Social communication is a very complex thing...

Post Reply New Topic

...if you think of it from a robot/computer programming perspective.

First of all, you need a bunch of "if" statements, and it is multi-dimensional too. Its not a simple "if the person say this then say that". You have to consider the emotions invovled. You have to consider the context and background of the conversation. You have to consider their personality. You have to consider their age. Some things are inappropriate to say at certain times.

And then there is the receiving end. Is the person making a joke? Is it sarcasm? Is it a figure of speech? Why would they say such things, given the situation they are in? What context/background does the person come from?

And then there are all these nonverbal cues.

I sometimes wonder if people will ever be able to make a robot that acts like an NT. I highly doubt it.

The problem with programming it is that an IF ... THEN is a purely logical construct. Humans are irrational by nature, not logical. Even Aspies, while often predisposed to more highly structured thinking, are still irrational with enough frequency to make IF ... THEN statements a tool of limited utility when trying to programatically imitate their behavior.

Social communication is a cpontradiction in terms.

What goes on inside an individual CPU is not actually all that bad - complex, yes, balancing all kinds of inputs and working with a very large data bank. But logical [IF you grant the premises, which is a very big IF].

Thing is, take any two - even two almost interchabgeable NTs - and you get handshaking attempts between Arabic Win XP running a freeware com program designed by a Russian to use Japanese and an OS 9 Mac doing Yiddish. The code is not accessible and the inputs often - even between NT twins - fail to match anything alreasdy in there.

Yes, quite! (Seeing as I'm in the process of reverse-engineering it...)


If/Then statements are what's used in a finite state machine, which has inherent limitations and are always linear. The non-linearity involved with decomposing, analyzing, and synthesizing human social communication requires a Turing-complete machine (TCM), which is much more complicated than If/Thens. The human brain is a TCM, and recent developments in artificial intelligence (AI) are starting to make computational TCMs a possibility.


I'm calling this affect dynamics. I'm basing it on the verbal, facial, and body affect that people present during conversation. I'm finding it rather difficult to implement, but still do-able.


Regressive analysis, part of AI.


I'm calling this the personality vector. It also includes culture, language, interests, etc.


This is one of the most difficult things to model. I'm using modal logic for this, since it's too difficult to run discovery directly. Eventually, my project might be able to actually list out the rules associated with this.


Affect dynamics plus sociolinguistic norms is the best bet here.


Probabilistic inference and neural networks here.


That's part 2 of my plan. First, I just want to get myself to act like an NT.

It looks like you got a head on your shoulders. You should consider working on AI; we're always looking for smart people.