Tag Archives: thought

The thought that counts.

Posted on by
11

Neurocritic asks a great question here, neatly provoking that which he would have defined – thought. What is thought, and what are individual thoughts? He quotes reports that we have an estimated 70,000 thoughts a day and justly asks how on earth anyone knows. How can you count thoughts?

Well, we like a challenge round here, so what is a thought? I’m going to lay into this one without showing all my working (this is after all a blog post, not a treatise), but I hope to make sense intermittently. I will start by laying it down axiomatically that a thought is about or of something. In philosophical language, it has intentionality. I include perceptions as thoughts, though more often when we mention thoughts we have in mind thoughts about distant, remembered or possible things rather than ones that are currently present to our senses. We may also have in mind thoughts about perceptions or thoughts about other thoughts – in the jargon, higher-order thoughts.

Now I believe we can say three things about a thought at different levels of description. At an intuitive level, it has content. At a psychological level it is an act of recognition; recognition of the thing that forms the content. And at a neural level it is a pattern of activity reliably correlated with the perception, recollection, or consideration of the thing that forms the content; recognition is exactly this chiming of neural patterns with things (What exactly do I mean by ‘chiming of neural patterns’? No time for that now, move along please!). Note that while a given pattern of neural activity always correlates with one thought about one thing, there will be many other patterns of neural activity that correlate with slightly different thoughts about that same thing – that thing in different contexts or from different aspects. A thought is not uniquely identifiable by the thing it is about (we could develop a theory of broader content which would uniquely identify each thought, but that would have weird consequences so let’s not). Note also that these ‘things’ I speak of may be imaginary or abstract entities as well as concrete physical objects: there are a lot of problems connected with that which I will ignore here.

So what is one thought? It’s pretty clear intuitively that a thought may be part of a sequence which itself would also normally be regarded as a thought. If I think about going to make a cup of tea I may be thinking of putting the kettle on, warming the pot, measuring out the tea, and so on; I’ve had several thoughts in one way but in another the sequence only amounts to a thought about making tea. I may also think about complex things; when I think of the teapot I think of handle, spout, and so on. These cases are different in some respects, though in my view they use the same mechanism of linking objects of thought by recognising an over-arching entity that includes them. This linkage by moving up and down between recognition of larger and smaller entities is in my view what binds a train of thought together. Sitting here I perceive a small sensation of thirst, which I recognise as a typical initial stage of the larger idea of having a drink. One recognisable part of having a drink may be making the tea, part of which in turn involves the recognisable actions of standing up, going to the kitchen… and so on. However, great care must be taken here to distinguish between the things a thought contains and the things it implies. If we allow implication then every thought about a cup of tea implies an indefinitely expanding set of background ideas and every thought has infinite content.

Nevertheless, the fact that sequences can be amalgamated suggests that there is no largest possible thought. We can go on adding more elements. There’s a strong analogy here with the formation of sentences when speaking or writing. A thought or a sentence tends to run to a natural conclusion after a while, but this seems to arise partly because we run out of mental steam, and partly because short thoughts and short sentences are more manageable and can together do anything that longer ones can do. In principle a sentence could go on indefinitely, and so could a thought. Indeed, since the thread of relevance is weakened but not (we hope) lost at each junction between sentences or thoughts, we can perhaps regard whole passages of prose as embodying a single complex thought. The Decline and Fall of the Roman Empire is arguably a single massively complicated thought that emerged from Gibbon’s brain over an unusually extended period, having first sprung to mind as he ‘sat musing amidst the ruins of the Capitol, while the barefoot friars were singing vespers in the Temple of Jupiter’.

Parenthetically I throw in the speculation that grammatical sentence structure loosely mirrors the structure of thought; perhaps particular real world grammars emerge from the regular bashing together of people’s individual mental thought structures, with all the variable compromise and conventionaljsation that that would involve.

Is there a smallest possible thought? If we can go on putting thoughts together indefinitely, like more and more complex molecules, is there a level at which we get down to thoughts like atoms, incapable of further division without destruction?

As we enter this territory, we walk among the largely forgotten ruins of some grand projects of the past. People as clever as Leibniz once thought we might manage to define a set of semantic primitives, basic elements out of which all thoughts must be built. The idea, intuitively, was roughly that we could take the dictionary and define each word in terms of simpler ones; then define the words in the definitions in ones that were simpler still, until we boiled everything down to a handful of basics which we sort of expected to be words encapsulating elementary concepts of physics, ethics, maths, and so on.

Of course, it didn’t work. It turns out that the process of definition is not analytical but expository. At the bottom level our primitives turn out to contain concepts from higher layers; the universe by transcendence and slippery lamination eludes comprehensive categorisation. As Borges said:

It is clear that there is no classification of the Universe that is not arbitrary and full of conjectures. The reason for this is very simple: we do not know what kind of thing the universe is. We can go further; we suspect that there is no universe in the organic, unifying sense of that ambitious word.

That doesn’t mean there is no smallest thought in some less ambitious sense. There may not be primitives, but to resurrect the analogy with language, there might be words.  If, as I believe, thoughts correlate with patterns of neural activity, it follows that although complex thoughts may arise from patterns that evolve over minutes or even years (like the unimaginably complex sequence of neural firing that generated Gibbon’s masterpiece), we could in principle look at a snapshot and have our instantaneous smallest thought.

It still isn’t necessarily the case that we could count atomic thoughts. It would depend whether the brain snaps smartly between one meaningful pattern and another, as indeed language does between words, or smooshes one pattern gradually into another. (One small qualification to that is that although written and mental words seem nicely separated,  in spoken language the sound tends to be very smooshy.) My guess is that it’s more like the former than the latter (it doesn’t feel as if thinking about tea morphs gradually into thinking about boiling water, more like a snappy shift from one to the other), but it is hard to be sure that that is always the case. In principle it’s a matter that could be illuminated or resolved by empirical research, though that would require a remarkable level of detailed observation. At any rate no-one has counted thoughts this way yet and perhaps they never will.

Zappiens unreads

Posted on by
13

attentionAre our minds being dumbed by digits – or set free by unreading?

Frank Furedi notes  that it has become common to deplore a growing tendency to inattention. In fact, he says, this kind of complaint goes back to the eighteenth century. Very early on the failure to concentrate was treated as a moral failing rather than simple inability; Furedi links this with the idea that attention to proper authority is regarded as a duty, so that inattention amounts to disobedience or disrespect. What has changed more recently, he suggests, is that while inattention was originally regarded as an exceptional problem, it is now seen as our normal state, inevitable: an attitude that can lead to fatalism.

The advent of digital technology has surely affected our view. Since the turn of the century or earlier there have been warnings that constant use of computers, and especially of the Internet, would change the way our brains worked; would damage us intellectually if not morally. Various kinds of damage have been foreseen; shortened attention span, lack of memory, dependence on images, lack of concentration, failure of analytical skills and inability to pull the torrent of snippets into meaningful structures. ‘Digital natives’ might be fluent in social media and habituated to their own strange new world, but there was a price to pay. The emergence of Homo Zappiens has been presented as cause for concern, not celebration.

Equally there have been those who suggest that the warnings are overstated. It would, they say, actually be strange and somewhat disappointing if study habits remained exactly the same after the advent of an instant, universal reference tool; the brain would not be the highly plastic entity we know it to be if it didn’t change its behaviour when presented with the deep interactivity that computers offer; and really it’s time we stopped being surprised that changes in the behaviour of the mind show up as detectable physical changes in the brain.

In many respects, moreover, people are still the same, aren’t they? Nothing much has changed. More undergraduates than ever cope with what is still a pretty traditional education. Young people have not started to find the literature of the dark ages before the 1980s incomprehensible, have they? We may feel at times that contemporary films are dumbed down, but don’t we remember some outstandingly witless stuff from the 1970s and earlier? Furedi seems to doubt that all is well; in fact, he says, undergraduate courses are changing, and are under pressure to change more to accommodate the flighty habits of modern youth who apparently cannot be expected to read whole books. Academics are being urged to pre-digest their courses into sets of easy snippets.

Moreover, a very respectable recent survey of research found that some of the alleged negative effects are well evidenced.

 Growing up with Internet technologies, “Digital Natives” gravitate toward “shallow” information processing behaviors characterized by rapid attention shifting and reduced deliberations. They engage in increased multitasking behaviors that are linked to increased distractibility and poor executive control abilities. Digital natives also exhibit higher prevalence of Internet-related addictive behaviors that reflect altered reward-processing and self-control mechanisms.

So what are we to make of it all? Myself, I take the long view; not just looking back to the early 1700s but also glancing back several thousand years. The human brain has reshaped its modus operandi several times through the arrival of symbols and languages, but the most notable revolution  was surely the advent of reading. Our brains have not had time to evolve special capacities for the fairly recent skill of reading, yet it has become almost universal, regarded as a natural accomplishment almost as natural as walking. It is taken for granted in modern cities – which increasingly is where we all live – that everyone can read. Surely this achievement required a corresponding change in our ability to concentrate?

We are by nature inattentive animals; like all primates we cannot rest easy – as a well-fed lion would do – but have to keep looking for new stimuli to feed our oversized brains. Learning to read, though, and truly absorbing a text, requires steady focus on an essentially linear development of ideas. Now some will point out that even with a large tome, we can skip; inattentive modern students may highlight only the odd significant passage for re-reading as though Plato need really only have written fifty sentences; some courteously self-effacing modern authors tell you which chapters of their work you can ignore if you’re already expert on A, or don’t like formulae, or are only really interested in B. True; but to me those are just the exceptions that highlight the existence of the rule that proper  books require concentration.

No wonder then, that inattention first started to be seriously stigmatised in the eighteenth century, just when we were beginning to get serious about literacy; the same period when a whole new population of literate women became the readership that made the modern novel viable.

Might it not be that what is happening now is that new technology is simply returning us to our natural fidgety state, freed from the discipline of the long, fixed text? Now we can find whatever nugget of information we want without trawling through thousands of words; we can follow eccentric tracks through the intellectual realm like an excitable dog looking for rabbits. This may have its downside, but it has some promising upsides too: we save a lot of time, and we stand a far better chance of pulling together echoes and correspondences from unconnected matters, a kind of synergy which may sometimes be highly productive. Even those old lengthy tomes are now far more easily accessible than they ever were before. The truth is, we hardly know yet where instant unlimited access and high levels of interactivity will take us; but perhaps unreading, shedding some old habits, will be more a liberation than a limitation.

But now I have hit a thousand words, so I’d better shut up.

Turing Test tactics

Posted on by
19

turing22012 was Alan Turing Year, marking the hundredth centenary of his birth.  The British Government, a little late perhaps, announced recently that it would support a Bill giving Turing a posthumous pardon; Gordon Brown, then the Prime Minister, had already issued an official apology in 2009. As you probably know, Turing, who was gay, was threatened with blackmail by one of his lovers (homosexuality being still illegal at the time) and reported the matter to the authorities; he was then tried and convicted and offered a choice of going to jail or taking hormones, effectively a form of oestrogen. He chose the latter, but subsequently died of cyanide poisoning in what is generally believed to have been suicide, leaving by his bed a partly-eaten apple, thought by many to be a poignant allusion to the story of Snow White. In fact it is not clear that the apple had any significance or that his death was actually suicide

The pardon was widely but not universally welcomed: some thought it an empty  gesture; some asked why Turing alone should be pardoned; and some even saw it as an insult, confirming by implication that Turing’s homosexuality was indeed an offence that needed to be forgiven.

Turing is generally celebrated for wartime work at Bletchley Park, the code-breaking centre, and for his work on the Halting Problem: on the latter he was pipped at the post by Alonzo Church, but his solution included the elegant formalisation of the idea of digital computing embodied in the Turing Machine, recognised as the foundation stone of modern computing. In a famous paper from 1950 he also effectively launched the field of Artificial Intelligence, and it is here that we find what we now call the Turing Test, a much-debated proposal that the ability of machines to think might be tested by having a short conversation with them.

Turing’s optimism about artificial intelligence has not been justified by developments since: he thought the Test would be passed by the end of the twentieth century. For many years the Loebner Prize contest has invited contestants to provide computerised interlocutors to be put through a real Turing Test by a panel of human judges, who attempt to tell which of their conversational partners, communicating remotely by text on a screen, is human and which machine.  None of the ‘chat-bots’  has succeeded in passing itself off as human so far – but then so far as I can tell none of the candidates ever pretended to be a genuinely thinking machine – they’re simply designed to scrape through the test by means of various cunning tricks – so according to Turing, none of them should have succeeded.

One lesson which has emerged from the years of trials – often inadvertently hilarious – is that success depends strongly on the judges. If the judge allows the chat-bot to take the lead and steer the conversation, a good impression is liely to be possible; but judges who try to make things difficult for the computer never fail. So how do you go about tripping up a chat-bot?

Well, we could try testing its general knowledge. Human beings have a vast repository of facts, which even the largest computer finds it difficult to match. One problem with this approach is that human beings cannot be relied on to know anything in particular – not knowing the year of the battle of Hastings, for example, does not prove that you’re not human. The second problem is that computers have been getting much better at this. Some clever chat-bots these days are permanently accessible online; they save the inputs made by casual visitors and later discreetly feed them back to another subject, noting the response for future use. Over time they accumulate a large database of what humans say in these circumstances and what other humans say in response. The really clever part of this strategy is that not only does it provide good responses, it means your database is automatically weighted towards the most likely topics and queries. It turns out that human beings are fairly predictable, and so the chat-bot can come back with responses that are sometimes eerily good, embodying human-style jokes, finishing quotations, apparently picking up web-culture references, and so on.

If we’re subtle we might try to turn this tactic of saving real human input against the chat-bot, looking for responses that seem more appropriate for someone speaking to a chat-bot than someone engaging in normal conversation, or perhaps referring to earlier phases of the conversation that never happened. But this is a tricky strategy to rely on, generally requiring some luck.

Perhaps rather than trying established facts, it might be better to ask the chat-bot questions which have never been asked before in the entire history of the world, but which any human can easily answer. When was the last time a mole fought an octopus? How many emeralds were in the crown worn by Shakespeare during his visit to Tashkent?

It might be possible to make things a little more difficult for the chat-bot by asking questions that require an answer in a specific format; but it’s hard to do that effectively in a Turing Test because normal usage is generally extremely flexible about what it will accept as an answer; and failing to match the prescribed format might be more human rather than less. Moreover, rephrasing is another field where the computers have come on a lot: we only have to think of the Watson system’s performance at the quiz game Jeopardy, which besides rapid retrieval of facts required just this kind of reformulation.

So it might be better to move away from general stuff and ask the chat-bot about specifics that any human would know but which are unlikely to be in a database – the weather outside, which hotel it is supposedly staying at. Perhaps we should ask it about its mother, as they did in similar circumstances in Blade Runner, though probably not for her maiden name.

On a different tack, we might try to exploit the weakness of many chat-bots when it comes to holding a context: instead of falling into the standard rhythm of one input, one response, we can allude to something we mentioned three inputs ago. Although they have got a little better, most chat-bots still seem to have great difficulty maintaining a topic across several inputs or ensuring consistency of response. Being cruel, we might deliberately introduce oddities that the bot needs to remember: we tell it our cat is called Fish  and then a little later ask whether it thinks the Fish we mentioned likes to swim.

Wherever possible we should fall back on Gricean implicature and provide good enough clues without spelling things out. Perhaps we could observe to the chat-bot that poor grammar is very human – which to a human more or less invites an ungrammatical response, although of course we can never rely on a real human’s getting the point. The same thing is true, alas, in the case of some of the simplest and deadliest strategies, which involve changing the rules of discourse. We tell the chat-bot that all our inputs from now on lliw eb delleps tuo sdrawkcab and ask it to reply in the same way, or we js mss t ll th vwls.

Devising these strategies makes us think in potentially useful ways about the special qualities of human thought. If we bring all our insights together, can we devise an Ultra-Turing Test? That would be a single question which no computer ever answers correctly and all reasonably alert and intelligent humans get right. We’d have to make some small allowance for chance, as there is obviously no answer that couldn’t be generated at random in some tiny number of cases. We’d also have to allow for the fact that as soon as any question was known, artful chat-bot programmers would seek to build in an answer; the question would have to be such that they couldn’t do that successfully.

Perhaps the question would allude to some feature of the local environment which would be obvious but not foreseeable (perhaps just the time?) but pick it out in a non-specific allusive way which relied on the ability to generate implications quickly from a vast store of background knowledge. It doesn’t sound impossible…