^[1] In recent discussion in ethics, a number of writers have sought to combine ‘cognitivism’ (the idea that values are genuinely ‘in the world’) and a recognition of the affectively toned character of much ethical experience. In this paper, I wish to consider whether John McDowell’s formulation of this view may be fruitfully generalised to the case of religious experience. In this way, I shall sketch a case for the idea that certain religiously significant values may be discerned in our affective experience and not otherwise. This question is of some interest because so much religious experience is, evidently, shot through with affective concerns. If there is a non-projective way of treating the affective dimension of religious experience, that would be of some importance therefore for an understanding of the character of such experience in general.

^[2] I shall begin by setting out McDowell’s proposal, before considering how his approach may complement that of two writers who are explicitly concerned with affectively toned religious experience, namely, William Alston and John Henry Newman. I shall then outline some objections to McDowell, derived from Simon Blackburn, and consider their force for an account of religious experience informed by the work of McDowell, Newman and Alston.

^[3] In his paper ‘Non-Cognitivism and Rule-Following’, John McDowell notes one kind of argument for non-cognitivism in ethics, one which postulates a parallel between our experience of values and of colours. [1] On this view, just as we can explain our colour experience by reference to qualities in the world which are themselves colourless (the ‘primary qualities’ of things), so we can explain our value experience by reference to qualities in the world which are themselves value-free. The conclusion to draw, so the argument goes, is that values, like colours, are not part of the fabric of the world; instead, they reflect not so much the character of the world as the character of the mind, and its way of representing to itself the nature of the world.

^[4] McDowell opposes this line of argument by challenging the distinction it seeks to draw between the element of value experience that can be attributed to a set of value-free qualities that are in the world, and the element that reflects the human subject’s contribution, its glossing of the world in the light of its needs and concerns. Perhaps it is possible to draw some such distinction in relation to colour experience, as when we suppose that light of a certain wavelength (where wavelength is understood in quantitative, colour-independent terms) gives rise to a certain kind of colour experience (seeing red, say). But, McDowell suggests, there should be no expectation that we can match up in the same sort of way value-neutral qualities in the world and various kinds of value experience. And in that case, this particular argument for non-cognitivism about values will fail, since the argument depends upon the idea that because some such pairing off is possible, we can trace our value experience to qualities in the world which are value-neutral. [2]

^[5] McDowell does not present any simple knock-down argument for the claim that we cannot match up value-free qualities in the world and kinds of value experience in the way required by this sort of argument for non-cognitivism. Instead he argues as follows:

^[6] Here McDowell is suggesting, I think, that the burden of proof in this debate rests on those who subscribe to the possibility of the ‘disentangling manoeuvre’. For if such disentangling were possible, then we would be able to grasp the extension of value terms independently of any appreciation of the very ‘concerns’ which give rise to the use of those terms, and why think that is at all likely? The thesis of the paper is then that arguments for non-cognitivism which depend on appeal to the disentangling manoeuvre fail to assume the requisite burden of proof.

^[7] An example of a quality which may fit McDowell’s proposal is that of being funny or amusing. This quality seems to differ from qualities such as being in motion or being hot in so far as it cannot be specified independently of human reactions. [4] Moreover, it also seems to differ from colour properties, such as the property of being red, even if we suppose that such properties cannot be specified independently of human subjective experience. For we do not have a ready way of grouping all the things that are funny independently of their tendency to evoke such a response, whereas we do have a ready way of grouping all the things that appear red independently of their tendency to evoke this response (by reference to the fact that these things reflect light of a certain wavelength). So to put the matter in McDowell’s terms, whereas the term ‘red’ has an extension which can be picked out in colour-neutral terms, the extension of the term ‘funny’ cannot be specified without reference to our responses of being amused.

^[8] These remarks are enough, I hope, to indicate in summary form the nature of McDowell’s proposal, and to provide some sense of why he thinks it worthy of belief: on this view, values should not be seen as the result of the mind’s glossing of qualities which are themselves value-free; instead, we should suppose that at least some values are genuinely ‘in the world’, and can be recognised by way of affectively toned evaluative experience. This proposal calls for further defence and clarification, and I shall return to various objections to McDowell’s account in the final section of the paper, but first I want to consider how that account might be extended to the case of religiously informed value experience. For this purpose I shall consider in turn the work of William Alston and John Henry Newman.

^[9] Both these authors (and especially Newman) acknowledge the affectively toned character of much religious experience, and this suggests an initial point of connection with McDowell’s account of value experience. For although McDowell does not say much explicitly on the point, it is clear that he is thinking of value experience in affectively laden terms. Hence he can write of the possibility that ‘we can learn to see the world in terms of some specific set of evaluative classifications, aesthetic or moral, only because our affective and attitudinative propensities are such that we can be brought to care in appropriate ways about the things we learn to see as collected together by the classifications’. [5] So here is an initial connection between McDowell’s account and those of Newman and Alston. Next I want to consider more closely the possibility that these various approaches can be brought into fruitful relationship.

^[10] In his book Perceiving God, William Alston examines what he calls ‘mystical perception’ or (equivalently) ‘direct perception of God’. In general, if one directly perceives X, then ‘one is aware of X through a state of consciousness that is distinguishable from X, and can be made an object of absolutely immediate awareness, but is not perceived’. [6] (Hence on this account, my awareness of the computer before me now is a case of direct perception.) In the course of his discussion, Alston considers the possibility that the state of consciousness through which we directly perceive God is purely affective in terms of its phenomenal content. He writes:

^[11] So while he allows that the phenomenal content of a genuine perception of God may be purely affective, Alston regards this possibility as a source of ‘nagging worry’. Why should he think of the possibility in these terms? At the beginning of the passage I have just cited, Alston characterises the affective dimension of such an experience as ‘various ways the subject is feeling in reaction to what the subject takes to be the presence of God’. It is striking that this formulation assumes that the feeling content of a mystical perception is a ‘reaction’ to (what is presumably) a feeling-neutral thought. On this view, it seems that feelings are being construed as rather like sensations (such as the sensation of being bruised), in so far as they do not themselves bear any intentional content (they are not about anything), albeit that they differ from sensations in so far as they are occasioned by a thought, rather than by the impact of an object upon the sense organs. Given some such conception of feelings, it is then easy enough to subject affective experience to the sort of disentangling manoeuvre that is envisaged in McDowell’s discussion: just as colour experience can be explained by reference to colour-neutral items in the world, so affective experience (on this view) can be explained by reference to the affect-neutral thoughts from which it derives. Now McDowell’s discussion appears to invite a rather different characterisation of the role of feeling, as I shall now argue.

^[12] We have seen that on McDowell’s view value experience should not be disaggregated into a value-neutral element that derives from ‘the world’ and a phenomenal element that reflects the mind’s contribution to the experience. Instead, we should suppose that the object of such an experience can only be understood in value-indexed terms. So on this view, there are values ‘in the world’. Moreover, as we have seen, on McDowell’s account, it is by way of our affective responses that we come to recognise these values. As Simon Blackburn puts the matter, on McDowell’s view, ‘our affective natures expand our sensitivity to how things are, on the lines of any mode of perception’. [8] This suggests a model according to which feelings are ways of taking stock of (evaluative) features of the world, and to that extent, themselves forms of thought. Indeed, it may be that our felt responses offer our only mode of access to certain values (just as in certain cases, our amused responses may offer our only mode of access to the quality of being funny).

^[13] If this is the right way to read McDowell, then his account does indeed pose a challenge to the model of affect that is implied in Alston’s remarks. For on McDowell’s picture, feelings are being represented as thoughts (in the sense of having intentional content, or being about something) in their own right, and not simply by virtue of their association with some thought by which they are caused. By contrast, as we have seen, on Alston’s account, feelings seem to be represented as in themselves thought-less, and as occasioned by (feeling-less) thoughts. It is, I suggest, this rather impoverished account of affect that leads Alston to remark that: ‘Our inability to specify any other sort [i.e. some non-affective sort] of non-sensory phenomenal qualities leads naturally to the suspicion that the experience is confined to affective reactions to a believed presence’. This suspicion is only ‘natural’, I suggest, given the assumption that affectively informed experiences can be (and in general ought to be) disaggregated into a thought component and a feeling component; and McDowell’s account challenges this distinction. So here is a first point where McDowell’s discussion proves to be relevant to Alston’s account of religious experience. If we adopt McDowell’s conception of affective experience (rather than Alston’s), it will be easier to see how a religious experience whose phenomenal content is purely affective may, even so, be veridical. Again, this is a matter of some significance, given that religious experience is so often infused by affective concerns.

^[14] In the passage we have been discussing, Alston seems to concede that there is rightly some initial scepticism about the trustworthiness of a mystical perception whose phenomenal content is purely affective. However, he goes on to suggest that such an experience could in principle be veridical after all, and in support of this claim he cites an analogy with sense perception: ‘even if, as seems possible, sensory phenomenal qualities are as subjective as affective qualities, that does not prevent them serving as a phenomenal vehicle of the perception of objective external realities.’ [9] And in that case, Alston asks, could we not suppose similarly that affects may serve as the ‘phenomenal vehicle’ for the recognition of mind-independent realities? Does this proposal suggest a more generous assessment of the role of affects in theistic experience? Here Alston does seem to allow that feelings may have intentional content: a recognition of the character of ‘objective reality’ can be realised in affective experience. However, a McDowell-inspired view might still be wary of Alston’s analogy, on the grounds that it assimilates sensory and value experience too quickly, and fails to draw attention to the distinction between, for example, colour and value experience: unless this distinction is noted, we might find ourselves allowing that affects have intentional content (just as our phenomenal-colour-informed experience of colour has intentional content), while failing to allow that affects represent features of ‘the world’. However, in fairness to Alston, he does indicate that he intends the analogy to apply with reference to the ‘perception of objective external realities’. Even so, while admitting the possibility of affects playing this sort of role, it is striking that he persists in trying to downplay them (as we shall see again shortly). This suggests to me that while Alston sees no objection of principle to this reading of the significance of affects, he thinks that in practice the model of affects as occasioned by thoughts and as themselves thought-less is truer to our experience (or preferable for some other reason). In that case, we might take McDowell’s account as a helpful corrective to the idea that, in standard cases, our evaluative experience should be disaggregated into a thought component and a further, affective component which derives from the thought component.

^[15] There is one further element of Alston’s account that comes into new focus, I think, when viewed through the lens afforded by McDowell’s discussion. Once again, Alston’s observations at this juncture have as their focus a perceived difficulty in the notion of mystical perception. He writes:

^[16] In this passage, Alston is responding to an (implied) objection to the trustworthiness of mystical perception. The objection maintains that there is a significant disanalogy between mystical perception and sensory perception, because only in the case of sensory perception do we have a language for recording the phenomenal content (or more exactly, ‘the basic phenomenal qualities’) of the experience. This disanalogy poses a threat to Alston’s project, I take it, in so far as it implies that religious experience may lack any (coherent) phenomenal content; for it that were so, then we might doubt whether such experience is really experience at all (or at any rate, experience which relates us to a coherent, objective reality). Interestingly, on this point, Alston’s reply takes the form of allowing the disanalogy, and seeking to explain it. (Contrast the strategy he uses in response to the thought that the phenomenal content of mystical perception may be purely affective: here he appeals to a point of similarity between mystical and sense perception.) [11]

^[17] In the passage just cited, Alston comments: ‘We know quite a bit about the ways in which sensory experience depends in a regular way on its physical, physiological, and psychological conditions.’ Now if McDowell’s account is on the right lines, then we might suppose that value and sensory experience are dissimilar in this respect. For while we can construct general correlations between states of the physical world and the sensory experiences to which those states are likely to give rise, we cannot (so easily) correlate states of the physical world (picked out in scientific or value-neutral terms) and the value experiences to which they are likely to give rise. Might this fact help to explain the difficulties of constructing a language to record the phenomenology of religious experience?

^[18] No doubt, there are quite a number of considerations which help to explain the difficulty of constructing such a language. Many commentators have thought (reasonably enough) that God is not an ‘object’ available for manipulation in the way that the objects of sense experience are, and we might suppose that it is for this reason that we have less opportunity to subject theistic experience to examination under controlled conditions. The consideration introduced by McDowell provides a further perspective on this idea. Given his remarks, we might suppose that the difficulties in providing a language for the phenomenology of religious experience reflect not just the fact that God is not a manipulable, spatio-temporally locatable object, but also the fact that religious experience is a kind of value experience. For value experiences cannot be mapped onto sets of physical conditions which constitute natural classes when characterised in scientific or value-neutral terms; and accordingly, it is a relatively complex matter to identify the physical stimulus conditions for a given kind of value experience.

^[19] Does the McDowell-style explanation add much to Alston’s? In the passage just cited, Alston envisages this possibility: perhaps the stimulus conditions for mystical perception ‘have to do with God’s purposes and intentions, and if so that gives us absolutely no handle on prediction and control’. Here Alston may be implying that at least on occasions, God brings about mystical perceptions miraculously, so that there is no possibility even in principle of pairing off such perceptions with stimulus conditions characterised in physical terms. But suppose we drop this assumption, and allow that if God brings about a mystical perception of himself in state of the world A, then he will bring about such a perception in a qualitatively indistinguishable state of the world B. Now on McDowell’s account, we have reason to think that, even under these conditions, it will still be relatively difficult to construct a language to record the phenomenology of religious experience. This is because even given this relationship of supervenience (whereby sameness of physical state implies sameness of mystical perception), there will still be no simple way of mapping mystical perceptions onto stimulus conditions, where these stimulus conditions are seen to constitute a natural class when identified in value-neutral terms. [12] And in that case, it will be difficult to move from one example of a stimulus condition for mystical perception to some more general class of conditions, knowledge of which would enable us to pin down the phenomenal character of mystical perception in a range of circumstances.

^[20] So an appeal to the evaluative character of theistic experience seems to constitute a different consideration from that cited by Alston and can therefore play a distinct, though potentially complementary, role within an account of the indescribability of the phenomenal content of theistic experience. This McDowell-inspired approach also constitutes at least a partial response to another, more standard objection to religious experience (an objection pressed by Anthony O’Hear and others). [13] This is the objection that such experiences are not epistemically reliable because they cannot be predicted. But the unpredictability of religious experiences, on the McDowell-style view, may reflect not so much the fact that they are ‘merely subjective’ (O’Hear’s conclusion) as the fact that they cannot be correlated with a set of easily circumscribed physical stimulus conditions. And anyone who wishes to be a (McDowell-style) cognitivist about, say, moral experience may be hard-pressed to explain why this fact alone should call into question the reliability of religious experience.

^[21] We might wonder whether these reflections throw any light on the comparative difficulty of providing a language to record the phenomenology of religious experience as compared with say moral or aesthetic experience. After all, these are all cases of value experience, so on the hypothesis under consideration, should we expect the phenomenology of each to be equally difficult to capture linguistically? I do not think this does follow, for there is some reason to suppose that the kind of value experience that is relevant in religious contexts makes it especially difficult to undertake the mapping of experiences onto value-neutral stimulus conditions. (So perhaps a revised version of O’Hear’s objection could be lodged here; but at any rate, this will have to be a new, more nuanced version of the objection.) After all, it is relatively easy to specify, in value-neutral terms, at least some of the conditions that are relevant to classifying an action as kind or cruel. But it is much more difficult to specify, in value-neutral terms, the circumstances which present proper stimulus conditions for a mystical perception. Even if the supervenience thesis holds for such experiences, it is very difficult for us to see how to move from one or two examples of acknowledged stimulus conditions to a larger class of stimulus conditions; by contrast, it is relatively easy to move from one or two examples of the stimulus conditions for, say, cruelty (where these conditions are picked out in value-free terms) to at any rate a class (one of several, we might suppose) of stimulus conditions that are relevant to this kind of (dis)value experience. This suggests that while an explanation of the difficulties in describing the phenomenal content of theistic experience can usefully deploy the idea that such experiences are evaluative in character, this idea will not provide a comprehensive explanation. For that purpose, further considerations will need to be adduced (addressing the question of why the mapping of value experiences on to stimulus conditions should prove particularly difficult in the case of theistic value experience).

^[22] In concluding my discussion of Alston, I want to return to one final strand of his case where, again, he is seeking to downplay the thought that the phenomenal content of theistic experience is purely affective. Having allowed the in-principle possibility of veridical theistic experience whose phenomenal content is purely affective, Alston goes on to try to show that this possibility need not be allowed as a matter of fact. His interest in sustaining this idea shows, I suggest, his ongoing reservation about the role of feeling in religious experience. Hence he writes: ‘To further shore up the supposition that mystical perception involves distinctive, nonaffective phenomenal qualia, we can advert to the doctrine of “spiritual sensations” that was developed in the Catholic mystical tradition’. [14] According to this tradition, as Alston expounds it, there are certain spiritual experiences which (although nonsensory) in some way resemble touch, and others which resemble taste or smell, and so on. [15] However, the examples Alston gives in support of his thesis appear to retain a strongly affective character. For instance, here is an excerpt from the passage he cites to illustrate the idea of a mystical perception which is reminiscent of the experience of smell (the passage is taken from St Teresa’s Interior Castle): ‘Understand me, the soul does not feel any real heat or scent, but something far more delicious, which I use this metaphor to explain.’ [16] The thought that references to ‘heat’ or ‘scent’ in this connection are strictly metaphorical, and that such experience is ‘far more delicious’ than any mundane experience, suggests that the quality in question is registered, at least in significant part, through an experience of delight, which implies that affects remain, to say no more, of fundamental importance to the phenomenology of the experience.

^[23] Whether or not this is the right way to read Teresa and others who appeal to analogies between spiritual experience and the various modalities of sensory experience, it remains true, I suggest, that ordinary folk who report religious experience (as distinct from elevated mystics such as Teresa) are more likely to describe their experiences in affective terms than in terms which suggest a non-affective, non-sensory ‘intuition’ of the divine. [17] Alston would no doubt reply that even if this is so, it may just reflect the relative poverty of our language for describing non-sensory, non-affective experience (see again his comments above). Nonetheless, if the subjects of theistic experience are heavily reliant upon reference to feelings to describe such experiences, that suggests that theists would do well to adopt a rather more sympathetic attitude than Alston towards the possibility that the phenomenal content of certain religious experiences is in large part (if not entirely) affective. (Again, it is of course this thought that provides the rationale for this paper.) Moreover, quite apart from these sociological considerations, the notion of a non-sensory, non-affective mode of intuition may well be more difficult to understand for philosophical reasons than the idea of an experience whose phenomenal content is basically affective. I shall not expand on this thought here, but perhaps some such conviction lies behind the shift in ethical cognitivism away from the kind of ‘intuition’ to which G.E. Moore appeals and towards the sort of position that McDowell espouses, where the role of feeling is more freely acknowledged.

^[24] The various arguments of this section suggest, I hope, that Alston’s view and McDowell’s can be fruitfully combined on certain points. The resulting approach will build on the thought that there is no easy correlation between value experiences in general and states of the world characterised in purely physical, non-evaluative terms. Alston’s account does not articulate precisely this idea, but is consistent with it. And, as we have seen, the idea can help to bolster Alston’s position on various points. First of all, it can help to explain why we do not have a more developed vocabulary for describing the phenomenology of religious experience. Secondly, it can help to rebut the objection that if the phenomenal content of religious experience is purely affective (a possibility which we should take seriously, I have argued), then such experience is best interpreted non-cognitively (since affects, at least typically, lack intentional content in their own right). And finally, appeal to McDowell’s discussion can help us to see how we might (if only for ad hominem purposes) allow a non-cognitivist reading of colour experience without thereby being committed to a non-cognitivist reading of religious experience. In these various ways, then, McDowell’s proposals can be grafted onto Alston’s, to produce an account which is more hospitable to the thought that the affective dimension of religious experience is cognitively significant.

^[25] In his Grammar of Assent, John Henry Newman draws a much cited distinction between having a ‘notion’ and having a ‘real image’ of God. He writes:

^[26] So the distinction between having a notion and having a real image of God amounts to the distinction between having a verbal appreciation of what God is like and having an understanding that is grounded in some direct, experiential encounter with God (having an ‘apprehension’ of ‘the object’, as Newman puts it). This suggests that Newman conceives of religious experience in much the same way as Alston: for the experience to which he is alluding is of God (or at any rate, of God’s ‘voice’), and not of something else as pointing towards God. He is also talking of a mediated apprehension, for it is by way of the data of conscience, Newman proposes, that we are able to experience God. Hence he can write of ‘this instinct of the mind recognizing an external Master in the dictate of conscience, and imaging the thought of Him in the definite impressions which conscience creates’; [19] or again he envisages that ‘in the dictate of conscience, …[an infant] is able gradually to perceive the voice, or the echoes of the voice, of a Master living, personal, and sovereign’. [20] So this account is certainly akin to Alston’s theory, and can be read as a more precise specification of that theory, in so far as Newman proposes that it is our experience in conscience in particular that provides the medium through which we become aware of God. Newman’s account also allows for the affective dimension of religious experience: ‘Conscience’, he writes, ‘…considered as a moral sense, an intellectual sentiment, is a sense of admiration and disgust, of approbation and blame: but it is something more than a moral sense; it is always what the sense of the beautiful is in only certain cases; it is always emotional.’ [21] So Newman is advancing, I suggest, a cognitivist interpretation of affectively toned religious experience (or ‘mystical perception’); and his proposal is therefore of the same general type as the Alston-McDowell model I sketched just now.

^[27] How might Newman’s discussion contribute to the further elaboration of this model? Alston, as we have seen, is struck by the difficulty of verbalising the phenomenology of religious experience. There is a related idea implied in Newman’s distinction between a real image and a notion of God. A notion involves, as we have seen, a verbally expressible understanding. By contrast, the content of a real image of God exceeds our powers of verbalisation. So Newman is proposing that there is a (religiously deeper) non-discursive understanding of God available only to those who have had relevant morally and affectively informed experience of God. Now Newman may think, with Alston, that it is difficult to describe the phenomenology of religious experience, but what he says is simply that it is difficult to capture verbally (in full) the understanding of God that is vouchsafed in such experiences. How should we understand the relation between these two views?

^[28] Appealing to McDowell again, there are reasons for thinking that Alston’s perspective on this point implies Newman’s. For the upshot of McDowell’s discussion is that the phenomenology of value experience is not dispensable for the purpose of identifying its object in the way that the phenomenology of colour experience is dispensable (from the perspective of some commentators) for the purpose of identifying its object. To rehearse again a point that will be familiar by now, on a projectivist reading of colour experience, the phenomenology of such experience fails to reveal what the world is really like, for the qualities in the world that give rise to the experience are colourless; by contrast, says McDowell, in the case of value experience, we cannot draw the same distinction between what appears (values) and the qualities ‘in the world’ which give rise to this appearance. So on this view, a difficulty in recording the phenomenology of value experience implies also a difficulty in recording the character of the ‘object’ which is the source of that experience. Hence, applying this idea to the case of affectively toned religious experience, a difficulty in describing the phenomenology of such experience (this is Alston’s point) implies a difficulty in describing the ‘object’ (God) which is the source of that experience—and this latter thought is the idea advanced in Newman’s proposal that theistic experience involves a real image (as distinct from a notion) of God.

^[29] So Newman’s cognistivist account of affectively informed religious experience complements the model that we have derived from Alston and McDowell, by advancing a claim that is implied (but not explicitly articulated) in the Alston-McDowell account, namely, the claim that there are difficulties in verbalising (in full) the character of the God who is revealed in affectively toned theistic experience. [22] It is worth noting that Newman’s proposal here involves the idea, one I have already attributed to McDowell, that feelings can be intrinsically intentional: that is, they can bear an intellectual content in their own right, and not simply a content that belongs more properly to the discursive (verbalisable) thoughts (or ‘notions’) with which they are associated—for a certain understanding of God is lodged, he is saying, in feeling, and not otherwise expressible. On this point, his approach offers a striking anticipation of the turn taken in recent philosophical discussion of the emotions. [23]

^[30] Does Newman’s proposal contribute anything further to the Alston-McDowell model? Evidently, Newman is thinking of religious experience as closely related to moral experience; the experience of God on his view turns out to involve an experience of oneself as morally accountable, where this accountability points towards a transcendent personal source to whom one is answerable. Hence he writes: ‘If, as is the case, we feel responsibility, are ashamed, are frightened, at transgressing the voice of conscience, this implies that there is One to whom we are responsible, before whom we are ashamed, whose claims upon us we fear’. [24] This is not to say, I think, that the phenomenology of religious experience is indistinguishable from that of moral experience. Religious experience in its fullest sense (in the sense relevant to forming a real image of God) requires not just an experience of conscience, but an experience of conscience as binding, and also an experience of someone as the source of its binding authority. But Newman’s remarks are evidently in keeping with the general strategy of this paper, of treating religious experience as a form of value experience, and one which is tied to moral experience (that is, the kind of value experience that is primarily in view in McDowell’s paper).

^[31] Newman’s discussion relates constructively to one other strand of the Alston-McDowell model. Here we return to the question of why it should be difficult to describe the God revealed in affectively toned theistic experience. In part, no doubt, Newman thinks of the ‘image’ of God realised in religious experience as unverbalisable because this is a familiar theme of the mystical tradition. But he also has a more philosophical kind of reason for advancing this idea. His thought is, I suggest, that it is difficult to express the content of a real image of God in verbal terms because such an image rests upon an encounter with a particular object. Here Newman is presupposing a distinction between knowledge which involves a kind of ‘acquaintance’ with its object (resting on a direct experiential encounter) and knowledge by description. Because knowledge by description trades upon the general categories of a language, he seems to think, it will never be able to capture in all its particularity the content of knowledge by acquaintance. On this view, value (including religious) experience is not (in this respect) radically different from sense experience: what I know by acquaintance of the colour of the apple before me also transcends what I can set down in words. That this is Newman’s teaching is implied, for example, in the passage I cited above, where he writes that a real assent (one involving a real image) is ‘an assent made with an apprehension, not only of what the words of the proposition mean, but of the object denoted by them’. Here we find that the notion/image distinction corresponds to a distinction between the sort of knowledge that turns on knowing the meanings of words (knowledge by description) and the sort that involves knowing an object first hand, by way of direct experiential encounter.

^[32] We might press Newman on the question of why we should suppose that the content of knowledge by acquaintance is, at least in part, unverbalisable. To revert to the case of seeing an apple, the physicist surely does have the necessary concepts to describe exhaustively what I grasp by acquaintance of a given apple’s dimensions, weight and so on, in so far as what I grasp in these respects can be expressed in quantitative terms. By contrast, we do not have the language to capture in all its particularity the apple’s phenomenal appearance as a mix of particular shades of red and green and so on. Why should we have the concepts required to describe the apple in terms of its dimensions and other quantitative properties but not in terms of its phenomenal appearance? This is a consequence, I suggest, of the simplifying power of quantitative descriptions: I can grasp conceptually the full number range required to express the mass, dimensions, velocity etc. of any material object. By contrast, I cannot grasp so simply, in terms of distinguishable concepts, the full colour range required to describe precisely the appearance of an apple. This is not to say that the qualitative appearance of the apple is in principle indescribable (we could after all invent a word for an apple of precisely this colour); but it explains why for practical purposes we lack the necessary richness of vocabulary.

^[33] So we can elaborate upon Newman’s account as follows: the content of a real image of God is not fully expressible in verbal terms because such an image involves knowledge by acquaintance; and, we might now add, knowledge by acquaintance is not fully expressible in verbal terms where it concerns phenomenal colours or, in general, where it concerns the qualitative appearance of things. We need to ask now of course: why is experience of God (of the kind implied in having a real image of God) more like an experience of phenomenal colours, in terms of its indescribability, than like an experience of quantifiable qualities? The Alston-McDowell account of value experience yields an answer to this question: we should think of the object of religious experience as a set of value-indexed qualities, and such qualities can not be expressed in simple quantitative terms. (Compare for example standard criticisms of utilitarian attempts to produce a quantitative calculus for settling ethical questions.) Moreover, whereas it might be possible to argue that the real content of an experience of colours is in fact knowledge of a set of properties whose character can be specified in purely quantitative terms, if McDowell is right, then the same cannot be said of a value-laden experience of God—for in this case we cannot get behind the appearances to identify their source in non-evaluative, non-qualitative terms. So McDowell’s treatment of value experience suggests that a real image of God will involve the sort of knowledge by acquaintance that cannot be fully expressed in quantitative terms; and thereby it suggests (with Newman) that the content of a real image of God will not admit of a precise verbal paraphrase. Here is a further example, then, of how McDowell’s approach can throw new light on a familiar picture of religious experience.

^[34] As we have seen, the Alston-McDowell model also has an account of why it should be difficult to verbalise the understanding of God that is realised in mystical perception. This account involves, first of all, the thought that it is difficult to identify the stimulus conditions of value experience, and therefore difficult to describe the phenomenology of such experience; it then adds that in the case of value experience (including religious experience), a difficulty in describing the phenomenology of the experience implies also a difficulty in describing the object of the experience. By contrast, Newman’s account (as we have developed it) starts from the thought that in general the qualitative appearance of things is difficult to describe (even if those things should be physical objects), and it goes on to claim that the ‘real object’ of a mystical perception can be described only by reference to qualitative appearances. These accounts are distinct but not in conflict; indeed, they both appeal to the thought that in the case of religious experience, there is no route behind the appearances to identify the real source of the experience.

^[35] Having argued that McDowell’s proposal can be related fruitfully to Newman’s and Alston’s discussions of religious experience, I want now to return to the question of this proposal’s plausibility. In his book Ruling Passions, Simon Blackburn offers a number of alleged counterexamples to McDowell’s cognitivist reading of affectively toned moral experience. [25] Here is one of his examples, which concerns the value term ‘cuteness’:

^[36] Now the ‘disentangling manoeuvre’ is not only possible here, Blackburn maintains, but morally required. He continues:

^[37] What should we make of this example? It seems to me that Blackburn is right about the importance of undertaking the ‘disentangling manoeuvre’ (and distinguishing ‘input’ from ‘output’) in this case. But this example trades on our sense that cuteness is indeed an inappropriate reaction; this is why we are reluctant to see it as grounded in an affectively toned ‘perception’, and why it is indeed ‘morally vital’ to substitute instead a projectivist account of the quality. So even someone who is, overall, in favour of McDowell’s account will agree with the appropriateness of a projectivist reading here, providing of course that he or she shares Blackburn’s moral assessment of cuteness. A more testing example would require a case where the response is deemed morally appropriate. It is precisely this sort of case, and the sense of its distinction from the cuteness kind of case, that generates the cognitivist account in the first place: the cognitivist appeals to the idea of an affectively toned perception of certain values ‘in the world’ in order to sustain a distinction between those value claims that are grounded in the nature of things and those (such as judgements of cuteness) that are best understood in projectivist terms.

^[38] So the cuteness example (along with others that Blackburn cites) is, I suggest, neutral between Blackburn’s view and the view of McDowell, since it is common ground that this example is to be construed in projectivist terms. If Blackburn is to make his case, he needs, rather, to show that he can give a better account than the cognitivist of our sense that some value claims are more worthy of endorsement than others. Naturally, Blackburn also has a view on this further issue. In particular, he proposes that we can take ourselves to have knowledge of ethical matters because we can grasp that some of our value assessments are incapable of being improved. Here he quotes Hume’s remark: ‘Temperance, sobriety, patience, constancy, perseverance, forethought, considerateness, secrecy, order, insinuation, address, presence of mind, quickness of conception, facility of expression; these and a thousand more of the same kind, no man will ever deny to be excellences and perfections’. [28] Blackburn comments:

^[39] Blackburn goes on to contrast this sort of judgement with his belief that the government ought to introduce a minimum wage. He notes that it is easy enough to imagine how he might improve his view on this issue, given for example his sketchy understanding of the relevant economic data and theory.

^[40] So Blackburn has an account of what makes some value judgements worthy of the title of knowledge. But from the perspective of a McDowell-style cognitivism, this account still fails, of course, to explain satisfactorily how a value judgement that constitutes knowledge is grounded in the nature of things. All we can say on Blackburn’s approach is that value judgements count as knowledge claims if they properly register the character of certain non-evaluative facts (facts recorded in economic theory, for example) and if our reactions to these non-evaluative facts are freed from ‘insensitivities, fears, blind traditions, failures of knowledge, imagination, sympathy’. [30] If these conditions are satisfied, then we cannot see how our perspective might be open to improvement, and that is sufficient reason for supposing that the perspective counts as knowledge. So on this view, all we have are certain non-evaluative facts and a reaction to them (a reaction we think beyond improvement). A cognitivist account of values will go one step further, and offer an explanation of the appropriateness of the reaction in cases of knowledge: in these cases, the cognitivist will say, the value judgement is cognisant not only of relevant non-evaluative facts, but subjects them to assessment in the light of relevant evaluative facts.

^[41] This dispute can hardly be settled here. But it is worth noting that from a theistic perspective, the cognitivist account of our value judgements seems superior, at any rate in the case of value judgements which concern God. On Blackburn’s approach, the judgement that God is good can be counted as knowledge if it is the sort of judgement we would reach once acquainted with relevant non-evaluative facts and providing that we have overcome failures of imagination, sympathy, etc. But on the theistic view, while it may be true enough that we would form the judgement that God is good under the conditions described, there is something further to be said: the reason why someone who is acquainted with relevant non-evaluative facts and who has purged themselves of various failures of affective response will conclude that God is good is because being in such a state enables a right appreciation of the fact of God’s being good. To suppose that this further step cannot be taken is to suppose that whether or not God is good is ultimately contingent upon our particular human sensibility (and our tendency to form the judgement that God is good under the conditions specified). But a theist will surely want to say that far from our judgements providing a criterion for the goodness of God, it is the goodness of God that provides a criterion for the adequacy of our judgements. It is only if our judgements issue in the claim that God is good (under the conditions specified) that our sensibility can be said to be trustworthy on this question. So while we cannot resolve here the question of whether the further explanatory step taken by the cognitivist about values is in general appropriate, it seems at any rate that some such step is required by traditional theism where judgements about the goodness of God are concerned.

^[42] There is a further key point of disagreement between McDowell and Blackburn that emerges in the original symposium exchange. Here we return to the analogy between value and comic experience that I invoked at the outset to make sense of McDowell’s proposal. Blackburn writes:

^[43] In reply to this line of argument, McDowell suggests that if the non-cognitivist endorses the shapelessness thesis, then ‘there need be no genuine same thing (by the non-cognitivist’s lights) to which successive occurrences of the non-cognitive extra [the felt response] are responses.’ He continues that in that case: ‘non-cognitivism must regard the attitude as something which is simply felt (causally, perhaps, but not rationally explicable); and uses of evaluative language seem appropriately assimilated to certain sorts of exclamation, rather than to paradigm cases of concept-application’. [32]

^[44] What should we make of this exchange? Suppose we take for granted that when characterised in non-evaluative terms, the classes of things which elicit our value (including our moral, aesthetic and religious) responses are shapeless. Both parties to the dispute accept this claim. McDowell continues: since the non-cognitivist does not recognise any evaluative properties ‘in the world’, he or she must infer that our value responses involve no consistent sensitivity to a particular class of things (although given the supervenience thesis, or some analogue of it of the kind that Blackburn could accept, there will still be some sort of patterned relationship between stimulus source characterised in non-evaluative terms and value response); and in that case, our value responses do not conform to standard cases of concept application. In reply, Blackburn wants to know why the cognitivist is not in the same boat, since he or she also subscribes to the shapelessness thesis. Here McDowell would say, I take it, that from the value perspective (but not otherwise), we can see our value responses as tracking shapely properties in the world. I think both sides have a measure of truth here. Blackburn is right to think that shapelessness does not refute his position: after all, why should the non-cognitivist (of all people) suppose that our value responses do track coherent, ‘shapely’ properties in the world? But McDowell is also partly right: he shows, I think, how shapelessness is consistent with the thought that our value experience does track genuine properties ‘in the world’, where the qualities in question are only discernible in the light of our value experience. If this is the right way of reading this exchange, then the result is a kind of stand-off. McDowell fails to establish that non-cognitivism is defeated by the shapelessness of moral and other value properties; and Blackburn fails to establish that cognitivism falls in the face of the same phenomenon.

^[45] However, this weaker reading of what McDowell has shown is sufficient, I suggest, for the model of affectively toned theistic experience that we have been exploring. For that model does not seek to provide a simple knock-down disproof of the non-cognitivist approach; rather, it is an attempt to turn aside certain objections to a cognitivist reading of affectively toned theistic experience. And in the ways we have explored, McDowell’s account does help to show how various features of religious experience do not after all establish any presumption in favour of a non-cognitivist view. It may be true that we cannot describe the phenomenology of religious experience, and cannot describe its object at all fully; it may be true that such experiences are unpredictable, and that their phenomenal content is purely affective; and it may be true that in some respects they resemble our colour experience; but by reference to McDowell, we can see that none of these considerations constitutes a decisive reason for adopting a non-cognitivist interpretation of affectively toned theistic experience. And if these familiar objections to the veridicality of affectively toned theistic experience can be rebutted, then it will be at any rate more difficult to sustain any presumption that such experiences are in general unreliable.

^[46] In this paper, I have sought to bring into fruitful relationship three strands of reflection: on the one side, John McDowell’s cognitivist reading of value experience and, on the other side, the view of affectively toned theistic experience that is expounded in the writings of John Henry Newman and William Alston. The resulting account of theistic experience is built around a number of focal claims, notably these: we should not expect the values disclosed in ‘mystical perceptions’ to be correlated with features of the world which constitute natural classes when characterised in non-evaluative terms; accordingly, there is no presumption that such experience has as its real object or source some non-evaluative feature of the world, rather than a genuine value quality; in turn these ideas suggest that the phenomenology of affectively toned theistic experience may be difficult to describe, and also that the understanding of God that is communicated in such experiences may be difficult to describe. This account is consistent with the general drift of what Alston and Newman have to say, and helps to answer certain questions that are raised by their discussions, including questions such as ‘Why should the phenomenology of religious experience not be more easily describable?’ (an issue for Alston) and ‘Why should the understanding of God achieved in religious experience not be more easily describable?’ (an issue for Newman). Finally, I have argued that Simon Blackburn’s critique of McDowell does not count against a ‘weak’ reading of his case, and that such a reading is all that is required for the model of affectively toned theistic experience that we have been considering. I have also argued that some of Blackburn’s counterexamples to McDowell are inconclusive (notably the ‘cuteness’ example), and that theists have their own reasons (quite apart from McDowell’s discussion) for favouring a cognitivist account of value judgements concerning God.

^[47] Undoubtedly feelings contribute in significant part to the religious experience of ordinary believers. Of course, such feelings can be difficult to describe, and what they reveal about God (if anything) can also be difficult to record in verbal form. But according to the case made in this paper, such experiences are not for that reason untrustworthy. Indeed, if (following McDowell) we favour a cognitivist reading of our value experience in ethics, then there is some reason to think that we ought equally to favour a cognitivist reading of affectively toned theistic experience, and to suppose that such experience offers a mode of access, in some cases our only mode of access, to religiously important values. [33]