LAGERWERF, Luuk (1998). Causal Connectives Have Presuppositions. Effects on
Coherence and Discourse Structure. The Hague: Holland Academic Graphics. 252
reviewed by Elisabeth Le, University of Alberta.
In his doctoral dissertation, Luuk Lagerwerf explores the lexical meaning of
causal connectives and the effect it has on discourse coherence and
structure. To do so, he answers the four following research questions:
1. What interpretations of causal or contrastive relations should be
2. How is lexical meaning of causal connectives represented?
3. How is lexical knowledge represented when a causal connective is used to
indicate causal coherence?
4. How do causal relations affect discourse structure?
First question: What interpretations of causal or contrastive relations
should be distinguished?
Causal relations can have a semantic or pragmatic interpretation. It is
semantic (locutionary meaning) when the condition contains a necessary part
for the result. When it is pragmatic (illocutionary meaning), following
Sweetser's distinction (1990), it can be epistemic (when it represents the
speaker's conclusion) as in (1), or it can be a speech act interpretation
(when the uttering of one clause is justified by the other clause) as in
(1) John loved her, because he came back.
(2) I ask you what you are doing tonight, because there's a good movie on.
Contrastive relations appear in three different interpretation schemes:
denial of expectation, semantic opposition and concession. These
interpretations can be characterized using Sanders, Spooren & Noordman's
cognitive primitives (1992) for the classification of coherence relations.
Thus, denial of expectation can be semantic, causal and negative:
(3) Connors didn't use Kevlar sails although he expected little wind;
or, pragmatic (epistemic), causal and negative:
(4) Theo was not exhausted, although he was gasping for breath;
or, pragmatic (speech act), causal and negative:
(5) Mary loves you very much, although you already know that.
Semantic opposition is semantic, additive and negative:
(6) Greta was single, but Prince was married.
while concession is pragmatic, additive and negative:
(7) A: Shall we take this room?
B: It has a beautiful view, but it is very expensive.
Second question: How is lexical meaning of causal connectives represented?
As for the second question, Lagerwerf claims that 'causal and causal
contrastive connectives have a presupposition in the form of an implication
that expresses causality' (111). This presupposition can be made in the case
of semantically causal and pragmatically causal interpretations. An example
for a causal contrastive connective in a pragmatic (epistemic)
interpretation is given in (8). In (8a) and (8b), 'although he sauntered to
the university' presupposes (8c). As (8c) is being contradicted in (8b),
(8b) is unacceptable.
(8) a. Sauntering makes you restful. Although he sauntered to the
university, Theo was exhausted.
b. ?Sauntering wears you out. Although he sauntered to the university,
Theo was exhausted.
c. Normally, if you saunter, you are not exhausted.
Third question: How is lexical knowledge represented when a causal
connective is used to indicate causal coherence?
Lascarides and Asher's (1991) system of Discourse Inference and Commonsense
Entailment (DICE) and Pustejovsky's description of a structured lexicon
(1991) are used as a framework to show the difference between explicitly
marked discourse relations and unmarked relations. Unmarked causal relations
have to be initiated by the lexical items to indicate causal coherence in
the text. However, marked causal relations, since they are expressed in an
implication presupposed by connectives (see question 2), do not need to be
constructed by lexical knowledge or world knowledge; but they are supported
by it. Thus, causal relations, whether marked or not, are established with
respect to lexical knowledge or world knowledge.
Fourth question: How do causal relations affect discourse structure?
In the Linguistics Discourse Model (LDM) (Polanyi, 1988; Prust, 1992), used
as a framework to represent the discourse structure, clauses are added to
the right hand side of the structure (right frontier), so that the order of
their occurrence in the discourse is preserved. Antecedents of unmarked or
marked propositional anaphors should occur on the right frontier. However,
antecedents of marked anaphors may not occur on the right frontier, if a
causal coherence relation connects them with it. Construction rules of LDM
have been changed and extended to reflect this.
It has been shown how causal relations are defined by their semantic and
epistemic interpretation, and therefore, how essential the interpretation of
the speaker's conclusion is for the study of coherence. As this speaker's
conclusion also indicates the way the speaker uses causal relations in the
argumentation, its study would allow to look at the use of argumentation to
achieve a communicative goal.
This dissertation is in some places rather technical, but Luuk Lagerwerf has
made his argumentation in general easier to follow by systematically
announcing what he is going to discuss in the next section, and then, by
summarizing it. Thus, the reader can easily follow the thread of the
argumentation. However, the merit of this dissertation is in the way it
relates semantics, pragmatics, discourse representation, argumentation,
computational linguistics, and the linguistic analysis of conjunction. By
showing how the analysis of causal connectives cannot be dissociated from
the context in which they occur, it opens the door to new interdisciplinary
studies in argumentation. This book can and ought to be read by anybody
interested in the argumentative discourse, even if they are not familiar
with models in computational linguistics.
Lascarides, A. and Asher, N. (1991). Discourse relations and defeasible
knowledge. In Proceedings of the 29th Meeting of the Association of
Computational Linguistics. ACL91.
Polanyi, L. (1988). A formal model of the structure of discourse. Journal of
Pragmatics, 12: 601-638.
Prust, H. (1992). On Discourse Structuring, VP Anaphora and Gapping. PhD
thesis, University of Amsterdam, Amsterdam.
Pustejovsky, J. (1991). Towards a generative lexicon. Computational
Sanders, T.J.M., Spooren, W.P.M.S. & Noordman, L.G.M. (1992). Toward a
taxonomy of coherence relations. Discourse Processes, 15(1): 1-35.
Sweetser, E. (1990). From Etymology to Pragmatics. Metaphorical and Cultural
Aspects of Semantic Structure. Cambridge University Press, Cambridge.
Elisabeth Le, Assistant Professor at the University of Alberta
Main research interests: coherence, argumentation, academic discourse