Follow the reluctant adventures in the life of a Welsh astrophysicist sent around the world for some reason, wherein I photograph potatoes and destroy galaxies in the name of science. And don't forget about my website, www.rhysy.net



Monday, 15 February 2021

I Like Locke (I) : Knowing What You Know

Okay people, you and I need to have a little talk.

I post a fair bit of philosophy stuff here, right ? I go on quite a bit about free will and the nature of mind versus matter, qualia, how we establish objectivity and all that, yes ? Yes, I do. You know it's true. Especially if you also follow Decoherency.

And yet... not a single bloody one of you has ever said to me, "You've never read John Locke ? You must go at once ! Here, take my car !".

You. Utter. Bastards.

Once you're done hanging your heads in shame, it's time for a trilogy on what I make of Locke's epic "Essay Concerning Human Understanding". I got it in this edition, which is an abridged version. Unlike most philosophy texts this one doesn't come with an extensive introduction, or indeed any introduction at all. So all I know about context is from the blurb on the back : that Locke himself advised the essay should be abridged, which is reassuring since I don't normally like abridgements.

The Essay, it must be said, is a bit rambling, but that's what you get in an era when everything had to be hand-written. I've chosen what I see as the three main themes of the work : (1) knowledge and reality; (2) free will and identity; (3) how belief relates to reason. Drawing all of Locke's related arguments together to from a more coherent whole seems like a worthwhile activity to me, so that's what I'm gonna do. Anyway, it's not as if I have anywhere better to go.

Although (surprisingly) Locke is infinitely more comprehensible than, say, Shakespeare, he does have the usual old style of long, flowing sentences. These read more like a recorded speech than modern text. There are some cases in which this makes it hard to discern if he means one thing or its exact opposite (though nowhere near as often as in Gibbon), but fortunately most of the time he clarifies things immensely with well-chosen examples. While eloquent, his long sentences don't lend themselves to easy quotations, so I'll largely be paraphrasing. If I misinterpret him, then that's obviously entirely my fault.

In this first post I'll look at the main topic of the Essay : knowledge. In this post truth age of fake news and bullshitting, fighting facts with fiction and vice-versa, it's surely more important than ever to understand what knowledge really is.


John Locke's Guide To Figuring Stuff Out : knowledge and reality



Everybody craves certainty. But can we ever really be certain of anything ? I'm not sure. And I'm not sure what Locke thought about it either.

What exactly do we mean by "knowledge" anyway ? Is it different from "certainty" ? Terms like knowledge, belief, faith, reason, certainty and opinion are all context-dependent and fraught with implicit assumptions. When we ourselves say, "I know this to be true", we usually count ourselves certain, or very nearly so. We generally say that this is quite different to the statement, "I think this is true". No real problem there.

But when someone else says to us, "I know this to be the case", unless that person has earned an exceptionally high level of trust, it doesn't have the same effect. We might be a bit more convinced of what they say than we were before, but we won't say we know it. We don't doubt that that other person thinks they know it, but we don't necessarily say, "Oh, you saw a mermaid eating a flaming avocado in your airing cupboard ? Are you sure ? Oh, you are sure ? Well I never ! I guess that's it for marine biology then." We view other people's "knowledge" as a sort-of opinion, no matter how fervent they seem in their sincerity.

Lots of caveats to this of course : it depends strongly on our relationship to the other person. For some interesting data on how much this matters, see this, along with some accompanying discussion here.

Locke is a bit confusing as to whether he means "knowledge" to mean true certainty or something else. But mainly he describes everything as ultimately "ideas", i.e. information : "we can have knowledge no further than we can have ideas". This renders true certainly uncomfortably impotent... but it's not all bad. After all, it's similar to my own notion that everything relies on certain basic assumptions about the world : ignore the assumptions that reality is objective, external, and (after a fashion) measurable and you basically say we can't know anything about the external world at all. 

The point is that those assumptions are themselves unprovable. So yes, we have only "ideas", but we are necessarily limited to those - so any discussion on knowledge simply has to work within that framework, not examine the unprovable assumptions themselves. And as Locke elaborates, our "knowledge" is certainly enough to be useful; if we are all suffering a dream, then it's a remarkably self-consistent one :

How vain, I say, it is to expect demonstration and certainty in things not capable of it; and refuse assent to very rational propositions, and act contrary to very plain and clear truths, because they cannot be made out as evident, as to surmount even the least pretence of doubting. He that, in the ordinary affairs of life, would admit of nothing by direct plain demonstration, would be sure of nothing in this world but of perishing quickly.

Elsewhere Locke is more unclear as to whether he means knowledge should be described as true certainty or only this more moderate, idea-limited variant. But I think his meaning is, overall, clear enough that I can gloss over this.


I think, therefore I'm certainly thinking, therefore, I can be certain, hurrah !

Locke doesn't tackle anything as mundane as how we form conclusions. Rather, he's concerned with the fundamental nature of our different types of ideas and whether we can call any of them "certain". His answer is here is happily unambiguous : yes. While all knowledge is a type of idea, not all ideas are knowledge. Just because ultimately everything we perceive happens inside our head (and is to a degree subjective) does not mean that some ideas aren't demonstrably correct and others demonstrably false. The idea of a mermaid eating avocado flambĂ© doesn't have any claim on certainty just because our knowledge of, say, wooden chairs is also ultimately a type of idea.  

But how ? Locke emphasises that much of what we "know" is really just opinion, and true knowledge is very limited indeed. But we do have some. How, then, is this possible ? How can we distinguish our imaginary fancies from proper knowledge ?

One answer is that you can definitely fight people using chairs, but try using mermaids instead and see how well that goes.

Locke boils this down to two sorts of certainty. One is subjective, internal, and inherently unprovable to anyone except ourselves, while the other concerns the outside world.

On the first, there are some things that we can truly "know" all by ourselves. One, following Descartes, is our own existence : I think therefore I am. If I think I'm thinking, I must (a) be thinking and (b) exist, otherwise I couldn't be thinking that I'm thinking*. Locke expands on point (a) to go beyond Descartes. He says, for example, if I think I'm thinking of a circle, I must actually be thinking of a circle. While I might be mistaken about some things that I more deeply believe (more on that in part three), when it comes to the fundamentals of my own existence or my more superficial thoughts, I have no reason for doubt. I can be as certain of those as I ever can be about anything.

* Compare and contrast this with the truly ludicrous notion of illusionism**, which literally says, "I'm not really thinking, I just think I'm thinking."
** Only on reviewing that earlier post when this one was nearly complete did I realise just how similar my stance is to Locke.

What of the certainty of the external world, then ? All knowledge, says Lock, is derived from perception and our thoughts about what we've perceived. There are some ideas so fundamental that they essentially atomic : indivisible and only achievable via our senses. You cannot conceive of whiteness, or colour, or softness, based only on description. While Plato didn't hold with the idea of knowledge as perception (since perception can be fooled), Locke circumvents the problem by going to the most basic aspects of what our senses tell us : qualia.

In this remarkably strong illusion, the thick blue lines are actually parallel.

Objects, according to Locke, have primary and secondary qualities. Primary qualities are attributes that they really have : size, mass, motion, that sort of thing. Secondary qualities are "powers" that the objects have to induce sensation : pain, heat, taste, yellowness etc. A fire feels warmer depending on our proximity to it, and only causes actual pain if we're very close indeed. That heat is a sensation doesn't negate the fact that the fire is composed of energy-rich material, but the sensation itself is not the same as the matter inducing it. There are no teeny-weeny pain particles in the fire that are out to get us. Pain isn't physical, but it derives from physical phenomena - so you can't stick your hand in a fire and not have it hurt. 

In other words, there are certain irreducible properties that you cannot simply imagine, or imagine away : pain being a good example (also solidity, heat, colour). When you remember pain, it hardly seems but a pale shadow of the real experience, and you likewise can't voluntarily simply choose to stop experiencing it. These are what Locke calls "simple ideas", the indivisible atoms of perception. 

The different nature of these "simple ideas" from our internal images (specifically the weakness and voluntary nature of the latter) is what assures us that they indicate something real is at work. Unfortunately, we can never have knowledge of them that we can directly impart to anyone else. We must know them by ourselves. Likewise there are "intuitive" truths, not in the sense of setting the clock on the microwave without needing to read the manual, but in being self-evident.

My hypothesis : electronics designers of the late 20th century created such convoluted user interfaces that the collective emotions of millions of disgruntled users echoed back through time, where they intersected the head of one John Locke. This caused him such anguish that he came up with the notion of "intuitive" knowledge as a means to prevent this agony far in the future. He failed.

This again would seem to agree with my general anti-"reality is not perception" stance. Yes, it's very difficult indeed, perhaps impossible in the strictest sense, to know the true "primary qualities" of anything, but that doesn't mean we can't get a meaningful view of reality anyway. We define things as being the source of our perception. We don't really need to know their "true" Platonic forms. It's enough to treat our view as incomplete, not wrong, unless something comes along to contradict us.

What Locke sadly doesn't discuss much of is disagreement. It would seem impossible to disagree on intuitive truths, whereas I've posited the "fire is hot" problem where such a thing does appear to occur. But I suppose there's no real difficulty here : such disagreements are likely mere appearances, the result of different heuristics and abstract generalisations that people use to analyse problems that can't actually be reduced to first principles, even if a quick overview would suggest otherwise. We will, however, return to this in part three.


Seeing is believing, but not the other way around

Abstract notions in particular are very interesting. You cannot directly perceive justice or responsibility or desire or luck*. They are generalised from specific, perceptible examples, but the notions themselves defy direct perception : yet nevertheless we can be said to know them, in some sense - or at least understand them. They are not "real" in the ordinary, physical sense of the word. Yet, says Locke, despite this we can indeed say we have true knowledge of mathematics and - in an interesting juxtaposition - morality. When we prove (in Euclidian geometry) that the angles of a triangle always add up to 180 degrees, we have done so for all time for all triangles. The actual physical existence of any particular triangle is irrelevant; a mathematical proof is indeed certain.

* A friend of mine had this idea for a superhero who can literally smell crime. I would watch that show.

Yes, fine, clever clogs, it's not true in non-Euclidean space. I'm making implicit assumptions, so you shut up.

It's hard to deny this*. Yet Locke then makes the leap that by the same token we can have certain knowledge of morality as well as mathematics, which is much harder to swallow. Unfortunately he doesn't provide any examples, so this throwaway comment is left unexamined. It seems to me that Locke's statement that knowledge is a type of idea, but ideas are not necessarily knowledge, means that we can accept his basic premise without having to accept an objective morality if we don't want to.

* If you don't agree that 1+1 = 2, then we've probably reached a genuine "fire is hot" problem, a fundamental disagreement at the axiomatic level with no clear resolution. Or in other words, you're a nutter.

Anyway, external perception is quite different from the internal perception that Locke calls intuition. This is the highest, most certain sort of knowledge of all, e.g. knowing your own thoughts, understanding abstract generalisations that cannot be perceived with the external senses. The ordinary perception from our senses is not the same. Here's one of those confusing instances were Locke allows "knowledge" to have different levels of confidence :

The notice we have by our senses of the existing of things without us, though it be not altogether so certain as our intuitive knowledge, or the deductions of our reason employed about the clear abstract ideas of our own minds; yet it is an assurance that deserves the name of knowledge. Nobody can, in earnest, be so skeptical as to be uncertain of the existence of those things which he sees and feels. At least, he that can doubt so far will never have any controversy with me, since he can never be sure that I say anything contrary to his own opinion.

Oh Locke, what a joker...

We might doubt that what we're seeing is correct (Plato's analogy that someone a long way off could be wrongly mistaken for Socrates), but not that we're seeing what we think we're seeing. I don't think there's any scope for doubt on this point. Less convincing is the argument we can know that our perception is basically accurate :

If any one say, "a dream may do the same thing, and all these ideas may be produced in us without any external objects", he may please to dream that I make this answer : 1. That it is no great matter, whether I remove his scruple or no : where all is but dream, reasoning and arguments are of no uses, truth and knowledge nothing. 2. That I believe he will allow a very manifest difference between dreaming of being in the fire, and actually being in it.

The first point is then an implicit statement that we simply have to make assumptions, which I have no argument with, but the second is harder to sustain. Just because our dreams to tend to have a distinctly dreamlike quality to them, it does not follow that this is always necessarily the case. I don't see any really rigorous grounds by which we can prove anything about the external world with true, absolute certainty. Better to take the assumption that we're not in a dream as a given, and accept that our definition of knowledge is forever constrained by that : as he says, otherwise all argument collapses anyway, and we'd have to give up and go home.

A stronger argument is perhaps Locke's assertion that we cannot have "simple ideas" except through direct perception. This does indeed give us very strong grounds for accepting the existence of an external, objective world (something must exist externally to induce sensations within us) but, not, frustratingly, truly certain knowledge of it. Our memories could be being continually deceived - yes, something like yellowness must exist, but that's about as far as it goes. Whether the object we're looking at really has the properties which cause us to perceive it as yellow, even whether chairs truly exist, is something that will remain forever less than perfectly certain.

Well, if nothing else, that surely explains why philosophers haven't managed world domination yet.

Also Finland is totally fake. The philosopher's approach to world conquest is simply to redefine "world conquest" and then go home for cocoa.



Intuitive knowledge helps explain why maths is hard

Even the "knowledge" we can have under standard assumptions is necessarily limited :

I cannot be certain that the same man exists now, since there is no necessary connection of his existence a minute since with his existence now : by a thousand ways he may cease to be, since [the time when] I had the testimony of my senses for his existence. And therefore, though it be highly probable that millions of men do now exist, yet, whilst I am alone, writing this, I have not that certainty of it which we strictly call knowledge, though the great likelihood of it puts me past doubt, but this is probability, not knowledge.

Our intuitive knowledge is limited in that it cannot tell us fundamentally new truths about the universe. We cannot spontaneously imagine colour if we've never perceived it before. Our perceptive knowledge is also limited because not everything is directly perceptible. And both of these sorts of knowledge cannot be taught : you can acquire a probabilistic assessment of them, but not knowledge of them except directly through your own faculties. Fortunately, we possess another faculty, that of reason, which allows for the third type of demonstrative knowledge.

First, Locke deserves a lengthy quote on intuitive knowledge :

Thus the mind perceives that white is not black, that a circle is not a triangle, that three are more than two and equal to one and two. Such kinds of truths the mind perceives at the first sight of the ideas together, by bare intuition; without the intervention of any other ideas : and this kind of knowledge is the clearest and most certain that human frailty is capable of. This part of knowledge is irresistible, and, like bright sunshine, forces itself immediately to be perceived, as soon as ever the mind turns its view that way; and leaves no room for hesitation, doubt, or examination, but the mind is presently filled with the clear light of it. It is on this intuition that depends all the certainty and evidence of all our knowledge. He that demands a greater certainty than this, demands he knows not what, and shows only that he has a mind to be a skeptic, without being able to be so.

Again echoing the need for basic assumptions that cannot be inherently proven. But obviously not everything is intuitive; knowledge is an idea but ideas are not necessarily knowledge. However, sometimes we can create chains of intuitive knowledge that demonstrate other truths which are not immediately obvious. So long as each link in the chain is solidly intuitive, this allows for so-called demonstrative knowledge.

Although things get very tricky if a "proof" is so complicated that hardly anyone can understand it.

This allows us to circumvent our innately limited comprehension - but only to an extent. As a simple example, I can instantly identify a square or a triangle or maybe a hexagon or even an octagon. Much more complex than that, though, and I'll run into trouble. A hundred-sided shape ? There's no way I can tell you it has a hundred sides just by looking at it, because my brain can't hold that much information all at once*. 

* I could perhaps learn to recognise specific hundred-sided shapes and do pretty well by comparing them this way, but I definitely could not learn the general form as I could so easily do with a triangle.

So am I stuck ? No ! Of course, for each side I can intuitively determine that they're straight* and where each side ends. So I can demonstrate the number of sides by counting them, allowing me to leapfrog my raw mental processing power. My stage of reasoning is at every stage perfectly intuitive, so I can know the shape has a hundred sides with every bit of certainty as I can know it has sides at all, even though I cannot immediately or directly intuitively determine the number of sides from a simple inspection. And this skill, to a large degree, I can teach to other people. Hurrah !

* With caveats. In most situations I can see this directly with sufficient accuracy. But even if something confuses me, as in the optical illusion from earlier, I can take numerical measurements to confirm what I think I'm seeing.

This too has limitations though. My usual analogy is mathematics. Counting is simple enough, but complex numbers, integration, matrix transformations, anything to do with tensors... at some point of complexity I hit a wall beyond which I cannot progress. "Tensors ?!?!", I cry. "TENSORS !?!? AAAAAARRRGHHH !!!". And then I run screaming from the room, never to return.

This very nice answer on Quora* may help explain why. If a problem requires a certain level of working memory and processing speed, I won't ever be able to understand it. I can use a lot of heuristics, such as learning what typical polygons look like, to help me do surprising feats of mental agility, but if a problem has a limit below which it cannot be further simplified, and my brain can't handle even that, then I'm forever limited : if I need to hold a hundred different things in my head all at once, then it sucks to be me. Admittedly, "different things" presents a bit of quandary, since the brain can use its clever heuristics to simplify enormously complicated things into more manageable units, but still, there's a limit to how much it can do this.

* I would somewhat dispute the assertion there that mathematical ability is a good proxy for intelligence, though, or at least I'd say that mathematical prowess at best constitutes only a very particular form of intelligence and certainly shouldn't be mistaken for wisdom.

Locke calls the process of assessing the connections between intuitive ideas our faculty of reasoning. This too is something not dissimilar to something I've mentioned before : that understanding is a kind of knowledge of the connections between things. Combining this with Locke and the Quora answer, we can perhaps now improve on this. The reason I can't understand some things is because they require a simultaneous knowledge of a larger number of connections than my brain will allow, and/or more rapid processing than I'm capable of. 

These feels pleasantly satisfying. However, it doesn't help with artificial intelligence, which can have "knowledge" of an unlimited number of connections but still a piss-poor understanding of, say, frogs. Only in part is this down to heuristics (with the brain's ferocious ability to simplify allowing us to achieve more than the sum of our parts, whereas computers are currently limited to brute force approaches) - this only gets us so far... I don't think a computer's level of "understanding" of those atomic-level ideas, those that require an inner awareness (like whiteness) is in any way comparable to mine. More on that in part two.

Honestly I don't understand how anyone ever got anything done without tea.


Beyond reason

But I digress. The point is that deductive reasoning can produce true knowledge, not opinion, and thus we can learn more than our pure senses and intuition tell us. There are limits though - some philosophical, some technical. The philosophical variety concern our knowledge of the primary qualities of objects : "And seek ye in vain for certain and universal knowledge of unperceived qualities in substances", says Locke, "We neither know the real constitution of the minute parts on which their qualities do depend; nor, did we know them, could we discover any necessary connection between them and any of the secondary qualities."

You might say, but we could measure the chemical composition of chocolate and determine which ones give rise to the taste of chocolate. But exactly how we have the sensation of taste, that entirely non-physical experience of having taste, remains utterly mysterious. "It's electrical currents in the brain !" you might say. Oh yes, and why does one current correspond to taste and other to anger or an obsession with fish, then ? Do all electrical currents correspond to experiences ? If not, why not ?

The technical limitations are easier. Locke doesn't say it directly, but implies that the more sophisticated our measuring equipment, the more information we have to analyse regarding various connections. True, some connections are unknowable, but not all. Reason, says Locke, is also what lets us judge whether connections are certain or only probable, how we judge fact to be distinct from opinion, and this constitutes the bulk of our knowledge.

For where the mind does not perceive this probable connection, where it does not discern whether there be any such connection or no; there men's opinions are not the product of judgement, or the consequence of reason, but the effects of chance and hazard, of a mind floating at all adventures, without choice and without direction.

But though we have, here and there, a little of this clear light, some sparks of bright [intuitive] knowledge, yet the greatest part of our ideas are such that we cannot discern their agreement or disagreement by an immediate comparing them. And in all these we have need of reasoning, and must, by discourse and inference, make our discoveries.

......


So that's Locke's view of the world, a thoroughly common-sense view that I enthusiastically support. Mind you, I've inevitably read it with my own biases, and I've attempted to remove any apparent contradictions to paint a more cohesive picture than might be had with a more thorough reading. But any contradictions there may be are nowhere near as fatal as they are, say, in Stoic philosophy. Reality is external, we know it through direct perception, we can generally distinguish it from dreams and suchlike without any fatal problems. Our true knowledge is highly limited but there is a thing we can legitimately call certainty, and we aren't constrained to mere probability. Yes, it relies on assumptions, but :

After all, why even bother having a discussion if you're convinced that none of it is real ?

The blurb on the back warns, though, that its selection traces the rise and fall of empiricism, so I'm looking forward to seeing how that turns out. But Locke has other ideas that should be explored first. So next time I'll cover that other Big Topic of philosophy : free will.

No comments:

Post a comment