Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: SUO: Re: Policy On Substitutions




I agree with Jon on the following point:

>Maybe I need to say that I mean "eliminating variables"
>in the sense of "explaining them away", in other words,
>more like "taking them off the list of basic concepts".

As I pointed out in several notes, we have to make a distinction
between the concepts that describe the domain of discourse (i.e.
the real world or some aspect of it that the SUO is trying to
address) and the concepts that are used to describe the language
that is used to talk about the world (such as KIF, CGs, SQL,
Express, Epistle, or even English).

The concept of Variable is used to describe certain languages,
but it is not necessary in all kinds.  The more primitive notion
is _coreference_, i.e. expressing the fact that two different
signs (words, symbols, images, etc.) refer to the same entity.
There are many different ways of showing coreference:

 1. Pronouns and other indexicals in English and other natural
    languages.

 2. Connecting lines (i.e. coreference links, lines of identity,
    ligatures) in various graph notations.

 3. Locally defined names, AKA variables, which were first
    introduced by Aristotle and popularized by Euclid.  They
    were never called "variables" until they were used to
    represent physical quantities in the 17th century, and in
    logic, it is incorrect to call them variables, since they
    don't actually "vary".

Bottom line:  variables don't exist except in certain kinds
of languages.  They would be discussed in the KIF standard,
but not necessarily in the SUO (except in those subparts of
the SUO that address the syntax of computer languages).

John