On what projects in Vietnamese

Abstract

By comparison with other areally- and typologically-related languages, the Vietnamese language disposes of a large and diverse set of (non-affixal) grammatical particles: these display interesting parallels with functional heads in familiar Western European languages. Most of these grammatical morphemes are ‘multifunctional’ in the sense that their meaning is largely—in some cases, exclusively—determined by their clausal distribution; alternatively, by their configurational relationship to other grammatical morphemes. In this paper, I document the distribution of these particles, working down the clausal spine. I also present a set of analyses of those cases where particles interact with one another, with each group considered in its own terms. Following this presentation, some broader implications of these analyses are briefly considered: it is suggested that a more satisfactory explanation of Vietnamese grammar can be found if it is assumed that grammatical meaning inheres in syntax, rather than in lexical representations.

Full text

Running head: ON WHAT PROJECTS IN VIENAMESE
1
!
!
!
On What Projects in Vietnamese
Nigel Duffield1
Konan University
duffield@center.konan-u.ac.jp
ORCID iD: 0000-0002-4797-9282
1 Department of English and American Literature and Language, Faculty of Letters, Konan
University, 8-9-1 Higashinada, Kobe, Hyogo 658-8501, Japan.

ON WHAT PROJECTS IN VIETNAMESE
2
On What Projects in Vietnamese
Abstract
By comparison with other areally- and typologically-related languages, the Vietnamese language
disposes of a large and diverse set of (non-affixal) grammatical particles: these display
interesting parallels with functional heads in familiar Western European languages. Most of
these grammatical morphemes are ‘multifunctional’ in the sense that their meaning is largely—in
some cases, exclusively—determined by their clausal distribution; alternatively, by their
configurational relationship to other grammatical morphemes. In this paper, I document the
distribution of these particles, working down the clausal spine. I also present a set of analyses of
those cases where particles interact with one another, with each group considered in its own
terms. Following this presentation, some broader implications of these analyses are briefly
considered: it is suggested that a more satisfactory explanation of Vietnamese grammar can be
found if it is assumed that grammatical meaning inheres in syntax, rather than in lexical
representations.
Keywords
Vietnamese phrase structure • Multifunctionality • Syntactic Cartography • Outer vs. Inner
Aspect • Tense and Clausal Negation
!

ON WHAT PROJECTS IN VIETNAMESE
3
On What Projects in Vietnamese
Introduction
This paper offers a cartographic study of Vietnamese clause structure. While the empirical
coverage of the paper is broadly commensurate with that of preceding versions, its goals are
more modest: for the most part it aims at purely observational—at best descriptive—adequacy, in
the sense of Chomsky (1965). Through a systematic exposition of the distribution and
interpretation of grammatical morphemes above and within the thematic verb phrase (or
predicate phrase) the primary goal of the article is to establish a body of syntactic facts that any
more adequate theoretical analysis should be able to account for.
Vietnamese is a particularly rewarding language to study since it exhibits ambiguous, and
contrary, morphological and syntactic properties. A naïve consideration of Vietnamese sentences
from the perspective of Western Indo-European languages suggests that Vietnamese is the
archetypal ‘isolating’ language variety (see, for example, Whaley (1997)) in as much as it is
devoid of bound inflectional morphology, and—aside from a (proclitic) reflexive element—also
lacks any bound derivational morphology (causative, passive, (de-)transitivizing morphemes)
that might plausibly be considered to be syntactically active; cf. Baker (1985, 1988). Temporal
relations are generally expressed through tense adverbials and conjunctions, or inferred
contextually; also, as we shall see, transitivity alternations are handled analytically. However,
even though it completely lacks affixation Vietnamese has a rich inventory of grammatical
morphemes—albeit these are optionally expressed: whenever these elements are syntactically
projected, their distribution and interpretation is strikingly familiar from a Western perspective,

ON WHAT PROJECTS IN VIETNAMESE
4
and correspondingly alien from the point of view of speakers of other South East Asian
languages, especially Sinitic varieties; see also Alves (1999).
Viewed in a more universalist light, Vietnamese reveals itself as a near-perfect blend of
the layered functional syntax that was originally motivated by inflectional categories of Western
languages—hence the anachronistic terms ‘INFL’, IP, AgrS, AgrO, etc.—with the syntactic
transparency of isolating East Asian languages, whose underlying structure is unobscured by
morphologically-driven head-movement. That at least is what is suggested by the data presented
here.
Fig. 1 below articulates the layers of grammatical meaning involved in a typical English
sentence with past time reference, such as in example (1a); examples (1b) and (1c) show the
Dutch and Vietnamese equivalents.
a. She read the book. (1)
b. Ze heeft het boek gelezen.
c. Cô ấy (đã) đọc quyển sách.
PRN.DEM2 (ASP) read CL. book
‘She (has) read the book.’

ON WHAT PROJECTS IN VIETNAMESE
5
Figure 1. Layers of meaning in an English indicative sentence
Intuitively, the English sentence comprises the following elements of grammatical
meaning. Working from the bottom up, we first encounter the root meaning of the lexical
predicate READ: stripped of its argument structure, READ (inherently) denotes an ACTIVITY of
interpretation or information transfer (cf. ‘she read the book/my mind/_ to her children/the
weather’). The addition of an object—i.e., the transitivization of READ—introduces the
possibility of a BOUNDED EVENT, though it does not by itself force a change of Aktionsart from
activity to accomplishment (cf. ‘She read the book for hours at a time.’). The addition of a
subject argument implicates a CAUSATIVE meaning of some kind: whereas previously this may
have been treated as a primitive thematic relation (‘Agent’, ‘Actor’) borne by the subject DP,
most recent syntactic analyses treat this added interpretation relationally, analyzing the subject as

ON WHAT PROJECTS IN VIETNAMESE
6
the argument of a covert causal predicate, e.g., v: see Baker (1997), also Copley & Martin
(2015). The examples in (1) also involve ASSERTION VALIDITY in the sense of Klein (1998): as
discussed below, finite indicative sentences involve the claim that a given proposition holds or
held—or does/did not hold, in the case of negative sentences. Finally, tense morphemes serve to
situate these meanings relative to the time of utterance in a particular universe of discourse. In
the cases at hand, the English preterite form read in (1a), unambiguously marks PAST TENSE,
whereas the Dutch present perfect form is ambiguous between a present perfect and a preterite
interpretation. As we shall see directly, the Vietnamese sentence in (1c) exhibits the same
ambiguity as is found in the Dutch example, thanks to the pre-verbal morpheme đã.
Though few would deny that these hidden meanings are a crucial part of clausal
interpretation, the idea that they correspond directly to syntactic functional projections is
disputed; even more controversial is the idea that this functional architecture is universal. In what
follows, however, I will suggest that these ideas are true, and that Vietnamese makes the best
possible case for them. Specifically, I will argue for the cartography of Vietnamese given in (2),
in which layered functional projections appear in two regions: (i) outside the thematic domain, in
what is conventionally termed the IP-domain; (ii) within the thematic vP, which—following
Travis (2010)—I will term the ‘Inner Aspect’ domain. (The Kleene star symbol in (2) denotes
iterable projections of the same semantic kind: for example, different sorts of aspectual
projection).
The rest of the paper is organized as follows. Section 2 is concerned with functional
projections in (2) above the thematic verb-phrase: the first subsection provides motivation for the
projection of Tense, (Outer) Aspect and Negation as autonomous syntactic positions, the second
part deals with Assertion (Mood) and Modality. In Section 3, the focus is on justifying the

ON WHAT PROJECTS IN VIETNAMESE
7
existence of functional heads within the thematic verb-phrase: Inner Aspect and causative v.
Section 4, which concludes the paper, considers more general implications of the observed
multifunctionality of grammatical morphemes in Vietnamese.
(2)
2. IP and vP Syntax in Vietnamese
2.1 Tense, Aspect and Negation
Following Pollock (1989), Chomsky (1989), and innumerable subsequent ‘Split-INFL’
treatments of clausal phrase-structure, I will assume that aspectual auxiliaries in both French and

ON WHAT PROJECTS IN VIETNAMESE
8
English are initially merged with the predicate phrase prior to merger of Negation and Tense, and
that surface word-order in finite, indicative clauses involving auxiliaries is the result of Asp-T
movement. No such movement is observed in English non-finite clauses, where T is filled by to
(see also Roberts 1993); in French non-finite clauses, this movement is assumed to be optional.
French and English are thus primarily distinguished by the lack of main-verb raising in finite
contexts, and by the consequent appearance (in English) of do-support in cases where lowering is
blocked, such as negative and emphatic contexts. The paradigms in (3) and (4) below illustrate
the parallels and contrasts between these two languages:2
a. She does not often read letters written by hand. (3)
b. She has rarely read a letter written by hand.
c. ?She claims to not have read such a letter.
d. She may not have read such a letter.
e. I suggest you not be here when I get back.
f. For you to have been living there for six years without learning the language is
disgraceful.
2 For present purposes, I ignore the French pre-verbal clitic negation marker ne. Pollock (1989)
explains its distribution in terms of obligatory Neg-T raising; however, the sole motivation for
this move would seem to be to account for its unexpected position. Similarly, I ignore the fact
that in English non-finite clauses negation more typically precedes English to; cf. (3c) above. It
is plausible to think that these anomalies are related.

ON WHAT PROJECTS IN VIETNAMESE
9
a. Elle ne lit pas souvent de lettres manuscrites. (4)
she writes NEG often DET letters handwritten
‘She doesn’t often read letters written by hand.’
b. Elle n’a pas récemment lu une lettre manuscrite.
she has NEG recently read a letter handwritten
‘She hasn’t recently read a letter written by hand.’
c. Elle regrette de ne pas avoir lu sa lettre.
she regrets to NEG have read his letter
‘She regrets not having read his letter.
d. Elle nie absolument (d’) avoir écrit la lettre.
she denies absolutely to have written DET letter
‘She absolutely denies having written the letter.’
To determine whether Tense and Aspect are syntactically projected in the same fashion in
Vietnamese, one needs first to address the more fundamental question of whether Vietnamese
has tense at all. The more descriptive literature on Vietnamese offers a full spectrum of views on
this matter. At one extreme, one finds the categorical denial of Nguyen (1998: 116),3 who insists
that ‘[T]rong tiếng Việt không có phạm trù thì… [In Vietnamese, there is no tense]’; at the other
end, Lo Cicero (2001) claims a direct correspondence between the French past, present and
future tenses and the Vietnamese elements đã, đang, and sẽ’:
‘La correspondance des temps verbaux du vietnamien aux français, si l'on peut dire ainsi,
est simple d’une manière générale: ‘đã’ exprime les temps du passé, ‘đang’—ou sans
3 See also Cao (1998:10), for a near identical claim.

ON WHAT PROJECTS IN VIETNAMESE
10
‘đang’—le présent, et ‘sẽ’ le futur, ces termes précedent les verbes vietnamiens. Pour
mettre en valeur le moment de l'action, la langue vietnamienne se sert donc de termes ou
marqueurs comme ‘đang, đã, sẽ.’4
The evidence presented directly below suggests that both of these views are mistaken,
though Nguyen’s position is closer to being correct, in the literal sense that there are no pure
tense morphemes in the language; Lo Cicero’s claim, by contrast, can readily be shown to be
empirically false. However, just because there are no tense morphemes in Vietnamese, this does
not mean that there is no TP projection.
Consider first, the examples in (5-7) below, which indicate the distribution of đã, sẽ, and
đang relative to three VP-external elements: the clausal negation marker không (5), manner
adverbials such as cẩn thận (‘carefully’) (6), and temporal adverbs, such as hôm qua
(‘yesterday’) (7). It should be clear that đã and sẽ display the same distribution as tensed
auxiliaries in English (finite) clauses: đang shows more variability, but it generally behaves
much more like an aspectual auxiliary (cf. English progressive be) than as a tense morpheme.
Notice also that just in negative contexts (5a) đã has an exclusively preterite, rather than perfect,
interpretation (a point to be returned to directly).
4 ‘Broadly speaking, there is a straightforward correspondence between French and Vietnamese
tense systems: đã expresses past tense, đang (or without đang) the present and sẽ the future, each
of these elements preceding the [lexical] verb. To indicate the time of an action, Vietnamese thus
uses markers such as 'đang, đã, sẽ.' (2001) [my translation: NGD]'

ON WHAT PROJECTS IN VIETNAMESE
11
a. Tôi (đã) không (*đã) làm việc đó. (5)
PRN PAST NEG ASP do job DEM2
‘I didn’t do that.’
b. Tôi (sẽ) không (*sẽ) làm việc đó.
PRN FUT NEG FUT do job DEM2
‘I will not do that.’
c. Tôi (đang) không (đang) ăn cơm.
PRN ASP NEG ASP eat rice
‘I am not having a meal.’
a. Tôi (sẽ) cẩn thận (*sẽ) viết lá thư này. (6)
I FUT carefully FUT write letter DEM1
‘I will write this letter carefully.’
b. Anh ấy (đã) cẩn thận (*đã) đọc quyển sách này.
PRN.DEM ANT carefully ANT read CL book DEM1
‘He (has) read the book carefully.’
a. *Anh Lài đã hôm qua giúp tôi. (7)
PRN Lai ANT yesterday help me
‘Lai helped me yesterday.’
b. ?[TP Anh Lài hôm qua đã [VP giúp tôi ]].
PRN Lai yesterday ANT help me
‘Lai helped me yesterday.’

ON WHAT PROJECTS IN VIETNAMESE
12
c. [ToPP Anh Lài thì [TP hôm qua [TP pro đã giúp tôi ]].
PRN Lai TOP yesterday ANT help me
‘Lai, he helped me yesterday.’
With respect to đang, the examples in (8) quickly dispel Lo Cicero’s claim that this is a
marker of present tense: examples (8a) and (8b) show that đang can readily appear in past and
future time contexts, respectively, while examples (8c) and (8d) express generic—timeless—
assertions. Taken together, these examples confirm that đang is an aspectual morpheme,
expressing durativity (progressive aspect).
a. Lúc đó, họ đang chơi quần vợt. (8)
place DEM2 PRN ASP play tennis
‘At that time, they were playing tennis.’
b. Sang năm, vào ngày này, chắc tôi đang làm ở Pháp.
enter year, enter day DEM1, sure I ASP work in France
‘By this time next year, I shall be working in France.’
c. Trẻ em đang biết rất nhiều điều không nên biết.
young PRN ASP know very much thing NEG should know
‘Young people know (lit. are knowing) a lot of things they shouldn’t.’
d. Hãy quý những gì mình đang có.
IMP treasure PL what self ASP have
‘Treasure what you have (lit. are having).’
The tenseless nature of đang is re-inforced by the grammatical examples in (9) and (10),
which show that đang, just like the English progressive be_ing, is fully compatible with perfect

ON WHAT PROJECTS IN VIETNAMESE
13
đã (9) and future sẽ (10a) morphemes; this immediately contrasts with the complementary
distribution of đã and sẽ, exemplified in (10b).
a. Lúc tôi đến, nó đã đang ngủ rồi. (9)
time I come, PRN ASP ASP sleep already
‘When I came, he had been sleeping.’
b. Vào giờ này tuần tới tôi đã đang nghỉ mát ở Hawaii rồi.
come hour DEM1 week next I ASP ASP holiday BE Hawaii already
‘By this time next week I will have been holidaying in Hawaii.’
a. Đừng gọi điện từ 7 đến 8 giờ. Lúc đó chúng tôi sẽ đang dùng cơm tối. (10)
NEG.IMP call tel. from 7 to 8 hour. time DEM2 PL I FUT ASP have. meal even.
‘Don’t call me between 7 and 8! At that time we shall be having dinner.’
b. Vào giờ này tuần tới tôi (*sẽ) đã nghỉ mát ở Hawaii rồi.
come hour DEM1 week next I FUT ASP holiday be Hawaii already
‘By this time next week I will have been on holiday in Hawaii.’
Hence, the distributional evidence of examples (5)-(10) is consistent with the idea that đã
and sẽ are pronounced in T, and that đang occupies some lower aspectual position. In so far as sẽ
has an invariant future interpretation, and appears devoid of aspectual interpretation, I will
assume that it is base-generated in this position. (This assumption is further supported by its
exclusion from Yes-No questions; see below.) By contrast, even though đã appears in T, its
primary interpretation is not as a past tense marker, but as an aspectual morpheme. This is
supported by two pieces of evidence. First, the future perfect and counterfactual examples in (11)
demonstrate that đã freely occurs in contexts incompatible with a past tense morpheme:

ON WHAT PROJECTS IN VIETNAMESE
14
a. Bằng giờ này năm sau, chị đã là giáo viên rồi. (11)
by time this year next, she ANT COP teacher already
‘By next year, she’ll already be working as a teacher instead.’
b. Đến cuối năm nay, tôi đã ra.trường.
arrive end year DEM1 PRN ANT go.out.school
‘I shall have graduated by the end of the year.’
c. (Nếu) ông nói với tôi sớm.hơn thì tôi đã săn.sóc đến việc ông.
(if) PRN say with me earlier TOP I ANT take.care work PRN
‘If you had told me about it earlier, I would have taken care of that business of yours.’
Second, the examples in (12) show that outside of negation contexts—compare (5a), and
(14) below—đã expresses anteriority (prior inception). As one expects of an aspectual
morpheme, the precise interpretation of đã is sensitive to the properties of the associated lexical
predicate: thus, with stative predicates such as sáng (12a) addition of đã signals a change of state
(the sky became brighter prior to the Topic Time); with achievement predicates such as thắng
(12b), đã carries the assertion that the event occurred prior to the utterance time, as well as the
implication that it had not occurred previously; whereas with activity and accomplishment
predicates (12c-d), đã signals that the inception of the event or activity precedes the Topic Time
(Klein 1994). See Phan (2013) for further discussion.
a. Ngoài đường trời đã sáng. (12)
out road sky ANT bright
‘It$got$bright$out$there.’!!![from&(Trần,#K.P.#(2008:73)].!

ON WHAT PROJECTS IN VIETNAMESE
15
b. Cuối cùng Andy Murray đã thắng cuộc. [from Phan (2013)]
Finally Andy Murray ANT win contest
‘Finally, Andy Murray won the contest.’
c. Tàu đã chạy.
train ANT run
‘The train has departed.’
d. Nó đã viết bài ở văn phòng.
3SG ANT write paper at office
‘He wrote a paper at the office. ’
Crucially, the perfect/anterior value contributed by đã should be distinguished from
perfectivity or completion (which is signaled by post-verbal particles, see below): except for
achievement predicates, where event completion prior to the utterance time is entailed by their
meaning, đã does not entail, or even imply completion of the event or situation. This is supported
by the absence of any completion entailment in the examples in (13) below (similarly, (12b) may
be true of a situation in which the paper has been started, but is still unfinished):
a.i Tàu đã chạy rồi mà giờ nó lại dừng. (13)
train ANT run already but now PRN again stop
‘The train has already run [= set off], but it has now stopped again.’
a.ii Tàu đã chạy rồi và giờ nó vẫn chưa dừng.
train ANT run already and now PRN still not.yet stop
‘The train has already run [= set off], and hasn’t stopped yet.’

ON WHAT PROJECTS IN VIETNAMESE
16
(14)
All of the preceding examples involving đã are consistent with the analysis sketched in
(14), in which đã is initially merged in Asp and subsequently raised to T. On this analysis, the
complementary distribution of đã and sẽ in (15a-b) exactly mirrors the complementarity of
modals with finite aspectual auxiliaries (or do-support) in English (15c-d).
a. Đến cuối năm nay, tôi (*sẽ) đã ra. trường. (15)
arrive end year DEM1 PRN FUT ANT go.out.school
‘I shall have graduated by the end of the year.’
b. Vào giờ này tuần tới tôi (*sẽ) đã nghỉ mát ở Hawaii rồi.
come hour DEM1 week next I FUT ASP holiday be Hawaii already
‘By this time next week I will have been on holiday in Hawaii.’

ON WHAT PROJECTS IN VIETNAMESE
17
c. *She will has graduated by the end of this year.
d. *Kerry might does indeed know the way to solve this.
To this point, the Vietnamese data are fairly mundane when considered from a theoretical
perspective: the interim conclusion must be that Vietnamese does have syntactic tense (TP), but
that the T position, when lexicalized, is primarily occupied by an aspectual morpheme. However,
matters become more interesting when negative contexts are taken into consideration. So
consider now the interactions and interpretive alternations involving đã, the default clausal
negation morpheme không (‘not’) and its perfect alternant chưa (‘not yet’), exemplified in (16)-
(18) below, many of which also feature the ‘assertion’ marker có.
First, the examples in (16) demonstrate the mutual compatibility of đã, không and có in
emphatic declarative clauses. While the distribution of these elements is exactly as predicted, the
interpretive restriction – already seen in (5a) – is not: here (i.e. in negative sentences), the
aspectual reading is canceled; đã is exclusively interpreted as a preterite marker.
a. Hôm qua anh ấy đã không có đến nhà chị. (16)
yesterday PRN DEM2 PAST NEG ASR arrive house prn
‘He didn’t go to your house yesterday.’
b. Trong bản khai, nó đã không có nói gì đến tổ chức cả.
in CLS state. PRN PAST NEG ASR say what about organiz. all
‘In his statement, he didn’t say anything at all about the organization.’
The observation that đã is unambiguously a past time marker in negative contexts is
originally due to Trinh (2005), who presents the minimal contrast in (17a) and (17b) below.
Trinh (2005) treats this restriction as a case of lexical homophony: on his account, there are two

ON WHAT PROJECTS IN VIETNAMESE
18
separate lexical entries in the Vietnamese lexicon—perfect đã1 and preterite đã2, each occupying
distinct underlying positions. In affirmative contexts, perfect đã1 is taken to be initially inserted
under PERF, an aspectual category close to—or just inside—the VP, and raised to T. Negation is
assumed to block this raising, leading to the insertion of an alternative morpheme—đã2—directly
to the T node, which yields the exclusive preterite reading in negative contexts.
a. Nó đã đọc sách. [perfect~preterite] (17)
PRN PAST/ASP read book
‘He read books/has read books.’
b. Nó đã không đọc sách.
PRN PAST NEG read book
‘He did not read books./*He has not read books.’
c. Nó chưa đọc sách. [perfect (negative)]
PRN not.yet read book
‘He has not read books.’
d. Nó đã chưa đọc sách. [past perfect (negative)]
PRN PAST not.yet read book
‘He had not read books.’
Duffield (2013a, b) offers an analysis that is very close in spirit to Trinh’s original
treatment, but which dispenses with homophony: on the revised (‘multifunctional’) analysis,
only one underspecified đã is lexically represented: this inherits a dual interpretation if inserted
in Asp and moved to T, but only has a temporal interpretation if inserted directly into T. Aside
from the massive redundancy implied by Trinh’s analysis—as Duffield points out, đã is wholly
typical of most Vietnamese functional morphemes whose interpretation varies systematically

ON WHAT PROJECTS IN VIETNAMESE
19
according to their structural position (không and có, for example, and most of the other
functional morphemes discussed in this article likewise have multiple interpretations)—the
homophony account fails to explain two cross-linguistic facts. The first observation is that
affirmative ‘present perfect’ forms in Modern Romance and Germanic varieties also display
perfect~preterite ambiguities: the addition of a temporal adverbial (gisteren ‘yesterday) to the
Dutch sentence in (1b), for example, automatically yields a preterite interpretation; alternatively,
with respect to Chinese, Lin (2005) claims that the aspectual particle le displays a similar
ambiguity. The second point to observe is that other languages also exhibit constraints on
negative/perfect interactions—see Matthews (1990), also Miestamo & Van der Auwera (2011):
once more in Chinese, le appears to be incompatible with both bu (18b) and meiyou (18c).
Hence, treating the Vietnamese interaction in terms of an arbitrary lexical specification would
seem to miss a significant crosslinguistic generalization.
a. ta qu le faguo. (18)
3SG go LE France
‘He went to France.’/ ‘He has been to France.’
b. *ta bu qu le faguo.
3SG NEG go LE France
‘He did not go to France.’ (Li 1999:235)
c. *ta meiyou qu le faguo.
3SG not.have go LE France
‘He hasn’t been to France.’ (Linda Badan, p.c.)
In fact, Trinh’s obervation needs to be understood as part of a wider paradigm of
interactions between past, perfect and negation morphemes in Vietnamese, a paradigm that

ON WHAT PROJECTS IN VIETNAMESE
20
(minimally) includes the pair of sentences containing chưa in (17c) and (17d) above. Example
(17c) shows that it is wholly grammatical to make a perfect/negative assertions—just as long as
chưa replaces both đã and không—while the acceptability of (17d) demonstrates that that there is
no structural incompatibility between đã and chưa, but makes clear that the negative constraint
still holds: rather than the expected perfect interpretation, đã chưa only admits a past perfect
reading (as indicated by the gloss).
Both previous proposals, Trinh (2005) and Duffield (2013a, b), share the assumption that
clausal negation blocks what would otherwise be a licit connection between T and a lower
functional head (Asp), (implicitly transferring to Vietnamese the Chomsky/Pollock (1989)
analysis of do-support as a failure of T-lowering). The obvious embarassment for both accounts,
however, is the fact that double insertion of đã—simultaneously above and below negation—is
ungrammatical, as shown by the examples in (19) below: indeed, in contrast to English or
French, unraised perfect đã is never possible.
a. *Anh-ấy đã không đã đến. (19)
PRN PAST NEG ANT come
‘He didn’t come.’
b. *Anh-ấy đã chưa đã đến.
PRN PAST NEGPRF ANT come
‘He hadn’t come.’
On earlier movement analyses, there is no way to explain this constraint other than
through appeal to the avoidance of syntactic haplology, which seems ad hoc at best: any strategy
involving morphological feature-checking seems even more unlikely given the isolating nature
of the language. In response to this difficulty, Phan & Duffield (2016) offer an alternative

ON WHAT PROJECTS IN VIETNAMESE
21
Nanosyntax solution, which covers the data in (15)-(17), and which also explains the
impossibility of double-đã insertion.
Nanosyntax (Starke 2009, Baunaz & Lander to appear, Caha, to appear) conceives of
lexical entries as competing pre-formed syntactic fragments—or L-trees—which can be
combined to derive syntactic representations (S-trees) in a one-to-many fashion; that is, a given
L-tree may fit a larger or smaller set of structural nodes. The various principles that determine
which L-tree best satisfies a particular section of a syntactic build—Superset Principle, Cyclic
Override Principle, Elsewhere Principle, etc.,—account for collocational restrictions and cases of
morphological suppletion, while dispensing with the need for head-movement operations. In the
case at hand, the lexical entries (L-trees) for đã and chưa and không are plausibly as in (20)
below; the relevant Nanosyntactic principles are given in (21):
(20)

ON WHAT PROJECTS IN VIETNAMESE
22
a. Superset Principle. An L-tree can match more than one S-tree: for a successful match, (21)
the L-tree must be the same size or larger than the S-tree (Caha 2009, 2014);
b. Elsewhere Principle. When more than a single L-tree can lexicalize an S-tree, the L-
tree with the fewest unused features is chosen;
c. Principle of Cyclic Override. Assuming that derivation is built bottom-up, later,
higher-level spellouts cancel out previous, lower-level spellouts (Lander 2016).
Phan & Duffield (2016) offers a fuller description of this analysis. Here, it suffices to
consider how the analysis derives: (a), the ambiguity of đã in affirmative contexts (17a); (b), the
preterite-only interpretation in negative contexts with không (17b); and (c) the blocking effects
of perfect chưa (17c,d).
• "#!$%&!'()&!*+!(++,-.($,/&!0&-+&'$!'*#$&1$)2!*#34!5&-+&'$5!,)!0-*6&'$&78!9%&!3&1,'*#!
*++&-)!$:*!:(4)!*+!)0&33,#;!*<$!$%&!0&-+&'$=!đã!*-!chưa=!>4!$%&!?<0&-)&$!0-,#',03&2!
>*$%!@A$-&&)!+*-!đã!,#!(#7!chưa!,#!BCDE!(-&!)<0&-)&$)!*+!?A$-&&!5&-+&'$58!9%&!
F3)&:%&-&!5-,#',03&!-&)*3/&)!$%&!'*.0&$,$,*#=!$%&!@A$-&&!+*-!đã!:,#)!*<$!),#'&!,$!%()!
+&:&-!<#<)&7!+&($<-&)8!5&-+&'$5!)0&33)!*<$!()!đã8!
• G)!+*-!$%&!0-&$&-,$&!,#$&-0-&$($,*#!*+!đã!,#!(++,-.($,/&!'*#$&1$)2!$:*!7&-,/($,*#(3!
)$&0)!(-&!,#/*3/&7=!
• At the first step, we also begin from PerfectP; as before, the best match in the
lexicon is the L-tree of đã, so PerfectP is spelled out by đã.
• Next, PastP is inserted on top of PerfectP. There is an L-Tree match for the whole
trunk PastP>PerfectP in the lexicon, which is spelled out by đã, over-riding the
first spellout. *đã đã is ruled out by the principle of Cyclic Override.

ON WHAT PROJECTS IN VIETNAMESE
23
• In case of past negative contexts involving không, there are once again two steps in the
derivation of this section:
• Beginning with NegP, there are two matches in the lexicon: không or chưa. The
Elsewhere Principle resolves the competition: the L-tree for không involves fewer
unused features, thus NegP spells out as không;
• Next, PastP is built on top of NegP (there being no match for the whole trunk section
PastP>NegP in the lexicon). Here, the lexicon only allows one scenario: NegP is
spelled-out by không, PastP is spelled-out by đã. This yields the correct word order
(đã precedes không), with the right interpretation (đã is interpreted as preterite only).
• Perfect negative contexts involving chưa also involve several steps in the derivation:
• For perfect negative chưa contexts (example 17c), the derivation starts from
PerfectP: the best match in the lexicon is the L-tree of đã, so PerfectP is spelled
out by đã. The derivation continues, building NegP on top of PerfectP: for this,
there is one match for the NegP>PerfectP in the lexicon, spelled out as chưa (the
L-tree for chưa). This higher spell-out chưa cancels out the previous spell-out đã
(so *chưa – đã is ungrammatical due to the Cyclic Override principle).
• In case of past perfect (17d), PastP is built on top of NegP>PerfectP. There is no
match for the whole trunk in the lexicon. Here, the lexicon only allows one
possibility: NegP>PerfP is spelled-out by chưa, and PastP is spelled-out by đã (by
the Superset Principle). *Double-đã (19) is ungrammatical due to Cyclic Override
principle.

ON WHAT PROJECTS IN VIETNAMESE
24
To a first approximation, this analysis derives the paradigm in (17) above. On the
understanding that certain issues concerning không remain unresolved, we may now turn
attention to functional categories situated lower within the IP domain, specifically, those
concerned with Assertion, Mood and Modality.
2.2 Assertion, Mood and Modality
For Klein (1998), the notion of Finiteness is viewed as composite of two separate
concepts: Tense and Assertion (aka ‘Assertion validity'): whereas Tense expresses the time of a
given event or situation, Assertion ‘marks the claim — the fact that the situation described by the
utterance indeed obtains, in contrast to the opposite claim (Klein 1998).’ (In other frameworks,
assertion validity seem to be close to the notion of ‘realis’, or event actualization; see, for
example Desclés (2016)). Klein draws attention to the intonation contrasts given in (22) to (24)
below: these show that in English, emphasis placed on finite auxiliaries may contrast either tense
(22b) and/or assertion validity (22c), whereas emphasis on tensed verbs serves only to contrast
tense and/or lexical meaning (23b,c); in order to focus assertion validity in sentneces without
aspectual auxiliaries, do-support must be invoked, as in (24). As discussed in Dufffield (2007,
2013), these paradigmatic distinctions imply that do-support is more than a ‘Last Resort’ strategy
to save Tense inflection; within the Pollock/Chomsky (1989) framework adopted here, they also
imply that—unlike tense features—assertion features may not be lowered (even in the absence of
negation).
The book was on the table. (22)
b. The book is on the table.
— No, the book was on the table.

ON WHAT PROJECTS IN VIETNAMESE
25
c. The book was not on the table.
— No, that’s wrong, the book was on the table.
a. John loved Mary. (23)
b. John loved Mary, but he doesn’t love her now.
c. John loved Mary, but he didn’t adore her.
The idea that he didn’t love her is plainly wrong: (24)
— John (really) did love Mary.
Exactly this (pure) assertion function is expressed in affirmative Vietnamese sentences by
the preverbal particle có, which is the next element encountered after Tense>Negation>Aspect,
moving down the clausal spine. The examples in (25) demonstrate the contrastive function of có,
including its compatibility with sentential negation (25b); those in (26) reinforce the point that có
occurs relatively low in the clause, even below the progressive auxiliary đang, but still above
manner adverbials (26c):
a. Anh có mua sách! (25)
PRN ASR buy book
‘He DID buy the book!’
b. Anh không có mua sách!
PRN NEG ASR buy book
‘He did NOT buy the book!’
c. Bà đâu có phải là nguời Hành-thiện!
PRN DAU ASR right cop person Hanh-Thien
‘She is NOT a native of Hanh Thien, I tell you!’

ON WHAT PROJECTS IN VIETNAMESE
26
d. Có (chứ!)
ASR (exclamative)
‘(He) did (indeed)!’
a. Tôi (đã) có (*đã) gặp anh Phòng một lần từ thời còn ở Tiên Phước. (26)
I ANT ASR ANT meet PRN Phòng one time from time still be.LOC Tien Phuoc
‘I met Phòng once when I was still in Tiên Phước.’
b. Tôi biết là [chị (đang) có (*đang) yêu một người].
I know COMP PRN PROG ASR PROG love one person
‘I know that she’s in love with someone.’
c. Anh ấy đã (?có) cẩn thận (?có) đọc quyển sách này.
PRN.DEM ANT ASR carefully ASR read CL book DEM1
‘He (has) read the book carefully.’
These data thus provide further prima facie evidence for the structure in (14) above. A
question that arises at this point, however, is whether assertion (validity) is best viewed as an
atomic property, or as one feature of a multivalued functional node. The latter conclusion is
suggested by the fact that có exhibits even greater multifunctionality than English auxiliary do:
as illustrated by the examples in (27), có also appears in Yes-No questions (27a), as a stand-
alone light verb (≈ ‘have’) (27c), and as an existential predicate in (27d) (with unraised subject;
note the clause-initial position of the future morpheme sẽ, indicating an empty {Spec, TP}).
What is more, có also occurs in negative imperatives, as will be seen directly:

ON WHAT PROJECTS IN VIETNAMESE
27
a. Chị (có) mua cái nhà không. (27)
PRN ASR buy CL house NEG
‘Did you buy the house?’ (not ‘You didn’t buy…)
b. Con uống thuốc chưa?
PRN drink medicine not.yet
‘Have you taken your medicine?’
cannot mean ‘You [child] have not taken your medicine.’
c. Cô ta sẽ có quá nhiều tiền mà không thể đếm được hết.
PRN FUT HAVE so much money RM not can count can all-up
‘She will have so much money that she won’t be able to count it.’
d. Sẽ có người cần anh. (song title)
FUT exist person need you
‘There will be someone who needs you.’
Notice in passing the (obligatorily) final position of the negation morpheme không in
construction with có in Yes-No questions. A similar brace construction is observed with
đã…chưa in (27b): significantly, in contrast to indicative constructions cf. (17d), đã in Yes-No
questions has a perfect, rather than past perfect, interpretation). Although space constraints
preclude full discussion of this Neg~Q alternation, which is analyzed in Duffield (2013a), the
facts presented there are most consistent with the phrasal movement analyses sketched in (28)
and (29) below:

ON WHAT PROJECTS IN VIETNAMESE
28
(28)
(29)

ON WHAT PROJECTS IN VIETNAMESE
29
These analyses are hardly immune to criticism: see, for example, Nguyễn Đình Hoà
(1997), Law (2015), for alternative accounts. The treatment in (29) of (perfect) đã…chưa in
(27b) is also hard to reconcile with the Nanosyntactic derivation of chưa perfect structures, such
as (17c) above.5 Given this, it seems likely that this analysis should be reconsidered. Whatever
supercedes it, however, will need to capture the intuition that medial and clause-final không are
instances of the same multifunctional morpheme, arguably of the same functional projection.
This is strongly suggested by the fact that there are no negative Yes-No questions in Vietnamese
(in direct contrast to tag questions), as evidenced by the paradigm in (30):
a. Anh ấy không đến *không/(có) phải không? (30)
PRN DEM NEG come NEG ASR right NEG
‘Isn’t he coming?/He isn’t coming, is he?’
b. Bạn chưa xem phim này *không/(có) phải không?
friend NOT.YET see film DEM NEG ASR right NEG
‘Haven’t you seen the film yet?/You haven’t seen the film yet, have you?’
c. *Mày không có xu nào không/(có) phải không?
PRN NEG have money which NEG ASR right NEG
‘Haven’t you got any money?/You haven’t any money, have you?’
5 The two analyses can be reconciled, but one would need to assume that chưa is attached only
after the PerfP has been raised out of the scope of NegP (otherwise the Cyclic Override Principle
should cancel the lower đã). If that were the case, however, then the Elsewhere Principle would
seem to favor insertion of không, rather than chưa, as the realization of bare negation.) There is
thus an obvious analytic tension here, which must be resolved.

ON WHAT PROJECTS IN VIETNAMESE
30
It should be clear that this constraint follows directly from the analyses in (28) above, as
does the fact that preterite đã is excluded from Yes-No questions: this is shown by the minimal
contrast between (31a) and (31b = 16a).6
a. Hôm qua anh ấy (*đã) có đến nhà chị không? (31)
yesterday PRN DEM2 PAST ASR arrive house prn NEG
‘Did he go to your house yesterday.’
b. Hôm qua anh ấy đã không có đến nhà chị.
yesterday PRN DEM2 PAST NEG ASR arrive house prn
‘He didn’t go to your house yesterday.’
Returning now to the other functions of có, the examples in (27) imply that có expresses
a mixture of modal (Mood) and argumental properties: on the one hand, it is often interpreted as
a realis marker Duffield (2013b), at the same time, as discussed in Duffield (2007, 2011), có
seems to function as an ‘Event Argument’, where the presence of có in a sentence implies a
specific event or situation (stage-level property). Klein’s own discussion (1998) of the concept of
assertion validity suggests that it should alternate with other modal (Mood) features. He writes:
6 It seems very likely—as Nguyễn Đình Hoà (1997) supposes—that Yes-No questions developed
diachronically from alternate bi-clausal ‘Yes-or-No’ questions involving predicate reduplication,
and in which có functioned as an emphatic affirmative marker (‘indeed-VP (or) not-VP’): that is
to say, the original construction involved ellipsis under identity of the second predicate phrase;
see also Law (2015). However, final không now appears fully grammaticalized within a
monoclausal structure in Modern Vietnamese, so some alternative treatment is required.

ON WHAT PROJECTS IN VIETNAMESE
31
‘It is plausible to assume that tense only marks that some arbitrary time span, for which
we keep the term TT, placed somewhere on the time axis, and that either ASS) or,
depending on the particular illocution, some other “modality marker” assigns a special
function to this time span. So, TT can be the time span for which a claim is made, but it
can also be the time span at which some obligation is put into force (or in whichever way
we want to analyse the role of the imperative) [emphasis mine].’
With this in mind, observe that imperative morphemes in Vietnamese occupy a sentence-
medial position very similar to that of có: to the right of the subject, (apparently) immediately
above the predicate phrase. The examples in (32) illustrate the distribution of the affirmative
imperative markers hãy and cứ (32a,b), as well as the negative imperative đừng chớ (32c); see
also (10a) above.
a. (Các anh) hãy đọc bài này! (32)
PL PRN IMP read lesson DEM1
‘Read this text!’
b. (Anh) cứ hỏi!
PRN IMP ask
‘Go ahead. Ask!
c. (Anh) đừng nói tó!
(PRN) NEG.IMP talk loud
‘Don’t speak loudly!’
Klein’s speculation regarding Mood implies that có should be in complementary
distribution with imperative morphemes. It turns out, though. that this is only true of affirmative

ON WHAT PROJECTS IN VIETNAMESE
32
imperatives: negative imperatives, as in example (33c) below, are perfectly compatible with có,
suggesting something like the phrase-structure trees in (34):
a. *Các anh hãy có đọc bài này. (33)
PL PRN IMP ASR read lesson DEM1
‘(youpl) Read this text!’
b. *Anh cứ có hỏi.
PRN IMP ASR ask
‘Go ahead. Ask!’
c. (Anh) đừng/chớ có nói tó!
PRN NEG.IMP ASR talk loud
‘Don’t speak loudly!’
Hence, though it may be no more than coincidence,Vietnamese có finds a further parallel
with English do-support in imperatives: excluded, in most English varieties, in affirmative
contexts (%‘Do come and see me next week’), but obligatory in negative contexts (‘Don’t be an
idiot!’/*‘Be not an idiot’). Nevertheless, exactly why do-support is necessary in English negative
imperatives – but not, say, in subjunctive contexts ((3e) above – remains unclear: see Pollock
(1989), compare Henry (1995).

ON WHAT PROJECTS IN VIETNAMESE
33
(34)
Now, if it is correct that (indicative) có and (imperative) hãy occupy the same low
‘modal’ position, one might wonder about the position of root modal auxiliaries (nên, được,
phải) which also appear low in Vietnamese structure, below negation (35a), and which—in
contrast to English—are fully compatible with both temporal (35b) and aspectual auxiliaries
(35c,d):
a. Những ai yêu bóng.đá thì (*nên) không (nên) bỏ lỡ clip này. (35)
several who love football TM should NEG should miss clip DEM1
‘Those who love football, (they) should not miss this clip.’
b. Tôi (*nên) sẽ (nên) làm gì nếu bị sa.thải?
PRN should FUT should do what if PASS fire
‘What should I do if I get fired?’

ON WHAT PROJECTS IN VIETNAMESE
34
c. Lẽ ra lúc này họ (*nên) đã (nên) đi rồi.
right out when DEM PRN should ANT should go already
‘He (should) have left already.’
d. Lẽ ra lúc này mình (?nên) đang (nên) làm một thứ gì đó.
right out time DEM PRN should DUR should do one thing what DEM
‘I should have been doing something at the moment/by now.’
Once more, the simplest assumption – viz., that Mood and modals compete for the same
structural position – proves to be incorrect. This time however, the shift is in the opposite
direction: as the Googled examples in (36) attest, interrogative có appears to the left of root
modals:
a. Em có nên mua Audi A4 2010 không ạ? (36)
PRN [+Q] should buy Audi A4 2010 NEG EXCL
‘Should I buy a 2010 Audi A4?’
b. Em có nên bỏ chồng không?
PRN [+Q] should leave husband NEG
‘Should I leave my husband?’
On the continuing assumption that indicative, interrogative and imperative có instantiate
the same functional node, the position of root modals implies the existence of one final
functional position, immediately above the thematic predicate phrase. Combining all of these
distributional facts yields the extended IP-cartography in (37) (with the proviso that some
projections are mutually exclusive):

ON WHAT PROJECTS IN VIETNAMESE
35
(37)
This concludes the basic description of the Vietnamese IP-domain. Before examining
functional categories within the thematic domain—that is to say, to the right of the underlying
subject—it is worth paying some more attention to modal auxiliaries, since they exhibit two
interesting properties: first, as indicated by the examples below, they are the only functional
categories that may appear both pre- and post-verbally in indicative clauses; second, they offer
near-canonical examples of grammatical multifunctionality in the sense of Travis, Bobaljik &
Lefèbvre (1998):7
7 The quoted text appears in a grant application submitted by the authors cited: for a variety of
reasons, the idea was not subsequently developed by them.

ON WHAT PROJECTS IN VIETNAMESE
36
‘A multifunctional functional category (MFC) is one that is ‘inherently underspecified
with the unspecified properties of the host head...[where]... syntax can provide additional
information not available in the lexical entry of the item. The lexical entry encode[s] the
intersection of the uses of the item...[d]ifferent senses [of a multifunctional item] follow
from the different head positions in which it occurs (Bobaljik, Travis & Lefèvbre 1998).’
Consider, then, the paradigm for được (≈ CAN) in (38), from Duffield (1998, 1999):
a. Ông Quang được mua cái nhà. [pre-verbal = root] (38)
PRN Q. CAN buy CL house
‘Quang was allowed to buy a house.’
b. Ông Quang mua được cái nhà. [immediately post-
PRN Q. buy CAN CL house verbal = aspectual]
‘Quang has bought (was able to buy) a house.’
c. Ông Quang mua cái nhà được. [post-vp = abilitative/
PRN Q. buy CL house CAN epistemic]
‘Quang is able to buy a house/
Quang may possibly buy a house.’
The examples in (38) demonstrate how Vietnamese word order quite clearly
disambiguates different readings of modal được: where the modal appears pre-verbally, the only
available reading is deontic, whereas in immediately post-verbal position (before the object

ON WHAT PROJECTS IN VIETNAMESE
37
noun-phrase), the reading is aspectual (perfective), instead of modal.8 A point worth mentioning
in passing is that unlike post-verbal modals, which compete for a single functional slot, pre-
verbal modals are iterable: where this happens, the higher modal (further to the left) is
interpreted epistemically (epistemic>deontic), as in (39):
a. Cô ấy nên được kiếm việc. (39)
PRN DEM should obtain find job
‘She should be allowed to find a job.’
b. *Cô ấy được nên kiếm việc.
PRN DEM obtain should look-for job
(as (39 a))
What is especially interesting about post-verbal được (38b) is that its modal meaning
seems to have been almost entirely bleached in this position: as the examples in (40) illustrate,
modals that have distinct semantics in pre-verbal position are interpreted in exactly the same way
whenever they are inserted post-verbally (as asserting that a particular event actually took place):
a. Họ làm nên việc lớn. (40)
PRN do ASP job big
‘They did (made) great things.’
8 For present purposes, I set aside the post-VP occurrence of được (38c). The main justification
for this is that whereas all of the other modal auxiliaries participate in the pre-verbal/postverbal
alternation—see examples (39)-(41) below—được is the only modal auxiliary to occur after the
VP. For analyses of final được see Duffield (1998, 1999); cf. also Simpson (1999), Simpson &
Wu (2000), for discussion of a similar phenomenon in Thai.

ON WHAT PROJECTS IN VIETNAMESE
38
b. Cô ấy kiếm được việc.
PRN DEM seek ASP job
‘She found a job.’
c. Cô ấy kiếm phải việc.
PRN DEM seek must job
‘She found a job.’
At first glance, this reading of the post-verbal modal might seem to be no more than
‘actuality entailment’ in the sense of Bhatt (1999), which arises quite generally when ability
modals combine with perfective aspect. So, for example, whereas (41a) only expresses potential,
(41b) suggests that the event was in fact realized.
a. Quang is able to buy the house. (41)
b. Quang was/has been able to buy the house.
c. Quang was/has been able to buy the house (for some time), but in the end he (has)
decided against it.
On the basis of the cancellability of this inference in English (41c), Hacquard (nd., 2009)
argues that the effect is an implicature, rather than an entailment. Vietnamese, however, supports
Bhatt’s original label, since consultants are agreed, for example, that (40b) and (40c) both mean
that the woman got the job: any ambiguity in the English translation is removed by the fact that
the potential reading is expressed by a post-VP modal (compare 40c). Whatever the correct label
may be, there is a consensus that perfective—as distinct from perfect—aspect is involved. As
will be discussed directly, it is unsurprising from a cartographic perspective that perfective
aspect is associated with a position immediately to the right of the lexical verb, what—adopting
the label of Travis (2010)—I will refer to as ‘Inner Aspect.’

ON WHAT PROJECTS IN VIETNAMESE
39
3. vP-internal Functional Syntax
In this penultimate section, I turn attention to those aspects of hidden meaning first
mentioned at the outset, that are most intimately associated with the argument structure of the
thematic core: (working this time from the bottom up) (i) the boundedness introduced by specific
objects, and (ii) the causativity reading implicated by the (underlying) subject DP. The facts
presented below suggest that these semantic properties are directly mapped to functional
structure, as in (42) below. Once again, thanks to the isolating character of Vietnamese, this
mapping is perfectly transparent in particular grammatical constructions.
(42)
Let us first consider the position of post-verbal perfective (or completive) morphemes in
Vietnamese, a set that including xong (‘finish’), ra (‘out’) and hết (‘end, up’); see Phan (2013)
for more extensive discussion. For purposes of illustration, we focus on the first of these

ON WHAT PROJECTS IN VIETNAMESE
40
elements. As shown by the alternation in (43) below, completive particles may either precede
(43a) or follow (43b) the object noun-phrase (here, the bare noun sách (‘books’)). Though both
orders are permitted, the change has immediate interpretive consequences: where the object
appears to the left of the completive particle, it is obligatorily interpreted as specific (in spite of
the absence of any determiner). This constraint is also reflected in the impossibility of fronting
overtly marked non-specific indefinites such as một cái xe (‘one car’), as shown in (44):
a. Nó đã đọc xong sách rồi. (43)
PRN ANT read PTC book already
‘He has finished reading (the) books.’
b. Nó đã đọc sách xong rồi.
PRN ANT read book PTC already
‘He has finished reading the books.’
a. *Nó sửa một cái xe xong rồi. (44)
PRN fix one CL car PTC already
‘He finished fixing one car.’
b. Nó sửa xong một cái xe rồi.
PRN fix PTC one cls car already
‘He finished fixing one car.’

ON WHAT PROJECTS IN VIETNAMESE
41
These interpretive contrasts find obvious and immediate parallels with sentences in
Germanic languages involving telic particles. In Dutch, for example, as discussed in van Hout
(1996, 2000), Thrift (2003), only specific objects may precede the telic particle op (≈ Vn. hết).9
a. *Het meisje eet koekjes op. (45)
the girl eats cookies PTC
‘The girl eats up cookies.’
b. *Het meisje eet brood op.
the girl eats bread PTC
‘?The girl eats bread up.’
c. Het meisje eet het brood op.
the girl eats the bread PTC
‘The girl eats the bread up.’
In previous analyses of OBJECT-PTC word order in Germanic (e.g., Borer 1994, Van Hout
2000, Johnson 1990, Travis 2010) it has been assumed that the specific object has been raised
from its base position to the Specifier of the aspectual projection headed by the telic particle. The
fact that finite verbs in Dutch raise out of the thematic domain—Dutch being a Verb-Second
language—as well as the complications that arise from restructuring in infinitival constructions
(Wurmbrand (2001) leave it unclear whether this aspectual projection in Dutch is external to vP,
as more generally assumed, or internal, as diagrammed in (42) above. Travis (2010) argues for
9 The Dutch~Vietnamese parallelism is not complete, since Dutch does not tolerate leaving
indefinites below op (in situ) in finite clauses (*Het meisje eet op koekjes). Presumably, this is
ruled out for independent reasons, perhaps having to do with Case.

ON WHAT PROJECTS IN VIETNAMESE
42
the latter position: if this is correct, then the Vietnamese and Dutch facts are directly assimilable
to one another.
The minimal contrasts in (46) offer further evidence of the aspectual function of xong, in
demonstrating that telic particles (46b) and quantified objects (46c) play an equal and
complementary role in signalling an entailment of completion. (Example (46a), where no such
entailment is observed, reinforces the claim, made earlier, that preverbal đã is a perfect, not
perfective marker: that is, it signals inception, not completion.)
a. Nó đã viết bài, nhưng vẫn chưa xong. (46)
PRN ANT write paper (but still NEG finish)
‘He wrote (= has started writing) the paper (but he hasn’t finished it yet).’
b. Nó đã viết xong bài (*nhưng vẫn chưa xong).
PRN ANT write PTC paper (but still NEG finish)
‘He wrote (up) the paper (*but he hasn’t finished it).’
c. Nó đã viết hai bài, (*nhưng vẫn chưa xong).
PRN ANT write two paper (but still NEG finish)
‘He wrote two papers (*but he hasn’t finished them yet).’
The examples just cited (43-44, 46) thus provide prima facie support for a vP-internal
aspectual projection in Vietnamese, such as that diagrammed in (42). Given this, one question
that now arises is whether this aspectual head is the same position occupied by the post-verbal
modals exemplified in (40) above.

ON WHAT PROJECTS IN VIETNAMESE
43
This turns out not to be the case. As discussed in Phan (2013), post-verbal modals10 are
fully compatible with post-verbal xong, ra, hết etc. Upon reflection, this is perhaps unsurprising
since the two elements serve different semantic functions: as Phan points out, unlike completive
particles which are restricted to lexically atelic predicates, post-verbal modals are more correctly
described as resultative particles. The examples in (47) below show that where both elements are
projected, they appear in a fixed order (completive precedes resultative):
a. Cuối-cùng nó cũng lau (*được) xong (được) cái bàn. (47)
final PRN also wipe RES ASP RES CL table
‘He finally finished wiping down the table.’
b. Họ đã tìm (*được) ra (được) cách chữa bệnh AIDS
3p PRN seek RES out RES way treat disease AIDS
‘They have found the cure for AIDS.’
While this ‘resultative-last’ constraint is unremarkable from a cross-linguistic
perspective, see for example, Nicol (2002), it is nonetheless striking just how transparently the
span of events~situations—from anticipation (sẽ) to inception (đã) to duration (đang) to
completion (xong), to result (được)—is iconically reflected in the cartography of Vietnamese
clauses; cf. Rizzi (2002), Cinque & Rizzi (2008), Ramchand (2008), see also Haiman (1980,
1984).
10 It should be acknowleged that post-verbal resultative CAN is an areal phenomenon—see, for
example, Enfield (2001, 2005), Cheng & Sybesma (2004); also Simpson (1998), Simpson & Wu
(2000), for clause-final CAN in Thai.

ON WHAT PROJECTS IN VIETNAMESE
44
Before concluding this examination of the functional syntax of Vietnamese, it is
necessary (for the sake of completeness) to consider the realization of the higher v node
diagrammed in (42), which, at least since the work of Hale & Keyser (1993) has generally been
taken to license the ‘external’ argument, or ‘underlying subject’ argument in regular finite
clauses. The Vietnamese examples considered thus far have all involved active, non-ergative,
and inherently causative, predicates: in all of these cases, I will assume the same V-v raising
analysis that is normally assumed for English.11 But what about contexts where cause is not
inherent, and is added in the course of the syntactic derivation? In the absence of derivational
morphology, how does Vietnamese do causativization?
Vietnamese has two analytic causative constructions: an unrestricted bi-clausal
construction involving the light verbs làm and cho (literally, ‘make give’) and a mono-clausal
construction involving just the first of these bare causative predicates (that is, làm alone). The
11 An obvious difference between the two languages is that this internal verb-raising ‘skips’ the
Asp node—and any other vP-internal functional categories, in Vietnamese apparently in
violation of the Head Movement Constraint (Travis 1984). This could be viewed as a significant
problem, unless—as seems reasonable—the HMC is taken only to apply to movement steps that
are morphologically-driven. As an aside, it is plausible to think that the distributional and
interpretive differences between Cantonese post-verbal dak (Cheng & Sybesma 2004) and
Vietnamese post-verbal được can be derived according to whether V incorporates with the modal
head on its way to v (Yes in Cantonese—V-R-v, No in Vietnamese V-v); see Duffield & Phan
(2011).

ON WHAT PROJECTS IN VIETNAMESE
45
examples in (48) and (49),12 which feature the reciprocal anaphor nhau as DP2 subject, offer
direct evidence of this bi-clausal vs. monoclausal contrast: cf. Kwon (1994). Example (48b)
shows that làm cho patterns with muốn (‘want’) in selecting a clausal complement: here, the DP1
subject antecedent họ is unable to bind nhau, while the lower verb khóc (‘cry’) can be modified
by a modal phải. Conversely, in (49b), the lower subject under làm may be bound by họ; on the
other hand, the lower predicate must now be bare. ((49c) shows that the re-introduction of cho
blocks A-binding.)
a. Họ muốn [là [John/*nhau sẽ sống hạnh phúc]] (48)
they want COMP John/e.o. FUT live happy
‘They want John/each other to live happily.’
b. Họ làm [(cho) [ John/*nhau phải khóc]].
they make give John/e.o. must cry
‘They made John/Each other have to cry.’
a. ?Họ nhìn thấy nhau làm việc. (49)
they see e.o. do work
‘They saw each other working.’
b. Họ làm nhau (*phải) khóc.
they make each other must cry
‘They made each other cry.’
12 The data in (48) and (49) are originally due to Tue Trinh (pers. comm.).

ON WHAT PROJECTS IN VIETNAMESE
46
c. *Họ làm [cho [nhau khóc]]
they make give e.o. cry
‘They made each other cry.’
Postponing any further analysis of the bi-clausal construction, consider now the
possibility that simple làm is a pure lexicalization of (causative) v in (42), with DP1 in its
specifier, as in (50).
(50)
This analysis immediately yields the prediction that transitive and strongly unergative
predicates cannot be embedded under simple làm, since the projection needed to license their
(DP2) subjects is already filled by DP1 (underlyingly). As the examples in (51), from Duffield
(2011), show, this prediction is fully borne out: in direct contrast to parallel cases where cho is
inserted, all of these examples are markedly unacceptable:

ON WHAT PROJECTS IN VIETNAMESE
47
a. *Tôi làm [đứa con gái giúp anh ấy]. (51)
I make CL CL girl help PRN DEM
‘I make the girl help him.’
b. *Tôi làm [đứa con gái nhảy].
I make CL CL girl dance
‘I make the girl dance.’
c. *Tôi làm [đứa con gái hát].
I make CL CL girl sing
‘I make the girl sing.’
d. *Tôi làm [đứa con gái ngủ].
I make CL CL girl sleep
‘I make the girl sleep.’
At the other end of the thematic spectrum, where the sole argument of the lower predicate
is a pure (change of state) Theme, simple làm causatives are fully grammatical: in such cases, as
shown in (52) and (53), an alternation is observed between a (preferred) post-verbal position for
DP2 and a marginal—though still acceptable—pre-verbal position:13
a. Tôi làm gẫy cái que. (52)
I make break CL stick
13 In this respect, Vietnamese diverges from its neighbors: Chinese shows no unergative-
unaccusative split in causatives: Thai, Lao, Khmer, and Burmese disallow inversion quite
generally.

ON WHAT PROJECTS IN VIETNAMESE
48
‘I broke the stick.’
b. Tôi làm rách tờ giấy.
I make torn CL paper
‘I tore the sheet of paper.’
a. ?Tôi làm cái que gẫy. (53)
I make CL stick break
‘I broke the stick.’
b. ?Tôi làm tờ giấy rách.
I make sheet paper torn
‘I tore the sheet of paper.’
In principle, there are a number of different ways of deriving this alternation: it might be
that the object shifts within the lower VP, as in (54.i) below; alternatively, the object might be
base-generated in {Spec,VP}, while the verb optionally raises to Asp (54.ii); or, given the
possibility of a resultative phrase below Asp, transitive predicates may generally raise this far
(54.iii) (with optional raising of DP2 to{Spec, Result} or {Spec, Asp} (54.iii). Other analytic
combinations cannot be excluded.

ON WHAT PROJECTS IN VIETNAMESE
49
(54)
(i) (ii)
(iii)

ON WHAT PROJECTS IN VIETNAMESE
50
However this alternation affecting Themes is ultimately analyzed, all of the analytic
options presented above offer a position for a DP2 argument ‘halfway’ between an Agent (the
argument of a CAUSE head) and a Theme (the argument of a root predicate/root+RESULT in
(54.iii)): if CAUSE/V can license its own argument, then there is no conceptual reason that ASP
should not be able to do the same. In fact, this is precisely what is argued in Travis (2000, 2010),
on the basis of the Malagasy data in (55) and (56) below, involving the transitivizing prefix –ha-
(Phillips 2001).
a. mijery ‘to look at’ ~ mahajery ‘to notice’ (55)
b. mandinika ‘to examine’ ~ mahadinika ‘to remark’
Travis and Phillips observe that this Malagasy prefix participates in two kinds of
derivational alternation: when attached inside an already transitive atelic stem such as mi-jery
(‘look at’) (55a), its effect is to alter the meaning to that of a bounded event (where the subject is
construed as an Experiencer instead of an Agent). By contrast, when –ha– is added to an
intransitive predicate such as tsara (‘beautiful’) the result is a transitive construction whose
subject is necessarily non-agentive—an ‘inadvertent cause’ (Duffield 2011, 2014a); hence,
voninkano (‘flowers’) (56b), but not Rabe (56c) (man’s name), can make the house beautiful.
a. Tsara ny trano. [from Travis (2000)] (56)
beautiful the house
‘The house is beautiful.’
b. Maha-tsara ny trano ny voninkano.
PRES.a.ha.beautiful the house the flowers
‘The flowers make the house beautiful.’
(literally, ‘..beautified the house’)

ON WHAT PROJECTS IN VIETNAMESE
51
c. *Maha-tsara ny trano Rabe.
PRES.a.ha.beautiful the house Rabe
‘Rabe make the house beautiful.’
If the same functional projection that licenses inadvertent causes in Malagasy is active in
Vietnamese, then it could be expected that the arguments of unaccusative (and other weakly
agentive) predicates would occur in this position: in simple làm causatives, unaccusative DP2
subjects should be licensed, but only pre-verbally—in contrast to the Theme DP2s in (52). Once
more, this prediction is directly borne out, as evidenced by the contrast between examples in (57)
and (58) below, which work in just the opposite direction from those in (52) and (53) above:
a. ??Tôi làm ngã thang-be. (57)
I make fall boy
‘I made the boy fall.’
b. ??Tôi làm biến-mất thang-be.
I make disappear boy
‘I made the boy disappear.’
a. Tôi làm thang-be ngã. (58)
I make boy fall
‘I made the boy fall.’
b. Tôi làm thang-be biến-mất.
I make boy disappear
‘I made the boy disappear.’

ON WHAT PROJECTS IN VIETNAMESE
52
Notice further that it is possible to embed an unergative predicate such as nhảy (‘dance’)
under a làm causative just as long as its DP2 subject is construed as non-agentive/non-volitional:
thus, though one cannot make someone dance or sing (*51b, c), it is quite possible to make
someone cry or laugh (non-volitionally) (59a,b), or even to make a puppet or doll dance (59c).
Once again, as required by the phrasal architecture in (60) below, the non-volitional argument
may only appear pre-verbally (assuming a bar on argument lowering).
a. Tôi làm đứa con trai khóc. (59)
I make CL CL male cry
‘I made the boy cry.’
b. Tôi làm đứa con trai cười.
I make CL CL male laugh
‘I made the boy laugh.’
c. ?Tôi làm con búp-bé nhảy/hát.
I make CL puppet dance/sing
‘I make the puppet dance/sing.’

ON WHAT PROJECTS IN VIETNAMESE
53
(60)
Finally it may be observed that the structure in (60) permits an explanation of the fact that
làm causatives are also acceptable where both DP1 and DP2 are construed non-agentively: the
grammaticality of examples such as those in (61)—which proved an embarassment for previous
analyses, in which only {Spec, IAsp} was available (Duffield 2011)—can now be handled if
{Spec, Res} is also allowed to license arguments, as in (62):14
14 It will not escape close attention that làm lexicalizes v in the trees above, whereas in (62)
below, it lexicalizes Asp. This is not as ad hoc as it may at first appear, since Vietnamese
causatives typically have a much more indvertent character than English analytic causatives.

ON WHAT PROJECTS IN VIETNAMESE
54
a. Con giớ làm thang-be ngã. (61)
CLS wind make CLS boy fall
‘The wind blew the boy over.’
b. Cái chuyện đó làm thang-be cười.
cls story dem make boy laugh
‘The story made the boy laugh.’
(62)
This inadvertence is also holds true of Thai causatives—see Vichit-Vadakan (1976)—and may
be a more widespread areal phenomenon.

ON WHAT PROJECTS IN VIETNAMESE
55
4. Conclusion
The principal goal of this paper has been to provide an observationally adequate description of
Vietnamese phrase-structure. By systematically plotting the distributions and varying
interpretation of functional categories in Vietnamese, all of which are expressed as free
morphemes, it has been possible to construct a detailed cartography of the ‘IP-domain’ above the
thematic predicate phrase, as well as to chart the functional structure to the right of the lexical
verb.
Two properties of the description are especially noteworthy. First, from a crosslinguistic
perspective, it is remarkable how directly the order of free grammatical morphemes in
Vietnamese aligns with functional projections established on the basis of diverse languages with
richer inflection. The more theoretically interesting point concerns the prolific multifunctionality
of grammatical morphemes in Vietnamese: with very few exceptions, each of the functional
morphemes discussed above15 can lexicalize more than one—usually several distinct—syntactic
position(s); when they do, their interpretation is significantly altered. It should be clear that
multifunctionality is a phenomenon, not a theory: though I have argued elsewhere that this
phenomenon is best handled in terms of ‘Anti-Projection’ (Duffield 2014c), that is evidently not
the only conceivable approach. Whatever tack is taken, however, it should be one that meets the
empirical challenges set by the data presented here.
15 There are many others besides these, including especially the wh~NPI~universal quantifier
elements ai, gì, etc., as well as the copula~complementizer~topic marker là. See Duffield
(2013b, 2014c) for further discussion.

ON WHAT PROJECTS IN VIETNAMESE
56
Acknowledgements
This paper summarizes the results of research carried out over nearly twenty years, and
presented in two invited talks given at CamCos3 (University of Cambridge), and TEAL9
(Nantes). This work would have been impossible without the invaluable help, advice and support
of Vietnamese native speakers in Montreal and elsewhere. My gratitude extends to my
professional colleagues, most especially to my co-author on several papers, Trang Phan, and to
Tue Trinh, both of whose suggestions and criticisms have significantly improved my
understanding of this language. As noted in the text, this paper has been extensively revised in
response to audience reaction and to reviewers' comments: I am very grateful to them for their
time and expertise, and to the editors for their forebearance.

ON WHAT PROJECTS IN VIETNAMESE
57
References
Alves, Mark. 1999. What's so Chinese about Vietnamese? In Papers from the ninth annual
meeting of the South East Asian Linguistics Society, edited by Graham Thurgood.
Berkeley: University of California.
Baker, Mark. 1997. Thematic Roles and Syntactic Structure. In Elements of grammar: handbook
in generative syntax, edited by Liliane M. V. Haegeman. Dordrecht, Boston: Kluwer
Academic Publishers.
Baunaz, Lena & Lander, Eric. To appear. Nanosyntax: the basics. In Lena Baunaz, Karen De
Clercq, Liliane Haegeman & Eric Lander (eds.), Exploring nanosyntax. Oxford: Oxford
University Press.
Bhatt, Rajesh. 1999. Covert Modality in Non-Finite Contexts. Ph.D. Thesis, University of
Pennsylvania. [2006. Mouton De Gruyter: Berlin].
Borer, Hagit. 1994. The projection of arguments. ms, University of Massachusetts, Amherst.
Caha, Pavel. 2009. The Nanosyntax of Case. PhD, University of Tromsø.
Cao, Xuân Hạo. 2003. Tiếng Việt. Mấy Vấn đề Ngữ Âm, Ngữ Pháp, Ngữ Nghĩa. Đà Nẵng: Nhà
xuất bản Giáo dục.
Cheng, Lisa L. S., & Sybesma, Rint. 2004. Postverbal ‘can’ in Cantonese (and Hakka) and
Agree. Lingua 114: 419-445.
Cinque, Guglielmo & Rizzi, Luigi. 2008. The cartography of syntactic structures. In CISCL
Working Papers on Language and Cognition, ed. V. Moscati, V. Vol. 2, 43-59. MITWPL
Copley, Bridget, and Fabienne Martin. 2014. Introduction. In Causation in grammatical
structures, edited by Bridget Copley and Fabienne Martin, 1-. Oxford: Oxford University
Press.

ON WHAT PROJECTS IN VIETNAMESE
58
Desclés, Jean Pierre. 2016. A cognitive and conceptual approach to tense and aspect markers. In
Aspectuality and temporality: descriptive and theoretical issues, ed. Guentchéva, Zlatka,
27-62. Amsterdam & Philadelphia: John Benjamins Publishing Company.
Chomsky, Noam. 1965. Aspects of The Theory of Syntax. Cambridge, MA: MIT Press.
Chomsky, Noam. 1989. Some Notes on Economy of Derivation and Representation. In MIT
Working Papers in Linguistics #10, edited by Itziar Laka and Anoop Mahajan, 43-74.
MIT.
Duffield, Nigel. 1998. Auxiliary Placement & Interpretation in Vietnamese. In M. C. Gruber, D.
Higgins, K. S. Olson and T. Wysocki (eds.) Proceedings of the 34th regional meeting of
the Chicago Linguistic Society, 34, 95-109.
Duffield, Nigel. 1999. Final Modals, Adverbs and Antisymmetry in Vietnamese. Revue
québécoise de linguistique 27: 92-129.
Duffield, Nigel. 2007. Aspects of Vietnamese Clausal Structure: separating tense from assertion.
Linguistics 45: 765-814.
Duffield, Nigel. 2011. On Unaccusativity in Vietnamese and the Representation of Inadvertent
Cause. In R. Folli and C. Ulbrich (editors), Researching interfaces in linguistics, 78-95.
Oxford & Cambridge, MA: Oxford University Press.
Duffield, Nigel. 2013a. 'Head-First: On the head-initiality of Vietnamese clauses. In eds. Daniel
Hole & Elisabeth Löbel, Linguistics of Vietnamese: an international survey, 127-155.
Berlin: de Gruyter Mouton.
Duffield, Nigel. 2013b. On Polarity Emphasis, Assertion and Mood in Vietnamese and English.
Lingua 137, 248-270.

ON WHAT PROJECTS IN VIETNAMESE
59
Duffield, Nigel. 2014a. Inadvertent Cause and the Unergative-unaccusative Split in Vietnamese
and English. In eds. Bridget Copley & Fabienne Martin, Causation in Grammatical
Structure, Oxford and Cambridge, MA: Oxford University Press.
Duffield, Nigel 2014b. Minimalism and Semantic Syntax: Interpreting Multifunctionality in
Vietnamese. In International conference on the linguistics of Vietnam in the context of
renovation and integration, 1090-1113. Hanoi. [LingBuzz/001919].
Duffield, Nigel 2014c. Shake Can Well. Talk presented in CamCos3, Cambridge.
[Lingbuzz/002119]
Duffield, Nigel & Trang Phan. 2011. What do Chinese L2 learners know about ‘Inner Aspect’
and Unaccusativity in Vietnamese: an experimental psycholinguistics approach. In
Nguyen Hong Con et al. (ed.) Proceedings of the international conference on linguistics
training and research in Vietnam, 388-412. Hanoi: Vietnam National University Press.
Enfield, Nicholas J. 2001. Areal grammaticalization of postverbal 'acquire' in mainland
Southeast Asia. In Proceedings of the eleventh meeting of the Southeast Asian Linguistics
Society.
Hacquard, Valentine. 2009. On the interaction of aspect and modal auxiliaries. Linguistics and
Philosophy 32, 279-312.
Hacquard, Valentine n.d. Actuality Entailments. ms. University of Maryland.
Hale, Kenneth, and Samuel J. Keyser. 1993. On argument structure and the lexical expression of
syntactic relations. In The view from Building 20: essays in linguistics in honor of Sylvain
Bromberger, eds. Kenneth Hale and Samuel J. Keyser. Cambridge, Massachusetts: MIT
Press.

ON WHAT PROJECTS IN VIETNAMESE
60
Henry, Alison. 1995. Dialect variation and parameter-setting: a study of Belfast English and
Standard English. Oxford: Oxford University Press.
Klein, Wolfgang. 1994. Time in language. London: Routledge.
Klein, Wolfgang. 1998. Assertion and Finiteness. In Issues in the theory of language acquisition.
essays in honor of Jürgen Weissenborn, eds. Norbert Dittmar and Zvi Penner, 222-245.
Bern: Lang.
Kwon, Nayoung. 2004. A semantic and syntactic analysis of Vietnamese causatives. Paper
presented at the Western Conference on Linguistics, UC, San Diego.
Johnson, Kyle. 1991. Object Positions. Natural Language and Linguistic Theory 9: 577-636.
Law, Paul. 2015. Negation-Final Yes-No Questions in Vietnamese. Conference Presentation
TEAL 9, University of Nantes.
Li, Mei. 1999. Negation in Chinese. PhD dissertation, University of Manchester.
Lin, Jonah W. 2005. Time in a language without tense: the case of Chinese”, Journal of
Semantics 23: 1-53.
Lo-Cicero, Minh Ha. 2001. L'expression du temps et de l'aspect du verbe vietnamien. The
Review of Vietnamese Studies 1:1-12.
Matthews, Stephen. 1990. A Cognitive Approach to the Typology of Verbal Aspect. PhD,
University of Southern California, Los Angeles.
Miestamo, M. & Van der Auwera, J. (2011). Negation and perfective vs. Imperfective aspect,
Chronos 22: 65-84.
Nguyễn, Đình Hoa. 1997. Vietnamese. Amsterdam/Philadelphia: John Benjamins Publishing
Company.
Nguyễn Đức Dân. 1998. Logích và Tiếng Việt: Nhà Xuất Bản Giáo Dục.

ON WHAT PROJECTS IN VIETNAMESE
61
Phan, Tran. 2013. Syntax of Vietnamese Aspect. PhD dissertation, University of Sheffield.
Phan, Trang & Duffield, N. 2016. A Nanosyntax Account of the Negation Constraint on the
Perfect Auxiliary in Vietnamese. Conference paper, Chronos 12, Caen University.
Phillips, Vivianne. 2001. The interactions between prefix and root: the case of maha in
Malagasy. In Formal Issues in Austronesian linguistics, edited by Vivianne Phillips,
Ileana Paul and Lisa Travis. Dordrecht: Kluwer Academic Publishers.
Pollock, Jean-Yves. 1989. Verb-Movement, UG and the Structure of IP. Linguistic Inquiry
20:365-424.
Ramchand, Gillian. 2008. Verb Meaning and the Lexicon: A First-Phase Syntax. Cambridge:
Cambridge University Press.
Rizzi, Luigi. 2002. Locality and the Left Periphery. In The Structure of CP and IP: The
Cartography of Syntactic Structures, eds. Adriana Belletti and Luigi Rizzi. Oxford:
Oxford University Press.
Roberts, Ian. 1993. Verbs and Diachronic Syntax. Dordrecht: Kluwer Academic Publishers.
Simpson, Andrew. 1998. Focus, Presupposition and light predicate raising in S.E. Asian. SOAS
Working Papers in Linguistics & Phonetics 8: 91-122.
Simpson, Andrew, and Wu, Xiu-Zhi Zoe. 2000. IP-raising, tone sandhi and the creation of
particles: evidence for PF-movement/cyclic Spellout. ms., SOAS/USC.
Starke, Michal. 2009. Nanosyntax: A short primer to a new approach to language. In:
Svenonius, Peter, Ramchand, Gillian, Starke, Michal & Taraldsen, Knut (eds.). Nordlyd
36.1. Special issue on Nanosyntax, 1–6. Tromsø: CASTL.
Trần, Kim Phượng. 2008. Ngữ pháp tiếng Việt, những vấn đề về thời và thể. Hà Nội: Nhà xuất
bản Giáo dục.

ON WHAT PROJECTS IN VIETNAMESE
62
Thrift, Erica. 2003. Object-drop in the L1 Acquisition of Dutch. Utrecht: LOT.
Travis, Lisa. 1984. Parameters and Effects of Word Order Variation. PhD dissertation, MIT.
Travis, Lisa. 2000. Event structure in syntax. In Events as grammatical objects, ed. by Carol
Tenny and James Pustejovsky, 145-87: Stanford: CSLI Publications.
Travis, Lisa. 2010. Inner Aspect: the articulation of VP. Dordrecht: Springer.
Trinh, Tue. 2005. Aspect of Clause Structure in Vietnamese. MA, Humboldt University.
Van Hout, Angeliek. 1996. Event Semantics of Verb Frame Alternations: A Case Study of Dutch
and its Acquisition. Ph.D. Dissertation, Tilburg University. Published 1998. New York:
Garland Publishing.
Van Hout, Angeliek. 2000. Event semantics in the lexicon-syntax interface: Verb frame
alternations in Dutch and their acquisition. In Carol Tenny and James Pustejovsky (eds).
Events as grammatical objects, 239-282. Stanford: CSLI Publications.
Vichit-Vadakan, Rasami. 1976. The Concept of Inadvertence in Thai Periphrastic Causative
Constructions. Syntax and Semantics, ed. by Masayoshi Shibatani, 459-76. New York:
Academic Press.
Whaley, Lindsay J. 1998. Introduction to typology: the unity and diversity of languages.
Thousand Oaks, CA: Sage Publications.
Wurmbrand, Susi. 2001/2003. Infinitives: restructuring and clause structure. Berlin: Mouton de
Gruyter.