LINGUIST List 9.1667

Mon Nov 23 1998

Qs: Compounds, Technical Corpus Analysis, Armenian

Editor for this issue: Karen Milligan <karenlinguistlist.org>


We'd like to remind readers that the responses to queries are usually best posted to the individual asking the question. That individual is then strongly encouraged to post a summary to the list. This policy was instituted to help control the huge volume of mail on LINGUIST; so we would appreciate your cooperating with it whenever it seems appropriate.

Directory

  1. Andreas Gather, Romance and German Verb + Noun Compounds
  2. alejandro curado fuentes, Technical Corpus Analysis
  3. David Harris, Availability of a General Amerind Reference

Message 1: Romance and German Verb + Noun Compounds

Date: Sun, 04 Oct 1998 23:14:16 -0700
From: Andreas Gather <ac.gathert-online.de>
Subject: Romance and German Verb + Noun Compounds

Dear Linguists,

I'm working on the morphological structure of Romance Verb + Noun
Compounds (henceforth VNCs) of the type fr. 'ouvre-bote', it.
'asciugamano', span. 'abrelatas', port. 'limpavidros', cat.
'obrellaunes' etc. This type of compounding seems to be fully
productive in all Romance languages with the exception of Roumanian.

(a) Clearly, in the overwhelming majority of the cases, the noun has
to be analyzed as satisfying the theme-role of the verbal constituent
or, to put it differently, as functioning as the direct object of the
verb. The semantic notion of theme and the syntactic notion of direct
object overlap to a very large extent, but perhaps are not fully
co-extensive (a point not directly relevant to my queries). What
complicates matters, is the fact that there are some VNCs which do not
fit the outlined pattern and where the nominal constituent does not
function as the direct object/theme of the verb. What I have in mind
here, are not so much the well-known and highly lexicalized cases as
it. 'marciapiede', 'batticuore' or (perhaps) fr. 'croque-monsieur',
where the noun seems to function as the subject of the verb, or
compounds like span. 'girasol'/fr. 'tourne-soleil'/it. 'girasole',
where the noun likewise does not bear a theme-role, but more recent
and fully transparent formations reported in the relevant literature,
as, for instance, fr. 'garde-boue' ('something that protects someone
from splashes'), span. 'protege-esquinas' ('something that protects
children against (sharp) edges' and NOT 'something that protects
edges') or fr. 'protge-pluie' (in the sense of 'umbrella',
i.e. 'something that protects from rain'). In the light of such
examples and the well-known partly or fully lexicalised cases, some
linguists, especially in the German speaking area, have argued for a
general word formation rule, covering all these cases and leaving the
exact relation between verb and noun underspecified and to be inferred
from pragmatics and language use. 

To my mind, this is not a viable solution because it does not allow
for an explanation of the fact that the theme/direct object examples
by far outnumber the other cases and that especially interpretations
of the noun constituent as the external argument are completely
rejected by native speakers (at least in neologisms and cases where
the semantics is still truely compositional). I think that cases where
the noun does not bear the theme-role should better be thought of as
belonging to the periphery of a more restricted word formation rule,
which only has to handle the core cases. But perhaps those cases that
I term 'peripheral' are more widespread than I think. The corpora of
examples that I studied and reactions of native speakers that I
tested, however, do not point in such a direction. If, nevertheless,
someone has reasons to think that I underestimate the role of
"non-canonical" examples, I would be glad if she/he would let me know.

(b) In some way related to the above problem is the question of
whether VNCs are subject to genuine morphological principles or
whether they can (and should) be handled by syntactic principles. The
striking correspondence between the basic Romance word order in the
syntax and the sequencing of verb + object in the VNCs has always been
noticed and often been alluded to in the context of the emergence of
these compounds in early Romance or late Latin, while they are
practically absent from classical Latin with a supposedly different
basic sentence pattern. Many of the Romance VNCs correspond to German
or English synthetic compounds of the type 'truck driver', 'thirst
quencher' and 'skyscraper'. At least for German, the hypothesis that
these compounds reflect syntactic structure can be upheld; the English
case is much more controversial (as can be seen by the syntactic
movement analysis proposed by Lieber (1992) and some of the reactions
it provoked). But note that even in German and English there are at
least superficially similar compounds like 'chain smoker' oder
'daydreamer', which do not correspond to the general pattern with the
non-head satisfying an argument of the (de)verbal head and which do
not reflect syntactic relations. In the relevant literature, these
compounds, however, are (as far as I know) treated separately from the
more common and semantically predictable cases and are not subsumed
under a more strictly defined notion of 'synthetic compound'.

I would appreciate very much if someone could give me information
about languages that have similar types of VN- or NV-compounding and
would let me know if these compounds are in line with or diverge from
the basic (underlying) word order in those languages. Moreover, as I
see things, a strict syntactic analysis of Romance VNCs could only be
defended if cases as the above mentioned 'protege-esquinas', which do
not mirror syntactic structure, are indeed to be analysed as
peripheral. (I leave aside here differences between truely syntactic
VPs and VNCs as far as the presence/absence of functional categories
is concerned; cf. 'ouvre-bote', but 'ouvrir une/la/cette
bote'/*'ouvrir bote'. Such differences may be explainable by
independently motivated mechanisms.)

(c) My third question concerns German VN compounds of the type
'Leuchtreklame' (lit.: 'shine sign'; 'neon sign'); 'Stechmcke' (lit.:
'sting mosquito'; 'mosquito'), 'Kriechtier'(lit.: 'creep animal'),
'Mietwagen' (lit.: 'hire car'), 'Heilpflanze' (lit.: 'heal plant';
'medicinal plant'), 'Wibegierde' (lit. 'know desire'; 'thirst for
knowledge'), 'Lesebuch' (lit.: 'read book'; 'reader'), 'Brechreiz'
(lit.: 'vomit irritation'; 'nausea'), 'Waschstrasse' (lit.: 'wash
street'; 'car-wash'), 'Strfaktor' (lit.: 'trouble factor';
'disruptive factor'), 'Putzlappen' (lit.: 'clean cloth'; 'cloth'),
'Lehrbuch' (lit.: 'teach book'; 'textbook'), 'Sprechanlage' (lit.:
'speak device'; 'intercom'), 'Singvogel' (lit.: 'sing bird';
'song-bird') etc. Similar compounds exist in English ('rattlesnake',
'playboy', 'callgirl' etc.), but are less widepread and obviously less
productive than in German. The German (and English) VNCs differ
fundamentally from the Romance VNCs. The latter are (at least at the
phonetic surface and according to some researchers morphologically as
well) exocentric, while the former are always endocentric (with the
head on the right). Furthermore, in the German compounds, the actual
relation between verb und noun is scarcely predictable on purely
linguistic grounds - a fact that seems to offer evidence for a truely
morphological analysis and to militiate against a syntactic
treatment. The inherent indeterminacy of the VN-relation in the German
VNCs can easily be seen from the examples given above, although in
many cases the nominal constituent, which is the head of the
construction, obviously satisfies the external (subject) argument of
the verbal (non-head) constituent - a pattern categorically excluded
in the Romance VNCs except for a few lexicalized and more or less
opaque cases, where the noun functions as the subject argument of the
verb (see above). The fact that German VNCs never qualify as
translations of Romance VNCs might be due to the fundamental
differences in headship assignment.

Cases of German VNCs where the noun functions as the theme of the verb
are possible (cf. 'Trinkmilch' (lit.: 'drink milk'; 'milk' (as opposed
to 'evaporated milk')), 'Trinkwasser' (lit.: 'drink water'; 'drinking
water')), although they do not seem to constitute the bulk of the
cases. Since I'm specialised in Romance linguistics and (though a
native speaker of German) I'm not too familiar with linguistic studies
of German, I would be glad if someone could point out to me if there
are in depth-analyses of the German VNCs (I only know of
Boase-Beier/Toman (1986) and some remarks on these compounds in
Toman's contribution to the Handbook of Morphological
Theory). Moreover, I don't know if there already exists a comparative
study of German and Romance VNCs.

In case I receive enough responses, I will, of course, write a summary.

Many thanks in advance

Andreas Gather
andreas.gatherruhr-uni-bochum.de

Address:
Dr. Andreas Gather
Ruhr-Universitaet Bochum
Romanisches Seminar
GB 8/133
Universitaetsstr. 150
44780 Bochum
Germany
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: Technical Corpus Analysis

Date: Mon, 23 Nov 1998 21:27:22 -0800
From: alejandro curado fuentes <alexcurdunex.es>
Subject: Technical Corpus Analysis

Does any one know about recent research on specialized (i.e. science and
technology) lexis based on corpus analysis?
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 3: Availability of a General Amerind Reference

Date: Mon, 23 Nov 1998 18:50:39 -0500
From: David Harris <dharrislas-inc.com>
Subject: Availability of a General Amerind Reference

I have been looking in vain for something resembling Comrie's _The World's
Major Languages_ which contains similar analyses of a wide variety of
Amerind languages, a group that is missing from Comrie's book. I would be
interested in finding a similar analysis of Armenian. Thanks,

Dave Harris
Washington, DC
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue