Seeking out Experience in the Causal Cracks
The following is on the theme of nomological danglers, namely laws that are imagined to explain the correlation between physical and mental states. These laws dangle in the sense that mental states, or their characteristic properties, are taken to be irreducible to the physical, and so any such laws would have to be brute – nothing can be said about how these properties depend on or arise from physical states. Herbert Feigl and J.J. Smart, who coined the term, thought that the existence of nomological danglers is implausible (Feigl 1958; Smart 1959). They are ugly addenda to an otherwise complementary set of natural laws. In addition they would have to be laws that relate complex brain processes to experience, whereas all other brute laws are fundamental, that is, they concern the basic constituents of nature, e.g., photons.
‘Nomological danglers’ names the theme insofar as the notion looms large in the background, though it is not addressed directly. Feigl and Smart saw the implausibility of nomological danglers as evidence in support of their identity theories. Below I do not defend any such theory. Instead I have the negative aim of showing that property dualists cannot offer a way of understanding experience as a natural phenomenon despite their assumption that it is. One of the most prominent exponents of property dualism is David Chalmers, for example. Assuming its epiphenomenal implications Chalmers notes that property dualism would in turn imply that any laws linking experience to the physical are likely to be nomological danglers. However, he does not see this as a fatal difficulty, remarking that rather than viewing experience as “floating free of processing in some way; a better picture…is a picture of experience sitting down among the causal cracks.” (1996, p. 160) Below I argue that, to extend the metaphor, the non-physicality of experience held by property dualism, because it entails that its characteristic properties are not part of the causal net, leads to experience slipping through the ‘cracks’ and escaping. How experience is a natural phenomenon becomes intolerably mysterious.
From a purely metaphysical perspective, where experience or consciousness (these terms are treated as synonymous) is abstracted from the observable details of the world that concern science, talk of its being non-physical is theoretically viable. In fact many defend some sort of property dualism. However, most agree that experience is a natural phenomenon, and as such they suppose that whatever theory of consciousness one might come up with, it must fit with our overall scientific worldview. The salient aspect of this worldview is the belief that all natural phenomena are open to scientific inspection, that is, nothing in nature can escape its scrutiny. That experiences are an exception in this way is intolerable insofar as it isolates them from the rest of nature. Now, property dualists too are naturalists in this way. However, scientific inquiry concerns the task of making nature intelligible by relating natural phenomena to one another. It is a basic assumption of this worldview that nothing in nature occurs in isolation and so all natural phenomena can be explained in terms of each other, without appeal to any extranatural causation. Experiential properties, thought of as non-physical, do escape the scrutiny of science in this sense and accordingly cannot be understood as natural.
Section 2 begins by noting that, in agreement with naturalism, all natural phenomena must stand inside the causal nexus. I then look at the relationship between experiential properties understood as non-physical and causation. While it is generally accepted that their non-physicality would make them causally irrelevant, Chalmers has argued that this is not necessarily the case. In reply I argue that there is little evidence to support his claim.
In section 3 I consider the suggestion that even if experiential properties are causally irrelevant they can still be thought of as standing within the causal nexus so long as we understand them as epiphenomena, and hence at least caused by the physical. In reply to this suggestion I argue that the cause-based notion of epiphenomenalism is questionable. Extending John Lachs’s original argument against this notion, I argue that any property of a physical body must have some causal efficacy, contrary to the epiphenomenalist’s claim that experiential properties have none at all. Also I consider Chalmers’s suggestion that perhaps experiential properties are brute, that is, they are not caused but are basic features of bodies originating with the fundamental entities posited by physics. I argue that there is no possibility of confirming or disconfirming this hypothesis. It is therefore scientifically useless and it cannot help us understand experience as a natural phenomenon.
In section 4 I take up the idea that experiential properties are best thought of as supervenient, and this at least allows us to understand them as natural. I argue that the metaphysical notion of psychophysical supervenience this idea is based on, as opposed to Donald Davidson’s original conceptual notion, still does not allow us to understand experience as a natural phenomenon. Supervenience offers no way of connecting experiential properties to other phenomena in this sense. I briefly consider John Searle’s conclusion that we cannot expect to relate experience to other phenomena in virtue of its subjective ontology. In reply I contend that understanding any phenomenon as natural requires relating it to other natural phenomena – this is foundational to natural science. Therefore Searle’s conclusion suggests that understanding experience as subjective has no place in our understanding of it as a natural phenomenon.
In section 5 I defend the pivotal claim that science aims to make nature intelligible by relating phenomena to one another, without necessarily requiring reduction. This fact entails that a scientific understanding of a phenomenon is formed on the basis of its fit with other phenomena, as ultimately evidenced by observation. I argue that the naturalism endorsed by the property dualists, Chalmers and Thomas Nagel in particular, assumes otherwise. They suppose that a priori conceptions of phenomena can take precedence so that science must accommodate them. Consequently their prescriptions for revisions of science to incorporate experience understood as non-physical specifically cannot be met.
Finally in section 6 I respond to the worry that denying the understanding of experience in phenomenological terms plays a role in understanding experience as a natural phenomenon is to dodge the so-called hard problem. Here I emphasise that my criticism of property dualism is not that it is false as such, but that assuming experiential properties are non-physical precludes the possibility of understanding consciousness as natural. This makes the metaphysical naturalism that Chalmers and Nagel assume empty in the sense that it is scientifically otiose.
- The Non-Physical and Causation
All natural phenomena, I have implied, are relatable to other natural phenomena. We would not tolerate any claim to the effect that there is some natural phenomenon that does not depend on any other phenomena. The dependence we assume to exist between phenomena in this sense is causal. In other words, we assume that everything in nature is causally dependent on something else. Nothing stands outside the causal nexus. Whatever causation amounts to, we understand it as the basic way phenomena – objects, events, properties etc. – are interrelated such that the world stands as a unity as opposed to being divided into isolated fragments. It is what Hume famously called the ‘cement of the universe’. The idea of a fragmented world must be seen as self-contradictory. It makes no sense to think of the world as composed of several worlds, or causally isolable realms. Even Descartes insisted that the independently existing mind, as he imagined it, must be causally related to the physical, hence his interactionism. An embarrassment for substance dualism, of course, was its failure to explain the causal interaction between mind and body. All natural phenomena, therefore, stand inside the causal nexus.
The central thesis of property dualism is that experiential properties, i.e., the phenomenological qualities that essentially characterise experience, are non-physical. That is to say, these properties are ontologically irreducible to any physical properties. There are no identity relations between properties understood as experiential and as physical. The clearest statement of this is perhaps given by Saul Kripke (1980) in his modal argument. There he argues that because experiential properties must be understood in terms of appearance their correlations with physical phenomena, e.g., pain with C-fibre stimulation, can only be contingent. Identity, however, is a necessary relation. Therefore, experiential properties cannot be identical with any physical properties.
Now given the causal closure of the physical, the non-physicality of experiential properties implies they have no role in any account of the causal relations between phenomena generally. So, for example, the act of removing my hand from close to a flame – as a physical phenomenon – is causally explained wholly in terms of physiological processes such as C-fibres firing, movement of muscles and so on. I cannot appeal to the phenomenological quality of hurtfulness in this case to explain my action. Experiential properties, therefore, must be causally irrelevant. That is indeed the conclusion of early Frank Jackson (1982) for example.
That said, property dualist Chalmers has argued that the causal irrelevance or epiphenomenalism of experiential properties is not implied by their non-physicality. He admits that a weak form of epiphenomenalism is likely the consequence of property dualism, which is roughly that while experiential properties may play some causal role in the background, appeal to them is at least superfluous with respect to explaining our behaviour. He considers four reasons for resisting epiphenomenalism. The first three he more or less discounts himself, so I will only briefly discuss them.
First, we might interpret phenomenon A causing phenomenon B simply to mean that there exists a regularity between these types of phenomena, roughly understood in Humean terms as merely a form of constant conjunction. Here we could perhaps invoke nomic principles that describe this regularity (1996, p. 153). But, as Chalmers notes, this notion of quasi-causation (my term) implies too much. For example, we could assume a quasi-causal connection in this sense between identical twins and hair colour.
Second, one might imagine routine causal overdetermination with respect to our behaviour. Whenever some physical state described by property, P, causes behaviour, A, it coincides with experiential property, Q, and we might be able to suppose that both P and Q causally explain A each and every time. Chalmers remarks that such overdetermination is regarded with suspicion, though he sees no principled reason for dismissing the suggestion. What is dubious about this suggestion, of course, is that we observe no such routine causal overdetermination in nature. Perhaps if we were to discover routine causal overdetermination in nature at some point we could revisit the idea. It holds very little promise as things stand.
Third, Chalmers thinks we know very little about causation as such, and so, following G.H. Rosenberg (1996), he speculates that there might be an underlying connection between experience and causation. The idea is roughly that what Hume called ‘the unknowable causal relation’ – a hypothetical unobservable force that connects events seen by us to be constantly conjoined at most – concerns experience. In this way experience may play a hidden role in causation generally. I call the imagined causal role of experience hidden to emphasise the lack of any empirical evidence in support of this hypothesis. Further, Chalmers acknowledges, the idea is extremely speculative (Chalmers 1996, pp. 152-53). He notes the unobservability of this connection between experience and causation still leaves open the possibility of a zombie world, that is, a world in which is physically indistinguishable from the actual world but where experience does not exist, where effectively causation is divorced from experience.
The way of denying epiphenomenalism that Chalmers is most attracted to, and presumably finds most plausible, is based on thinking of experiential properties as intrinsic. He argues that it is reasonable to assume the fundamental entities posited by physics, such as electrons, possess intrinsic properties. While our physical theories are based solely on relational properties, we can think of the fundamental entities physics posits in their own right, that is, independent of the relational properties that concern physical theories (ibid, pp. 153-56). He reasons there must be something more to electrons etc. than is described by their relational properties alone. These intrinsic properties relate to the internal nature of any such entities. Any such intrinsic properties are indirectly causally relevant, as causation is a relation between entities with these intrinsic properties. Thus, if we think of experiential properties as intrinsic we can understand them as causally relevant after all.
The problem is again that experiential properties by this measure are mysterious. Chalmers himself describes them as “the hidden intrinsic properties that constitute phenomenology.” (ibid, p. 155) Here, I take a property F of some object, a, to be intrinsic if Fa is true even when a is considered in isolation, i.e., everything else in the world to which it could stand in relations is cast aside. Further any experiential properties must be regarded as in principle unobservable given their non-physicality. Therefore, it is impossible to observe experiential properties so conceived directly, or indirectly in terms of their relations with other phenomena, since they have none. Chalmers acknowledges this consequence of this construal of experiential properties, remarking that “[b]ecause consciousness is not directly observable in experimental contexts, we cannot simply run experiments measuring the experiences that are associated with the various physical processes, thereby confirming or disconfirming various psychophysical hypotheses.” (ibid, p. 215) Indeed, he concludes that the only scientifically respectable method there is left to judge the truth of any theories we might construct concerning experiential properties so understood is inference to the best explanation (IBE). He adds that this “means that a theory of consciousness will have a speculative character not shared by theories in most scientific domains.” (ibid, p. 218) It would be more accurate to state that all theories of the natural sciences demand the possibility of empirical testing, here denied to a theory of consciousness. While IBE is an integral, and perhaps indispensable, tool in the methodology of natural science, it is no good on its own since any such theory is about nature, i.e., the world as we find it, not how we think it most plausibly is.
What we see with all four defences of property dualism against the charge of epiphenomenalism provided by Chalmers is that none of the hypotheses on which they are founded are open to empirical verification or falsification. Perhaps they are each metaphysically viable insofar as they have not be shown to be false a priori. However, they are all scientifically otiose, or useless even, in virtue of their empirical untestability. They dangle nomologically, as again Feigl and Smart would put it. This fact matters when we think of experience as a natural phenomenon quite generally. More particularly with respect to the charge of epiphenomenalism, none of Chalmers’s strategies enable us to observe experience as cause in action. The experiential ingredient is always veiled.
In the case of causal overdetermination, for example, if experiential property Q as well as physical property P is relevant in explaining A, then to test this hypothesis one has to isolate Q from P and see if on its own it must be realised for A to arise. But, the idea of separating Q from P simply makes no sense. That is because Q must be considered a property of a physical thing. Q depends on the physical. Any being that has Q, as a physical entity must also possess P in the causal explanation of its action A, given that this action is physical in nature. Or again, in the case of the suggestion that experiential properties’ causal relevance can be explained by their being intrinsic properties, their causal relevance holds in virtue of their being parasitic on the physical. Here Chalmers conjectures that causation as a relation between physical entities would amount to a relation between things ultimately instantiated in virtue of their intrinsic properties (ibid, p. 154). But there is nothing substantive we could ever say about these properties in their own right. There is no independent evidence for their existence. They remain irredeemably hypothetical entities.
- Whence Do Experiential Properties Arise?
We see, therefore, that there is no convincing way of understanding experiential properties, thought of as non-physical, as causally relevant. While it may be hypothetically possible that property dualism does not entail epiphenomenalism, there can be no independent evidence or testable implications that supports this possibility. However, even assuming epiphenomenalism there still appears to be no difficulty in supposing experiential properties to be natural, where, recall, it is their falling outside the causal nexus that would make it impossible to understand them as natural phenomena. After all, though they are not causally relevant themselves, as epiphenomena they are understood to be physically caused and thus part of the causal nexus. Alternatively it has been argued, e.g., by Chalmers, that they are fundamental properties of nature. Below I show how both these strategies are unconvincing.
Property dualism must anchor experience to the physical given that the view is defined in terms of experience being a feature of material things. But by itself this requirement is very weak. It is compatible with holding that everything has experiential properties, i.e., experience is a fundamental feature in the world much like the fundamental features posited by physics. This fact has encouraged some to entertain seriously various versions of panpsychism with respect to experience (see Chalmers 1996 & Strawson 2006). That said, there is no evidence that experience is in any sense a ubiquitous feature of things. As it stands, we have to assume it results from certain complex arrangements of matter as realised by sentient organisms. Nonetheless, these considerations present a dichotomy. The property dualist can suppose that either (1) experiential properties are physically caused or, despite there being no evidence for it, (2) they are fundamental, in principle at least, and hence ubiquitous – this is to suppose that they are brute, that is, not caused by anything. Let us look at (1) and (2) in turn.
John Searle usefully spells out the sort of relationship between the experiential and the physical that concerns us in (1). He understands experiential properties as causally supervening on physical properties. By this he simply means that when one thing, A, has a set of physical properties, φ, and that is the cause of a particular experiential property, ψ, it follows that if some other thing, B, has φ it must have ψ, all else being equal. But it does not follow conversely that when both A and B have ψ they both have φ. There could be various causes of a thing’s having ψ. This asymmetric dependence relation holds between solidity, for example, and molecular structure; a piston’s molecular structure, say, causally explains its solidity, so we can say that its solidity causally supervenes on this structure. The solidity of an icecube, on the other hand, is causally explained by a different molecular structure. But crucially Searle rejects property dualism on the grounds that experiential properties supervening on physical properties in this sense implies that experiential properties must in fact be causally efficacious, and hence physical by his lights. He writes:
The solidity of the piston is causally supervenient on its molecular structure, but this does not make solidity epiphenomenal; and similarly, the causal supervenience of my present back pain on the micro events in my brain does not make the pain epiphenomenal. (1992, p. 126)
This rejection of epiphenomenalism certainly agrees with commonsense, which suggests that any phenomenon which is an effect must itself be a cause or causally efficacious. But Searle offers no independent reasons for this assumption. An argument for it has been offered by John Lachs, so let us look at that.
First, however, we need to distinguish between two types of epiphenomenalism, what one might call supervenience- and cause-based. Jaegwon Kim charges Davidson’s anomalism monism – taken as supervenient physicalism – of implying epiphenomenalism on the grounds that if a physical state, S1, is both the cause of another physical state, S2, and also determines by supervenience some mental state or property, M, then, he argues, M plays no causal role, i.e., is epiphenomenal, because S1 by itself causally explains S2, i.e., it preempts M. This describes supervenience-based epiphenomenalism. Our concern on the other hand is with cause-based epiphenomenalism, the idea that M is caused by S1 rather than supervening on it, but it is nonetheless causally inert.
Lachs (1963) argues that we think of events, or phenomena more generally, to entail one another causally. In this sense causation is analogous to reason. For example, given any proposition p, there must be some other proposition or conjunction of propositions that logically entail it. Conversely, for any p there must be some proposition or conjunction of propositions that p entails. On the standard model of causation we expect phenomena to entail each other in an analogous manner. That being the case, epiphenomenalism is ruled out since it amounts to the claim that some phenomena do not causally entail any others, against this expectation (see 1963, pp. 142-43). In addition Lachs argues that any phenomenon that we think of as an effect we must also think of as a cause. There is no conceptual distinction between an effect and a cause in this sense (ibid). In other words, causes and effects belong to the same ontological category. This fact again rules out epiphenomenalism, since it assumes cause and effect can belong to distinct ontological categories.
These two arguments appeal to our intuition, how we think causation must be. However, it remains possible to reply that while epiphenomenalism offends our understanding of causation this fact does not rule it out. Perhaps, the epiphenomenalist might argue, it turns out that cause and effect can belong to different ontological categories and that sometimes an effect does not entail a cause, against our intuition.
However, the overall argument above can be strengthened by thinking of causal entailment in terms of the transitivity of causal power, or more exactly, the idea that the causal powers of some body are inherited from the causal powers of its constituent elements at the microstructural level. A physical property as such is one that is causally relevant – it relates to a body’s power to react causally. Solidity, for example, is a physical property defined by a body’s disposition to be stable or rigid, i.e., to resist alteration in shape. This power to resist is explained by the microstructure of its component molecules, e.g., the solidity of a metal is explained by the covalent bonds between its molecules. In this sense a metal inherits its solidity from the causal powers of its constituent elements. We are meant to imagine, on the other hand, that some non-physical property, ψ, is explained by the causal powers of the constituent elements of the body that realises it, but that ψ inherits none of these causal powers, since it has none of its own. How, we must ask ourselves, can a body have such a property? It makes no sense to suppose that a body’s constituent elements can configure themselves to realise a property that inherits none of the causal powers responsible for it. The causal powers that explain the realisation of ψ cannot disappear or evaporate in any sense – they must have some causal consequence. If ψ is a property that results from a particular configuration of matter then it must inherit the causal powers that brought it about. Hence a non-physical property such as ψ is impossible. Therefore, as Searle claims, if experiential properties causally supervene on the physical properties of some object then these properties must be causally relevant, just like solidity.
A reply to this objection might be that while an experiential property is brought about by the brain processes specifically, it plays no role in the brain’s operations, i.e., it is a byproduct of these physical processes. That is all that is meant by an experiential property being causally irrelevant or impotent. This idea, however, turns out to be empty. Imagine two persons, A, who is conscious and B, her behavioural duplicate who is non-conscious. The epiphenomenalist means to assert that A and B are not physically identical – to explain the experiential difference between them – but they are functionally identical. The difficulty with this suggestion is that the difference is purely stipulative. While A is said to differ from B in realising experiential properties, because A is behaviourally identical with B there is no way to determine objectively that there is this difference between them. To paraphrase Feigl, understood as purely subjective events, A’s realising experiential properties and B’s not realising them, would be “in principle not intersubjectively and independently confirmable.” (1958, Ch. 5, §B, p. 61) Both A and B would sincerely report there being something that it is like for them each to see red, for example. There could be no objective method of determining that A’s report is true and B’s false. In reply one might argue that the physical difference between A and B enables us to determine objectively this experiential difference. We could imagine on a separate occasion, say, isolating in someone that neuronal activity associated with painfulness alone, where the pain behaviour is left unaffected by its absence. So once we identify B in particular as lacking this neuraonal activity we can conclude she is non-conscious in this repect at least. Further, the identification process can be expanded to account for all experiential properties. But this will not work. It is not possible to determine when some victim’s realisation of painfulness is removed if, as required, there is no behavioural difference.
Moving on to (2), perhaps we can think of experiential properties as fundamental, that is, almost everything, from the subatomic level up, realises experiential properties in the same way as almost everything is extended or has mass. By this measure these properties can be understood as natural without assuming they are caused. They are simply brute features. Here it is assumed that because the hypothesis is not self-contradictory, i.e., it is not clearly false a priori, it is alive or viable. But given that our aim is to understand experience as natural, this hypothesis is of no use at all. It has no testable implications whatsoever, and that matters.
Imagine positing a subjective particle that, like purported non-physical experiential properties, is taken to be objectively undetectable. Call it the ‘ψ-particle’. It has, let us suppose, the peculiar power of being self-intimating, i.e., presenting itself to itself. When these particles combine in the right way this self-intimation is amplified, and makes the body they comprise subjectively present – present to itself – as opposed to objectively present, i.e., existent as such. Similar suggestions have been bruited by those sympathetic to panpsychism, in terms of experiential properties rather than particles. One might also think of Jean-Paul Sartre’s distinction between en-soi and pour-soi to ground the idea a bit more, where ψ-particles exist for themselves, i.e., pour-soi. Thus we might explain the emergence of a subject by appeal to ψ-particles. This ‘theory’ is internally consistent. Moreover, it is all inside, that is, it says nothing about the rest of nature. The existence of ψ-particles would not depend on any other aspect of nature. The theory is thus logically independent of our scientific theories and our beliefs more generally.
Likely you would be sceptical of the theory. It is after all arbitrarily constructed. Why, however, ought we to consider (2) seriously? It is likewise logically independent of science – whether or not it is true makes no difference to the rest of our scientific theories, none of which either are implied by or imply the existence of such fundamental experiential properties. Of course, the reason some are tempted to take it seriously is that it ostensibly rescues property dualism. It allows for the possibility of maintaining that experiential properties are non-physical but still part of nature. However, if the property dualist hopes to naturalise experience understood as non-physical in this way she must not only show that experiential properties are hypothetically natural. She must also show how they relate to the rest of nature; otherwise the dualist’s manoeuvre is ad hoc – wholly justified in terms of its rescuing the theory. It has no more merit scientifically than the arbitrary ψ-particle theory.
- The Metaphysics of Exceptionalism
There appears to be one more option open to the property dualist, namely to assume experiential properties supervene naturally on the physical. This involves a shift from a conceptual understanding of supervenience to a metaphysical one. In the case of causal supervenience, for example, some experiential property is said to supervene on a set of physical properties because this set causes a body to have the experiential property. Thus supervenience merely describes the asymmetry of this particular dependence relation, i.e., causation. In the case of natural supervenience some experiential property is said to supervene on a set of physical properties as such, not because of any independently conceived dependence relation such as causation. In other words, supervenience is promoted from a way of describing, i.e., as asymmetric, a dependence relation to naming one in its own right.
Interestingly, it is a conceptual understanding of supervenience that spawned its metaphysical cousin as a result of a misinterpretation of it. Donald Davidson was the first to hypothesise psychophysical supervenience in his ‘Mental Events’. There he remarked the anomalism of the mental he advances “is consistent with the view that mental characteristics are in some sense dependent, or supervenient, on physical characteristics.” (1970, p. 214) Essentially his claim is that if two events have the same physical description, then they must be describable exactly the same way in mental terms, but the reverse does not hold. He understands that there are no event types, only token events – they are by nature unrepeatable concrete particulars. Further, the nature of the way we describe events in physical terms contrasts importantly with how we describe them in mental terms. We cannot tolerate the possibility of two events having the same physical description but varying in their causal efficacy. A physical description captures an event as cause and effect. On the other hand, we can tolerate the possibility that two events falling under the same mental description vary causally. The fact that both John and Mary, say, believe that there is a burglar in the house and this belief explains why John called the police, does not entail that Mary called the police, everything else being equal. This is the basis of anomalism of the mental, i.e., the view that mental talk cannot be understood in nomological terms – the mental does not fall under strict laws.
Supervenience in this sense is clearly used to describe the asymmetric conceptual dependence relation between events under their mental and physical descriptions. For Davidson an event is physical insofar as it can be described in physical terms, which he takes to be the case for all events. Likewise every event is mental given that any one can be described in mental terms. Davidson, therefore, in no way postulates any metaphysical dependence relation between the mental and the physical. Nonetheless, he has been interpreted as suggesting that the mental supervenes on the physical in a metaphysical sense. Most notably, Jaegwon Kim embraces this understanding of supervenience, exploring the idea in great detail (see, for example, Kim 1984). Charitably one must suppose that Kim and other readers of Davidson (e.g., Horgan 1993; Marras 2007) see him as having to assume some such metaphysical dependence given that, in line with the general consensus, they view the physical as a metaphysical category and not merely as a way of describing events. Their realist inclinations steer them away from such a seemingly blatant nominalism. There is no evidence, however, that Davidson was thinking of the physical in these metaphysical terms (see Davidson 1993). At any rate, understanding supervenience qua metaphysical relation might at least allow us to think of experiential properties as natural phenomena; though, as I argue below, this strategy does not work.
Here we observe the exceptionalism afforded to experience assumed to be natural. No other natural phenomenon is thought to be metaphysically determined in this sense. Supervenience has not been used in any serious manner to account for phenomena in the natural sciences. Indeed, how could it given that the aim of science is to explain natural phenomena in terms of each other? To assert that some property, P, supervenes on some set of base properties, N, explains nothing about how N determines P. It is dressed up as a scientific claim insofar as supervenience is thought of as underwriting psychophysical laws. Chalmers describes the situation thus: “The position we are left with is that consciousness arises from a physical substrate in virtue of certain contingent laws of nature, which are not themselves implied by physical laws.” (1996, p. 125) It so happens, by this measure, that when a certain physical substrate is realised by a body, e.g., in brain processes, that body realises some experiential property. There is no reason for this in terms of the substrate as such. In a world with distinct psychophysical laws (where the same physical laws hold) things could have turned out differently, e.g., the property could be realised by magnesium reacting with water, or even two sticks rubbing together. Contingency here implies accident, since in general P‘s supervening naturally on N is not the result of laws governing N, i.e., physical laws. So again, it is an ad hoc manoeuvre to invoke natural supervenience to explain how non-physical properties arise. There are no other, i.e., independent, reasons for hypothesising psychophysical supervenience. Indeed, as noted above, it was not originally hypothesised by Davidson to do any metaphysical heavy work in this sense.
Chalmers himself does not worry about the charge of adhocism. He sees any psychophysical supervenient laws as fundamental and so it is natural to think that the ‘existence’ of such laws is brute. Further, he explains that “dependence of experience on the physical cannot be derived from physical laws, so any final theory [of ‘everything’] must include laws of this variety.” (ibid, p. 127) According to Chalmers, therefore, we must think of these psychophysical laws as applying at a fundamental level, parallel to, but independent of, the laws of physics concerning subatomic particles etc. One thing that makes this idea implausible, that has been pointed out by others (e.g., Hardcastle 1997 and Searle 1992), is that experience is a biological phenomenon. Nothing except relatively few complex organisms show any evidence of consciousness. It is very difficult to join Chalmers in supposing that everything we have found to be conscious is coincidentally biological. This would have to be coincidental because there is nothing about the fundamental make up of conscious organism that is any different from such objects as rocks and chairs. Now, Chalmers does not worry about this difficulty either because he speculates that full-fledged experience, as realised by complex organisms like ourselves, happens to require a level of organisation that is only found in such organisms, and even then only occasionally. He invokes what he calls the principle of organisational invariance. This is to hold essentially that experiential properties supervene on the fine-grained functional organisation of a system (1996, pp. 248-49). Of course, this manoeuvre is in the end no less ad hoc than the suggestion that experiential properties supervene on physical properties, and for the same reasons.
But, even if we were to grant that experience merely coincides with the biological, the property dualist is still confronted with the major problem of being unable to relate experience to the rest of nature. Talk of psychophysical laws suggests a causal relationship between experiential and physical properties. Nomological relations are ordinarily assumed to be causal in nature. To assert, for example, that there is a gravitational force of attraction between bodies, b and c, proportional to their respective masses and inversely proportional to the square of the distance between their centres of gravity, is to make a causal claim, namely that the universal force of gravity causes bodies b and c to move toward one another, everything else being equal. But as Chalmers admits experiential properties understood as supervenient bear no causal relations with the physical – their supervening on certain physical properties is never implied by physical laws.
Now Chalmers claims that supervenient psychophysical laws describe how experiential properties are related by brute dependence to the physical. Further, he tells us that “nothing about this view contradicts anything in physical theory; rather, it supplements that theory.” (ibid, p. 128) Any theory of consciousness that appeals to such psychophysical laws is certainly compatible with our scientific theories, but only in the trivial sense that it is logically independent of them. As well, the falsity of these laws would make no difference to our scientific theories. However, for a theory based on psychophysical laws to supplement the natural sciences, as Chalmers claims it would, it would have to add to it, extend it. But such a theory would stand isolated from the sciences. It neither would support the sciences nor would the sciences support it. Accordingly, there is no reason to think such a theory would allow us to understand consciousness as a natural phenomenon; for to manage that it must permit us to relate experience to the rest of nature, and that is exactly what is denied it in virtue of its logical independence from the natural sciences.
The property dualist may embrace the exceptionalism of consciousness that these facts imply. This exceptionalism is spelled out clearly, again by Searle – though he explicitly rejects dualism. He tells us nonetheless that
it is a general feature of such reductions that the phenomenon is defined in terms of the “reality” and not in terms of the “appearance.” But we can’t make that sort of appearance-reality distinction for consciousness because consciousness consists in the appearances themselves. (1992, pp. 121-22)
No understanding of experience in natural scientific or physical terms is possible, therefore, because experience is appearance itself, that is, a subjective phenomenon. While we can reduce heat, understood in terms of how it feels to us, i.e., how it appears, to molecular kinetic energy, it makes no sense to reduce appearances themselves in this way. Here Searle shares Kripke’s assumption that experience can only be understood as appearance. In other words, it is never possible to explain experience in terms of other phenomena.
To the extent that the dualist considers experience as natural the in principle impossibility of understanding it in naturalistic terms is intolerable. A phenomenon is understood to be natural in the sense that it is relatable to other phenomena. That is not to think of a natural phenomenon as that which in fact bears relations to other phenomena; rather it is to think of a phenomenon’s being natural in a normative sense, namely that we aim to understand it qua natural by relating it to other phenomena. By contrast, for example, God is understood as supernatural, thus we do not aim to understand ‘it’ in terms of its relation to other phenomena. The theistic God at least is thought of as standing outside nature.
That said, the notion of relatability alludes to the unity of science, to the idea that all sciences are fundamentally interconnected so that a phenomenon under one science, e.g., the biological phenomenon of a bacterium, can be understood in terms of another. Hence the dream of full-blooded intertheoretic reduction, where all phenomena are imagined to be ultimately reducible to the entities posited by physics. The notion of relatability I appeal to here is not this strong notion of reducibility, where a phenomenon, Φ, is said to be reducible to some set, χ, of more basic phenomena such that Φ is nothing but χ-phenomena. I am not assuming a phenomenon is natural on condition that it can in principle be reduced to the fundamental entities of physics in this way. Instead, in saying all natural phenomena are relatable I mean that no natural science concerns phenomena that cannot be causally related to phenomena studied (that fall under) another science. For example, Darwin’s and Wallace’s theory of natural selection makes speciation a natural process partly because we can explain the random mutations it appeals to in terms of the chemistry of DNA. On the other hand, traditional creationism fails to offer a satisfactory naturalistic understanding of speciation since it presents no proper way of relating speciation to other phenomena.
Searle, however, thinks of experience as causally relevant but not understandable in naturalistic terms. This fact contradicts my equating the causal relevance of a phenomenon with understanding it in terms of other phenomena. Indeed, Searle holds that while experience is causally relevant – recall his denial of its being epiphenomenal – it is ontologically distinct and hence irreducible to the physical. He explains that experiential states and processes “have a special feature not possessed by other phenomena, namely, subjectivity.” (1992, p. 93, my emphasis) He maintains that we can identify the physiological causes of an experience, e.g., pain, but insists that we cannot identify the causally relevant physiological processes as pain. In this sense pain is, according to Searle, a causally emergent phenomenon (see ibid, pp. 115-16). And again, he reasons that pain and all other experiential states are ontologically irreducible in this sense because of their subjective nature, because of their first-personal existence. How do these essentially subjective phenomena arise from physical processes? Searle thinks no answer is available to us; it is a mystery in other words. This sort of mysterianism about experiential properties, I am arguing, is intolerable. Searle effectively holds both that pain is caused by physical processes and that it is in principle impossible to explain how. These two claims are contrary. To say ‘A causes B’ is to assert that ‘B occurs because of A’, that is, ‘A explains B’, at least in principle. By Searle’s light experience, understood as an essentially subjective phenomenon, is miraculous. There is no reason why experience is subjective, it just is. Indeed, Searle’s position shows that the subjective aspect of experience can play no role in understanding it as natural.
- Under the Aegis of Naturalism
Almost every philosopher subscribes to naturalism in the broadest sense, i.e., the view that there are no extranatural forces at work in the world. Chalmers, for example, takes his dualism to be naturalistic precisely because “it posits that everything is a consequence of a network of basic properties and laws,” (1996, p. 128). We have seen that the basic properties and laws he posits concerning experience are addenda in the sense that the natural sciences do not need to include them at all for their theories to remain viable. The problem with this strategy, again, is that these addenda bear no relations to other natural phenomena. Another dualist, Thomas Nagel, recognises this difficulty, writing that “adding extra peculiar ingredient like qualia,…to the otherwise magnificently unified mathematical order of the physical universe…does not answer to the desire for a general understanding of how things fit together.” (2012, p. 15) So far so good, but Nagel goes on to assert that the natural order he alludes to here is inherently intelligible; that is to say, the scientist does not make nature intelligible but rather she uncovers its underlying intelligibility (ibid, p. 17).
Implicit in Nagel’s view is the quite reasonable assumption that there is a gap between how the world really is and how science describes it. In addition one may also assume there are no extrascientific facts about nature that we can grasp and thereby determine how our scientific theories fall short in this sense. In other words, there are no truths about the natural world we can grasp independently of science, that we can hold without need to test them against experience, i.e., by application of the scientific method. However, Nagel’s claim that nature is inherently intelligible tempts him to think there are such extrascientific facts. If nature is itself ‘rationally structured’ then our reason alone might enable us to grasp some truths about nature. One such truth, according to Nagel, is that experiential properties do not fall under physical laws – they are not physical. The view that the natural order is rational in this strong sense I will call metaphysical naturalism. It stands in opposition to scientific naturalism, namely the view that how we reason the world to be cannot by itself count as conclusive evidence, rather how the world is in any respect must ultimately be shown to agree with observation. I call this naturalism scientific to indicate that it reflects scientific practice – to do science is to aim to explain and predict phenomena in accordance with observation. Below I argue that the metaphysical naturalism assumed by property dualists like Chalmers and Nagel leads them to prescribe revisions to scientific practice that cannot be accommodated.
Nagel, for example, argues that it is a mistake to think that Darwinism and our physical theories conjointly offer a satisfactory account of mentality in general. He states that “more is needed to explain how there can be conscious, thinking creatures whose bodies and brains are composed of those [fundamental chemical and physical] elements.” (ibid, p. 20) Why? Because according to Nagel we can determine a priori that the subjectivity of experiential properties cannot be subsumed under our physical theories (Nagel 1986, pp. 25-27). Chalmers essentially reaches the same conclusion about the non-physicality of experience by a priori means as well. Both claim, therefore, that science falls short of providing a totally inclusive theory of reality, i.e., a so-called ‘theory of everything’. Crucially, however, neither sees the non-physicality of experience as an extrascientific fact in principle, that is, a fact that it is impossible to incorporate into the sciences. They assume science can in principle subsume experience so understood under its theories by expanding them in suitable fashion, hence their naturalism. Again, for Chalmers this involves positing additional psychophysical laws. For Nagel it instead involves reconstruing the laws of nature in teleological terms (Nagel 2012, pp. 122-26).
To illustrate the distinction between metaphysical and scientific naturalism consider a historical example concerning Descartes’s objection to the demonstration of a vacuum. Early experiments aimed at creating a vacuum were performed in his day first by Evangelista Torricelli (a student of Galileo) and soon after by Blaise Pascal. The experiments very basically involved observing that the weight of a mercury column in a glass tube, sealed at its top, left an empty space there after the mercury had dropped. On hearing of the results from Pascal Descartes was sceptical. He had previously reasoned that a vacuum is impossible because space or extension is in essence matter. He thought it is self-evident that matter is identical with space, hence where there is no matter there can be no space. Since a space was observed at the top of the tube it must have been filled with matter (see Garber 1992, pp. 136-43).
Descartes’s scepticism, then, was motivated by his already having a conception of space based on his a priori reasoning. In this sense Descartes exhibited an attitude coinciding with metaphysical naturalism. He was confident that his reasoning allowed him to determine the real nature of space – an extrascientific fact – and this took precedence over contradictory empirical evidence. Indeed he likely thought that the experiments’ result was best explained by the entry of a fine ether through the relatively coarse pores of the glass of the tube that fills the space. His objection to the experimenters’ conclusion that likely a vacuum existed at the top of the tube showed he believed quite generally that reason alone can provide an experimentally operative conception of some aspect of nature. Observations ought to accommodate his a priori conception of space. By contrast, Pascal demonstrated an inclination to form his conception of space according to his observations. If the presence of a vacuum best explains the experiments’ result then one forms an understanding of space that accords with it. This attitude can be generalised as operating by the following precept that coincides with scientific naturalism: Think of a phenomenon according to its best observable fit with other phenomena. So, famously, if the two-slit experiment demands that we think of particles as waves, so be it.
Chalmers and Nagel think that the natural sciences fail to account for experience as they understand it really to be. Thus, in accordance with their metaphysical naturalism, they see the need for revisions to the sciences in order to incorporate experience as they understand it. From a scientific naturalist’s perspective, however, our understanding of experience ought to follow its fit with other phenomena, that is, how we think of experience is best formed by relating it to the rest of nature. Indeed, this approach, I have been arguing, is the basis for understanding experience as a natural phenomenon. It is recognisable as the approach taken by Daniel Dennett, for example. Dennett argues that the only viable way of explaining consciousness is from the third-person perspective. No progress can be made in understanding experience if we insist it is essentially subjective and therefore invulnerable to standard scientific scrutiny (Dennett 2005).
There is, then, an underlying tension in property dualism. The dualist offers a priori arguments against the possibility of experiential properties being physical, as falling under the laws of physics. At the same time she adheres to naturalism and insists that experience is in some sense an integral part of nature, though it is special in virtue of its non-physicality. This exceptionalism leads Chalmers and Nagel to prescribe revisions of distinctive sorts to science in order to incorporate experience. But their prescriptions are misplaced because they assume science can accommodate their a priori understanding of experience. We have seen, however, that a priori conceptions of natural phenomena have no precedence over our observations of nature. Science operates by adopting those conceptions of phenomena that agree with our observations, since the fundamental aim of science is to make nature intelligible by relating natural phenomena to each other.
- Metaphysics versus Science
The property dualist’s reply to the claim that we cannot understand experience as natural when its characteristic properties are thought of as non-physical might well be that that is too bad for science. If, following Nagel, we take scientific naturalism – what he calls ‘materialist naturalism’ – to be the view that everything including experience can be incorporated into the physical sciences, then this view is simply wrong. Nagel writes: “Perhaps the natural order is not exclusively physical; or perhaps, in the worst case, there is no comprehensive natural order in which everything hangs together – only disconnected forms of understanding.” (2012, pp. 15-16) My argument seems only to confirm that one or both of these disjuncts are likely true. However, this response misses the point I am making.
My claim is that in order to understand experience as a natural phenomenon we must relate it to other phenomena. To do science is to build connections between forms of understanding. This is not something that scientists would like to achieve as Nagel assumes above, that is, take our understandings of phenomena and see if we can find connections between them. That is to get things backwards. Scientific concepts originate with making such connections. So, while experience has a subjective dimension to it – no one seriously argues that we are in the dark, i.e., non-conscious in the same way as a rock is – thinking of experience in this way does not help us understand it in scientific terms, that is, in terms of its relations with other phenomena. Talk of experiential properties generally in a metaphysical context makes sense. We have an intuitively clear understanding of experiences as phenomenological. But such talk is scientifically otiose, again given that it cannot allow us to connect experience to the rest of nature. Another historical example makes this point clearer.
For many centuries mathematicians operated with an intuitive concept of continuity, i.e, in spatiotemporal terms. However, by it we cannot fully make sense of the notion of the infinitesimal. If, like Zeno, we think of an arrow’s motion as continuous in this spatial sense, then we understand that no matter how many times we divide up the period of its flight it is always possible to divide it more. If we were to suppose, on the other hand, that such divisions really reach a limit, it is hard to see how each infinitesimal period, i.e., each instant, can have any magnitude, since if it did we could imagine its being divisible still. A problem arises for example with demonstrations of calculus that appeal to the infinitesimal and so face a paradox, famously spelled out by George Berkeley. Determining the derivative of a function – each of which has direct applications to dynamical models of natural phenomena – involves both assuming an infinitesimal point as having magnitude and as having none (see Berkeley 1992, p. 175).
These paradoxes are dissolved by abandoning thinking of continuity in spatiotemporal terms. Mathematicians (e.g., Karl Weierstrass) began to construe continuity in formal terms, that is, strictly as a relation between numbers. One might say that they define continuity operationally. Thus, the infinitesimal is informally thought of as an arbitrarily small quantity, that is, as small as one likes. In this way a function, f(x), represented by a curve is said to be continuous at x = a if for any value of f(x) approaching this value as a limit, i.e., f(a), there is some positive number ε, i.e., ε > 0, such that │ f(x) – f(a)│< ε , then there is always some number δ, where similarly δ > 0, such that │x – a│< δ. This definition of continuity in terms of limit does not invoke spatiotemporal concepts, it is purely formal. Calculus conceived in this thoroughly formalistic manner avoids the difficulties pointed out by Berkeley; expunged of any empirical content in this way the infinitesimal, specifically, is no longer thought of as a magnitude.
There is a useful parallel to be drawn with Wilfrid Sellars’s distinction between the manifest and scientific image (Sellars 1997, pp. 79-85). He points to the evolution of a commonsense picture of the world based on everyday talk of physical objects and their characteristics, as well as on our folk psychology. This manifest image of the world is in the process of being replaced in part by a scientific image or picture – to provide a more complete understanding of phenomena – where, in Sellars own words, “the scientific account of ‘what there is’ supersedes the descriptive ontology of everyday life.” (ibid, p. 82). But Sellars does not see the replacement of the one picture by the other as a denial of the existence of the everyday entities scientific talk might replace – this process has no such ontological consequences. Within the framework of our commonsense talk it would be absurd to deny, for example, that the chair before me exists, that all that really exists is a complex of molecules. Similarly, the mathematicians’ replacing of the spatiotemporal concept of continuity with a formal concept does not entail the abandonment of our spatiotemporal concept that belongs to a manifest picture of the world, despite the paradoxes it generates. And again, crucially, giving up talk of experience as an irreducibly subjective phenomenon within a scientific framework is not a call for the forthright denial of subjectivity.
If one seeks to gain an understanding of experience as a natural phenomenon, nothing is served by thinking of it as characterised by non-physical properties. On the face of it this suggestion sounds like a dodge, an attempt to sidestep the difficulty of explaining the phenomenology of experience as such – what is called the ‘hard probem’. However, the claim is not that experience is a purely objective phenomenon and so it is futile to try to understand it qua subjective. I have argued, rather, that to think of experience in phenomenological terms does not help us to understand it as a natural phenomenon. Science, quite generally, does not aim to explain phenomena as they fit with our intuitive understanding of them. It aims instead to provide ways of conceiving of phenomena, of understanding them, that permits us to relate them to one another. Our intuitive understanding of some phenomenon is a barrier to this aim, albeit perhaps sometimes a necessary starting point, since it is based on how we think the world is rather than how it is according to the best evidence of our observations. Over all, the complaint is that property dualists, such as Chalmers and Nagel, illegitimately treat science as the handmaiden of metaphysics. Their attempts to harness science for their metaphysical ends cannot succeed.
© Hugh Alcock
Berkeley, G. (1774/1992), The Analyst. De Motu and The Analyst, Jesseph, D.M., ed. & trans., Dordrecht, Netherlands: Kluwer Academic Pub.
Chalmers, D.J. (1996). The Conscious Mind, Oxford: Oxford Univ. Press.
Davidson, D. (1980). Mental Events. Essays on Actions and Events, Oxford: Clarendon Press, pp. 207-227.
Davidson, D. (1993). Thinking Causes. Mental Causation, Heil and Mele, eds. Oxford: Clarendon Press.
Dennett, D.C. (2005). A Third-person Approach to Consciousness. Sweet Dreams: Philosophical Obstacles to a Science of Consciousness, Cambridge, MA: MIT Press, pp. 25-56.
Feigl, H. (1958). The “Mental” and the “Physical.” Minnesota Studies in the Philosophy of Science, II.
Garber, D. (1992). Descartes’ Metaphysical Physics, Chicago: Univ. of Chicago Press.
Hardcastle, V. (1996). The Why of Consciousness: A Non-Issue for Materialists. Journal of Consciousness Studies 3, No. 1: 7-13.
Horgan, T. (1993). From Supervenience to Superdupervenience: Meeting the Demands of a Material World. Mind 102: 555-86.
Jackson, F. (1982). Epiphenomenal Qualia. Philosophical Quarterly 32: 127-136.
Kim, J. (1984). Concepts of Supervenience. Philosophy and Phenomenological Research 45: 153-76.
Kripke, S. (1980). Naming and Necessity, Oxford: Blackwell Pub.
Lachs, J. (1963). Epiphenomenalism and the Notion of Cause. The Journal of Philosophy, vol. 60, No. 6: 141-46.
Marras, A. (2007). Kim’s Supervenience Argument and Nonreductive Physicalism. Erkenntnis, vol. 66, No. 3: 305-327
Nagel, T. (2012). Mind and Cosmos:Why the Materialist Neo-Darwinian Conception of Nature is Almost Certainly False, Oxford: Oxford Univ. Press.
Rosenberg, G.H. (1996). Consciousness and Causation: Clues Toward a Double-Aspect Theory. Manuscript, Indiana Univ.
Sartre, J-P. (1956). Being and Nothingness, Barnes, H.E., trans, NY: Washington Square Press
Searle, J. (1992). The Rediscovery of Mind, Cambridge, Mass: MIT Press.
Sellars, W. (1997). Empiricism and the Philosophy of Mind, Cambridge, MA: Harvard Univ. Press.
Smart, J.J.C. (1959). Sensations and Brain Processes. The Philosophical Review, 68(2), pp. 141-156.
Strawson, G. (2006). Realistic Monism: Why Physicalism Entails Panpsychism. Journal of Consciousness Studies, 13, No. 10-11, pp. 3-31.