D-linkability of Mandarin wh-nominals: A morphosyntactic account

Abstract

To appear in Proceedings of NACCL-36

Full text

D-Linkability of Mandarin Wh-Nominals:
A Morphosyntactic Account
Fangning Ren and Qiuhao Charles Yan
University of Cambridge & Queen Mary University of London
This paper centres around two major issues: (i) examining the correlations between D-
linkability of Mandarin wh-nominals and their clausal distributions, and (ii) investigating
the relationship between the D-linkability of these wh-nominals and their morphosyntactic
structures. Our analysis distinguishes three types of wh-nominals: inherently D-linked
which-complexes, ambiguously D-linked what-complexes, and stubbornly non-D-linked
what-simplex. We demonstrate that, ‘D-linking’ is a precondition of wh-fronting rather
than the result. Therefore, we argue, in the vein of a split-DP hypothesis, that wh-nominals
in Mandarin Chinese spell out different functional heads within the DP structure, and that
the D-linkability of these wh-nominals depends on whether the wh-item is realised at the D
head.
1 Introduction
It has been well known since Pesetsky 1987 that a characteristic that differentiates which-
phrases from others is D(iscourse)-linking. As Pesetsky argues, which-phrases are aggres-
sively D-linked, picking an entity or a subset of entities from a set of alternatives known
to each interlocutor by the time the speech occurs. By contrast, canonical wh-phrases like
what or who are non-D-linked, which can be asked out of the blue.
The idea of D-linking has some precursors as well as descendants in the literature. For
instance, Katz and Postal (1964) claim that what or who questions an indefinitely marked
domain, while which questions a definitely marked one. Kuroda (1969) argues that, by
raising a which-question, the speaker expects an answer with a definite nominal expres-
sion, whereas no such expectation is readily available for what-questions. Erteschik-Shir
(2007) proposes that the corresponding element in an answer to the D-linked wh-phrases
is construed as a restrictive focus. It is fair to say that the majority of literature focuses on
the interpretational differences among these wh-phrases, and attention has not been paid
enough to the interaction between the morphosyntactic structures of these phrases and D-
linking (with the exception of, e.g., Hirose 2003 and Boeckx and Grohmann 2004, among
many others).
In this paper we pursue an account for the correlation between D-linkability of wh-
nominals and their distributions in a wh-in-situ language, namely, Mandarin Chinese. We
1

claim that wh-phrases bearing a D-linking reading are not restricted to which-phrases in
Mandarin Chinese. Instead, we make a distinction among three types of Mandarin wh-
nominals in terms of D-linkability (in the sense of Pesetsky 1987, 2000 and Dayal 2017),
as illustrated in (1).
(1) Three types of Mandarin wh-nominals w.r.t. D-linkability
a. Which-complexes: Inherently D-linked, e.g., na-(yi)-ben shu ‘which book’
b. What-complexes: Ambiguously D-linked, e.g., shenme shu ‘what book(s)’
c. What-simplex: Stubbornly non-D-linked, e.g., shenme ‘what’
By relating wh-nominals’ D-linking status (i.e., whether they are D-linked or not) with their
distributions (i.e., whether they stay in situ or move to the clausal periphery), we argue that
wh-words in Mandarin Chinese spell out different functional heads within the DP structure,
and that the D-linking status of wh-nominals derives from the D head’s interaction with
other functional heads, namely, whether D participates in a head movement chain. We
propose that, to license overt wh-fronting, wh-nominals must be D-linked in their base
position. That is, D-linking is the precondition of wh-fronting rather than the result.1
The paper is organised as follows. Section 2 shows the distinct wh-nominals’ D-
linkabilities and their correlations with wh-fronting. Section 3 compares which-complexes
with their what-counterparts. Section 4 outlines our proposal and account for the varying
D-linking status of wh-nominals. Section 5 further discusses D-linkability as a precondition
of wh-fronting. Section 6 concludes the paper with two questions left for future research.
2 D-linkability and wh-fronting
2.1 In-situ wh-nominals
The default way of forming an information-seeking question in Mandarin Chinese is spelling
out the wh-nominals in the base-generated, argument position.2The three types of wh-
nominals are able to form wh-in-situ questions, as shown in (2).
(2) a. Ni
2SG
xihuan
like
na-(yi)-ben
which-one-CLF
shu?
book
‘Which book do you like (among the presupposed alternatives)?’
b. Ni
2SG
xihuan
like
shenme
what
shu?
book
‘What (kind of) books do you like (among the presupposed alternatives)?’
1A similar idea has been taken by Pan 2007, 2011 for wh-topicalisation in Mandarin Chinese.
2There are lots of ways of licensing in-situ wh-nominals proposed in the literature. See Pan 2019 for a
concrete literature review. As to more detailed discussions, see Huang 1982, Cheng 1991, Aoun and Li 1993,
and Tsai 1994, among many others.
2

c. Ni
2SG
xihuan
like
shenme?
what
‘What do you like?’
The felicity of these questions varies under different contexts. In-situ which-complexes like
(2a) are systematically D-linked, thereby cannot be licensed by the out-of-the-blue context
(3c). The in-situ what-simplex like (2c), by contrast, is always non-D-linked, leading to the
infelicity of its occurrence under (3a) and (3b). Interestingly, in-situ what-complexes like
(2b) are ambiguous between a D-linked reading and a non-D-linked one, as evidenced by
the felicitous results observed across all three contexts given in (3).
(3) a. Context 1: D-linking
Zhangsan and Lisi, two book lovers, went to a bookstore together. Zhangsan
asked Lisi . . .
→(2a): felicitous ✓; (2b): felicitous ✓; (2c): infelicitous ✗
b. Context 2: D-linking
Zhangsan: ‘I bought three books for my friend today: A Tale of Two Cities,
Pride and Prejudice, and The Old Man and the Sea.’ Lisi asked Zhangsan . . .
→(2a): felicitous ✓; (2b): felicitous ✓; (2c): infelicitous ✗
c. Context 3: Out-of-the-blue
Zhangsan and Lisi, two strangers, met each other for the first time in their
shared kitchen. Zhangsan asked Lisi . ..
→(2a): infelicitous ✗; (2b): felicitous ✓; (2c): felicitous ✓
In view of the above observations, we generalise as follows.
(4) Mandarin wh-nominals can appear in situ, regardless of their D-linking status.
2.2 Ex-situ wh-nominals
Another way of forming questions in Mandarin Chinese is fronting the wh-nominals to
the clausal left periphery (5), though there exists some restriction with respect to their
interpretations (see also Pan 2007, 2011).
(5) a. [Na-(yi)-ben
which-one-CLF
shu]i
book
ni
2SG
xihuan
like
ti?
‘Which book do you like (among the presupposed alternatives)?’
→felicitous ✓in (3a) and (3b); infelicitous ✗in (3c)
b. [Shenme
what
shu]i
book
ni
2SG
xihuan
like
ti?
‘What (kind of) books do you like (among the presupposed alternatives)?’
3

→felicitous ✓in (3a) and (3b); infelicitous ✗in (3c)
c. *Shenmei
what
ni
2SG
xihuan
like
ti?3
‘What do you like?’
→infelicitous ✗in (3a), (3b) and (3c)
(5a) indicates that the D-linking status of which-complexes remains unchanged even
when fronted to the clause-initial position, as demonstrated by the identical licensing con-
texts observed when these which-complexes are in situ. What-complexes become un-
ambiguously D-linked when they occur clause-initially, as evidenced by (5b). However,
fronting what-simplex overtly to the clausal left periphery is barely acceptable (5c). Un-
surprisingly, they are infelicitous in all the given contexts. Based on what are observed in
(5), the second generalisation we derive is given in (6).
(6) Ex-situ wh-nominals in Mandarin Chinese are fixed with a D-linked reading.
2.3 Wh-fronting as ¯
A-movement
What is the syntactic nature of the reordering of wh-phrases? The diagnostics showing
¯
A-properties, including, but not limited to, parasitic gaps (7), ATB movement (8), and
reconstruction effects (9), all point to a conviction that overt wh-fronting in Mandarin Chi-
nese is ¯
A-movement. We are agnostic about the feature in the left periphery that triggers
wh-fronting in this study. We refer readers who are interested in this question to the more
relevant discussions in Ren 2023, 2024 and citations therein.
(7) Parasitic gaps
a. [Na-(yi)-ben
which-one-CLF
shu]i
book
Zhangsan
Zhangsan
mai-le
buy-PFV
tizhihou
after
meiyou
NEG
kan-guo
read-EXP
ti?
‘Which book (among many) hasn’t Zhangsan read after buying it?’
b. [Shenme
what
shu]i
book
Zhangsan
Zhangsan
mai-le
buy-PFV
tizhihou
after
meiyou
NEG
kan-guo
read-EXP
ti?
‘What book(s) (among many) hasn’t Zhangsan read after buying it?’
(8) A(cross)-T(he)-B(oard) movement
a. [Na-(yi)-ge
which-one-CLF
huati]i
topic
Zhangsan
Zhangsan
taolun-guo
discuss-EXP
tiLisi
Lisi
mei
NEG
taolun-guo
discuss-EXP
ti?
‘Which topic (among many) has Zhangsan discussed but Lisi not discussed?’
3Shenme ‘what’ in this sentence is only licensed to have the exclamative interpretation (e.g., ‘It is not right
that ...’), rather than an interrogative reading (Yang, 2021).
4

b. [Shenme
what
huati]i
topic
Zhangsan
Zhangsan
taolun-guo
discuss-EXP
tiLisi
Lisi
mei
NEG
taolun-guo
discuss-EXP
ti?
‘What topic(s) (among many) has Zhangsan discussed but Lisi not discussed?’
(9) Partial/total reconstruction
a. [Zijij/k
self
de
POSS
na-ben
which-CLF
shu]i
book
Zhangsanj
Zhangsan
xiang-zhidao
want-know
[tiLisik
Lisi
du-le
read-PFV
ti]?
‘Which of hisj/k books does Zhangsanjwonder that Lisikread?’
b. [Zijij/k
self
de
POSS
shenme
what
shu]i
book
Zhangsanj
Zhangsan
xiang-zhidao
want-know
[tiLisik
Lisi
du-le
read-PFV
ti]?
‘What hisj/k books does Zhangsanjwonder that Lisikread?’
2.4 Interim summary
The observations above are summarised in (10).4Which-complexes are constantly D-linked,
and felicitous only under the D-linking contexts. What-complexes are quite flexible with
respect to their D-linking status while staying in situ. However, they become exclusively
D-linked once getting fronted. The what-simplex always behaves non-D-linkably, thereby
prohibited to occur clause-initially.
(10)
D-link.+in situ O.o.t.B+in situ D-link.+ex situ O.o.t.B+ex situ
which-com. ✓ ✗ ✓ ✗
what-com. ✓ ✓ ✓ ✗
what-sim. ✗ ✓ *✗*✗
We hypothesise that wh-nominals in Mandarin Chinese are distributed along a D-linkability
spectrum. At one extreme are which-complexes, and at the other is the what-simplex, with
what-complexes positioned in between.
3 Comparing na with shenme
In this section we make three comparisons between na involved in which-complexes and
shenme involved in what-complexes and the simplex. Coming up, it will be shown that
although both na ‘which’ and shenme ‘what’ are treated as wh-items, their morphosyntactic
properties are essentially different. This further suggests that these wh-items in Mandarin
Chinese receive separate analyses concerning their D-linkabilities, which we will turn to in
Section 4.
4For the sake of space, ‘out-of-the-blue’ is abbreviated to ‘O.o.t.B’, ‘complexes’ to ‘com.’, and ‘simplex’
to ‘sim.’.
5

3.1 Dependence on functional elements and interpretations
The first distinction between na ‘which’ and shenme ‘what’ lies in their dependence on
other functional elements within the nominal structure as well as the corresponding inter-
pretations. As shown in (11), na ‘which’ cannot stand alone as an independent constituent.
It must co-exist with a classifier but optionally with a numeral (when the number is ‘one’) or
an NP. Since there is a rich inventory of classifiers in Mandarin Chinese, some of them are
considered informative enough to denote an entity independently. When the NP is phonet-
ically null, it is almost always the case that the set of referents denoted by the which-(one)-
CLF combo is visually perceived by the interlocutors. Note that the individual-denoting
classifier used in the which-complexes, as given in (11), predetermines that the answer
to the question must denote an individual. Consequently, a kind-denoting nominal like
xiaoshuo ‘novel’ is ruled out as a valid answer.
(11) Q: Ni
2SG
xihuan
like
kan
read
na*(-(yi)-ben)
which-one-CLF
(shu)?
book
‘Which book do you like to read?’
A: Wo
1SG
xihuan
like
kan
read
Zhanzheng-Yu-Heping
war-and-peace
/ *xiaoshuo.
novel
Individual-denoting reading: ‘I like to read War and Peace (among a set of
alternatives).’
*Kind-denoting reading: ‘I like to read novels (in comparison with other
genres of books).’
As illustrated in (12), shenme ‘what’ can directly select an NP and form a constituent
without appealing to other functional elements, unlike na ‘which’ discussed above. Such
a constituent results in an interpretation equivalent to either what type/kind of NP or which
NP, corresponding to a kind-denoting or individual-denoting reading, respectively. This is
because shenme ‘what’ forms a linkage to the previous discourse by either setting a criterion
of classification (known to the interlocutors) of the set of NPs which it quantifies over (Kiss,
1993), or picking out a member among a set of pre-established alternative individuals.
(12) Q: Ni
2SG
xihuan
like
kan
read
shenme
what
shu?
book
‘What books/kind of books do you like to read?’
A: Wo
1SG
xihuan
like
kan
read
Zhangzheng-Yu-Heping
war-and-peace
/ xiaoshuo.
novel
Individual-denoting reading: ‘I like to read War and Peace (among a set of
alternatives).’
Kind-denoting reading: ‘I like to read novels (in comparison with other gen-
res of books).’
Conversely, the what-simplex can be free-standing, as demonstrated in (13). Interpretation-
6

wise, since it does not bear a linkage to the discourse, the bare what in the question can be
resolved, by default, by an NP with a generic reading.
(13) Q: Ni
2SG
xihuan
like
kan
read
shenme?
what
‘What do you like to read?’
A: Wo
1SG
xihuan
like
kan
read
shu
book
/ manhua
manga
/ baozhi.
newspaper
Generic reading: ‘I like to read books/manga/newspaper.’
3.2 Parallel to demonstratives
The second distinction is illustrated by comparing the wh-items to canonical demonstra-
tives. The Num-Cl sequence can be sandwiched by the wh-item and the nominal in which-
complexes (14b) but not in what-complexes (14c). This suggests that which-complexes
function similarly to demonstratives, as demonstrated in (14a).
(14) a. zhe/na
this/that
(yi)/ji
one/several
ben
CLF
shu
book
‘this/that/these/those books’ Dem Num Cl N
b. na
which
(yi)/ji
one/several
ben
CLF
shu
book
‘which book(s)’ Which Num Cl N
c. *shenme
what
(yi)/ji
one/seveal
ben
CLF
shu
book
Int.: ‘what book(s)’ *What Num Cl N
Additionally, both demonstratives and which can be combined with the plurality marker
xie, as shown in (15a) and (15b). This property, however, does not apply to the what-
complex (15c). This is probably because the kind-denoting nature of what-complexes is
conceptually plural by default (Dayal, 2004). Adding the plurality marker to what, in this
sense, seems redundant and unnecessary, thereby leading to a clash.
(15) a. zhe/na-xie
this/that-PL
shu
book
‘these/those books’ Dem-PL N
b. na-xie
which-PL
shu
book
‘which books’ Which-PL N
c. *shenme-xie
what-PL
shu
book
7

Int.: ‘what books’ *What-PL N
3.3 Local ordering restrictions
The third aspect for comparison regards the positions of the wh-words along the nominal
extended projection. Within the DP structure, the hierarchical position of na ‘which’ is
always higher than other adnominal elements such as the numeral, the classifier, and the
adjective, as exemplified in (16). It turns out that na ‘which’ consistently acts as an ‘edger’
of nominal phrases in Mandarin Chinese, similar to its role in many other languages dis-
cussed by Cinque (2005) and Giusti (2018).
(16) a. na
which
yi
one
jian
CLF
bai
white
chenshan
shirt
‘which piece of white shirt’ Which Num Cl A N
b. *yi
one
jian
CLF
na
which
bai
white
chenshan
shirt
Int.: ‘which piece of white shirt’ *Num Cl Which A N
c. *yi
one
jian
CLF
bai-de
white-DE
na
which
chenshan
shirt
Int.: ‘which piece of white shirt’ *Num Cl A Which N
As we previously mentioned, the Shenme-Num-Cl-N sequence is never allowed, whereas
Num-Cl-Shenme-N is perfectly fine, as demonstrated in (17). Moreover, it is worth noting
that, when both shenme ‘what’ and the adjective are hierarchically lower than the numeral
and the classifier, the order between them seems to be flexible, as illustrated in (17b) and
(17c).
(17) a. *shenme
what
yi
one
jian
CLF
bai
white
chenshan
shirt
Int.: ‘what piece of white shirt’ *What Num Cl A N
b. yi
one
jian
CLF
shenme
what
bai
white
chenshan
shirt
‘what piece of white shirt’ Num Cl What A N
c. yi
one
jian
CLF
bai-de
white-DE
shenme
what
chenshan
shirt
‘which piece of white shirt’ Num Cl A What N
The relative orders between these two wh-items and the Num-Cl sequence is pretty telling:
shenme ‘what’ would be merged hierarchically lower than na ‘which’, as shown in (18).
The two must be positioned on opposite sides of the Num-Cl sequence: na ‘which’ must
precede the sequence, while shenme ‘what’ must follow it.
8

(18) a. na
which
yi
one
ben
CLF
shenme
what
shu
book
‘which book of what kind’ Which Num Cl What N
b. *shenme
what
yi
one
ben
CLF
na
which
shu
book
Int.: ‘which book of what kind’ *What Num Cl Which N
4 A morphosyntactic account for D-linkability
4.1 Theoretical assumptions
In what follows, we propose a morphosyntactic account for wh-nominals with respect to
their distinct D-linkabilities. To this end, we rely on two key theoretical underpinnings.
The first is the assumption that the DP structure, where the wh-phrases are realised, is split
by the presence of an nP layer (Guardiano, 2012; Roberts, 2017). According to Roberts
(2017), a similar bi-phasal structure exists in the nominal domain, analogous to what has
been proposed for the clausal domain (Chomsky, 2000, 2001). This parallel is shown in
(19).
(19) a. Clause structure (Chomsky, 2000, 2001)
[CP C[TP T[vPv[VP V]]]]
b. Nominal structure (Roberts, 2017)
[DP D[QP Q[nPn[NP N]]]]
A further point made by Roberts is that C and D are external phase heads interacting with
the higher categories or discourse contexts. On the other hand, vand nare ‘inward-looking’
internal phase heads that assign functional properties to the lexical roots. A similar split-
DP structure is proposed by Guardiano (2012), without resting on the parallel to the clause
structure as well as phasehood.
The second assumption is that the bi-phasal nominal structure also contains a tree-
splitting function that affects the interpretation of nominals. It was observed that English
bare plural subjects of stage-level predicates are ambiguous between a generic and an exis-
tential reading. Those of individual-level predicates, by contrast, only receive a generic in-
terpretation. This being the case, Diesing (1992) argued that the asymmetry between these
predicates regarding subject interpretation boils down to different mappings between the
syntactic structures and the logical representations. Splitting the IP into two distinct zones
at LF, Diesing mapped VP into the ‘nuclear scope’, and the remaining IP-level elements
into the ‘restrictive scope’. She further partitioned the generic reading from the existential
reading by positing that the semantics of the latter consists of a restrictive set defined by a
restrictive clause and a nuclear scope. The semantics of the former, however, requires only
a nuclear scope. Thus, anything remaining in the VP has an existential reading, while those
involved in the IP-level subtree can have a generic one. The distinction between the two
9

types of predicates can then be explained by manipulating the syntactic positions of the sub-
jects at LF. Subjects of the stage-level predicates can take the position of either Spec,IP or
Spec,VP, and those of the individual-level predicates only occupy Spec,VP.5This analysis
forms the basic mechanism of what Diesing referred to as the Mapping Hypothesis.6
In this study, we leverage the second hypothesis in the nominal domain from two per-
spectives. First, we postulate that the DP structure can be divided into an nP and a DP-level
subtree, with nP resisting D-linking. Second, the ambiguous D-linking status of what-
complexes arises from their flexibility in either remaining in nP or moving to DP. The split-
DP structure we argue for is illustrated in (20). Following Cheng and Sybesma (1999),
we posit that Num and Cl are individual functional heads along the nominal extended pro-
jection, being sandwiched between D and n. For ease of exposition, all the specifiers are
omitted, but structurally they are still in the tree.
(20) DP
D
Num
Cl nP
nNP
4.2 Analysis
We propose that the D-linkability of wh-nominals depends on whether spelling out the wh-
items at the D head. Starting with which-complexes, recall that na ‘which’ has to precede
all the other functional elements in the DP structure. Thus, the D head seems to be the only
place to spell out na ‘which’, as shown in (21). Such a configuration offers an account for
na’s parallel to demonstratives and its consistent D-linking status simultaneously.
5The terminologies used by Diesing (1992) are easily translatable into more modern ones like Spec,TP,
Spec,vP, or Spec,VoiceP. Here we ignore any nuanced differences between them.
6The interpretational distinctions between strong and weak determiners can also be accounted for in the
same vein. See also Berman 1991.
10

(21) DP
D
na
‘which’ Num
Cl nP
nNP
As to what-complexes, we argue that both nand D are able to spell out shenme ‘what’.
Spelling out shenme ‘what’ at nis supported by the fact that it can be positioned between
the Num-Cl sequence and an NP. When nrealises shenme as in (22), the what-complex is
non-D-linked, since the Num-Cl-Shenme-N sequence is prohibited from being fronted to
the left periphery, as exemplified in (23).
(22) DP
D
Num
Cl nP
n
shenme
‘what’
NP
(23) *[Yi-ben
one-CLF
shenme
what
shu]i
book
Zhangsan
Zhangsan
xiang
want
mai
buy
ti?
Int.: ‘One token of what book does Zhangsan want to buy?’
Alternatively, when the what-complex is unambiguously D-linked, we posit that shenme
‘what’ undergoes head movement from nto D, and gets spelled out at D as with na ‘which’,
as demonstrated in (24). This also captures the fact that the Shenme-Num-Cl-N order is
11

ruled out, as the separate spell-out of the Num-Cl sequence would violate the Head Move-
ment Constraint (Travis, 1984).
(24) DP
D
shenme
‘what’ Num
Cl nP
nNP
Turning to the what-simplex, we argue that N head-moves and adjoins to n, where they
together spell out the bare shenme ‘what’, as given in (25). Since the what-simplex is ex-
poned by N-n, it cannot undergo further head movement to D. That is, the head movement
in the lower nP-level subtree is unable to reach other functional heads in the higher DP-level
subtree. Thus, the what-simplex is stubbornly non-D-linked. Empirically we think that our
proposal is reasonable, since it helps rule out the ungrammatical Shenme-A sequence, e.g.,
*shenme bai ‘*what white’. Theoretically we admit that we cannot give a more satisfactory
account at the moment for the reason why N-nstops moving higher to the DP-level subtree.
We leave this for further research.
(25) DP
D
Num
Cl nP
N+n
shenme
‘what’
NP
12

5 D-linkability as a precondition of wh-fronting
Let’s simply comment on the relationship between D-linkability of wh-nominals and wh-
fronting. As mentioned in Section 2.2, only D-linked wh-nominals are able to be fronted to
the left periphery. The fact suggests that there must exist some relation between these two
phenomena. That is, either wh-fronting feeds D-linking, or D-linking feeds wh-fronting.
The former relation ‘wh-fronting ≺D-linking’ has several puzzling issues. Why is
what-simplex never able to be fronted in the first place? How do we explain the fact that
which and what-complexes can be D-linked while staying in situ? What is the point of
triggering wh-fronting to the left periphery if the target is already D-linked?
The approach we pursue in this work strongly favours the latter relation ‘D-linking ≺
wh-fronting’. We provide an adequate account for the varying D-linking status of the in-
situ wh-nominals. D-linkability, in our sense, is best regarded as the licensing condition
of fronting wh-nominals. This idea successfully predicts the distributions of wh-nominals
with respect to their D-linkabilities.
6 Closing remarks
In this paper we have looked through D-linkabilities of three types of wh-nominals in Man-
darin Chinese, with a particular focus on their relations with overt wh-fronting as well
as their morphosyntactic structures. We have demonstrated that wh-nominals behave dif-
ferently with respect to D-linkability, that is, complex wh-phrases are able to receive a
D-linking interpretation, contrasting with the what-simplex which canonically resists a
linkage to the discourse. This further speaks to our observation that the D-linkability of
wh-nominals correlates with their frontability: only which- and what-complexes can un-
dergo fronting, whereas the what-simplex cannot. We argue that being D-linked requires
the D head in the nominal structure realises the wh-items. Therefore, the D-linking status
of wh-nominals whose D-linking property is dynamic (i.e., the what-complexes) can be
fixed through morphosyntactic maneuvers within their internal DP structures (i.e., undergo
head movement from nto D).
We leave two relevant remarks for future research. One of them concerns the D-linking
status of wh-phrases other than which and what in Mandarin, e.g., shei ‘who’. As far as
we can tell, shei, which looks like a wh-simplex on the surface, can undergo wh-fronting
as well. Our account predicts that the internal structure of shei ‘who’ is in effect more
similar to the what-complexes. The other remark is related to the licensing conditions of
wh-fronting. So far we have only known that being D-linked is one condition for Mandarin
wh-nominals to undergo overt movement to the left periphery. It is worth studying what
other conditions could be, in particular at the syntax-semantics interfaces.
13

References
Aoun, Joseph, and Y.-H. Audrey Li. 1993. Wh-elements in situ: Syntax or LF? Linguistic
Inquiry 24:199–238.
Berman, Stephen. 1991. On the semantics and logical form of wh-clauses. Doctoral dis-
sertation, University of Massachusetts Amherst.
Boeckx, Cedric, and Kleanthes K. Grohmann. 2004. SubMove: Towards a unified account
of scrambling and D-linking. In Peripheries: Syntactic edges and their effects, ed. by
David Adger, Cécile De Cat, and George Tsoulas, 241–257. Dordrecht: Kluwer.
Cheng, Lisa L.-S. 1991. On the typology of wh-questions. Doctoral dissertation, MIT.
Cheng, Lisa L.-S., and Rint Sybesma. 1999. Bare and not-so-bare nouns and the structure
of NP. Linguistic Inquiry 30:509–542.
Chomsky, Noam. 2000. Minimalist inquiries: The framework. In Step by step: Essays on
minimalist syntax in honor of Howard Lasnik, ed. by Roger Martin, David Michaels, and
Juan Uriagereka, 89–155. Cambridge, MA: MIT Press.
Chomsky, Noam. 2001. Derivation by phase. In Ken Hale: A life in language, ed. by
Michael Kenstowicz, 1–52. Cambridge, MA: MIT Press.
Cinque, Guglielmo. 2005. Deriving Greenberg’s Universal 20 and its exceptions. Linguistic
Inquiry 36:315–332.
Dayal, Veneeta. 2004. Number marking and (in)definiteness in kind terms. Linguistics and
Philosophy 27:393–450.
Dayal, Veneeta. 2017. Deconstructing the D in D-linking. Talk presented at DP60 work-
shop, MIT.
Diesing, Molly. 1992. Indefinites. Cambridge, MA: MIT Press.
Erteschik-Shir, Nomi. 2007. Information structure: The syntax-discourse interface. Ox-
ford: Oxford University Press.
Giusti, Giuliana. 2018. Demonstratives as arguments and modifiers of N. In Atypical
demonstratives: Syntax, semantics and pragmatics, ed. by Marco Coniglio, Andrew
Murphy, Eva Schlachter, and Tonjes Veenstra, 23–55. Berlin: de Gruyter.
Guardiano, Cristina. 2012. Demonstratives, word order and the DP between syntax and
semantics: Crosslinguistic research. Studies in Greek Linguistics 32:100–115.
Hirose, Tomio. 2003. The syntax of D-linking. Linguistic Inquiry 34:499–506.
14

Huang, C.-T. James. 1982. Logical relations in Chinese and the theory of grammar. Doc-
toral dissertation, MIT.
Katz, Jerrold J., and Paul M. Postal. 1964. An integrated theory of linguistic descriptions.
Cambridge, MA: MIT Press.
Kiss, Katalin É. 1993. Wh-movement and specificity. Natural Language & Linguistic
Theory 11:85–120.
Kuroda, Sige-Yuki. 1969. Attachment transformations. In Modern studies in English:
Readings in transformational grammar, ed. by David A. Reibel and Sanford A. Schane,
331–351. Englewood Cliffs, NJ: Prentice-Hall.
Pan, Victor Junnan. 2007. Interrogation and quantification: The role and the function of
the interrogative particles and elements in Mandarin Chinese. Doctoral dissertation,
University of Nantes, Nantes.
Pan, Victor Junnan. 2011. ATB-topicalization in Mandarin Chinese: An intersective oper-
ator analysis. Linguistic Analysis 37:231–272.
Pan, Victor Junnan. 2019. System repairing strategy at interface: Wh-in-situ in Mandarin
Chinese. In Interfaces in grammar, ed. by Jianhua Hu and Haihua Pan, 133–166. Ams-
terdam: John Benjamins.
Pesetsky, David. 1987. Wh-in-situ: Movement and unselective binding. In The represen-
tation of (in)definiteness, ed. by Eric J. Reuland and Alice G. B. ter Meulen, 98–129.
Cambridge, MA: MIT Press.
Pesetsky, David. 2000. Phrasal movement and its kin. Cambridge, MA: MIT Press.
Ren, Fangning. 2023. Mandarin overt wh-fronting as focus movement. Proceedings of the
Linguistic Society of America 8:5491.
Ren, Fangning. 2024. Mirativity and embeddability: The case of Mandarin Chinese. In
Proceedings of the 25th Seoul International Conference on Generative Grammar, ed.
by Tae Sik Kim, 213–226. Seoul: The Korean Generative Grammar Circle & Dongguk
University.
Roberts, Ian. 2017. The Final-over-Final Condition in DP: Universal 20 and the nature
of demonstratives. In The Final-over-Final Condition: A syntactic universal, ed. by
Michelle Sheehan, Theresa Biberauer, Ian Roberts, and Anders Holmberg, 151–185.
Cambridge, MA: MIT Press.
Travis, Lisa deMena. 1984. Parameters and effects of word order variation. Doctoral
dissertation, MIT.
15

Tsai, Wei-Tien Dylan. 1994. On economizing the theory of A-bar dependencies. Doctoral
dissertation, MIT.
Yang, Barry C.-Y. 2021. Two types of peripheral adjunct WHATs. Concentric 47:61–92.
16