Nature of Intelligence: Semantics

A counter-argument to myself

We are mostly surrounded by connectionist and statistical forms of artificial intelligence these days. Another theory Ibelieve in i.e. that these ‘connectionist machines are implicitly implementing symbolc ai or turing-machines’ implictly also relies on a very ‘behaviouralism’ or ‘functionalism’ kind of intelligence. Why? Because then it doesn’t matter ‘how’ something is implemented; in such a world the only difference between a artifical neural net, a Turing Machine, and perhaps, a biological neural net, is just the substrate: at the end of the day, each of them are capable of doing any kind of computation possible, and consequently any kind of intelligence. I believe that the idea that human intelligence is a collection of Turing Machines, and is a very ‘functional’-ism idea.

But is there a possibility at all that true intelligence does requires semantic qualia-like understanding and manipulation of the subject matter at hand?

And these machines: i. if they are lacking in qualia, ii. and, if they’re actually lacking ‘semantic’ understanding: Could that rule them out from being truly intelligent? (As opposed to human intelligence.)

Informally, I feel like, if a machine is a collection of fixed TMs or algorithms, I find it quite difficult to believe that it will actually have any qualitative experience or qualia. Because most of these machines are doing is symbol pushing. This, of course brings us to to the hard problem: where and how does qualia come forth?

With neural nets, perhaps because of their internals being somewhat mysterious, I feel like maybe there is a chance that they could possibly implicitly have qualia?

Could intelligence be a ‘meaning’-oriented phenomenon? Tied up with the a animal’s survival and their capacity for meaning?

When I am going through a hard chapter in a next book or a hard logical problem, and unable to parse it’s ‘meaning’ quickly and feel a frustration –> maybe that is what is intelligence (or it’s lack). Maybe I could do a rule-following and push symbols around to get the right answer, and perhaps that is what is more akin to what machines are doing.

Maybe intelligence requires the subjective feeling of ‘understanding’ something, of ‘doing’ something, of qualia.

Could that be explained away by the TM-idea? If I am getting stuck in a logical problem, or a programming puzzle, maybe I do not have the right ‘mental’ algorithm to parse the problem correctly? Maybe the problem or the textbook hasn’t described it in a eloquent manner? maybe I am tired?

But then, is ‘understanding’ then a huge component of intelligence? I am re-thinking everything.