Re: cognitive != observable?

> Or maybe the ones that bring a reply, and reply to that. To explore the
tree
> of conversations, to see which brings more conversation? Is conversation
> itself the goal? To extend the banter?

It's an interesting point. Women, for instance, often engage in "talking for
the sake of talking", because they like the feeling that everyone is in
dialogue with each other. The actual words themselves don't matter much;
it's the act of banter itself that makes them feel close. Women even
sometimes feel that they wish they had more problems so that they could
start more conversations and have more to talk about. Not to solve the
problems, of course -- that's besides the point. They look forward to the
getting-together and the talking-about-the-problems and
having-the-coffee-together part.

Men tend to value content more. For them, there should be some point or end
goal to conversation directly attributable to the information in the talk.
When men have nothing to gain from talking, they say little -- fathers and
grown-up sons and brothers can go for years without picking up the phone,
and when they do it's usually to say "How ya doing? I'm fine, yeah. Anything
new with you? No? Well, nothing's new here. Okay, see ya." Women find that
odd; they could easily exchange talk about what their respective pasts have
been like even if nothing new arises in informational terms. To not do so
would be to risk appearing unsupportive. In short, most men value data, and
most women value feelings.

It was actually for the above reasons that (in earlier times, so I've been
told) that male scholars weren't sure if most women even had souls, but were
instead zombies, because their behavior appeared circularly defined. I
personally prefer to believe the situation exists because of evolutionary
and biological necessity -- a species composed entirely of male minds or
only of female minds would be at a competitive disadvantage to others. As to
why each person tends not to contain both mindsets, I imagine that there
exists a mutual exclusion about them.

It perhaps begs the question of whether we should split the Turing Test into
male and female versions. At what point does an AI convince a group of women
that it is one of them? Would it be a subset of a male AI's capabilities, or
some other set of functionality altogether? Could an AI exhibit both male
and female conversational styles without the supposed mutual exclusion?

Ray