Twenty years ago, when I first became involved in the theology and science dialogue, it was possible to ask whether there was really anything for scientists and theologians to talk about. It is important to remember that some of the most powerful influences in the development of modern theology, such as Immanuel Kant and Friedrich Schleiermacher, had argued that religion and science have nothing to do with one another.

Various intellectual strategies for insulating theology from science have made use of body-soul dualism. Put crudely, science can study the body but the soul is the province of theology. Such strategies, however, have become problematic in that neuroscientists are now studying all of the human faculties once attributed to the soul.

I would argue that Christians who have not already done so ought to join philosophers and neuroscientists in adopting a physicalist account of the person [1]. The problems with dualism, in my judgment, are insurmountable. First, it may well be conceptually impossible to give an account of mind-body interaction--how can something non-material interact causally with material entities? Second, while neuroscience can never prove that there is no mind or soul, it is increasingly clear that, to quote Laplace out of context, we have no need of that hypothesis. Finally, in addition to being unnecessary on biblical or theological grounds, dualism is theologically undesirable due to its penchant for distorting Christian priorities. Briefly, what I mean here is that the adoption of dualism gave Christians something to care about (their souls) in place of Jesus' primary concern, which was the Kingdom of God.

There are problems with physicalism, also. Most of the problems come down, in one way or another, to the issue of reductionism [2]. If humans are essentially bodies, can we still understand ourselves to have features once attributed to an immaterial mind or soul, such as rationality, morality, and free will?

What I intend to do this evening is to suggest the outlines of an approach to the problem of rationality. Here is the problem in brief: If humans are purely physical entities, how can it fail to be the case that their thoughts are determined by physical laws, and if so, what happens to our conception of rationality? This problem is discussed in the philosophical literature as the problem of mental causation. Philosopher of mind Jaegwon Kim expresses it as a dilemma: he argues that mental properties will turn out to be reducible to physical properties unless one countenances some sort of downward causation. But such downward efficacy of the mental would suggest an ontological status for the mental that verges on dualism [3]. My plan in this paper is to suggest a strategy for solving the problem of mental causation, exactly by providing an account of the downward efficacy of the mental that leaves an ontologically physicalist account of the human person intact.

Problems with Nonreductive Physicalism

Figure 1 
(click to enlarge). The problem confronting the physicalist can be evoked using a simple diagram (see Figure 1).

Here M1 and M2 represent mental states or properties; P1 and P2 represent physical states or properties [4]. The arrow from P1 to P2 represents a causal relation, and I use the dollar sign to represent the supervenience relation. The concept of supervenience is widely used in philosophy of mind to describe the relation between mental events or properties and brain events or properties. It was first employed to describe the relation between moral and descriptive attributes, and so we might call it a constitutive or "in virtue of" relation. St. Francis has the moral property of goodness in virtue of his particular set of actions and character traits, and hence the property of being good supervenes on these non-moral properties. Similarly, it is hypothesized, one is experiencing mental event M1 in virtue of undergoing a particular set of brain events, P1. There is considerable controversy over an exact definition of 'supervenience' but for present purposes I need not pursue this [5].

The diagram, then, expresses the assumption of causal closure at the physical level--that is, every physical event (in this case, the neurobiological event P2) has a sufficient physical cause. What, then, of the relation between the mental events, represented by the dashed arrow and the question mark?

The dilemma for the nonreductive physicalist comes down to this. Mental properties can be taken to have causal efficacy insofar as they supervene on physical properties and those subvenient physical properties are causally efficacious. But if the physical properties are causally efficacious, what causal work is left for the mental properties? We seem to be left with a new version of epiphenomenalism. Thus, I intend here to sketch out the basics of an argument for the compatibility of reasoned connections at the mental level with causal connections at the neurobiological level. To do so I shall turn, eventually, to the concept of downward causation.

First, let me make it clear that I am reframing Kim's question. Kim speaks in terms of mental and physical properties of events: if the physical property is causally sufficient, what is left for the mental property to do? I want to argue that this way of describing the problem misses the crucial issue. The crucial issue is whether the sequence from M1 to M2 is a reasoned sequence or merely a causal sequence. So, for example, I say to you "5 times 7." You all think "35." Did that happen because it's true that 5 X 7 = 35 or because a causal process in your brain made you think it?

Given that we presuppose the truth of 5 X 7 = 35, that is, that it is rational to think "35" when I say "5 times 7," we can again reframe the question: how can we reconcile an account in terms of reasons with a physicalist account of the mental without giving up on the causal closure of the physical? Colin McGinn asks: "How, for example, does modus ponens get its grip on the causal transitions between mental states?" [6]. I would rephrase his question as follows: "How does modus ponens get its grip on the causal transitions between brain states?"

Some Resources from Philosophy and Neuroscience

So I have identified the question of mental causation with the question of whether sequences of events (events simultaneously mental and neural) fall into patterns that are recognizably rational. And if so, how can we reconcile this with the supposed causal closure of the physical world? A hint about where I am going in this paper: Notice that a calculator obeys the laws of physics and the laws of arithmetic. This is because it has been built in such a way that its causal processes model arithmetic transformations. The calculator has been structured in such a way that (any token instance of) a series of triggering causes--pressing the "5" key, the "times" key, the "7" and the "equals"--causes the machine to display a "35."

Fred Dretske has introduced a distinction between triggering and structuring causes, which he illustrates by means of the following example: "A terrorist plants a bomb in the general's car. The bomb sits there for days until the general gets in his car and turns the key. . . . The bomb is detonated (triggered by turning the key in the ignition) and the general is killed." The terrorist's action was the structuring cause, the cause of its being the case that turning the key sets off the bomb [7].

So for many purposes it is an oversimplification to represent a causal sequence simply as a series of events: E1 --> E2 --> E3. Instead we need to think of two series of events: those leading up to the triggering of the effect as well as those leading up to the condition under which T is able to cause E. Figure 2, adapted from Dretske's diagram, is intended to represent these intersecting strings of triggering and structuring causes:

Figure 2 
(click to enlarge).

Here the Ts represent a series of triggering causes and the Ss represent a series of structuring causes leading to the ongoing condition C such that T is able to cause the effect E.

We shall have come a long way toward the goal of this paper if we can explain how the brain becomes structured so that, in the happy case, causal processes model rational relations. So the question is, are there significant enough analogies between the human brain and a calculator such that we can plausibly assume that the "wetware" has been structured in such a way that its causal processes model or instantiate rational sequences? The disanalogy, of course, is that the calculator has been intentionally designed by a rational agent. Can we provide a plausible account, based on what we now know of neurobiology, as to how such rational structuring might occur without having to presuppose rational agency? I believe that the answer is yes. But here my account necessarily becomes somewhat speculative due to the tentativeness and incompleteness of neuroscientific explanations.

The physicalist assumption is that a mental event, such as thinking of the number 5 or thinking of Grandma, supervenes on a neural event. I chose the thought of Grandma as an example in order to allude to a controversy in neuroscience. The question was whether it was reasonable to assume that brains come equipped with individual neurons designated for recognizing patterns--that is, whether there is a "grandmother neuron" devoted to recognition of this one particular elderly woman, and other cells for each pattern that the brain is able to distinguish.

It is now believed that most recognition tasks depend on the activation of large nets or assemblies of neurons rather than on the firing of individual neurons. Here is an account by neuroscientist Alwyn Scott:

[L]et us consider what may appear to be a simple memory: that of your grandmother. Most of us are conscious of our grandmothers. But how? What series of neural links, of connections to connections, allow us to conjure up those dear old dames?

Presumably an image of grandma may come to mind, and that would appear to involve the optic lobes. But other parts of the memory relating to voice would have originated, presumably, in the temporal lobes. These recollections are connected to others related to things she said and did, the way her house smelled on Thanksgiving Day, the colors of her kitchen, and so on. Because the memory of your grandmother is no doubt imbued with emotional overtones, those cells, whose locations are not known, would also need to be activated. And finally there is the not inconsequential linguistic task of matching the word "grandmother" to one elderly or even long-deceased human female who happened to be the mother of one of your parents. It is difficult, if not impossible, to see how a single so-called "grandmother" cell would manage to bind all of these components of a complex memory together [8].


Thus, even the simplest of mental events probably supervenes on the activation of a vast network of interconnected neurons. The concept of a "cell assembly" was introduced by Donald Hebb, and its formation is described as follows: "Any frequently repeated, particular stimulation will lead to the slow development of a 'cell-assembly,' a diffuse structure comprising cells . . . capable of acting briefly as a closed system. . . . [9]"

It is in the training of such assemblies that we begin to see downward causation. It is better described as downward causation from the environment to the brain rather than mental causation, but insofar as intentionality or reference is an essential ingredient in rationality we have here the beginnings of an account of the rational structuring of the brain. Before pursuing this line of thought, however, we need to explore the concept of downward causation, a concept that Arthur Peacocke has made familiar in the theology and science literature.

Footnotes for the first three sections

[1] There is controversy over the proper term to use for an account of the person that denies that we are composites of body and something else. 'Physicalism' is the term most often used by philosophers who hold this view.

[2] 'Reductionism has a variety of related meanings. Methodological reductionism is a strategy in science that seeks to understand an entity by studying its parts. Epistemological or theoretical reductionism is the thesis that sciences above physics in the hierarchy of the sciences can or should be related to lower-level sciences by means of definitions or bridge laws. The important issue here is causal reductionism, the thesis that the behavior of an entity is determined by the behavior of its parts.

[3] See Jaegwon Kim, "The Myth of Nonreductive Materialism," in The Mind-Body Problem, Richard Warren and Tadeusz Szubka, eds., (Oxford: Basil Blackwell, 1994), 242-60.

[4] The relations among events, properties or states of events or entities, and descriptions all need to be worked out more carefully. I shall not attempt to do so here. I merely note that there may be a problem with Kim's use of these terms: For Kim an event is the instantiation of a property at a time. But if there is a supervenient mental property and a physical property, both instantiated at t, how are we to know what is the relation between M1 and P1? Is it identity? If so we seem to face all of the well-known problems with the various versions of the mind-brain identity thesis. If it is mere correlation (as Kim's definitions of supervenience would suggest) then we seem to have psychophysical parallelism or a new version of epiphenomenalism unless M1 is hypothesized to play a causal role in producing M2/P2. But this looks suspiciously like mind- brain interactionism. So I am inclined to emphasize that there are only two events here, neutrally referred to as e1 and e2, but susceptible of both mental and physical descriptions. To argue this, though, I would need to spell out, first, an account of supervenience different from Kim's, and second, deal with the problem of the relation between properties and descriptions.

[5] Supervenience is understood by Kim and many others in terms of co-variation of properties: there can be no mental difference without a physical difference. Let me insert a demurral here: this account of supervenience does not seem to me to capture the original sense of the term--the dependence of the supervenient on the subvenient--and it does seem to ensure reducibility of supervenient properties.

[6] Colin McGinn, "Consciousness and Content," in Ned Block, Owen Flanagan, and Güven Güzeldere, eds., The Nature of Consciousness: Philosophical Debates (Cambridge: Cambridge University Press, 1997), 255-307; 305.

[7] Fred Dretske, "Mental Events as Structuring Causes of Behavior," in John Heil and Alfred Mele, eds., Mental Causation (Oxford: Clarendon Press, 1995), 121-136; 122-3.

[8] Alwyn Scott, Stairway to the Mind: The Controversial New Science of Consciousness (New York: Springer-Verlag, 1995), 78.

[9] Quoted by Scott, op. cit., 81.