In Causation in Psychology John Campbell offers an original, intriguing picture of mental causation. Specifically, in Chapters 1 and 2 he argues that we have a conception of causal mental processes – processes such as, for example, the chain of thoughts and feelings that led from Sally's humiliation to her depression, or the train of thought that led to Billy's decision to accept the job offer – and that such processes are accessible to us by means of our capacity for imaginative (or empathic) understanding of others. In Chapter 3 Campbell argues that these causal mental processes are unique to humans and are relevant to explaining both the sense in which we are free as well as what matters to us most about human interaction. In the fourth, final chapter Campbell examines how the view he has been developing bears on the mind-body problem.

The book is very rich: each chapter is packed with intriguing arguments, thought-provoking examples, and insightful observations. It is impossible here to do justice to all of them. I'll focus on the view Campbell defends in the first two chapters – namely, that we have a conception of causal mental processes, and that we are able to gain knowledge about such processes by means of imaginative understanding. I'll describe some aspects of Campbell's argument for the view, and while doing so, attempt to explain what the conception in question involves, and what Campbell means by ‘imaginative understanding’. One issue that readers may find puzzling is the fact that Campbell moves between claims about what our conception of causation involves and what causation is – thus, the argument in question is meant not only to establish a claim about our conception of mental causation, but also about mental causation itself. Campbell doesn't explain explicitly why he thinks this is justified. I'll indicate, very roughly, a line of thought that might explain this. I'll end by raising a few questions.

A key to Campbell's argument is the thought that we can gain insight into what causation is by looking at the ways in which we acquire knowledge about causation. Campbell starts by pointing out that our ways of coming to know general causal claims – such as ‘salt causes heart disease’ and ‘humiliation causes depression’ – and our ways of coming to know singular causal claims – such as ‘Sally's high salt intake caused her heart disease’ and ‘Sally's humiliation caused her depression’ – are, in certain respects, independent of each other. In the case of general causal claims – both those concerning the mental and those concerning the physical – we derive causal information from statistical evidence about what happens under interventions. The most prominent method by which we establish such claims is that of randomised controlled trial (RCT) where individuals are randomly assigned to one of two (or more) groups; the individuals in one group, but not the other, undergo a certain treatment; and ‘if there is a difference in the incidence of the outcome between the two groups, then we conclude that the treatment caused the outcome’ (p. 42). Campbell points out that these ways of finding out about general causation aren't limited to the physical domain. In the social sciences, RCTs are regularly used as a way of finding out about general causation in the mental domain. Furthermore, from a young age we develop our common-sense knowledge of generalities regarding mental causation on the basis of relevant statistical regularities (much of this being done implicitly) (p. 22).

Campbell notes that the practices by which we learn about general causation manifest a close link between our conception of causation and the effect of intervention. He argues, however, that our conception of causation cannot be captured only in terms of interventions – specifically, in terms of counterfactuals concerning what would happen in case of intervention (e.g., as in James Woodward's account, see Woodward, Reference Woodward2003) – since the notion of a causal process (that isn't reducible to counterfactuals concerning interventions) is also part of our conception. He argues in Ch. 1 §4 that certain aspects of our practice of using RCTs – both in the physical and the mental domain – indicate that our conception includes a notion of a causal process. In Ch. 2 he goes on to argue that our conception of singular mental causation requires a conception of a causal mental process. Campbell assumes that a good case has already been made in the literature for holding that we have a conception of a causal physical process. And two of the considerations he puts forward in Ch. 2 involve applying, in the case of singular mental causation, considerations that have been used to argue for a process conception of physical causation. I will focus on the second of these (Ch. 2 §3), which concerns the way in which we acquire knowledge of singular mental causation.

Campbell first highlights the fact that in several cases of singular physical causation we can immediately see that A caused B (e.g., that the cue shot caused the red billiard ball to go into the pocket) by tracking the spatio-temporal paths of objects and the effects of interactions between them (e.g., collisions). Importantly, in such cases we are normally unable to discover the causal relation merely on the basis of general causal claims or laws. Moreover, in some of these cases the circumstances make it rather unlikely that the occurrence of the actual cause would lead to the occurrence of the effect, thus there are no corresponding general causal claims which the singular case is an instance of. (Cases of the latter type have been used in the literature on physical causation to argue that we need a notion of causal process that cannot be analysed in terms of regularities.) Campbell argues that something similar holds for singular mental causation. We have the ability to trace specific mental processes of others by means of imaginative understanding, while we aren't in a position to discover the relevant causal relation (e.g., between what initiated the process and its product) merely on the basis of general causal claims (the most immediate evidence for this is that we don't know general claims that would enable this).Footnote ¹

To explain what he means by imaginative understanding, Campbell appeals to Jaspers’ distinction between ‘subjective and objective psychopathology’. The latter involves reliance on observed regularities, whereas the former is characterised by Jaspers as follows: ‘We sink ourselves into the psychic situation and understand genetically by empathy how one psychic event emerges from another’ (quoted by Campbell, p. 73). This highlights two aspects that are important to Campbell. First, the emphasis is on understanding ‘the dynamics of the mind’ – how one mental event (or state, or any other relevant type of factor) emerges from, or is generated by, another – thus, in this sense, it involves tracing mental processes. Second, the causal relation between the mental events isn't discovered on the basis of known regularities; rather, one imaginatively takes the other person's point of view, in the particular circumstances, in order to understand how that person's mental events generated others (Campbell seems to use ‘imaginatively’ as interchangeable with Jaspers’ ‘by empathy’).

Following Jaspers, Campbell uses ‘meaningful’ to characterise the processes that imaginative understanding makes accessible to us. Note that this isn't meant to give us a handle on which processes are in question, independently of their being processes we can access in this way. To clarify what kinds of processes are in question, Jaspers and Campbell give examples of meaningful processes – e.g., chains of thoughts linked by rational relations, understandable ways experiences generate certain emotions, etc. – contrasting them with chains of thoughts and feelings of which we can't make any sense. Importantly, though, when we are using our capacity for imaginative understanding to find out, say, what caused Billy to decide to accept the job offer, it isn't sufficient to identify mental events in Billy's mind that may generate such a decision in a meaningful way; rather, we are looking for the ones that actually generated the decision. To trace the actual causes, one's exercise of one's imagination has to be well grounded in ‘the tangible facts (that is, […] the verbal contents, cultural factors, people's acts, ways of life, and expressive gestures)’, (Jaspers' words, quoted by Campbell, p. 76). The last point is important for two reasons. First, what we are tracing is the causal process that actually led to the relevant outcome. The existence of meaningful relationships of the relevant types between one's mental events/states (e.g., a belief and desire that justify a certain intention) isn't sufficient for there to be a causal relation between them. Second, grounding in ‘tangible facts’ is what enables us to (correctly) trace the process which actually took place, and thus acquire knowledge of such processes.

With these clarifications in hand, let's return to Campbell's claim that we are able to trace (and thus gain knowledge about) specific mental processes of others by means of imaginative understanding, while we aren't in a position to discover the relevant causal relation merely on the basis of regularities. To persuade the reader that this is the case, he points to familiar everyday examples in which we seem to be able to trace (and thus know about) each other's mental dynamics – e.g., we often take ourselves to be able to follow the line of thought of our interlocutor, or to find out what gave rise to an observed action – though it doesn't seem to us that we discover, or are able to discover, what caused what on the basis of regularities. (Campbell takes it for granted that in all the relevant cases we are applying our capacity for imaginative understanding. As I'll mention later, this raises further questions as to what imaginative understanding involves.) I take it that the reader is expected to recognise that this is so in their own case, and should thus recognise that, in that sense, it is part of our conception of mental causation that there are causal mental processes and that we are able to trace them by imaginative(/empathic) understanding.

Now, one might be persuaded by Campbell that this is indeed part of our conception of mental causation but be suspicious of the claim that imaginative understanding can and does provide us with knowledge of causal mental processes. Moreover, Campbell takes his characterisation of our conception of mental causation to be also a characterisation of mental causation itself – in particular, he also concludes that there are causal mental processes characterised by their meaningfulness. And again, one might wonder what justifies this.

Regarding the latter question, Campbell doesn't explain explicitly how he is thinking about the relationship between our conception of causation and causation itself (such an explanation would have been useful). But the discussion in Ch. 4, pp. 178–80, suggests that Campbell holds something like the following view. Claims about ontological relationships – between what we regard as causal mental processes and causal physical processes or regularities (e.g., that the former are reducible to the latter) – which have no bearing at all on how we are able to gain knowledge about mental ones, or on what such knowledge entails for us (in non-philosophical contexts), do not add anything significant to our conception of what mental causation is. In Campbell's words, such claims belong to ‘a kind of ontological tidying up that tends to preoccupy philosophers but doesn't have much significance beyond that’ (p. 179). This suggests that if our capacity for imaginative understanding can and does provide knowledge of causal processes, and in addition, as Campbell argues in Ch. 2, we have no other epistemic access to occurrences of such processes, then the characteristics of our conception of these causal processes that are manifested in our use of this capacity are characteristics of these causal processes.

I want to focus on the first conjunct of the antecedent of this conditional suggestion. Campbell argues that we take our capacity for imaginative understanding to be knowledge-providing and finds it unproblematic to move from this claim to the claim that this capacity is knowledge-providing. How can such a move be justified? It is possible that Campbell has in mind something roughly like the following. Our taking our capacity for imaginative understanding to be knowledge-providing means, at the very least, (a) that we look for and respond to further tangible facts that support or undermine the results of our exercises of imaginative understanding, and (b) there are cases in which we treat the results of exercises of the capacity (i.e., ascriptions of mental processes to others) as known – that is, we are willing to, and do, rely on them in our interaction with others, often in situations in which the correctness of the ascriptions really matters to us (e.g., when it is relevant to whether we should reconsider our relationship with a person, punish or reward them, etc.). Now, in the physical case, taking our perceptual capacity to be knowledge-providing in these respects (i.e., using it with sensitivity to further perceived aspects, and relying on what we take ourselves to know in our interaction with the environment) seems to constitute a link between (i) our conceptions of physical objects, of their causal interactions, etc., including our conception of how we may gain knowledge about them, and (ii) what such objects, interactions, etc., and our ways of gaining knowledge about them are – a link that grounds the former in the latter. The suggestion is that something similar is true about mental causation.

A natural worry here is that there seems to be a significant disanalogy between the two cases. It seems that our interaction with our physical environment is much more likely to reveal to us mistakes we might make when using our perceptual capacity (e.g., false beliefs formed hastily or on the basis of perceptual illusions) than our interaction with others is likely to reveal mistakes made when we use imaginative understanding. Thus, there seems to be room for arguing that our interaction with others doesn't link our conception of causal mental processes with what it is meant to be a conception of in a way that grounds the former in the latter. There clearly are things to say in response. However, there is no space here to reconstruct and evaluate Campbell's response. Thus, somewhat unfairly, I'll end by indicating, very briefly, some of the questions I encountered when attempting to reconstruct his response.

It seems reasonable that one's knowledge of one's own mental dynamics plays a role in enabling us to use imaginative understanding to find out about the mental dynamics of others – in particular, I'm inclined to think that it plays some role in our grasp of chains of thoughts and feelings as causal, and that at least in some cases it plays a role in making chains of mental events meaningful (in the relevant sense) to one. (This is consistent with denying that we gain knowledge about the mental processes of others by analogy.) Moreover, such self-knowledge seems crucial to the possibility of using others’ reports on their mental dynamics in evaluating our ascriptions to them. However, Campbell focuses, entirely, on our ability to find out about the mental processes of others and doesn't mention self-knowledge in the discussion of this ability. It is thus unclear how Campbell is thinking about knowledge of mental dynamics in one's own case, and why our capacity to have such knowledge doesn't figure at all in the discussion of our conception of causal mental processes. (To emphasise, I'm not suggesting that our capacity for mental self-knowledge is independent of our capacity for knowledge of others’ mental dynamics, but rather that they are mutually dependent.)

Another question concerns what counts as an exercise of the capacity for imaginative understanding. Such exercises are supposed to involve imaginatively taking the other person's point of view (in specific circumstances), and it seems that the clearest examples of doing so involve a conscious exercise of the imagination, explicitly taking into account relevant tangible facts – e.g., a historian attempting to trace the chain of thoughts and feelings that caused the actions of a historical figure. When considering whether we take imaginative understanding to provide us with knowledge (and whether it actually does), cases of this type might not immediately strike us as knowledge-providing. Perhaps this is because, in such cases, we need to uncover and select the tangible facts to be taken into account, and it may seem that, for various reasons, we can easily miss crucial facts. I think that some of the everyday cases that Campbell points to (e.g., following one's interlocutor, understanding what gave rise to an observed action, etc.) provide better examples of cases for which, in non-philosophical contexts, it seems rather implausible to claim that we never have knowledge of singular mental causation. The clearest cases seems to me to be ones in which (i) the relevant aspects of the circumstances are, in some sense, given to us (e.g., we know the other person well, we are both at the same location, focusing together on a certain topic, etc.), (ii) the other person acts in a way which we take to have only one likely mental cause in the circumstances, and (iii) it seems to us immediately clear what that mental cause was. But in such cases, no explicit exercise of imagination seems to be involved. Would Campbell count these as involving the capacity for imaginative understanding? It seems that he should. But if so, it would be good to hear a bit more about the sense in which it can be said to be so.

The last question I'll mention here is about the relationship between imaginative understanding and knowledge of relevant regularities. Campbell highlights the respects in which knowledge of singular causation and knowledge of general causation are independent of each other, and only occasionally mentions ways in which they interact. In discussing imaginative understanding Campbell emphasises that when we exercise the capacity, we don't discover what caused what (merely) on the basis of known regularities. But, arguably, known regularities (perhaps very local, and with ceteris paribus clauses) do play some role in imaginative understanding. The cases, mentioned above, in which we effortlessly ascribe singular mental causes seem to require such general knowledge, and it also seems that our ability to rely on relevant tangible facts presupposes such general knowledge. This seems consistent with holding that we don't discover what caused what (merely) on the basis of known regularities, and with holding that it isn't possible for us to derive singular causal claims from regularities. My guess is that Campbell wouldn't rule out that knowledge of generalities plays some role in imaginative understanding – but this is only a guess.

There are many more interesting issues, points, and arguments in the book. My hope is that I've managed to convey some of the interest, sophistication, and originality of the picture Campbell offers and of the ways in which he argues for it. Anyone with interest in mental causation would benefit from thinking through the various considerations and ideas presented in this book.Footnote ²