query: associative plurals via noun-verb disagreement

[David Gil]

Dear all,

I am greatly appreciative of all the examples that you have and are 
continuing to send in; once the dust settles (it doesn't seem to have 
yet) I will post a summary. But in the meantime, some terminological 
issues have arisen which, I think, need addressing ...

Grev Corbett complains that calling the construction in question 
"associative plural" is misleading because "the associative is 
orthogonal to number"; in another message addressed just to me he likens 
the term "associative plural" to "dative plural", which drives home the 
independence of the two concepts in question. I agree with Grev that the 
concepts of associative and number are orthogonal; in fact, in a paper I 
gave many years ago (alas, never published), I argued exactly this, 
citing as evidence cases where the associative plural is expressed as a 
transparent combination of associative plus plural. However, in many 
other cases there is a dedicated associative plural marker, just as some 
languages may have a specific dative plural marker, thereby fusing the 
expression of the two independent concepts. And in such cases, I think 
terms such as "dative plural" or "associative plural" are quite 
appropriate, in that they express both the independence of the concepts 
but also their formal combination.

A couple of people, including Nick Evans in the general LINGTYP 
discussion, suggested that what I was looking for was what has been 
called the "inclusory construction", by Ruth Singer and others. No it 
isn't. Associative plurals, including the particular subtype that I am 
after, overlap with inclusories, but the two are essentially different 
kinds of beasts. (The following remarks are essentially an amplification 
of Misha Daniel's latest message, which makes roughly the same point.)

The defining feature of an inclusory construction is that it contains 
two referring elements, one of which wholly contains the other in its 
reference (hence "inclusory"); prototypical examples are constructions 
such as the following from Riau Indonesian (where /kami/ wholly includes 
/John/ in its reference):

(1) Kami sama John
1PL:EXCL together John
'John and I/us'

In contrast, the defining feature of an associative plural is that its 
plural reference is constituted by a single member of the plural set, to 
which all other members of the set are associated, as in the Japanese:

(2) John-tati
John-PL
'John and his associates'

Clearly (1) and (2) are different construction types: (2) is not an 
inclusory, and a moment's reflection will reveal that (1) is not an 
associative plural: the reference of the pronoun /kami/ is determined by 
previous linguistic and/or extralinguistic context, rather than being 
constituted by /John/ -- sentence (1) does *not *mean 'John and his 
associates, including the speaker'.

Now at least some of the examples that have been offered to me in the 
discussion of the last couple of days are clearly of the inclusory type, 
as in (1), and not of the associative type, as in (2); in particular, 
this seems to be the case for all the examples cited by Ruth Singer that 
I have been able to access from assorted downloaded talk handouts. (I 
must confess that I have not yet had the opportunity to check Nick 
Evans' Bininj Gun-wok grammar or Sasha Aikhenvald's Manambu, as both are 
currently checked out of our library!)

But although inclusories and associatives are different things, they may 
partially overlap. Consider the following example from Papuan Malay:

(3) John dong
John 3PL
'John and his associates'

Constructions such as (3) exemplify one of the most common subtypes of 
associative plurals, formed with a 3rd person plural pronoun. The 
semantics is clearly associative; it means the same thing as (2). But is 
it also an inclusory? Well, the answer to that hinges on whether the 
reference of the 3rd person plural pronoun /dong/ includes /John/, and 
this is a question that is remarkably difficult to answer. Asking native 
speakers doesn't help: I tried, and their answers were inconsistent and 
inconclusive. Analytically, the grammar of Papuan Malay allows (3) to be 
analyzed either as a zero-marked coordination 'John and them', or as a 
zero-marked prenominal genitive 'John's them'. Under the coordination 
analysis it isn't an inclusory, but under the genitive analysis it is. 
(In reality I would rather opt for an analysis that was indeterminate 
with respect to the coordination/genitive dichotomy, but that's another 
story.) So for the time being, the interim conclusion is that 
constructions like (3) are clear associative plurals, while their 
membership in the class of inclusory constructions is something that 
needs to be sorted out by further work.

In principle, the same is also true for the similar constructions in 
Roon that got this discussion started:

(4) John si-(berif)
John 3PL-(laugh)
'John and his associates (are laughing)'

What is clear is that this is an associative plural construction. What 
is less clear is whether this is also an inclusory construction. For it 
to be one, one would have to establish the following two points: (a) 
that the agreement marker /si-/ has independent reference, and (b) that 
its reference includes /John/. At present I don't know how to adjudicate 
these questions. With regard to the former point, it is indeed true for 
Roon that the verbal agreement markers are almost identical in form to 
the independent pronouns, so a case could conceivably be made that they 
retain their referential properties even when occurring as prefixes. 
(But this case would harder to make for other languages, such as the 
Russian dialect and the Maltese that Grev cites in his book on Number, 
where the agreement morphology is formally unrelated to the independent 
pronouns.) But with regard to the latter point, the same issues arise 
for Roon and its likes as do for Papuan Malay and its likes where the 
construction is periphrastic, and, again, I know of no straightforward 
way of answering them.

So to conclude, constructions such as the Roon example in (4) clearly 
represent a kind of associative plural, which, in lieu of a better term 
(any suggestions?), might be referred to as Associative Plural by 
Disagreement. But their status as inclusory constructions is at best 
questionable.

However interesting this terminological discussion has been, I would 
still greatly appreciate more data, and especially negative data from 
all the boring run-of-the-mill languages you're familiar with which 
(like Alemannic; thanks, Frans!) do NOT have the ASPD construction.

Thanks,

David

Acknowledgment: much of my understanding of these issues is due to a 
lengthy email correspondence with Misha Daniel last month, for which I 
am very grateful. (Though any shortcomings are entirely my own.)

-- 
David Gil

Department of Linguistics
Max Planck Institute for Evolutionary Anthropology
Deutscher Platz 6, D-04103 Leipzig, Germany

Telephone: 49-341-3550321 Fax: 49-341-3550119
Email: gil at eva.mpg.de
Webpage:  http://www.eva.mpg.de/~gil/