Wednesday, 9 March 2022

seth's vision

on it 'looking as if' the sun goes round the earth

Anscombe and Wittgenstein
      by Dave McKean
A gloriously pithy little dialogue between Wittgenstein and Anscombe goes like so: 

Wittgenstein: ‘Why do people say that it was natural to think that the sun went round the Earth rather than that the Earth turned on its axis?’

Anscombe: ‘I suppose, because it looked as if the sun went round the Earth.’

Wittgenstein: ‘Well, what would it have looked like if it had looked as if the Earth turned on its axis?’

On this, in chapter 4 of his 'Being You', Anil Seth offers a gloriously muddled take:

In this delightful exchange between Wittgenstein and his fellow philosopher (and biographer) Elizabeth Anscombe, the legendary German thinker uses the Copernican revolution to illustrate the point that how things seem is not necessarily how they are. Although it seems as though the sun goes around the Earth, it is of course the Earth rotating around its own axis that gives us night and day, and it is the sun, not the Earth, that sits at the centre of the solar system. Nothing new here, you might think, and you’d be right. But Wittgenstein was driving at something deeper. His real message for Anscombe was that even with a greater understanding of how things actually are, at some level things still appear the same way they always did. The sun rises in the east and sets in the west, same as always.

What's gone wrong here? Leave aside the fibs about Wittgenstein being German and about Anscombe being his biographer. Leave aside too the unhappy idea that 'really', or 'of course', it's the earth spinning on its axis, and not the sun orbiting the earth, that accounts for there being night and day. (Doesn't it just depend - as Einstein points out in his little introduction to relativity theory - on which you stipulate as your reference frame?!) Also put aside the notion that it's any more than a truism that the sun, rather than the earth, sits at the centre of the solar system. Focus instead on the fact that Wittgenstein's lesson for Anscombe is neither that 'how things seem is not necessarily how they are' nor that 'even with a greater understanding of how things actually are, at some level things still appear the same way they always did'. His point is instead that it in truth no more seems to us as if the sun goes round the earth than it seems as if the earth goes round the sun! Wittgenstein's emphasis isn't here on the 'no more'; it's instead on the presumption that 'seems as if' is here being deployed with any meaning at all. Anscombe's point, to repeat, is that she was caught up in a mere illusion of sense.

Not only is Wittgenstein offering neither the surface nor the deeper message which Seth ascribes to him, but the lessons Seth suggests are in truth ruled out by Wittgenstein's actual lesson. (Something can't meaningfully be said to seem one way rather than another, with or without greater understanding coming into it, if 'seem' is used without meaning.) Anscombe makes all this perfectly clear, by the way, in the very paragraph from which Seth takes his extract:

The general method Wittgenstein does suggest is that of 'shewing that a man has supplied no meaning [or perhaps: "no reference"] for certain signs in his sentences'. I can illustrate the method from Wittgenstein's later way of discussing problems. He once greeted me with the question: 'Why do people say that it was natural to think that the sun went round the earth rather than that the earth turned on its axis?' I replied: 'I suppose, because it looked as if the sun went round the earth.' 'Well,' he asked, 'what would it have looked like if it had looked as if the earth turned on its axis?' This question brought it out that I had hitherto given no relevant meaning to 'it looks as if' in 'it looks as if the sun goes round the earth'. My reply was to hold out my hands with the palms upward, and raise them from my knees in a circular sweep, at the same time leaning backwards and assuming a dizzy expression. 'Exactly!' he said. In another case, I might have found that I could not supply any meaning other than that suggested by a naive conception, which could be destroyed by a question. The naive conception is really thoughtlessness, but it may take the power of a Copernicus effectively to call it in question.

If it seems peculiar, at this point in our discussion, that Seth should just ignore Wittgenstein's actual lesson, I hope it shan't by the discussion's end. Or at least, that it won't seem peculiar for him. For what we find, again and again as we read his chapter, is him doing precisely what Wittgenstein was teaching us to not do: he (like many a neuropsychologist colleague of his) uses familiar terms but fails to assign them meanings in the novel contexts in which they're redeployed. (Perhaps he's somehow assumed not that the meaning of words in sentences is a function of their use-in-context, but - to borrow a metaphor a friend of mine once offered - that they carry self-contained meanings about with them in little semantic rucksacks on their backs.) And just as he projects his own favoured scheme for representing the interactions between sun and planets onto that scheme's objects - so that he now imagines it makes ready sense to proffer that 'really' or 'of course' the earth goes round the sun - so too does he without demur project his own (rather peculiar, anthropomorphising) forms of description onto the operation of the perceptual system, mistaking this for the proffering of insight into the essential nature of perceptual reality contact itself. 

Let's consider some examples.

'I open my eyes and it seems as though there's a real world out there'

Here's how Seth continues his discussion:

As with the solar system, so with perception. I open my eyes and it seems as though there’s a real world out there. Today, I’m at home in Brighton. There are no cypress trees like there were in Santa Cruz, just the usual scatter of objects on my desk, a red chair in the corner, and beyond the window a totter of chimney pots. These objects seem to have specific shapes and colours, and for the ones closer at hand, smells and textures too. This is how things seem.

Although it may seem as though my senses provide transparent windows onto a mind-independent reality, and that perception is a process of ‘reading out’ sensory data, what’s really going on is – I believe – quite different. [all italics in original]

What does Seth mean here by 'it seems as though there's a real world out there'? Recall that the ordinary use of either 'seems as though' or 'looks as if', in the context of talk about perceptual judgement, is either i) to distinguish veridical perception from perceptual illusion or hallucination, or ii) to express self-conscious caution. The point of i) saying 'it looked as if there was an oasis there' is to make clear that here we're instead talking of a mirage. (And we understand what visual illusions are precisely by contrasting them with the ordinary business of seeing what's actually going on about us.) The point of saying ii) 'well it looks as if the golf ball's gone into the hole' is to make it clear that, being 50 rather than 2 yards away, we can't see it very well so may be wrong. Yet when it comes to 'I open my eyes and it seems as thought there's a real world out there' it's obvious that Seth is using the phrase 'it seems as though' in neither of these senses. He's not i) contrasting cases of misleading visual appearance with cases of ordinary perceptual encounter that take in how things are - since he's talking about ordinary, non-illusory, perceptual experience. It's not as if he could intelligibly say 'But it's all illusion, all a perceptual maya veil' since then we've lost the contrast ('veridical perception') which gives the concept of 'illusion' any content. And he's not ii) talking about being meaningfully cautious in perceptually less-than-ideal circumstances, since he's imagining our just opening our eyes in the day time and seeing whatever's right in front of us. So what does Seth mean by 'it seems as though'? 

Well, he doesn't tell us. He uses words outside of their normal sense-affording context of application, but fails to spell out how he's instead using them. It's as if he were trying to extrapolate from the manifest intelligibility of 'The partly submerged pencil looks bent' to a putative intelligibility for 'This ordinary pencil right in front of me, in ordinary lighting, in a circumstance in which there's no hint of anything being awry, looks straight'. But any such transfer of meaning is illusory; the only illusion we here encounter is (not sensory but) one of meaning. Or, well, perhaps Seth does intend something specific, something else, with his words! But speaking as the reader of a book which is presumably written in order to be understood - it would've been nice to have been told what it was!

'commonsense' ... or 'out there'?

As a foil for his own 'perception as controlled hallucination' view (we'll get to this later), Seth offers  something called a 'commonsense' or 'how things seem' conception of perception. This conception sees the world as being 'out there', and holds that our senses are 'windows' onto this 'external' reality, windows looked out of by 'the self', an 'I behind the eyes' which 'receives' and 'processes' 'sense data' in order to 'build an inner picture of an outside world'. ... And yet, and of course, if we're sitting in our study looking at the objects on the desk or the red chair, we don't experience these objects as 'out there'. After all, it doesn't 'seem' to us, we don't take it as commonsense to suppose, that we're somehow trapped inside our own bodies or heads! We might use the 'out there' locution to articulate what's beyond the front door, but such talk presupposes for its very intelligibility some perceivable 'in here' with which we may contrast it. And yet, in the case of looking at the paraphernalia on one's desk, there's no perceivable 'in here' to offer an intelligibility-providing contrast. (Thank goodness, right? Think how gruesome it'd be if the eyes pointed inwards.) And we look with, rather than through, our eyes. (What would I even look through my eyes with?! We may have adult teeth waiting in the wings to replace our milk teeth, but we don't have further eyes behind the alleged windows provided by our ordinary eyes.) What it is about this view that warrants its description as 'common sense' - rather than as something which, in a different sense of the idiom, really is rather 'out there' - is utterly unclear. It instead looks to me far more like what in C18th philosophy, and in even today's sciences of perception, but not for the man on the street, is fairly common nonsense. 

Dan Dennett
In truth Seth himself acknowledges that this conception is no more than conceptual confusion writ large: he notes that it gets us caught up in what Dennett calls the 'fallacy of 'double transduction'' whereby inner images are invoked to explain perception of the world, leaving us with the equally problematic issue of how we see these inner images. What Seth doesn't do, however, is question whether the question to which this 'inner image' view of perception offers an answer has a cogent sense.  Now, one way to mobilise the 'how do we see?' question, one way to give it at least the appearance of intelligibility, is to imagine first that we really are somehow stuck inside our own skulls, and therefore forced to perceptually reconstruct a now external world using images that appear on the retina. And now the question 'well how do we do that?!' will - to say the least - appear pressing. But undo, avoid, this alienated conception of our perceptual encounter with the world, and it's none too obvious that there's a question remaining which requires the provision of an alternative answer. If 'how do we see?' is to be understood as inviting an answer in neurological terms, then all well and good.  But a psychological or an epistemological answer? What, exactly, is the psychological or epistemological problem that the question is addressing?! I can't myself see one, and so don't see what it is that a psychology of vision is here supposed to be doing. But perhaps there is a good question hereabouts? Well, I'm all ears: do tell!

perception: 'generated by the brain'?

Part of this allegedly 'commonsense' view has it that we ordinarily think that 'A coffee cup out there in the world leads to a perception of a coffee cup generated within the brain.' And it turns out that whilst Seth will dispute a neurophysiological outside-in or bottom-up theory - one in which the neurological events enabling perception follow a unidirectional cascade from eyeball to striate cortex etc - he actually agrees with (what in fact is) this only allegedly 'commonsense' view that perception is 'generated within the brain'. 

Nobody (sane) could disagree that we're dependent for our perception on the activity of the brain. In order to see - to actually see, that is - a cat, the following ingredients are required: 1 medium sized cat, a few ounces of light, 1 or 2 retinae, and a goodly pinch both of optical nerve, and of striate cortex, activation. (Or, if all we're after is a mere cat hallucination, we may leave out the first 4 ingredients.) From none of this does it follow, however, that perception is intelligibly described as 'generated within' the brain. Perception just isn't the right kind of thing to be 'generated' anywhere. (In truth, and en passant: whilst I just offered the above 5 ingredients as a recipe for perception, we ought to acknowledge that, to truly be counted a perceiver, one must also enjoy animate life: a body, and all the neural and physiological movement control apparatus to sustain such a life. But let's not get into that here!)

Now, if we'd subscribed to what Seth construes as the 'commonsense' view - that perception involves inner images - we can I think imagine thinking of perception as being generated, since its in the nature of the coming about of images that we do talk intelligibly of their generation. Absent some such conception, however, and it's hard to see what talk of 'generation' could be getting at, let alone talk of 'generation within the brain'. Compare my paying for my shopping. Paying, like perceiving, is something I do. But must the paying be generated somewhere, perhaps somewhere within me? No, of course not. This isn't to doubt that it happens - it most certainly does (I'm not a thief). It's instead to doubt that there's any clearly intelligible role for talk of 'generation' here. So too, I suggest, for the idea that perception is 'generated by' - or as Seth also says, 'a construction of' - the brain. Perception is not the name of an entity or process; it's instead an action - which is to say, or by which I mean: it's something I do. The action relies on the generation of ATP, neurotransmitters, etc., sure. And when I do perceive, all sorts of activity obtains in my CNS - which activity certainly is generated. But my perceiving is neither such activity itself, nor some further activity generated by it; it's not itself activity in this sense - though in a different sense we might describe it as an activity.   

This idea of perception as something somehow 'coming about' within us pops up throughout Seth's discussion, and colludes there with other peculiar ideas such as that our perceptual consciousness is not of the world but of what is generated by neurological processes: 'Whenever we are conscious, we are conscious of something, or of many things. These are the contents of consciousness. To understand how they come about...[we should consider the functioning of the brain].' The idea here seems to be that what you are conscious of, when you are perceptually aware of something, is something which comes about inside the brain. This however is straightforwardly mistaken: what you're conscious of, when you actually perceive something, is something that's on the desk in front of you! And if you want to know how pencils, laptops and coffee cups come about - well, don't ask a neuropsychologist! (To voice this is not to engage in either naive or sophisticated or philosophical or psychological theorising about perception: it's merely to remember how to use the word 'perceive'.) ... Or perhaps Seth had in mind what philosophers call the 'intentional' rather than 'material' objects when talking about consciousness's 'contents'? But, well, that hardly helps, since we aren't conscious of intentional objects: they're logical constructs, not perceptibilia.

the brain: 'constantly making predictions'?

When presenting his opponent, Seth frequently merges together two different stories about perception into one. The first narrative is an empirically false theory: a neurological story about perception being enabled by a merely bottom-up stream of neuronal activity that begins at the retinae and moves on to the striate cortex, activating 'feature detectors' as it goes along its merry afferent way. The other is not even false: it's the philosophically confused notion of a perceiving person as allegedly being in the desperate epistemic predicament of having to reconstruct the glories of the visual scene from the meagre data to be found on the retinae - as if we were all somehow trapped inside our own skulls. Bring these together and we arrive at the idea of the brain now being involved in somehow solving this poor person's predicament. (As far as I can tell we also find something like this latter muddle in Helmholtz, Gregory, Frith, and in much of contemporary cognitive neuroscience.)

Rather than separating out the empirically testable from the philosophically confused, Seth's own positive alternative is also something of an unholy hybrid. One aspect of it is a scientifically intriguing story - this has visual perception neurologically underwritten not only by afferent 'bottom-up' enervation deriving from retinal stimulation but also, and more importantly, by a complex 'top-down' set of central processes. Unfortunately however, and so far as I could tell (I've read chapter 4 carefully but rather skipped about in the rest of the book), Seth tells us absolutely nothing at all about the actual neurological details. (Perhaps he thought the reader just wouldn't find this of interest. ... I admit to finding this all a great shame. To return to Wittgenstein for a moment: recall his discussion with Bouwsma about the difference between the ghastly pop-sci writing of the likes of Jeans or Eddington, and the patient, empirically detailed, well-grounded, accessible writing found in Faraday's Chemical History of the Candle. The former tends toward the sensational and merges inchoate philosophical claims with poorly elucidated empirical details; the latter deploys careful plain prose to describe actual empirical details, making a circumscribed matter truly intelligible to a lay reader. Speaking for myself I found Seth's popular book ( - I'll own that I know nothing of his actual neuroscientific contributions, and have no idea if I'd even understand them - ), with its lack of actual neuroscientific detail and its sweeping, conceptually awry, claims about consciousness, to be rather more Jeans than Faraday.) The other is the philosophically confused notion articulated above: of an anthropomorphised brain in the dismal epistemic predicament of somehow fathoming an external world from the confines of its own bony cavern.

Here's the story as he develops it:

[T]he brain is constantly making predictions about the causes of its sensory signals, predictions which cascade in a top-down direction through the brain’s perceptual hierarchies (the grey arrows in the image opposite). If you happen to be looking at a coffee cup, your visual cortex will be formulating predictions about the causes of the sensory signals that originate from this coffee cup.

What is it for a brain to make predictions? Well, let's recall first what it is for a creature to make predictions. To predict you must first be able to 'dict' - i.e. verbally communicate, describe what is happening right now. After this you must be able to enjoy the kinds of thoughts about the future which are afforded to such language users as have mastered tensed verbs. Animals and young children can anticipate or expect (we ascribe anticipation to them on the basis of their ongoing coping with a changing world, on the basis of their lack of surprise by changes, etc.) but not predict. Such mastery of tensed verbs is of course not simply a matter of being able to say certain things. It's rather a skill which requires its performer to be culturally situated, to act and react in such ways as warrant us talking of intention and agency, to be able to use a whole lot of other language too, to mean what one says, and so on. Literally none of these things are possible for non-human animals or pre-linguistic children, let alone any of their behaviourally inert, non-vocal, non-verbal internal organs. 

So we're forced to conclude that - since it'd be a nonsense to say that a brain is actually expecting or anticipating, let alone predicting, anything - Seth must either i) be in a terrible muddle, or ii) be, wittingly or unwittingly, using the word 'predict' in a special way. Let's adopt interpretative charity and assume the latter. Seth is really talking, we might say, about 'prediction2'. The question arises: what are the criteria governing its use; what are its ascription conditions? When shall we say of a natural process - e.g. one occurring in the brain - that it constitutes a prediction2? Shall it be that, say, a leaf predicts2 the sun's position if it turns the leaf not to where the sun now is but to where it will maximise photosynthesis in twenty minutes? Shall we say that a pancreas (let's imagine, I've no idea how they work...) predicts2 what insulin shall be required to digest the food that's being eaten if it releases it not in response to current blood sugar levels but instead in response to mastication or smell? Well, we can do! That's just fine! A perfectly innocent metaphor which we perfectly well understand. Unfortunately, however, Seth doesn't tell us what he means by 'predict', so his theory remains either cogent yet occult, or, if he was intending his word in the normal sense, incoherent.

It might be argued: 'But surely scientists extend or creatively misuse words all the time; scientific theorising is riddled with metaphor, and this is all for the better!' It's important to note that with that I have absolutely no quarrel. All I ask is that the metaphor user or new sense deployer stop at some point to explain what this new or extended use is. Otherwise we just shan't know what's being talked about.

'perception as controlled hallucination'

Hippolyte Taine
The central claim of Seth's presentation is that hallucination is 'controlled hallucination'; he takes the idea from Frith (who presumably got it from Taine). We arrive at this prima facie extraordinary claim by pushing the above-described hybrid account to its limits. On the one hand we accept a new empirical theory of vision, one which gives pride of place to an array of 'top-down' activity in the striate cortex, and which correlatively diminishes the role of 'bottom-up', relentlessly afferent, neural activity. On the other hand we accept the mapping of such inner activity in the CNS onto that conceptually corrupt epistemological story: which has us stuck in our own heads, which has perception generated within the brain, which somehow either 'identifies' perceptual acts with their most central neurobiological substrates or sees the latter as 'generating' the former. These central processes, recall, are not under the direct control of afferent sensory stimuli. They are brought into causal connection with such stimuli, but what serves to ready one's bodily movement (i.e. ultimately to underwrite one's very understanding of one's environment) has a lot more to do with internally-generated, rather than afferently controlled, neural activity.

So how do we get to 'controlled hallucination'? Well, having accepted i) the unhelpful philosophical idea that internal neural processes either are or generate perceptual experiences, whilst ii) promoting the valuable empirical idea that perceptual activity is made possible by largely top-down rather than bottom-up neurological processes, we arrive at the idea that actual perceptual experiences exist or arise independently of their objects! It's this odd admixture of epistemology and neuroscience that results in a perfectly legitimate (if radically under-described in 'Being You') neurological story taking on such extraordinary epistemological garb.

The truth, of course, is that far from veridical perception being helpfully theorised in terms of hallucination, it is necessarily that against which hallucination can be understood for what it is. Again, only on an 18th century picture according to which experiences go on somehow 'inside us' could anyone even begin to think of hallucination as the stuff of perception itself. We only have to recall what it is to hallucinate - to have an experience which in some ways if not in all is for one as if one were actually perceiving something despite the fact that one isn't - to realise what a non-starter it is. Perception, considered epistemologically, has to do with our 'openness' to the world, our capacity to 'take in' how things are. Those are epistemological metaphors, mind you: they've nothing to do with subpersonal goings on in the retinae, optical nerve, or striate cortex. There is simply no reason to superimpose the two pictures, any more than there's reason to superimpose the facts about the control of movements by the motor cortex onto a story about acting on intention. (A point Susan Hurley made back in 1998.) Neither the intentions nor the perceptual experiences are 'inside' us; action is not 'output' and sensation is not 'input'; the subpersonal is not the instantiation of the personal.

Søren Overgaard has recently made the point that something like a performative contradiction appears to lie at the heart of the 'perception is hallucination' view, just as it famously does for the 'eliminative materialist' and the 'logical positivist' projects. The cognitive scientist, after all, presumably got the evidence for his view by inspecting brain scans, reading papers, etc. But according to Seth he in fact experienced a hallucination of these scans, papers, etc. It is of the essence of hallucination,  however, that it provides no actual knowledge of the world about one. So the theory would appear to be based upon no actual evidence. Now clearly this is a rather cheeky objection. But it's important to understand it for the challenge it is, which may be put like this: 'You say perception is hallucination, but obviously, in any ordinary sense of the word, it is not. So please, please, tell us what instead you mean by 'hallucination'!'

'chairs aren't red'

Not content to leave us with the paradox of perception as hallucination, Seth next ventures into the notion that chairs (which ones? ... red ones...) aren't red. We get there by way of some standard-issue and utterly sound observations regarding the context- and light-levels- influenced nature of our perceptual judgements combined, once again, with some decidedly C18th philosophy - this time to do not only with a conception of perception as something happening inside us but also with 'secondary qualities' (colour, sound, etc.) not truly  belonging to objects. How now does the story go?

Seth begins by pointing out that our capacity to judge the colour of objects is enabled not only by our responsivity to light of this or that wavelength, but also by a range of contextual factors including the form of illumination in play. 'Take a white piece of paper outdoors and it still looks white, even though the light it reflects now has a very different spectral composition'. So far, so good: it's not simply that different lighting and contexts can mislead us as to the actual colour of a thing (as assessed by taking it outside into daylight); it's also that we can correctly perceive that something is the same colour even when the wavelength of the light reflected from it alters considerably.

From this, however, Seth argues - not against the reductionist notion that wavelength is a happy index of colour, but - that the 'brain infers' the 'invariant property' of 'the way in which the paper reflects light', and that this 'inference' or 'best guess' 'appear[s] in our conscious experience'. We finally arrive at the subjectivist notion that 'colour is not a definite property of things-in-themselves', but instead 'a useful device that evolution has hit upon' to keep track of objects in changing lighting conditions. But how do we get here? How do we get to the idea that colour is a property not of things themselves but is instead 'the subjective, phenomenological aspect' of the 'mechanisms of perception' deployed by the brain to keep track of objects through keeping a bead on the 'way-in-which-they-reflect-light'? 

So far as I can tell, the thought tacitly underlying Seth's theory here owes nothing to science and everything to scientism. It's the thought, that is, that the 'real' properties of things are what's discernible by natural scientific investigation. We can I think see this in such peculiar lines as the following: 'Chairs aren't red just as they aren't ugly or old-fashioned or avant-garde.' But of course, chairs can in fact truly be ugly or old-fashioned or red, so it's obvious he's not using the term 'is' or 'really' in the normal way. With this deviation in play he effectively invites us to simply leave behind our ordinary use of 'real' to distinguish the actual colour of things from the colour they merely appear to have under certain odd lighting conditions or behind a hastily added disguise. But, well, why would we use the word 'real' in this new way? How does it help rather than confound? Which scientific facts about vision are illuminated rather than obscured by such a usage? I must here own that I just can't think of a cogent reason to speak of red chairs not being red.

Seth arrives at the thought that 'we assume that we each see the world in roughly the same way, and most of the time perhaps we do. But even if this is so, it isn’t because red chairs really are red'. Here we're rather back to where we started, with a use of the words 'assume' and 'same way' without a manifest sense in the context of their applications. I assume that you and I both see the red chair as red? Well, sure, I typically assume that you've nothing wrong with your eyes. But what does it even mean: to see a red thing as red? (To not be making a mistake about its colour perhaps?) Or: what's even meant by 'same or different way' here? Once again, Seth puts words together in what look to be grammatically well-formed sentences - but we're left with no understanding of what they mean in the particular context of his prose.