Nature of clefting

Let S be a simple sentence containing a major constituent F of category C which bears the sentence focus. Then clefting S is a complex operation consisting of the following steps:

  1. F is extracted from S, leaving an open place there.
  2. A new main clause S1 is created, with F as its nominal predicate.
  3. The rest of S becomes an open subordinate clause S2 which resumes F.

The communicative function of clefting is major emphasis and typically contrastive focus on F. It differs from mere contrastive stress on F within S by its explicit syntactic coding. Sentence-clefting is, in fact, the most explicit syntactic strategy of contrastive focus.

There are two varieties of clefting, the cleft-sentence and the pseudo-cleft-sentence, depending on the syntactic category and function of S2.

Cleft-sentence

General structure

A cleft-sentence is a complex sentence of the structure of S3 in the following diagram.1 The phoric relation between F and the empty place that it leaves in S2 is symbolized by the index i.

Cleft sentence
[[[ F ]C.i]S1[... [Pron/∅]C.i ...]S2]S3
expletive/zero subjectnon-verbal predicateempty place
focus expression[extrafocal clause]
[non-verbal clause][open clause]
[main clause][dependent clause]

illustrates a cleft sentence:

.It is yesterday [ that I met her ].

The subject of the main clause S1 is semantically empty. It depends on the syntax of the language whether it requires an expletive subject there, just as it depends on the language whether it requires a copula with non-verbal predicates. contains both. If both are missing, the main clause of a cleft sentence reduces to the focused constituent. This is the case, e.g., in Yucatec Maya, as in .

.ma'teechint'an-ik=i'
YucNEGthou[ SBJ.1.SGcall-INCMPL ]=NEGF
It is not you I'm calling.(MUUCH_139)

The extrafocal clause commonly takes the form of a complement clause subordinated by the universal subordinator ('that') (again, lacking from Yucatec). It is open in the sense that it lacks one of its arguments or satellites, viz. F. The order of S1 and S2 is not crucial, although it is usually this order.2

If all the segmental material that codes the clefting is zero, as it is in , the cleft-sentence may differ from a simple sentence only by the positioning of F. It is necessarily at the margin of S2. It is by far the most common order for S2 to follow S1, but it is not excluded for S2 to precede S1.

While the extra-focal clause of a cleft-sentence generally has the same internal structure as an open complement clause, its syntactic function in the main clause is less clear. It behaves in some respects as the subject of S1, but differs from normal subjects in being obligatorily extraposed. Nor does it have any other syntactic function in S1. Much less can it be analyzed as a relative clause whose head would be F. For instance, the F of is not a possible head of any relative construction. If the focus constituent is a nominal expression, as in , it is commonly a (semantically) definite NP. This could only take a non-restrictive relative clause. However, there is no intonation break and no pause at the clause boundary of S1 and S2.

The focus constituent of a cleft-sentence can have more different syntactic functions in S2 than the relativized position in a relative clause. Among other things, the focus may be an adverbial (as in ) or a predicate nominal. It is rather difficult to orient a clause towards an adjunct and impossible to orient it towards the predicate. As a consequence, extrafocal clauses, although open, are not oriented in principle.

Cleft-sentences are part of a semasiological grammar of complex sentence constructions. From an onomasiological point of view, they belong in the functional domain of information structure rather than the domain of nexion.

Pronominal interrogatives

In a pronominal interrogative sentence, the interrogative pronoun has focus function. Such a sentence is often coded as a cleft-sentence, as in :

.a.Qui est-ce [ qui me parle ]?
FrenchWho is speaking to me?
b.Qu'est-ce [ que tu penses ]?
What do you think?

The interrogative pronouns qui ‘who’ and que ‘what’ represent the F of the diagram, but occupy initial position in S1. In speech, they are univerbated with the rest of S1. In this way, the inherited interrogative pronouns (which are phonologically very weak indeed) are currently being renewed as [kjɛski] 'who' and [kɛskə] 'what'. In many languages, the interrogative pronouns share a morphemic or submorphemic element of a similar origin.

Pseudo-cleft sentence

The general structure of the pseudo-cleft sentence is shown in the following diagram:

Pseudo-cleft sentence
[[... [Pron/∅]C.i ...]S2 (Cop)[ F ]C.i]S1
[subject] non-verbal predicate
empty place focus expression
[ [free relative clause] non-verbal clause]
[ [dependent clause] main clause]

is an example.

.Teech=e'chéen[ ba'xk=ameet-ik ]=e'chéenlek'ool=o'
Yucyou=TOPjustwhatIPFV=SBJ.2do-INCMPL=TOPjustDEMcorn.soup=D2
As for you, all you do is to make the sauce.
(Lit.: ... just what you make is just the sauce)
(SANTO_058)

has two sentence-initial constituents which are left-dislocated topics, marked by the enclitic =e'. The second of these is a free relative clause. The focal constituent is represented in the relative clause by a relative pronoun of interrogative origin, which has direct object function.

The definitional properties of the pseudo-cleft sentence are:

  1. As in any cleft-construction,
  2. The non-focal portion is a free relative clause.
  3. The non-focal portion is the subject of the sentence.
  4. The empty position in the relative clause is coreferential with the predicate of the main clause. Consequently, the main predication states identity of the subject with the predicate nominal.

The default order of the two clauses is the one indicated in the diagram. However, initial position of the main clause (as in the cleft-sentence diagram above) is not excluded. Moreover, the relative clause is often a pronominal one, and the pronoun is often based on the interrogative pronoun. Again, this is not definitional.

The cleft-sentence and the pseudo-cleft sentence have much in common:

The definitional properties that distinguish the cleft-sentence and the pseudo-cleft sentence are the following:

Many languages have only one of these two constructions. Latin is among them: it only has pseudo-cleft sentences, and these are rare in the corpus. (With its free main constituent order, it has little use for clefting.) It is no coincidence that pronominal interrogative sentences, as in , figure prominently among the extant occurrences:

.Quidestigiturquodfieripossit?
Latinwhatbe(PRS):3.SGconsequently[ REL:NOM.SG.Ndo.PASS:INF.PRScan:SUBJ.PRS:3.SG ]
What then can be done?(Cic. Verr. 1, 1, 32)

This is, at the same time, an example where the focus is the initial constituent of the main clause and the latter introduces the entire construction. Both features are explicable by interrogative syntax.

In languages which have both cleft- and pseudo-cleft sentenes – with English among them –, there is the possibility of a functional contrast:

Where the relative clause of the pseudo-cleft sentence works with an interrogative pronoun, the difference between a pseudo-cleft sentence and a dependent interrogative clause may be minimal. is a celebrated example (in the 1970s) of this structural similarity:

.a.What lay on the table was the tissue.
b.What lay on the table was the issue.

.a is a pseudo-cleft sentence; #b is an interrogative clause plus matrix clause. The diagnostic properties for this analysis are provided by the prosody:

The information structure is therefore:


1 The term 'non-verbal clause' is intended to cover clauses both with a nominal and with a copula predicate.

2 This means that the distinction between clefting and pseudo-clefting is not made on the basis of clause order. Instead, the criterion is the nature of the extrafocal clause: in the case of pseudo-clefting, it is a free relative clause.