Re: Schyns Comms 01-03 abdi benson braisby

From: HARNAD Stevan (
Date: Sun May 03 1998 - 19:26:04 BST

> From: Whitehouse Chantal <>
> About the commentary on Schyns et al. by Braisby. Towards the end
> (paragraphs 6, 7 & 8) he starts talking about "pragmatic factors"
> > conceptual flexibility needs explication via constraints on
> > acceptable inductions (i.e., combinations of features)2 and we
> > suggest that this can be achieved by pragmatic factors which guide
> > the appropriate combination of fixed features.
> > the classifications supported by a fixed set of conceptual (or
> > semantic) features can be augmented by pragmatically-motivated
> > operations upon that set. Whether categorisation exploits semantic
> > .or pragmatically derived attributes will therefore depend upon the
> > task, context and the kind of object involved.
> > Pragmatic factors also explain findings that Schyns et al. take to
> > undermine fixed features.
> What exactly are these pragmatic factors?

Braisby is vague, would be my first answer. But you can think of
"pragmatic" as it is used in linguistics to distinguish "syntax,"
(formal rules for combining symbols), "semantics" (the meanings of the
symbols) and "pragmatics": the situation or context in which the
symbols are used. Or the speaker's intended meaning, as opposed to the
sentence's meaning alone (every sentence carries some assumed context;
except possibly a definition like "a bachelor is an unmarried man" --
though you can imagine that being misconstrued out of context too):

The phrase "Would you mind?", despite your knowing its syntax (it's a
question, beginning with a verb) and its semantics (we know what each
of the words it is composed of means) is not understandable unless you
know that the person who said it to you had just said "I'll be an hour
and a half late." If he had instead said: "I believe you're sitting in
my seat," then you would understand the phrase in another way.

You can think of those 3 words as being able to do two different jobs
for which we would have needed different words if it had not been for
pragmatics. Pragmatics -- i.e., the conditions under which words are
uttered, and of course what we know about those conditions -- provide
extra constraints on what something might mean. Well, in the same
sense, Braisby suggests, you can get more mileage out a fixed set of
features from the context in which they occur, and what you know about

For example, knowing that you are looking for an embedded letter cuts
down on the number of features you need too look for in a scramble of
geometric shapes.

Context is best thought of as "the context of confusable alternatives."
Things are always categorised relative to other things. The features
you need to use to sort, say, plants vs. animals, are very different
from the ones you need to sort medicinal herbs. The features need to be
the ones that do the job: that reliably separate the members form the
nonmembers. But there are always pragmatic assumptions. the only reason
stripes are a good enough way of telling apart horses and zebras is
that there are no striped horses or stripeless zebras. Otherwise you'd
have to use other features, such as face-shape.

> One final question I wanted to ask was, to get a good mark on the exam
> is it enough to have read the target article, commentaries, and
> skywriting and incorporate these with your own thoughts to produce an
> answer? Or is it necessary to read extra references (e.g. those
> referenced in the target article)?

Absolutely no need! But I do advise reading all the printed commentaries,
rather than just the ones I had in the hidden electronic directory.

Cheers, Stevan

Professor of Psychology
Director, phone: +44 1703 592582
Cognitive Sciences Centre fax: +44 1703 594597
Department of Psychology
University of Southampton
Highfield, Southampton

This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:20 GMT