Newsgroups: comp.ai.philosophy
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!news.sprintlink.net!news.dorsai.org!news.ilx.com!psinntp!psinntp!psinntp!scylla!daryl
From: daryl@oracorp.com (Daryl McCullough)
Subject: What's Grammar, Anyway?
Message-ID: <1995Feb10.041713.23514@oracorp.com>
Organization: Odyssey Research Associates, Inc.
Date: Fri, 10 Feb 1995 04:17:13 GMT
Lines: 78

There has been extended discussion in this group about grammar, about
whether it is innate, or whether natural language actually has a
grammar, or not. But what exactly is grammar, and is it worth worrying
about? I know that for a long time I thought that grammar meant things
like worrying about dangling participles and split infinitives---rules
that existed only to give English teachers something to complain about
when your essays were otherwise absolute gems. But I've come to
appreciate grammar a lot more in the last ten years or so, largely
because of a wonderful book, _The Grammatical Man_, by Jeremy
Campbell. It's been years since I've read it, and so my memory might
be playing tricks on me, but here is how I remember Campbell's
description of what grammar is.

The primordial purpose of a grammar is the transmission of information
in the presence of noise. To understand this, try to imagine a
language that developed completely without grammar. If you were a
Cro-Magnon, there might only a relatively small number of things that
you would ever want to say: "I'm hungry", or "Let's mate". Why not
just have one word for each different thought? Who needs sentences
anyway? Well, there are problems. There's always the background noise
of the roaring mastodons, and since Oog lost all of his teeth, his
pronunciation has left something to be desired. If each word expresses
a different complete thought, and you mishear a word, then you will
get nothing at all. You might start fixing dinner when Oog meant to
tell you to run from the saber-toothed tiger.

The primary purpose of a grammar is to build error-correction into
language. If you mishear a word, you can still often tell by its
location in a sentence what part of speech it is, say a noun, and
often the context is enough to figure out what the word was supposed
to be. Maybe you might *not* figure out the word at all (or maybe
you didn't mishear it, it was simply a word you didn't recognize).
Still, if the speaker was following some grammatical rules, much of
the meaning is recoverable even in the presence of uncertainty.
A great example of this is the Lewis Carrol poem, Jabberwocky, which
starts off "T'was brillig, and the slithy toves did gyre and gimbal in
the wabe..." In spite of the fact that *none* of the content words
are recognizable, we can guess at possible meanings, because the poem
adheres to familiar English grammar.

The error-correcting ability of grammar has usefulness beyond human
communication. Campbell points out that DNA itself is a language,
describing how to construct a creature. In the presence of mutations,
DNA has the same problem with noise as human language does. How do you
prevent a single, small change from completely changing the message
(so that a dog's baby turns out to be an octopus, instead of a puppy)?
The answer, again is grammar. There is a grammar for mammalian DNA,
and this grammar specifies that a mammal must have certain
features--one head, not two, two eyes, not four, etc. Noise in the
transmission can make changes to parameters (how big the head is, how
far apart the eyes are, etc.) but the grammar builds in enough
redundancy that there is almost no chance for a single mutation to
completely violate the basic mammalian design.

The secondary purpose of grammar is to give a structured, controlled
way to explore new possibilities. Once you've got the basic plan for a
sentence down, with nouns, adjectives, and verbs, etc., it is
relatively easy to extend the language---just add new words of the
appropriate type. The grammar constrains the form of what is said, but
it allows infinite variation in the content. In the same way, the
basic mammalian plan coded in the grammar for DNA specifies in broad
outlines what it means to be a mammal, but within that plan, it is
possible to have bats, and horses, and dolphins. This secondary purpose
is in some sense a lucky side-effect of the first.

So, grammar makes language robust and hardy, yet it allows (and
supports and encourages) the exploration of brand new statements never
before uttered.

Daryl McCullough
ORA Corp.
Ithaca, NY






