Chapter 4 The Cleft Pronoun and Cleft Clause

Chapter 4 The Cleft Pronoun and Cleft Clause

Chapter 4

The Cleft Pronoun and Cleft Clause

This chapter focuses on the nature of the cleft pronoun and the cleft clause, and on the syntactic relation that holds between the four subcomponents of the cleft construction. It will be argued (1) that the cleft pronoun has referential status; (2) that the cleft clause is a relative clause; (3) that the cleft pronoun and the cleft clause function as a discontinuous constituent at the level serving as input to pragmatic interpretation; and (4) that the clefted constituent and the cleft clause form a syntactic constituent. I will suggest, finally, that all four of these requirements are satisfied by assuming a structure along the lines of (1) as the S-structure representation of the cleft construction:

(1)

IP it

j

I’ was

k

VP t

k

VP

NP who

i

George

CP

j

IP

4.1

The cleft pronoun

Although most analysts consider the cleft subject pronoun to be an expletive, dummy pronoun which is a mere grammatical filler with no semantic content, this view has occasionally been challenged. Thus, Bolinger 1972b takes the position that the cleft pronoun has ‘low information but not vague reference,’ and Gundel 1977 proposes that the cleft pronoun makes ‘pronominal reference to the topic of the sentence.’ Borkin 1984 adopts the view that the initial it ‘suggests the already known existence of a referent,’ with the proviso that the intended referent is generally

‘clarified’ only in conjunction with the information expressed in the cleft clause. The purpose of this section is to present evidence in favor of the view that the cleft pronoun has semantic content.

76

4.1.1

T

H

-

CLEFTS

As support for his referentiality thesis, Bolinger points to the existence of clefts with pronouns other than it as subject (that, the stuff, this, they, those, these, and we ), such as those in (2), concluding that,

‘in view of this range of possibilities, no special status can be conceded to it, beyond the fact that it is unadulterated ‘identity,’ uncommitted to number, person, or semantic content.’

(2) a.

What are you so upset about? — It’s (that’s, the stuff’s) MY money you’re spending!

b.

What difference does a little dispute make now and then? — ‘OK, but this was with his BOSS that he was having the argument.’

Other linguists have recognized the existence of th-clefts without drawing the conclusion that the cleft pronoun has referential content. Jenkins 1975 notes the similarity of that-clefts such as (3b) both to the it-cleft such as (3a) and to there-clefts such as (3c). He argues that additional similarities between there-clefts and existential-there sentences such as (3d) support the derivation of the latter as an instance of the former.

(3) a.

It’s Bill Smith (who is) standing on the corner.

b.

That’s Bill Smith (who is) standing on the corner.

c.

There’s Bill Smith standing on the corner.

d.

There’s a man standing on the corner.

Wirth 1978 also recognizes the existence of th-clefts, treating the cleft pronoun as the determiner of a subject noun phrase containing an empty head noun restrictively modified by the extraposed relative (cleft) clause, and viewing the cleft pronoun it as a pronominal allomorph of the definite determiner the—essentially the position that I will be arguing for below.

By far the most extensive discussion of th-clefts is given by Ball 1977, 1978. Ball coined the term ‘th-cleft’ to refer to clefts containing pronouns other than it (i.e. this, that, these, those, they ), and argued that th-clefts and it-clefts share all syntactic and semantic properties and thus constitute a syntactically unified class. As Ball points out, the th-clefts in (4), like the it-clefts in (5), exhibit the alternation between specificational (in the (a) examples) and predicational (in the (b) examples) interpretations that was discussed in Chapter 3:

(4) a.

But this is Ethel Schuster we’re talking about, not Jeanette.

[Telephone conversation, 7/89] b.

Make no mistake, that was not a grudge that Nancy Reagan was holding as she chatted about Donald T. Regan in a recent radio interview.…

[The New York Times, 10/22/89, E7]

77

(5) a.

It was the maid he asked, not the landlady.

[Knox, The Footsteps at the Lock, p. 136] b.

Because it’s a country that we have here, and all of Canada is involved.

[B. Mulroney, The Vancouver Sun, 7/30/90]

In what sense can the cleft subject pronoun be said to be referential? I suggest we take the neuter form of the pronoun as a clue. Although neuter pronominals cannot ordinarily be used to refer to people—cf. (6a)—they may be so used in identification sentences, as in (6b):

(6) a.

I ran into at the supermarket. He/*it/*this/*that was buying food for a party.

b.

A: Who is it/this/that?

B: It/this/that is my brother-in-law, John Smith.

Neuter demonstrative pronouns can also be used to refer to people in ‘caricature’ sentences, which are used to assert that the referent has a certain characteristic property:

(7) a.

[Did Bush put the speculation to rest about his involvement in the Iran-Contra scandal?] …George Bush did it again last night. I mean, this is a man who is

the Harold Lloyd of American politics. He finds himself out on perches, hanging onto the hand of a clock on top of a New York building, and somehow he clambers back in. I guess the question I have is, how does he get up there?

[Mark Shields, The McLaughlin Group, 1/9/88] b.

[Was Howard Baker a good choice to replace Regan as Reagan’s Chief of Staff?]

…I think he’s a great choice. The fundamental problem of the President is going to be dealing with Congress. This is a man who can negotiate, who

understands conciliation, who’s got probity and understands what is correct

and prudent, and what is crazy in this administration. We need some settling down, Fred.

[Morton Kondracke, The McLaughlin Group, 2/28/87]

The pronouns in both (6b) and (7) are clearly referential in some sense, though in neither case do they have quite the character of the ordinary pronoun in (6a). The speaker expects the addressee to realize who is being referred to, but does not expect the addressee to be already familiar with all relevant properties of the referent—the point of the utterance is to inform the addressee of one such property. Note also that both identification and caricature sentences, like clefts, are copular sentences; and, more specifically, copular sentences of the predicational subtype:

(8) a.

This is a man who can negotiate.

e <e,t> b.

This is my brother, Bill.

e <e,t>

I suggest that the predicational cleft pronouns in (4b) and (5b) are also neuter referential pronouns

(of type e). Furthermore, if we adopt the proposal of Williams and Partee that the referential and

78 predicative arguments of the copula can occur in either order (for discussion see section 4.1, and

Chapter 3 above), we can view the specificational cleft pronouns in (4a) and (5a) simply as predicative pronouns (of type <e,t>). Both referential (e-type) and predicative (<e,t>-type) pronouns are ‘referential’ in the broad sense of specifying a referent in the universe of discourse.

4.1.2

T

H

-

CLEFTS AND COGNITIVE STATUS

The predicate denoted by the cleft clause of an it-cleft or th-cleft can have any cognitive status appropriate to definite expressions in general, as the examples of th-clefts from natural discourse in

(9)-(21) below show. The predicate denoted by the cleft pronoun + cleft clause is speaker activated in (9)-(12), addressee activated in (13)-(14), extralinguistically activated in (15)-(17), and familiar but not activated in (18)-(21):

Speaker activated:

(9) M: There IS something that happened here that you might not know about.

N: Fred Lukermann resigned.

M: Oh, that’s right—you talked to Karen.

N: No, this was JeaNETTE who told me—I talked to her last Sunday.

[telephone conversation; 10/89]

(10) K: Michael asked me last FALL to teach 1005.

(11)

N: Oh?

K: I mean, this was the beginning of fall QUARter when he asked me.

[conversation, 12/31/89]

I wasn’t surprised by the massacre in China. [pause]

This is not

I

OWA we’re talking about. —This is a DIFFERENT SO

CI

ETY.

[Eric Severeid, interview on CSPAN by Brian Lamb, 12/31/89]

(12) NF: …And then, one morning, about three or four or five mornings before I was due to get out, I was lying in bed and someone, one of, one my fellow soldiers came by and and shook my bed and said, Come on Fredzo, get up… and the Sergeant himself said, ‘Leave him alone, he’s too short.’

KF: Hmm.

NF: I mean, the, that was the platoon sergeant that said that. I call that a pretty good guy.

[Frederickson tapes, Christmas 1988]

79

Addressee activated:

(13) NH: I fly to Michigan on Thursday and come back late on Friday. But I have to leave again on Saturday already . . . .

Mom:

When IS that on T

HURSDAY

that you

GO

?

NH: 2:25

[telephone conversation, 2/14/89]

(14) [Reading the address on an envelope]

Mom: ‘Okabena, Iowa’ (laughs)

Neil: Does it—is that what it says?

Mom: Yeah.

Neil: Oh boy.

Mom: ‘Minnesota 56161.’

Dad: Who is that who wrote that, that put Iowa on it?

Neil: Okabena, Iowa.

Mom: Homer, Homer York; Homer.

Dad: Oh.

[Frederickson Tapes, Christmas 1976]

Extralinguistically activated

(15) They react sharply, though, to the suggestion that they are mercenaries. ‘That’s

the French flag you see flying over there,’ Pierre Dufour, a former legionnaire, pointed out. ‘This is part of the French Army. In any event, dating back to the

Crusades, foreigners have always fought under different flags in Europe.’

[New York Times, 5/9/90, A4, ‘Are Glory Days at an End for the Foreign

Legion?’]

(16) RZ: Oh, that was the gaRAGE I saw first.

[conversation, 7/20/90, i.e. not the house]

(17) ‘Pardon me, sir,’ he said. ‘If you can spare a few moments, I’d like a word with you.’

Popple turned quickly and then smiled, holding out his hand. ‘Captain Vachell, isn’t it? Of course, if there’s any way at all I can help, I’ll do it. This is a terrible

thing that’s happened, terrible.

[Elspeth Huxley, Murder at Government House, p. 83]

Familiar but not activated.

(18) K: [answering phone] Linguistics.

N: Hi!

K: Hi!

N:

Was that a sign or a plaque that was on that bridge?

80

(19)

K: Wow! Uh. . a sign or a plaque. . .I would say it was. . a sign. . that’s a good question. I’d say it’s a plaque cause a sign is too ambiguous. A plaque has to be attached to the building. And this was, attached to the bridge. I’d say a plaque is more specific.

[telephone conversation, 6/8/89]

… ‘My good friend and parishioner Botts Tempe told me about it. He was catching for the over-forties when the call came in.’

‘Oh, that was your picnic he was at.’ The comment was surprised out of her, and she regretted it immediately.…

[Dead in the Scrub, B.J. Oliphant]

(20) …Mr. and Mrs. Nev Barnes. She bakes bread and pies and sells them, and he snitches some of the proceeds and buys hooch from a bootlegger named

Henrietta…”

Was that her bread at breakfast?

‘Yes. Salt-rising. You ate four slices.’

[Death of a Dude, Rex Stout]

(21) NH: That’s the reason I don’t want to go to Miami!

BP: Yeah. Wasn’t that somewhere in Southern Florida where they thought those

people got AIDS from bug bites — getting bit a hundred times a night or something, because the place was so roach infested?

[conversation, 2/89]

4.1.3

A

PRAGMATIC CO

-

OCCURRENCE CONSTRAINT

It is important to notice, however, that choice of cleft pronoun is not entirely free. The substitutions in (22) show that there is a pragmatic assymmetry between this-clefts and that-clefts. Thisclefts can be appropriately used only if the predicate denoted by the cleft clause is activated, while

that-clefts require merely that the predicate be familiar:

(22) a.

Wasn’t it/#this/that somewhere in Southern Florida where they thought those

people got AIDS from bug bites…?

b.

Was it/#this/that a sign or a plaque that was on that bridge?

c.

Oh, it/#this/that was your picnic he was at.

d.

Was it/#this/that her bread at breakfast?

Cleft pronouns thus conform to the general pragmatic constraints on the use of different forms of referring expression which are encoded in the Givenness Hierarchy, repeated in (23) from Chapter 2:

81

(23)

Givenness Hierarchy

in uniquely focus > activated > familiar > identifiable >

{

it



that



this

this N





{ { type identifiable



a N

 indefinite this N





Crucially, however, cleft pronouns are subject to the necessary conditions associated with the determiners rather than the pronominals. For example, in (18) above, the cleft pronoun meets necnot activated. The information encoded in the cleft clause supplements the deictic information encoded in the cleft pronoun in precisely the way that the nominal information in a noun phrase supplements the deictic information encoded in the determiner. The cleft pronoun thus seems to function more like a determiner co-specifying the predicate denoted by the cleft clause than like a pronominal referring independently to the predicate.

Note also the relative lack of restriction on it as a cleft pronoun: the predicate denoted by the cleft clause may have any cognitive status allowable in clefts in general. The pronoun and the cleft clause taken together can thus be seen to have the referential properties of full noun phrases with definite articles (i.e. the N in the Givenness Hierarchy).

1 This observation lends further support to the hypothesis that the cleft pronoun and cleft clause function as a discontinuous constituent.

The infelicity which results from the substitution of a proximal pronoun for a distal pronoun when the referent fails to be included in the ‘speaker’s context space,’ as in (24), shows that cleft pronouns also obey the ‘speaker activation condition’ on proximal demonstratives, discussed in

Gundel, Hedberg and Zacharski 1989.

2

1

Cf. the proposal of Wirth 1978 that the cleft pronoun it is the pronominal allomorph of the definite determiner the, and also the proposal of MacClaran 1982 that corresponding demonstrative determiners and pronominals are allomorphs. The more recent proposal of Abney 1987 that the determiner is the head of the noun phrase (the ‘DP hypothesis’)—and can hence can appear either with or without a complement—is clearly also relevant.

2

The ‘speaker activation condition’ accounts for the inappropriateness of using a proximal demonstrative to access a referent that is not construable as included in the ‘speaker’s context space’, as in (i):

(i) A:

B:

Have you seen the neighbor’s new dog?

Yes, and that/??this dog kept me awake all night.

Motivated exceptions include clarification questions and shared discourse topics.

82

(24) a.

It/#these/those are my cigarettes you got there buddy.

[Ball 1978] b.

It/?these/those are not independent, parallel assumptions that you’re dis-

cussing.

[Ball 1978]

The speaker-activation condition is sometimes obscured, however, by the fact that many addresseeactivated th-clefts are clarification questions. It is generally the case that a proximal demonstrative can be felicitously used to refer to an addressee-activated referent as a signal that the speaker’s contribution is not intended as an interruption, as illustrated in (25a). Substitution of a proximal pronoun for the attested distal pronoun in (25b) would thus transform the utterance into a clarification question like the addressee-activated this-cleft in (25c):

(25) a.

Mom:

Neil:

Is this your truck?

No, the beat-up old lousy van I have to drive.

[Frederickson Tapes, New Year’s Eve 1975] b.

When is it/this/that on Thursday that you go?

c.

A: Ray and Chris are both Psych majors, and she’s marrying one.

N: This is Chris you’re talking about, right?

A: Yeah.

[Frederickson Tapes, New Year’s Eve 1975]

Substitution of a distal for a proximal cleft pronoun can also result in infelicity, perhaps because such a substitution would generate the conversational implicature that the speaker considers the referent to be exclusive to his/her own context space:

(26) a.

I wasn’t surprised by the massacre in China. It/this/#that’s not Iowa we’re

talking about. It/this/#that’s a different society.

b.

A: Oh that’s right—you talked to Karen.

B: No it/this/?that was Jeanette who told me—I talked to her last Sunday.

c.

A: It may be that fanatical Muslims are even more stupid than fanatical

Christians.

B: But ?it/these/?those are students who are rioting.

4.1.4

S

PECIFICATIONAL AND PREDICATIONAL TH

-

CLEFTS

Ball 1977 argues that th-clefts as well as it-clefts exhibit the specificational-predicational distinction.

She argues further that in th-clefts with plural clefted constitutents, the distinction is morphologically encoded in the form of the cleft pronoun. In specificational clefts, the cleft pronoun is invariant with respect to person and number features of the clefted constituent—it is always third-person and singular:

83

(27) a.

It’s the contras who have cried uncle.

[McLaughlin Group, 3/25/88] b.

This is not language teaching problems that we’re talking about.

[Ball 1978] c.

This is Ford and Kissinger we’re dealing with, not two boy scouts.

[Ball 1978] d.

That was Mom and me that, and other people who laughed about ‘glossy balls.’

[Frederickson tapes] e.

That was two chicken dinners I ordered, not two shrimp boats.

[Ball 1978] f.

That was our right-wingers who got us into that you see.

[Geluykens 1983]

In predicational th-clefts, on the other hand, the cleft pronoun agrees in person and number with the clefted constituent—and is thus morphologically plural if the clefted constituent is plural.

This can be seen in examples (28)-(32), which exhibit all the characteristics of predicational clefts identified earlier in §3.3—i.e., the clefted constituents are indefinite and/or contain contain adjectives bearing primary accent:

(28) Yeah, I have the same concern about whether we’re getting the straight dope from these sleazeballs, these Iranian nuts who — They’re just fanatics who are holding him. If they’re concerned about the guy, why don’t they let him go, rather than insisting on 100 hostages.

[JG, McLaughlin group, 3/27/87]

(29) BP: It may be that fanatical Muslims are even more stupid than fanatical Christians.

EM: But these are STUdents who are rioting.

[Conversation, looking at a newspaper photograph, 2/14/89]

The examples in (30) involve non-demonstrative cleft pronouns. As would be predicted, all of the

they-clefts in the data can be independently identified as predicational:

(30) a.

They were the saddest little letters you wrote me.

[Ball 1978] b.

Thus they have to be surface segments that they’re talking about.

[Ball 1978] c.

They are royal horses you are catching. Let them go.

[Ball 1978] d.

They were English hands that dragged him up to the tree of shame.

[Ball 1978] e.

We are erstwhile friends and néighbors who are fighting with each other.

[Bolinger 1972b]

84

The examples in (31) contain plural proximal demonstrative cleft pronouns, and those in (32) contain plural distal ones. Again the clefted constitutents are indefinite and/or contain accented adjectives:

(31) a.

These aren’t floor samples that you’re saving on, but the best seats in the house.

[Ball 1978] b.

These are serious charges that you’re making.

[Ball 1978] c.

No, these aren’t human beings one deals with on Saturday nights.

[Ball 1978] d.

These are serious grammatical problems that he’s touching there.

[Ball 1978]

(32) a.

Seeing is believing! Those are real eyeglasses that Mickey is wearing.

[Ball 1978] b.

Those are not independent, parallel assumptions that you’re discussing.

[Ball 1978] c.

Those are my cigarettes you got there, buddy.

[Ball 1978] d.

Those were T-tests he was requesting.

[Ball 1978] e.

Those were Varick’s pearls about her neck.

[Ball 1978]

In sum, the morphological alternation exhibited by the cleft pronouns of plural th-clefts constitutes strong evidence in support of the hypothesis that the specificational/predicational distinction is applicable to clefts as well as to pseudoclefts, 3 which in turn supports the hypothesis that the cleft pronoun has referential content.

4

4.1.5

E

XPLETIVE PRONOUNS MORE GENERALLY

It is interesting to note that th-clefts are not confined solely to English. Examples from German are shown in (33), and examples from Dutch are shown in (34): 5

3

The restriction of person-number agreement to predicational cleft pronouns is consistent with their analysis as individual-level (type e) referring expressions. (cf. also the discussion in Heggie 1988 about the lack of agreement features in French clitics which denote propositions—i.e., propositions are invariantly referred to with the third person singular clitic pronoun le.)

4

It would be interesting to explore the ‘deictic harmony’ displayed across the cleft sentence as a whole, in order to explain the strong correlations which evidently exist between proximal/distal cleft pronoun, non-past/past tense copula, and activated/familiar cleft clause (as well as correlations between activated/familiar status of the cleft clause and nonpast/past tense or first/non-first person status of the cleft clause itself).

5

The examples are taken from Smits 1989, who notes concerning German clefts that ‘instead of expletive es the demonstrative pronoun das can be used in its deictic sense when the focus is a NP,’ and about Dutch clefts that ‘instead

85

(33) a.

es war der Chef der diesen Wagen gekauft hat it was the boss who this car bought has

‘it was the boss who bought this car’.

b.

das bin ich,den du dort am Rande siehst that am I, whom you there at edge see

‘that am I whom you see there at the edge.’

(34) a.

het was mijnbroer die Marie op heterdaad betrapte it was `my brother whom Marie [redhandedcaught]

‘it was my brother whom Marie caught redhanded.’ b.

dat was zijnvrouw die ik net aan de telefoon had.

that was his wife whom I just now ontelephone had

‘that was his wife whom I had on the telephone just now.’

Sentential subject extraposition constructions in English can also occasionally contain a demonstrative subject pronoun instead of it :

(35) … he said, ‘We don’t like slant-eyed Eskimo bastards around here.’ Funny thing, him saying that. Until I was about twenty I always thought our eyes were the proper shape…

However, with this guy I wasn’t taking.

That’s odd you don’t like slant-eyed Eskimo bastards,’ I said. ‘I don’t object to tall, skinny half-breed bastards.’

[Scott Young, Murder in a Cold Climate, p. 82]

It is worth noting, however, that expletive it cannot always be replaced with a demonstrative.

Demonstrative pronouns are impossible as subjects of ‘weather’ and ‘raising-verb’ sentences in any context:

(36) a.

It/*this/*that is raining this evening in Ithaca, Elmira and Cortland.

b.

It/*this/*that appears that Cuomo will run for a third term as Governor.

Expletive cleft subjects also pattern differently from raising and weather verb subjects in (standard)

French and Russian. Thus, the masculine singular pronoun il is used in weather and raising-verb sentences in French, while the demonstrative pronoun ce is used in clefts:

(37) a.

Il/*ce neige it/*that snows

‘It’s snowing’ b.

il/*ce me semble que tu as tort.

it/*that to.meseems that youhave wrong

‘It seems to me that you are wrong.’ of expletive het the demonstrative pronoun dat … can be used in its deictic sense when the focus is a NP, especially a person.’

86 c.

*il/c’est un livre queje vois.

*it/that’s a book that I saw

It was a book that I saw.

A similar pattern is found in Russian. Weather and raising-verb sentences obligatorily lack an overt subject, whereas clefts contain the demonstrative pronoun éto:

6

(38) a.

(*èto)morozit

(that)is-freezing

‘(It is) freezing’ b.

(*èto) kazetsjacto on usel

(that) seems that he left

‘(It) seems that he left’ c.

èto Ivana javidel that Ivan+accI saw

‘that/it (was) Ivan I saw.’

I conclude that the pragmatic properties of th-clefts support the hypothesis that the cleft pronoun has a critical role to play in the interpretation of clefts, and that the cleft pronoun and cleft clause function as a unit during pragmatic interpretation. Th-clefts also provide additional support for the hypothesis that clefts, like other copular sentences, have predicational as well as specificational interpretations.

4.2

The cleft clause

This section presents arguments in favor of the hypothesis that the cleft clause is structurally a relative clause, a view which I defend against traditional as well as contemporary attacks. Crosslinguistic evidence is cited which suggests that the structural similarity of relative and cleft clauses is universal in scope.

4.2.1

S

IMILARITIES BETWEEN CLEFT AND RELATIVE CLAUSES

The examples in (39) show that cleft clauses and restrictive relative clauses are structurally very similar: both contain a ‘gap,’ both permit an identincal range of complementizers, and both permit

‘pied piping’: 7

(39) a.

It was the man Ø/that/who/whom I saw first.

8

6

The Russian examples are taken from Gundel 1977. The Russian copula is not expressed in the present tense.

7

Note, however, that Rochemont 1986 argues that pied piping is less generally acceptable in clefts than in ordinary relative clauses.

87

The man Ø/that/who/whom I saw first won the race.

b.

It was the book Ø/that/which I read first.

The book Ø/that/which I read first was about horses.

c.

It was the young man whose dog died last summer.

The young man whose dog died last summer is still upset.

d.

It was the red folder in which we found the clue.

The red folder in which we found the clue was in my office.

The similarity between cleft clauses and relative clauses is not restricted to English.

9

In languages which have distinct complementizers for introducing relative clauses and sentential complements, it always seems to be the relative complementizer which is used in cleft clauses. This is true, for example, in Swedish, Irish, and Arabic, as shown respectively in (40)-(42):

10

,

11

(40) a.

jag såg den lampan som Anders köpte i Umeå

‘I saw the lamp that Anders bought in Umea.’

8

Delahunty notes dialectal variation in the choice of wh- complementizer in clefts in English. Some dialects accept who

(e.g. that of Heggie 1988) only in subject clefts, while others accept who(m) in object clefts as well. Some nonstandard dialectal features are the use of what as cleft complementizer in Irish and Scottish dialects, as in (i) and (ii) from

Delahunty 1982:131, and the lack of a complementizer even in subject clefts as in (iii) and (iv) from The Unpleasantness

at the Bellona Club, by Dorothy Sayers (24, 38):

(i) It was a falling tree what killed him.

(ii) It was the accident what upset him.

(iii) It was Weston was at the door all morning, my lord.

(iv) It was a gentleman spoke to me.

9

See also Schachter 1973 for discussion of similarities between focus constructions and relative clauses in several languages.

10

The Swedish data come from Smits 1989 (pp. 429, 433), the Irish data from McCloskey 1979 (p. 110), and Chung and McCloskey 1987 (218, 222). The Arabic data comes from Maher Bahloul (personal communication), who points out that only NP clefts of this sort are possible. Data from Smits 1989 (p. 56) indicates that the Zurich dialect of Swiss

German consistently uses wo (< ‘where’) in relatives, but daz (‘that’) in sentential complements, as shown in (i), but no data on clefts is presented:

(i) d e r ankx e

[wo d e

göti aakee hæt [daz mer em nο gon æs e]]

hæt gruuzig k s mökxt the butter that the g.f. claimed has that one it still could eat has horrible smelled

‘the butter that my godfather claimed that one could still eat it smelled horrible.’

11

Rizzi 1987 (ch. 2: 60) discusses the phenomenon of distinct relative and sentential complementizers, mentioning also

Hebrew relative ?asher versus sentential she. He suggests that such relative complementizers exhibit agreement either with their specifiers or with the head NP of the relative clause. Although he doesn’t discuss clefts in this context, the cleft facts are clearly relevant.

88 b.

det var denna rock som han köptel

‘It was this coat that he bought.’ c.

Ingrid sade att det var denna rock som han köpte

‘Ingrid said that it was this coat that he bought.’

(41) a.

Is é Seán aL thigeann ‘na bhaile

Copula him

COMP come home

‘It’s John that comes home.’ b.

an bhean aL chuir isteach air … the woman

COMP put in on-it

‘the woman that applied for it…’ c.

Dúiurt sé go dtiocfadh sé said he

COMP would.come he

‘He said that he would come.’

(42) a.

?inna dzan huwa ?alladhi: xaraja

FOCUS

John.

ACC he.

NOM who.

3M

left

‘It was John who left.’ b.

?ar-rajulu ?alladhi: xaraja…

DEF

-man.

NOM who.

3M

left

‘the man who left…’ c.

dhannat nansi ?anna dzan xaraja thought.

FEM

Nancythat John.accleft

‘Nancy thought that John left.’

In Cakchiquel, as in other Mayan languages, clefts, relative clauses, and wh-questions require a special ‘focus antipassive’ verb stem and agreement pattern when the ‘gap’ corresponds to a transitive subject.

12

(43) a.

ya riox-ri ri s - qu - cap - o rie

Cop we-dem Rel Asp-1pAbs- catch-Foc.ap them

‘It was we who caught them.’ b.

riox ri s - qu - cap - o

rie

we Rel Asp-1pAbs- catch-foc.ap them

‘we who caught them…’ c.

rIt s - Ø - a -bix ci riox s - e - qa - cap rie you Asp-3sAbs-2sErg-say that we Asp-3pAbs-1pErg-catch them

‘You said that we caught them.’

12

For discussion, see Hedberg 1989.

89

4.2.2

J

ESPERSEN

S ARGUMENTS

Despite such clear similarities, analysts have sometimes denied that cleft clauses are relative clauses.

Thus, in arguing against his earlier ‘transposition’ analysis, Jespersen 1937 suggests that the introductory that in clefts is a subordinating conjunction rather than a relativizer.

13 As evidence he notes that adverbial clefts in Danish use the sentential complementizer rather than the relative complementizer, and in ‘Vulgar Danish’ the sentential complementizer is used even in NP-clefts:

(44) a.

Det er her slaget (at) skal stå.

‘It is here (that) he must come’ b.

Når var det (at) han døde?

‘When was it (that) he died?’ c.

Det var Jens at jeg saa.

‘It was Jens that I saw’

Examples from Smits 1989 confirm this observation for PP and sentential adjunct clefts in Danish

(45) and for PP clefts in Norwegian (46). In Swedish (47), som is used instead of att, but there is a preference for no complementizer at all:

(45) a.

det var over droningen, (at)/*som han skrev en bog

‘it was about the queen that he wrote a book.’ b.

det var fordi at musikken var så høj, (at)/*som han ringede til politiet.

‘It was because the music was so loud that he called the police.’

(46) a.

det var om Etiopia at/*som han skrev en bok

‘It was about Ethiopia that he wrote a book.’ b.

det var på kjøkkenet at/*som Tarald bygde et skip

It was in the kitchen that Tarald built a ship’

(47) a.

det var om Etiopien som Pelle skrev en bok.

‘It was about Ethiopia that Pelle wrote a boot.’ b.

det var i köket (som) Anders byggde en båt

‘It was in the kitchen that Anders built a boat.’

Although these facts are certainly very interesting, they show only that NP clefts and oblique clefts are not structurally indistinguishable. This is clear already from the fact that some languages permit only NP-clefts (e.g. Arabic, cf. footnote 10 above). The Scandinavian data in (45)-(47) shows that

13

Jespersen 1947: ‘Is that (que ) a relative word, or is it the same ‘conjunction’ that we have in I think that he died here,

Je crois qu’il mourut içi ?’

90 complementizer choice in oblique clefts is subject to dialectal variation. However, we cannot draw conclusions about NP clefts from instabilities exhibited only by oblique clefts.

4.2.3

G

ENERATIVISTS

ARGUMENTS

Recent generative accounts of clefts have also denied relative-clause status to the cleft clause.

Delahunty 1982, Rochemont 1986, and Heggie 1988 assume that to be the basic complementizer in clefts, with wh-complementizers inserted only stylistically, by ‘analogy’ with relative clauses.

14

Delahunty argues that the existence of that-complementizer clefts with no wh-complementizer counterparts supports the analysis of that-complementizers as basic in clefts. This argument is unconvincing, however, given that CP’s, AP’s, AdvP’s, and PP’s are not ordinarily modified by relative clauses as NP’s are, there is no reason to expect that non-NP clefts should display the full range of complementizer choice open to NP relatives and clefts. (It would make more sense to view non-NP clefts as peripheral analogs of NP clefts, rather than the other way around.)

Delahunty and Rochemont claim additionally that where and when are not permitted as complementizers in clefts. They note the existence of where- and when-clause copular sentences such as those in (48a) which, unlike clefts, lack that-clause (48b) and simple sentence counterparts (48d) and permit preposing of the wh-clause (48c).

15 I would add that such sentences also lack pseudocleft counterparts (48e).

(48) a.

It was March/sunny when we arrived.

b.

*It was March/sunny that we arrived.

c.

When we arrived, it was March/sunny.

d.

*We arrived March/sunny.

c.

*When we arrived was March/sunny.

But it cannot be concluded from the existence of a class of adverbial-clause sentences which are not clefts, that no adverbial-clause sentences are clefts. Even if the when-clause sentence in (49a) is not a cleft, those in (50a)-(52a) do fit the paradigm associated with clefts.

14

Note that this is a crucial assumption for Rochemont 1986 since on his analysis, the clefted constituent is moved from its D-structure position inside the cleft clause first into COMP and then into post-copular ‘contrastive focus’ position, leaving behind an intermediate trace in COMP. For Delahunty and Heggie, cleft clauses, like relative clauses, contain an empty operator which is moved into COMP at S-structure. Heggie makes crucial use of the proposal that the cleft operator is empty to account for syntactic differences between clefts and pseudoclefts (cf. the discussion in Chapter 3 of

Heggie’s Null Operator Generalization).

15

Delahunty and Rochemont note that some post-copular elements (e.g. late) may appear in both the cleft and the noncleft paradigms. Cf. also Quirk and Greenbaum 1972, p. 416: ‘A wh-pronoun cannot be used at all in cleft sentences where the focal element is an adverbial.’

91

(49) a.

It was still only three when he awoke and he knew he wouldn’t get any more sleep that night.

[Ruth Rendall, Speaker of Mandarin, p. 37] b.

*It was still only three that he awoke c.

When he awoke, it was still only three.

d.

*He awoke still only three.

e.

*When he awoke was still only three.

(50) a.

Michael asked me last fall to teach 1005…I mean

THIS

was the beginning of fall QUARter when he asked me.

[KS, 12/31/89] b.

T

HIS

was the beginning of fall QUARter that he asked me.

c.

When he asked me, it/*this was the beginning of fall quarter.

d.

He asked me at the beginning of fall quarter.

e.

When he asked me was the beginning of fall quarter.

(51) a. A: When is it heavily used?…

B: I think — as Nip pointed out earlier on — it’s June when the pres sure is on for the examination.

A:

It’s the summer term when you’re really stuck.

C: Yes, it is, yes, yes.

[Geluykens 1983, C48] b.

It’s the summer term that you’re really stuck.

c.

When you’re really stuck, it’s the summer term.

d.

You’re really stuck in the summer term.

e.

When you’re really stuck is the summer term.

(52) a. A: Now where did I hear that from?

B: Probably me on the phone was it. It was the day AFTer when I RANG, and we…

[Geluykens, 1983, C4] b.

It was the day AFTer that I RANG c.

When I rang, it was the day after.

d.

I rang the day after.

e.

When I rang was the day after.

92

The where-clause sentences in (53) and (54) also fit the pattern of true cleft sentences.

(53) a.

It is here where the hearty French established a settlement along the frothy

St. Lawrence River and survived the first relentless winter.

[Insight Guide to Canada, p. 15] b.

It is here that the hearty French established a settlement along the frothy St.

Lawrence River c.

*?Where the hearty French established a settlement along the frothy St. Lawrence

River, it was here.

d.

The hearty French established a settlement here along the frothy St. Lawrence

River e.

Where the hearty French established a settlement along the frothy St. Lawrence

River was here.

(54) a.

Yet it is precisely on this point where Ahlquist and I must part company.

[Minneapolis Star and Tribune, Op-ed page, 8/22/87] b.

Yet it is precisely on this point that Ahlquist and I must part company. c.

*?Where Ahlquist and I must part company, it is precisely on this point.

d.

Where Ahquist and I must part company is precisely on this point.

e.

Ahlquist and I must part company precisely on this point.

Moreover, it is difficult to see how the Delahunty/Rochemont/Heggie analysis could be extended to account for genitive clefts with the complementizer whose in a non-ad hoc way, since these do not permit the use of that :

(55) a.

It was the criminal division of the Justice Department whose bungled

investigation of Watergate led to the call for a special prosecutor in the first place.

[Minneapolis Star and Tribune, Op-ed page, 8/25/87] b.

*It was the criminal division of the Justice Department that’s bungled investigation of Watergate led to the call for a special prosecutor in the first place.

c.

It was Captain Grant whose men sang for the queen.

[Smits 1989: 301] d.

It is the mayors whose income the government is trying to lower now.

[Smits 1989: 301] e.

It was Van Buren whose legacy has caused every vice president to rail at the unfairness of his political fate.

[Haynes Johnson, Mpls Star & Trib, 2/16/88]

I conclude, therefore, contrary to Delahunty, Rochemont and Heggie, that cleft clauses and relative clauses exhibit exactly the same range of complementizers. This makes it possible to abandon the rather mysterious suggestion that the appearance of overt wh-operators in clefts is due to a special stylistic mechanism that operates on clefts by analogy with relative clauses. I adopt

93 instead the far simpler hypothesis that cleft clauses are just relative clauses, as suggested by the crosslinguistic similarities between cleft and relative clauses discussed in §4.2.1 above.

16

4.3

The relation between the cleft pronoun and cleft clause

I concluded in §4.1 that the cleft pronoun and the cleft clause function as a unit during pragmatic interpretation. In this section I will present and attempt to rebut arguments which have been raised against the hypothesis that the cleft pronoun and cleft clause constitute a syntactically discontinuous unit. Recall from §3.1 that the most commonly proposed version of what I termed the

‘extraposition approach’ to the structure of clefts views the relation which holds between the cleft clause and the cleft pronoun to be identical to the relation which holds between an extraposed relative clause and its head NP. A second version, proposed in Gundel 1977, views the relation which holds betwen the cleft clause and the cleft pronoun to be identical to the relation which holds between a right-dislocated constituent and its pronominal antecedent. 17

Jespersen 1937 levels against his own earlier ‘transposition’ analysis the criticism that clefts in languages such as Italian lack an overtly expressed cleft pronoun:

(56) quando é un santo che parla, é il Signore che lo fa parlare

‘When it is a saint that speaks, it is God that makes him speak.’

However, as Halvorsen (1978:25) notes, ‘this objection would presumably seem less important to a transformational linguist. A transformational linguist would permit the presence of an underlying it which would function as head for the relative clause.’ Gundel 1977 views the lack of a cleft pronoun in null subject languages as support for her right-dislocation analysis of clefts, since lack of an overt pronominal antecedent for the right-dislocated topic is a general characteristic of right-dislocation in such languages.

A second objection against the extrapositon analysis is leveled by Ball 1977, who argues that non-extraposed restrictive relative clauses headed by pronouns are limited to archaic proverbs, as in

(57).

(57) a.

He who hesitates is lost.

16

Delahunty claims that cleft clauses are essentially distinct from relative clauses in Irish, since the indirect relative strategy, with a resumptive pronoun and the complementizer aN is unavailable in clefts. McCloskey 1979, on which

Delahunty bases his claim, is neutral with respect to this issue. Kari Swingle informs me that she believes the indirect as well as direct strategy to be available in Irish clefts. In any case, the similarities between cleft and relative clauses in Irish outweigh any differences, and thus Irish fails to support Delahunty’s conclusion that cleft clauses and relative clauses are distinct.

17

I am not concerned here with the issue of whether such syntactic relations should be encoded in the grammar procedurally by means of movement rules or purely declaratively by principles of surface structure interpretation.

94 b.

He who lives by the sword, dies by the sword.

c.

Let he who is without sin among you cast the first stone.

While it is certainly true that relative clauses are only marginally headed by personal pronouns in

English, they are quite commonly headed by plural demonstrative pronouns, as in (58):

(58) It has great potential value for those who must read technical documents.

[message from electronic news group]

Moreover, restrictive relative clauses can be headed by plural personal pronouns in Irish (McCloskey

1979: 23), as in (59), and in Japanese, even by singular definite pronouns (Fukui 1986:205), as in

(60):

(59) Sibh-se aL tá tinn, gabhaigi ‘na bhaile you(pl)

COMP are sick go home

(60) Tokyo-no biru-no okuzyoo kara mita

-Gen building-Gentop from (I) saw

Haree-suisei-wa smog-no tame bonyariot

Halley’s Comet-Top smog-Gen due.to faintly nigotte ita ga, Okinawa-no Naha-de

mita

-in saw

sore-wa yozora-ni kukkirito kagayaite-ita.

it-Top night.sky-in vividly shining was.

‘Halley’s Comet that (I) saw from the top of a building in Tokyo was blurred by night sky.’

Thirdly, Jespersen takes the absence of ‘comma’ intonation and pause between the clefted constituent and the cleft clause as evidence against an extraposition analysis. However, there is also no prosodic break before extraposed sentential subjects or extraposed relative clauses: 18

18

This criticism has perhaps been more justifiably raised against Gundel’s 1977 right-dislocated pseudocleft analysis of clefts—since right-dislocated constituents are often preceded by a pause. Gundel 1977 accounts for the lack of a prosodic break in clefts with a reduction rule applying only to clefts. In defense of her analysis, note that Lambrecht

1981 distinguishes two types of right-dislocation structures: unplanned ‘afterthoughts’, which are preceded by a prosodic break; and planned ‘antitopics’, which are not preceded by a break. Cleft clauses may then function as ‘antitopics.’

Gundel’s analysis has also been criticized (e.g. in Carlson 1983) on the grounds that right-dislocated pseudoclefts and clefts have different discourse functions. However Gundel’s point is that the two constructions have identical topic-comment structures — not that they are functionally identical in every respect. Perhaps confusion has arisen on this point because right-dislocated pseudoclefts seem generally to be predicational rather than specificational, at least in the examples from my data:

(i) Maybe you’re thinking, this is elistist, what that man’s saying.

[public radio pledge drive, 10/22/89, WSKG]

95

(61) a.

It’s unfortunate that Republicans keep winning the elections.

b.

Nobody would drink instant coffee who knew anything about espresso.

[Reinhart 1980]

Fourthly, Jespersen points out that cleft clauses but not extraposed relative clauses may lack overt complementizers — compare (62b) and (63b): 19

(62) a.

It was the Colonel who I was looking for.

b.

It was the Colonel I was looking for.

[Jespersen 1947] c.

*It who I was looking for was the Colonel.

(63) a.

A man was living here who the police were looking for.

b.

*A man was living here the police were looking for.

c.

A man who the police were looking for was living here.

However, this difference could result from clashing discourse conditions on ellipted complementizers and extraposed relative clauses. It may be the case that complementizer omission is subject to a general constraint requiring elliptical material to be contextually recoverable (i.e. to be activated). In contrast, extraposed relative clauses, unlike cleft clauses, are always unactivated (c.f. Huck and Na

1990). Consistent with this hypothesis is Prince’s (1978) observation that the complementizer cannot be omitted in a cleft clause that bears primary accent (which is typically unactivated.).

Finally, Jespersen objects that ‘the almost universal agreement with regard to person and number of the verb in the relative clause with the immediate antecedent points in the same direction, i.e. against the transposition theory.’ I would counter, however, that the universality of such agreement patterns is not entirely clear. Whether or not the verb of a subject-gap cleft clause agrees with the clefted constituent varies dialectally, and depends in part on the assignment of nominative case to the clefted constituent. Thus, Akmajian 1970 identifies the three dialects of

American English illustrated in (64), which differ primarily in the morphological case associated with a pronominal clefted constituent. Dialect I requires non-nominative case on all clefted constituents; Dialect II requires nominative case when the clefted constituent denotes the subject of the cleft clause; and Dialect III displays a mixture of the two strategies.

(64) I.

i.

It’s me who is/*am responsible.

ii.

It’s I who *am/*is responsible.

(ii) I think that’s exactly right, what Bob says — he’s got a themeless campaign.

[PB, The McLaughlin Group, 2/19/88]

(iii) But that was a big mistake, what we saw, this business of apologizing, saying it was bureaucrats.

[MK, The McLaughlin Group]

19

Ball 1977 objects also that extraposition would be obligatory in clefts, but optional in non-clefts—compare (62c) and

(63c).

96 iii.

It’s me/*I who you saw.

iv.

It’s you and me/*I who are/*is responsible.

II. i.

It’s me who *is/*am responsible.

ii.

It’s I who am/*is responsible.

iii.

It’s I/*me who you saw.

iv.

It’s you and I/*me who are/*is responsible.

III. i.

It’s me who is/*am responsible.

ii.

It’s I who am/*is responsible. iii.

It’s me/*I who you saw.

iv.

It’s you and me/I who are/*is responsible.

The verb of a subject-gap cleft clause always agrees in number with the clefted constituent, but only agrees with it in person if the clefted constituent is nominative. Akmajian argues that the personagreement system of Dialect I supports his extraposition analysis since the verb of the cleft clause can here be viewed as agreeing with the cleft pronoun instead of the clefted constituent.

In Middle English also the verb of the cleft clause agreed with the cleft pronoun, while the copula agreed with the clefted constituent, as shown in (65a). A similar pattern can be seen in

Modern German in (65b) and Dutch in (65c):

20

(65) a.

It am I that loveth so hote Emilye the brighte.

loves-3.sg

b.

Ich bin es der immer die Rechnungen bezahlt

I am it who always the bills pays-3.sg.

‘I am it who always pays the bills’ c.

Ik benhet die haar al die bloemen stuurt

I am it who here all those flowers sends-3sg.

‘I am it who sends all those flowers’

Extensive dialectal variation in the morphological case of the clefted constituent is dramatically displayed in the modern Scandinavian languages. Compare the possibilities available in Danish

(66a), Norwegian (66b), and Swedish (66c).

21

(66) a.

Det var *jeg/mig som købte denjakke

‘It was me that bought this coat.’

20

The Middle English example is from Chaucer, quoted in Jespersen. The German and Dutch examples are from Smits

1989.

21

The Scandinavian examples are from Smits 1989. Note that the Scandinavian languages don’t exhibit subject-verb agreement.

97 b.

Det var jeg/meg som kjøpte denne frakken

It was I/me that

It was I/me that bought this coat..’ c.

Det var jag/*megsom köpte denna rock

It was I/*me that bought this coat

‘It was I that bought this coat.’

Positive evidence in favor of the cleft clause as an extraposed relative clause can be found in the Scandinavian languages. Smits 1989 observes that the cleft clause must follow the finite verb in subordinate clefts in German and Dutch. Since the verb ordinarily appears in final position in subordinate clauses, the postverbal cleft clause must occupy an extraposition position. The example

(67a) is an example from German, to which I would assign the structure in (67b):

(67) a.

Jutta sagtdaß esdieserWagen war den sie saysthat it this car was which she kaufen wollte to.buywanted

‘J. says that it was this car that she wanted to buy.’ b.

CP daß

IP es

I’

I VP

VP CP

NP war den sie t

i

kaufen wollte dieser Wagen

Note further that the syntactic similarity between the initial pseudocleft clause and the final cleft clause is even more pronounced in Icelandic than it is in English:

(68) a.

¶a∂ var Olafur/Olaf sem Maria sá

(Smits 1989:347) that was Olaf-

NOM

/O

LAF

-

AC that Maria saw

‘It was Olaf that Maria saw.’ b.

¶a∂ sem María sá var Olafur/*Olaf

(Smits 1989: 348) that that Maria saw was Olaf-

NOM

/O

LAF

-

ACC

98

‘The one that Maria saw was Olaf.’

I have argued in this section that the arguments which have been advanced in the literature against the extraposition analysis of clefts are not compelling. Having argued in §4.1 that the cleft pronoun and cleft clause function as a unit in pragmatic interpretation, I conclude, therefore, that the relation holding between the cleft pronoun and the cleft clause should be encoded in the syntactic analysis of clefts—in other words, that the expletive approach to cleft structure should be abandoned in favor of the extraposition approach.

4.4

The relation between the clefted constituent and cleft clause

In this section, I present arguments that the clefted consituent and the cleft clause function syntactically as a unit—more specifically, that the cleft clause is a subconstituent of the VP.

Delahunty 1982 presents five constituency arguments in support of the hypothesis that the cleft clause is a subconstituent of the verb phrase. The clefted constituent and cleft clause act as a unit with respect to VP-deletion (69a), right-node-raising (69b), parenthetical formation (69c), VPconjunction (69d), and VP-preposing (69e):

(69) a.

I said that it should have been should have been.

b.

It could have been—and it should have been— contract .

c.

It must have been, in my opinion, d.

It must have been Fred that kissed Mary but e.

?I said that it was Bill that argued the case and

Delahunty observes that his VP-preposing example is of dubious acceptability. Since more acceptable examples can be constructed, however, as shown by the equally acceptable status of (70a) and (70b), I conclude that VP-preposing is indeed permissible in clefts, and thus that the cleft clause is indeed a constituent of the verb phrase:

(70) a.

I said it would be a conservative who’d win, and it certainly was.

b.

I said that I would finish by September, and

It is worth pointing out that the first four constructions Delahunty evokes do not entirely exclude extraposed relative clauses and right-dislocated phrases with subject antecedents from appearing in the position filled by the cleft clause in clefts:

(71)

Extraposed relatives:

99 a.

A man with blond hair, and a woman did with brown hair.

[Culicover and Rochemont 1990] b.

Nobody would—and nobody could— thing about espresso .

[c.f. Reinhart 1980] c.

Nobody would ever, in my opinion, who knew anything about espresso .

d.

Nobody could drink instant coffee and who knew anything about

(72) e.

*I said a candidate would win who had charisma, and candidate did.

Right-dislocated phrases:

a.

I said that he should have should have.

b.

I could have — and I should have — the phone call .

c.

It might, I suppose, .

d.

They’ e.

?*We predicted that they’d win, and

The ungrammaticality of (71e) and the extreme awkwardness of (72e), as compared to (70a) leads me to conclude, however, that the relation which holds between the cleft pronoun and the cleft clause is not identical to the relation which holds between an extraposed relative clause and its NP head, or between a right-dislocated phrase and its antecedent pronoun, thought all three relations have a great deal in common.

22

4.5

The syntactic structure of clefts

In the preceding chapter I argued that the copula plays the same semantic role in clefts that it plays in copular sentences in general since clefts, like other copular sentences, have predicational as well as specificational interpretations. In §4.1, I argued that the cleft pronoun is pragmatically contentful and varies in form depending on the cognitive status of the information expressed in the cleft clause;

22

Note also that the clefted constituent alone can be preposed:

(i) I said that it was Bill who believed that the Earth is flat and

[Delahunty 1982:101]

(ii) it was that Nigel stood on the edge of the lock, and began a one-sided conversation with the lifeless figure in the canoe.

[Footsteps at the Lock, p. 117]

100 and secondly, that the cleft pronoun is morphologically invariant in specificational clefts, while in predicational clefts, it agrees with the clefted constitutent at least in number and sometimes also in gender and person. In §4.2, I argued that the cleft clause is structurally a relative clause, which functions as a unit with the clefted constituent for purposes of VP-preposing, but functions as a unit with the cleft pronoun for purposes of semantic and pragmatic interpretation.

I propose, therefore, that the syntactic structure of clefts should be viewed along the lines of the structure shown in (71):

(71)

IP

NP

j

this

I’

I

I’ VP

V

k

is t

k

VP

NP a country

OP

i

C

CP

j

C’

IP that

I would claim that the structure in (71) is compatible with either a specificational or a predicational interpretation. If the postcopular NP is interpreted referentially, the sentence is interpreted specificationally, and the subject pronoun is morphologically invariant. If the postcopular NP is interpreted predicatively, the sentence is interpreted predicationally, and the subject pronoun agrees with the clefted constituent at least in number and sometimes in gender and person.

I assume that the internal structure proposed for the cleft clause would be uncontroversially accepted as the internal structure of a restrictive relative clause. As discussed in §4.2, my proposal differs from recent generative proposals for the structure of clefts in viewing the presence of an overt wh-operator in a cleft clause to be a full-fledged alternative to an overt that-complementizer, just as it is in ordinary relative clauses. Since I assume that predication is mediated by the copula, I don’t need to adopt any special mechanism to directly associate the clefted constituent with the missing argument of the cleft clause, though some sort of mechanism may be needed to accomodate agreement and case-marking relations that are found to hold between the clefted constituent and the

101 verb of the cleft clause. I adopt without modification current assumptions concerning V-to-I movement of the copula (c.f. Pollock 1989).

23

The most original, and least developed, aspect of the proposed structure is the coindexation of the cleft pronoun and cleft clause. It is not necessary, of course, to assume that the cleft clause is generated inside the subject NP and then moved to the end of the sentence. The cleft clause can instead be generated directly in VP-adjunct position, as long as it is coindexed with the cleft pronoun, or otherwise associated with it, at the level of representation which serves as input to full pragmatic interpretation. There are various theoretical avenues to explore in making this relation explicit. A purely semantic account might be adopted along the lines of Wittenburg’s (1987) proposal for extraposition-from-NP in discourse representation theory, or a syntactic account might be adopted along the lines of Culicover and Rochemont’s (1990) analysis of extraposition-from-NP involving their ‘Complement Principle.’ It might also be possible to extend to clefts, Chomsky’s

(1988) proposal that a ‘Principle of Full Interpretation’ forces the post-copular NP in an existential

there sentence, and the sentence sentential subject in an extraposition sentence, to raise and adjoin to the expletive subject pronoun at LF. Finally, it would be intriguing to investigate the implications of

Abney’s (1987) ‘DP hypothesis’ for encoding the functional parallels between determiners and expletives which were discussed in §4.1.3. above. Exploration of these alternatives, however, must be left for future research.

23

Note that the copula is not preposed with the clefted constitutent and cleft clause in VP-preposing examples such as

(68a). Though I have presented examples only of clefted NP’s and AP’s, I assume that clefted PP’s and adverbs and CP’s can also be accomodated. It would also be possible to assume that the cleft pronoun, the cleft clause, the cleft complementizer, the operator and the gap all receive the same index: the pronoun and the clause by the expletiveassociate relation, the clausal (CP) node with its head C by feature percolation, C with the operator by specifier-head agreement, and the operator with the gap by the operator-variable relation.

Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement