Re: What is the definition of "definition"?

Eliezer Yudkowsky (
Sun, 02 Feb 1997 01:52:23 -0600

[John K. Clark:]
> Well that's a real nice start Eliezer, but you didn't complete it. Talking
> about "a symbol" won't do, I want you to explain "the symbol", after all,
> I didn't ask a general question but a very specific one. So please explain
> all the information necessary for the formation of the symbol "definition".

I think it's a bit misleading, quoting only my quickie definition:
> >Definition (n): An explanation intended to convey all information
> >necessary for the formation of a symbol.

You should at least note that this was the summary only and that I also
included an explanation of all the cognitive objects associated with
"explanation", "information", "formation", and "symbol". This
explanation was in turn the information necessary for the formation of
the symbol "definition", as you say.

I could also point to the subjective aspect of watching something being
defined, and that would also be sufficient for defining the symbol
"definition". It would not, however, reduce the term "definition" to
its causal constituents but instead ask you to accept the experiential
aspect of definition as a conceptual primitive, which I deemed an unfair
"out" - even though definition actually *is* a conceptual primitive, we
can still attach a scientific explanation to our experience of that

In other words, pointing to the subjective aspect would indeed define
the symbol of "definition" and give it a referent, but not the
*explanatory* (rather than experiential) referent you were looking for.

So my answer is, I did explain all the information necessary for the
formation of the symbol "definition". I gave a detailed description of
all the cognitive events associated with defining something. Of
*course* the quickie definition is circular: It uses the phrase
"formation of a symbol", which unless defined further, simply means:
"defining a symbol". The only way to break this circularity is to point
to the cognitive processes involved in symbol formation, which can be
defined (presumably) in terms of basic Turing processes. Now, *those*
you can't define except circularly.

In any case, it strikes me as being a straw-man switch to call the
summary "circular" and ignore the three paragraphs of cognitive science.

--       Eliezer S. Yudkowsky

Disclaimer:  Unless otherwise specified, I'm not telling you
everything I think I know.