LINGUIST List 9.1712

Fri Dec 4 1998

Disc: Morphosyntax

Editor for this issue: Jody Huellmantel <jodylinguistlist.org>

Directory

1. Larry Trask, Re: 9.1692, Disc: Morphosyntax
2. Paul Llido, Re: 9.1692, Disc: Morphosyntax

Message 1: Re: 9.1692, Disc: Morphosyntax

On Tue, 1 Dec 1998, Martin Haspelmath wrote (LINGUIST 9.1692):

" But if the criteria for defining nouns, verbs, etc. in English are
 different from the criteria used for Latin, then how do we know that
 English Nouns and Verbs have anything to do with Latin Nouns and Verbs
 (I use capitalization to indicate language-particular categories)?
 Obviously, the only answer is that they express similar notions. The
 conclusion is that THERE ARE NO universal syntactic categories. What
 is universal is the broad pragmatic functions and conceptual
 distinctions expressed by language, as well as distributional patterns
 expressed by implicational universals."



There are two separate questions here. The broad one is this: are
there any universal syntactic categories? The narrow one is this: do
English and Latin share any syntactic categories? I propose to look
first at the narrow one.

Using criteria appropriate to each language, we can set up an English
word-class containing words like `man' and `house', and a Latin
word-class containing words like <vir> and <domus>. We then find two
things: first, the two classes contain words expressing many of the
same meanings, and, second, both classes contain most of the words
which denote classes of physical entities, like my examples. In
apparent contrast to Martin, I think this is already good evidence for
identifying the two as the class of nouns.

But there's more. In both languages, we can identify phrases and
heads, though again by somewhat different criteria. We then find that
words like English `man' and words like Latin <vir> typically function
as the heads of phrases. And, in both cases, the resulting phrases
perform similar functions. For example, both can function as
subjects, where subjects are identified by Keenan-style criteria, such
as verbal agreement, sentence position, and ability to undergo
raising. We therefore have further evidence that both languages
possess a class of noun phrases, and that the heads of noun phrases in
both languages are nouns.

I guess I would agree with Martin that none of this constitutes
out-and-out proof that the class of English nouns must be identified
with the class of Latin nouns. But I do think it constitutes a very
powerful case that this is so: no contrary identification appears to
be at all plausible, and hence refusing to make this one seems
perverse.

But does it follow that the class of nouns is universal? No, it does
not, of course: this is an empirical question which must be pursued
language by language. In fact, it appears that nouns and verbs are
the only word-classes that anybody wants to defend universal status
for. The universality of nouns and verbs was affirmed by Sapir,
queried by Whorf, and denied by Hockett, mainly on the basis of the
Wakashan and Salishan languages of the Pacific Northwest of North
America. But further work on these languages, perhaps most notably by
Bill Jacobsen, has called Hockett's interpretation severely into
question. And Paul Schachter, in his article in the 1985 Shopen
volumes entitled _Language Typology and Syntactic Description_,
asserts firmly that, on the basis of the evidence currently available,
we may safely conclude that recognizable and distinguishable classes
of nouns and verbs do indeed appear in all spoken languages. (I have
seen no work on sign languages in this connection.)

Of course, I haven't read the Croft book referred to elsewhere by
Martin, and maybe I'll be persuaded otherwise when I do read it. But,
for the moment, I'm happy to go along with Schachter.

Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larrytcogs.susx.ac.uk

Message 2: Re: 9.1692, Disc: Morphosyntax

***********************************************************************
**************************************************** Paul C. LLIDO *
******************************* e-mail: llidopgusun.georgetown.edu *
**** Georgetown University (Graduate School - Dept. of Linguistics) *
***********************************************************************


 Paul Llido writes (LINGUIST 9.1664):
 
 "The terms, "nomen", "verbum", "adverbum", etc. definitely were a
 legacy of latin grammarians. Their use in English as primitives needs
 to be defined distributionally."


 Martin Haspelmath replies (LINGUIST 9.1692):
 
 " But if the criteria for defining nouns, verbs, etc. in English are
 different from the criteria used for Latin, then how do we know that
 English Nouns and Verbs have anything to do with Latin Nouns and Verbs
 (I use capitalization to indicate language-particular categories)?
 Obviously, the only answer is that they express similar notions. The
 conclusion is that THERE ARE NO universal syntactic categories. What
 is universal is the broad pragmatic functions and conceptual
 distinctions expressed by language, as well as distributional patterns
 expressed by implicational universals."

It is my intuition too that there would be no universal syntactic
categories as defined in terms of morphological comparative
distribution. In practice, I would use a different terminology taken
from the language itself for the syntactic categories. For me,
universals (not pure concepts) lie in the field of ontology and
speakers of a language in a linguistic culture through time reflect
these in their language: e.g. ontological_substance -
grammatical_substance, ontological_space - grammatical_space,
ontological_gender - grammatical_gender, ontological_time -
grammatical time, etc, etc. There too are psycho-social communicative
relationships between people and these too are reflected in pragmatic
functions. But I would maintain that there are no universal
distributional patterns but distributional patterns common to
languages within a certain family of languages and across certain
related families: headedness in morphosyntax whether left (IE) or
right (Japanese) or top-to-bottom (Chinese).

 "The same reasoning applies to all other morphosyntactic categories, as
 has been argued persuasively by Bill Croft in recent work (see also
 his 1991 book "Syntactic categories and grammatical relations", U of
 Chiacgo Press)."

Good... This would bolster my suggestion that grammatical time in English
is reflected differently from grammatical time in Latin given the
morphosyntactic evidence. But I'd like to go back to the discussion on
[house] as noun and [house] as verb.

 "Since all languages perform basically the same tasks, there are a lot
 of similarities and many universals, but there are no universal
 morphosyntactic features."

...and lots of dissimilarities too.