[Lingtyp] one N vs CLF N

[David Gil]

On 13/05/2019 19:46, Martin Haspelmath wrote:
> On 13.05.19 14:34, David Gil wrote:
>>> David mentions the difference between "syntax" and "morphology", and 
>>> while this is traditionally considered important, Silva merely talks 
>>> about "a singulativizing and individuating marker". But *markers* 
>>> are not always affixes, so there is no need to decide whether 
>>> Vietnamese classifiers are affixes. All that matters is that they 
>>> are markers (i.e. they are bound forms which are not roots), and 
>>> this is not in question.
>> But there is still a problem here.  In many languages, e.g. 
>> Malay/Indonesian, Tagalog, bare nouns (a) have general number, and 
>> (b) may occur in construction with what seems like a numeral 'one', 
>> but without a numeral classifier, as follows:
>>
>> ONE N
>>
>> Presumably we would not wish to call the ONE word a singulative, but 
>> how do we distinguish between it and the Vietnamese CLF? The fact 
>> that ONE can be replaced by TWO, THREE ... SEVENTEEN ...  ? But then 
>> there are surely languages that have the ONE N construction but with 
>> a "limited" numeral system, e.g. with only ONE and TWO.  The 
>> difference between CLF N and ONE N seems to be that the former is 
>> "grammatical" in a sense that the latter is not, but it's not clear 
>> to me how to objectivize the notion of grammatical ... other than by 
>> recourse to a notion of wordhood.
>>
>> David
>
> Here I think that the term *marker* (also used by Silva Nurmio) is 
> useful – a marker is a bound form that is not a root (i.e. not a 
> minimal noun, verb or adjective).
>
> In most languages, the numeral 'one' is not a bound form – it can 
> stand on its own, e.g. in answers to questions (How many daughters do 
> you have? One). I would assume that this is also the case in Malay and 
> in Tagalog. Thus, these forms are not singulative markers. By 
> contrast, numeral classifiers cannot be used on their own (in many 
> languages), so then they are markers (and thus potentially singulative 
> markers).
>
> Measure terms used with mass nouns like German "korn" (in Reis-korn 
> 'grain of salt', mentioned by Harald Haberland) are generally free 
> forms, and thus not singulative markers either.
>
> Martin
>
Martin's suggestion is reasonable. There will still be problematical 
cases of classifiers that can stand alone "as nouns", e.g. Vietnamese 
"con" which is both classifier and noun, but such intermediate cases are 
inevitable, and I certainly don't have a better suggestion to offer.

-- 
David Gil

Department of Linguistic and Cultural Evolution
Max Planck Institute for the Science of Human History
Kahlaische Strasse 10, 07745 Jena, Germany

Email: gil at shh.mpg.de
Office Phone (Germany): +49-3641686834
Mobile Phone (Indonesia): +62-81281162816