The Discourse Functions

The Discourse Functions

of Reverse Wh-Clefts

Judy Delin

Centre for Research in Communication and Language

and

Department of English Studies

University of Stirling

U.K.

Overview

• RWHC: Some Givens

• Suggested Discourse Functions

• Additional features of RWHC

• Distinguishing TC/SF and CC/IP

RWHC

• Interpretation of RWHC in a model

of discourse

The Study

• 302 spoken examples from LLC

• Headless only (not ‘the one that..’

‘the thing that…’)

• Ball’s (1991) DEM-be wh

• Use Collins’ (1991) study of

this data as a point of comparison

Some examples

I mean my idea would be to teach language and hire somebody to do the literature sort of thing you see

whereas if Roy Peel went I suppose he’d do the literature and hire somebody to do the language

that’s what it would amount to isn’t it

so we ended up saying end of September twentieth of September and that’s when it’s going to be

England’s trouble over the miners took precedence over Ireland’s problems with Sunningdale and that’s why the election was called at that time

and you can get your spread of marks just from that alone

that’s where it’s coming from

so I said well I’m terribly sorry but you know

that’s how it’s going to be

TC/SF and CC/IP RWHC

• Informative Presupposition

or CC RWHC: tend to be non-what; prefer discourse-new presuppositions

(e.g. 13)

• ‘Stressed Focus’ or TC tend

to be what; prefer discourse-new

presuppositions (e.g. 7)

These are processed differently, and perform different functions

Suggested Functions of RWHC

Weinert and Miller (1996): rwhc is

a thematizing construction

That’s what worries me

That worries me

Wrapping Up, Ending and Summarizing (Hedberg 1988, Collins 1991, Weinert and Miller 1996)

IP and SF not always distinguished in these accounts
Distinctive Features of RWHC

Additionally (Oberlander and Delin 1996):

• Reference

• Case role allocation

• Information structure

• Discourse role

Reference

• Both ITC and RWHC can presuppose information that is neither discourse-old nor hearer-old: i.e many RWHCs are ‘comment-clause’ or IP

• ITC clefted constituent refers more often to single element; RWHC to propositions or sets of propositions

• Resolutionof reference of deictic

element is different: ITC has an

intermediary referent

so we ended up saying end of September twentieth of September and that’s when it’s going to be

so we ended up saying end of September twentieth of September and it’s then that it’s going to be

Case Role Allocation

• Range of wh-words means that RWHC anaphorically specifies material and allocates it a new case role in relation to the verb (examples

in 17)

Discourse Role

• Prevalence of discourse deixis -- not shared by ITC – allows previous

information to be re-appropriated and commented upon affectively

or metalinguistically within a new case frame.

• Deictic thematization of previous material in order to comment on it

directly is a much more direct device than the postponement of the

relevant equation necessitated in ITC.

• This transaction management information contributes background infomation, rather than moving a discourse forward in narrative terms.

Discourse Interpretation: SF/TC RWHCs

oh so THAT’S what you were talking about when I came in

OE: You were talking about something when I came in

CS: That is that thing

Presupposed OE is matched in memory and integrated, reference time updated appropriately

CS finds its place in temporal order of discourse via OE (talking)

CS overlaps the current reference time pinpointed by OE

‘That’ singles out a preceding referent

Cleft performs a ‘summative’ or ‘closing’ function

Cf. Delin and Oberlander (1995)
Discourse Interpretation: IP/CC RWHCs

this is what really differentiates England Scotland and Wales from the northeast of England

OE: Something differentiates England Scotland and Wales from the northeast of England

CS: this is that thing

Presupposed OE is looked for in memory and finds no match

CS is then integrated into the temporal order of the discourse; OE is only indirectly related (accommodated) to CS by inference, at some time during CS being true

Cleft as a whole does not update reference time, as states don’t do this: whole cleft content cannot therefore be directly related to the time-line of narrative or meta-narrative

Does not introduce metalinguistic or summative information; introduces a background discourse segment, but one that is immediately ‘sidelined’ or closed.

Summary

• ITC and RWHC have a lot in common in terms of information structure

• RWHC is different in terms of reference and case role allocation from ITC

• Distinctive anaphoric flavour of TC/SF RWHC, given Delin and Oberlander’s

model of cleft processing, gives ‘summative’ function: tends to be WHAT RWHCs

• CC/IP RWHC, which tends to be non-WHAT, introduces short background discourse segments; their content tends not to be taken up in subsequent discourse, which typically reverts to the pre-RWHC content

Information Structure: The Clefted Constituent

• Overwhelming preference for this or that

• Other NPs: it or which

• Anaphoric reference may be

extended – discourse deixis

• RWHC much more likely to specify abstract or lengthy discourse antecedents than ITC;

ITC much less likely to have

this/that as clefted constituent;

RWHC much less likely to have

non-neuter pronouns

• Any New ones? Only 6% of our corpus had prosodic nuclei on the clefted constituent. None are New, and not all are even contrastive:

Programming computers, THAT’S what I do

• Most, even NPs, are ACTIVATED or IN FOCUS, and a very few INFERRABLE

Cataphora and the Clefted Constituent

• Collins (1991) highlights referential

vagueness of RWHC clefted constituent

• Many are modified post-hoc: refer

both ways

B: you need sideboards

that’s what your bedroom lacks is a is a

dressing table on which there is a mirror

and that’s essentially what a sideboard is

C: how can you determine the price of things

without knowing how many copies they’re going to sell

B: that’s what I don’t know about publishing

selling books as a commercial racket

you can estimate how many words it’s going to be and you can I mean

there is a whole field called market research and there are people who go about doing it