nl-kr-request@CS.ROCHESTER.EDU (NL-KR Moderator Brad Miller) (12/01/88)
NL-KR Digest (11/30/88 20:56:45) Volume 5 Number 31
Today's Topics:
quirky case
The History of Syntax
Re: incorporation, prediction, and explanation
Re: References On Mass Terms
Submissions: NL-KR@CS.ROCHESTER.EDU
Requests, policy: NL-KR-REQUEST@CS.ROCHESTER.EDU
----------------------------------------------------------------------
Date: Mon, 7 Nov 88 12:27 EST
From: Avery Andrews <munnari!fac.anu.oz.au!andaling@uunet.UU.NET>
Subject: quirky case
The basic reference for Complementation in Icelandic, including
the evidence that there is Raising of Quirky-case subjects to
object position, is:
Thrainsson, Hoskuldur (1979) On Complementation in Icelandic.
Garland Press.
An LFG analysis of the phenomenon is to be found in:
Andrews, Avery (1982) On the Representation of Case in Modern
Icelandic, in Bresnan (ed), The Mental Representation of
Grammatical Relations.
See also:
Maling, Joan, and Annie Zaenen (to appear) Modern Icelandic Syntax.
Academic Press.
Avery Andrews
andaling%fac.anu.oz@seismo.css.gov
------------------------------
Date: Mon, 14 Nov 88 13:51 EST
From: fisher@spch.csc.ti.com
Subject: The History of Syntax
Re: recent messages about natural language unix, Chomsky
discovering CFL, etc.:
Chomsky did not "discover" CFL.
What's now known as "CFL" is a version of what linguists have called
"immediate constituent", "IC", "constituent structure", or "phrase
structure" grammar.
IC analysis was probably done by Panini, the first linguist,
about 1200 B.C. I remember seeing a letter to the editor of CACM
once that claimed that "Backus-Naur" Form should really be called
"Backus-Naur-Panini" Form.
In modern times, IC analysis was first formulated by the linguist
Bloomfield in his book LANGUAGE (1933) and then elaborated by the
linguists Wells (1947) and Bazell (1953) [Percival p. 232].
In Bloomfield's earlier AN INTRODUCTION TO THE STUDY OF LANGUAGE
(1914), there is a clumsy attempt to state these principles of
binary splits, the hierarchical decomposition of sentences, etc.
The immediate source of Bloomfield's syntactic theory was
Wilhelm Wundt, one of the founders of Psychology and a substantial
contributor to early Linguistics. In the section of his monumental
work VOLKERPSYCHOLOGIE (1900) titled "Die Sprache", he clearly
states the basic principles of IC analysis, including a geometric
diagramming system that amounts to a binary branching tree diagram.
Wundt was aware that his conception of syntax was quite different
from his predecessors, from Priscian to Paul [Percival p. 233].
From the thirties to the fifties there was a concensus among
linguists on forms of grammar that were equivalent to IC analyis,
although quite a few linguists added their special touches,
such as Tagmemic Theory or Stratificational Grammar [Postal 1967].
Far from "discovering CFL", Chomsky's first widely known work
(Chomsky 1957) was a formalization and blast AGAINST the
prevailing linguistic model, which was IC analysis. Nor was he
the first to suggest an explicit quasi-mathematical generative
model for natural language syntax: one thing he criticizes is
Hockett's earlier (1955) suggestion that probabilistic finite
state machines might be a good model for syntax. At about the
same time, Chomsky originated the field of mathematical linguistics
with his famous article in the journal of the IRE (Chomsky 1956).
Chomsky (and others, such as Lees and Postal) argued that a new
development - syntactic transformations - was needed to overcome
the deficiencies of syntax as it was then practiced. At about
the same time, Victor Yngve (and probably others) in his work
on Mechanical Translation began using IC grammars whose nodes
were complexes of feature-value pairs. Yngve (ca. 1966) even
created a new computer language to handle this syntax - COMIT.
The arguments of Chomsky et al. won the day for transformational
grammar. After that scientific revolution, the only professional
linguists using non-transformational IC analyses were anthropologically
oriented linguists who found it convenient in the first stages
of their studies of hitherto-unstudied languages. Computer
Science people, not having to meet the stringent criteria imposed
on the natural science of Linguistics, kept regular and CFL
grammar theory alive in this winter of our discontent. And lately
IC grammars with various methods of augmentation have once again
arisen to vanguish Transformational-Generative grammar, although at
least one current linguist has been known to say "I just can't
have any respect for a grammar that doesn't include RAISING."
REFERENCES:
[Chomsky 1956] "Three Models for the Description of Language",
I.R.E. Transactions on Information Theory IT-2, No. 3 (1956).
[Chomsky 1957] SYNTACTIC STRUCTURES, by Noam Chomsky,
JANUA LINGUARUM Series Minor No. ??, Mouton & Co.,
The Hague, 1957.
[Percival] "On the Historical Source of Immediate Constituent
Analysis", by W. Keith Percival, 1967 originally, reprinted
in SYNTAX AND SEMANTICS Vol. 7, "Notes from the Linguistic
Underground", ed. James D. McCawley, Academic Press, NY, 1976.
[Postal 1964] CONSTITUENT STRUCTURE: A STUDY OF CONTEMPORARY
MODELS OF SYNTACTIC DESCRIPTION, by Paul Postal, originally
Part II of the International Journal of American Linguistics,
Vol. 30, No. 1 (1964); reprinted by Indiana U., 1967.
[Yngve ] SCIENTIFIC AMERICAN, Vol. ?, No. ?, 1966 ??
------------------------------
Date: Wed, 2 Nov 88 16:05 EST
From: Rick Wojcik <rwojcik@bcsaic.UUCP>
Subject: Re: incorporation, prediction, and explanation
Walter Rolandi writes:
WR> But if linguistics were to address the question of the circumstances under
WR> which things are said, the predictions of linguistics might be of interest to
WR> people other than linguists.
Can you clarify what you mean here, Walter? Those who study pragmatics
certainly do concern themselves with the circumstances under which things are
said. What aspects of the circumstances are ignored by linguists, and why do
you think those aspects are important? (BTW, use the term 'linguist' to refer
to linguists in general. If you are just complaining about generative
linguists, then say so.)
WR> ...behavior can be engineered by manipulating environmental events.
WR> When you are satisfied that some behaviors are determined by environmental
WR> manipulation, you might start to wonder about the causes for all types
WR> of behavior, including complex social behaviors like language. By this
WR> time, you will be ripe for Verbal Behavior.
No one doubts that behavior can be affected or controlled by manipulating the
environment. The question in my mind is what kinds of problems you expect to
be able to solve by studying verbal behavior. Do you expect to be able to
find triggers in the environment for such things as passives and cleft
sentences? Just what kinds of predictions would you like linguistic theory to
be able to make?
--
Rick Wojcik csnet: rwojcik@boeing.com
uucp: uw-beaver!ssc-vax!bcsaic!rwojcik
------------------------------
Date: Wed, 9 Nov 88 08:20 EST
From: w.rolandi <rolandi@gollum.UUCP>
Subject: Re: incorporation, prediction, and explanation (long)
In response to Rick's:
WR> But if linguistics were to address the question of the circumstances
WR> under which things are said, the predictions of linguistics might be
WR> of interest to people other than linguists.
>Can you clarify what you mean here, Walter?
Sure Rick. I think that the body of knowledge that is linguistics is
somewhat obscure and scarcely known to anyone outside of the field.
When linguists speak, mostly just linguists listen. I am suggesting that
this would not be the case if the discipline were to address questions
which, if answered, would be of interest to the general scientific community.
More specifically, I am suggesting that linguistics will take a giant
step towards science when the field embraces causal analysis and
controlled experimentation. This is the stuff of all real science. As it
now stands, linguistics seems hopelessly preoccupied with an artifact of
verbal behavior: the form of the written sentence.
>Those who study pragmatics
>certainly do concern themselves with the circumstances under which things are
>said. What aspects of the circumstances are ignored by linguists, and why do
>you think those aspects are important?
To the extent that pragmatics investigates the influences of location,
audience, and other environmental variables upon the probability of
given verbal behavior, pragmatics then, is a very good thing. Ironically,
I was originally attracted to the field because I was told that in
pragmatics, the variable of context was the major subject of investigation.
Silly me, I supposed that to mean that pragmaticists were busy producing
a body of data that defines the causal relationships between the
people, places, and objects around us that effect the contents of what
we say. What I found instead was a lot of pompous intellectualizing on
the nature of "communication". _hat I have seen is little more than
mentalistic, philosophical, even literary analysis of things like
"information", "messages", and "intentions".
What's wrong with pragmatics and linguistics in general is that neither
field addresses any issues of scientific import. Who cares about
anti-homophones, u-umlaut and y (or rather, u-double-dot and y), and
embedded, recursive prepositional phrases? When you know all about
these things, what do you know? I want to know why people say the
things that they say and why their utterances take the forms that they do.
I want a scientific answer. Is it unfair or unkind of me to ask this
of linguistics?
>(BTW, use the term 'linguist' to refer
>to linguists in general. If you are just complaining about generative
>linguists, then say so.)
I stand accused.
WR> ...behavior can be engineered by manipulating environmental events.
WR> When you are satisfied that some behaviors are determined by environmental
WR> manipulation, you might start to wonder about the causes for all types
WR> of behavior, including complex social behaviors like language. By this
WR> time, you will be ripe for Verbal Behavior.
>No one doubts that behavior can be affected or controlled by manipulating the
>environment. The question in my mind is what kinds of problems you expect to
>be able to solve by studying verbal behavior. Do you expect to be able to
>find triggers in the environment for such things as passives and cleft
>sentences? Just what kinds of predictions would you like linguistic theory to
>be able to make?
The words that come to mind in a given conversation are part of the
"environment" of verbal production. Just as the presence of a certain class
of people in one's audience will affect the things one says, so will
the presence of certain classes of words. Some words are classed by the
articles that proceed them. Some words are classed by the actions with
which they are associated. Others are classed by the effect they have
on the listener. I don't think that the study of verbal behavior will merely
uncover "triggers in the environment". (In fact, I think your statement
represents a common, push-pull mechanistic conceptualization of behavioral
theory). I think linguistics has wasted enough time illuminating
"such things as passives and cleft sentences". I want it to aspire to
substance: give me a causal explanation of verbal behavior.
Walter Rolandi
rolandi@ncrcae.Columbia.NCR.COM
NCR Advanced Systems Development, Columbia, SC
------------------------------
Date: Wed, 9 Nov 88 21:32 EST
From: Greg Lee <lee@uhccux.uhcc.hawaii.edu>
Subject: Re: incorporation, prediction, and explanation (long)
From article <165@gollum.UUCP>, by rolandi@gollum.UUCP (w.rolandi):
" In response to Rick's:
" WR> But if linguistics were to address the question of the circumstances
" WR> under which things are said, the predictions of linguistics might be
" WR> of interest to people other than linguists.
"
" >Can you clarify what you mean here, Walter?
"
" Sure Rick. I think that the body of knowledge that is linguistics is
"...
If we merely address the questions without answering them, although
we might stir up a little interest, it wouldn't last. And there isn't
any immediate prospect of finding answers for the problems Walter
is interested in. Wanting to know those things isn't enough.
I agree with Walter's diagnosis of why it is that linguistics is so
little relevant to an understanding of human behavior. The few
things linguists do understand a little about are outre matters
that are of little interest generally. It's too bad.
I don't agree with the implication(?) that there is a royal
road to understanding language which will zip us right past
this troublesome analysis of details of expresssion that the
best working linguists are caught up in. Hey, we've been that
route -- glossematics, stratificational grammar -- theories
that claimed to offer *the* answer, but turn out to be notations
for making condensed reports of the facts.
Greg, lee@uhccux.uhcc.hawaii.edu
------------------------------
Date: Thu, 10 Nov 88 06:13 EST
From: Celso Alvarez <sp299-ad@violet.berkeley.edu>
Subject: Re: incorporation, prediction, and explanation (long)
In the discussion about the status of linguistics as a science,
Walter Rolandi (in article <165@gollum.UUCP> rolandi@gollum.UUCP) has raised
the issue of causality in verbal behavior. He seems to relate the scientific
status of a discipline with its ability to describe cause-effect processes.
His core question seems to identify the description of causality with the
unveiling of an indeterminate 'reason' for the forms that verbal behavior take:
WR>I want to know why people say the things that they say and why their
WR>utterances take the forms that they do. I want a scientific answer.
WR>Is it unfair or unkind of me to ask this of linguistics?
The "why" of behavior is a psychological one, and that's not what all
disciplines looking at verbal behavior are about. It is true that in the
literature treating a number of problems which provisionally we can group under
the label of 'pragmatics' we find abundant references to categories such as
"intentions". But fundamental notions of pragmatics such as those of
'indirectness' in speech acts, illocutionary force, felicitousness, etc.,
are based on a description of the non-linguistic and linguistic
*conditions* under which given forms of speech take place. If you look
especially at the work of conversation analysts and ethnomethodologists,
their carefulness in leaving aside the psychological motivations for behavior
and in describing merely the form and organization of talk is truly so
complete that it at times becomes exasperating.
I have to admit that I will disappoint myself again by adding more noise to
this discussion. I personally would disagree that only by finding the "why"
one would find a "scientific answer". On the one hand, I agree with Clay
Bond that only by looking at cognition we'll be in the right track of
discovering the "why" of verbal behavior. On the other, I'm not so sure that
this "why" is (or should be) the primary objective of a discipline focusing on
verbal behavior. My question, on the contrary, is neither the "why"
(psychological motivations) or the "what for" (communicative intent) of
verbal interaction, but the *how*. But this *how* is not a matter of
probabilities, as Rolandi seems to suggest:
WR>To the extent that pragmatics investigates the influences of location,
WR>audience, and other environmental variables upon the probability of
WR>given verbal behavior, pragmatics then, is a very good thing...
I believe, instead, that the how is a matter of speech in action. A model
that conceives of verbal interaction as the site for the interplay of a number
of pre-defined variables ignores one of the basic characteristics of talk: the
constant renewing of communicative context. Rather than adhering to or
violating pre-established patterns (or "norms") of behavior, each time an
utterance is produced interactants put into play their resources and construct
a system of interaction in which each participant's conversational move is
partly a function of, and has an *effect* upon other participants' moves.
This is not to deny the fact that regularities exist in the ordering of talk,
but to emphasize the *dynamism* of verbal interaction. A type of causality is,
yes, inherent in the ordering of talk, and this is apparent in the seminal
notion of 'preference organization' of conversational moves, by which replies
follow questions, denials follow requests, etc. However, I would like to
clarify some points about the nature of this causality: the *effect* of
conversational moves upon others' linguistic behavior should
be seen in terms of the *generation of a given range of permissible
alternatives in speech production*. I don't think that pragmatics or
sociolinguistics can prove that I uttered a "Yes" *because* you said to me
"Would you like and ice-cream?". An interactional approach can tell us,
however, that, my utterance, due to the fact that it occupies the "second-part"
slot in an adjacency pair that started with what I recognize as a 'question'
type of speech act, is in turn to be interpreted as a 'reply' type of speech
act. I might as well have uttered "Later", and the sequential organization
of the pair would be identical.
From this perspective, however, 'reply' and 'question' are not psychologically
motivated acts, but socially constructed categories which participants
themselves are able to identify and, usually, label. The ultimate reason
(Rolandi's "why") of my choices in my reply must be looked for in my personal
motivations. But that's not what an interactional approach is about.
That is why it is also risky to assert that "the presence of a certain class of
people in one's audience will affect the things one says" (<WR).
^^^^^^^^^^^
As for the manipulation of environment in controlled experiments, this indeed
has been done in social psycholinguistic studies, and I have no reason to
doubt that their findings are sound and useful. But to me the most interesting
body of data comes from the ethnographic method. The naturalistic recording
of oral data continues to be, to me, the main most useful, and richest source
for the study of verbal behavior. I don't think that this approach would
solve problems such as the (psychological?) "triggering" of passive
constructions, nor that this is its main objective. But it does illuminate,
yes, certain aspects of the production of talk by finding correlations between
linguistic forms and non-linguistic context (e.g. passive constructions are
apparently more frequent in -- and, to a certain extent, a marker of -- more
'formal' types of discourse).
Finally, it is true that some of the literature on verbal behavior resembles
what Rolandi denounces as "(even) literary analysis" (I would say textual
analysis) in its approach and basic categories. Is this bad, though? Another
part of its analytical apparatus comes from communication theory, without which
the study of speech in communication would simply be an absurd enterprise!
Is this also bad? Rolandi thinks so, I don't.
Celso Alvarez
sp299-ad@violet.berkeley.edu
------------------------------
Date: Mon, 14 Nov 88 12:02 EST
From: Rick Wojcik <rwojcik@bcsaic.UUCP>
Subject: Re: incorporation, prediction, and explanation (long)
In article <2602@uhccux.uhcc.hawaii.edu> lee@uhccux.uhcc.hawaii.edu (Greg Lee) writes:
>I don't agree with the implication(?) that there is a royal
>road to understanding language which will zip us right past
>this troublesome analysis of details of expresssion that the
>best working linguists are caught up in...
I agree with Greg and Walter (to some extent) that linguists have a
problem with their tendency to seek explanations independently of other
sciences. The field has become quite insular. But the other side of the
coin is that others--psychologists, in particular--have sought to explain
linguistic behavior without proper grounding in linguistic theory. Such
people risk reinventing the wheel or, worse yet, the flat tire. :-)
Walter's complaint can easily become a justification for ignoring a rather
vast, diverse body of knowledge. What we need is for more Walter Rolandis
to come in and show us how to achieve superior results from the
application of his methodology.
--
Rick Wojcik csnet: rwojcik@boeing.com
uucp: uw-beaver!ssc-vax!bcsaic!rwojcik
------------------------------
Date: Mon, 14 Nov 88 20:05 EST
From: The Cat in the Hat <davidbe@sco.COM>
Subject: Re: incorporation, prediction, and explanation ([not so] long)
[ I've finally gotten around to replying to this...portions may be edited]
[ for easy digestion.]
rolandi@gollum.UUCP (Walter G. Rolandi) said:
-In response to Rick's:
-WR> But if linguistics were to address the question of the circumstances
-WR> under which things are said, the predictions of linguistics might be
-WR> of interest to people other than linguists.
-
->Can you clarify what you mean here, Walter?
-
-Sure Rick. I think that the body of knowledge that is linguistics is
-somewhat obscure and scarcely known to anyone outside of the field.
-When linguists speak, mostly just linguists listen.
I agree with this part to some extent. The same can be said of most
disciplines, though. Linguistics is just one example.
- I am suggesting that
-this would not be the case if the discipline were to address questions
-which, if answered, would be of interest to the general scientific community.
This I do not agree with. Linguistics would be less obscure and more widely
known if it were to address questions and issues of interest to the
general public. Not exclusively, mind you, but still of some interest.
-
-I want to know why people say the
-things that they say and why their utterances take the forms that they do.
I'd like to know that too.
-I want a scientific answer. Is it unfair or unkind of me to ask this
-of linguistics?
I don't care about a strict scientific answer. You're looking more at
(if there is such a thing) psycholinguistics; a study of WHY we say what
we mean, rather than HOW we say it.
What I'm looking for (IMHO) is a way of bringing items like word meanings,
language change, grammars, and much of the body of linguistics, down to
a level above that of the general public, but not so high that it's out
of reach. Because linguistics deals with something all people have
experience with (language) it should be less difficult to condense and
simplify the "interesting and fascinating" parts of it. Mathematics is an
example. The basics of fractals can be easily explained, but a large
amount of background is necessary to begin to realize HOW fractals can be
produced and utilized. This shouldn't be necessary with linguistics.
All of the above is personal, unstudied opinion. I may be completely wrong.
Linguistics may be as unfathomable a field as much-higher mathematics.
In which case I stand corrected. (What is the origin of that phrase anyway?)
* * * * * * *
On a slightly related topic...is this the newsgroup to ask about word
and phrase origins? Or to talk about slang, or modern language, or about
how advertising is introducing new words into the English language (or
American idiom)?
Or would something like {talk,alt}.words be more appropriate?
--
David Bedno (aka The Cat in the Hat) Now appearing at: davidbe@sco.COM -OR-
....!{uunet,decvax!microsoft,ucbvax!ucscc}!sco!davidbe -OR-
At home: 408-425-5266 At work: 408-425-7222 x5123 (I'm probably here...)
Disclaimer: Not SCO's opinions. At least not that they've told me.
Complete Romantic Cynicism: The only healthy response to today's society.
------------------------------
Date: Sun, 20 Nov 88 10:03 EST
From: w.rolandi <rolandi@gollum.UUCP>
Subject: Re: incorporation, prediction, and explanation (long)
In response to Rick's:
>>Such environmental variables can be operationalized and quantified.
>>The behaviors in question can be understood as a function of covariation
>>with those variables.
>Would you please explain what it is that you think you'll find or predict with
>such an approach?
The probability and function of a given utterance or class of utterances.
>Will you be able to account for the difference between
>active and passive sentences by studying the environmental variables
>surrounding speech acts?
Possibly but I doubt surrounding speech acts alone control active/passive
constructions. These are probably a function of audience. (See Skinner
on multiple causation).
>What about when someone is writing a letter, and he
>switches from active to passive sentence structure?
Again, audience is the most likely causal variable. One would have to
observe the behavioral history of the speaker/writer when either employing
or seeing/hearing employed active vs. passive constructs. What you are
asking is what I would call a good experimental question. Shall we go
for the NSF grant together? 8-).
>Do you really expect to
>find some environmental correlate to explain the behavior? Good luck.
Yes, I do. But let me remind you that all parts of a sentence make up
the speech production environment. (See Skinner on tacting and autoclitics)
Walter Rolandi
rolandi@ncrcae.Columbia.NCR.COM
NCR Advanced Systems Development, Columbia, SC
------------------------------
Date: Thu, 3 Nov 88 12:56 EST
From: Graeme Hirst <gh@ai.toronto.edu>
Subject: Re: References On Mass Terms
>I am doing research on knowledge acquisition from NL text. I am in
>need of references on MASS TERMS. If anyone has any references they
>would be most helpful.
The work of Harry Bunt, published in his book "Mass terms and
model-theoretic semantics (Cambridge UP), would be a start. There is
also work by Francis Jeffry Pelletier, University of Alberta. He is a
philosopher interested in NL issues; try looking him up in the
appropriate indexes.
\\\\ Graeme Hirst University of Toronto Computer Science Department
//// uunet!utai!gh / gh@ai.toronto.edu / 416-978-8747
------------------------------
Date: Mon, 7 Nov 88 14:06 EST
From: Halvorsen.pa@Xerox.COM
Subject: Re: References On Mass Terms
An interesting and detailed analysis of mass terms can be found in Jan Tore
Loenning's "Mass Terms and Quantification", pp 1-52, Linguistics and
Philosophy, Vol 10, No. 1, February 1987. He has done more on mass terms
and on the semantics of plural, and you can try to reach him at:
m_loenning_j%use.uio.uninett@TOR.nta.no, or
m_loenning_j@use.uio.uninett.
--Per-Kristian
------------------------------
End of NL-KR Digest
*******************