- Donald Davidson by Steve Pyke, 1990
Whether reasons are the cause of actions is a debate concerning a central issue in the philosophy of mind. Specifically, the issue at hand is that of mental causal efficacy, i.e. whether the mental can cause events in the physical world. This issue is crucial as it forms one of three elements in the inconsistent triad of the mind-body problem; if the mental is causally efficacious, then, either physics must not be causally closed, or the mental and physical must not be distinct. The debate on mental causation is largely grounded in a Humean assertion that events described as cause and effect must be both logically and temporally distinct, and related by a strict law, perhaps awaiting discovery by “hard” science (Tanney, 2009). Thus, if reasons and actions are causally related, they must fall under this description. In this essay, I will outline an argument from Julia Tanney (ibid) on why reasons may not be the causes of actions and respond to this via Donald Davidson’s Actions, Reasons, and Causes (1963). I will then explore Davidson’s position, known as anomalous monism, in more detail, including its potential shortcomings. I will conclude the essay by considering some relevant work done in psychology and neuroscience.
Julia Tanney’s (2009) position is that reasons and actions are not causally related. Her argument is largely inspired by the work of the later Wittgenstein and, thus, takes a linguistic form. Tanney concedes that, where the terms cause, reason, and explanation, are used interchangeably, denying the causal efficacy of the mental also denies its explanatory power (ibid). This follows very simply from an example: if the mental is not causally efficacious, then the cause of my eating a sandwich cannot be my hunger, which is a mental state. However, it is obviously wrong to take this as meaning that my hunger does not explain my eating of the sandwich. This leads Tanney (ibid) to claim that explanations figuring in everyday mental descriptions (‘reason-explanations’) must be, in some way, different to explanations that appear in physical sciences (‘causal explanations’). Her argument begins by stating that reason-explanations serve to characterise performances or events as actions (i.e. performances or events behind which there is some intention). Indeed, there is no purely physical re-description possible that is capable of characterising actions as intentional, including neurological descriptions of brain events. Further, she suggests that the Humean construction of the cause-and-effect relationship requires that the effect-event be fully characterised already and that the causal relationship merely explain the chain of events that leads to the effect in question. Thus, reason-explanations are, by definition, different to causal explanations, as they serve to re-characterise the effect in question (ibid). Tanney then clarifies that reason-explanations are ‘non-causal, context-placing explanations’ (ibid). She uses the example of a student, leaving the class early, being confused as to why a teacher wrote the word ‘cat’ on the board. This is explained by another student: ‘ “Because she was writing ‘catalyst’ – you left the room before she completed the word.” ’ (ibid). In this example, only one puzzling event occurs (the writing of ‘catalyst’), and the event is explained by placing it into a context in which it is no longer confusing. Thus, although the writing of the word ‘catalyst’ in a linguistic sense “caused” the writing of the word ‘cat’, these two events are not logically or temporally distinct and, thus, do not fit under the Humean description of cause-and-effect. Writing ‘catalyst’ still explains the writing of the word ‘cat’, however. This, Tanney (ibid) argues, can be taken to be the same kind of explanation that occurs when reasons are used to explain actions. Hence, even where explanations ascribe beliefs, desires, or any other mental states, these need not be causal (in a Humean sense) in order for a successful explanation to be made. Tanney points out that assuming reasons are ‘logically independent, temporally antecedent, causally efficacious events’ (ibid) commits the philosopher to postulating the existence of an event that is unobservable, thereby making the hypothesis untestable. For example, if we were to monitor a subject’s brain activity with the view of forming neuro-mental bi-conditional bridging laws for reasoning, to form correlations we would require some way of empirically verifying correct reasoning other than asking the subject – since, how else would we know if the subject was herself correct in describing her own reasons (ibid)? Tanney thus argues that reasons are not causal concepts awaiting discovery, since they ‘discharge their explanatory role without designating anything; let alone causally efficacious states or events’ (ibid). Whilst compelling, Tanney’s position leaves us with an important question: if reasons aren’t causal, then what is the cause of our actions? To answer this, one may use Tanney’s arguments to augment an eliminative materialist position – although she resists this temptation herself, insisting that her argument is purely linguistic (I shall return to this idea later). Alternatively, Tanney’s hypothesis may be rejected entirely.
Donald Davidson’s influential article Actions, Reasons, and Causes (1963) anticipates several of the points made by Tanney, although perhaps unsuccessfully. Davidson (ibid) argues that reasons are the causes of actions. Specifically, ‘primary reasons’ – the combination of an agent’s ‘pro-attitude’ towards an action and ‘believing… that his action is of that kind’ – are causal. His claim makes a distinction between normative reasons (that a person may have for acting) and motivating reasons (that lead to the action) (Alvarez & Zalta, 2016). To illustrate this distinction, suppose a driver signals with their arm upon approaching a turn: clearly the driver has a normative reason to signal (he is approaching the turn), however, in this case, the driver’s motivation for signalling is that he sees his friend on the corner, and not the imminent turn. This example shows how a person can have non-causal normative reasons for acting. Hence, says Davidson (1963), correct reason-explanations must entail a causal relationship. Additionally, Davidson (ibid) rejects the notion that correct reason-explanations merely re-contextualise the action by insisting that primary reasons must also correctly justify the action. He asserts that explaining action via re-contextualisation is not sufficient since, firstly, the context contains both the reason and the action and, secondly, actions can occur outside their usual context and may require further non-contextual explanation (ibid). Tanney’s (2009) response to this argument is that, in the case of incorrect reason-explanations, the context is merely poorly understood or insufficiently explained. She also asserts that the possibility of incorrect reason-explanations is irrelevant to the fact that these explanations function differently in a linguistic sense to causal explanations (ibid). It isn’t clear whether either position is more correct here, since Davidson’s insistence that primary reasons are correct, justifying causal reasons, seems to reject the premises of Tanney’s argument. If Davidson’s position is assumed to be correct, however, it must show how reasons and actions are related in a law-like sense via the Humean cause-and-effect model. He addresses this via his larger philosophical position: anomalous monism.
Davidson describes anomalous monism in his work, Mental Events (1970). This position is characterised by three postulates:
- At least some mental events interact causally with physical events;
- Events related as cause and effect fall under as precise, explicit, and exceptionless laws as possible; and
- There are no strict laws relating mental events and their relationship to physical events.
At first glance, these three postulates appear incompatible; according to the Humean cause-and-effect criteria; how can the mental cause, yet not be relatable to, the physical? To explain, Davidson (1963) shows that there is no need to name a specific law for a full causal explanation. For example, suppose that a hurricane, reported online by Reuters, causes a major catastrophe, that is then reported on the BBC evening news. In this example, there are two distinct causes (the hurricane and the ensuing catastrophe) and two distinct effects (the two reports), yet there is no strict law relating the kinds of causes to the kinds of effects: the catastrophe caused the BBC evening report, but this isn’t subject to a strict law, since the BBC may have reported it online instead, or even not reported it at all. Davidson (ibid) argues that this is also true of reasons and action – that the absence of a strict law relating the reason to the action does not require that the reason be non-causal. This explanation is still insufficient, however, in explaining how the mental and physical might interact causally. To address this, in Mental Events Davidson (1970) defines mental events as those which come under a mental description. He then claims that every mental event is identical with some physical event, but that physical events cannot be ascribed to mental events in a law-like fashion. In doing so, Davidson arrives at a reductionist, token-identity, supervenience theory of mind; the mental is ‘supervenient’ on the physical in that it depends upon it, yet cannot be used to predict it (Yalowitz & Zalta, 2005). Hence, according to Davidson (1970), primary reasons can be granted causal efficacy since both pro-attitudes and their related beliefs are mental events supervenient on events that are clearly causal – namely, physical events. By associating mental events to physical ones in this way, Davidson appears to grant mental events causal efficacy whilst maintaining that physics is causally closed and that the mental and physical are distinct; an apparent solution to the inconsistent triad of the philosophy of mind. This is a problematic position, however.
Jaegwon Kim, in The Many Problems of Mental Causation (2002), summarises the problems with Davidson’s position well. Perhaps his most significant objection is that, by identifying every mental token as supervenient on some physical token, the causal relevance of mental events is doubtful. For example, if you take an aspirin because you are in the mental state of having a headache, which is supervenient on some non-descript physical (neurological) state, then, is the actual cause of you taking the aspirin the mental state or the neurological state? If mental states are granted causal efficacy by being supervenient on physical events, then it appears, in the above example, that the neurological state caused you to take the aspirin. If this is true, then the mental state has no causal relevance (Heil, 2003). A similar point can be argued by taking advantage of the multiple-realisability implicit in Davidson’s supervenience model; if one mental state can be realised by multiple, distinct physical states, and these physical states have distinct causes, then the causal power of the mental is dependent upon its supervenience on the physical. Hence, it doesn’t make sense to grant causal power to the mental in the first place (Kim, 2002). Another of Kim’s objections is called ‘The Problem of Extrinsic Mental Properties,’ (ibid) and essentially adopts a Wittgensteinian argument that, in order for mental states (such as belief, or desire) to differ in their content, they must also be contextually dependent. For example, consider two frogs from different planets: these may be atomically identical creatures and, when subject to a dark patch projected on a screen, react identically by licking the screen in search of prey – that is, they are in identical physical and mental states. However, whilst one frog – by virtue of having lived on Earth – is under the illusion that the dark patch is a fly, the other frog – by virtue of having lived another planet – may be under the illusion that the dark patch is a ‘schmy’ (a tiny bat-like creature that constitutes the second frog’s prey on the other planet). Thus, two identical frogs, in identical physical states, have representational mental states that demonstrably differ in their content. Since mental states are contextually dependent then it follows that reasons are also contextually dependent, which implies that the mental is not intrinsically causally efficacious (ibid). These three arguments show that Davidson’s model, whilst initially compelling, does not sufficiently illustrate that the mental is causally efficacious and, thus, that reasons can be the causes of actions.
It is possible that research done in the fields of psychology and neuroscience may resolve the debate as to the causal efficacy of the mind. There have been many interesting investigations to this end, but none of them are wholly conclusive. The work of Michael Gazzaniga on split-brain patients is particularly interesting in this regard. Split-brain patients have had the corpus collosum – the part of the brain joining the two hemispheres – severed to some degree, and exhibit unusual behaviours. By examining these patients, Gazzaniga, Wolford & Miller (2000) were able to identify that the visual fields from the left and right eyes are independent and administered to by the right and left brain hemispheres respectively, and that the left hemisphere of the brain is responsible for language operations and narrative creation. In one experiment, when split-brain patients were asked to explain decisions they made based on images isolated in a single visual field, they were only able to form narratives based on the image processed by the left hemisphere of the brain. In order to explain decisions based on images in the other visual field, the subjects ‘confabulated’ their responses – they invented demonstrably incorrect reasons for their choices to suit their purposes (Gazzaniga & LeDoux, 1978). This suggests that their reasons did not cause their actions but rather served as a linguistic context-placing tool, in agreement with Tanney’s model. A similar example is that of brain-surgery patients who, when undergoing electrical stimulation to their motor-cortex, confabulated reasons for moving their limbs such as ‘to gain your attention’, despite these movements being caused by the doctors administering the electrical signals (Hirstein, 2009). Further experiments have shown that, when unaware of their movements, patients would not report feeling a desire to move when the motor-cortex is electrically stimulated (Yong, 2009). This suggests that the reasons patients gave for moving when their motor-cortices were stimulated were not the result of electro-stimulated ‘desires’ in the brain but were, perhaps, merely context-placing explanations. On the neuroscientific front, studies of biological and artificial neural nets indicate that mental concepts are representable as partitions in multi-dimensional vector spaces generated by synaptic activation; a concept compliant with dynamical systems theory (Bickle et al, 1999). Patricia Churchland (1990) and other eliminativists have taken this neuroscientific research as indicating that there is no semantic structure consistent with concepts such as belief in the brain. Since Davidson’s position is contingent on the existence of beliefs, and their correlation to physical events, this casts doubt on his position. All of these investigations appear to support Tanney’s position, however, none of them are conclusive or necessarily complimentary. Nonetheless, they provide interesting perspective on the work of the two philosophers considered in this essay.
There is no definitive answer as to whether reasons are the cause of actions. Tanney’s account of reasons as non-causal, context-placing explanations provides us with a compelling mechanism to dispute the causal nature of the mental. However, by insisting that reasons must be correctly justificatory (considering that actions are multiply-realizable), Davidson somewhat dodges Tanney’s objections. His position, however, as a reductionist, token-identity, supervenience of mind, does not sufficiently show the causal relevance of the mental or how supervenience theory can account for the contextual dependence of mental content. Work done in psychology and neuroscience, particularly the study of split-brain patients and other confabulation examples, and studies of neural nets, appear to support the notion that reasons play a linguistic role only. These studies aren’t entirely conclusive, however. Perhaps the only conclusion to be drawn from this debate is that a marriage of neuroscience and philosophy is necessary for an eventual complete account of mental causation.
Alvarez, M. & Zalta, E.N. (ed.) (2016) Reasons for Action: Justification, Motivation, Explanation. The Stanford Encyclopedia of Philosophy (Winter Edition 2016). Available from: https://plato.stanford.edu/entries/reasons-just-vs-expl/[Accessed 21st April 2017].
Bickle, J., Mandik, P., Landreth, A. & Zalta, E.N. (ed.) (1999) The Philosophy of Neuroscience. The Stanford Encyclopedia of Philosophy (Summer 2012 Edition). Available from: https://plato.stanford.edu/entries/neuroscience/ [Accessed 21st April 2017].
Churchland, P.S. (1990) Neurophilosophy: Toward a Unified Theory of the Mind/Brain. Journal of Philosophy. 87 (2):93-102. Available from: doi: 10.2307/2026883 [Accessed 25th April 2017].
Davidson, D. (1963) Actions, Reasons and Causes. The Journal of Philosophy, 60(23), pp.685-700. Available fromt: https://bibliotecamathom.files.wordpress.com/2012/09/actions-reasons-and-causes.pdf [Accessed 21st April 2017].
Davidson, D. (1970) Mental Events In: Rosenthal, D.M. (ed.) The Nature of Mind. New York, Oxford University Press, pp. 247-256.
Gazzaniga, M. & LeDoux, J. (1978) The Integrated Mind. 1st ed. New York: Spring Science+Business Media. Available from: https://books.google.co.uk/books?hl=en&lr=&id=dhhRBAAAQBAJ&oi=fnd&pg=PT6&dq=Gazzaniga,+M.+S.+and+LeDoux,+J.+E.+1978.+The+integrated+mind&ots=-HFHQZyZmu&sig=XeR-IQzKsPYwhsO6fKMNkGLm5tU – v=onepage&q=chicken&f=false [Accessed 21st April 2017].
Gazzaniga, M., Wolford, G., & Miller, M.B. (2000) The Left Hemisphere’s Role in Hypothesis Formation. The Journal of Neuroscience. 20 (6) RC64. Available from: http://www.jneurosci.org/content/jneuro/20/6/RC64.full.pdf [Accessed 25th April 2017].
Gordon, R.M. & Zalta, E.N. (ed.) (2009) Folk Psychology as Mental Simulation. The Stanford Encyclopedia of Philosophy (Fall 2009 Edition). Available from: https://plato.stanford.edu/archives/fall2009/entries/folkpsych-simulation/ [Accessed 20th December 2016].
Heil, J. (2003) Mental Causation. In: Stich, S. P. & Warfield, T. A. (eds.) The Blackwell Guide to Philosophy of Mind. Malden, MA, USA, Blackwell Publishing, pp215-235. Available from: doi:10.1002/9780470998762.ch9 [Accessed 25th April 2017].
Hirstein, W. (ed.) (2009). Confabulation: Views From Neuroscience, Psychiatry, Psychology and Philosophy. Oxford, Oxford University Press.
Kim, J. (2002) The Many Problems on Mental Causation (Excerpt). In Chalmers, D.J. (ed.) Philosophy of Mind: Classical and Contemporary Readings. Oxford, Oxford University Press.
Malpas, J. & Zalta, E.N. (ed.) (2015) Donald Davidson. The Stanford Encyclopedia of Philosophy (Fall 2015 Edition). Available from: https://plato.stanford.edu/entries/davidson/ [Accessed 25th April 2016].
Tanney, J. (2009) Reasons as Non-causal, Context-placing Explanations. In: Sandis, C. (ed.) New Essays on the Explanation of Action. London, Palgrave Macmillan UK, pp. 94-111. Available from: doi:10.1057/9780230582972 [Accessed 25th April 2017]
Yalowitz, S. & Zalta, E.N. (ed.) (2005) Anomalous Monism. The Stanford Encyclopedia of Philosophy (Winter 2014 Edition). Available from: https://plato.stanford.edu/entries/anomalous-monism/ [Accessed 21st April 2017].
Yong, E. (2009). Electrical stimulation produces feelings of free will. National Geographic. Available from: http://phenomena.nationalgeographic.com/2009/05/07/electrical-stimulation-produces-feelings-of-free-will/ [Accessed 21st April 2017].