The day is coming when a single carrot, freshly observed, will set off a revolution.
Cézanne
It’s not what you look at that matters, it’s what you see.
Henry David Thoreau
People neglect the reality of the illusory world
Huang Po
What You See Is What You Don’t Get
Did you know, dear reader, that you have something called an Orienting Reflex? First mentioned by the Russian physiologist Ivan Sechenov in his 1863 book Reflexes of the Brain and later more extensively described by Pavlov (the one with the dogs). The Orienting Reflex, in simple terms, for these are the terms the Director understands best, is “an organism’s innate reaction to a novel stimulus”. If, whilst reading these paltry words someone were to creep up behind you and make a loud noise, you would engage in a complex behavioural repertoire of sensory, skeletal-muscular, and autonomic responses. For this reason, Pavlov also called this phenomenon the ‘What is it?’ response.
In the 1950s this response was studied systematically by the Russian scientist Evgeny Sokolov, particularly with regard to the phenomenon called “habituation”, which refers to a gradual “familiarity effect” and the subsequent reduction of the Orienting response. So reading these paltry words in an environment where people were frequently creeping up on you and making loud noises would not provoke the same ‘What is it?’ response if you were reading them in a library. Not that anyone goes to libraries any more. But that is a matter for another day.
Scientists have beavered away on this issue and now have a pretty good understanding of the neurological basis for the establishment of the orienting reflex. It seems that the response to novelty has an instinctual basis. Something I’m sure our dear friend Prof Dawkins would explain makes perfect sense in terms of predators not becoming prey and all that kind of thing. But I have my concerns. Mainly that this view of how we respond to the world ‘out there’ can lead us astray when it comes to really making sense of what this world is and therefore our relation to it. The recent and continuing furore over the disruptive march of AI might be a good way to illustrate my point.
The early pioneers of AI imagined the world in a fairly Sokolovian way, namely that there’s the world out there and it’s made out of objects and so we navigate the world by building an internal model of that external world. We make maps and plans that constitute a manipulation of that internal representation and then we act out the representation; responding to things to the extent that they conform or don’t, to our internal models or maps.
It’s kind of the old empirical idea, philosophically speaking: sense data is in some way given to us from ‘outside’ and from that sense data we build these models ‘inside’ and then that’s how we think and that’s how we operate in the world. And most people would agree that’s pretty much how it is. But the problem is, that is not (bold for emphasis) how it is at all. And that’s why we don’t have these all-purpose robots, despite being promised them for ages.
Those AI researchers tried to build these robots that could navigate in the external world and the way they tried to do that was by attempting to give them models of the external world. The environments we’re speaking of were fairly basic because after all you can’t model the entire world. Which is kind of the point, but we’ll get to that later. Maybe.
So they built ‘toy’ environments that the AI robots could potentially model and then tried to have them do simple things in the toy environment. What they discovered was that they couldn’t even get the robots to see the toy environment to start with. Because there are no ‘simple’ toy environments. Any environment is almost infinitely complex. Let’s say you have an environment that’s made out of pyramids and columns and spheres; just plain geometric forms. Well then you’d still have the problem of variant lighting, for example. Is a pyramid in the morning the same as a pyramid at midday, or five minutes past midday, or three seconds past midday? How much illumination change is necessary before the object isn’t the same object? And it’s no good saying we’ll just ignore illumination when it comes to identifying objects because without illumination you wouldn’t be able to see anything at all. Obvs. But we do manage to infer the stability of an object across transformations of illumination. So how? The answer is, as far as my meagre reading has taken me, nobody really knows. In fact nobody really understands how we’re able to perceive objects at all. It should be pretty much impossible when you think about the endless number of things you can do with a single image of anything.
Added to this difficulty, for what it’s worth, is my very strong hunch that trying to model perception in this way is a non-starter precisely because that is not what perception is.
From around the 1960s or so, a very similar problem emerged in the world of literary criticism when our friends the postmodernists came along and pointed out how it was impossible to derive a single so-called canonical meaning from any given text. There are a multiplicity of ways to interpret any given text just as there are any given pyramid. And if you were so inclined you could multiply that problem. Forget a single text for example, you could try a single paragraph or a sentence, or even a word. Each of these ‘objects’ are susceptible to an endless number of interpretations. The problem is that to perceive something there has to be an end to it. And this is precisely why I have never really sorted out my bookshelves.
Let me explain.
Picture the Director typing these paltry words in his study, surrounded by books arranged on shelves in higgledy-piggledy fashion. Now you might think that it is all fairly self-evident how I ought to arrange them. But dear reader, it’s not. I know what you’re thinking, you’re thinking, ‘Come on! Don’t be telling me that there’s a near infinite number of ways to arrange those books!’ To which I can only respond that this is the axiomatic structure of your a priori perceptions manifesting itself as self-evident fact to you. Obvs. I mean, what if I were to decide to organise the books according to colour, or thinness, or age, or thickness of paper, or the area of the blank space exactly half an inch below the 9th page in the third chapter..? To which, you respond, ‘Well that’s a silly way of organising your books!’ To which I would reply, how exactly do you know it’s silly?’ The thing is you can’t claim that it’s ‘self-evidently’ silly because the self-evidence of the silliness is actually the mystery. And you know how the Director loves a mystery.
Back to those all-purpose robots the younger Director was promised were just around the corner. It turns out, the corner might be a tiny bit closer to being rounded (although it never will be). Those clever scientists have decided that part of the way to solve the problem of perception in AI is by embodying cognition, which means incarnating artificial intelligence in an embodied structure. Solving the problem of perception in this way means duplicating the process of evolution in hardware. We shall watch this space.
It might be time for a little anecdote. This one occurred many years ago when my father was in a fairly specialised type of hospital. For reasons we do not need to go into at this point of the anecdote, the door to my father’s room had to be kept permanently open. Now this was all very well except for the fact that one patient also undergoing similar treatments kept on wandering into my father’s room causing all sorts of consternation in the process. This episode came to mind when reading for this week’s paltry words when I came across something called ‘utilisation behaviour’.
Let me explain…
What happens if you’re afflicted by this neuropsychological condition is that you lose the ability to inhibit your motor response to the presentation of an object. Now what does that mean? Well, normally we might think that it goes something like this:
object → thought about object →motor response to object
But it doesn’t go like this; the object itself announces its utility in the perception of it.
Our eyes map patterns of arrays onto our visual system but part of your visual system is actually your motor output system and so that when I look at this bottle in front of me, it’s not
see bottle → hand grip → drink
Bottle is ‘hand grip’ and ‘hand grip’ is ‘drink’. Anyone with ‘utilisation behaviour’ loses the ability to inhibit the motor response. If you had this condition and I put a bottle in front of you, you would pick it up and drink from it. And if you walk down a hallway and there’s a door open you will go through the door and it’s not because you see the object door and think ‘door’ and then think ‘walk through’ and then walk through. It’s that you perceive ‘door’ as ‘walk through place’ and if you lack inhibition, you can’t stop acting out the perception. I’m not sure this would have helped my father understand the serial (and often cereal) trespasser into his room but I like to think he would have found it interesting.
So the point is, we don’t see objects and infer meaning, we see meaning and infer objects. The primary object of perception is meaning, not objects. To try to strip the world of meaning and see it as populated by meaningless objects, turns out to be bad science. Which is ironic. But nice too.
Which leads us to a point I may have mentioned before, it is a mistake to confuse our maps of the world with the actual world as it presents to us. It is better to seek to see Cézanne’s carrot, rather than our representation of it, if you will. Better because, for one reason, it seems to be how we’re made.
But the idea that something comes into being for us through our attention to it rather than our already conceived theory/map/idea of it seems to run very much counter to the general tendency of our world’s way of approaching things (including those who are building those robots). I invite you to pop in to a Tiffinian poetry class for further illustration of this point that abstraction and categorisation take us away from true perception.
Speaking of poetry, our friend Wordsworth, throughout his writings, returned again and again to the importance of paying attention to how we pay attention. He writes of the difference between his experience of the Lake District, its mountains, lakes, waterfalls, rocks etc. as a child, compared to his experience of it when he was much older. When he was young, the environment spoke to him as a living presence and he was somehow intoxicated by it when he was in its presence. As he got older he found he couldn’t help generalising what he saw and reducing the mountains, lakes, waterfalls etc. to the categories of ‘mountains’, ‘lakes’ and ‘waterfalls’. In his Ode on the Intimations of Immortality he writes, shades of the prison house close around the growing boy talking about his own experience; that this prison house of abstraction, generality and theory takes over from the immediacy of experience. And if Wordsworth is suggesting we need to try and get away from representation and back into the presence of the moment, well I think it’s something important to entertain.
In many ways we could probably blame Descartes and his chums for this tendency of ours to see a world ‘in here’ (Director gestures at head) and a world ‘out there’ (Director gestures around the room). To our conventional way of approaching the world, perception and indeed all experience is believed to be divided into two essential ingredients: a subject on the inside and an object on the outside. This belief that perception is divided into these two essential ingredients underpins our entire world culture. It underpins the way we think, feel, act, perceive, use language and relate to one another. Which is a shame.
Enshrined in this belief is the idea that there is an outside world made of something called ‘matter’ and an inside self made of something called ‘consciousness’ or ‘awareness’. When William Blake was asked about the nature of perception he famously wrote; If the doors of perception were cleansed every thing would appear to man as it is, Infinite. For man has closed himself up, till he sees all things thro’ narrow chinks of his cavern. Much like Wordsworth’s prison, Blakes’ cavern speaks of how our conventional way of understanding perception cuts us off from the world which we pretend we are perceiving, when we should rather be seeing ourselves and the world as one and the same.
I fear we have run out of our time. Maybe we shall sail these waters in other columns but let me sign off for now with the words of that other Romantic poet, Tom Waits, who wrote movingly on the business of perception in San Diego Serenade:
I never saw the morning ’til I stayed up all night
I never saw the sunshine ’til you turned out the light
I never saw my hometown until I stayed away too long
I never heard the melody until I needed the song
I never saw the white line ’til I was leaving you behind
I never knew I needed you until I was caught up in a bind
I never spoke “I love you” ’til I cursed you in vain
I never felt my heart strings until I nearly went insane
I never saw the east coast until I moved to the west
I never saw the moonlight until it shone off of your breast
I never saw your heart until someone tried to steal, tried to steal it away
I never saw your tears until they rolled down your face
Until next time, Happy Reading/Bringing the world into being!
This week’s Director’s tip is prompted by the recent inclement weather.
Muddy Umbrella?
Give it a rub with a rag dipped in methylated spirits. Unless you are in possession of a silk umbrella. Such delicate specimens should be rinsed gently in a pint of hot water in which a tablespoon of sugar has been dissolved.