Philosophy delights in those aspects of the world that initially seem obvious and natural, but which on reflection turn out to be deeply mysterious. The mental causation of behaviour is one such phenomenon. Nothing could be more obvious than that our minds matter—that our beliefs and desires, and our perceptions and thoughts ultimately have a causal impact on our behaviour. And yet it has proved notoriously difficult to explain just how this could be the case.
The problem of mental causation has morphed and fragmented over the years. In its original guise, it was the problem of how a non-physical mental substance or property could causally interact with the physical brain. The obvious solution to this version of the problem was to adopt a thorough-going materialism of some kind, with the consequence that mental phenomena are identified with properties of the brain from which they inherit their causal efficacy.
With the advent of functionalism in the later years of the last century, this “obvious” solution ran into difficulties. If mental phenomena are multiply-realizable, as the orthodox construal of this metaphysical position seems to imply, then mental properties can’t be identified with properties of the brain after all; and since the latter do all the causal work insofar as behaviour is concerned, the problem of mental causation re-asserts itself in a different form (Kim 1992; Crane 1995). This version of the problem of mental causation, which seems to generalise beyond the realm of the mental to all multiply-realizable phenomena, is still keenly debated in philosophy (Kim 2000, 2005; Hohwy 2008).
There is yet another rendering of the problem, however, that revolves around the causal efficacy of the specifically representational properties of mental phenomena. This third version typically arises in the philosophy of mind from the conjunction of three widely accepted theses about mental phenomena and their physical realization in the brain:
The content causation problem
Mental phenomena are causally efficacious of behaviour in virtue of their representational contents.
The representational contents of mental phenomena are not determined by the intrinsic properties of the brain.
The brain is causally efficacious of behaviour in virtue of its intrinsic properties.
The first of these theses is a fundamental tenet of both folk psychology and the computational theory of mind that has been constructed on its foundations. It is simply common sense that our perceptions and thoughts are about various aspects of the world in which we are embedded. It is also commonsense that mental phenomena causally interact with other mental phenomena and bodily behaviour in a fashion determined by their content—i.e., how they represent the world as being. Fodor refers to this as the “parallelism between content and causal relations” (1987).
The second thesis is widely accepted because most contemporary philosophers think that the representational properties of mental phenomena are determined at least in part by factors beyond the brain. This is the conclusion drawn from a number of famous thought experiments implicating twin-earth, arthritis, and various species of tree (Putnam 1975; Burge 1979, 1986). But, even more importantly, the second thesis seems to be an entailment of the most popular approach among philosophers for explaining how the representational properties of mental phenomena are determined. This is the conjecture that mental phenomena are contentful in virtue of their causal relations with those aspects of the world they are about (Adams & Aizawa 2010).
The final thesis is consistent with all we know about the brain basis of behavioural causation. While the brain enters into complex causal relations with aspects of the environment via multifarious sensory channels, our best neuroscience informs us that the changes to musculature that constitute our behavioural responses are wholly determined by the intrinsic properties of the brain to which they are causally connected.
In conjunction, these three widely accepted theses form an inconsistent triad. This generates a distinct and narrower version of the problem of mental causation: How can mental phenomena be causally efficacious of behaviour in virtue of their representational contents if these contents are not determined by intrinsic properties of the brain? In what follows, I shall refer to this as the content causation problem. This is the version of the problem of mental causation with which I shall be concerned in this paper (see e.g., Kim 2006, pp. 200–202).
There are some philosophers who seek to resolve the content causation problem by rejecting either the first[1] or the third[2] of the theses composing the inconsistent triad. However, the most popular response has been to reject or at least modify the second thesis. This leads to the the narrow content program:
The project of developing an account of mental phenomena according to which (at least the causally relevant component of) their representational properties are determined by intrinsic properties of the brain.
There are a number of different proposals about narrow content in the literature. Two of these have been particularly prominent. One is Fodor’s suggestion that narrow contents can be unpacked as “functions from contexts to truth conditions” (1987, Ch. 2). The other is that narrow content is determined by “short-armed functional roles” (Block 1986; Loar 1981, 1982). But these (and other) proposals have been roundly criticised for failing to capture the relational character of mental content:
The main charge has been that narrow content, as construed in these accounts, is not real content. When one thinks of an apple, what one thinks about is not a role or a function, but a fruit. Real content must put the subject in cognitive contact with the external world. […] A water concept, for example, must involve a relation between the thought wherein the concept is deployed and some worldly property or kind, presumably having to do with water. The problem with narrow content, construed as short-armed functional role or as a function from contexts to wide contents, is that it is not clear how it could involve any such relation. (Kriegel 2008, p. 308)
At this point, however, we seem to butt up against a classic paradox. On the one hand, those theories that appear to capture the relational character of mental content (i.e., causal theories) hold that content is not wholly determined by the intrinsic properties of the brain and, hence, imply that it isn’t causally efficacious of behaviour. On the other hand, theories with the potential to account for the causal efficacy of mental content (i.e., narrow content theories), fail to capture its relational character. A solution to the content causation problem thus requires something that prima facie appears impossible: an explanation of the relational character of mental content that invokes only the intrinsic properties of the brain. Little wonder then that many philosophers despair of ever finding a solution to this puzzle.[3]
It is reasonable to hazard, however, that one of the main barriers standing in the way of a more productive treatment of the content causation problem is the radically underdeveloped understanding of mental content with which contemporary philosophy operates. In the foregoing quotation, for example, Kriegel characterises the relational character of content in terms of a subject’s “cognitive contact” with the external world; yet he readily admits elsewhere that this notion is “not altogether transparent” (Kriegel 2008, p. 305). This is typical of the literature on this topic, which has become accustomed to describing content using the notoriously vague language of aboutness. While this language might capture our commonsense intuitions about mental phenomena, its imprecision may prevent us from discerning the lineaments of candidate solutions to the content causation problem.
This last point, at least, gives us the motivation for intruding yet another discussion into this already crowded philosophical space. The foundational conjecture upon which this paper is based is that the apparent insolubility of the content causation problem issues from an impoverished and unenlightening account of the relational character of mental content. Furthermore, this impoverishment is largely a consequence of the dyadic conception of mental representation that has hitherto conditioned most philosophical thinking in this area. By contrast, a minority of philosophers has argued that mental representation is more properly analysed as a triadic relation. Triadicity, I will argue, yields a richer and ultimately more illuminating account of the relational character of mental content. Armed with this alternative treatment, we are in a position to assess the content causation problem anew. On the one hand, this novel viewpoint confirms the worry philosophers have expressed that causal theories of mental content are impossible to reconcile with the brain-based causation of behaviour. On the other hand, and much more positively, the triadic conception reveals a path that, from the perspective of content causation at least, looks more promising. The proposal that we travel down this path will undoubtedly face resistance, since it requires us to rehabilitate an approach to mental content that is unpopular in contemporary philosophy. But this approach, I shall conclude, seems unavoidable if we are to explain how mind matters.