I mean, Ethics: Not Just a County Next to Thuffolk

by Simon Smith

The flashbacks are finally beginning to fade. That boiling sea of faces, at once horrified and fascinated, witnesses to a pedagogic train crash, no longer haunt my dreams. That succubus of sex and ethics, demon of a Thursday afternoon in late November, is banished once again; and in its wake, our thoughts turn, once again, to the matter of philosophical fragmentations.

Just before that forest of wildly thrashing limbs, exercised frantically in an indescribable series of arcane gesticulations, blotted out the sun, the point we had reached, or almost reached, was that ethics is ethics, no matter the context or point of application. This is not to say that context adds nothing at all to our moral reasoning; only that, if our basic principle is, say, not to harm others, then that ought to be upheld no matter what the circumstances. Of course, the circumstances will, to some degree, dictate the manner in which our principle is upheld; the specific way in which we avoid harming others will depend largely on what we are doing at the time.

One field of doings, as it were, which always seems to face greater demands for moral scrutiny is technology. And rightly so, one might say, given the faith we place in it to solve all our problems and make us happy, despite the lack of evidence that it can do any such thing. There is, as Susan Pinker shows in The Village Effect (Atlantic Books, 2014), very little evidence to support the commonly accepted dogma concerning the value of exposing children to computer technology at the earliest possible opportunity; quite the opposite, in fact. Given how easy touch-screens are to use – easy enough for most adults to get to grips with – it’s not clear what advantage there might be for a three-year old to be able to do likewise rather than, say, active reading with a real human being.

One especially obvious, because potentially dramatic, field which seems to demand some hard moral thinking is that of Artificial Intelligence. As it happens, I don’t hold any special fears regarding the “rise of the machines”, either for them or us. Even if the Dotard in Chief hasn’t actually sucked the world into a nuclear firestorm long before it becomes an issue, the application of intelligence, artificial or otherwise, might well be a good idea. We’ve tried everything else and look where we’ve ended up. All around the world, people live and die in poverty; wars are fought over nothing very worthwhile and, wherever possible, ignored; mendacity, corruption, and outright criminality remain exceedingly popular; there’s every chance that we are polluting the world beyond habitability; and the Americans have a game-show host for a president. Surely now, more than at any other time in history, intelligence, any intelligence, has to be worth a try.

That said, the sheer temperamentally of much modern technology, downright obtuseness on occasion, do sometimes make me wonder.

Speaking philosophically, which is to say, less cynically, it remains to be seen whether the conception of intelligence which lies at the heart of AI research will ever be rich enough to present any real moral concerns. At present, it seems rather too nebulous to warrant the kind of alarm some people – including made-up people like Elon Musk and the Pope – are keen to generate. Where specificity has crept into the conceptualising, the analogical extension of “intelligence” has been degraded to the point where it’s hardly recognisable. More worrying is the way that degraded analogy is projected back upon its original source: i.e. the persons who constructed it in the first place.

During the latter part of the 20^th Century, it became fashionable to anthropomorphise computer technology, in direct contravention, nota bene, of IBM’s own code of conduct. Despite the rampant anthropomorphism, there remains a world of difference between “memory” and “data storage”, a difference which carelessness and linguistic ineptitude have all but erased.

This marks a dismal, though not particularly surprising, failure to understand how analogies work. It is, sadly, a symptom of the naïve realism that continues to pollute so much modern thought. (Naïve but never innocent, no matter what Peter Byrne thinks, never innocent; realism is always, ultimately, pernicious.) At the root of it is the assumption that, between analogue and analogate, there is a simple one-to-one correspondence: x is like y, therefore y is like x to the same degree and in the same way.

That’s “analogate”, by the way, not, as I assumed for many years, “analgate”; that, I suspect, would require a very different theory of language. And lubricant, lots of lubricant.

So x is to y as y is to x. Well yes, of course. Or rather, no, not really. That’s not how analogies work. Analogies aren’t mirror-images; they refract when they, or rather we, reflect. It is just possible that my love really is like a red, red rose (prickly and covered in greenfly); and I’m sure you bear more than a passing resemblance to a summer’s day (hot, sweaty, and buzzing with flies). Such analogies cannot simply be reversed, however. It would be a different thing entirely to suggest that a particular Tuesday afternoon in August was just like you.

So much seems perfectly obvious; except that it clearly isn’t, otherwise we wouldn’t be so quick to forget it. What began as a way of describing computers – data storage, it works a bit like our memory; so all the stuff you put in, the computer sort of “remembers” it – has turned back upon us. It’s like the threefold law of return in medieval witchcraft; language is a lot like magic in many ways. A computer’s data storage works a bit like our memory, therefore memory works a bit like data storage. Well the logic is sound enough but the idea is still cobblers. (A prime example, here, of how logic can be sound without actually being correct; demands for what necessarily must be so often are though.)

Cobblers, it may be, but there we are, stuck with the now widely held supposition that our memories, and by extension our brains, work like computers. The next step is a simple one: we just need to let those rather vague, allusive words slip out of sight; “sort of” and “a bit” goes first, of course, then “like”. No matter that these are what made the analogy work in the first place. Once they’ve been properly sublimated and repressed, we can conveniently forget that we were ever talking analogically at all and pretend that we’re really being terribly precise and scientific and technical. My, aren’t we clever?

Well, no, not really, because along with forgetting that we started out talking analogically, we’ve also, quite carefully, forgotten what we did to make the analogy work in the first place. It is unlikely that the clever dick who came up with the analogy between brains and computers actually meant it to be taken literally; the idea that those who first heard it would have taken it that way seems highly improbable.

In order to make the analogy work at all, we had to strip it down to its bare essentials, pull out all those bits that made it a genuine human experience. Stuff like consciousness and agency, all the various ways in which we ourselves are involved in our recollections; then there’s the whole web of connections, interlaced ideas and images, not to mention the way all our senses can get involved, sounds and smells can be especially evocative. And what about how we remember things? Not just as bits of data, that’s for sure; more like sudden flashes, lightening in a thunderstorm; or images of people, places, and events; sometimes a memory is a whole narrative, rich with layers of interpretation and personal perspective, wherever our attention, our consciousness came most sharply into focus. Sometimes it’s just a feeling. Just? Oh hardly that. Do you remember the smell of school dinners? Or the Sunday night feeling you used to get when you were a child? What about the lead-up to Christmas in years gone by? Or when you had to say “goodbye” to someone terribly, terribly important to you. Not “just”, oh no.

The point here is that what we remember and how we remember are vastly more complex than what a computer does with its data. In order to apply the notion of “memory” to a machine, or any other object for that matter, all the human, all the personal, aspects of it have to be stripped away.

Of course, there’s nothing wrong with doing that, it can be a very informative process, especially when it comes to diagrammatising the physical universe – oh yes, and just where did you think all those images of energy and process and function actually came from? Indeed they did not; Hume was absolutely spot on there.

That doesn’t change the fact that the extension of (severely watered down) personal analogies is an incredibly informative and therefore valuable process. Only, it can become a problem if we forget that we were talking analogically in the first place, if we mistake our analogies for literal truths. Obviously, doing that in this case wouldn’t make the slightest bit of sense. Computers don’t do all the things we do when we remember things. Of course they don’t. Unless we convince ourselves that the stripped down analogy of memory is all that memory really is, in and of itself; unless, that is, we take that stripped down analogy and return it unto its source, re-apply it to the place from which it came. Then, we might be making a serious mistake.

There, if you like, is the real moral problem at the heart of modern technology: the ease with which it allows us, even encourages us, to objectify ourselves and others, to treat real human beings as machines. And what a very old problem that is; Descartes’ legacy inviting us to step outside the world of real experience and reduce everything – and everyone – left inside to automatons. As observers and describers, we, of course, remain above such crude mechanics. But here’s the new twist: we don’t just treat people as things, actual objects that we might, just possibly, have to encounter in the real world and so think about how we conceive of them. Now we treat them, and ourselves, as idealised objects. We’ve become our own diagrammatic fictions. And all in the name of science, of truth, of the advancement of human knowledge. How very clever of us.

But can that really be right? Those realities that can only be seen and understood, which, frequently, can only be thought, by means of analogies and diagrams, can they be more true, more real than the human beings we encounter everyday? Can they be more real than the people who taught us to speak and think and construct diagrams in the first place?

Mathematics is the language of the sciences, it provides access to the to the world as it really is, allegedly. Mathematics is a logical system of signs and symbols. Those symbols have no natural or actual corollary in the world; if they did, they’d be no good to science. What, I wonder, makes us so sure that a world which can only be conceived of and made sense through those signs and symbols is more real than the people who sit across from us at the dinner table?

Oh dear, I seem to have digressed quite lamentably and given myself a case of the vapours in doing so. We started out well enough, ruminating on the fragmentation of philosophy, specifically ethics. Now here we are grumbling about the apparently universal failure to understand how analogies work and the objectification of people that follows. Quite evidently, our ruminations lack proper focus and attention; a consequence and continued effect, no doubt, of having taught classes on sex and ethics. Post-Traumatic Stress Disorder is, I believe, the right phrase.

We shall, once again, attempt to return to the matter in hand once all thoughts of talking about sex in front of sixty first-year undergraduates have returned to the dark and shadowy places of the human psyche.

…No! It’s in the trees! It’s coming!