Why the problem of consciousness won’t go away

Guest post by Michael Lopresto

[Michael is a PhD student at The University of Adelaide (Australia). His thesis explores evolutionary psychology from a much neglected empiricist perspective, arguing against the dominant nativist paradigm. He argues against the existence of specialised modules for things like language acquisition and moral sense, and tries to ground cognitive function in connectionist computational models (also a much neglected approach, nowadays). Jesse Prinz and Kim Sterelny are major influences.]


One of the main tasks of philosophers is to solve what are sometimes called “location problems.” We start with a conception of reality that we’re happy with, namely the description of the world that comes out of physics. This conception is both strongly supported by empirical evidence, and conceptually very clear. The problem is that it’s hard to square many aspects of our everyday experience with it. How is it — for instance — that morality, or meaning, or mathematics can exist in a purely physical world? In come the philosophers (hopefully to make matters better rather than worse) to try to “locate” these things into whatever conception of the world we’re happy with. To solve a location problem is to understand how it is that the thing we’re trying to locate — be it moral properties or anything in our manifest experience — fits conceptually into our picture of the physical world.

My project here is to ask whether it’s possible to locate consciousness in the physical world. That is, can we locate phenomenal properties in the physical world? My thesis is that given our conception of the physical world, it is in fact extremely difficult to locate phenomenal properties within it. 

So, how are we to understand phenomenal properties? Phenomenal properties (or simply experiences) are defined by what it’s like to have those properties (an expression made famous by Thomas Nagel is his seminal 1974 paper “What is it like to be a bat?”). There is something it’s like to experience the redness of red, to have a visual experience of a yellow lemon, to feel pain, and to hear the music of Beethoven. This is the phenomenon — the felt quality of experience — that I’m trying to distil. Note that this is different from other mental phenomena such as representation (content or aboutness). Representational properties are defined by  their ability to be true or false. Representation allows an organism to get around in the world, to behave intelligently and promote its survival. Now, giving a naturalistic account of representation — that is, an account that is both physical and makes no reference to antecedent representations — is an interesting and extremely challenging task (try talking about a physical thing that represents some object, without talking about a prior representation). The most popular accounts that philosophers have given have been in terms of causal-covariation (brain states get to be about certain objects in virtue of causally-covarying with them in the right way, over the organism’s lifetime, or its evolutionary history); or resemblance (brain states get to be about certain objects by resembling them). 

Still, neither theory makes any mention whatsoever of phenomenal properties: it is perfectly coherent (and plausible) to explain mental representation without any mention of phenomenal properties. This is important to note because mental representation is what explains behaviour — they are objects in the brain that have causal powers and bring about the property of aboutness that is ever so important if an organism is to avoid predators and find nutrients and mates. 

Our visual experience of a lemon has both representational and phenomenal properties: to explain our ability to represent lemons we need not posit anything phenomenal, as causal-covariation or resemblance will suffice, providing us with all the causal powers we need to respond to lemons in the right way. Our phenomenal experience of lemons and other such things is essentially defined by what it’s like to have those experiences, and those experiences can’t be defined in terms of any causal role the experience may happen to accompany. Representation is defined by function and structure, and consciousness is defined phenomenally.    

Now that we’re clear about our conception of the phenomenal (and how it’s distinct from the representational), what about the physical? Physics gives us a conception of reality as purely functional and structural in character; that is, defined by its dispositional properties. We have empirical knowledge of the external world because the world impinges causally upon our senses and our instruments of measurement. We have empirical knowledge via perception, and our percepts are caused by the dispositional properties of the objects we’re perceiving. For example, when we perceive the yellow lemon, we perceive its yellow colour, it’s oval shape and so forth. Such considerations give us the content of physicalism: the metaphysical thesis that everything is physical, where “physical” is defined by what physics tells us about, namely, dispositions. Mass is defined by resistance to acceleration, charge is defined by how a particle behaves in an electric field, and so forth. 

Now that we have a clear conception of consciousness, as defined by phenomenal properties or what-it’s-likeness, as well as of the physical world, as defined by function and structure, is it possible to locate consciousness in the physical world? It seems to me that it’s not. Function and structure only gives rise to more function and structure, and what it’s like to have an experience can only be understood in terms of what it’s like to have an experience. Some philosophers certainly have tried to understand experience in nonexperiential terms, such as those of function and structure. Daniel Dennett has proposed that we should understand experience as verbal reports of experiential states (which certainly don’t refer to experiences as I’ve conceived them). Alva Noë and Kevin O’Regan have suggested that experience is a sensorimotor function. Michael Tye and Fred Dretske have argued that experience is a representational relation between mind and external world. And there are many other such attempts to “functionalize” experience. Antti Revonsuo has quite ingeniously given a counterexample to many of these views: dreaming. When we dream, we have rich experiences independently of any sort of embodiment or behaviour, we undergo a sort of paralysis, and there are no verbal reports, no sensorimotor function, and no causal-representational-connectedness with the external world. 

As it happens, dreaming is also an excellent counterexample to some misguided people who claim that the mind is not representational. Our dreams are about bungee jumping and nights of passion and losing teeth — at least some of my dreams are. Revonsuo sees consciousness as a biological phenomenon, and from such a perspective attributes certain functional properties to experience. Dreaming, for example, has an adaptive advantage in Revonsuo’s view, because dreaming functions as threat simulation. Dreams tell us possibilities about what can go wrong, and what to look out for. I find it quite plausible to say that dreaming evolved in organisms to simulate threats. Where I would perhaps disagree with Revonsuo is that threat simulation is a property of experience as such. Couldn’t dreams be representational but unconscious? It seems to me that experience adds no substantial content to the representational aspect of a dream. 

The point is, Revonsuo has provided a concrete counterexample to some theories of consciousness (possibly all theories that attempt to understand experience in non-experiential terms). Are there abstract counterexamples that would apply to all such theories that try to account for consciousness in terms of function and structure? I think there are, in the form of two very well known thought experiments: the zombie argument and the knowledge argument. 

The zombie argument asks us to conceive of a physical duplicate of ourselves, thus preserving all the same dispositional properties, but who is lacking consciousness. Our zombie’s representational faculties are identical, so it will be behaviourally identical to ourselves, with all the same abilities to discuss the existence of consciousness, make the same verbal reports, discriminate between stimuli in the same way, and so forth. Is such a scenario coherent? If it is, then physicalism is false, as physicalism says that consciousness is to be located in the dispositional properties of the world. 

So, are zombies conceivable? If we distinguish between two sorts of conceivability, positive and negative, we are in a better place to see. Positive conceivability requires knowing what it takes for something to be true: “2 + 2 = 4” may be a good example, or if you’re really clever, Fermat’s Last Theorem. Some examples of things that are not positively conceivable are inconsistent objects, such as the impossible triangle. When I try to positively conceive of the impossible triangle, it’s like looking at an object in the dark with a flashlight: you can see any two consistent sides of the triangle, but the three sides together, creating the inconsistency, are obscured from view. Negative conceivability requires being able to detect a contradiction in some hypothesis, say, “2 + 2 =5”. Negative conceivability is perhaps most relevant here, as it seeks to find incoherence. Is the conjunction of physicalism and zombiehood incoherent? Physicalism says that if certain dispositional properties are the case, then certain phenomenal properties will be the case. When we conceive of their being no phenomenal properties, but the very same dispositional properties, we will never detect incoherence because phenomenal properties are defined by what it’s like, and dispositional properties are defined by their causal and spatial relations. So it would seem that such considerations tell us that we can’t locate consciousness in the physical world. 

The knowledge argument is about Mary the super neuroscientist, who grasps the complete set of physical facts regarding how the brain works, and particularly about colour vision and the physics of light. However, Mary was raised in a black and white environment. Upon being released, it seems that Mary will learn something completely new, say, what it’s like to experience red. What it’s like to see red is a fact that Mary doesn’t know before her release. However, physicalism says that complete physical knowledge is complete knowledge simpliciter, and if physicalism is true, then Mary ought to learn nothing at all upon release — which few physicalists have had the courage to say. 

The knowledge argument has provoked many responses (many of which probably make matters worse), but I’ll only deal with a few here. It has been said that the knowledge argument starts from epistemological premises and finishes at a metaphysical conclusion: premises about what Mary knows, and a conclusion about what’s not physical. Some people have argued that this inference is invalid. I think this response is mistaken. If you know that the metal bar in front of you is one meter long, then it follows that the metal bar is one meter long. Another response has been that Mary doesn’t learn any new facts upon release, she only acquires new abilities. Mary may have complete physical knowledge of the brain before her release, but she may not know how to swim or ride a bike. This deficit in her knowledge doesn’t lead us to any metaphysical conclusions, so why should any other case? The reason is that physical knowledge is factual — it has entailments about what is true about the world — whereas abilities are know-how and have no such entailment relations. David Lewis for example, argued that upon seeing red, Mary could then imagine red things and identify red mail-boxes and the like. But I think this analysis has things exactly the wrong way around. Knowing what it’s like to see red explains your ability to imagine and identify red things, it’s not that your ability to imagine and identify explains your knowing what it’s like to see red. 

Consideration of the zombie argument and the knowledge argument gives us a strong reason to suppose that we’ll never be successful in locating consciousness in the physical world, as we’ve conceived it. Given this predicament, we can either reconceive what it is to be a phenomenal property, or we can reconceive the physical world. Physicalists have typically tried to reconceive phenomenal properties, but in many cases it seems to me that they’ve made matters worse than they were to begin with — saying that consciousness is outside the head, for example. But another option is to reconceive the physical world. It’s possible that perception tells us only about the dispositional nature of the physical world, and that introspection tells us about the intrinsic nature of the physical world. This gives us some insight into why it is that we can’t locate phenomenal (non-dispositional) properties in physical (dispositional) properties, and why it is that the zombie and Mary thought experiments are coherent.

[Editor's note: for a different perspective on Mary and the zombies, see this recent RS post by Massimo]