[Lexicog] circular definitions

[Patrick Hanks]

I'm sorry to say that Rudy Troike's question about circular definitions and
subsequent interventions (in particular by Bill Poser, mentioning Anna
Wierzbicka) have got me thinking again.  This lexicographylist really does
eat up one's weekend!

There are many things that, as a monolingual lexicographer, I admire about
Wierzbicka. Her work on semantic primitives is not among them. The only
thing that is admirable about it is that, unlike so many philosophers of
language who talk vaguely and suggestively about "semantic primitives", W.
actually lists explicitly her set of semantic primitives, so we can evaluate
them for what they are.  Wierzbicka's dictionary of English Speech Act
Verbs, despite the bizarre style of its definitions (in a language
approximating primitive-ese), is full of thought-provoking linguistic
insights, especially in the footnotes to each entry, which are written in
English, not primitive-ese. In practice, the primitives don't work in
definitions for humans; Wierzbicka acknowledges that (in order to be
comprehensible),she had to make some compromises by using ordinary
English words as well as primitives in her definitions; and I'm told by
computational linguists that her primitives are unusable for computing
purposes too.

It is very debatable whether "semantic primitives" really exist. An
alternative view is that "each word is its own primitive". Primitives are
generally taken as implying some sort of ontology arranged in a a semantic
hierarchy, with a nice little set of primitives at the top and all the words
arranged in tree structures derived from them. But in fact, the best
evidence from corpus analysis is that semantic relations among words are
mostly shadowy, vague, and partial (not neat and hierarchical).  Yes, I know
a canary is a bird is a creature is a physical entity....  But that idea can
be, and has been, overextended.  Pace WordNet, is it really true to say that
an idea is a concept?  Perhaps it is equally true to say that a concept is
an idea. Or both. Or neither.

If I may be provocative for a moment, let's ask ourselves what is actually
wrong with "circular definitions".  Admittedly, defining "bull" as a male
cow and "cow" as a female bull is logically offensive, and not much help if
you don't know either term. But are dictionaries written for logicians? If
you an ordinary dictionary user and you know one term but not the other
(unlikely, I know -- but no less unlikely than the postulated dictionary
user who does not know either term), and if you know "male" and "female",
then the definition is about as good as it gets.

In actual fact, the objection is to reciprocity rather than circularity.
Insofar as a monolingual dictionary aspires to record the inventory of the
language, it could be said that all definitions are circular (i.e. they all
define words in terms of other words, which are in the dictionary -- except
those that escape into some other language -- New Latin, for example:
"thrush" defined as Turdus turdus).  From the point of view of ordinary
dictionary users looking up an unfamiliar word, which type of definition is
more useful?  The "Turdus turdus" type, or the (circular)  "male cow" type?

I think the relationship between dictionary definitions and logic needs a
lot more careful thought than it has had up to now.  Dictionary writers have
been far too easily caught in the mousetrap of formal logic. Could it be
that a useful rule of thumb, which might be paraphrased as "try to explain
rare words in terms of more common ones" -- has been hijacked and
overextended by logicians or, worse still, wannabe logicians? Knee-jerk
kowtowing to logic leads to -- has led to -- some really awful definitions
(definitions which fail to explain), of a kind only too familiar to users of
traditional monolingual dictionaries.


Patrick Hanks
Berlin-Brandenburg Academy of Sciences
and Brandeis University

----- Original Message -----
From: "William J Poser" <billposer at alum.mit.edu>
To: <lexicographylist at yahoogroups.com>
Sent: Saturday, March 06, 2004 2:21 AM
Subject: [Lexicog] circular definitions


>
> Someone who has written at some length about circular definitions and
> how to avoid them, and generally about the "language of definition"
> and semantic primitives and their role in lexicography is Anna Wierzbicka
> at the Australian National University (home page:
http://arts.anu.edu.au/linguistics/arts_fac/staff/wierzbicka.html). Her
publications include:
>
> 1972. Semantic Primitives. (Frankfurt: AthenC$um).
>
> 1980. Lingua Mentalis. (Sydney: Academic Press).
>
> 1985. Lexicography and Conceptual Analysis. (Ann Arbor: Karoma).
>
> 1988. The Semantics of Grammar. (Amsterdam: John Benjamins).
>
> 1991. Cross-cultural Pragmatics. (Berlin: Mouton De Gruyter).
>
> 1992. Semantics, Culture and Cognition. (Oxford: OUP).
>
> 1996a. Semantics, Primes and Universals. (Oxford: OUP).
>
> 1996b. Understanding Cultures through their Key Words. (Oxford: OUP).
>
> I have read some of her work and found it interesting. I don't know
> how much impact she has had on practical lexicography. My feeling is
> that she has some good ideas but that it would be difficult to
> implement them fully and live up to her standard, particularly
> when creating a bilingual dictionary of a language one is still
> struggling to understand.
>
> Bill
>
> --
> Bill Poser, Linguistics, University of Pennsylvania
> http://www.ling.upenn.edu/~wjposer/ billposer at alum.mit.edu
>
>
>
> Yahoo! Groups Links
>
>
>
>
>




Yahoo! Groups Links

<*> To visit your group on the web, go to:
     http://groups.yahoo.com/group/lexicographylist/

<*> To unsubscribe from this group, send an email to:
     lexicographylist-unsubscribe at yahoogroups.com

<*> Your use of Yahoo! Groups is subject to:
     http://docs.yahoo.com/info/terms/