive power; possession of this capacity tends to
justify, I will argue, epistemic priorities and belief systems which have
it.
Alien beliefs may be regarded as occupying a position in the logical
space of the Christian community’s belief system ranging from medial
to peripheral. For a claim to be alien is not necessarily for it to be
rejected by the community; beliefs are “alien” only in not being central.
Nor are alien beliefs those external to or outside of the church’s belief
system. If it makes sense to talk of beliefs “external to” a
comprehensive belief system like that of the church, external beliefs
will presumably be those which a
5. For a suggestion that a notion of “assimilative power” might have this sort of
epistemic function, see George A. Lindbeck, The Nature of Doctrine (Philadelphia:
Westminster, 1984),
p. 131. For a critical analysis of Lindbeck’s use of this notion, see my essay, “Absorbing the
World: Christianity and the Universe of Truths,” Theology and Dialogue, ed. Bruce D.
Marshall ( Notre Dame: University of Notre Dame Press, 1990), pp. 69–102. I now think
this essay goes wrong in supposing that much sense is to be made of the notion that there
are true beliefs “external” to a Christian view of the world, though I still think that its
estimation of Lindbeck’s views is basically correct.
6. The term is William Christian’s, and I am applying it in his sense: “A claim that what
is proposed in some assertion is true, or that some course of action is right, is an alien
claim, with respect to some community, if and only if what is proposed in the claim is not
an authentic doctrine of that community,” where an “authentic doctrine” of a community
is one which that community holds it is “bound to teach” – that is, what are here
1
conceived as a community’s “central” beliefs. The notion of “alien” claims is thus
parasitic upon that of claims central to a communal belief system. Doctrines of Religious
Communities: A Philosophical Study (New Haven: Yale University Press, 1987), pp. 145, 74.
Trinity and
1
community’s epistemic priorities require it to reject, whereas there will
be numerous alien beliefs the community will want to include. As the
church’s central, identity-constituting beliefs will tend to be distinctive
to it rather than shared with other communities, so also the church’s
alien beliefs will tend to be those which it shares with other
communities. The correlation, though, is not absolute; for example
some convictions near the center of the church’s belief system are
also central for the Jewish people, and there might be beliefs which
are distinctive to the church, but which are not central, and so are
alien.7
Of particular interest for present purposes are beliefs which are not
only alien for the church, but novel: claims which the Christian commu-
nity and its members have encountered as live options for belief, but
about whose epistemic status the community has on the whole not yet
come to a decision. Novel claims are “live” options for belief when (a) they
come with reasons attached which make the claims in some degree
persuasive, and (b) they impinge closely enough on the Christian
community’s identity-forming belief and practice to make deciding
about their truth worth the trouble for the community or its members.
For the Christian community to assimilate these novel beliefs is for
the community to find a way of holding them true, a plausible
interpretation of the novel claims which is at least consistent with the
church’s central beliefs. Assimilative and inclusive power obviously go
together. Novel beliefs which have been assimilated are thereby
included in the community’s belief system, though inclusion need not
involve any explicit or deliberate process of assimilation. When the
epistemic status of the beliefs has been decided, whether by
assimilation into the community’s belief system or rejection from it,
they are no longer novel.8 By definition all novel beliefs start out as
alien, and most will remain that way, but whether they do depends on
how their epistemic status is settled. Novel beliefs may arise which the
community eventually not only assimilates, but takes to be central; for
many Christians, this is the case with, for example, the sinlessness and
bodily assumption of Mary.
Since novel beliefs will for the most part already be held by people
7. For the case of Jews and Christians see the last section of this chapter. The converse
case might be exemplified by beliefs bound up with the practice of devotion to a
particular saint, which only highly socialized members of the Christian community are
likely to have, but which the community does not regard as central to its belief system.
8. The line between novel beliefs or practices and those whose truth or rightness has been
decided is not fixed; a belief or practice may seem decided for a long time, only to strike
the community as novel once again (as, for example, the practice of ordaining only men
1
to the church’s pastoral ministry once seemed settled, but has now widely been reversed).
Trinity and
1
outside the community, the question about whether they can be
assimilated is generally a question about whether the church can
share a particular stretch of belief with at least some of those who are
not its members. That is, at any rate, our chief concern here.
A community whose belief system displays an ongoing capacity to
assimilate novel beliefs will naturally be one which shows itself
capable of changing its established beliefs, and so of holding its beliefs
in an epistemically responsible way. Pressure to change beliefs currently
held very often comes from encounters with novel beliefs for which
good reasons can be offered, though of course it can come in other
ways as well (such as by continuing communal reflection on the
meaning and implications of its own most central beliefs – it was
chiefly in this way, for example, that the church rejected its own
previous tolerance of slavery). Rejecting such novel beliefs does not, of
course, require changing established communal convictions, but
assimilating them often will. Assimilation often requires, that is,
rejecting established beliefs inconsistent with the novel claims to be
taken in.
In cases of this kind, the more closely novel claims impinge on
the community’s central beliefs, the more difficult they will be to
assimilate. Should a community be able to accept novel claims only
at the cost of rejecting beliefs which are epistemically primary for it,
this would not exhibit the assimilative power of that community’s
belief system, but signal the end both of the belief system and the
community whose identity depends on holding true the rejected
beliefs. A communal belief system will thus have assimilative power
to the extent that it can incorpo- rate novel truth claims without
making the community give up its own most central beliefs, and so
abandon its identity-constituting epistemic priorities.
Encounters with novelty

The church’s engagement with novel beliefs and belief systems has
been ongoing since its earliest days. As the full corpus of Aristotle’s
writings made its way into the Latin West in the twelfth and thirteenth
centuries, the Western church found itself confronted by a belief
system whose explanatory power seemed to exceed that of any
philosophical outlook which the church had previously confronted
and assimilated. This com- pelling belief system appeared not only to
be at odds with teachings long regarded as indispensable by the
church (such as the immortality of the soul and the creation of the
1
world in time), but to support the teachings of
Trinity and
1
a rival religion – Islam – whose theologians were practiced at using it as
an apologetic tool. In this encounter the strategy which prevailed in the
church after long struggle was to assimilate Aristotle, by showing that
his philosophy can be accepted in virtually all of its particulars without
generating inconsistency with any central Christian beliefs, and indeed
that Aristotle in some important respects strongly supports Christian
claims. Were such inconsistency to arise it would, all but a few were
agreed, be too bad for Aristotle. But there seemed good reasons to accept
Aristotle’s physics, metaphysics, psychology, and cosmology over the
available alternatives, even though this meant changing well-established
(mostly Platonistic) habits of thought. In a series of highly technical
debates (on issues like the creation of unformed matter and the number
of intellec- tual souls) Christian thinkers in the medieval west became
widely con- vinced that they could have it both ways: their commitment
to Christian teaching did not require them to give up being Aristotelians
(that is, to reject as false the results of a textually persuasive
interpretation of Aristotle).9 This close reading of Aristotle in the
Christian community left neither the ancient philosophy nor the western
church’s view of the world unchanged, though the changes were of
different kinds. The church’s success at taking Aristotle in created a
transformed and much expanded Aristotelianism, arguably superior to
other assimilations of Aristotle, and it greatly altered the ways in which
the Christian community and its members would explain and
defend their most central
beliefs, though not the beliefs themselves.
Sometimes, however, assimilation of novel beliefs does lead to com-
munal reassessment of central convictions. The geological discoveries
of the eighteenth century, and Darwin’s development of an
evolutionary biological theory in the nineteenth, made it seem
incredible, despite what Christians had long tended to suppose, that
the stories in Gen. 1–11 could be true – that the origin of the world and
humanity could plausibly have taken place as they describe. One
widespread (though by no means universal) Christian response to the
community’s encounter with this particular set of novel beliefs has
been to propose a shift in the literary genre to which these stories are
assigned. They are not to be taken as science or history, but as myths,
sagas, or realistic (but in this case fic- tional) narratives. As such they
can still be mined for beliefs about God,
9. For analysis of one example of this procedure – Aquinas’s assimilation of

Aristotle’s account of matter to the creation narrative of Gen. 1 – see Marshall,
1
“Absorbing the World,” pp. 90–7.
Trinity and
1
the world, and humanity which remain abidingly central to the
Christian community: that in every respect the world’s existence
depends upon God’s creative will, that among all of God’s creatures
human beings are created in and for a distinctive and uniquely
intimate relationship with God, and so forth.10 But reading Gen. this
way also means taking some beliefs long regarded as relatively central
to a Christian understanding of the world as simply false: that Adam
and Eve were the first humans, that the rest of humanity is biologically
descended from them, and that the lost condition from which
humanity is redeemed by the triune God can be traced in a quasi-
causal way to this first pair.
Here the assimilation of alien and novel beliefs has meant a restruc-
turing of the communal belief system relatively close to the center,
since beliefs once seen as bound up with the most central ones are not
merely relegated to the periphery, but discarded altogether. Perhaps in
part for this reason the genre shift which funds the restructuring still
meets with resistance in the Christian community. Those who support
the shift argue, however, that it brings needed clarity, distinguishing
those beliefs which are genuinely central and indispensable to
communal identity from those which have simply been around for a
long time. One form of this argument, suggested by Kierkegaard and
exploited by theologians like Reinhold Niebuhr and Paul Tillich,
maintains that the Christian doctrine of original sin makes much more
sense if Adam and Eve are not regarded as the historical first parents
of humanity than if they are; shorn of its cumbrous historical
trappings, the doctrine becomes at once a genuinely central Christian
teaching and a truth about human existence per- spicuous to any
honest and attentive analysis of self – indeed the indispensable means
by which modern people can make sense of Christian teaching at all.11
Karl Barth gives the argument a quite different form, in which the
10. The suggestion that these texts have manifold meanings beyond their sense as
descriptions of events in nature and history is, of course, an ancient Christian
commonplace. And since finding sense in the texts does not depend on taking them as
science or history, supposing that they propose true and important beliefs about God and
humanity does not require supposing that those beliefs are of a scientific or historical sort
– a point generally upheld by those ancient interpreters who also assume their scientific
and historical veracity, and emphasized by those who, like Origen, entertain some doubts
on this score.
11.See, e.g., Reinhold Niebuhr, The Nature and Destiny of Man (New York: Scribners, 1941),
vol. I; Paul Tillich, Systematic Theology (Chicago: University of Chicago Press, 1951–63),
vol. II. On the centrality of sin and its conceptual variants as an apologetic device in
modern Protestant theology, see Hans Frei, The Eclipse of Biblical Narrative (New Haven:
Yale University Press, 1974), pp. 124–30.
1
impossibility of appealing to Adam and Eve to justify any account of
the human condition helps to bring out a cardinal point of which the
tradition had (he supposes) at best only partial sight: that an adequate
grasp of the human condition requires (rather than being required by)
a grasp of Jesus’ identity as the redeemer of humanity, by contrast
with whom our own condition emerges as a “bizarre antitype” and so
as sin.12 In cognate fashion Karl Rahner argues that christological
assimilation of an evolutionary view of the world highlights rather
than diminishes the unsur- passability of Jesus Christ.13 For both
Barth and Rahner, decentralizing traditional claims about Adam and
Eve as historical figures has the effect of emphasizing the centrality of
the church’s christological and trinitarian beliefs – that body of beliefs
whose epistemic status must remain stable if the attempt to take in
novel claims is to display the assimilative power of a belief system
rather than to prove its undoing.
This sort of effort to include and assimilate alien and novel
beliefs stems from the church’s own most central commitments – from
the narratives which identify Jesus. These narratives require that
those who hold them true should not simply be prepared to cope with
unfamiliar claims, but should actively seek out novel beliefs and try to
find a place for them in a Christian vision of the world. As Paul puts it,
Christians are called to “take every thought captive to obey Christ” (II
Cor. 10:5). To hold true at all the narratives which identify Jesus and
the Trinity calls for an always ongoing communal effort by Christians
to interpret and assess whatever novel claims they encounter, and
whatever they think they already know, by trying to find a place for
them in the world created through and for Jesus Christ, and put at
peace with God by the blood of his cross – the world of the triune
God, who makes all things new by drawing them into the interior
space of his own life through Jesus’ passion and resurrection, where
nothing has exactly the place we would otherwise have thought.14 The
imperative to take in alien beliefs is therefore not imposed on the
Christian community from without, but goes with holding a Christian
view of the world in the first place. It calls for an effort of
interpretation
12. Barth argues this point extensively in the volumes which make up his doctrine of
reconciliation; the quoted phrase is from Kirchliche Dogmatik IV/3 (Zurich: Evangelischer
Verlag, 1932–67 [for the complete work]), p. 426 (ET, p. 369).
13.See his “Christology within an Evolutionary View of the World,” Theological
Investigations (New York: Seabury and Crossroads, 1961–), vol. 5, pp. 157–92.
14. This way of putting the matter is suggested by Hans Urs von Balthasar, Herrlichkeit, vol.
I: Schau der Gestalt (Einsiedeln: Johannes Verlag, 1961), p. 594; see p. 127 (ET: The Glory of the
Trinity and
1
Lord, vol. I: Seeing the Form, trans. Erasmo Leiva-Merikakis [San Francisco: Ignatius Press,
1982], pp. 615, 135).
1
and assessment which will be ongoing as long as there are novel beliefs
to encounter, that is, until the end of time.15
But the church’s primary criteria of truth will not furnish the
primary reasons for holding most beliefs true. The Christian
community and its members will naturally be able to decide the truth
of relatively few proposals for belief by direct appeal to those beliefs
which are epistemically primary for the church, since beliefs
themselves inconsistent with one another will often alike be consistent
with that narrative and its interpretation. Confronted with alien or
novel beliefs, we will for the most part be unable to tell whether they
conform to the Christian community’s epistemic priorities except by a
process of reasoning which strives to work out inferential connections,
often of a complex and indirect sort, between the beliefs in question,
the reasons we have for holding them, and those beliefs which are
centrally Christian. Only by striving to include them, in other words,
will the Christian community and its members be able to tell whether
alien and novel beliefs cohere with their central convictions. And since
all beliefs have to be consistent with the narratives which identify
Jesus and the Trinity in order to be held true, the scope of the Christian
community’s assimilative effort has to match the scope of the
community’s epistemic priorities: both must be universal and unre-
stricted.16
The epistemic force of inclusive power
But why should success at including alien beliefs and assimilating

novel ones count in favor of the Christian community’s belief system?
Conversely: why suppose that in order to be justified in holding their
15. It might be thought that this effort to assimilate the novel, to “take every thought
captive”, reflects an illicit “totalism” which theology ought to reject (on which see
Emmanuel Levinas, Totality and Infinity, trans. Alphonso Lingis [Pittsburgh: Duquesne
University Press, 1969]). If “totalism” is the view that all possible beliefs have in some
sense already been assessed, so that there can be no genuinely novel beliefs and, on that
account, no possibility that we will have to change our epistemic priorities, then the
present argument is anti-totalistic: there will be novel beliefs until the eschaton, and they
might lead us to change our minds at any point. If, by contrast, “totalism” is the view that
there may be beliefs with unrestricted epistemic primacy, beliefs whose epistemic scope is
the whole open field (perhaps to a large extent yet unexplored) of possible belief, then in
this sense theology is totalistic, because Christian belief is itself totalistic; it is not even
possible to say what counts as “Christian” belief except by reference to the entire open
field of possible belief, across which some beliefs have a decisive epistemic role.
16. Under the limitations of actual practice the truth value of some beliefs will of course
be left undecided, since they seem too remote from the community’s chief concerns to
make deciding about their truth worthwhile. At the same time any belief, no matter
how remote it might seem, could become one whose truth value urgently required
deciding.
Trinity and
1
own most central beliefs, Christians need to maximize agreement pre-
cisely with people who do not hold those beliefs? The notion of
inclusive power calls for just such agreement; the greater the
assimilative capacity of a belief system, the more extensively its
holders will find themselves sharing beliefs proposed by people
outside the community. But surely the gospel radically calls into
question all of our beliefs, and in particular all of the epistemic
priorities we humans are otherwise inclined – or tempted – to hold.
This, it might seem, precludes extensive agreement in belief between
the members of the Christian community and those who have other
epistemic priorities, or at least rules out any epistemic need for such
agreement. In that case failure of inclusive power would not count as a
reason to suppose the Christian community’s central beliefs were false,
nor would successful inclusion of novel beliefs count as a reason to
hold them true.
In order to see why inclusive and assimilative power count in the
epistemic favor of a system of belief structured by the church’s
priorities, and also why assimilative failure counts against it, we need
to reflect a bit further on the connections between meaning, truth, and
belief. More exactly, we need to bear in mind the ties which bind the
way we interpret a speaker’s sentences, the truth value we assign to
them, and the extent of our agreement with the speaker in the
assignment of those truth values (the extent of our agreement, that is,
with the speaker’s beliefs). Is it even conceivable that we could have a
set of epistemic priorities, and with that a belief system, which
permitted (let alone required) us to disagree with other speakers most
of the time?
Such a scenario could be conceived only if it were possible to
interpret a speaker’s utterances correctly on the whole, and still find
the speaker for the most part mistaken, his utterances (and so the
beliefs we attribute to him) generally false. A workable but non-
circular theory of interpretation suggests, however, that this is not
possible. If we are to interpret plausibly, we will have to find speakers
for the most part correct in the truth values they assign to sentences,
which is to say that we will for the most part have to agree with them.
Radical interpretation has to presume, in other words, that all speakers
have mostly the same set of beliefs. The more extensively we find
ourselves disagreeing with others – holding false the sentences they
hold true, and conversely – the more we will have to suspect that we
have misunderstood them. Under the guid- ance of the principle of
1
charity, we will have to reinterpret their words so as to agree with
them on the whole, though not of course at every point.
Trinity and
1
With regard to the issue at hand, this implies that epistemic
priorities which would permit us to find the utterances of others
massively false are not simply undesirable, but impossible to sustain. If
we have epistemic priorities which seem to be placing us in
increasingly broad disagreement with other speakers, we will
eventually have to take one of two possible paths. (1) We can
reinterpret sentences which we hold false but others generally hold
true, so as to find them more in agreement with our epistemic
priorities, and so more generally true. Or (2) we can hold false the
beliefs which require such broad disagreement with other speakers,
that is, we can change our epistemic priorities. In one way or the other
we will have to find ourselves in agreement, for the most part, with
other speakers.
These options mutually limit one another, and so cannot simply be
chosen at will. Just because plausible interpretation has to optimize
agreement among speakers, epistemically troublesome sentences can
be reinterpreted only to the extent that revised assignments of
meaning tend to diminish, rather than increase, disagreement with
those who hold them true. We cannot, for example, reinterpret the
sentences which constitute evolutionary theory so as to bring them in
line with Christian epistemic priorities, if our reinterpretation tends to
increase rather than diminish our ascription of falsity to the sentences
held true by other speakers (including, though not only, speakers who
hold evolutionist beliefs). The result would be a less, rather than more,
plausible interpretation of the troublesome sentences.
The economic alternative to bad interpretation, as we observed in
the previous chapter, is to revise the assignment of truth values. What
guides the assignment of truth values, as the current problem brings
out, is the principle of charity, which optimizes agreement. If a
particular set of epistemic priorities can only be sustained at the cost
of massive disagreement with other speakers, and so at the cost of
interpreting their utterances badly, the only plausible interpretive
course is to hold false the beliefs which make up those epistemic
priorities, and with that to revise the priorities themselves. In this way,
the failure of our epistemic commitments to yield broad agreement
with other speakers, and so with beliefs which for us are alien and
novel, counts against those commitments; conversely, success at
generating agreement – inclusive power – counts in their epistemic
favor.
It may seem, however, that granting the inclusive power of
1
Christian beliefs turns out not to support their unrestricted epistemic
primacy, but
Trinity and
1
instead to eliminate it. We have argued all along that for central
Christian beliefs to be epistemically primary across the board means
that all other beliefs have to agree (be consistent) with them in order to
be regarded as true. But the argument just developed for the epistemic
relevance of a belief system’s inclusive power turns on the thought
that for these Christian beliefs themselves to be held true, they have to
agree on the whole with the rest of what everyone, including
Christians, mostly believes. This seems an unhappy result, since we
were looking for reasons for holding central Christian beliefs which do
not create epistemic subordination.
This conflict, however, is apparent rather than real. The need to
optimize agreement among speakers plays a different role in the
admission of beliefs into our overall view of the world than does the
requirement that any belief we hold must be consistent with our
epistemic priorities. Conflict would arise only if we assigned the
function of epistemic priority first to one set of beliefs, then to another
and different set. In order to dispel the appearance of conflict, we
therefore need to distinguish clearly the epistemic role played by
“most of ” our beliefs from the role of those which we take to be
primary.
We can admit and retain in our belief system only those sentences
which, in our ongoing effort to interpret the utterances of other human
beings, allow us to share most of the beliefs we attribute to them. But
the requirement that we optimize agreement is a coarse net; it will
filter out countless possible beliefs, but will admit countless more,
including many which are in conflict with one another. The need to
optimize agreement gives us nothing to go on when it comes to
resolving these conflicts
– the ones which arise between interpreted sentences which have
already passed the test of agreement with most of the beliefs we and
other speakers have to share in order to make sense of one another. In
order to have a largely coherent system of beliefs, we will regularly
have to decide between contradictory claims alike compatible with
most of the convictions we and others share. We will therefore have
to fail of agreement with beliefs maintained by lots of other speakers,
either because we have no opinion on the matter, or because by our
own epistemic standards we find their beliefs false – as we will
inevitably do in some cases, precisely in order to agree with them on
the whole.17 We can agree with others on a
17. See Davidson’s comments on this point in “A Coherence Theory of Truth and
1
Knowledge,” Truth and Interpretation: Perspectives on the Philosophy of Donald Davidson, ed. E.
LePore (Oxford: Blackwell, 1986), p. 318.
Trinity and
1
host of trivialities sufficient in breadth to establish that we are
interpreting each other well (“Snow is white,” “Grass is green”), and
still disagree about what each of us takes to be most important (“On
the third day Jesus rose” – or “did not rise” – “from the dead”; “God
was” – or “was not” – “in Christ reconciling the world to himself ”).
When either of two conflicting beliefs would allow us to agree with
most of what our interlocutors hold true, and so be good interpreters,
we need epistemic priorities to resolve the conflict – to decide which of
the beliefs, if either, to accept.18
The unrestricted epistemic primacy of central Christian beliefs thus
seems quite compatible with the requirement that a plausible set of
epistemic priorities exhibit inclusive power: no contradiction arises
from saying both that an interpreted sentence can be held true only if it
is consistent with the church’s central beliefs, and that the same
sentence can be held true only if it is consistent with most of the beliefs
we attribute to others. Both are, in other words, necessary conditions
for deciding about the truth of beliefs; neither is, by itself, sufficient.19
Indeed the present objection (that we have unwittingly eliminated the
epistemic primacy of central Christian beliefs) itself points up the need
for concrete epistemic priorities – particular beliefs by adverting to
which we make decisions about the truth of other beliefs. If agreement
with “most” of our beliefs were sufficient as well as necessary for
deciding what belongs in our belief system, then all of our beliefs
would be epistemically subordinate to most of our beliefs. But this is
absurd. If all of our beliefs were epistemically subordinate, there
would be no beliefs left over for them to be subordinate to.
None of this presents any obstacle to the thought that the gospel
itself is radically novel, given the ways we human beings are otherwise
inclined to look at the world. The epistemic novelty of the church’s
central beliefs does not depend on the apparently hopeless suggestion
that the gospel is opposed to most or all of the beliefs which the rest of
humanity holds true, but only on the contrast between the gospel and
the epistemic priorities human beings are otherwise inclined to have.
The gospel does not call all or even most of our beliefs into question
(“Snow is white,” and so
18. In practice we do not, of course, first decide which beliefs to admit and then decide
upon our epistemic priorities; rather in seeking to understand one another we
inevitably test the priorities we already have against the limit of excessive
disagreement.
19. There would, of course, be a contradiction if consistency with central Christian
beliefs were supposed to be a sufficient condition for the truth of all other beliefs. But, as
we have argued, it neither can nor need be that sort of condition.
1
forth), and it is not for this reason that the gospel overturns our settled
epistemic priorities. The gospel does, however, conflict with some of
the conventional beliefs which are closest to it in logical space, and in
particular with the inferences we are inclined to draw from some of
the obvious beliefs we all share (beliefs, for example, about what
happens when we die, as observed in the previous chapter). Since we
will readily use these conventional beliefs as standards by which to
judge the truth of the novel claims of the gospel, believing the gospel
requires a reversal of the epistemic priorities we would otherwise
have, but it does not require that we reject most of the beliefs we
would otherwise have.
Here the traditional theological distinction between nature and
grace helps clarify matters – or, more precisely, the distinction between
the status integritatis in which God created human beings (not to be
confused with the entirely hypothetical state of “pure nature”) and the
status corruptionis in which human beings now live. Fallen human
beings want to reject the triune God (even granted that, at another
level, they implicitly desire God), and so inevitably fashion epistemic
priorities inconsistent with the gospel by which God identifies and
gives himself to us. So for fallen humans the novelty of the gospel
requires a reversal of epistemic priorities – and with that holding false
some beliefs we are otherwise deeply inclined to hold true – if we are
going to believe the gospel at all. This means that in the nature of the
case, believing the gospel requires a change of heart as well as a
change of mind (a point to which we will return in the next chapter).
Were human beings still in their original created condition, without
sin, the gospel – which is to say, the Spirit’s disclosure to human beings
of “the deep things of God” (see I Cor. 2:10) – would not meet the
resistance of epistemic priorities already opposed to it; such resistance
presupposes sin. But the novelty of the deep things of God would still
be beyond the epistemic capacities of nature. The epistemic priorities
we could otherwise have (say, those given by Quine’s and Davidson’s
occasion sentences) would not be opposed to these novel beliefs, but
would leave us with no basis for deciding about them one way or the
other. In order to be justified in holding these beliefs true at all, we
would have to take them as epistemically primary. So in either case, in
the status integritatis as well as the status corruptionis, the novelty of the
deep things of God requires the epistemic primacy of the beliefs by
which we apprehend them.20
20. Thus, to recall a precedent, Aquinas’s insistence (Summa theologiae II–II, 5, 1, c) that even
Trinity and
1
Adam and Eve in paradise needed faith to know the Trinity; see the discussion in
Marshall, “Faith and Reason Reconsidered,” The Thomist 63/1 (1999), p. 14.
1
The limits of epistemic judgment
The usefulness (as distinguished from the logical force) of inclusive

power as a test of epistemic priorities lies in its employment of shared
beliefs in order to assess the epistemic rights of those which are not
shared. Others need not agree with our epistemic priorities, but need
only be sufficiently diligent, in order to discover whether those
priorities tend to display inclusive power over time. It has long been
observed that reasoned argument is possible only on the basis of
convictions shared by all sides.21 The present claim is that even when
it comes to beliefs of unrestricted epistemic primacy, the convictions
we share with those who do not hold these primary beliefs will always
provide ways of testing the claims we take to be primary, a test whose
force can be acknowledged regardless of one’s epistemic priorities.
But the same feature which makes this a readily sharable test also
limits its effectiveness. Failure of assimilative power rules out
epistemic priorities, and with that the belief systems structured by
those priorities. Lots of conflicting beliefs, however, can pass the test of
agreement with “most of ” what we and others alike hold true. This
creates the possibility of conflicting epistemic priorities. Successful
assimilation over time of alien and novel beliefs seems to be a feature
of manifold actual belief systems which conflict with one another at
crucial points, not least in their epistemic priorities; this includes a
number of religions. The justificatory force of assimilative power is
thus strictly negative; it can eliminate beliefs and epistemic priorities
as false, but it cannot fix the truth value of those which remain – it
cannot individuate a single true set of epistemic priorities, still less a
single true belief system.
That we will always share enough beliefs with one another genuinely
21. So Thomas Aquinas, for example, argues that when it comes to first principles (e.g., in
metaphysics or, above all, theology), one “enters into argument with a person who denies
one’s own principles if the adversary concedes at least something, but if he concedes
nothing it is impossible to argue with him, although one can dismantle the reasons he
gives” – that is, we can show that our first principles do not involve absurdity or
contradiction with what we otherwise have good reason to believe (Summa theologiae I, 1, 8,
c). Thomas thought of this as an idea which came to him from Aristotle, in particular
from Metaphysics IV, 7–8 (1012a 18–b 10). Commenting on this passage, Thomas makes a
suggestion strikingly similar to some of the Davidsonian considerations of which we
have made use. Should we find another disagreeing with us in seemingly peculiar ways
about what there is, we “can ask that he grant that the words mean something; without
this the argument vanishes (disputatio tollitur).” In Duodecim Libros Metaphysicorum Aristotelis
Expositio, ed. M.-R. Cathala and R. Spiazzi (Turin/Rome: Marietti, 1964), 4, 17 (no. 740).
That is: if we do not mean the same thing by our words – if we are not talking about the
same thing – then we do not disagree (“disputatio tollitur”); if we do mean the same thing
by our words, then we will find, Thomas goes on to suggest, that we cannot hold utterly
Trinity and
1
divergent beliefs.
1
to test one another’s epistemic priorities does not, therefore, entail that
we will always be able to settle conflicts of belief, and in particular con-
flicts about which beliefs ought to have epistemic primacy. “Settle,”
that is, in a fashion acceptable to all sides in an argument; we
inevitably resolve some disputes by appeal to beliefs and epistemic
priorities which are not shared with at least some of the people with
whom we are talking, and so do not elicit their assent. There need be
nothing irrational in such appeals to unshared beliefs. They merely
point up the epistemic limits of what is shared, the room for rational
disagreement which even massive agreement leaves open.
In the nature of the case, differences of epistemic priority (especially
when the priorities structure comprehensive belief systems, like relig-
ions) are differences over what count as good reasons to hold or reject
beliefs, particularly in disputed cases. The apostle Paul argues that
belief in Jesus’ resurrection makes a good, indeed decisive, reason for
believing in the resurrection of those who have faith in Jesus, but he
also acknowledges that his modus ponens may – to their detriment – be
the Corinthians’ modus tollens: their skepticism about the future
resurrection might lead them to deny Jesus’ resurrection as well (see I
Cor. 15:16–17). Within the limits of plausibility established by inclusive
power, adherents of conflicting beliefs and epistemic priorities will
often differ not only about what is true, but about what confers or fails
to confer epistemic right.
Commensurability
This last thought is sometimes expressed by saying that beliefs and
epistemic priorities can turn out to be “incommensurable.” If this
term simply labels the sort of conflict between beliefs which cannot be
resolved except by appeal to the disputed beliefs themselves, then the
foregoing argument for the epistemic significance of assimilative
power allows for “incommensurable” beliefs and epistemic priorities.
But the term is often used to suggest two much stronger claims, both
of which are ruled out if appeal to assimilative power has any
epistemic force.
1. It is sometimes argued that beliefs and systems of belief (or “con-
ceptual schemes”) can be incommensurable in that the adherents of
one communal belief system simply cannot understand the adherents
of another communal belief system. It is not merely that the two
communities hold different beliefs – one either rejects or suspends
judgment on various beliefs proposed by the other – but that one
Trinity and
1
community has no access, in whole or in part, to the contents of the
other community’s
1
beliefs; the conceptual resources available to the one have no
equivalents for what is affirmed in the beliefs of the other. The
assumption that (at least for humans) having beliefs depends on
having a language connects the notion of incomprehensible beliefs to
that of untranslatable languages: holding beliefs we cannot
understand goes together with speaking a language we cannot
interpret, one with no potential equivalents in our own linguistic
stock.22 Should this strong notion of incommensurability prove out,
the power to include the beliefs of others, to count them right most of
the time, will be of little use as a test for our own epistemic priorities.
The very beliefs we will seek to assimilate, namely ones we do not
already have, may well turn out to be convictions which, as non-
adherents, we cannot understand, let alone decide whether to count
right.
Everyone acknowledges that dramatic differences of both belief and
meaning sometimes come to light. The question for our purposes is
whether these support the claim that there can be incomprehensible
cultural outlooks or untranslatable languages: beliefs we can
understand only by holding them, communal convictions which we
necessarily distort unless we adhere to them, or languages which
permit their users to hold sentences true which we cannot interpret.
The arguments against this strong notion of incommensurable
beliefs are now widely familiar. In order to say that the beliefs of
another community belong to a worldview which is for us alien or
foreign, we have to know what their beliefs are – we have to
understand them. Beliefs we cannot comprehend are obviously beliefs
we cannot classify as either foreign or domestic. Now it seems that we
can only attribute beliefs to others, and discriminate among their
beliefs in the fine-grained way we routinely do, if we understand what
their words mean. Speakers of a language which we cannot interpret
(or translate) using the resources of our own tongue will therefore
likewise be creatures whose beliefs, to the extent that we cannot fix the
meaning of the sentences they hold true, we
22. So for example John Milbank, following Alasdair MacIntyre, argues that “in certain
cases, the signifying terms of one cultural outlook simply cannot be translated into the
signifying terms of another, without betrayal and distortion,” and infers from this that
“for the insider” to a religious or other cultural tradition, “without belief, understanding
can only be partial. So any claim to full understanding on the part of the outsider must
negate the alien tradition’s own self-understanding.” Theology and Social Theory: Beyond
Secular Reason (Oxford: Blackwell, 1990), pp. 340–1. For the background in MacIntyre, see
especially Whose Justice? Which Rationality? ( Notre Dame: University of Notre Dame
Press, 1988), pp. 370–88. With Milbank’s concern that theology avoid captivity to
“secular reason” – that theology keep its epistemic priorities straight – I concur; the
Trinity and
1
question at present is whether the notion of incommensurability helps make this point.
1
can judge neither incommensurable nor commensurable with our own.
They will, indeed, be those to whom we have no apparent grounds for
attributing beliefs in the first place. The idea that we might run across a
communal worldview so foreign we could not understand it seems
like an idea we do not really have. Given what we mean by a worldview
(a big collection of mostly coherent beliefs, expressible in a language),
our interpretive failure would eventually signal not that we had
encountered a radically foreign worldview, but no worldview at all.
Advocates of radical incommensurability try to handle this problem
by saying that we do not comprehend the beliefs of an alien
community by translating their language into our own; rather we
become “bilingual,” we learn to speak their language like natives.23
Moving the sup- posedly untranslatable language into the head of the
interpreter does little, though, to make sense of the notion of radically
incommensurable beliefs. Suppose that I, a speaker of L1, acquire native
fluency in L2 (I need
not, of course, learn L2 primarily by translating it into L1). If I, the speaker
of L1, cannot interpret L2 in L1, then I can no more tell whether or not
the beliefs I have when I hold sentences in L2 true are
incommensurable with those which I have when I use L 1 than I could
when L2 was spoken by somebody else. Likewise I the speaker of L1
will have no way to tell whether my interpretation of what I say as a
speaker of L2 betrays and distorts my L2 utterances. In fact if I cannot
translate these languages into one another, then when I speak L 2, I, the
speaker of L1, will not know what I am saying. The beliefs I have as a
speaker of L2 will consequently be incomprehensible to the speaker of
L1, that is, to me. This assumes, of course, that “I” refers to the same
thing when I speak L1 and L2. If it does not, then we are back to the
more familiar problem of alien speakers
whom we do not understand; moving the two speakers into the same
head has not helped make sense of the notion of radically
incommensurable beliefs.
It therefore also proves difficult to get a fix on the claim, which
often seems to drive appeals to radical incommensurability, that others
might have convictions we could understand only by sharing them, or
conversely that we (for example, we Christians) might have
convictions which others must accept in order to understand them. As
it turns out, either we need not share any particular belief of another in
order to understand that belief, or sharing the belief will leave us as
much in the
Trinity and
1
23. “Bilingual” is Milbank’s phrase (Theology and Social Theory, p. 341); MacIntyre calls
this, oddly, learning a “second first language” (Whose Justice? Which Rationality? pp. 364,
374).
1
dark about its content as we were in the first place. To reject the beliefs
of others is by itself neither to distort their view of the world nor to
lack comprehension of it, but simply to disagree with it. To differ in
other ways from their beliefs (by suspending judgment on convictions
they cherish, for example) is neither to betray nor to misunderstand
their view of the world, but simply to hold a different one.
Of course we often enough fail to understand one another.
Sometimes this happens because of differences in what we mean,
sometimes because of differences in what we believe. We are practiced
at coping with both kinds of difference, against the background of that
massive sameness of meaning and belief assumed by the principle of
charity, which funds plausible interpretation in the first place.24
Our concern here is with understanding the beliefs of others, with
grasping the meaning of the sentences they hold true.
“Understanding” is often used in a broader way, as when we try to “get a
feel for” the situation and outlook of another. This affective and perhaps
volitional sort of “understanding” surely does depend on sharing to some
extent the beliefs of those we hope to “understand,” assuming that
emotions are belief dependent. If the foregoing argument is correct,
then sharing the beliefs of others and so getting a feel for their situation
depends on grasping the contents of their beliefs, which one need not
share the beliefs to do; we can understand the beliefs of others well
enough to know whether we also want to get a feel for their situation.
Even this affective sense of understanding does not necessarily require
conversion to the disputed beliefs of the other, since there is no one-to-
one correlation of affections with beliefs, and we will in any case share
most of the other’s beliefs. There is also such a thing as willful
misunderstanding and distortion of others, a deliberate refusal to
grasp the sense of their utterances and beliefs. This we normally regard
as a vice. It does not support the notion of incommensurable beliefs; if
distortion were unavoidable then we would presumably not hold others
morally accountable for engaging in it.
2. A less radical idea of incommensurability turns not on the notion
of meaning, but of justification or warrant. It picks up the suggestion
24. Coping with difference successfully – understanding one another – does not require
that there always be a clear way of deciding whether understanding is best served by
taking differences as matters of meaning or of belief. As Davidson observes, “When
others think differently from us, no general principle, or appeal to evidence, can force us
to decide that the difference lies in our beliefs rather than in our concepts” (“On the Very
Idea of a Conceptual Scheme,” Inquiries into Truth and Interpretation [Oxford: Clarendon
Press, 1985], p. 197).
Trinity and
1
that differences in epistemic priority entail differences in what count as
good reasons for holding beliefs, and proposes that any belief system
with reasonably clear epistemic priorities will always succeed by its own
lights. Decisive criticism of such systems is therefore impossible; they
will always turn out true by their own criteria of truth, and include the
beliefs which conform to those criteria. As an epistemic test inclusive
power is not just limited, but useless; all belief systems will pass it.
Just as the requirements of plausible interpretation limit the extent
to which our beliefs can be supposed to differ from those of the
speakers we are trying to understand, so the same requirements limit
the degree to which we and they can differ on what count as good
reasons for beliefs. Since we have to assume for the most part that our
beliefs are the same as theirs, and good reasons are relations of
evidence, support, agreement, and so forth, among beliefs, there is no
chance that we can differ massively from them on what to count as
good reasons, any more than we can differ massively on what to
believe. Combine this with the earlier obser- vation that even the most
central beliefs will generate only a fraction of the primary (that is, most
decisive) reasons for holding sentences true within a total belief
system, and it becomes obvious that any belief system, no matter how
comprehensive, will always be able to fail by its own standards.25
Conceiving such epistemic failure is not difficult. It threatens when-
ever beliefs which a community has good reason to hold, or which
seem too obvious to deny, imply the negations of the community’s
most central convictions. Various strategies for meeting such threats
present themselves. The community can try to assimilate the alien
beliefs, arguing that they do not in fact imply the negations of its
central beliefs (as with Aristotle for the Christian community in the
thirteenth century). It can argue that the beliefs negated by the novel
claims are not really central for it (as with evolution in the nineteenth
century), though there are evident limits to pursuing this strategy;
eventually the community will run out of beliefs to decentralize. It can
mount a counterargument seeking to show that the reasons for
rejecting the troublesome beliefs are better than the
25. A communal tradition, as MacIntyre argues, can undergo an “epistemological

crisis,” and may find that it has suffered “defeat in respect of truth” at the hands of an
“alien tradition,” when the latter provides the resources (which it lacked) for solving its
epistemological crisis (Whose Justice? Which Rationality?, pp. 361, 365). MacIntyre’s analysis
of epistemological crises has bite, to be sure, only if his analysis of radical
incommensurability is wrong – only if we can understand, without (or without yet)
belonging to, traditions which might defeat us, including, presumably, those most
different from our own.
1
reasons for accepting them (as Christians in the thirteenth century did
not so much with Aristotle as with Muslim interpretations of
Aristotle). Or the community can grant that the problem is insoluble,
that it has no insight into the way in which novel claims which it has
good reason to believe hang together with its most central convictions,
though it is con- vinced that they must. This may involve a
determination to keep working on the problem in the hope that greater
light may come, but the difficulty may also be one whose solution we
should not expect in advance of the eschaton. The decision to live with
this sort of epistemic tension can be made only sparingly. Though
armed truces can under some circumstances go on indefinitely, if the
front on which one attempts to fight this delaying action becomes too
broad, or the battle goes on too long, outright collapse becomes more
imminent – less metaphorically, giving up central beliefs increasingly
looks like the only plausible course.
Strong objections
This suggests that the strongest objections to Christian belief – those
most telling to adherents of a Christian view of things – will not stem
from arguments whose force depends heavily on acceptance of rival epis-
temic priorities. On the assumption that the Christian community’s
central beliefs cannot be its epistemic priorities, modern theology has
often thought it necessary to revise the church’s central beliefs
precisely in order to accommodate alternative epistemic commitments.
But such calls for revision rarely carry the force of necessity suggested
by their advocates. Ceteris paribus, the more clearly the demand for
revision relies on alternative epistemic commitments, the more easily
the Christian community will be able to reject the demand – to take it
as one of those places where we can count our interlocutors wrong in
particular without endangering our agreement with them in general.
The strongest objections to Christian belief, as to any
comprehensive belief system, will rather stem from matters on which
everybody agrees – the more obvious the agreement, the better. A
really forceful objection to the Christian community’s belief system will
be one where the community’s central beliefs lead it to expect a
particular state of affairs which manifestly fails to obtain. Take, for
instance, the eucharistic disunity of the church. The New Testament
seems to propose that the temporal missions of the Son and the Spirit
from the Father have created a single human community, united chiefly
by the eucharist (see I Cor. 10:16–17), in a way visible to all (see Jn. 17:21,
Trinity and
1
23). The bonds which are visibly to unite this
1
temporal community are nothing less than those which eternally unite
the Son with the Father and the Spirit in love (see Jn. 17:11, 21, 26). If
the gospel (the narrative of the missions of the Son and the Spirit from
the Father) is true, therefore, we should expect to find in the world a
single community embracing all who celebrate the eucharist.
But we find no such community. Instead, obvious to Christians and
non-Christians alike, multiple communities celebrate the eucharist
while (in many cases) refusing to share it with the others. Multiple
communities recognize one another as Christian communities, but
lack the visible eucharistic bond which the missions of the Son and the
Spirit were supposed to create in the world by giving the world a
share in their own eternal bond of love. The existence of this
community would not, of course, demonstrate the truth of the gospel,
but its absence bespeaks the gospel’s falsity. Though perhaps not the
sort of objection which those who do not share the church’s beliefs are
likely to press, here we find a novel state of affairs which bears quite
directly on the church’s most central beliefs, and which the Christian
community has had relatively little success in assimilating. Whether
the church nonetheless has resources to cope conceptually and
practically with the epistemic problem posed by its own divisions is
therefore a particularly pressing question for Christian belief, a
question not evaded but intensified by regarding the church’s
trinitarian identification of God as epistemically primary.26
Taking the Christian community’s central beliefs as epistemically
primary without restriction does not, then, prevent alien and novel
claims from counting as good reasons against the church’s central
beliefs, calling for arguments in reply. Assimilative power is not
useless as an epistemic test of this (or, mutatis mutandis, any other)
community’s belief system; success at assimilating novel beliefs will
have to be earned, and does not come automatically once communal
epistemic priorities are fixed. And this means that the unrestricted
epistemic primacy of the beliefs which identify the Trinity is quite
compatible with the now commonplace notion of rationality to
which we alluded at the outset: the
26. On the argument of this paragraph see my essay, “The Disunity of the Church and the
Credibility of the Gospel,” Theology Today 50/1 (1993), pp. 78–89, but especially Ephraim
Radner, The End of the Church: A Pneumatology of Christian Division in the West (Grand
Rapids: Eerdmans, 1998). The end of the church as a coherent community will come,
Radner provocatively argues, not from failure or inability to maintain its epistemic
priorities, but as the divinely willed outcome of the divided church’s willful failure to
conform, in its life and practice, to what it claims its highest priorities to be.
Trinity and
1
view that being reasonable does not require (or even permit) us to
doubt all of our beliefs at once, but does require that we be prepared to
give up any particular belief, even the most central – and so be
prepared to change our epistemic priorities.
Locating conditions under which there would be good reasons to
give up the Christian community’s most central beliefs, and thereby to
abandon its epistemic priorities, is not especially difficult. Consider the
following scenario.27 A letter is discovered in an ancient
Mediterranean, now Turkish, village, addressed to one Paul, formerly
Saul, of Tarsus.
I can hardly believe we got away with it. The place where we hid
the body was so obvious, and it took so long before we could finally
get rid of it, that I’m amazed no one discovered it. And that story we
cooked up about seeing him alive after they crucified him – not just
once, but for forty days! Admittedly a few Athenians thought this
was pretty funny, but it’s astonishing how many people have
believed it. So let’s press on to Rome and see how far we can carry
this thing. Be careful, and write when you are able . . . As ever, Peter
That Jesus is not dead, though he once was, is utterly central to the
Christian community’s belief system. The New Testament also seems
to link the very possibility of having the belief that Jesus is risen to
hearing the testimony of others – ultimately, of a small group who did
not hear about it from anyone else, but met the crucified one in his
risen flesh, and were taught about the resurrection from the source.28
So if we have good reason to believe that those upon whose testimony
we depend in order to
27. Which I owe to a suggestion of Michael Root.

28. On this see Heinrich Schlier, “Kerygma und Sophia. – Zur neutestamentlichen
Grundlegung des Dogmas,” Die Zeit der Kirche: Exegetische Aufsätze und Vorträge I, 4th edn
(Freiburg: Herder, 1966), especially pp. 215–17 (on I Cor. 15:1–11). Modern exegetical and
theological discussions of Jesus’ resurrection have, to be sure, often vehemently contested
the idea that justified belief in the resurrection entails belief in the veracity of a handful of
first-century Jews whom the New Testament presents as having seen the risen Jesus –
have contested, in other words, the thought that anything depends on the truth of the
claim that Schlier offers as an interpretation of Paul: “The resurrection of Jesus Christ
occurred in front of witnesses, precisely in the sense that the risen one unmistakably
appeared to them as such” (p. 215). It is perhaps not accidental that objection to this idea
tends to go together with what we have argued are implausible interpretations of the New
Testament resurrection texts, viz., those which try to avoid attributing “is risen” to the
same referent to which they attribute “was dead,” and so try to avoid seeing the
resurrection as an event which happens to the dead Jesus. On such interpretations one will
of course not expect anyone to have seen the crucified Jesus, risen, so believing in the
resurrection does not depend on believing in the veracity of ancient witnesses who claim
to have done so. If, by contrast, rising to life (however more precisely construed) is an
event which happens to the dead Jesus, having the right to believe in the occurrence of
this event would seem to depend on either seeing its unambiguous result – Jesus’ risen
1
flesh – or on hearing about that flesh from those who have.
Trinity and
1
believe in Jesus’ resurrection were lying or deceived, we have good reason
to think that what they said is false.
Of course the counterfactual scenario we have just imagined would
be really forceful only if it could be developed much more elaborately
than we have the space to do here. Naturally one Christian response to
this fancied discovery would be to question the provenance of the
letter, and to argue that it could not be attributed to its putative author.
But suppose that scientific tests on the paper and ink ruled out a
modern origin for the letter, that it fits with our best historical
knowledge about Paul’s jour- neys, and so forth. The Christian
community would then find it increasingly difficult to deny the
authenticity of the letter without rejecting wide stretches of seemingly
obvious belief common to Christians and non-Christians alike;
eventually the failure of assimilative power brought on by trying to
maintain the belief that Jesus is not dead would call for dropping the
belief altogether.
That we can state conditions under which Christians would be pre-
pared to give up the belief that Jesus is risen does not, however,
require that this belief be held tentatively, or that it be regarded as
uncertain and therefore corrigible. What it rules out is taking certainty
and incorrigibility as descriptive rather than normative concepts.
Foundationalism took these to be descriptive notions. Confronted by
the right kind of evidence, a functioning mind could not help having
empirically evident or self- evident beliefs. Certainty and
incorrigibility were made parasitic on these types of belief; eliminating
the latter eliminates the former.
The Christian tradition, however, has generally taken the church’s
central beliefs to be certain and incorrigible, while explicitly denying
that these beliefs are self-evident, empirically evident, or even very
widely held. This position treats certainty and incorrigibility as norma-
tive concepts.29 To hold the Christian community’s central beliefs at
all is, in virtue of their content, to regard them as giving the holder a
share in
29. Thus Luther: “This is the reason why our theology is certain: because it tears us away
from ourselves and puts us outside of ourselves, so that we may rely not on our own
powers, conscience, sense, person, or works, but on that which is outside of ourselves –
that is, on the promise and truth of God, which cannot lie” (WA 40/I, p. 589, 25–8). But
equally Aquinas: “A human being is much more certain about what he hears from God,
who cannot be deceived, than about what he sees by his own reason, because his reason
can be deceived” (Summa theologiae II–II, 4, 8, ad 2). Therefore, “The believer’s assent to
what belongs to the faith is greater and more stable even than assent to the first principles
of reason” (In I Sent. pro., 1, 3, iii, c). For a discussion see “Faith and Reason Reconsidered,”
1
especially pp. 15–16.
Trinity and
1
God’s own creative knowledge – a veritable God’s-eye view of the
world.30 We who hold these beliefs should therefore regard them as
certain and incorrigible – incapable of turning out to be false – not
because we have evidence which would compel anyone in our
situation to hold the same belief, but because the triune God’s creative
knowledge, in which he gra- ciously gives us a creaturely share by way
of adherence to the church’s epistemic priorities, cannot be deceived or
corrected. To acknowledge conditions under which Christians would
have to give up their own most central beliefs is not, therefore, to
concede that these beliefs are uncertain and corrigible, but to express
the confidence, rooted in the contents of these beliefs themselves, that
these conditions will never be met.
A test case: religious diversity
The election of Israel

Christians are not, of course, likely to be alone in having this sort of
confidence in the truth of their most central beliefs. Other
communities, in particular other religious communities, will no
doubt often have reasons, linked to the contents of their own most
central and identity- forming beliefs, for supposing that these
beliefs cannot turn out to be false. Encounters with the beliefs and
practices of other religious communities pose special problems for a
theological account of how to decide what is true.
One problem is epistemic. Encounters with other religious commu-
nities confront the church not simply with alien and novel beliefs, but
with alternative epistemic priorities. Other religious communities will
order the whole field of possible belief differently than the church
does, around different convictions of unrestricted epistemic scope.
How then
30. Some philosophers regard the notion of a “God’s Eye point of view” as an essentially
humorous one, useful for mocking the pretensions of epistemologies they reject (the
quoted phrase is Putnam’s, Reason, Truth and History, p. 49; see Richard Rorty,
“Wittgenstein, Heidegger, and the Reification of Language,” Essays on Heidegger and
Others,
p. 59). The humor is supposed to lie in the pretense that we can leap out of our
conceptual schemes, our social practices, our skins, or whatever, in order to obtain a
more surely veridical view of the world than these philosophers allow. But in order to
obtain a God’s- eye view of the world we need do nothing so dramatic; we need merely
hold true the narratives which identify Jesus and organize the rest of our beliefs
accordingly. The Holy Spirit’s gift of wisdom thereby gives us “an exhaustive view from
on high, which makes our perspective that of the Trinity” (M. M. Philipon, Les dons du
Saint-Esprit [Paris: Desclée de Brouwer, 1964], p. 137). This is, to be sure, seeing “in a
1
mirror dimly” (I Cor. 13:12), but it is sight nonetheless; we will one day see better, but we
are not on that account now blind.
Trinity and
1
can either community (though our present focus is of course on the
Christian side of the encounter) genuinely learn from the other –
assimilate the other’s beliefs and change its own – without sharing
the other’s epistemic priorities, and thereby losing its own identity?
The encounter of conflicting religious truth claims also poses a
moral problem. This sort of epistemic confrontation never takes place
in a social and political vacuum, and rarely in a situation where power
among the communities is more or less equally distributed. When one
community wields vastly more power than another, it will face the
temptation to mistake its political, economic, military, or even sheer
numeric domi- nance for an epistemic triumph. Assuming epistemic
victory naturally tends to foreclose the possibility of constructive
dialogue between communities, since those who make this
assumption will tend to suppose that they can have nothing to learn
from the putatively defeated community, or, less harshly, that they
have already assimilated whatever there is of value in the other
community’s view of the world. And while we need not, of course,
abuse those from whom we suppose we have nothing to learn, this
supposition does regularly serve as justification for abuse.
The historically dominant attitude of the church to the Jewish
people is a particularly striking case of a larger religious community
taking social and political power over a smaller one for epistemic
triumph. We may usefully reflect, therefore, on whether the church can
learn from the Jewish people about matters at the heart of its faith,
without having to give up its own epistemic priorities – and without
demanding that the Jewish people give up theirs, so that learning
becomes the occasion for epistemic triumph. Should the epistemic
outlook we have developed so far allow for an epistemically more
constructive and morally less proble- matic relationship of the church
with the Jewish people, the plausibility of the account would
presumably be strengthened; should it not, there would be good
reason to find the account unsatisfactory.31
For present purposes we will limit ourselves to brief consideration
of a single topic, the election of Israel.32 Jewish belief in Israel’s
election, to
31.It might be objected that the distinctive relation of the church to Israel, such that
the church claims to worship the same God as the Jewish people, and takes Jewish
scripture as the word of this God, makes the church’s encounter with Israel a poor test
case for the genuinely interreligious consequences of the epistemic proposal for which
this book argues. At least as worshipping communities, however, the church and the
Jewish people appear to regard each other as sharing no members in common. In that
quite basic sense – at the point where their epistemic priorities are most clearly on
1
display – it seems fair to call the relation between the church and the Jewish people an
interreligious one.
32. For a fuller account of some of the issues discussed in this section, see my essays “The
Trinity and
1
summarize roughly, seems to include at least the following basic
elements. (1) God freely chooses Israel – the descendants of Abraham,
Isaac, and Jacob – from among the nations of the world for an enduring
relationship with himself, one both uniquely intimate and intended
for the ultimate good of all the nations. Israel responds to God’s
election, but does not originate it. (2) God gives the Torah to Israel;
its obligations provide the chief content of this unique relationship,
and their fulfill- ment is central to its aim. The gift of the Torah is thus
inseparable from, but not identical with, Israel’s election. (3) God’s
choice of Israel and his revelation of the Torah to her distinguish Israel
from the nations.33
Belief in Israel’s election is not only among the historically central
doctrines of the Jewish community, but is clearly regarded by some
Jewish thinkers as having unrestricted epistemic primacy. Michael
Wyschogrod, for example, argues that the most basic feature of Jewish
identity is neither a distinctive system of beliefs nor the observance of
a distinctive legal code, however indispensable both are. Rather, “the
foundation of Judaism is the family identity of the Jewish people as
the descendants of Abraham, Isaac, and Jacob. Whatever else is added to
this must be seen as growing out of and related to the basic identity of
the Jewish people as the seed of Abraham elected by God through
descent from Abraham.”34 Thus “the most important part of the
whole [of Judaism] is the existence of the Jewish people as the earthly
abode of Hashem,” such that “everything else must be seen in this
light. Only because it is true is everything else true.”35
Remarks of this kind suggest that in Judaism, at least as construed
by a Jewish thinker like Wyschogrod, belief in Israel’s election is the
primary criterion of truth.36 When it comes to deciding about truth,
God’s unshakable electing love for Israel forms that conviction within
the open
Jewish People and Christian Theology,” The Cambridge Companion to Christian Doctrine, ed.
Colin Gunton (Cambridge: Cambridge University Press, 1997), pp. 81–100, and “Truth
Claims and the Possibility of Jewish-Christian Dialogue,” Modern Theology 8/3 (1992), pp.
221–40.
33. This characterization is drawn in particular from David Novak, The Election of Israel
(Cambridge: Cambridge University Press, 1995).
34. Michael Wyschogrod, The Body of Faith: God and the People Israel, 2nd edn (Northvale, New
Jersey: Jason Aronson, 1996), p. 57. 35. The Body of Faith, pp. 223, 118.
36. Or at least one of the primary criteria. David Novak argues that Wyschogrod’s account
implicitly subordinates Torah to election, whereas the two ought to be regarded as equally
basic to Jewish identity and to Judaism; see The Election of Israel, pp. 241–8. For our
purposes this dispute is not decisive, since both locate election among Judaism’s primary
beliefs. In modern times some Jewish thinkers have, to be sure, argued that the Jewish
people and the Jewish religion would be better off without the traditional doctrine of
Israel’s election; for a critical treatment of one such proposal from a Jewish perspective,
1
see David Novak, “Mordecai Kaplan’s Rejection of Election,” Modern Judaism 15 (1995), pp.
1–19.
Trinity and
1
field of possible beliefs which the faithful Jew is most unwilling to give
up or reinterpret, and correlatively that with which all other belief and
practice must at least be consistent in order to be held true or regarded
as right.
This way of understanding the epistemic structure of Judaism dis-
plays an obvious similarity to the account of Christianity’s basic epis-
temic structure proposed here. In each case unlimited epistemic
primacy is ascribed to a complex body of beliefs which identifies a
particular – the Jewish people as the object of God’s electing love and
his dwelling place in the world; the resurrection and exaltation of the
crucified Jesus of Nazareth as the self-identification of the triune God
for the world.37
This structural affinity highlights, of course, a criteriological dis-
agreement. If Wyschogrod and I are right about our respective
communities, Jews and Christians both assign unrestricted epistemic
primacy to descriptions of particulars. But we ascribe this decisive
epistemic role to descriptions of different particulars. Once adequately
picked out and described, these two particulars – Abraham’s descendants
as God’s elect, Jesus crucified and risen – cannot be mistaken for one
another; they are manifestly discernible (fail to have all the same
properties) and therefore not identical. Since descriptions of two
different particulars are epistemically primary for the two
communities, the epistemic conflict between these communities seems
irreducible: there is no chance that, so long as each survives, the two
communities will agree on the decisive criteria by which to evaluate
beliefs – including, of course, each other’s beliefs.
Does this mean that the two communities are fated never to learn
from, and perhaps never even to understand, one another? Or for
present purposes, does ascribing unrestricted epistemic primacy to the
narratives which identify Jesus foreclose the possibility that
Christians could learn from Jews, even about matters at the heart of the
Christian faith, and by the same token raise the specter of abuse
directed at those from whom we cannot learn?
Conceiving religious diversity

Recent attempts, theological and otherwise, to cope conceptually with
the plurality of religions have regularly tended to suppose that the
answer to these questions must be yes. If different religious
communities are genuinely to learn from one another, such that one
comes to share
1
37. Wyschogrod is aware of this affinity, and takes Karl Barth as his model for a
Christian theology which displays this sort of epistemic structure. See The Body of Faith,
pp. 75–81.
Trinity and
1
beliefs important to another, they cannot finally decide about truth in
disparate ways, still less by appeal to descriptions of different
particulars. Rather they will already have to share not only many of the
same beliefs, but at least implicitly the same epistemic priorities.
This point is sometimes put as a need to avoid religious
“exclusivism.” To ascribe unrestricted epistemic primacy to
descriptions of a particular, and with that ultimate significance to the
particular so described, is eo ipso exclusivistic. It requires those who
make such an ascription to regard those who do not as either ignorant
or mistaken about ultimate matters. Exclusivism is at once
epistemically overcommitted and morally proble- matic; it embodies,
so the argument goes, just the sort of attitude which leads people to
suppose that others cannot teach them anything.
The suggestion that different religious communities in fact share the
same epistemic priorities can be developed in a variety of ways.38 It might
be argued that underlying their apparent diversity of belief and
practice, all religions finally have the same objective aim – “God,”
“reality,” or “the transcendent,” for example. Or it might be argued
that all religious belief and practice expresses the same basic
experience – of “love,” for example, or, once again, “God,” “reality,” or
“the transcendent.” Some versions of each approach hierarchize the
various religions according to the relative adequacy with which they
lead toward the common aim or express the common experience.
These tend to identify the aim or experience in relatively definite and,
in the case of theological versions, predominantly Christian, terms.
Other versions resist hierarchy and attempt to interpret all religions as
equally adequate, although only partial, accounts of the common aim
or expressions of the common experience. These tend to identify the
common aim or experience in relatively vague terms, such as “reality,”
or “the transcendent.” Epistemic primacy belongs, so these accounts
variously imply, to those beliefs which locate the common aim or
express the common experience. The assumption that this aim or
experience is shared by all religions serves as warrant for the claim
that they all at least implicitly have the beliefs which locate it, and
indeed at least implicitly regard them as primary in decisions about
truth.
Such strategies for coping with religious diversity need not rule out
the possibility that the shared beliefs to which epistemic primacy
belongs
1
38. For an argument in support of the typology briefly sketched in this paragraph, see J. A.
DiNoia, O. P., The Diversity of Religions: A Christian Perspective (Washington: CUA Press,
1992). The approaches I here characterize as hierarchical DiNoia calls “inclusive,” those
which resist hierarchy he labels “pluralistic.”
Trinity and
1
identify and describe a particular. But one can see why the proposals
they make for the common religious aim or experience generally fail to
include any identification and description of particulars. These propo-
sals have to render plausible the suggestion that diverse communities
not only share (at least implicitly) various relevant beliefs, but (at least
implicitly) regard these shared beliefs as epistemically primary. Beliefs
whose contents identify and describe particulars (God’s elect people
Israel; Jesus crucified and risen) are those which different religious
communities are most evidently prone not to share, let alone jointly
regard as epistemically primary. So a search for common criteria of
truth will naturally tend to settle on beliefs whose content is abstract
or indeterminate enough to make credible the claim that they are
shared. The reference of the term which is supposed to locate the aim
of the common quest or the content of the common experience will be
correspondingly vague. Such expressions may be usable as singular
terms (as is “the transcendent,” or even “reality” or “mystery,” which
are singular in the way “water” or “blue” is singular), but they will
almost surely not enable us to identify any particular (even singular
terms whose reference is entirely clear may fail to support the
identification of a particular, like “the first dog to be born at sea”).39
These approaches rightly strive to account for the possibility that
one religious community could come to share beliefs held by another.
It is not clear, though, that they actually succeed.
In various ways these approaches reject the idea that descriptions of
particulars can be epistemically primary in religious belief systems,
and instead interpret these systems so that referring terms of a vague
or general kind, with more or less rich descriptions attached, take on
this epistemic role. We locate a community’s epistemic priorities,
however, by finding out which beliefs that community is least willing
to give up or modify. Proposals about a community’s epistemic
priorities are thereby proposals about its identity. If descriptions of
different particulars are epistemically primary for the Jewish and
Christian communities, then interpretations of Judaism and
Christianity which seek to dislodge these descriptions from their
epistemic place are proposals for basic changes in the identity of the
two communities. But this seems an odd way to think about a
mutually instructive dialogue between the two communities. It would
require Jews to stop being Jews, and Christians to stop being
39. To use P. F. Strawson’s example of a “pure individuating description”; see Individuals
(London: Methuen, 1959), p. 26.
1
Christians, in order for a dialogue to take place; a Jewish–Christian
dialogue would not be the result. This seems not so much to account
for the possibility of genuine dialogue between the two different
traditions as (borrowing a metaphor from Donald Davidson) to
arrange a shotgun wedding between them – one which, moreover,
does not marry the original mates.
To this kind of objection it might, of course, be replied that descrip-
tions of particulars are not in fact central to the belief systems of the
Jewish and Christian communities (or of other historic religious tradi-
tions). The sort of interpretation which wants to centralize such
descriptions, rather than that which aims to move them to the
periphery, actually distorts the identity of the communities involved.
Even were such suggestions empirically plausible, they would not
help with the matter at hand. Decentralizing descriptions of particulars
in the belief systems of religious communities does not solve the
problem of finding a morally tolerable approach to epistemic
encounters between them. It defines the problem out of existence, and
then declares it solved. By suggesting that religious communities
which share no members nonetheless have, at least implicitly, the same
primary criteria of truth, this approach in effect proposes that religious
diversity is epistemically superficial: when it comes to criteria of truth,
there is no warrant for the existence of separate communities (there
may of course be warrants of other kinds). This thought fails to explain
how we can learn from those with whom we have genuinely basic
disagreements, religious or otherwise – in particular, disagreements
about which beliefs are epistemically primary. The assumption, on the
contrary, seems to be that we can only come to share the beliefs of
those with whom we already agree on the most basic matters, those
who are basically like us, at least when it comes to beliefs. Other
communities with whom we assume agreement may, of course, fail to
realize how little difference there is between us. They may actively
resist the suggestion that they already share our beliefs – a hazard of
assigning to them implicit beliefs, convictions they are unaware of
having, and which they may see as inconsistent with their own explicit
epistemic priorities.
Nor does epistemically decentralizing descriptions of particulars
succeed in avoiding religious “exclusivism.” Any belief system struc-
tured by epistemic priorities of unlimited scope is the same size as any
other. Each includes the same totality of possible belief; worldviews
differ not in being more or less inclusive, but in assigning truth values
Trinity and
2
differently across the totality of possible belief which they already
share. Religious belief systems which ascribe epistemic primacy to
descriptions of a particular, like Judaism and Christianity, admit in
principle no more or fewer beliefs, are no more inclusive or exclusive,
than those which do not – however differently they assign truth values
in individual cases. And to assign truth values differently than others
is to regard others as mistaken. If this is exclusivism, outlooks which
epistemically decentralize descriptions of particulars no more
succeed in avoiding it than those with highly particularistic priorities.
Error about matters of ultimate significance is still attributed to some,
and not to others; we differ only as to whom.
We should perhaps ask, therefore, whether a religious community
which grants unrestricted epistemic primacy to descriptions of a
particular is in fact incapable of acquiring beliefs from another
religious community. In the present case: is it possible after all for the
church to learn from Jews to believe in the election of Israel?
Christological inclusion of Israel’s election

The Christian community now widely, and in some cases officially,
teaches the permanence of Israel’s election.40 Repudiating the claim
that God has rejected or abandoned the Jews has evident moral
implications; it counts as a strong reason (though of course not the only
one) for oppos- ing the church’s historic hostility toward the Jewish
people, and for changing Christian practice accordingly.
Through most of its history the Christian theological mainstream
had, of course, affirmed God’s election of Israel and his gift of the
Torah to her, but had maintained that both the election and the gift
were tempo- rary. The church, so the historically dominant view goes,
has superseded
– taken the place of – Israel as God’s elect, and sacramental worship
has, roughly speaking, taken the place of the Torah in structuring the
basic relationship between God and his chosen people. A genuinely
post- supersessionist understanding of election in Christian theology is
therefore one which affirms the permanence of Israel’s election, and
with that
40. Thus Vatican II: “According to the Apostle the Jews remain most dear to God on
account of their fathers, since God does not repent of his gifts or his call [see Rom. 11:28–
9]” (Nostra Aetate, § 4; see Lumen Gentium, § 16). Norman Tanner, S. J., ed., Decrees of the
Ecumenical Councils, 2 vols. (London: Sheed and Ward, 1990), vol. II, pp. 970, 861. For a
collection of Protestant statements see The Theology of the Churches and the Jewish People
(Geneva: WCC Publications, 1988). For more on some of the issues discussed in the next
2
several paragraphs, see “The Jewish People and Christian Theology.”
Trinity and
2
the permanence of the divinely willed distinction between Jew and
Gentile. The rallying point for this renewed understanding of Israel’s
election in Christian theology has been Rom. 9–11, where Paul’s effort
to come to grips with the relationship between Israel and the church
culmi- nates in a vigorous affirmation of God’s unshakable love for and
fidelity to Abraham’s children (see Rom. 11:28–9). Indeed Paul seems to
have a doctrine of Israel’s election so strong that it might not be
possible to conceive a stronger one. On his account it seems impossible
that Jews can stray so far as to lose their election, since those Jews who
have already done what Paul takes to be the worst thing they can do –
reject the Messiah – are nonetheless “beloved for the sake of their
forefathers” (Rom. 11:29).
However, just as there were many ways of being a supersessionist
in Christian theology, so there are many ways of not being one. In
order to get beyond supersessionism, some theologians argue that the
Christian community has to rearrange its long-standing epistemic
priorities. The church has to give up, in particular, the conviction of
Jesus’ universal primacy, and with that the epistemic priority of the
Trinity. Perhaps the most obvious way to make this point is to argue
that Israel’s God has established two covenants or saving
arrangements in the world, one for Jews (through election and Torah)
and the other for Gentiles (through Jesus).
Though stemming from understandable motives, this adjustment
of epistemic priorities looks like an unpromising strategy for post-
supersessionist Christian theology. Any religious community will
naturally resist the suggestion that it give up its most central, identity-
forming convictions, and will likely endure even discomfiting associa-
tions of these convictions (in this case, supersessionism) if it sees
rejection of its own most central beliefs – and with that infidelity to its
God – as the only alternative. So a successful post-supersessionist
theology will have to show that belief in Israel’s permanent election is
at least compatible with, and if possible more strongly implicated in,
the unrestricted epistemic primacy of the narratives which identify
Jesus, and thereby in the doctrines of the incarnation and the Trinity.
Here there is room for only two brief suggestions. Belief in the
permanence of Israel’s election, and so in the permanence of the
distinction between Jew and Gentile, surely seems at least compatible
with belief in the universal primacy of Jesus of Nazareth, and so with
the epistemic primacy of the narratives which identify him. Nothing
about the one belief manifestly contradicts the other, requiring us to
2
choose between
Trinity and
2
them, though a consideration of the complex entailments of each set of
beliefs would be necessary in order to see if this appearance of
consistency were borne out in the long run. Paul, at any rate, evidently
thinks the two convictions are compatible, and so commits those who
regard his writings as scripture to find ways of spelling out their
compatibility.
In so doing, it ought to be possible to make a case for a tighter
relationship between Jesus’ unrestricted primacy and the election of
Israel than mere compatibility. According to the traditional Christian
doctrine of incarnation, for example, in the person of the Logos God
makes his own the flesh of the particular Jew, Jesus of Nazareth. God’s
ownership of this Jewish flesh is permanent. In the end, when all flesh
shall see the glory of the Lord, the vision of God will, so the
traditional Christian teaching goes, be bound up ineluctably with the
vision of this Jew seated at God’s right hand. So in willing his own
incarnation, it seems that God wills the permanence, indeed the
eschatological permanence, of the distinction between Jews and
Gentiles. But Jesus cannot be a Jew, or be identified as such (as he will
be even in the eschaton), all by himself, in isolation from his people.
He is a Jew, like any other, only in virtue of his descent from Abraham,
and thus in virtue of his relationship to the Jewish people as a whole.
And this suggests that in owning with unsurpassable intimacy the
particular Jewish flesh of Jesus, God also owns the Jewish people as a
whole, precisely in their distinction from us Gentiles; he cannot own
the one without also owning the other. The two forms of ownership
are not identical – the one involves nothing short of union, the other
something like indwelling – but neither are they totally disparate. As
both Jewish and Christian theologians have sometimes observed, the
Christian doctrine of incarnation is an intensification, not a
repudiation, of traditional Jewish teaching about the dwelling of the
divine presence in the midst of Israel.41
In the end this Christian teaching clearly says more than the Jewish
doctrine of Israel’s election, and more than Jews can accept. But for
present purposes the key point is that the Christian doctrine of God’s
incarnation does not say less than the Jewish belief in the eternal
election of Abraham’s children. That the Christian community could
come to share this central Jewish belief, and thereby change its own
historic teaching, may readily be understood in light of the foregoing
account of the way communal belief systems may assimilate novel or
alien claims.
2
41. For some penetrating observations on this point from the Jewish side, see Wyschogrod,
The Body of Faith, pp. 211–15.
Trinity and
2
Since a community’s primary criteria of truth will generate
relatively few of the reasons the community and its members will have
for holding beliefs, a viable communal belief system will be one which
allows those who hold it to take in novel or alien claims when they
have reason to do so, yet without giving up their own epistemic
priorities. In the present case, however, the bond is much closer than
this. Israel’s permanent election seems not simply to be compatible
with the Christian community’s epistemic priorities, but actually to
be required by them. The Christian rejection of supersessionism
seems like a case of an initially alien claim – one central to another
community but rejected by the church – which the Christian
community finds it has good reason not only to assimilate, but to
locate near the center of its own belief system.
God’s faithfulness to his election of the Jewish people is not, of
course, exactly a novel claim for Christians; belief in it goes deep
enough into the community’s system of beliefs to have not only
textual, but scriptural fixity. A soon largely Gentile church more or
less thought it could reject this belief without sacrificing plausible
interpretation of its own scriptural text. The claim had to become
novel again in order for the church to interpret its scriptures and order
its belief system more coherently – in order for the church to have
truth commitments on this matter which matched its epistemic
priorities.

SplitPDFFile 124 To 195

Uploaded by

Copyright:

Available Formats

SplitPDFFile 124 To 195

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

SplitPDFFile 124 To 195

Uploaded by

Copyright:

Available Formats

5

The epistemic primacy of belief in the Trinity

108 An investigation of its practices has suggested that the church

Jesus’ universal primacy

Jesus undertakes his journey from Bethlehem to Golgotha, so the

7. On this see the discussion in chapter 9, pp. 251–8.

Epistemic right as christological coherence

What then of deciding about truth? We undertook this exploration of

Identiﬁcation of Jesus as epistemic trump

Unrestricted epistemic primacy

19. On this see chapter 6.

Epistemic primacy and highest truth

The Father’s epistemic role

Epistemic priorities, truth commitments, and plausible

The unrestricted epistemic primacy of those beliefs which are most

Interpreting Jesus’ resurrection

Epistemic priorities and alien claims

Fideism and epistemic priorities 141

The suggestion that the church’s trinitarian identiﬁcation of God

1. This way of putting the point is Wittgenstein’s; see On Certainty, ed. G. E. M.

3. See again Wittgenstein, On Certainty.

Inclusion and assimilation

The success of the Christian community’s belief system at coping with

Encounters with novelty

9. For analysis of one example of this procedure – Aquinas’s assimilation of

The epistemic force of inclusive power

But why should success at including alien beliefs and assimilating

The usefulness (as distinguished from the logical force) of inclusive

25. A communal tradition, as MacIntyre argues, can undergo an “epistemological

27. Which I owe to a suggestion of Michael Root.

A test case: religious diversity

The election of Israel

Conceiving religious diversity

Christological inclusion of Israel’s election

You might also like