Jay asked:

This is a question that came up on the language forum at http://www.a-i.comand is still under debate
on both sides: Does conversation measure intelligence? Can a being that cannot communicate be
said to be intelligent? Or does the responsibility lie in the being that is receiving the information?
(From Answers page 13.)

============

I have another slant on this. I don't think that it could be fairly said that conversation measures
intelligence, for there are many sorts of intelligence (see for example Howard Gardner's book
"Multiple Intelligences"), and verbal intelligence is only one of them. Here I agree with Steven Ravett
Brown, who gives some nice examples of other intelligences.

However, I think that there is good evidence that thought and language are more closely tied than his
answer might imply. Firstly, though, we have to look a little at what we mean by language.
Conversation as verbal interaction clearly does not capture all of language, as the examples of sign
language, written language and body language all attest.

Mary Midgley (in a number of books — e.g. "The Ethical Primate", "Utopias, Dolphins and
Computers") maintains, correctly I think, that the most intelligent animals are those (including
humans) that have highly developed social structures, and that having such structures are crucial to
the development of intelligence. These social structures are clearly implicated in the development of
languages — it is hard to see how solitary animals could possibly develop language. Whether you
can call the sorts of communications in which the more intelligent non-human animals engage
languages is a moot point — I'm inclined to say not — but they are necessary precursors. Intelligence
and social (especially linguistic) interaction seem to go hand-in-hand.

Lev Vygotsky, the Russian psychologist, in his "Thought and Language", maintained that it is in the
engagement in conversation with more experienced others that the ability to think well develops, and
there have been a raft of studies to back him up. The sort of ability to use intelligence that humans
manifest, especially but by no means exclusively in the logico-linguistic domain, seems to me to be
highly dependent on language. In fact, to tie this to your AI example, I cannot see how a robot could
possibly become intelligent in anything like the human sense (including having morals and probably
emotions) if it did not 'grow up' in a community of language users.

P.S. to Steven's comments: I believe that studies have shown that dolphins have self-recognition, and
also that they have the concept that another can have false beliefs — a characteristic hitherto only
seen in humans above the age of about 4, and some of the higher apes.

Tim Sprod