Kai asked:

a. Would a Turing test that tested aesthetic experience be possible? Why or why not?

b. Would this kind of Turing test be a more accurate way to assess human intelligence (or human consciousness) or less accurate or as accurate as the standard Turing test? Why or why not?

Answer by Jürgen Lawrenz

Your question might have been relevant if it was based on a sound (i.e. verifiable) scientific theory of intelligence, consciousness and aesthetic experience. Regrettably all we really possess is a little well-founded philosophical speculation, which would dismiss these assumptions with a tired handwave and reference to ‘category error’.

As for any scientific merit of the Test, you must surely be aware that it measures (if that’s the word!) the degree of gullibility of easily fooled people, who are deceived day in, day out by fakes and imposters for as long as history remembers. That’s asking a lot! In science, you’ll remember, subjective impressions are not allowed, because they are not quantifiable. So that’s pretty much the end of this story, I would say.

Apart from this, nothing more is required than a minimum awareness of the mechanics of computer programming. Forget the hype, but take a cold hard look at it. The you’ll see that every such algorithm is human intelligence flowcharting procedures that can be implemented by a machine. A century ago the acme of this was the steamship. Today’s computer belongs into the same category, except that electricity replaced steam and the functional aspects became miniaturised and more diverse. But one as the other implements its procedural routines on the basis of clear-cut and well-understood causal principles.

So here as there, it is pretty absurd to ask for a machine to ‘acquire’ autonomous consciousness, intelligence and aesthetic sensibility, simply through the agglomeration of ever more bits and pieces and branches on the flowchart!

Accordingly the Turing Test, as it relies on human gullibility for its results, has no truly scientific merit. Philosophically it has none at all: being deceived is hardly a philosophical subject. Even speculatively it works only on the proviso that you first reduce all human specificity to quantifiable features, i.e. regarding life processes in an exclusively molecular context. But this is hardly the proper way of differentiating a rock from a rodent, as you’ll surely agree. There is an ineradicable internal conflict between the cause-and effect mechanisms that apply to non-living objects and the spontaneity of living creatures. Which spells out that you have the choice between thinking of yourself as a human creature or as a biological computer. But the computer does not have that choice.