First impressions

It is easy to be unimpressed by the idea that facts in nature support claims about God in the way that facts in nature support claims about human minds, and to think also that we have a clear understanding of what the steps amount to. G. E. Hughes expresses the standard response: ‘the step from observation of the world to belief in God is greater in degree or different in kind from the step from observation of the bodies of others to the belief that, for example, they are in pain’.Footnote ³

Impressions can be misleading. Our assessment of the asymmetries or similarities between the two cases will depend on how we understand the relevant concepts and relations. The step from observation of the world to belief in God may seem big and problematic, but maybe this is because our metaphysical framework guarantees that this is so, and this framework could well be unjustified. The step from observation of another's body to the belief that he is, say, in pain certainly seems much smaller and unproblematic, but this could be because we have abandoned the metaphysical framework which would make it difficult to traverse, and accepted with Wittgenstein that the human body is ‘the best picture of the human soul’.Footnote ⁴

Our knowledge of other minds

Two considerations suggest the epistemological problem of other minds – considerations which, I believe, pose no threat once we get our metaphysics straight. The first concerns the distinction between first- and third-personal access to the mental – my pain as opposed to your pain; the second is the distinction between third-personal access to the mental and access to other external things – your pain as opposed to your hair colour. Both distinctions must be preserved if we are to have a satisfactory account of our knowledge of other minds – the first because it allows that we are concerned with other minds, the second because it captures the distinguishing features of this relation.

Let us begin with the first distinction. My knowledge of my own mental states is different from my knowledge of the mental states of others. It is different in the sense that it is not mediated by behavioural evidence or bodily appearance. Rather, it is immediate. This is why the question, ‘How do you know?’ seems so absurd in the first-person case.

That there is a difference in these ways of knowing has no immediate sceptical or metaphysical implications. However, it is a familiar complaint that the substance dualist – pejoratively called the ‘Cartesian’ – exploits this asymmetry to recommend his metaphysics, claiming that mental facts can be accessed only from a first-personal point of view, a third-personal perspective providing, at best, a lowly behavioural surrogate. The difficulties of this picture are familiar.Footnote ⁵ Less familiar is the rejoinder that this ‘Cartesianism’ can be rejected by the substance dualist, allowing that mental facts are equally accessible from a third-personal perspective and respecting embodied human life.Footnote ⁶

I shall return to these claims below. For the moment it suffices to make the following points. First, even if we accept, as we surely must, that first-person ascriptions of the mental carry some epistemological authority, this is compatible with the idea that the relevant facts are equally accessible from a third-personal perspective, the distinguishing feature of this mode of presentation being that it must be based on behavioural evidence. Second, although such a position presupposes that the mental states of others are bodily realized, and in this respect requires that we acknowledge the embodied nature of human life, the underlying metaphysics of this picture are unclear. In particular, it is not ruled out that such a position is compatible with substance dualism.

The idea that the mental states of others are not accessible in the way that things like hair colour are might be thought to revert to the problematic ‘Cartesian’ picture. Indeed it does if we take it to imply that the mind of the other lies beyond anything we could experience. However, there is an alternative way of interpreting the relevant thought, namely, that an observational model omits the distinguishing features of this relation. As Colin McGinn puts it: ‘direct object perceptual reports “I saw the pain in his foot” seem definitely wrong … . I do not believe that others’ mental states can be objects of perception at all.'Footnote ⁷

Our inability to observe the mental states of others does not imply that we cannot tell whether another person is, say, in pain. We can detect things that we cannot directly observe, such as the subatomic particles we detect by observing the behaviour of macroscopic bodies. So our focus must be on the relation ‘seeing or telling that p’, and more specifically, on whether it raises a distinctive difficulty when p involves the thoughts and feelings of another.Footnote ⁸

I can tell much about you by looking. I can tell that you have black hair and green eyes, and, given normal conditions of perception, I can see these properties for what they are without inferring their presence from anything more basic. Things become more complicated when we turn to mental states. For regardless of our answer to the question of how best to characterize the behaviour we can observe, mental states are not open to observation in quite the same way.

Why is this so? To say that it is because they are her mental states is not quite right, because we can say that of any property without implying that it is unperceivable, for example, hair colour. Rather, it has to do with what distinguishes mental states from properties like this, and a distinguishing mark of the mental is that it has a first-personal as well as a third-personal mode of presentation. This does not imply that a third-personal mode of presentation is incompatible with detecting the mental state. Nor does it imply that first-personal access to the mental can be understood on the model of observation.Footnote ⁹ It does suggest, however, that being able to tell that another person is in pain or has a particular thought, requires that we base our judgement on what can be observed from that perspective, namely, the bodily appearance and behaviour of the person to whom the state belongs.

The explanatory inductive model

I want now to examine an approach which takes on board these conclusions and offers a model to explain our knowledge of other minds.Footnote ¹⁰ On this account, we are to draw an analogy with the way we acquire knowledge of unobservables in science. Just as the physicist posits things like electrons to explain the behaviour we observe at the macro-level, so, too, we can best explain human behaviour by positing minds and the mental.Footnote ¹¹

This seems plausible. We often try to figure out why others behave as they do, and tend to do so by reference to what they are thinking or feeling – she jumped out of the window because she was unhappy in love, high on LSD, or anxious to avoid the flames. When we give such explanations, we assume that the behaviour is different from the motions of billiard balls. If we did not make this assumption, our explanations would be as absurd as saying that the billiard ball is moving because it is in the throes of a manic episode. So, when I explain your behaviour, I presuppose that the behaviour is mindful, and try to figure out which particular mental state best explains it.

It is natural to protest that such explanation does not go deep enough. For we need to know not merely whether we are capable of determining in particular cases the details of a person's mental life, but whether we are entitled to adopt this mode of explanation at all. The explanatory inductive approach is targeted at this deeper problem. If we accept all this, several things follow. First, our interest in other minds becomes a theoretical interest, on a par with our interest in sub-atomic particles. Second, it is an interest which may have to be abandoned in light of a rival theory which does not involve the positing of minds and the mental, but which better explains the behavioural data.Footnote ¹² Third, and consequently, we must characterize the data in terms which are not mind-involving, for it is only by so doing that we can grant the possibility of its being better explained in some other way.

Each of these implications raises difficulties. The idea that our interest in other minds is theoretical takes us some distance from our ordinary dealings with one another. We relate to others in terms which involve various emotional and practical commitments – commitments which are quite irrelevant to our dealings with sub-atomic particles. Furthermore, as Hilary Putnam insists,Footnote ¹³ it is difficult to make sense of the idea that these interpersonal interactions are in principle dispensable – to be abandoned in the light of a better theory. This idea becomes remotely plausible if we characterize our observational base in terms which are not mind-involving, but if we make this move, we are led back to a version of the difficulty we face if we try to characterize the movements of a billiard ball in mind-involving terms. For mindless behaviour is not best explained by positing minds and the mental.

So we face a dilemma if we apply the explanatory inductive approach to our knowledge of other minds, both horns of which undermine the cogency of this approach. On the first option, we characterize our observational data – human behaviour – in mind-independent terms. This respects the distinction between observational data and theory, but leaves no way of justifying the inference. The second option is to characterize our data in terms which are already mind-involving, but this presupposes the truth of the hypothesis in question, so that the model is abandoned in all but name.

What I have said implies that observational data and theory are to be held apart on the explanatory inductive approach, that there is an alternative in which they are not held apart, and that this alternative is preferable as far as knowledge of other minds is concerned. The idea that observational data and theory can be separated has been challenged on the ground that all observation is theory-laden to some extent. But what does this mean? It has been taken to mean that all observation is concept-involving. However, the concepts in question are not the theoretical concepts of science. Rather, they are those which are operative at the level of ordinary thought and experience. These concepts do not explain what we observe. On the contrary, they reveal what we observe and in this respect play a pre-suppositional rather than an explanatory role in our experience.Footnote ¹⁴

I shall assume that this account is correct, and that a ‘theory’ in this sense is presupposed in our experience of the world. But what is its scope? And how does it relate to scientific theories? First, and this concerns the issue of how we are to comprehend the distinction between scientific theories proper and the kind of ‘theory’ which is presupposed in ordinary thought and experience, scientific theories, in so far as they serve to explain what we observe, can be set apart from observational data. Thus, when we abandon one scientific theory for another, we abandon a certain way of explaining things. We do not, however, abandon the things we seek to explain. By contrast, if we assume that the concepts operative at the level of ordinary thought and experience are revelatory rather than explanatory of the things to which we relate at that level, they can be abandoned only at the cost of abandoning those very things.

This returns us to the question of how we are to understand the scope of these concepts, and in particular, whether it can be stretched to accommodate the mind-involving behaviour to which we respond when we interact with and try to make sense of one another. I have suggested that an explanatory inductive approach faces a difficulty in making sense of our interactions with others, and a satisfactory alternative requires that we comprehend human behaviour in terms which already presuppose our entitlement to the concepts of mind and the mental. This suggests that such concepts do not have the status of a scientific theory, that we could abandon them only at the cost of rendering unintelligible our capacity to relate to others, and hence, that they form part of the background to our intuitive dealings with reality. Our question then is whether there is an alternative way of vindicating our right to work with these concepts, and in this way to allow that others have inner lives. This alternative must be consistent with the idea that their inner lives are not perceived in the way that we perceive things like hair colour, and that their perceived behaviour is not reducible to billiard-ball motion.

The expressive model

A more promising model comes with the introduction of the notion of expression,Footnote ¹⁵ the expression in play when mental states are revealed in bodily appearance or behaviour, for example, a facial or behavioural expression of pain. Behaviour here is not mere bodily motion, for this would take us back to the problematic idea that the mental life of others is cut off from anything we can observe, and imply that the behavioural expression of the mental is, at best, a rather poor surrogate for something that must elude a third-person perspective. As Wittgenstein puts it, the genuine expression would become a sum of the expression and something else.Footnote ¹⁶ Behavioural expression already presupposes this something else.

Before considering this in more detail, we should clear up potential misunderstandings. First, the model allows for differences among mental phenomena; their characteristic modes of expression need not conform to a single type. Second, these modes of expression are not restricted to the evincing which occurs when, say, someone gives expression to their pain – which also occurs in babies and animals – but includes the communication of intentional states. Third, mental phenomena are not to be reduced to their characteristic modes of expression, and so no concession is made to behaviourism. The model simply allows that we can get in cognitive touch with the mental states of others, and interpret mindful behaviour for what it is.

On this model, what is directly available to the third-personal perspective is mentally expressive behaviour as opposed to the inner state itself. The point of saying that this behaviour is mentally expressive is that it is of a different order from that of billiard balls. Saying that it is directly available to us stresses that we need not proceed from a more basic level of mere bodily motion. Finally, in saying that what is directly available is mentally expressive behaviour, we respect the difference between first- and third-personal access to the mental, for what is directly available to the third-personal perspective is not the inner state itself, but a person's expression of it.Footnote ¹⁷

It is consistent with this model that interpreting mentally expressive behaviour is no easy matter. Interpreting behaviour which is not mentally expressive can be complex too. The difference is that mentally expressive behaviour, unlike mindless motion, has its source in another consciousness. It involves the communication of an inner life – a communication which, although it is our main source of knowledge of the other, may be refused, withdrawn, or designed to deceive.

The idea that mentally expressive behaviour does not always wear its nature on its face might be thought to conflict with the claim that the mental states of others are directly available in their expression. But the conflict is merely apparent. The point of saying that mental states are directly available in their expression is to entitle us to view the behaviour as mind-involving. By contrast, the point of saying that this behaviour does not always wear its nature on its face is to allow that our interpretations are fallible, that we may fail to make sense of what someone is doing, thinking or feeling. This concession retains an element of inaccessibility on the part of the other, thereby avoiding crude behaviourism. However, it is perfectly compatible with the possibility of getting things right, provided that we insist on the first point, namely, that the behaviour is mindful and appreciable as such.

For example, I wrongly suppose that you are in terrible pain. I am presented with mindful behaviour that merely seems expressive of pain. It does not matter whether we say that we have a genuine expression of pain without genuine pain or that we do not have a genuine expression of pain. However we describe it, we must assume that the behaviour is mindful. Without this presupposition the interpretative task cannot be completed, and once this presupposition is in place, the possibility of error remains permanent but harmless.

Diagnosis

Where does this leave the question whether we are entitled to assume that those around us have inner lives? This entitlement remains a mystery as long as we insist on the explanatory inductive approach, and this approach involves a conception of behaviour we have reason to reject. On this conception, behaviour is not mind-involving and stands opposed to mental states for which it provides, at best, dubious evidence. The alternative is to work with a richer conception of human behaviour – one which allows that others' mental states can be directly apprehended in behavioural form.

This alternative becomes viable once we stop thinking of the mind as a thing apart from anything we observe. Once we embrace it, the move from observation of a person's body to the belief that they are, say, in pain, becomes the familiar step that we take in our everyday dealings with one another, the step from observed mindful behaviour to the belief that it is expressive of this specific mental state as opposed to that.

The idea that an expressive model becomes viable once we stop thinking of the mind as a thing apart from anything we observe reminds us that our options are dictated by our conceptual presuppositions. This point is hardly remarkable. However, it leads to some more interesting questions, which will be important when we get on to God. Why do we think in terms which set the mind apart from anything bodily/behavioural? What do these terms amount to? Are they acceptable?

John McDowell identifies the underlying framework as one in which the concept of a human being is displaced from its focal position in our experience and replaced with that of a human body.Footnote ¹⁸ On this conception, we are presented in experience with human bodies, human bodies providing us with bodily and behavioural information which stands opposed to the realm of mind and the mental. As to why we succumb to this framework, McDowell blames it on an objectifying view of reality or nature, precluding the possibility that bodies and behaviour can be genuinely mind-involving.Footnote ¹⁹

This explanation seems uninformative. Talk of objectifying reality is just another way of describing the phenomenon to be explained, and even if we go further, as McDowell does, and blame it on something like scientism, this just raises the question of why we accept such a position. It is unclear that there is an answer to this question that does not simply circle back on the framework we are trying to explain.

This circularity need not undermine the value of the explanation. Maybe we can gradually enrich our understanding of the relevant pattern of thought, and address the more important question whether we ought to be thinking in this way. We should be prepared to question these conceptions of mind, body, and behaviour, and embrace an alternative which focuses on the embodied human being – a being whose mental states can be accessed in behavioural and bodily form.

Such a focus will be incompatible with any position which treats mind and body in mutually exclusive and ‘splintered’ terms,Footnote ²⁰ and, for philosophers like McDowell, this involves abandoning traditional Cartesianism.Footnote ²¹ I have noted already that a certain kind of substance dualist will resist this implication on the ground that his position can accommodate the embodied human being. Charles Taliaferro, for example, endorses ‘integrative dualism’, a position which grants the ‘materially conditioned, embodied nature of personal life’, whilst allowing that the person and the body are separable individuals. As such, it is ‘juxtaposed to versions of dualism that depict the person as unduly remote and splintered’.Footnote ²²

It seems clear that the integrative dualist could exploit an expressive model of mind. Thus, he can avoid the epistemological problem of other minds and agree with McDowell that the concept of a human being has a focal position in our experience. But is there any justification for insisting on the ‘dualist’ aspect of this integrative position by claiming that person and body are metaphysically distinct? For if we make this move there is a clear sense in which we are committed to claiming that the mind is a thing apart from anything bodily/behavioural. We shall be in a better position to address this question after considering the case of God.

The problem of God

Let me return to Hughes's claim that the step from observation of the world to belief in God appears greater in degree or different in kind from the step from observation of the bodies of others to the belief that they are in pain.

God is not an object of perception, and His presence is not obviously detectable in the materials which are available to perception. So someone might well be persuaded that the step from observation of the world to belief in God is problematic. An analogous problem presents itself for other minds, and there is a way of addressing it which implies that belief in other minds involves an explanatory induction. The worry here is that disbelief in other minds may turn out to be the most rational response to the evidence. Here we can appreciate a difference between the two cases. As Hughes notes, there are no serious non-believers in other minds, but the serious non-believer in God is an everyday acquaintance.Footnote ²³ A serious non-believer in God admits all the empirical facts about the world, yet denies that God exists. We must consider then whether the theist has room for manoeuvre.

The explanatory inductive approach to God

Hughes's characterization of the atheist suggests that a principled distinction can be made between God and the world. Certainly, it seems easy to describe the world in terms which are not God-involving and unclear what a God-involving description would amount to. Furthermore, the suggestion occurs in certain ways of thinking about arguments for the existence of God. Thus, Tillich says that ‘in arguments for the existence of God, the world is given and God is sought’,Footnote ²⁴ and Frederick Ferre says that we are ‘going from facts about the world to facts about God’.Footnote ²⁵

Given the importance of the distinction between observational data and theory in the explanatory inductive approach, it is unsurprising that this model has been applied to God, treating His existence as a hypothesis which best explains the existence and nature of the world. It is unsurprising also that this hypothesis has been rejected by the non-believer.

We can appreciate the dialectical situation by focusing on two philosophers – Richard SwinburneFootnote ²⁶ and John MackieFootnote ²⁷ – who agree that this model can be applied to God, but disagree over the cogency of the hypothesis.Footnote ²⁸ For Swinburne, God provides the best explanation of the existence and nature of the world. The world incorporates not merely facts open to scientific investigation but also the good, the beautiful, the evil, the miraculous, and the numinous. He takes all this evidence together and concludes that it would be ‘strange and puzzling’Footnote ²⁹ and highly improbable in the absence of God, a wholly good, omnipotent personal Being.

Mackie, by contrast, denies that the world requires a theistic explanation, arguing that it is better explained without reference to God. First, he sees nothing about the world that demands a theistic explanation, and much that resists it – the existence of evil, for example. Second, he holds that we can explain scientifically everything that needs explaining.Footnote ³⁰ Explanations in physics comprehend the behaviour of macroscopic bodies, and explanations in psychology promise to accommodate our experience of goodness, beauty, and even God Himself.

Presented in this way, the dialectic displays a version of the dilemma which arises when the explanatory inductive approach is applied to other minds. We there faced two alternatives. We can characterize our observational base without reference to mind and the mental, leaving open the possibility of a rival theory which better explains the data. If we do this, the explanation fails to go through. Or we can characterize our observational base in mind-involving terms. This captures our intuitive experience of human behaviour, but it undermines the model by presupposing the hypothesis.

The dilemma can be applied to God. If we follow Mackie, we can treat the existence of God in inductive explanatory terms, for our observational base can be characterized neutrally. To Mackie, however, a theistic hypothesis is implausible: nothing about our observational base – the world – requires reference to God.

Swinburne holds likewise that the explanatory inductive model can be exploited to account for the world, but unlike Mackie, thinks that a theistic hypothesis is preferable. He takes the existence of God to be intrinsically plausible, views the world in terms already charged with religious significance, and regards Mackie's level of explanation as deficient. All this suggests that the evidence is not unequivocal as far as Mackie and Swinburne are concerned, and Swinburne's conception of it loads the dice in favour of theism. Mackie is like someone who views human behaviour as mere bodily motion and rejects the hypothesis that there are other minds; Swinburne is like someone who views human behaviour in mind-involving terms, and sees the other minds hypothesis as the only adequate way of explaining what he observes.

As regards other minds, we should insist that human behaviour is mind-involving. As Putnam said, there is just no alternative in the field, and, crucially, we have a model which makes this conception intelligible. God seems quite different, not least because we can make sense of the idea that the world we observe is not God-involving, and that He does not exist. Are we not returned then to the idea that the step from observation of the world to belief in God is of a different order to the step from observation of a person's body to the belief that they are in some mental state or other, and that it is a step we are not justified in taking?

A deeper analogy?

One reason for caution is that some theologians imply that the two cases have more in common than our previous arguments might suggest. In particular, it has been claimed that we should avoid the conception of God as a supernatural entity, set apart from the natural world, a conception that conspires to make Him unreal.Footnote ³¹ The structure of this line of thought is familiar from the case of other minds. For there is a tendency to make a similar mistake in that case too, by treating the mind as a thing apart from anything bodily or behavioural. And the framework proposed for God in light of these theological considerations is similar to the framework we recommended for other minds. There may yet be scope for claiming that disbelief in God is as far off-field as disbelief in other minds.

We have seen that, as regards other minds, a position which respects embodied human life may be compatible with substance dualism. On this view, mind and body function as a unity, but are metaphysically distinct. This position may ultimately be indefensible. However, the possibility of such a move becomes relevant to our theological case if it turns out that it is fundamental to any plausible conception of God that He is metaphysically distinct from the natural world.

My preliminary characterization of the proposal might be thought to rule out the idea that God is metaphysically distinct from the natural world. For the negative claim is that we must reject the view that God is a supernatural entity. However, we need to think carefully about what this rejection amounts to. The claim is twofold. First, God is not an entity. Second, God is not a supernatural entity.

One way of spelling out the first requirement is to say that God cannot be viewed in terms appropriate to entities within the natural world. This requirement would be respected by dualists and non-dualists alike, for it is agreed by all sides that a reduction along these lines would compromise the reality of God. It is unclear, however, whether it follows that God is not an entity in some other sense, and whether a rejection of this idea does not compromise the reality of God in the opposite direction.

The second claim – that God is not a supernatural entity – is equally problematic, for there is no straightforward answer to the questions of how we are to interpret the term ‘supernatural’, and of whether the proposal involves or ought to involve its repudiation. It remains open at this stage that we can grant with our theologians that God has some kind of worldly presence – in this respect, we distance ourselves from the position which treats the God/world relation in splintered and mutually exclusive terms. What is not yet ruled out is that such a concession is compatible with the idea that God is metaphysically distinct from the world.

An expressive model of God

We can begin to spell out the proposal by returning to the analogy with other minds, and considering how far we might exploit an expressive model of God. First, we must allow that our relation to God can be modelled on our relation to other persons, and hence, that God is to be understood in personal terms. Second, our observational data – the world – must admit of a richer characterization than that implied by the explanatory inductive model, at least as that model is understood by Mackie. Third, it must be taken to be expressive of God as a mind.

That God is to be understood in personal terms is familiar, but it is equally familiar that, in itself, this description is little more than a placeholder for a solution.Footnote ³² One reason for caution is that, in itself, it imposes no particular metaphysical constraints. Swinburne, for example, takes it to imply that God exists as a disembodied person – an implication he is happy to accept, and would doubtless extend to beings like ourselves. Mackie, by contrast, objects that there is no evidence that disembodied persons exist, and that our experience of persons is limited to embodied human beings. The implication here is that persons are embodied and so God is not a person.

I have argued that we must accept that human persons are embodied, but have left open the possibility that this is compatible with substance dualism. Thus, as far as the Mackie/Swinburne disagreement applies to the case of human beings, we can agree with Mackie that our experience of persons is limited to embodied human beings, whilst denying that it follows that persons cannot exist in disembodied form. But what are the implications for the case of God? If we accept that God is a person, what kind of metaphysical constraint does this impose? Does it follow that God is embodied or disembodied or both of these things?

The theologian John Robinson argues that it is not a requirement on thinking of God in personal terms that we think of Him as a disembodied person, and that if we do accept such terms of description we play into the hands of the non-believer who sees no worldly evidence for such a Being. His related worry is that we have no way of making sense of the idea that we stand in a personal relationship to God, on the assumption that we cannot be personally related to a disembodied person utterly distinct from anything in the world. He poses the following question: ‘Might not the image of this super-Person distract from the reality-in-relationship it was seeking to express by turning people's eyes upwards or outwards to a Being for whose existence the evidence was to say the least, doubtful, instead of focussing attention on “the beyond in the midst”?’Footnote ³³

The idea that personality is inextricably tied to personal relationship is plausible, but what does it mean to say that we are to focus on the ‘beyond in the midst’? And how is this intended to make sense of the idea that we can stand in a personal relationship to God? Given the supposed difficulties of treating God as a disembodied super-Person, we can hazard a suggestion of what Robinson is getting at, and relate it to other minds. In both cases there is a tendency to oppose the relevant x to the data we experience, and in both cases, we are to suppose, there is hope of a more satisfactory alternative if we can characterize the data in terms which are already x-involving. As far as other minds are concerned, this amounts to focusing on a living human being as opposed to a disembodied mind or a mere body and allowing that mental states admit of behavioural/bodily expression. As far as God is concerned, we might suppose, it involves focusing on a God-involving world as opposed to a disembodied God or a Godless world, and allowing that His mind admits of behavioural/worldly expression.

Some worries

The general idea promises to maintain a distinction between God and the world whilst granting Him a worldly presence. However, there are difficulties. One general worry is whether this shift in framework is really justified in a theological context. The motive for welcoming such a move in the case of other minds is that we avoid certain epistemological difficulties, and can make sense of our intuitive dealings with others. It is easy to see how a similar vindication might proceed for God. It will be easier to make sense of our dealings with God if we reject the idea that He is an inaccessible super-Person behind the scenes. However, there are two problems with this response. First, it will carry little weight with a non-believer who denies that our experience should be described in God-involving terms. Second, it may be challenged by one who is happy to bring God into play, but worries about the details and perhaps even the intelligibility of the proposed model.

The model implies that God is to be treated as a God-world unity, and that the aspects of this unity are intelligible only by reference to this more primitive whole.Footnote ³⁴ To use Peter Strawson's terminology, we are concerned with one two-sided thing rather than two one-sided things. If we accept this, we can allow a distinction between God's mind and body, just as in the case of a human person, but we must presumably say that His body is the world, and that His mind finds expression in the world – the world as the self-expression of God.Footnote ³⁵

There are difficulties lurking in these claims, some of which can be sidestepped by allowing that analogies need not be perfect fits. I shall focus on one potential difficulty relevant to what has gone before. I have stressed that, in the case of human beings, we can distinguish between mentally expressive behaviour and mere bodily motion. By acknowledging this distinction we see how problematic it is to subsume all our behaviour under the latter category, and we have a metaphysical framework which entitles us to resist this objectifying approach. If we accept an expressive model of God, it may seem that we risk fudging this distinction. An expressive model allows us a richer conception of the world than that which is permitted by the explanatory inductive approach, at least as that approach is understood by Mackie. Moreover, this richer conception is ours once we grant that the world is God-involving, or expressive of the mind of God. I have been reticent about the precise meaning of this claim, but one of the things it might mean is that mental activity is all-pervasive. Now if we accept that mental activity is all-pervasive, and if we assume that the term ‘mental’ is not used entirely equivocally as applied to man and to God, we surely risk undermining the distinction within the world between mentally expressive behaviour and mindless motion, albeit in the opposite direction from that which motivated my initial argument. We have what might be termed an exclusively subjectivist approach to reality. And we are now saddled with a position which, according to McDowell, cannot be taken seriously by educated people, a position which takes us back to the medieval view that nature is ‘a book of lessons for us’.Footnote ³⁶

McDowell again

McDowell insists that we distinguish between the intelligibility operative at the level of the things we investigate scientifically – the motions of billiard balls and electrons, say – and the intelligibility of genuinely human interactions. The difficulty, McDowell believes, is that there has been a tendency to assume the ‘disenchanted’ conception of nature presupposed in the scientific outlook and to extend it to everything. We are then left with the ‘objectifying’ approach – an approach which cannot make sense of the idea that human behaviour is different in kind from billiard ball motion. McDowell insists, however, that ‘we had better not aspire to put the lost enchantment back into the merely natural world’, and warns against a ‘regress into a pre-scientific superstition’. This is ‘a crazily nostalgic attempt to re-enchant the natural world’, reinstating the idea that ‘the movements of the planets, or the fall of a sparrow, is rightly approached in the sort of way we approach a text or an utterance or some other kind of action’.Footnote ³⁷

As regards human behaviour, McDowell rejects the assumption that what distinguishes us from billiard balls is something supernatural – a Cartesian soul, for example. Such an assumption would play into the hands of one who insists that we can buy such a distinction only at the cost of retreating into a position that is intolerably mysterious. We are to assume no doubt that wholesale re-enchantment of nature would be similarly problematic. So McDowell requires that we ‘naturalize’ human behaviour. However, ‘naturalizing’ human behaviour demands not that we comprehend it in terms appropriate to billiard balls – this would be to comply with the framework under attack. Rather, it is ‘naturalized’ in the sense that it is shown to derive from our nature as human beings. And precisely because we have rejected the picture of human beings as either free-floating souls with a precarious foothold in the animal realm or as creatures whose behaviour is no different from that of billiard balls, there is no obvious reason why such a position should offend the sensibilities of the empirically minded philosopher. The only concessions he needs to make is that human behaviour is of a different order to that of billiard-ball motion, that we can accommodate this difference in perfectly sober metaphysical terms, and that we can do so by allowing that the realm of the natural includes human reality.

It should be clear from what I have said that the substance dualist could challenge this pejorative characterization of his position. Let me leave that worry on one side, however, and focus upon the theological implications. As McDowell sees it, there are good and bad ways of re-enchanting the natural world. We re-enchant it in a good way when we acknowledge the distinction between mindful behaviour and bodily motion, while resisting supernaturalism. We re-enchant it in a bad way when we fudge this distinction in the opposite direction from scientistic fudging, by treating all behaviour as mindful. This would be a regress into pre-scientific superstition, eliminating the distinction we seek to accommodate.

Our question then is whether the view that the world is expressive of the mind of God involves such a regress, and hence, whether it is a view that cannot be countenanced by educated people. Some versions of this position are open to this criticism. One could insist that thunder expresses God's anger, that sunshine means that He is in a good mood, and that the odd hurricane indicates theological mania. Such a position could not be taken seriously by educated people, for it is simply an extension of the ‘God as man in the sky’ picture. It is on a level with the interpreting of billiard balls in mind-involving terms. But where does this leave the expressive model? Is it a pre-scientific superstition or can it be defended?

The defence must meet four conditions. First, the model must give due weight to science. Second, it must respect the distinction between bodily motion – the province of scientific investigation – and our mentally expressive behaviour. Third, it must not compromise the status of mentally expressive behaviour by subsuming it under the mentally expressive behaviour of God. Finally, we must show that there are good reasons for regarding the world as expressive of God.

Jantzen on God's self-expression

Among the various ways of addressing these issues, I shall focus on Grace Jantzen's account in God's World, God's Body. Jantzen believes that everything is God's self-expression. She distinguishes this from the view that there are aspects of God's behaviour and body – i.e. of the world and its workings – wholly independent of Him, claiming that this deviates from a properly Christian framework.Footnote ³⁸ There is, she argues, a crucial disanalogy between God's relationship to the world and our relationship to our bodies: ‘(w)e are embodied only incompletely, in the sense that our bodies are only partially expressive of ourselves, our desires and attitudes … we cannot by taking thought make one hair white or black’.Footnote ³⁹ The claim that we are embodied only incompletely allows for aspects of ourselves beyond our control – hair colour, etc. But things are not like this with God: everything is in His control. We must consider then whether Jantzen can allow that there is bodily motion at work in reality, upholding a distinction between bodily motion and human expressive behaviour (our second condition), and between our mentally expressive behaviour and God's (our third condition).

Jantzen accommodates these distinctions by her notion of ‘transcendence’. Human persons are transcendent in the sense that they are open to reason, emotion, responsibility, and ‘the mysterious infinity’, and so cannot be fully understood in mechanical or physiological terms. Transcendence in this sense does not imply disembodiment. We are embodied beings who cannot be fully comprehended scientifically. Thus the opposite of ‘transcendent’ is not ‘immanent’ but ‘reducible’.Footnote ⁴⁰ Human persons are transcendent because they are not reducible to matter, and not fully comprehensible in the terms of physiology or mechanics. But they can be partially comprehended in these terms. The picture is familiar from McDowell.

Jantzen next turns to God, claiming that just as we cannot be fully explained in scientific terms, neither can the universe of which we are a part:

Again, the theme is familiar from McDowell: mechanism isn't everything and not everything can be fully comprehended scientifically, though some things can be comprehended scientifically. But, unlike McDowell, she introduces God. God is, or may be, related to the observable universe in the way that the human mind is related to the body. Neither the person nor the mind is ‘reducible’ to bodily behaviour and physiology; the person is transcendent in that it expresses a mind and the mind is transcendent in that it transcends the bodily aspect of the person. Analogously, neither the observable universe nor God is reducible to sheer mechanism; the universe is transcendent in that it does (or may) express God, and God is transcendent in that He transcends that aspect of the universe which can be comprehended in mechanistic terms. God does not, however, transcend the universe in the sense that He is separate from it or could exist without it. God can no more dispense with the world than the mind can dispense with the body.

The idea that some aspects of reality can be comprehended scientifically allows Janzten to give due weight to science and hence to satisfy our first condition. However, it remains to be seen whether her preferred theological terms of description compromise this conclusion, whether the underlying expressive theory forces a retreat into the position that McDowell wants to avoid, and whether such a position is as intellectually disastrous as he assumes.

Let me begin by considering how Jantzen proposes to accommodate the mentally expressive behaviour of human beings (our third condition). Her response is to be found in a discussion of how there can be individuality and autonomy if God is in control of every aspect of the universe. Jantzen tries to lighten the blow by reminding us that the problem is inevitable for any properly Christian position. The problem could be solved if we allowed that the world and things in it have an origin and existence utterly independent of God, but this move is prohibited.Footnote ⁴² In any case, she believes that it is unnecessary, and sketches a solution compatible with the idea that everything is God's self-expression.

The solution eludes us, she argues, if we focus on God's omnipotence and infinity to the exclusion of his love. If we consider God's love, and the limitation love imposes on power, we see that it involves granting autonomy to His creatures. Hence, ‘He is the One who is All, yet to whom the Many owe their freedom and their selves’,Footnote ⁴³ and she quotes with approval John Macquarrie's claim that ‘Being itself lets be the beings’.Footnote ⁴⁴

How then do we express God if we are autonomous? Presumably we must say that what we are is part of God's self expression, and so, perhaps, is what we involuntarily suffer, such as falling in love, but that what we do voluntarily is not. This promises to make sense of the idea that our mentally expressive behaviour has its source in us rather than in God, and it also reminds us that the relation between God and man is personal. So it looks as if Jantzen can allow that there is a distinction between our mentally expressive behaviour and that of God (condition 3). And, given her previous claim that there are aspects of ourselves which elude scientific explanation, she can allow that our mentally expressive behaviour counts as one such aspect and is to be distinguished from all that can be adequately comprehended scientifically (condition 2).

Much of this would be acceptable to a McDowellian naturalist. For he agrees that there are aspects of nature which cannot be comprehended scientifically, and that mentally expressive human behaviour falls into this category. This much grants him the right to agree with Jantzen that human persons and nature are transcendent in the sense she defines. However, he would reject her assumption that the essential missing ingredient is explanation along theological lines.

Jantzen's account satisfies, then, the first three conditions. But now we face condition 4 and the question whether there are any good reasons for introducing God. It would be a fatal disadvantage if the reference to God constituted a retreat back to pre-scientific superstition. But if pre-scientific superstition involves ignoring science then Jantzen can sidestep this complaint. Nor is it any objection to say that God makes the world too enchanted. For the fundamental message of the McDowellian naturalist is that the world is enchanted, and that we can appreciate it as such provided that we reject scientism. However, McDowell would balk at the idea that the world is enchanted in a God-involving way.

It may look as if we have reached an impasse. We have two ‘enchanted’ conceptions of nature, but McDowell refuses to go the whole theological hog. One option is to say that the difference is merely terminological. Both McDowell and Jantzen are rejecting a scientistic conception of nature, we can describe the non-scientistic alternative in theological terms, but nothing really hangs on it. On this view, presenting the world as God's self expression is just a fancy way of saying that human reality cannot be exhausted scientifically. God makes no difference. And if God does make a difference, theistic expressionism seems skewered on the other horn of the dilemma: what reason could there be to believe that the world expresses God?

But what does it mean to say that God makes no difference? Even if God is just a placeholder for aspects of nature resisting scientific explanation, He makes a difference in some sense. He makes a difference in just the way that we make a difference. But that, one might say, is just a fancy way of saying that it is human reality that makes the difference, that enchants the world. But this raises further questions: are we clear about the nature of human reality, in particular about its limits? How do we know that our mental capacities are not God-involving? Can we be sure that ‘atheism’ will not leave us with an impoverished conception of human reality? Such doubts might suggest both that God makes a difference and that there may be reason to believe in Him.

Naturalism and dualism

Both Jantzen and McDowell want to re-enchant nature, arguing that the first step is to reject scientism. Both stress our embodied nature and believe that substance dualism disregards it. Taliaferro agrees in respecting human embodiment, but denies that this conflicts with substance dualism, with the metaphysical distinctness of person and body.

Jantzen and McDowell share a conception of nature expansive enough to accommodate the distinction between ourselves and other things, but refuse to spell out the distinction in substance dualist terms. Jantzen characterizes this conception of nature in theological terms, but her conception of God is likewise shorn of substance dualism. So for both the natural world has aspects which cannot be comprehended scientifically. Jantzen goes further than McDowell by describing these aspects in God-involving terms, but refuses to treat God as a separable substance.

We now face two disputes. The first concerns the question of how far we can expand the limits of the natural world, the second concerns that of whether these limits should be set against something else. In the human case, this something else is a separable soul, in the worldly case, it is a separable God. McDowell rejects this second move in both contexts. Yet if Taliaferro is right to suppose that the substance dualist can accommodate everything McDowell wants to say about the embodied nature of human being, then we lose the motive for rejecting it on these grounds, and are left with the question of whether there is any further reason for rejecting it or any justification for accepting it. McDowell could repeat his complaint that it is just too philosophically problematic, but on what ground? The usual arguments – the problems of interaction and individuation – tend to depend on considerations which are persuasive only to one who is already convinced that substance dualism is a non-starter.Footnote ⁴⁵ Furthermore, it would be open to the substance dualist to complain that McDowell's own preferred alternative introduces equally devastating difficulties. After all, it is not obvious that McDowell's conception of human nature is without difficulty. Hence the lure of scientific naturalism.

So we secure no immediate philosophical advantage by settling with McDowell's conception of nature. Furthermore, if the question of whether McDowell should expand the limits of nature in a theological direction remains open, then we might be tempted to conclude that there is but a knife edge between Jantzen's theological naturalism and the view that God is metaphysically distinct from the natural world. Yes, there are versions of this position which lead to difficulties analogous to those which arise if we treat the person/body relation in mutually exclusive and splintered terms. However, we can avoid this implication by taking as our model the integrative dualism which Taliaferro endorses for human being. On this account, we grant the insights of Jantzen's conception of the God/world relation whilst insisting that He is metaphysically distinct from the world.Footnote ⁴⁶

We find ourselves in a similar predicament to that we faced when trying to decide between McDowell's and Jantzen's brand of naturalism. Furthermore, it is tempting to conclude that, in this new context, too, it is really just a matter of terminology. For both sides agree that God makes an essential difference, and hence, that McDowell's brand of naturalism doesn't go far enough. Does it really matter then whether we describe this distinguishing feature in terms which commit us to substance dualism? Given the inadequacies which surround any attempt to describe the nature of God, one might suppose that adjudication is impossible, and that nothing important hangs on the issue. However, it should be clear from what has been said that this is not the case. For if we accept Jantzen's account, then we must deny that God is separate from the universe or could exist without it.

It is this claim that marks the essential difference between her position and that of Taliaferro, and the question of how we are to choose between these options becomes the question of whether there are any good reasons for accepting this dualist conception of God. If these reasons are forthcoming, and if we accept, as I think we should, that theology contains important implications for the philosophy of mind, then we shall be forced to rethink our understanding of human nature along the lines recommended by Taliaferro. This extra step may be surprisingly harmless, and no more mysterious than anything envisaged by McDowell.