richw@ada-uts (11/03/86)
I found a passage in Stoy's, "Denotational Semantics: The Scott-Strachey
Approach to Programming Language Theory" (pp.8-9) which I thought others
might find amusing: (Rich Wagner)
Incidentally, the fuzziness of the boundary [between syntax and
semantics] is much worse when dealing with natural languages.
There are many situations (for example, "Time flies like an
arrow", compared with "Fruit flies like a banana") where semantic
questions (the existence or otherwise of fruit flies and time
flies) affect the parse. In other examples (e.g. "Our mothers
bore us") the probable characteristics of the objects described
affect the syntax. The situation is much easier with artificial
languages -- we design them ourselves, so we can take care to
avoid such horrors. This may indicate that "language" is the
wrong word to use for the objects of our study, and that perhaps
the word "notation" would give a more accurate impression of
what we are about: we do not normally talk about "the language
of tensors", or "Leibnitz's language for the integral calculus".
But we are bedevilled by over-inflated jargon in computing
(usually implying unwarranted anthropomorphisation), and we must
learn to live with it.