Sunday, December 21, 2008

Compatibility of se, khong, co

Here's an entry to the VGP site that I've been working on recently. Any help or advice anyone might be able to give would be gratefully received:

Compatibility of sẽ/đã and có

Some consultants report an incompatibility of the tense-markers sẽ/đã with the assertion marker có. A commentator on Duffield (2007), for instance, found example (19b) below unacceptable, and attributed this ungrammaticality to the fact "in general, sẽ and đã are incompatible with có." Another consultant also noted his dispreference for sentences containing both elements simultaneously.

(19) a. Hôm qua anh ấy đã không có đến nhà chị.
yesterday PRN DEM ANT ASR go house PRN
‘He didn’t go to your house yesterday.’

b. Hôm qua anh ấy đã có đến nhà chị không?
yesterday PRN DEM ANT ASR go house PRN khong
‘Did he go to your house yesterday?’
‘*He didn’t go to your house yesterday?’

While not disputing these consultants' intuitions, it may be that the source of the unacceptability lies elsewhere. Of course, I have a personal stake in this: if it is the case that these two elements compete for the same syntactic position, this threatens to undermine the general analysis presented in Duffield (2007), which claims that sẽ and đã are tense elements inserted under T, whereas có is located in a separate functional head, ASR. See Fig 1.

Yet, even leaving aside any theoretical agenda, there are reasons to doubt that these two types of element are in complementary distribution. First, an internet search on the strings {"tôi sẽ có" + V/ "tôi đã có" + V} reveals numerous examples of sentences in which sẽ or đã immediately precedes có + V. The following are representative:

1a. Sư rất vui vì tôi đã có gặp Sư trong lần về thăm nhà năm trước, 2002. [, accessed 16/12/08]
Sư very happy because I ant asr meet Su in time return visit home year last, 2002
'Su is very happy because I visited when I returned home for a visit last year, in 2002.'

b. Tôi đã có gặp anh Phòng một lần từ thời còn ở Tiên Phước. [; accessed 16/12/08]
I ant asr meet prn Phong one time from time still be-in Tien Phuoc
'I met Phòng once when I was still in Tiên Phước.'

c. Tôi đã có đi vòng quanh thế giới nhiều lần. []
I ant asr go around world much time
'I have been around the world many times.'

2a. Tuy nhiên nếu ai nhìn ra tiềm năng của đầu tư trong lĩnh vực này thì sẽ có biết phải làm gì! [, accessed 21/12/08]

b.Hôm nay là một ngày mới, ta sẽ có biết bao nhiêu cơ hội để làm được những điều tốt đẹp!”. [, accessed 21/12/08]

c. Nhưng sẽ có biết bao nhiêu người trên Thế Giới nhìn vào đó. [ 81950199dc2a7b1b46f2203209421b5a, accessed 21/12/08]

The examples in (1) and (2) certainly suggest no real incompatibility between sẽ/đã and có in affirmative indicative contexts. Instead, the problem in (19b) may lie with an incompatibility between không and (just in case functions as an assertion marker preceding the matrix predicate, rather than the matrix predicate itself. As noted in Duffield (2007), shows striking formal and functional parallels with English do, which can also function either as a grammatical auxiliary in negative, interrogative and emphatic contexts ("do-support") or else as a main verb. The speculation that Vietnamese works in the same way is supported by a second internet search that reveals many thousands of hits for the exact sequence "tôi đã không có" + NP, as illustrated in (3), but very few for the same sequence immediately followed by a verb, functioning as the matrix predicate (i.e. "tôi đã không có + V"): cf. the examples in (1) above. [The only exception is—predictably enough (!)—được: see section NN below]

3a. Nhưng tính tôi, đã không có tình cảm,... [, accessed 16/12/08]
but I ant neg have feeling
'But I had no feelings...'

b. Bố tôi đã không có con trai để cùng chia sẻ [, accessed 16/12/08]
father I ant neg have child male in order to share
'My father had no sons to share.'

As for "sẽ không có" this is vanishingly rare. Indeed, a Google search revealed exactly one hit for the string "sẽ không có" (a Googlewhack, in net terminology!). It is also telling, I think, that this is a counterfactual context, supporting the idea that an assertion marker có is necessarily tied to past tense contexts (see Duffield 2007):

4a. Nếu Nguyễn Thanh Quang cấm tôi vì lý do chính đáng, hay làm theo ý kiến cộng đồng, tôi sẽ không có nói gì.
[ắn_cho_Tin_nhắn_cho_người_quản _lý/Liebesapfel, accessed 21/12/08]
If NTQ had banned me for a good reason, or was following the opinion of the community, I would not have said anything.'

The comparative rarity of any "đã không có + V" strings could be explained if, in general, không and (ASR) có competed for the slot. [This would, however, necessitate a re-analysis of VP-ellipsis constructions in Vietnamese: see section NN, Duffield (in prep.), cf. Duffield (2007).]

A more radical alternative explanation for the unacceptability of (19b) may be that Y-N questions in Vietnamese involve a "Kaynian approach": sentence-final không occupies the ASR-head, while the rest of the sentence forms a tenseless topic constituent. See Duffield (in prep., for further development of this idea). Whichever is correct, however—and the two are not mutually exclusive analyses—the data in (1)-(3) above imply no real incompatibility between Tense and Assertion elements.