638

Geluykens R (1994). The of discourse anaphora Levinson S C (1995). ‘Three levels of meaning.’ In Palmer F in English: evidence from conversational repair. Berlin: (ed.) Grammar and meaning. Essays in honour of Sir Mouton de Gruyter. John Lyons. Cambridge: Cambridge University Press. Himmelmann N P (1996). ‘Demonstratives in narrative 90–115. discourse: a taxonomy of universal uses.’ In Fox B A Lyons J (1977). ‘, space and time.’ In (ed.) Studies in anaphora. Amsterdam/Philadelphia: (Ch. 15), vol. 2. Cambridge: Cambridge University Press. John Benjamins. 206–254. 636–724. Huang Y (2000). Anaphora. A cross-linguistic study. van Hoek K (1997). Anaphora and conceptual structure. Oxford: Oxford University Press. Chicago: University of Chicago Press. Kleiber G (1994). Anaphores et pronoms. Louvain-la- Yule G (1981). ‘New, current and displaced entity .’ Neuve: Duculot. Lingua 55, 41–52. Langacker R W (1996). ‘Conceptual grouping and pro- Ziv Y (1996). ‘Pronominal reference to inferred antece- nominal anaphora.’ In Fox B A (ed.) Studies in anaphora. dents.’ Belgian Journal of 10, 55–67. Amsterdam/Philadelphia: John Benjamins. 333–378.

Discourse Domain P A M Seuren, Max Planck Institute for Psycholinguistics, Nijmegen, The Netherlands (2) d–n [a | King(a) // Rich(n)]

ß 2006 Elsevier Ltd. All rights reserved. or ‘n is rich,’ where n stands for ‘the a such that a is a king.’ In (1), a is an existential ; in A discourse domain D is a cognitive space for the (2), a is a definite determiner – a function from the middle-term storage of the information conveyed by [[King]] to an (the refer- subsequent utterances. The notion of discourse do- ence value). The change is due to address closure, main is part of the theory of discourse semantics, represented as //, which takes place when an open which holds that the interpretation of utterances is address is called on by a subsequent definite term. codetermined by the information stored in the D at Address closure enables the semantic distinction be- hand. For an utterance u to be interpretable, u must tween the open address in (3a) and the closed address be anchored in a given D. The anchoring of u in (3b): requires, at least, that all referring expressions in u, (3a) John has few clients who are dissatisfied. (open including anaphoric , link up with an ad- address) dress in D that represents the object or objects re- (3b) John has few clients. And they are dissatisfied. ferred to. The information conveyed by the main (closed address) predicate in the sentence S underlying u is then distributed over the relevant addresses. This process Some addresses are subdomains, representing what is called the incrementation of S or i(S). When S has been specified as someone’s belief (hope, knowl- presupposes P, P must be incremented before S. edge, etc.), as being possible or necessary, or the When D does not yet contain i(P), i(P) is supplied alternatives of an or-disjunction. There is an intricate post hoc by accommodation (see ), system of interaction between the main Do and sub- unless blocked. Ds are to a condition of con- domains Ds. Information within Do can be called on sistency and a condition of cognitive support, ensur- in any Ds, unless blocked by contrary information in ing the compatibility of accommodated increments Ds (downward percolation). Likewise, unless blocked with available world knowledge, unless D specifies in Do, accommodated in a Ds are also otherwise. accommodated in Do (projection). A D contains object addresses, domain addresses, A D may also contain instructions constraining its and instructions. Object addresses represent real or further development. , as in not-S, is an fictitious objects. A (singular or plural) object address instruction banning i(S) from D. This explains the is created, in principle, by an existentially quantified incoherence of example (4), where the it sentence, say, There is a king, resulting in a labeled calls on an address that has just been banned: address (where n is an arbitrary natural number): (4) !John has no car. It is in the garage. (1) d–n [a | King(a)] The study of discourse domains is still in its infancy. or ‘there is an a such that a is a king’ (disregarding Yet it already provides explanations for phenomena tense). The sentence He is rich results in: that have so far remained obscure. Discourse Markers 639

See also: Anaphora, , , ; Bibliography Definite and Indefinite; versus Connotation; Discourse Semantics; Donkey Sentences; Presupposi- Seuren P A M (1985). Discourse semantics. Oxford: tion; Projection Problem. Blackwell.

Discourse Markers J R Trillo, Universidad Auto´ noma de Madrid, Madrid, contemporary French; Bazzanella (1990) did the Spain same for contemporary Italian, and Fraser-Gupta ß 2006 Elsevier Ltd. All rights reserved. (1992) for Singapore colloquial English. Other stud- ies delved into the use of a specific discourse marker in a language, giving a full description of its different Discourse markers can be defined as those elements, functions; cf., for example, Tognini-Bonelli (1993) on such as you know, I mean, well, oh, m, you see, look, ‘actually’ in British English. listen, that have a distinct prosodic entity, tend not to Since then, several authors have tried to name, have a specific semantic meaning, and contribute to classify, and describe the nature and function of scaffold the pragmatic coherence of interaction (see these seemingly ‘elusive’ elements. In my opinion, Scaffolding in Classroom Discourse). Broadly there are three main approaches to the study of dis- speaking, discourse markers have often been de- course markers: the conversational, the grammatico- scribed by their absence of traditional linguistic prop- syntactic, and the discourse-cognitive. erties, since neither their use nor their meaning seem The conversational approach concentrates on the to resemble any of the traditional linguistic cate- role of discourse markers in the structure of conver- gories. In fact, these elements have been typically sation. For example, Schegloff (1984), who called discarded from rigorous linguistic studies, and in them ‘‘continuers,’’ stated that they are used by different languages have been assigned the general the speaker to show the listener that his/her speech vague name of ‘particelle’ (Italian), ‘fillers’ (English), forms a coherent whole in progress. In the same vein, or ‘muletillas’ (Spanish) – both the English and the Schiffrin (1987: 31) provides the following definition: Spanish terms evidence these markers’ function of ‘‘markers are sequentially dependent elements that ‘filling’ or ‘supporting’ discourse. In some sociolin- bracket units of talk.’’ Both authors (along with guistic domains, discourse markers were margina- others) agreed on the essentially spoken nature of lized because they were traditionally attributed to these elements; from their studies, one is led to con- incompetent and incoherent speakers of a language clude that discourse markers are better analyzed from (Watts, 1989). an interactional than from an ideational perspective Schenkein (1972) and Jefferson (1978) were among (following Halliday’s (1994) macrofunctions of lan- the first that became convinced of the need to study guage) (see Systemic Theory). In fact, Schiffrin (1985: these vague elements that seem to distort syntax and 281) pointed out that: ‘‘discourse markers ... help have a multiplicity of undefined meanings. As an speakers express interactional alignments toward instance, consider how Dik (1989: 45) characterized each other and enact conversational moves’’; else- these elements as ‘extra-clausal constituents’; their where she stated (Schiffrin, 1987) that the presence role in natural language was described as follows: of discourse markers in a conversation helps the me- ‘‘any natural language text can be exhaustively chanics of turn-taking, the organization of speech divided into clauses and extra-clausal constituents acts, the structuring of discourse ideas, the interactive ... which are neither clauses nor part of clauses.’’ In structure of participants, and the presentation of other words, discourse markers, for this author, be- information. long to the realm of spoken language and need to be The second model corresponds to the grammatico- studied only with reference to the parameters of syntactic approach to discourse markers. Among spoken language. others, Knott and Dale (1994: 45) described discourse In the 1980s and early 1990s, some authors studied markers as a ‘‘reasonably homogeneous group’’ that the general presence of discourse markers in several tends to be formed by ‘‘simple linguistic expressions languages, with no special reference to any grammat- [their italics] ...that have become simplified because ical or syntactic model. Thus, for instance, Vicher they correspond to constructs that are in continual and Sankoff (1989) described discourse markers in use when we process text.’’ In their description of