Daniel Stoljar’s new book Ignorance and Imagination makes a case for slug-like ignorance as the solution to our problems with consciousness. The real reason we have trouble with reconciling conscious experience with brute physical reality, he says, is that we just don’t know enough. It’s not that we need to crack some subtle piece of metaphysics; nor is it true that our minds aren’t up to the job, as Colin McGinn suggests; nor yet is it a matter of finding some new and exotic additions to physics: there are just some key facts we don’t have. The problem is epistemic. This seems a little like the position of John Searle, who says that consciousness arises from some properties of the brain we don’t understand yet, but which are likely to be perfectly accessible to further scientific research.
To dramatise his point, Stoljar tells the story of slug-like creatures living on a vast mosaic pavement. There are two kinds of mosaic pieces: triangles and segments of circles. The slugs have sensory apparatus which can detect triangles or complete circles, but not segments of circles. Faced with this set-up, the slug theorists divide into several schools of thought. Some say that although the pavement contains squares, hexagons, and many other shapes, there are two fundamental elements: triangles, and circles. All the other shapes can be analysed down into triangles: the circles are an entirely separate affair and just have to be accepted as a feature of the pavement in their own right. Others say that ultimately circles can be reduced to triangles just like all the other shapes, so there is only one basic kind of thing in the pavement. A more radical group says that indeed triangles are all there are, and moreover the impression that there are circles is a mere illusion. Stoljar exploits the analogy cleverly, introducing slug equivalents for several of the most celebrated arguments about consciousness. The point is that the slugs can never get it right, because they simply don’t know that all the circles are made up of segments, nor that other shapes made out of segments appear in various places.
The analogy is amusing and enlightening, but I’m not sure it actually serves Stoljar’s purpose perfectly: it seems to me that the dualist slugs are actually right. Their knowledge of one of the two basic elements of the pavement is faulty, but they are correct in thinking that there are two basic elements. But Stoljar’s main argument does not suffer from the same weakness. He defines the basic issue as being what he calls the logical problem. There are three plausible propositions involved.
· There are experiential truths (ie true statements about experiences)
· If there are any experiential truths, every experiential truth is entailed by some nonexperiential truth.
· If there are experiential truths, not every experiential truth is entailed by some nonexperiential truth
You can choose to believe any two of these, but whichever you choose, they will contradict the other one. I don’t know whether this really deserves to be called the logical problem, but it is a neat formulation of a central problem, and it allows Stoljar to embark on a well-organised survey of the considerations which make each of the propositions attractive. Few would seriously want to deny the first. If you want to believe in physicalism, the doctrine that everything in some sense comes down to physics, you have to adopt the second proposition. Arguments for the third are a little more complex. Stoljar presents two: the conceivability argument and the knowledge argument. The first is about the conceivability of having the physics without the experience: arguments from the possibility of zombies, or of inverted spectra. The knowledge argument is exemplified by the story of Mary the colour scientist.
Stoljar’s treatment is careful and comprehensive, and one way or another he touches on most of the main arguments connected with consciousness in making his case. In terms of meaty pieces of discussion, the book is certainly value for money: but not overweight – Stoljar’s prose is mercifully clear and straightforward, and there were only a couple of moments when I found myself wishing for a quick cut to the chase.
In fact, there was at least one place where I wanted more detail. One of the weaker areas, in my view, is Stoljar’s consideration of conceivability. It is altogether implausible to me that conceivability straightforwardly implies possibility: the former concerns what can happen in my head while the latter concerns what can happen in the world out there. I think these arguments are only plausible if a restricted sense of possibility is specified. The fact that I can conceive of breaking the bank at Monte Carlo does not by any means imply that that feat is possible for me: but my ability imagine it might just imply that it is logically possible. I thought this aspect was not sufficiently acknowledged. Moreover, where Stoljar quotes examples of others using arguments from conceivability, I wasn’t altogether convinced.
This arises with the historical precedents he quotes. As one means of making his thesis plausible, he offers examples of cases where people have drawn false philosophical conclusions because they simply lacked some key facts. Descartes said (not in English, but still) that it was “not conceivable that a machine should give an appropriately meaningful answer” to questions. Stoljar reads the argument as proceeding from Descartes’ inability to conceive of such a thing to its impossibility: but Descartes, he says, just didn’t know that certain kinds of linguistic competence are perfectly realisable from mere physics (after all, we set out to be physicalists, and that implies that we ourselves are mere machines at the end of the day). I’m not entirely sure that amounts to mere factual ignorance, but more fundamentally I’m not convinced Descartes actually meant to offer that argument. I suspect he was merely saying that it was just obvious that a machine couldn’t answer intelligently.
Of course, what we really want to know is: what are these facts we need to know in order to understand consciousness? Stoljar’s chosen mission here is to give a pared-down, generalised version of the epistemic view, so he does not offer a definite answer on this. He claims that to do so would be paradoxical, since you can’t tell people unknown truths (yes, but come off it, Stoljar, you could tell us the truths which were unknown up until just now!). By way of further enhancing the attractions of his thesis, he does expound one possible kind of answer, which he calls Russellian. This is that the vital things which we don’t know concern the categorical properties of objects – the inner, thing-in-itself properties which physics does not bear on. Stoljar notes that this line of reasoning might lead towards panpsychism, if we suspected that the categorical properties were in fact experiential. Although this has an undoubted neatness, panpsychists will be disappointed to hear that Stoljar takes this panpsychist leaning as one of the counts against the Russellian speculation. Personally, I prefer not to believe the Russellian line, in spite of its appeal, because it seems unlikely to me that we could ever gain any useful knowledge of the supposed categorical properties.
In the end, I think Stoljar achieves victory through the modesty of his ambitions: he doesn’t even aim to prove that his view is true, he merely wants us to accept it as a plausible possibility. I do: but at the same time the nature of the chasm between experience and physics still leads me to think something more radical is really required.