Newsgroups: alt.fan.cecil-adams,alt.usage.english,sci.lang
Path: cantaloupe.srv.cs.cmu.edu!rochester!hood.cc.rochester.edu!news.acsu.buffalo.edu!dsinc!spool.mu.edu!howland.erols.net!www.nntp.primenet.com!nntp.primenet.com!netcom.com!misrael
From: misrael@scripps.edu (Mark Israel)
Subject: Re: WARNING  Popperesque Paradigm shift approaches
Message-ID: <misraelE2Dpzn.GxJ@netcom.com>
Sender: misrael@netcom14.netcom.com
Supersedes: <misraelE2Dpnv.GFC@netcom.com>
Organization: The Scripps Research Institute, La Jolla, California, USA
References: <32ADF69E.25E1@sn.no> <58mvu7$77p@riscsm.scripps.edu> <32B1CD8A.217C@scruznet.com>
Date: Sat, 14 Dec 1996 01:11:47 GMT
Lines: 63

In article <32B1CD8A.217C@scruznet.com>, darwin@scruznet.com (Mike Wright) writes:
> In article <58mvu7$77p@riscsm.scripps.edu>, misrael@scripps.edu (Mark Israel) writes:

>> I believe that humans grasp language chiefly through semantic 
>> associative retrieval, and that syntactic parsing is done only as 
>> a last resort.

   Please note that I said "only as a last resort", *not* "never".

> By "semantic associative retrieval", do you mean something like a
> combination of knowing the meaning of words and looking at context? 

   Well, not *just* words.  Our mental lexicon stores whole phrases
as well.

> Assuming that, then translating should mostly be a matter of 
> looking up definitions of unknown vocabulary.

   That is, indeed, what I spend most of my time doing when I
translate.  If the text I'm translating uses a syntactic construct
I'm not familiar with, I find I can usually still infer the meaning
from pragmatic considerations.  When I'm stumped, I don't wish I 
had a grammar-book; I wish I had a corpus of related phrases to
search.

> In practice, this doesn't seem to work very well. The problem is that
> "context" is determined, in part, by the "rules".

   The rules we apply first aren't even syntactic!  We use intonation
patterns or punctuation to decide where clauses begin and end.

> Without knowing the part played by word order in English syntax, 
> how would one distinguish the meanings of "the dog bit the man" 
> and "the man bit the dog"?

   Typically, it takes us much *longer* to grasp an uncommon
statement like "The man bit the dog" than a common statement like
"The dog bit the man."  If syntax were what we looked at first,
this wouldn't be the case.  But actually, what we access when we
see "The dog bit the man" is our knowledge about dog's biting
humans.

   How is it that we understand "Is it legal for a man to marry his
widow's sister?" as "Is it legal for a man to marry his deceased
wife's sister?"  We don't formally analyse the relationships:  we 
just let the neurons for the associated concepts fire, and grope our 
way to a plausible meaning.

> On the other hand, I understand that word order is irrelevant to the
> classical Latin versions of these sentences, since the relationship of
> each of the nouns to the verb is shown by inflection.

   It's partly inflections and partly word order.  If the subject and
object are both neuter nouns ending in -um, then one assumes SOV word
order.  Ancient Greek had freer word order than Latin.

   But that's *my* kind of syntax!  You can find all you need in a 19th-
century prescriptive grammar, and you can summarize it on a set of Key
Facts cards.  You don't need huge tomes full of starred and unstarred 
sentences, which is what linguists have spent the 20th century writing.

--
misrael@scripps.edu			Mark Israel
