Chinese wh conditionals

LANGUAGE AND LINGUISTICS 124753-798 2011

2011-0-012-004-000322-1

Do Chinese Wh-conditionals Have Relatives in Other Languages

Qiong-peng Luo1 and Stephen Crain2

Nanjing University1 Macquarie University2

Previous studies on Chinese wh-conditionals leave several issues unresolved These studies fail to capture the definiteness effects in wh-conditionals they fail to address apparent violations of the novelty condition and they overlook some genuine readings We propose to resolve these issues by analyzing wh-conditionals as a special instance of topic-comment structures with an identity relation between pairs of referring expressions Chinese wh-conditionals can refer to either a single situation or to multiple situations leading either to a definite interpretation or to a generic interpretation respectively The choice between these interpretations is determined by the contextpragmatics Typologically Chinese wh-conditionals are linked to correlatives both semantically and structurally Key words Chinese wh-conditionals correlatives uniqueness free relatives

1 Introduction

The semantics of wh-indefinites and wh-conditionals in Chinese has been a topic for Chinese linguistics for several decades In this study we offer a novel analysis of wh-conditionals by setting the phenomenon in a broader typological context A typical wh-conditionalcalled lsquobare conditionalsrsquo in Cheng amp Huang (1996)always contains a pair of matching wh-phrases one in the antecedent clause and the other in the consequent clause1 These conditionals have a matching requirement namely the wh-

This research was conducted as part of the Australian Research Council Centre of Excellence

in Cognition and its Disorders (CE110001021) It was funded in part by two ARC Discovery Project grants (DP1096160 DP0879842) and by a grant from the National Philosophy and Social Science Foundation of China (09BYY022) We would like to thank two anonymous reviewers of Language and Linguistics for insightful comments on an earlier version of the current draft Parts of the paper were presented at Sinn und Bedeutung 15 (September 2010 Saarbruumlcken Germany) and we are indebted to members of that audience for useful comments Tom Roeper Klaus von Heusinger and Zhiguo Xie We have also benefitted from numerous discussions with members of the language acquisition lab at Macquarie University Rozz

Qiong-peng Luo and Stephen Crain

754

phrases in the antecedent and consequent clauses must be identical in number form and reference In the present paper we add the further observation that Chinese wh-conditionals can come with some definiteness flavor akin to free relatives in English

The most frequently cited account of Chinese wh-conditionals by Cheng amp Huang (1996) treats wh-expressions in wh-conditionals as indefinites This analysis is somewhat problematic however because it appears to violate the novelty condition which requires each indefinite to introduce a novel entity into the domain of discourse yet only the first of the pair of indefinites in wh-conditionals serves this function In part to circumvent the problem Chierchia (2000) proposes that wh-indefinites in Chinese are indefinite pronouns (ie pronominals) As such they can appear in both the antecedent and in the consequent of wh-conditionals without violating the novelty condition However Chierchiarsquos account fails to explain why wh-indefinites display Principle C effects which suggests that wh-indefinites are r-expressions rather than pronominals

We propose to circumvent these problems by making two assumptions One is that wh-indefinites encode uniqueness just as indefinites do in donkey conditionals (a Russellian semantics for uniqueness is adopted to capture the definiteness effect) The second assumption is that wh-conditionals constitute identity statementsa special kind of topic-comment structure that is not subject to the novelty condition The second assumption reconciles the tension between the novelty condition and Principle C We contend that this novel analysis receives cross-linguistic support Several similarities between Chinese wh-conditionals and correlatives are explained (i) both structures display definiteness effects (ii) both are subject to a matching requirement and (iii) both are open to definitegeneric interpretations

We offer a unified analysis for these apparently unrelated constructions that crop up in historically unrelated languages (eg Chinese vs Hindi) Lacking lexical items like Hindi bhii and English ever or an explicit conditional word like if Chinese wh-conditionals are a cross between being correlatives and conditionals On the one hand when a definiteness flavor is added wh-conditionals strongly resemble free relatives (as well as correlatives in languages such as Russian Vietnamese Hungarian etc) in that they make reference to a particular (unique) individual in a particular situation On the

Thornton Katherine Demuth Nobu Akagi Aijun Huang Anna Notley Alessio Ursini Hao Yin Min Liao Miao-ling Hsieh Peng Zhou and Likan Zhao Finally we would like to thank Hao Yin Miao-ling Hsieh and Yatin Mahajan for supplying critical judgment data

1 In Cheng amp Huang (1996) conditionals (which they call Chinese donkey sentences) are partitioned into two classes bare conditionals and ruguo-conditionals While the former requires the presence of a pair of identical wh-phrases (one in each clause) the latter only allows a pronominal-anaphor (eg overt pronouns empty categories) We focus on bare conditionals in this paper


755

other hand when the context establishes a plurality of situations the unique individuals denoted by the wh-phrase get relativized to situations and the identity of the referent is not known or not relevant This reading involves a universal quantification over situations and is semantically akin to Hindi correlatives with bhii The major insight in Cheng amp Huang (1996) is preserved intact on the current account namely that wh-conditionals are donkey conditionals and are open to a generic interpretation The cross-linguistic comparison we provide also indicates that all natural languages draw from a basic ontology of pluralities and atoms that are operative for nominal quantification though its realization is constrained by language-specific factors This comparison between Chinese wh-conditionals Hindi correlatives and English free relatives provides evidential support for the account of wh-conditionals presented in this study and sheds further light on the study of correlatives in general

The paper is structured as follows Section 2 discusses the (absolute) unique reading and definiteness effect of wh-conditionals Section 3 surveys the problems concerning the novelty condition and Principle C that arose in previous analyses in which wh-phrases were analysed as genuine indefinites A Russellian semantics for uniqueness and wh-indefinites is advanced in Section 4 and Section 5 applies this to Chinese wh-conditionals to (a) capture the uniquenessdefiniteness effects and (b) reconcile the tension between the novelty condition and Principle C Section 6 investigates the generic interpretations of wh-conditionals taking into consideration certain modal implications (ignorance indifference and free choice) Section 7 sketches a unitary account for wh-conditionals A full comparison between Chinese wh-conditionals and correlatives (with special reference to Hindi correlatives) is offered in Section 8 and Section 9 points out some residual issues

2 Chinese wh-conditionals definiteness and beyond

The seminal paper by Cheng amp Huang (1996) contains a summary of the typical properties of one kind of conditional structure in Chinese They call these lsquobare conditionalsrsquo because they lack an overt conditional word like ruguo lsquoifrsquo or an overt adverb of quantification (Cheng amp Huang 1996132) We shall refer to them as wh-conditionals Here is the summary offered by Cheng amp Huang

(1) Properties of wh-conditionals a The (donkey) anaphor must take the form of a wh-word b The (donkey) wh-word must be identical to the wh-word in the antecedent clause c There must be an element in the consequent clause referring back to the wh-word

in the antecedent clause


756

Typologically Chinese wh-conditionals are exceptional because they require an identical wh-word in both clauses (sometimes called the lsquomatching effectrsquo) Even minor variations are unacceptable Example (2) below illustrates a typical wh-conditional in Chinese while (3) illustrates violations of the matching effect

(2) Shei xian lai shei xian chi who first come who first eat Lit lsquoIf X comes first X eats firstrsquo

(3) a Shei xian lai shenme ren xian chi who first come what person first eat

b Shei xian lai tongyang de ren xian chi who first come the same DE person first eat

Cheng amp Huang (1996) analyse wh-phrases in wh-conditionals as variable-denoting indefinites that are bound in wh-conditionals by a covert universal necessity operator which provides universal quantificational force According to Cheng amp Huang example (2) has the semantic representation in (4) where lsquoxrsquo stands for the variable denoted by the wh-indefinites and the universal quantifier is the semantic spell-out of a covert necessity operator If the wh-conditional has multiple wh-indefinites these are unselectively bound by the quantificational expression as in Heim (1982) and Kamp (1981)

(4) x (x comes first x eats first) (Cheng amp Huang 1996132)

On the analysis by Cheng amp Huang (2) presumably means everybody who comes first eats first This semantics has a plurality commitment It is committed to multiple comers and eaters However intuitively (2) is true if for one particular situation say at Annrsquos birthday party there turns out to be exactly one individual who comes first and eats first Notice in this regard that if there is exactly one man in the room it is semantically odd to say something like Everybody in the room studies kangaroos even if the one person studies kangaroos Yet as Kadmon (1990) has shown donkey conditionals typically have a unique reading (eg If there is a doctor in London he is Welsh = Kadmonrsquos example (42))

The unique reading of wh-conditionals correlates with the definiteness effect The definiteness effect is illustrated in example (5) which shows that the wh-indefinite in the antecedent clause of a wh-conditional can be referentially linked to a partitive expression in the consequent clause2

2 Examples like (5) can be found in Google and we have been found them to be acceptable to

the native speakers we have consulted We thank Hao Yin and Maggie Liao for their judgments


757

(5) Shenme ban biaoxian hao shenme ban de sanfenzhiyi jiu keyi dedao what class perform well what class DE one third then can get jiangli reward3 lsquoOne third of whatever class that perform(s) well will get a rewardrsquo

Example (5) casts doubt on the idea that wh-indefinites in Chinese wh-conditionals are genuine indefinites because the wh-phrase is used as the complement of a partitive with the form lsquoNP of wh-NPrsquo According to the well-known Partitive Constraint partitive structures prohibit a singular ordinary indefinite from being used as the complement DP (cf Jackendoff 1972 Barwise amp Cooper 1981) The constraint is illustrated in (6)

(6) most of four girls two of three students none of some men each of some professors one third of some class the top ten of some class

The constraint dictates that the complement DP of a partitive must be a plural definite as in (7)

(7) all of the boys two of the students four of the gifts you gave to Mary on Christmas many of my friends one third of the class the top ten of the class that performs well

Not surprisingly ordinary partitives in Chinese show the same contrast as indicated in (8)

(8) a zhe-ge ban de yixie tongxue this-CL class DE some student lsquosome students of this classrsquo

3 Abbreviations used in this paper are as follows

1SG first person singular 3PL third person plural ACC accusative ASP aspect marker AUX auxiliary CL classifier CLpl plural classifier CLsg singular classifier DAT dative DE de DEM demonstrative ERG ergative HAB habitual IMPERF imperfective LOC locative NEG negative marker OBL oblique PASS passive PERF perfect PFV perfective REL relative TOP-PRT topic particle


758

b yi-ge ban de yixie tongxue one-CL class DE some student lsquosome student of a classrsquo

c san-ge ban de qian shiming three-CL class DE top ten lsquothe top ten of three classesrsquo

In view of the Partitive Constraint the expression shenme ban lsquowhat classrsquo must be definite in examples like (5) above Later we shall provide further evidence showing that wh-conditionals are semantically akin to free relatives (FRs) which are widely analysed as definite descriptions

Another observation by Cheng amp Huang supports our claim about definiteness They observe that the existential word you lsquohaversquo can be inserted before the wh-phrase in the antecedent clause of a ruguo-conditional (if-conditional) but not in the antecedent clause of wh-conditionals The following examples are from Cheng amp Huang (see also Chierchia 2000)

(9) a Ruguo you shei yao jian wo ni jiu jiao ta deng yi deng if have who want see me you then ask he wait a while lsquoIf somebody wants to see me please ask him to wait for me for a whilersquo

b (You) shei xian lai shei xian chi have who first come who first eat

The expression you lsquohaversquo is often analyzed as the overt realization of the existential quantifier in Chinese Cheng amp Huang attribute the oddness of (9b) to a semantic conflict between universal quantification and existential quantification Since on their analysis the wh-phrases in wh-conditionals are variables bound by a universal necessity operator examples (9b) and (10b) are both unacceptable for the same reason

(10) a You yi-ge ren zai huayuan li have one-CL person at garden LOC lsquoThere is a man in the gardenrsquo

b You mei-ge ren zai huayuan li have every-CL person at garden LOC Intended lsquoThere is every one in the gardenrsquo

(11) a There isare a man two men in the garden b There is everyman in the garden

However it should be noted that if wh-phrases are treated on a par with definite descriptions as we have argued then the oddness of (9b) is also explained Definites are


759

incompatible with the existential you as indicated in (12) and (13)

(12) There is the man in the garden

(13) You zhe-ge ren zai huayuan li have this-CL man at garden LOC lsquoThere is the man in the gardenrsquo

To summarize First Chinese wh-conditionals are open to unique interpretations (eg shei xian lai shei xian chi is true if there turns out to be exactly one individual who comes first and eats first) And second Chinese wh-conditionals exhibit definiteness effects These empirical observations suggest that just like indefinites in donkey conditionals in English wh-indefinites in Chinese wh-conditionals are unique indefinites The issue is how the universal interpretation of wh-conditionals can be captured and to what extent a unitary treatment (for both the unique reading and universal reading) is possible granted that Cheng amp Huangrsquos seminal observation holds Another puzzle is the matching effect To explain this Cheng amp Huang propose a Parallelism Constraint on Operator Binding (PCOB)

(14) Revised PCOB (Cheng amp Huang 1996139) In a tripartite structure of quantification Q [A] [B] [X1 X2 hellipXn] (where n1) are variables in A

For every variable in A there must be an identical variable in B

According to (14) the examples (3a) and (3b) (ie who comes first what man eats first and who comes first the same person eats first) are ruled out because the anaphoric elements in the consequent clauses are not lexically identical to the wh-indefinites in the antecedent Later we attempt to provide a more straightforward account of the matching effect

3 The novelty condition and Principle C

As noted earlier Cheng amp Huangrsquos analysis of wh-phrases does not conform to the novelty condition proposed by Heim (1982) Ordinary indefinites introduce novel entities into the discourse as example (15) indicates

(15) If a mani comes first a mani eats first

If wh-phrases are ordinary indefinites then each wh-phrase should introduce a novel entity to the discourse In Chinese wh-conditionals however the wh-phrase in the


760

antecedent and the one in the consequent are identical in reference In short the wh-phrase in the consequent of a conditional introduces a familiar referent rather than a novel one As Chierchia (200017) points out this makes for an inconsistent semantics for wh-indefinites

(16) a wh-words must introduce a novel variable in the antecedent of a conditional b wh-words must introduce a non-novel variable in the consequent of a conditional

To solve this paradox Chierchia proposes that wh-indefinites in Chinese are indefinite pronouns (ie pronominals) A pronominal can be used as a discourse anaphor A simple example is (17)

(17) If a mani comes first hei eats first

This explains why wh-indefinites can appear in the consequent clause of wh-conditionals without violating the novelty condition At first glance this seems to be a reasonable solution Some issues remain however First if wh-phrases in Chinese are indefinite pronouns (ie pronominals) we expect them to always introduce a familiar discourse referent in both the antecedent of a conditional and in the consequent But a wh-phrase in the antecedent of a conditional does not seem to require a linguistic antecedent On Chierchiarsquos account moreover wh-phrases are expected to be subject to Principle B (because they are pronominals) However wh-phrases in general appear to display Principle C effects Consider the following examples

(18) a Sheii shuo tai xihuan wo who said he like me lsquoWho said he likes mersquo Johni said hei likes me Peterj said hej likes me hellip

b Tai shuo sheii xihuan wo he said who like me lsquoWho did he say likes mersquo hei said Johni likes me hej said Peterj likes me hellip

Example (18b) is a case of strong crossover The contrast between (18a) and (18b) indicates that Chinese wh-phrases are r-expressions rather than pronominals The following examples adopted from Bruening amp Tran (2006) constitute further support

(19) a Tai shuo sheii xihuan wo meimei he said who like my sister lsquoWho did he say like my sisterrsquo

b Tai zongshi shuo sheii tai xihuan wo meimei he always said who heshe like my sister lsquoHei always says whoihei likes my sisterrsquo


761

c Sheii (yaoshi) shuo tai sheii xihuan wo meimei wo jiu zou ta who (if) say he who like my sister I then hit he lsquoIf somebodyi says heiwhoi likes my sister I will hit himrsquo

As the examples in (18) and (19) indicate we face a dilemma On one hand if wh-phrases are pronominals as on Chierchiarsquos proposal then we require an explanation of why wh-phrases display Principle C effects On the other hand if wh-phrases are not pronominals we need to explain why wh-phrases in the consequent clause remain anaphorically linked to wh-phrases in the antecedent

This leaves us to address the following unresolved issues 1 How to capture the uniquenessdefiniteness effect of wh-conditionals 2 How to reconcile the tension between the novelty condition and Principle C On the

one hand wh-indefinites in the consequent clause behave like pronominals but ordinary wh-indefinites display Principle C effects like r-expressions

3 What is the typological status of Chinese wh-conditionals We shall discuss 1 and 2 next In discussing 3 we demonstrate some striking similarities between Chinese wh-conditionals and correlatives (our focus will be on Hindi correla-tives) We contend that Chinese wh-conditionals are typologically linked to correlatives This is the topic of the last section of the paper

4 Indefinites and uniqueness

The definite reading of wh-conditionals is most apparent when a unique referent is being established The uniqueness effect is evident in the example in (20) where the interlocutors A and B are discussing next yearrsquos US presidential election

(20) A Shei neng huanjie shiye wo jiu xuan shei who can alleviate unemployment I then vote who Lit lsquoWho can alleviate unemployment I then vote for whomrsquo B Ni xuan shei you vote who A Wo xuan Obama I vote Obama a I will vote for everyone that can alleviate unemployment b I will vote for whoever can alleviate unemployment

In example (20) the particular situation under discussion is the presidential election


762

The use of a wh-conditional is felicitous in this situation But notice that Arsquos remark in (20) cannot mean I will vote for everyone that can alleviate unemployment Voters can cast a ballot for only one candidate So the semantic counterpart of Arsquos remark in (20) is an -ever Free Relative (as in (20b)) rather than a structural representation with a universal quantifier (as in (20a))

Free Relatives (FRs) are definite descriptions that denote maximal entities FRs can either be singular (eg the thing the babysitter tells you) or plural (eg the sum of the things that the babysitter tells you) (cf Jacobson 1995 Dayal 1997 among others) In (20A) the wh-indefinite picks up a singular individual But wh-indefinites can also pick up a maximalized individual To see this recall the example shenme ren xian lai shenme ren xian chi lsquowhat personpeople first come what personpeople first eatrsquo Suppose there are three persons Li Si Ma Liu and Zhang Qi who simultaneously arrive first at the party Then the sentence in question is true if and only if all the people that come first also eat first However not all wh-indefinites can pick up a (plural) maximalized individual just as not all wh-indefinites can be anaphorically linked to a partitive Wh-indefinites of some kinds can be such as ones of the form lsquowhich-CLPL NPrsquo (when followed by a collective noun) but other wh-indefinites cannot be anaphorically linked to a partitive such as shei lsquowhorsquo Witness the following contrast between (21) and (22)

(21) Shei xian lai shei zhongjian de sanfenzhiyi jiu keyi chi dangao who first come who among DE one-third then can eat cake Lit lsquoWho first come one third of who can eat the cakersquo

(22) Shenme ren xian lai shenme ren zhongjian de sanfenzhiyi what person first come what person among DE one third jiu keyi chi dangao then can eat cake Lit lsquoWhat person first come one third of what person then can eat the cakersquo

Whether a wh-phrase can be entered into a partitive construction (lsquoNP of which-NPrsquo) or not depends on the type of classifier (singular or plural) and the type of head noun (collective or atomic) A which-phrase with a plural classifier or a collective noun can always enters into a partitive construction

In Chinese when the classifier ge is used the whole NP picks out a singular entity and when xie is used the NP picks out a plural entity Examples (23)-(24) indicate that the distinction between singularityplurality of wh-phrases is lexicalized to some extent


763

(23) Na-ge ren xian lai na-ge ren zhongjian de sanfenzhiyi which-CLsg person first come which-CLsg person among DE one-third jiu keyi chi dangao then can eat cake

(24) Na-xie ren xian lai na-xie ren zhongjian de sanfenzhiyi which-CLpl person first come which-CLpl person among DE one-third jiu keyi chi dangao then can eat cake

Table (25) shows some basic number (singularplural) typology for wh-phrases in Chinese

(25) SingularPlural distinction and Chinese wh-phrases

wh-phrases Plural Singular Notes shei lsquowhorsquo na-ge lsquowhich-CLsgrsquo na-xie lsquowhich-CLplrsquo na-li lsquowherersquo shen-me lsquowhatrsquo Depending on the head noun and context

So the complement DP of a partitive must be a plural definite It is reasonable to assume that for Chinese the basic ontological distinction between pluralitysingularity is operative despite the fact that Chinese lacks productive number morphology in nominal quantification

It appears that Cheng amp Huangrsquos semantic representation assigned to wh-conditionals only captures half the story According to Cheng amp Huangrsquos semantic representation in (27) (26) is true in a model where more than one person comes first and eats first However Chinese speakers also judge (26) to be true in a model where there is exactly one person who comes first and eats first As we have just shown this is because shei lsquowhorsquo tends to pick out a singular entity rather than a plural one

(26) Shei xian lai shei xian chi who first come who first eat lsquoIf X comes first X eats firstrsquo

(27) x (x comes first x eats first)

The fact that (26) is true in a model where there is exactly one individual indicates that we should include uniqueness into the semantics of wh-conditionals as in (28) This amounts to saying that the definite effect of wh-conditionals is related to uniqueness


764

(28) Shei xian lai shei xian chi is true in a model where there is a unique individual who comes first and eats first

In the literature it has been reported that in certain circumstances an indefinite can have a unique reading (cf Evans 1980 Kadmon 1990 among others) Kadmon (1990) notices that the uniqueness effect is associated with singular cross-sentential anaphora

lsquosuppose I need to borrow a chair [hellip] Leif has ten identical chairs and he is willing to lend any of them You can now say [(29)] to me [hellip] In this situation the NP a chair does not refer to a unique chair [hellip] when anaphora is attempted however the uniqueness effect always shows up Consider [(30)] in the same situation and be sure that you are completely unable to distinguish any one of Leifrsquos chairs from his other chairs [hellip] Many speakers cannot use [(29)] in such a situation [hellip] [(30)] is only felicitous [hellip] [if] they are referring to a chair which is uniquely identified by some propertyrsquo (Kadmon 1990279-280)

(29) Leif has a chair (30) Leif has a chair It is in the kitchen

There are multiple ways to encode uniqueness into the semantic representation We follow Brasoveanu (2007 2008) who adopts a Russellian treatment The Russellian analysis of definites consists in presuppositions of existence maximality and singleton This is demonstrated below

(31) The chair Leif brought is wobbly

(32) x (chair (x) amp brought (x Leif) amp y (chair (y) amp brought (y Leif) y =x) amp wobbly (x))

existence maximality singleton uniqueness

In set-theoretic terms (32) can be rewritten as (33)

(33) X (X amp X = y chair(y) amp brought (y Leif) amp X=1 amp wobbly (X)) existence maximality singleton uniqueness


765

In (33) the singleton is contributed by the singular morphology The presuppositions of existence maximality and singleton give rise to a uniqueness reading We propose the wh-phrases in wh-conditionals are unique indefinites with two semantic components existence and maximality Unlike English Chinese lacks the morphology to mark number specification in nominal quantification so whether a wh-phrase denotes a singleton or a plurality is left either to the context or to the lexical semantics of wh-phrases as illustrated in (34) and (35)

A wh-phrase denotes a singleton (34) Shei xian lai

X [X amp X= y person (y) amp first_comes (y) amp X=1] existence maximality singleton uniqueness

A wh-phrase denotes a plurality (35) Na-xie ren xian lai

which-CLpl person first come X[X amp X= y person (y) amp first_comes (y) amp X1]

existence maximality plural

When the wh-phrases denote a plurality a maximality operation turns the plurality into a maximalized individual in the spirit of proposals by Link (1983) and Grosu amp Landman (1998) In lattice-theoretic terms if a and b are individuals then the sum of a and b (written as a b) is also an individual If a b and c are individuals the sum of a b and c (written as a b c) is also an individual and so on The technical details are immaterial here (see Link 1983 Grosu amp Landman 1998) For our purposes we simply propose the following definition of the maximality operation

(36) For any poset (partially order set) = ltX gt MAX () = sup(X) where a b iff a b = b

The sup(X) is the least upper bound of X Example

(37) X = a b c a a b b a b c a c ababc ac abc bc abc sup(X) = abc

The MAX operation always returns a plural individual In this sense it is still singular


766

and the wh-phrases generate a unique reading Kadmon (1990) offers examples that show that uniqueness is related to maximal collections In (38) for example the plural pronoun they refers to the maximal collection of the chairs that Leif owns

(38) Leif has four chairs They Three of them are in the kitchen (Kadmon 1990 ex (24))

This uniqueness-based account offers a straightforward explanation for the

definiteness effect in Chinese wh-conditionals Our analysis predicts that (39) is ambiguous it allows both the distributive reading and collective reading

(39) Shenme ban biaoxian hao shenme ban de sanfenzhiyi jiu keyi dedao jiangli what class perform well what class DE one-third then can get reward lsquoOne third of whatever class(es) that perform well get(s) a rewardrsquo

On the distributive reading example (39) means for each class that performs well one third of its members will be rewarded On the collective reading example (39) is compatible with the situation that for some class none of its members get rewarded while for some other classes all of the members get rewarded It should be noted however that not all Chinese wh-conditionals display this ambiguity When a different wh-phrase is used the ambiguity disappears Example (40) lacks the ambiguity

(40) Na-ge banji biaoxian hao na-ge banji de sanfenzhiyi jiu keyi which-CLsg class perform well which-CLsg class DE one-third hen can dedao jiangli get reward lsquoOne third of whichever class that performs well gets a rewardrsquo

The difference between (39) and (40) is that the expression lsquowhat classrsquo in (39) can be either singular or plural whereas lsquowhich-CLsg classrsquo in (40) can only be singular The difference is not marked by number morphology but is lexicalized

5 Identity in wh-conditionals

We have shown the definiteness effect can be captured by assuming that wh-indefinites encode uniqueness as illustrated in (41) However two difficulties remain First a wh-conditional always involves a pair of matching wh-phrases How can we account for the matching effect in Chinese wh-conditionals like (42) Second indefinites are subject to the novelty condition Why then does the novelty condition fail to apply


767

(41) Leif has a chair It is in the kitchen

(42) Shei xian lai shei xian chi who first come who first eat

We start by dealing with the novelty condition Consider example (43)

(43) An American president won a Nobel prize in economics and an American president

won a Nobel prize in peace

A natural interpretation of (43) is that there are two distinct American presidents one who won a Nobel prize in peace and the other who won a Nobel prize in economics Why cannot one and the same American president have won both prizes We propose that this is a reflection of the pragmatic Principle of Cooperation and its submaxim of informativeness First compare (43) and (44) Notice that if (44) is true then so is (43)

(44) An American president won a Nobel prize in economics and a Nobel prize in peace

What links (44) and (43) is an asymmetric entailment which we shall define as follows

(45) A proposition p asymmetrically entails a proposition q iff s (s p s q)

The definition says that p asymmetrically entails q iff whenever p is true q must be true but not vice versa It is easy to see that (44) asymmetrically entails (43)

Now suppose that exactly one man came in and sat down As in the previous examples this entails (46)

(46) A man came in and a man sat down

So if the speaker knows that if the same person performs both actions the Principle of Cooperation would compel himher to use (47) instead of (46)

(47) A man came in and sat down

When the hearer hears (46) he must assume by virtue of the Principle of Cooperation that the speaker conveys certain information that is not expressed by (47) otherwise the speaker would use the more informative (47) instead This lsquomissingrsquo information not conveyed is that the second indefinite expression in (46) introduces a novel entity into the discourse If the second indefinite expression introduces a previously mentioned


768

entity into the discourse this violates the Principle of Cooperation We derive the novelty condition using the implicature of disjointness

(48) Implicature of Disjointness lsquo[asome NP1 asome NP2 ]rsquo where NP1 and NP2 are identical in form and NP1 does not c-command NP2 is infelicitous if lsquo[Asome NP3 amp ]rsquo asymmetrically entails lsquo[Asome NP1 asome NP2 ]rsquo in the case NP3 NP1 and NP2 are identical in form

Notice however that (48) says nothing about wh-indefinites and wh-conditionals in Chinese Obviously wh-conditionals violate (48) since the wh-indefinites in the antecedent and consequent clauses are not only identical in form but also identical in reference

A straightforward solution to this dilemma is to treat wh-conditionals in Chinese as identity statements This provides us with a way to reconcile the tension between the novelty condition and Principe C Overt identity statements are immune to the novelty condition as the examples in (49) illustrate

(49) a A man who drinks alcopops is a man who gets a hangover b Cicero is Tully c That is Uncle Bob (pointing to a picture)

In (49a) the first occurrence of the indefinite lsquoa manrsquo does not c-command the second However the novelty condition does not apply and the indefinite expressions assume identity of reference We assume that in identity statements the novelty condition is overridden because there is an overt identity operation that forces expressions to pick up the same referent In other words the novelty condition is an Elsewhere Condition (EC) which applies only when it can

We already proposed that wh-indefinites encode uniqueness just as indefinites do in donkey conditionals Following Link (1983) and Grosu amp Landman (1998) we propose the antecedent wh-indefinites are subject to a -operation where stands for uniqueness

(50) Shei xian lai shei xian chi who first come who first eat The antecedent [[ shei xian lai]] = x person (x) amp first_comes (x)

A covert identity operation applies to the wh-indefinite in the consequent clause giving rise to the identity reading as in (51)


769

(51) Shei xian lai shei xian chi [who first come]x x [first eat [y [person (y) amp y=x]]]4

On the proposed account the antecedent wh-indefinite shei (xian lai) lsquothe unique person who comes firstrsquo binds the variable x by λ-abstraction The wh-anaphor is interpreted as a definite description introducing variable x that is identical to the one previously introduced in the antecedent

It has been proposed since Cooper (1979) that donkey anaphors should be interpreted as generalized D-type pronouns (cf Heim amp Kratzer 1998 Elbourne 2005) The D-type pronouns contain both a definite description and a free relation variable R which helps fix the referent of the definite description Cooper assumes R is provided by pragmatic saliency This idea has been challenged by Heim (1990) who observes that donkey anaphora is subject to a condition which she dubs the Formal Link Condition According to this condition donkey anaphors requires an explicit linguistic antecedent (eg every man who has a wife is sitting next to her vs every married man is sitting next to her)5

Chinese wh-conditionals constitute supporting evidence for the D-type pronoun analysis of donkey anaphors Instead of looking for any linguistic antecedent the wh-anaphor looks for an identical antecedent in order to fix its referent This analysis yields the correct semantics for wh-conditionals as indicated in the following derivation

(52) a the antecedent [[ shei xian lai]] = x (person (x) amp first_comes (x)) b the wh-anaphor [[ shei (xian chi)]] = y (person (y) amp y = x amp R(y)) c R x first_comes (x) d (after identity operation) the wh-anaphor [[ shei (xian chi)]] = x (person (x) amp first_comes (x)) e the consequent z first_eats(z) (x (person (x) amp first_comes (x))) = first_eats (x (person (x) amp first_comes (x))) f [[ shei xian lai shei xian chi]] = 1iff the unique individual who comes first is the unique individual who eats first

On the proposed account wh-conditionals are semantically akin to free relatives in English Despite the structural differences Chinese wh-conditionals and English free

4 BE of identity X Y [X=Y] where type (X) = type (Y) Leibnizrsquos Law of Identity lsquoX = Yrsquo is true iff every predicate that is true of X is also true of Y 5 Patel-Grosz amp Grosz (2010) distinguish strong pronouns from weak pronouns (eg German

demonstrative pronouns vs personal pronouns) According to them English donkey pronouns are parallel to referential pronouns and are contextually resolved by means of saliencyaccessibility of an antecedent in the sense of Cooper (1979) Only the strong pronouns are subject to a (strict) Formal Link Condition


770

relatives share a common semantics Attesting to this is the fact that English free relatives can be translated as identity statements (cf Moltmann 2010)6

(53) Whoever comes first eats first = the first comer is the first eater I donrsquot like whatever you bought = the thing(s) you bought is(are) the thing(s) I donrsquot like

There is independent support for this analysis The pair of wh-indefinites in Chinese wh-conditionals are not only identical in reference but they are also identical in form and number This lsquomatching requirementrsquo follows for free on the present account We assume without the copula to mark identity in Chinese wh-conditionals identity of form is a prerequisite to identity of reference Even minor variations are unacceptable as shown in (54)


b Shei xian lai tongyang de ren xian chi who first come the same Nom person first eat

Not surprisingly we find the same form-matching restriction is also operative in English identity statements Consider the following examples

(55) a A man who drinks alcopops is someone who gets a hangover b A man who drinks alcopops is the same person who gets a hangover c A man who drinks alcopops is the man who gets a hangover

To summarize the identity relation is responsible for the identity in reference between the wh-indefinites and the inapplicability of the novelty condition in wh-conditionals This identity reading renders wh-conditionals semantically akin to identity statements

This analysis faces potential problems however For one thing wh-conditionals do not resemble other identity statements which typically involve the copula7 In Chinese

6 A reviewer correctly points out that the structural differences between Chinese wh-conditionals

and English free relatives may be superficial Citko (2001) proposes that in a simple free relative like John ate what Mary cooked the single instance of what is an argument of both ate and cooked However due to Deletion under Identity the lower copy what gets deleted at PF The difference between Chinese and English viewed in this light is unlike English the two copies of the wh-indefinite must stay at PF in Chinese We thank the anonymous reviewer for bringing this to our attention

7 Again we thank one of the anonymous reviewers for this observation


771

a prototypical identity statement is marked by the copular shi as in (56)

(56) TaNei-ge ren shi Zhang San hethat-CL person be Zhang San lsquoHeThat person is Zhang Sanrsquo

However the expression shi (and its ilk including jiu-shi) is generally disallowed in wh-conditionals regardless of the position of the wh-anaphors

(57) a Shei xian lai (jiu) shi shei xian chi (subject wh-anaphor) who first come jiu be who first eat b Shei xian jinlai (jiu) shi wo xian da shei (object wh-anaphor) who first come-in jiu be I first beat who

At first glance this fact may seem to undermine our claim that wh-conditionals are identity statements But there are independent reasons why the copular shi cannot be inserted in wh-conditionals

A combination of semantic and syntactic factors conspire to prohibit shi from appearing in wh-conditionals Instead of being the marker for identity Chinese copular shi has an additional predicate-creation function Tham (2008) proposes that in equative copular sentences shi is the lexical realization of the type-shifting pred in the sense of Chierchia (1998) and Partee (1986) Its major function is to map entities to their corresponding properties Tham suggests that in equative copular sentences both the pre-copular and the post-copular arguments are referential ie of type e When shi is combined with an argument of type e this creates a predicate of type lte tgt which can be properly combined with another entity of type e To illustrate

(58) [The richest man in the city]e [is Lee Kar-shinge]ltetgt (Tham 200863) (59) Nominal Predicativization (ibid 69)

All noun phrases may combine with the copular to form a predicative VP [shi NP]VP

whose denotation is pred (NP) where pred is a function that takes referential entities to their corresponding properties

If Tham is on the right track we have an explanation for why shi cannot appear in wh-conditionals The consequent clause of wh-conditionals is propositional not of type e If so predicate-creation using shi is not needed for wh-conditionals

It is also reassuring for our analysis to observe that across languages identity statements do not require the copular For example den Dikken (2006) proposes that the copula in English performs double duties it is both a supporter of tense aspect and agreement and it is a linker However he observes that when an NP occupies a TP-


772

external topic position the copula cannot occur This happens in Hungarian The following example is den Dikkenrsquos example (95) in Chapter 5

(60) a (van) Janos (van) a legjobb diak (van) is Janos is the best student is

b (van) a legjobb diak (van) Janos (van) is the best student is Janos is

As (60) indicates the third-person singular present-tense form lsquobersquo mdash that is van mdash may not surface no matter where it is placed and regardless of the relative order of subject and predicate den Dikken attributes this to the fact that unlike English Hungarian is discourse-prominent such that the prime determinants of word order are information-structural properties Like Hungarian Chinese is a discourse-prominent language In a topic-comment construction when the pragmaticsdiscourse is clear about the relationship between the topic and comment the copular shi is not needed even if two referring expressions have the same reference Here is an illustration

(61) Situation Two students are talking about a particular individual Zhang San a Zhang San (a) ta xian chi Zhang San TOP-PRT he first eat lsquoAs for Zhang San he first eatsrsquo

b Zhang San (a) shi ta xian chi Zhang San TOP-PRT be he first eat

c Zhang San (a) jiu SHI ta xian chi Zhang San TOP-PRT jiu be he first eat

Example (61) indicates that when the referential link between Zhang San and ta lsquohersquo is clear from the context ta refers back to Zhang San and the copular shi is not required If shi is inserted an additional focus effect is introduced and the copula is stressed In most such cases jiushi rather than shi is preferred

The present considerations have encouraged us to seek a novel syntactic analysis for wh-conditionals Based on a comparison of structures across languages we have reached the tentative conclusion suggesting that wh-conditionals are topic-comment structures in which the antecedent wh-indefinite is the topic and the comment is included in the consequent clause8 We believe this analysis is on the right track for several reasons First treating wh-conditionals as topic-comment structures permits us to derive the identity in reference between the wh-indefinites thereby resolving the 8 This may also help explain the source of uniqueness ie topical elements are always definite in

some sense We leave this issue open in this study


773

tension between the novelty condition and Principle C Second the analysis provides a motivated explanation for the matching requirement in wh-conditionals On this analysis the wh-indefinite antecedents are topics and indefinites cannot be marked as topics (due to wh-morphology) lsquothe intuitive idea is that topic-comment sequencing presupposes that the comment is about the topic It requires hellip every topical discourse referent introduced in the topic updated to be picked up by an anaphoric element in the comment updatersquo (Bittner 2001) This also proves to be the case for wh-conditionals ie the anaphoric wh-phrase cannot be omitted

(62) a Shei xian lai (shei) xian chi who first come who first eat

b Shei xian lai wo da (shei) who first come I beat who

This move to treat wh-conditionals as topic-comment structures is welcomed by recent crosslinguistic research Recently it has been frequently proposed that conditionals in general are topic-comment constructions (cf Lewis 1973 Bittner 2001 Schlenker 2004 Ebert Endriss amp Hinterwimmer 2008 among others) By treating wh-conditionals as topic-comment structures we are able to link wh-conditionals with correlatives which are widely assumed to be topic-comment structures (Bittner 2001) Correlatives are composed of a left-adjoined relative clause that contains a wh-indefinite and a main clause that contains a demonstrative pronoun Like Chinese wh-conditionals the wh-indefinites and the demonstrative pronoun (a strong pronoun) are identical in reference and a copula cannot be inserted between the relative clause and the main clause in correlatives We postpone the remainder of our typological discussion to the final section9

9 Analyzing wh-conditionals as topic-comment structures also explain the asymmetry between

the identity statements lsquoX is Yrsquo and lsquoY is Xrsquo As a reviewer points out if identity conforms to Leibnizrsquos Law then lsquoX is Yrsquo should be logically equivalent to lsquoY is Xrsquo However this does not seem to be the case for wh-conditionals The following two sentences are not judged to be semantically identical

(i) a Shei xian lai shei xian chi b = Shei xian chi shei xian lai We think however that the asymmetry is due to discoursepragmatic factors of topic-comment structures Despite identity in reference there is a temporal ordering between topical and comment expressions The topical elements are always given information and the comments provide new information Conceptually new information follows given information Any juxtaposition of these therefore leads to a conceptual conflict in information structure as illustrated by the asymmetry between (ia) and (ib)


774

6 On the generic interpretation of wh-conditionals

So far we have looked at cases in which wh-conditionals like shei xian lai shei xian chi (lsquowho comes first who eats firstrsquo) are used to refer to a particular situation In this case wh-conditionals have the absolute unique reading in the sense of Kadmon (1990) However wh-conditionals are also open to a generic interpretation For example shei xian lai shei xian chi could also refer to a general phenomenon for each case the first comer is the first eater and for different cases there are (presumably) different individuals involved In this section we propose that wh-conditionals have additional features when used as general statements and that these added features support our claim that wh-conditionals should be treated on a par with FRs 61 The ignorance implication

Wh-conditionals have ignorance implications (ie the speakeragent expresses epistemic uncertainty about the identity of the referent denoted by the wh-phrase or more plainly the speakeragent does not know who has the property P) This is depicted in (63)

(63) a Shei xian lai shei xian chi who first come who first eat b The person who comes first eats first but I donrsquot know who will be the one that

comes first

Two linguistic tests reveal this ignorance reading (a) the namely test and (b) the copular test First consider the namely test (cf Dayal 1997) When a phrase meaning lsquonamelyrsquo (which is to identify the referent denoted by the wh-phrase) is being inserted in a wh-conditional the sentence becomes odd as shown by (64)

(64) Shei xian lai (wo zhidao shi Zhang San) shei xian chi who first come I know be ZS who first eat

In (64) lsquowo zhidao shi Zhang Sanrsquo (lsquoI know it is Zhang Sanrsquo) is used to identify the referent denoted by the wh-phrase (64) is bad The copular test also indicates wh-conditionals have some ignorance implication (cf Condoravdi 2005)

(65) a Huangdi xuanze-le na-ge nuumlhai na-ge nuumlhai jiu hen buxing emperor choose-ASP which-CL girl which-CL girl then very unlucky lsquoWhichever girlwhoever that is chosen by the emperor is very unluckyrsquo


775

b Huangdi xuanze-le na-ge nuumlhai na-ge nuumlhai jiu shi Anni emperor choose-ASP which-CL girl which-CL girl then be Annie Lit lsquoWhichever girlwhoever that is chosen by the emperor is Anniersquo

This ignorance implication is reminiscent of that of English singular -ever FRs (66) are from Dayal (1997) and (67) are adopted from Condoravdi (2005)

(66) a Whatever Mary is cooking namely ratatouille has tons of unions b Whatever Mary is cooking be it ratatouille latkes or goulash has tons of onions

(67) a The book Mary likes was Language Acquisition b Whatever book Mary likes was Language Acquisition c Whatever book Mary likes was certainly not Language Acquisition

62 The indifference implication

Wh-conditionals also have implications of indifference That is the speakeragent does not care who has the property under consideration Example (68a) reveals such an implication as represented in (68b)

(68) a Shei zuihou lai wanhui shei xi wan who last come party who wash dish

b the person who arrives last for the party washes the dishes Counterfactual implication it could be anyone else that washed the dishes if he was the last person for the party

Example (69) is suggestive of the semantic similarity between wh-conditionals and -ever Free Relatives in English

(69) a Whoever arrived last for the party washes the dishes

b the person who arrived last for the party washes the dishes Counterfactual implication it could be anyone else that washed the dishes if he was the last person for the party

63 Free choice implication

Wh-conditionals also have some free choice implication under certain circumstance Consider the following example Context the university requires 50 credits for a bachelorrsquos degree and Mary has


776

already got 47 credits To fulfill the universityrsquos requirement Mary has to get 3 more credits There are three courses Mary can register for for this purpose Each course has 3 credits The following sentence is felicitous in this situation

(70) Ni xuan na-men kecheng na-men kecheng jiu keyi rang ni biye you choose which-CL course which-CL course then can let you graduate lsquoWhichever course you take can let you graduatersquo

Again English -ever FRs also have this semantic feature

(71) a Any course you take can let you fulfill the universityrsquos requirement = b Every course you (may) take can let you fulfill the universityrsquos requirement

To summarize wh-conditionals carry modal implications as listed in (72)

(72) a ignorance the speakeragent does not know the referent denoted by the wh- phrase

b indifference the speakeragent does not care about the referent denoted by the wh-phrase

c free choice there is freedom of choice over the domain of individuals denoted by the wh-phrase

64 Deriving the genericuniversal interpretation

Although wh-indefinites still introduce some uniqueexactly one individual it seems clear that this unique individual is being relativized to worldssituations in this case The truth condition of (68a) can be illustrated by (73) for instance

(73) For any party the speaker believes the following holds If Zhang San arrives last for the party Zhang San washes the dishes Li Si arrives last for the party Li Si washes the dishes Wang Wu arrives last for the party Wang Wu washes the dishes Ma Liu arrives last for the party Ma Liu washes the dishes Zhang San and Li Si arrive last for the party Zhang San and Li Si wash the dishes hellip

This closely aligns wh-conditionals with -ever Free Relatives (FRs) in English Dayal (1997) argues that -ever FRs in English always involve some universal quantification over identity alternatives to the worlds of evaluation She proposes that a typical -ever FRs in English have the following semantics10

10 von Fintel (2000) proposes a slightly different analysis for -ever FRs He notices that some


777

(74) Whatever Mary is cooking uses onions

(75) a Whateverj [IP[ hellip tjhellip]] = Qif (w)(s)[Q(i) (x P(x)(i))] where P is the property derived by abstracting away xj in the IP denotation b f(w)(s) = wrsquo p[s believes p(w) p(wrsquo)] for a world of evaluation w and speaker s f(w)(s) is the set of worlds in which the speakerrsquos beliefs about w hold c A world wrsquo f(w)(s) is an identity alternative iff there exists a wrsquorsquo f(w)(s) such that x P(wrsquo)(x) x P(wrsquorsquo)(x)

According to Dayal FRs denote the set of properties that their referents have in any relevant world Without tense or aspect the world w remains free The condition (75b) says that modal base represents the speakerrsquos belief about w Assuming that the free relative is used felicitously every world in the set will have a unique referent for the free relative This follows from the standard presupposition associated with definites The condition (75c) makes explicit the notion of identity alternatives (i-alternatives) It characterizes a world as an alternative iff it can be distinguished from another world solely on the basis of the denotation of the free relatives This is the condition about non-Rigidity and variation on the referent In this semantics example (76a) below is semantically represented as (76b)

(76) a Mary is cooking something Whatever she is cooking uses onions (Dayal 1997 ex (27)) b i-Altf(w)(s) [uses-onions (i) (x [cooking (i)(x)(m)])] c i-alt1 x[cooking (i)(x)(m)] = ratatouille i-alt2 x[cooking (i)(x)(m)] = lentils i-alt3 x[cooking (i)(x)(m)] = goulash hellip

(76b) says that as far as the speakerrsquos belief is concerned the dish being cooked by

-ever FRs carry indifference implications while some carry ignorance implications To consider the following example (von Fintelrsquos ex (17))

(ii) a I grabbed whatever tool was handy b I grabbed the tool that was handy and if a different tool had been handy I would

have grabbed that According to von Fintel what kind of implication an -ever FRs can carry depends on the modal base when the modal base is epistemic variation results in the ignorance reading a counterfactual modal base yields indifference von Fintel leaves it open how the modal base is linguistically introduced Chinese wh-conditionals may provide a good suggestion Despite the semantic commonalities between wh-conditionals and FRs wh-conditionals have some CONDITIONAL form which can be interpreted either epistemically or counterfactually A further bridge between conditionals and free relatives thus is being established We have to leave this issue to others and future works


778

Mary uses onions in all relevant i-alternatives (for each world there is a uniqueexactly one dish that Mary cooks) The speaker believes the assertion to be held regardless of the identity of the dish Let the main clause denote Q and the FR denote P where both P and Q are properties Property Q holds of the entity by virtue of its meeting description P and the relationship between P and Q is essential ie not accidental It does not allow an interpretation where the assertion is based on a particular entity as witnessed by ungrammaticality of (77)

(77) Whatever Mary is cooking namely ratatouille uses onions (ibid ex (29))

We adopt this same semantics for Chinese wh-conditionals in order to derive the generic interpretation We assume the unmodified wh-conditionals all contain a null generic operator GEN over world variables (that is definites may acquire universal properties in a generic context) And wh-phrases are concepts ie from possible worlds to individuals

(78) [[ shei xian lai]] = i x[person (x) amp first_comes (x)](i)

(79) GEN PQi-Altf(w)(s))P(i)Q(i) where (i) f(w)(s) is the set of worlds the speakerrsquos belief hold and (ii) a world wrsquof(w)(s) is an i-alternative iff there exists some wrsquorsquo such that x[P(wrsquo)(x)] x[P(wrsquorsquo)(x)]

The unique referent denoted by the wh-phrase is being relativized to worlds ie for each world there is exactly one individual involved in it And quantification is over the worlds rather than individuals This semantics derives the generic interpretation of wh-conditionals straightforwardly

(80) a [[ shei xian lai shei xian chi]] = 1 iff if(w)(s))first_eat (i) (x (person (x) amp comes_firsts (x) (i)) b As far as the speakerrsquos belief is concerned the first comer is the first eater c For each world w the first comer in it is the first eater in it

What (80) expresses can be illustrated as follows

(81) For any dinnerparty the speaker believes the following holds If Zhang San comes first Zhang San eats first Li Si comes first Li Si eats first Wang Wu comes first Wang Wu eats first Ma Liu comes first Ma Liu eats first Zhang San and Li Si arrive come first Zhang San and Li Si eat first hellip


779

The generic readings of wh-conditionals have been captured as desired on this approach

7 Are Chinese wh-conditionals ambiguous

The previous discussion unambiguously invites the inference that Chinese wh-conditionals are ambiguous Semantically Chinese wh-conditionals are akin to English FRs which have two varieties plain FRs and -ever FRs English plain FRs are argued to have a prima facie definiteunique interpretation while -ever FRs have some universal quantification interpretation Dayal (1997) proposes the universal quantificational force of -ever FRs is contributed by ever which adds some modality to the semantic representation and renders FRs to be interpreted attributively

11 A plausible assumption extending to Chinese wh-conditionals is that Chinese wh-conditionals conflate this distinction (between plain FRs and -ever FRs) and are always open to two interpretations Chinese lacks a lexical item like ever for the generic interpretation and sometimes only the contextpragmatics can tell which reading is the most salient one

In a neutral context a wh-conditional might be open to two interpretations The following (82) can be semantically represented either by (83a) and (83b) (83a) is about the definite reading while (83b) is the reading involving universal quantification over possible worlds Semantically (83a) is akin to plain FRs while (83b) to -ever FRs

(82) Shenme ren xian lai shenme ren xian chi what person first come what person first eat lsquoWho comes first eats firstrsquo lsquoWhoever comes first eats firstrsquo

(83) a eats-first (x [person (x) amp comes_firsts (w)(x)]) b i f(w)(s)eats-firsts (i)(x (person (x) amp comes_firsts (x)(i))

(84) a Who comes first eats first b Whoever comes first eats first

We again credit Cheng amp Huang (1996) with the relevant insight As noted earlier they analyze wh-conditionals on a par with donkey conditionals in English The ambiguity between definite and generic readings of wh-conditionals is also present in English donkey conditionals As Kadmon (1990) also observes donkey conditionals have both an absolute unique (definite) reading and a universal reading She distinguishes one-

11 Donnellan (1966) distinguishes two uses of definites referential vs attributive According to

Dayal the primary semantic function of ever is to force the FRs to be read attributively Otherwise FRs always receive a referential absolute unique reading


780

case conditionals from multi-case conditionals The following examples in (85a-b) are Kadmonrsquos (1990) examples (36) and (37) respectively and (86a-b) are her examples (35) and (41) respectively

One-case conditionals (85) a If there is a doctor in London and he is Welsh then we are all set b If it is true that a man walked in and that he decided to stay then Sally will be pleased

With stative antecedent clauses (85a-b) say that if the state described in the antecedent is true (85a) or that it is a fact that certain events have already taken place (85b) the consequent is also true These one-case conditionals contrast with multi-case conditionals

Multi-case conditionals (86) a If a man walks in and he sits down Sally is pleased b If a semanticist hears of a good job she applies for it

For instance example (86a) has simple present tense and a non-stative verb in the antecedent clause so it can be used to quantify over multiple cases of instances of a man walking in and sitting down ie it means that Sally is pleased about each instance (Kadmon 1990297) Example (86b) can also be used as a general statement ie whenever a semanticist hears of a good job she applies for it The semanticists and the jobs involved (simultaneously) can be multiple since no absolute uniqueness is being implicated However the uniqueness is relativized to situations (see Heim 1990)

We propose that the choice between a definite reading and a generic reading is regulated by the context in Chinese When the context is unspecified about the fixation of the referent it has a generic reading and involves universal quantification over the speakerrsquos belief worlds When the context imposes some absolute uniqueness requirement of the referent it has a definite reading Consider the following example

(87) Wo wangji ni jie-le ji-ben shu gei wo le I forget you lend-ASP how-many-CL book to me ASP danshi ni jie-le shenme gei wo wo jiu huan gei ni shenme le but you lend-ASP what to me I then return to you what ASP I donrsquot remember how many books you lent me but a Irsquove returned to you whatever books you lent me b Irsquove returned to you the books you lent me c Irsquove returned to you all the books you lent me

In (87) when the antecedent specifies a particular casesituation (ie book-lending case


781

by you to me) the generic reading is no longer the salient one And the ignorance and indifference implications also disappear (87) can be used to simply express the speaker has returned all the books the addressee lent to him

By patterning wh-conditionals with donkey conditionals we manage to preserve Cheng amp Huangrsquos (1996) insight We leave the studies on the deeper links between free relatives and donkey conditionals for some future work 71 A comparison with Lin (1996)

As pointed out in the literature a pronoun (or definite description) in the consequent clause of wh-conditionals can refer back to the wh-phrase in the antecedent To explain the distribution of wh-phrases and pronouns Lin (1996) takes inspiration from Kadmonrsquos distinction between one-case conditionals and multi-case conditionals Lin argues that a wh-indefinite can occur in the consequent clause in wh-conditionals if the antecedent clause denotes a multi-case situation When the antecedent denotes a one-case situation a pronoun is used He proposes the following condition for the distribution of pronouns in wh-conditionals

(88) Condition on donkey pronouns in wh-conditionals A donkey pronoun in a wh-conditional is felicitous only if it picks out a unique referent (Lin 1996250)

We agree that when an overt pronoun is used in the consequent clause of wh-conditionals it picks out a unique referent But we remain dubious about the distinction between one-case and multi-case According to Lin a wh-indefinite should not appear in the con-sequent clause of example (89) where the antecedent denotes a one-case situation To the contrary it has been pointed out that both a wh-indefinite and an overt pronoun can appear in the consequent clauses (cf Pan amp Jiang to appear)

(89) Shei shang xueqi na di-yi-ming sheita zhe xueqi jiu keyi dang who last semester got top-one who heshe this semester then may serve banzhang monitor

lsquoIf who got the first last semester he may serve as the monitor for the class this semesterrsquo lsquoWhoever got the first last semester may serve as the monitor for the class this semesterrsquo

The reason that (89) equally allows a wh-indefinite and an overt pronoun in its consequent clause is that (89) is indeed equivocal about the fixation of the referent In a situation


782

where there is only one class and only one student who got the top-first a uniqueness requirement (ie existential presupposition that there is exactly one student who got the top-first last semester) is established so a pronoun can be used In a situation where there are several classes and each class has one who got the top-first the identification of the referent is unspecified and the conditional makes a general statement The use of a wh-indefinite is licensed in that situation Since the intralinguistic information is insufficient to identify the referent the context is open to both interpretations Therefore both the pronouns and wh-indefinites can occur in the consequent clause This is not surprising As Kadmon (1990) admits lsquomany conditionals have both one-case and multi-case interpretationsrsquo (Kadmon 1990297)

8 Wh-conditionals and correlatives some typological consequences

81 Correlatives a first look

The Chinese wh-conditionals we have studied have three features (a) they have a strict matching requirement (b) they have some definite flavor (c) they are open to both the definite and universal interpretations Are these features unique to Chinese Correlatives appear to have these features Correlatives are widely attested in most Indo-Aryan languages (Hindi Sanskrit Nepali Bangla Hittite etc) Slavic and some East-Asian languages such as Korean and Vietnamese) Correlatives compose of a left-adjoined free relative (RC) and a main clause The left-adjoined free relative is referentially linked to a nominal correlate (which always takes the form of a demon-strative pronoun) (COR) in the main clause In the Indo-Aryan languages this is shown by a demonstrative requirement Schematically most correlatives have the following structure

(90) [RChellipwh(x)hellip]i [CORhellip(DEM)-yihellip] (Liptaacutek 2009)

The similarities between Chinese wh-conditionals and correlatives are hard to miss even at first glance Consider the following correlative examples (from Leung 2009)

(91) a [Wie jij uitgenodigd hebt] die wil ik niet meer zien [Dutch] who you invite have that-one want I no-longer see lsquoI donrsquot want to see the one you have invited any longerrsquo

b [Kogo ljublju] togo poceluju [Russian] whom love thatone willkiss1SG lsquoI will kiss who I loversquo


783

c [Aki Koran jot] azt ingyen beengedtek [Hungarian] who early came thatACC freely pvadmitted3PL lsquoThose who come early were admitted for freersquo

d [Khyodra-s gya gare njos yod na] nga-s de bsad pa yin [Lhasa Tibet] you-ERG yak what buy AUX if I-ERG that kiss past AUX lsquoI killed whatever yak you boughtrsquo

e [Ai nau] nay an [Vietnamese] who cook that-person eat lsquoWhoever cooks eatrsquo

f hellip[terra pulli qui nascentur] eos in terram deprimito [Latin] earth scions who willspring these in Earth pressIMPERF lsquohellippress into the earth scions that will spring from the groundrsquo

The examples in (91) are all correlatives (from a bunch of languages that are historically unrelated to each other and to Chinese) They share the same structure as (90) The RC contributed by a free relative is anaphorically linked to a definite expression in the COR As the Vietnamese example (91e) indicates they have the same truth condition as an identity statement (whoever cooks eat = the person who cooks is the person who eats)

Hindi correlatives occupy an important position in the study of correlatives (cf Dayal 1996 Grosu amp Landman 1998 Liptaacutek 2009 and references therein) A typical correlative in Hindi contains a left-peripheral relative clause which is anaphorically linked to a (possibly phonetically unrealized) definite nominal correlate in the clause that follows the relative clause The following well-cited illustrative example is from Srivastav (1991)

(92) [Jo laRkii khaRii hai] vo lambii hai wh girl standing is that tall is Lit lsquoThe girl who is standing there is tallrsquo lsquoWhichever girl is standing there is tallrsquo

What is interesting here is the referential link between the correlate DP lsquothat girlrsquo and the left-peripheral relative clause led by a wh-expression In (92) vo lsquothatrsquo picks out the same referent as the relative clause and occupies the same argument slot There is some independent evidence that the correlatives are different from headed relatives (please refer to Liptaacutek 2009 for more details)

In a correlative the correlate DP always follows the left-peripheral relative clause This contrasts with headed relatives Headed relatives either occur next to the nominal head they modify or they occur to the right of it at some distance Unlike correlatives however a relative clause that originates from a headed construction can never precede the modified nominal The following examples are Liptaacutekrsquos (6a-c)


784

(93) a John called somebody [who he knows from school] yesterday b John called somebody yesterday [who he knows from school] c [Who he knows from school] John called someone yesterday

Hindi correlatives contrast with Hindi headed relatives In Hindi headed relatives the relative occupies a clause-internal position which is necessarily right-adjacent to the head noun (cf 94a) Such relative clauses cannot be non-adjacent to their head in sentence-internal positions (cf 94b)

(94) a Vo laRkii [jo khaRii hai] lambii hai that girl wh standing is tall is lsquoThe girl who is standing is tallrsquo

b Vo laRkii lambii [jo khaRii hai] hai that girl tall wh standing is is

This contrasts with correlatives In correlatives the relative clause can be non-adjacent to the nominal it modifies as shown in example (95)

(95) [Jo vahaaN khaRii hai] raam us laRii-ko jaantaa hai (Liptaacutek 2009 ex (4)) wh there standing is Ram that girl-ACC know is Lit lsquoRam knows the girl who is standing therersquo

Another notable difference between correlatives and headed relatives concerns the distribution of the nominal element they modify In correlatives the nominal element can be spelled out either inside the relative clause as shown in (92) or inside the correlate as shown in (96a) below or both inside the relative clause and in the correlate phrase simultaneously as shown in (96b)

(96) a [Jo khaRii hai] vo laRkii lambii hai (Liptaacutek 2009 ex (9)) wh standing is that girl tall is

b [Jo laRkii khaRii hai] vo laRkii lambii hai wh girl standing is that girl tall is Lit lsquoWhich girl is standing that girl is tallrsquo

Headed relatives do not have the option of allowing the nominal element to appear both in the head position and in the relative clause as shown by the contrast between (97a) and (97b) below

(97) a Vo laRkii [jo khaRii hai] lambii hai that girl wh standing is tall is


785

b Vo laRkii [jo laRkii khaRii hai] lambii hai that girl wh girl standing is tall is

82 Definiteness in Hindi correlatives

Hindi correlatives also have some definiteness flavor The definite nature of correlatives can be demonstrated by the fact that the correlate DP must be definite or universal (92b) (Dayal 1996 Grosu amp Landman 1998 Liptaacutek 2009) Numerals ordinary indefinites are not allowed to introduce the correlate DP

(98) a [Jo laRkii khaRii hai] vo lambii hai wh girl standing are that tall is lsquoWhich girl is standing that is tallrsquo (lsquoWhoever standing there is tallrsquo)

b [Jo laRke KhaRe haiN] vedonosabdokuchadhiktam lambe haiN wh girl standing are thosebothalltwofewmost tall are lsquoWhich boys are standing theybothalltwofewmost are tallrsquo

As (98b) illustrates the relative can only be anaphorically linked to a definite or universal DP (like lsquothosersquo lsquobothrsquo lsquoallrsquo) but not an indefinite DP (like lsquotworsquo lsquofewrsquo etc) Also like Chinese wh-conditionals the correlate DP can be a partitive12

(99) [Jo laRke KhaRe haiN] unme se do unme se kuch lambe haiN wh boys standing are among them two among them some tall are lsquoThe boys are standing there two of them some of them are tallrsquo

Grosu amp Landman (1998) take this fact to argue that there is a maximalization operation applying to the relative clause of the correlative The maximalization operation generates a maximal individual which can only be referred back by a definite or a universal but not by an indefinite The notion lsquomaximalizationrsquo is sort of descriptively misleading however note that the correlate DP can be a partitive But if we treat the referential link between the RC and the correlate DP as a result of semantic identity this maximali-zation effect follows for free 83 The matching requirement in Hindi correlatives

Another commonality between Hindi correlatives and Chinese wh-conditionals is the matching requirement The Hindi correlatives are subject to a matching requirement which requires the relative morphemes in the relative clause and the demonstrative

12 We thank Yatin Mahajan for the judgment


786

morphemes in the main clause to be identical in number and reference Leung (2009) states this matching requirement as follows

(100) Matching requirement for correlatives a The number of relative morphemes in the correlative equals the number of

demonstrative morphemes in the main clause b The mapping between the set of relative morphemes within the correlative clause

and the set of demonstratives within the main clause is bijective ie for every relative morpheme within the correlative clause there is exactly one demonstrative morpheme within the main clause such that relative morpheme and demonstrative morpheme are coindexed

In other words the correlate DP must be identical to RC in reference in addition to being identical in number

(101) A Rel that bears an index i maps onto a Dem that bears an index i a Rel that bears an index j maps onto a Dem that bears an index j and so on

This matching requirement can be illustrated by the following examples (Leung 2009 317-318) (102a) below is out because there is no Dem in the correlate and (102b-c) is out because the Dem and Rel are not identical in number

(102) a [Jis larke-ne sports medal jiit-aal] academic medal-bhii jiit-aa wh boy-ERG sports mdedal win-PFV academic medal-also win-PFV

b [Jis larke-ne jis larki-ko dehka] us larki-ko piitaa gayaa wh boy-ERG girl-ACC saw DEM girl-ACC beaten was lsquoA girl who a boy saw was beatenrsquo

c [Jo laRkii jis laRke-ke saath khelegii] vo jiit-jaayegii wh girl wh boy-OBL with play she win-PERF lsquoA girl who plays with a boy will winrsquo

Chinese wh-conditionals also have this requirement as we have already discussed in the previous sections The matching requirement says the wh-expressions in the antecedent and consequent must be identical in number form and reference The Chinese wh-conditionals counterparts of (102) are all unacceptable as expected

(103) a Na-ge nanhai de-le yundong jinpai de-le xueshu jinpai which-CL boy get-ASP sport medal get-ASP academic medal

b Na-ge nanhai kanjian na-ge nuumlhai na-ge nuumlhai jiu bei da which-CL boy saw which-CL girl which-CL girl then PASS beat


787

c Na-ge nuumlhai he na-ge nanhai wan na-ge nuumlhai jiu neng ying which-CL girl and which-CL boy play which-CL girl then will win

84 The definite and universal readings of correlatives

Correlatives permit both definite and universal interpretations The following example (104) in Hindi has a definite reading

(104) definite reading- Hindi correlative (the example is from Dayal 1996) Jo laRkii lambii hai vo khaRii hai wh girl tall is that standing is lsquoThe girl that is standing is tallrsquo

Some correlatives (especially ones with multiple wh-indefinites) have mixed definite universal readings

(105) mixed definiteuniversal readings- Hindi correlatives (ibid) Jis laRkii-ne jis laRke-ke saath khel-aa us-ne us-ko haraa-yaa wh girl-ERG wh boy-with together play that-one-ERG that-one-ACC defeat

lsquoEvery girl that played against a boy is such that (she played against exactly one boy and) she defeated the one boy she played againstrsquo

lsquoThe girl that played against the boy defeated the boyrsquo

Dayal (1995) notes that correlatives always receive a universal interpretation when we switch from episodic to habitual morphology

(106) universal reading with habitual morphology- Hindi correlatives (Dayal 1995) Jo laRkii lambii ho-tii hai vo khaRii ho-tii hai wh girl tall be-HAB is that-one standing be-HAB is lsquoA tall girl (generally) stands eg on buses with very little leg room between seatsrsquo

Example (106) has a reading that generalizes over situations in which there is exactly one girl who is tall About each such situation the tall girl in it stands So for different situations there are (possibly) different tall girls involved The habitual morphology forces the sentence to be read as involving some quantification over situationscases (rather than over individuals)

More needs to be said about Hindi correlatives There are two types of correlatives in Hindi one with the morpheme bhii and one without it The correlatives with bhii readily allow the universal interpretation while the correlatives without bhii are subject to a salient definite reading (without habitual morphology) Dayal (1997) analyzes bhii on par with English lsquoeverrsquo in -ever FRs That is bhii introduces some modality dimension


788

to the semantic representation Also unlike Chinese Hindi correlatives have internal heads and the verbal morphology is transparent with respect to number marking so correlatives in Hindi come out either with a definite reading or a quasi-universal reading The following examples with bhii are from Dayal (1997)

(107) [Jo bhii laRkii mehnat kar rahii] vo fafal hogii wh ever girl effort making is she successful will be lsquoWhichever girl that is making an effort will be successfulrsquo lsquoAny girl who makes an effort succeedsrsquo

So Hindi is more like English which has some overt quantificational strategies (it has a habitual morphology to force a reading as involving quantification over situations see (106) it has a lexical item bhii that contributes modality to semantic representation like English ever which contributes to a reading where the referent denoted by the wh-phrase is unknown or not relevant see (107)) Chinese lacks these strategies and the choice between the readings is left largely to contextpragmatics Cross-linguistically Romanian is more like Chinese Brasoveanu (2007) notes that in Romanian which reading (definite or universal) is available for a particular correlative depends on world knowledge and pragmatic factors (eg accidentalsporadic vs non-accidentalhabitual nature of the situations under discussion) The following Romanian example (108) has a salient definite reading while (109) has a salient universal reading (both examples are from Brasoveanu 2008)

(108) Definite interpretation ndash Romanian Care fata si-a-uitat ieri haina pe aceea o-cauta hata-l ei wh girl her-DAT-forgotten yesterday coat-the PE that-one her-look-for father her lsquoThe father of the girl that forgot her coat yesterday is looking for herrsquo

(109) Universal interpretation ndash Romanian PE care om l-a-interogat securitate-a in acela nu-mai-am incredere PE wh person him-ACC-inter security-the in that-one not-any more trust lsquoI donrsquot trust any person interrogated by the secret police anymorersquo

Semantically the difference between the definite vs universal readings of correlatives (and Chinese wh-conditionals) boils down to a difference in granularity level of the quantification (see Brasoveanu 2007) The quantification can be coarse-grained ie we lsquocollectivelyrsquo quantify over topical casessituations which boils down to quantifying over topical individualsand the correlative clause (the consequent clause in the case of Chinese wh-conditionals) is predicated about these individuals This yields the definiteunique reading Alternatively the quantification can be fine-grained ie we lsquodistributivelyrsquo quantify over the topical casessituations introduced in the relative


789

clauseand the correlative clause (or the consequent clause) is predicated of each of such casessituations This yields the universal interpretation (Presumably) some languages (like Hindi English) lexically (or morpho-syntactically) specifies which granularity level in quantification is being adopted while some other languages (like Chinese Romanian) turn to pragmaticsdiscourse to decide the granularity level in quantification The accidentalsporadic vs non-accidentalgeneric nature of the situationscases under consideration also matters We leave a fuller implementation of this idea to future works

9 Chinese wh-conditionals as correlatives

Cheng amp Huang (1996) offer three arguments against Chinese wh-conditionals as genuine correlatives We shall examine these arguments one by one in light of the new evidence gathered in this study According to Cheng amp Huang first wh-conditionals are not typical relatives They argue that a typical relative clause in Chinese is marked by the relative marker de while this is not the case for wh-conditionals

(110) Wo mai-le [Jinyong xie (de) shu] (Cheng amp Huang 1996 ex (78)) I buy-ASP Jinyong write REL book lsquoI bought books that Jinyong wrotersquo

Example (110) is a typical headed relative (lsquobooks that Jinyong wrotersquo) As shown earlier correlatives are not like headed relative clauses It does not follow from the fact that Chinese wh-conditionals differ from typical relatives to the conclusion that they are not similar to correlatives On the present account Chinese wh-conditionals are more like headless free relatives than headed relatives Interestingly Cheng amp Huang themselves also comment on the commonality between headless free relatives and wh-conditionals The following example (111b) is their example (80)

(111) a What you see is what you get b Ni kandao shenme ni jiu dedao shenme you see what you then get what lsquoIf you see X then you get Xrsquo

The second argument Cheng amp Huang offer concerns tenseaspect specifications

They argue that wh-conditionals are restricted to certain tenseaspectual specifications that make them different from Hindi correlatives The verb in the consequent clause cannot bear completive aspect in Chinese wh-conditionals but it can in Hindi correlatives They provide the following examples (the sentences under (112) are their (81) while (113) is their (82))


790

(112) a Shei lai shei jiu hui zhong jiang who come who then will win prize lsquoWhoever comes shehe will win a prizersquo

b Shei lai shei jiu zhong-le jiang who come who then win-ASP prize

(113) Jis laRkiiNe ji laRkeKO dekhaa usNE usKO passand kiyaa wh girl-ERG REL boy-ACC saw DEM-ERG DEM-ACC liked lsquoWhich girl saw which boy she liked himrsquo

Why do Chinese wh-conditionals disallow a completive aspect marker in the consequent clause According to Cheng amp Huang Chinese wh-conditionals are genuine conditionals and the consequent clause of a conditional always express a situation posterior to rather than an anterior to the situation expressed by the antecedent clause So a future marker hui is OK while a completive marker is out Typical relatives do not have this require-ment This is the contrast between (112) and (113) (Note in (113) the main clause contains a verb with past tense)

We think Cheng amp Huangrsquos example (112b) is odd is because of a disagreement in aspect between the antecedent clause and consequent clause not due to tenseaspect restriction on conditionals In their example the antecedent bears a default present tense while the consequent bears a completive aspect This mismatch in tenseaspect agree-ment is responsible for the oddness of the example as in the following English example

(114) Whatever you bring didnrsquot please Mary

If the antecedent clause and the consequent clause match in tenseaspect the sentence becomes grammatical just like Hindi correlatives even if both clauses bear past tenseaspect markers When the antecedent clause bears completive aspect the conse-quent clause also needs to bear the same aspect in order to ensure agreement and the sentence is fine Consider (115)

(115) An old man is recounting his life Rensheng jiushi zheyang life just-be like this ni xuanzhe-le shenme shenghuo ni jiu xiguan-le shenme shenghuo you choose-ASP what life you then accustom-ASP what life Lit lsquoLife is always like this you choose what life you grow accustomed to what lifersquo lsquoYou have been accustomed to whatever life you have chosenrsquo

Further supporting observations are due to Huang (2008) She also observes that wh-conditionals (which she calls Chinese Bare Conditionals (CBS)) can admit non-future


791

tenseaspect in the consequent clause contrary to Cheng amp Huangrsquos claim She remarks lsquoif uttered to describe a past event CBS can bear a perfectivecompletive aspect which can be overtly or covertly indicated by a perfective morphemersquo Consider her example (17) (repeated here as (116) below cf Huang 2008698-699)

(116) Zuotian shei zai xuanpiao shang Dawei jiu tou-le shei yesterday who is ballot on-top David then vote-PERF who lsquoYesterday David voted for whoever appeared on the ballotrsquo

As (116) indicates in the presence of time adverbials such as zuotian lsquoyesterdayrsquo and natian lsquothat dayrsquo wh-conditionals freely allow the verb in the consequent to be modified by the perfective morpheme

The third argument by Cheng amp Huang is about the proportional problem of donkey conditionals According to them ordinary donkey conditionals have both a symmetric reading and an asymmetric reading while Hindi correlatives reportedly do not have this ambiguity Chinese wh-conditionals resembling English donkey conditionals have this ambiguity The following (117) is their original (75) (the glosses and English translation are also theirs)

(117) Shei yan shei shei tongchang jiu hui xiang shei who play who who usually then will resemble who lsquoIf X plays the role of Y then usually X will resemble Yrsquo

In a neutral context the most salient reading for (117) is the subject asymmetric reading Suppose there are ten actors Suppose also that while nine of the actors each play one role and resemble that role the tenth plays ten roles and does not resemble any of those roles The sentence is true in this scenario But the symmetric reading (namely actor-role pairs) would predict it to be false The symmetric reading is possible according to Cheng amp Huang only when a topic is being specified If the speaker first says lsquowith respect to the actors and characters (roles) in the playrsquo then the sentence will be true Topic choice plays an essential role here

In our view Chinese wh-conditionals are ambiguous because of discourse factors In a neutral context when the topic is not being specified the wh-conditionals are unambiguous Recall that this discourse prominence is an important feature for Chinese We attribute the lack of ambiguity in Hindi to the fact that Hindi is not as discourse-prominent as Chinese If topic choice were as easily accomplished in Hindi as it is in Chinese we would expect Hindi correlatives to be ambiguous And this seems to be the case as attested by the Hindi example (105)


792

To summarize we have re-examined the three arguments offered by Cheng amp Huang (1996) that argue against Chinese wh-conditionals as correlatives We have tried to deflect the force of their arguments using new evidence we have gathered in the course of the present study

10 Concluding remarks

Previous studies treat wh-conditionals as a proper instance of universal donkey conditionals and derive the generic reading via unselective binding We argue wh-conditionals encode uniqueness and are open to both generic and definite interpretations We adopt a Russelian semantics to capture the definitenessuniqueness effect and derive the generic reading via generic quantification over possible worlds where the (unique) identity of the referents denoted by the wh-phrases is not known or not relevant This analysis brings wh-conditionals close to unique donkey conditionals as discussed in Heim (1990) and in Kadmon (1990) and close to free relatives and correlatives as discussed in Dayal (1997) and in von Fintel (2000) Despite being conditional in form Chinese wh-conditionals turn out to be semantically akin to correlatives (and free relatives) generally Typologically they may appear to be strange relatives of a third kind (Grosu amp Landman 1998) But the many commonalities in form and meaning between Chinese wh-conditionals and Hindi correlatives speak out for a common analysis

Even granting the similarities between wh-conditionals and correlatives however differences between them need to be explained In correlatives anaphoric expressions (eg pronouns demonstratives even epithets) appear in the consequent whereas a wh-phrase always appears in wh-conditionals We must admit that we do not have a definitive account of this difference Solutions however are on the table After discussing the commonalities between wh-movement and correlative proforms Izvorski (1996) contends that correlative proforms should be treated as wh-phrases in the syntax It has been proposed that the ordering (clause-initial or not) of the correlative proforms in correlatives is correlated with the same parametric variation about the overt vs covert movement of wh-phrases In languages with overt wh-movement the correlative proform is always fronted to the highest Spec of CP (ie clause-initial) The following Bulgarian example is from Izvorski (1996 ex (34))

(118) a [Kolkoto pari iska]i tokkovai misli ce ste i dam how-much money wants that-much thinks that will her give-1SG lsquoShe thinks I will give her as much money as she wantsrsquo b [Kolkoto pari iska]i misli ce ste i dam tokkovai c [Kolkoto pari iska]i misli ce tokkovai ste i dam


793

The overt movement of the correlative proforms is subject to island constraint (see Izvorski 1996 ex (35))

(119) [Kakto im kazah]i takai cuh (sluha) ce sa postapili how them told-1SG that-way heard-1SG the-rumor that are done lsquoI heard (the rumour) that they had acted the way that I had told them to dorsquo

In languages with no overt wh-movement by contrast the correlative proform can stay in-situ and does not have to appear clause-initial The following Hindi example is due to Srivastav (1991)

(120) [Ram-ne Sita-ko jo kitaab dii]i Bill-ne Sara-ko vohi dikhaa-ii Ram-ERG Sita-DAT wh book give-PERF Bill-ERG Sara-DAT DEM show-PERF lsquoBill showed to Sara the book that Ram gave to Sitarsquo

As Izvorski contends in addition to the covertovert movement distinction between languages further variations in wh-movement (ie extraction from indicative vs sub-junctive clauses possibility vs prohibition of left-branch extraction whether or not topics are allowed to precede the wh-word) obtain in the case of correlative proforms These facts provide further support for the position that the correlative proform is treated as a wh-phrase by the syntax Also the fact that in Chinese-type correlatives (ie wh-conditionals) the wh-words always remain in-situ may not turn out to be so typo-logically surprising But we must leave this possibility for future research


794

References Barwise Jon and Robin Cooper 1981 Generalized quantifiers and natural languages

Linguistics and Philosophy 42159-219 Bhatt Rajesh and Pancheva Roumyana 2006 Conditionals The Blackwell Companion

to Syntax Vol 1 ed by Martin Everaert amp Henk C van Riemsdijk 638-687 Malden Blackwell

Bittner Maria 2001 Topical referents for individuals and possibilities Proceedings of SALT 11 ed by Rachel Hastings Brendan Jackson amp Zsofia Zvolenszky 36-55 Ithaca CLC Publications

Brasoveanu Adrian 2007 Structured Nominal and Modal Reference New Brunswick Rutgers University dissertation

Brasoveanu Adrian 2008 Uniqueness effects in correlatives Proceedings of Sinn und Bedeutung 12 ed by Atle Groslashnn 47-65 Oslo ILOS

Bruening Benjamin and Thuan Tran 2006 Wh-conditionals in Vietnamese and Chinese against unselective binding Paper presented at the 32nd Annual Meeting of the Berkeley Linguistics Society February 10-12 2006 Berkeley University of Cali-fornia

Cheng Lisa L-S and C-T James Huang 1996 Two types of donkey sentences Natural Language Semantics 42121-163

Chierchia Gennaro 1998 Reference to kinds across languages Natural Language Semantics 64339-405

Chierchia Gennaro 2000 Chinese conditionals and the theory of conditionals Journal of East Asian Linguistics 911-54

Citko Barbara 2001 Deletion under identity in relative clauses Proceedings of NELS 31 ed by Minjoo Kim amp Uri Strauss 131-145 Amherst GLSA

Condoravdi Cleo 2005 Not knowing or caring who Manuscript Palo Alto Palo Alto Research Center

Cooper Robin 1979 The interpretation of pronouns Syntax and Semantics Vol 10 Selections from the Third Groningen Round Table ed by Frank Heny amp Helmt S Schnelle 61-92 New York Academic Press

Dayal Veneeta 1995 Quantification in correlatives Quantification in Natural Language ed by Emmon Bach Eloise Jelinek Angelika Kratzer amp Barbara H Partee 179-205 Dordrecht Kluwer

Dayal Veneeta 1996 Locality in Wh-quantification Questions and Relative Clauses in Hindi Dordrecht Kluwer

Dayal Veneeta 1997 Free relatives and -ever identity and free choice reading Proceedings of SALT 7 ed by Aaron Lawson amp Enn Cho 99-116 Ithaca CLC Publications


795

den Dikken Marcel 2006 Relators and Linkers The Syntax of Predication Predicate Inversion and Copulas Cambridge MIT Press

Donnellan Keith D 1966 Reference and definite descriptions Philosophical Review 753281-304

Ebert Christian Cornelia Endriss and Stefan Hinterwimmer 2008 Topics as speech acts an analysis of conditionals Proceedings of the 27th West Coast Conference on Formal Linguistics (WCCFL 27) ed by Natasha Abner amp Jason Bishop 132-140 Somerville Cascadilla Proceedings Project

Elbourne Paul D 2005 Situations and Individuals Cambridge MIT Press Evans Gareth 1980 Pronouns Linguistic Inquiry 112337-362 von Fintel Kai 2000 Whatever Proceedings of SALT 10 ed by Brendan Jackson amp

Tanya Matthews 27-39 Ithaca CLC Publications Grosu Alexander and Fred Landman 1998 Strange relatives of the third kind Natural

Language Semantics 62125-170 Gu Chloe 2009 Maximalization and the definite reading in Mandarin wh-conditionals

Paper presented at the 40th Annual Meeting of the North East Linguistic Society (NELS 40) November 13-15 2009 Cambridge MIT

Heim Irene 1982 The Semantics of Definite and Indefinite Noun Phrases Amherst University of Massachusetts dissertation

Heim Irene 1990 E-type pronouns and donkey anaphora Linguistics and Philosophy 132137-177

Heim Irene and Angelika Kratzer 1998 Semantics in Generative Grammar Malden Blackwell

Huang Yahui Anita 2008 Presuppositions in Chinese bare conditionals Proceedings of the 20th North American Conference on Chinese Linguistics (NACCL-20) Vol 2 ed by Marjorie K M Chan amp Hana Kang 691-704 Columbus East Asian Studies Center The Ohio State University

Izvorski Roumyana 1996 The syntax and semantics of correlative proforms Pro-ceedings of NELS 26 ed by Kiyomi Kusumoto 133-147 Amherst GLSA

Jackendoff Ray S 1972 Semantic Interpretation in Generative Grammar Cambridge MIT Press

Jacobson Pauline 1995 On the quantificational force of English free relatives Quantification in Natural Language ed by Emmon Bach Eloise Jelinek Angelika Kratzer amp Barbara H Partee 451-486 Dordrecht Kluwer

Kadmon Nirit 1987 On Unique and Non-unique Reference and Asymmetric Quantifi-cation Amherst University of Massachusetts dissertation

Kadmon Nirit 1990 Uniqueness Linguistics and Philosophy 133273-324 Kamp Hans 1981 A theory of truth and semantic representation Formal Methods in

the Study of Language ed by Jeroen A G Groenendijk T M V Janssen amp Martin


796

J B Stokhof 277-322 Amsterdam Mathematisch Centrum Leung Tommi 2009 On the matching requirement in correlatives Correlatives Cross-

linguistically ed by Anikoacute Liptaacutek 309-342 Amsterdam amp Philadelphia John Benjamins

Lewis David K 1973 Counterfactuals Cambridge Harvard University Press Li Y-H Audrey 1992 Indefinite wh in Mandarin Chinese Journal of East Asian

Linguistics 12125-155 Lin Jo-wang 1996 Polarity Licensing and Wh-phrase Quantification in Chinese

Amherst University of Massachusetts dissertation Link Godehard 1983 The logical analysis of plurals and mass terms a lattice-

theoretical approach Meaning Use and Interpretation of Language ed by Rainer Baumluerle Christoph Schwarze amp Arnim von Stechow 302-323 Berlin amp New York Walter de Gruyter

Link Godehard 1998 Algebraic Semantics in Language and Philosophy Stanford CSLI

Liptaacutek Anikoacute 2009 The landscape of correlatives an empirical and analytical survey Correlatives Cross-linguistically ed by Anikoacute Liptaacutek 1-48 Amsterdam amp Phila-delphia John Benjamins

Moltmann Friedericke 2010 Identificational sentences and the objects of direct per-ception Manuscript Paris IHPST

Neale Stephen 1990 Descriptions Cambridge MIT Press Pan Haihua and Yan Jiang (to appear) NP interpretation and donkey sentences in

Chinese Journal of East Asian Linguistics Partee Barbara H 1986 Noun phrase interpretation and type-shifting principles Studies

in Discourse Representation Theory and the Theory of Generalized Quantifiers ed by Jeroen A G Groenendijk Dick de Jongh amp Martin J B Stokhof 115-143 Dordrecht Foris

Patel-Grosz Pritty and Patrick Grosz 2010 On the typology of donkeys two types of anaphora resolution Proceedings of Sinn und Bedeutung 14 ed by Martin Prinzhorn Viola Schmitt amp Sarah Zobel 339-355 Vienna University of Vienna

Schlenker Philippe 2004 Conditionals as definite descriptions (a referential analysis) Research on Language and Computation 23417-462

Schwarz Florian 2009 Two Types of Definites in Natural Languages Amherst Uni-versity of Massachusetts dissertation

Srivastav Veneeta 1991 The syntax and semantics of correlatives Natural Language and Linguistic Theory 94637-686

Tham Shiao Wei 2008 The semantic category of the subject NP in Mandarin specifi-cational copular sentences Journal of East Asian Linguistics 17161-82


797

Tredinnick Victoria 2005 On the Semantics of Free Relatives with -ever Philadelphia University of Pennsylvania dissertation

[Received 27 October 2010 revised 19 February 2011 accepted 14 March 2011] Qiong-peng Luo College of Liberal Arts Nanjing University Nanjing 210093 China jucathy2009gmailcom Stephen Crain Australian Research Council Centre of Excellence in Cognition and its Disorders Macquarie University Sydney NSW 2109 Australia stephencrainmqeduau


798

漢語 Wh-條件式在別的語言中可否有（從句）親屬

羅瓊鵬1 Stephen Crain2

南京大學1

Macquarie University2

本文討論了當前對 Wh-條件式的研究中的幾個問題(1) Wh-條件式的

ldquo有定性現象(2) 把 Wh-詞處理為無定項與ldquo新穎性條件的矛盾

(3) Wh-條件式真正的語義解讀本文認為(1) Wh-無定項位於條件式中

時它們是ldquo獨一的無定項(2) Wh-條件式是一種特殊的ldquo話題ndash陳述

結構並且在兩個指稱項中存在ldquo等同的關係(3) Wh-條件式可以描述

單數或者複數的情景並導致兩種不同的解讀有定性解讀或者全稱性解

讀這兩種解讀由語境語用決定從類型學的角度來看漢語 Wh-條件式

和關聯關係從句 (correlatives) 在句法和語義上存在諸多相似之處

關鍵詞漢語 Wh-條件式關聯關係從句獨一現象自由關係從句

ltlt ASCII85EncodePages false AllowTransparency false AutoPositionEPSFiles true AutoRotatePages None Binding Left CalGrayProfile (Dot Gain 20) CalRGBProfile (sRGB IEC61966-21) CalCMYKProfile (US Web Coated 050SWOP051 v2) sRGBProfile (sRGB IEC61966-21) CannotEmbedFontPolicy Error CompatibilityLevel 14 CompressObjects Tags CompressPages true ConvertImagesToIndexed true PassThroughJPEGImages true CreateJobTicket false DefaultRenderingIntent Default DetectBlends true DetectCurves 00000 ColorConversionStrategy CMYK DoThumbnails false EmbedAllFonts true EmbedOpenType false ParseICCProfilesInComments true EmbedJobOptions true DSCReportingLevel 0 EmitDSCWarnings false EndPage -1 ImageMemory 1048576 LockDistillerParams false MaxSubsetPct 100 Optimize true OPM 1 ParseDSCComments true ParseDSCCommentsForDocInfo true PreserveCopyPage true PreserveDICMYKValues true PreserveEPSInfo true PreserveFlatness true PreserveHalftoneInfo false PreserveOPIComments true PreserveOverprintSettings true StartPage 1 SubsetFonts true TransferFunctionInfo Apply UCRandBGInfo Preserve UsePrologue false ColorSettingsFile () AlwaysEmbed [ true ] NeverEmbed [ true ] AntiAliasColorImages false CropColorImages true ColorImageMinResolution 300 ColorImageMinResolutionPolicy OK DownsampleColorImages true ColorImageDownsampleType Bicubic ColorImageResolution 300 ColorImageDepth -1 ColorImageMinDownsampleDepth 1 ColorImageDownsampleThreshold 150000 EncodeColorImages true ColorImageFilter DCTEncode AutoFilterColorImages true ColorImageAutoFilterStrategy JPEG ColorACSImageDict ltlt QFactor 015 HSamples [1 1 1 1] VSamples [1 1 1 1] gtgt ColorImageDict ltlt QFactor 015 HSamples [1 1 1 1] VSamples [1 1 1 1] gtgt JPEG2000ColorACSImageDict ltlt TileWidth 256 TileHeight 256 Quality 30 gtgt JPEG2000ColorImageDict ltlt TileWidth 256 TileHeight 256 Quality 30 gtgt AntiAliasGrayImages false CropGrayImages true GrayImageMinResolution 300 GrayImageMinResolutionPolicy OK DownsampleGrayImages true GrayImageDownsampleType Bicubic GrayImageResolution 300 GrayImageDepth -1 GrayImageMinDownsampleDepth 2 GrayImageDownsampleThreshold 150000 EncodeGrayImages true GrayImageFilter DCTEncode AutoFilterGrayImages true GrayImageAutoFilterStrategy JPEG GrayACSImageDict ltlt QFactor 015 HSamples [1 1 1 1] VSamples [1 1 1 1] gtgt GrayImageDict ltlt QFactor 015 HSamples [1 1 1 1] VSamples [1 1 1 1] gtgt JPEG2000GrayACSImageDict ltlt TileWidth 256 TileHeight 256 Quality 30 gtgt JPEG2000GrayImageDict ltlt TileWidth 256 TileHeight 256 Quality 30 gtgt AntiAliasMonoImages false CropMonoImages true MonoImageMinResolution 1200 MonoImageMinResolutionPolicy OK DownsampleMonoImages true MonoImageDownsampleType Bicubic MonoImageResolution 1200 MonoImageDepth -1 MonoImageDownsampleThreshold 150000 EncodeMonoImages true MonoImageFilter CCITTFaxEncode MonoImageDict ltlt K -1 gtgt AllowPSXObjects false CheckCompliance [ None ] PDFX1aCheck false PDFX3Check false PDFXCompliantPDFOnly false PDFXNoTrimBoxError true PDFXTrimBoxToMediaBoxOffset [ 000000 000000 000000 000000 ] PDFXSetBleedBoxToMediaBox true PDFXBleedBoxToTrimBoxOffset [ 000000 000000 000000 000000 ] PDFXOutputIntentProfile () PDFXOutputConditionIdentifier () PDFXOutputCondition () PDFXRegistryName () PDFXTrapped False CreateJDFFile false Description ltlt ARA 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 BGR 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 CHS ltFEFF4f7f75288fd94e9b8bbe5b9a521b5efa7684002000410064006f006200650020005000440046002065876863900275284e8e9ad88d2891cf76845370524d53705237300260a853ef4ee54f7f75280020004100630072006f0062006100740020548c002000410064006f00620065002000520065006100640065007200200035002e003000204ee553ca66f49ad87248672c676562535f00521b5efa768400200050004400460020658768633002gt CHT ltFEFF4f7f752890194e9b8a2d7f6e5efa7acb7684002000410064006f006200650020005000440046002065874ef69069752865bc9ad854c18cea76845370524d5370523786557406300260a853ef4ee54f7f75280020004100630072006f0062006100740020548c002000410064006f00620065002000520065006100640065007200200035002e003000204ee553ca66f49ad87248672c4f86958b555f5df25efa7acb76840020005000440046002065874ef63002gt CZE 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 DAN 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 DEU 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 ESP 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 ETI ltFEFF004b00610073007500740061006700650020006e0065006900640020007300e4007400740065006900640020006b00760061006c006900740065006500740073006500200074007200fc006b006900650065006c007300650020007000720069006e00740069006d0069007300650020006a0061006f006b007300200073006f00620069006c0069006b0065002000410064006f006200650020005000440046002d0064006f006b0075006d0065006e00740069006400650020006c006f006f006d006900730065006b0073002e00200020004c006f006f0064007500640020005000440046002d0064006f006b0075006d0065006e00740065002000730061006100740065002000610076006100640061002000700072006f006700720061006d006d006900640065006700610020004100630072006f0062006100740020006e0069006e0067002000410064006f00620065002000520065006100640065007200200035002e00300020006a00610020007500750065006d006100740065002000760065007200730069006f006f006e00690064006500670061002e000d000agt FRA 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 GRE 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 HEB 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 HRV (Za stvaranje Adobe PDF dokumenata najpogodnijih za visokokvalitetni ispis prije tiskanja koristite ove postavke Stvoreni PDF dokumenti mogu se otvoriti Acrobat i Adobe Reader 50 i kasnijim verzijama) HUN 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 ITA ltFEFF005500740069006c0069007a007a006100720065002000710075006500730074006500200069006d0070006f007300740061007a0069006f006e00690020007000650072002000630072006500610072006500200064006f00630075006d0065006e00740069002000410064006f00620065002000500044004600200070006900f900200061006400610074007400690020006100200075006e00610020007000720065007300740061006d0070006100200064006900200061006c007400610020007100750061006c0069007400e0002e0020004900200064006f00630075006d0065006e007400690020005000440046002000630072006500610074006900200070006f00730073006f006e006f0020006500730073006500720065002000610070006500720074006900200063006f006e0020004100630072006f00620061007400200065002000410064006f00620065002000520065006100640065007200200035002e003000200065002000760065007200730069006f006e006900200073007500630063006500730073006900760065002egt JPN ltFEFF9ad854c18cea306a30d730ea30d730ec30b951fa529b7528002000410064006f0062006500200050004400460020658766f8306e4f5c6210306b4f7f75283057307e305930023053306e8a2d5b9a30674f5c62103055308c305f0020005000440046002030d530a130a430eb306f3001004100630072006f0062006100740020304a30883073002000410064006f00620065002000520065006100640065007200200035002e003000204ee5964d3067958b304f30533068304c3067304d307e305930023053306e8a2d5b9a306b306f30d530a930f330c8306e57cb30818fbc307f304c5fc59808306730593002gt KOR ltFEFFc7740020c124c815c7440020c0acc6a9d558c5ec0020ace0d488c9c80020c2dcd5d80020c778c1c4c5d00020ac00c7a50020c801d569d55c002000410064006f0062006500200050004400460020bb38c11cb97c0020c791c131d569b2c8b2e4002e0020c774b807ac8c0020c791c131b41c00200050004400460020bb38c11cb2940020004100630072006f0062006100740020bc0f002000410064006f00620065002000520065006100640065007200200035002e00300020c774c0c1c5d0c11c0020c5f40020c2180020c788c2b5b2c8b2e4002egt LTH ltFEFF004e006100750064006f006b0069007400650020016100690075006f007300200070006100720061006d006500740072007500730020006e006f0072011700640061006d00690020006b0075007200740069002000410064006f00620065002000500044004600200064006f006b0075006d0065006e007400750073002c0020006b00750072006900650020006c0061006200690061007500730069006100690020007000720069007400610069006b007900740069002000610075006b01610074006f00730020006b006f006b007900620117007300200070006100720065006e006700740069006e00690061006d00200073007000610075007300640069006e0069006d00750069002e0020002000530075006b0075007200740069002000500044004600200064006f006b0075006d0065006e007400610069002000670061006c006900200062016b007400690020006100740069006400610072006f006d00690020004100630072006f006200610074002000690072002000410064006f00620065002000520065006100640065007200200035002e0030002000610072002000760117006c00650073006e0117006d00690073002000760065007200730069006a006f006d00690073002egt LVI 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 NLD (Gebruik deze instellingen om Adobe PDF-documenten te maken die zijn geoptimaliseerd voor prepress-afdrukken van hoge kwaliteit De gemaakte PDF-documenten kunnen worden geopend met Acrobat en Adobe Reader 50 en hoger) NOR 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 POL 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 PTB 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 RUM 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 RUS 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 SKY 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 SLV ltFEFF005400650020006e006100730074006100760069007400760065002000750070006f0072006100620069007400650020007a00610020007500730074007600610072006a0061006e006a006500200064006f006b0075006d0065006e0074006f0076002000410064006f006200650020005000440046002c0020006b006900200073006f0020006e0061006a007000720069006d00650072006e0065006a016100690020007a00610020006b0061006b006f0076006f00730074006e006f0020007400690073006b0061006e006a00650020007300200070007200690070007200610076006f0020006e00610020007400690073006b002e00200020005500730074007600610072006a0065006e006500200064006f006b0075006d0065006e0074006500200050004400460020006a00650020006d006f0067006f010d00650020006f0064007000720065007400690020007a0020004100630072006f00620061007400200069006e002000410064006f00620065002000520065006100640065007200200035002e003000200069006e0020006e006f00760065006a01610069006d002egt SUO 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 SVE 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 TUR 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 UKR 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 ENU (Use these settings to create Adobe PDF documents best suited for high-quality prepress printing Created PDF documents can be opened with Acrobat and Adobe Reader 50 and later) gtgt Namespace [ (Adobe) (Common) (10) ] OtherNamespaces [ ltlt AsReaderSpreads false CropImagesToFrames true ErrorControl WarnAndContinue FlattenerIgnoreSpreadOverrides false IncludeGuidesGrids false IncludeNonPrinting false IncludeSlug false Namespace [ (Adobe) (InDesign) (40) ] OmitPlacedBitmaps false OmitPlacedEPS false OmitPlacedPDF false SimulateOverprint Legacy gtgt ltlt AddBleedMarks false AddColorBars false AddCropMarks false AddPageInfo false AddRegMarks false ConvertColors ConvertToCMYK DestinationProfileName () DestinationProfileSelector DocumentCMYK Downsample16BitImages true FlattenerPreset ltlt PresetSelector MediumResolution gtgt FormElements false GenerateStructure false IncludeBookmarks false IncludeHyperlinks false IncludeInteractive false IncludeLayers false IncludeProfiles false MultimediaHandling UseObjectSettings Namespace [ (Adobe) (CreativeSuite) (20) ] PDFXOutputIntentProfileSelector DocumentCMYK PreserveEditing true UntaggedCMYKHandling LeaveUntagged UntaggedRGBHandling UseDocumentProfile UseDocumentBleed false gtgt ]gtgt setdistillerparamsltlt HWResolution [2400 2400] PageSize [612000 792000]gtgt setpagedevice


754

phrases in the antecedent and consequent clauses must be identical in number form and reference In the present paper we add the further observation that Chinese wh-conditionals can come with some definiteness flavor akin to free relatives in English

The most frequently cited account of Chinese wh-conditionals by Cheng amp Huang (1996) treats wh-expressions in wh-conditionals as indefinites This analysis is somewhat problematic however because it appears to violate the novelty condition which requires each indefinite to introduce a novel entity into the domain of discourse yet only the first of the pair of indefinites in wh-conditionals serves this function In part to circumvent the problem Chierchia (2000) proposes that wh-indefinites in Chinese are indefinite pronouns (ie pronominals) As such they can appear in both the antecedent and in the consequent of wh-conditionals without violating the novelty condition However Chierchiarsquos account fails to explain why wh-indefinites display Principle C effects which suggests that wh-indefinites are r-expressions rather than pronominals

We propose to circumvent these problems by making two assumptions One is that wh-indefinites encode uniqueness just as indefinites do in donkey conditionals (a Russellian semantics for uniqueness is adopted to capture the definiteness effect) The second assumption is that wh-conditionals constitute identity statementsa special kind of topic-comment structure that is not subject to the novelty condition The second assumption reconciles the tension between the novelty condition and Principle C We contend that this novel analysis receives cross-linguistic support Several similarities between Chinese wh-conditionals and correlatives are explained (i) both structures display definiteness effects (ii) both are subject to a matching requirement and (iii) both are open to definitegeneric interpretations

We offer a unified analysis for these apparently unrelated constructions that crop up in historically unrelated languages (eg Chinese vs Hindi) Lacking lexical items like Hindi bhii and English ever or an explicit conditional word like if Chinese wh-conditionals are a cross between being correlatives and conditionals On the one hand when a definiteness flavor is added wh-conditionals strongly resemble free relatives (as well as correlatives in languages such as Russian Vietnamese Hungarian etc) in that they make reference to a particular (unique) individual in a particular situation On the

Thornton Katherine Demuth Nobu Akagi Aijun Huang Anna Notley Alessio Ursini Hao Yin Min Liao Miao-ling Hsieh Peng Zhou and Likan Zhao Finally we would like to thank Hao Yin Miao-ling Hsieh and Yatin Mahajan for supplying critical judgment data

1 In Cheng amp Huang (1996) conditionals (which they call Chinese donkey sentences) are partitioned into two classes bare conditionals and ruguo-conditionals While the former requires the presence of a pair of identical wh-phrases (one in each clause) the latter only allows a pronominal-anaphor (eg overt pronouns empty categories) We focus on bare conditionals in this paper


755








756












757











758













759













760












761











762








763












764












765













766












767

















768











769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













755








756












757











758













759













760












761











762








763












764












765













766












767

















768











769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













756












757











758













759













760












761











762








763












764












765













766












767

















768











769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













757











758













759













760












761











762








763












764












765













766












767

















768











769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













758













759













760












761











762








763












764












765













766












767

















768











769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













759













760












761











762








763












764












765













766












767

















768











769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













760












761











762








763












764












765













766












767

















768











769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













761











762








763












764












765













766












767

















768











769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













762








763












764












765













766












767

















768











769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













763












764












765













766












767

















768











769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













764












765













766












767

















768











769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













765













766












767

















768











769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













766












767

















768











769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













767

















768











769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













768











769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













769










770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













770














771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













771













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













772












773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













773









774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













774





comes first






775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













775















776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













776















777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













777










778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













778











779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













779













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













780










781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













781










782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













782










783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













783











784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













784













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













785












786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













786















787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













787














788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













788








789










790











791









792







793







794


















795



















796



















797




798



南京大學1













789










790











791









792







793







794


















795



















796



















797




798



南京大學1













790











791









792







793







794


















795



















796



















797




798



南京大學1













791









792







793







794


















795



















796



















797




798



南京大學1













792







793







794


















795



















796



















797




798



南京大學1













793







794


















795



















796



















797




798



南京大學1













794


















795



















796



















797




798



南京大學1













795



















796



















797




798



南京大學1













796



















797




798



南京大學1













797




798



南京大學1













798



南京大學1












Documents

Chinese wh conditionals