The Turing Test concerns machine intelligence. The question is: when can it be said about a machine that it is “intelligent”? Answer: when we cannot tell from observing its conversation that the machine is a machine.
Trying to think about this test, I have identified three areas that seem essential to me personally in any human interaction – three features of personality which I look for in every human encounter. All three are readily tested through conversation.
To some extent perception can be trained. For instance, looking at lots of song birds and comparing one’s findings with entries in ornithological guides trains the “mind’s” eye: it teaches the looker to look for features such as “rump” and “wing coverts” which an untrained looker might not notice (having no clue what he is looking at/for); or: looking at lots of very fine details (say, magnitude +6 stars with one’s naked eye) teaches one the trick of looking at things by not looking at them directly but rather by focusing one’s sight just to the right or left of the object (in order to engage one’s peripheral vision); and: smelling lots of roses teaches one not only that roses of different colors smell differently, but that a rose cannot be smelled too long before the brain no longer detects the smell (usually about 20-40 seconds), after which the nose must be “washed”; and that one can improve one’s sense of smell by pouting one’s lips (so that the upper lip creates a kind of “funnel” under one’s nostrils).
But all of these are techniques; and are useless if their owner is not interested to look/ smell/ taste/ observe; and then interested/able to reflect on what s/he sees. This kind of curiosity for the world around us is linked to something Konrad Lorenz called “exploratory instinct” and ascribed to all mammals (mice and hamsters in particular). But it is clear that not all mammals possess it: a great number of human beings are perfectly uninterested in observing and learning. And when they do (as tourists in my city do, for instance) they are perfectly happy to follow a manual (notice only what is pointed out to them).
Yet, to be in any way interesting a person must be able to tell us something new about the world, something they have not read in a book, or heard on TV, but something they saw and realized themselves. Otherwise, why listen to them in the first place? (And without listening The Turing Test cannot be performed).
Authentic aesthetic and/or emotional response
Another thing I look for in people is their ability to surprise me with original, novel, and authentic responses to the world. By authentic I don’t mean heartfelt, but – their own. i.e. ones not borrowed from others (Mom, friends, teachers, TV). I do not only ask them their opinions of things or love stories (or “art stories”), but also look at their clothing, accessories, apartments and furniture. By this measure most people are unoriginal in every way: they furnish their homes at IKEA and paint the walls white; they dress like they see others dress; their professional life is dictated to them by the market (“plastics”); and their emotional life is a copy of what they have seen on TV and read in romantic novels. If you ask them why they do this, or feel that, they shrug: as far as they are concerned, that’s how it is and there is nothing they can (or care to) do about it. Generally, my interlocutors are surprised when they are told they could do something/feel about something differently; prodded to say why not they can’t say why not, merely resist in a kind of panicky, animal, unthinking refulal: it is simply unthinkable. If you think about it, this is how computer generated characters in fantasy action games behave: they behave in some way and you cannot argue with it.
I am interested to talk to people who live their lives “differently” – who do not marry, or reproduce, for instance; or who do not live all their lives where they were born; or who do not buy a 42 inch flat screen TV when everyone else does – and generally do not buy anything when everyone else does; or who do not take a mortgage; or who opt out of the state pension program; or who do not own a car; or who wake up before daybreak; or who do not go to the beach on Labor Day; or who, during rush hour, when all traffic goes zig, drive the only car in the opposite lane, zagging; or who don’t know who won last night on penalty shots and genuinely don’t care; or who marry a woman twice their age, or live with two.
But this in itself is not interesting: a lot of non-conformist behavior is hard-wired – and hard-wired actors can’t tell you why they are doing what they are doing. These are not so interesting, no matter how odd their course of life.
The really interesting people are the ones who are doing odd things – or normal things, but oddly – because of a calculation: people who have thought about their objectives and then plotted their own course because that is what they wanted to do and this was the best way to get there. (They are called “autotelic”).
I can safely say that on these three measures, a very large majority of human beings would fail the Turing Test. Indeed, to an observer applying these three measures to his test, most of us would appear to be automata engaged in an elaborate deception to produce the (false) impressions that we are independently observant, sensitive and autotelic, that we have a taste, or emotions, or cunning; that we are, in other words, actually human. But this deception is easily exposed: put your ear to their forehead and listen carefully: you will distinctly hear the low murmur of the cooling fan.
If you’re working in aesthetics, psychology, or musicology and know a little math and programming you can do this rather easily:
Introduction: New, previously unheard melodies of Chopin, Mozart, Monteverdi, Schubert and Vivaldi are instantly recognizable to the ear of anyone who has heard a hundred others. The reason why must be that the melodies of each composer are ruled by a unique algorithm (which is what our brains recognize subconsciously); such an algorithm must be expressible as a polynomial (i.e. ax + bx2 +cx3…). (Given how easily our brains learn to recognize each composer, it is unlikely to be a multivariable polynomial — ie. ax + py + bx2 +sy2 +cx3 ty3…).
Method: plot n Mozart melodies on a graph; then use graph-fitting software to approximate the values of a, b, c etc. Each polynomial thus generated can be said to be a property of the brain of the particular composer — his cognitive “signature”.
Interest: Think about it: would it not be fascinating to know that Mozart’s melodies all have the structure of, say, 1x + (1/2)x2 +(1/3)x3 ?
A free PhD follow-up idea:
Calculate signature algorithms of the A class composers (Bach, Chopin, Mozart, Monteverdi,Vivaldi) to illustrate the general procedure. Then show that it is impossible to generate by this method similar signature polynomials for B class composers (Berlioz, Brahms, Janacek, Tchaikovsky etc.) — and conclude that their music can therefore be mathematically demonstrated to be directionless and haphazard.