Whither Discourse n.nd Speech Acts? 
Barbax~ J. Gross 
Harvard University 
Cambridge Massachusetts 02138 
In addition to the questions initially posed on discourse theory and speech acts 
(and listed in the proceedings), the panelists were asked to consider how their beliefs 
about discourse and speech acts had changed in the last five or so years, and in partic- 
ular to consider the following two questions: 
Of what you believed about discourse five years ago, what now 
strikes you as most wrong? 
What are you less confused about now than five years ago? 
The panelists papers and a glance through the proceedings of the previous two 
TINLAP meetings suggests that we have in fact learned a good deal and that research 
in discourse and speech act theory has become much more focused in the last five or 
so years. Several theories of discourse have been proposed to account for the struc- 
ture and processing of extended sequences of utterances. In a recent paper 
\[Grosz&Sidner86\], Candy Sidner and I present one such theory and discuss a range of 
alternative ones. Other research has provided partial accounts for the interpretation 
and generation in discourse of individual expressions, clauses, and interclausal connec- 
tions (e.g., \[McKeown85\], \[Appelt85\], \[Groszeta183\], \[Mann&Thompson83\] 
\[Sidner83\], \[Webber83\],\[Hobbs79\]). These theories differ in the predictions they 
• make about discourse structure and coherence, and about the processing of various 
types of linguistic expressions (e.g., cue phrases, definite descriptions, and pronouns). 
Initial psychological investigations into various aspects of these theories are being car- 
ried out. 
One of the most striking changes is that research on discourse and speech act 
theories now typically addresses more specific questions, a fact illustrated by the papers 
for this panel. In addressing the role of research on speech for discourse, Hirschberg 
describes results that reveal the crucial role of information provided in the speech sig- 
nal; she argues that simpler solutions to many discourse problems are available only if 
one does not ignore speech. Perrault discusses various key problems that arise in tak- 
ing seriously the intuitions underlying "speech act theory," in looking in detail at what 
it means to say that utterances are actions that most directly affect the mental state of 
the participants in the discourse; he indicates the variety of approaches being taken to 
specific problems within this framework and their connection to work more generally 
in natural-language semantics. 
Another clear difference from the previous two TINLAP meetings is the more 
general acknowledgement of the pervasiveness of the role of intentions and plans in 
discourse. These affect not only the interpretation of individual utterances, but also of 
extended sequences of utterances and of individual phrases within an utterance. This 
recognition has led to a close examination of theories of planning and action; it has 
become clear that work in AI on planning is insufficient to support discourse 
86 
processing. Wilensky argues for much more complex notions of plans and goals to 
support discourse processing; he describes a range of discourse problems requiring 
these and indicates research directed toward developing them. Other recent research 
(e.g., \[Pollack86\], \[iautz&Allen86\]) has examined the adequacy of current models of 
planning for plan recognition and developed alternative models better able to support 
discourse processing. 
In brief, we seem to have gotten much closer to understanding what the prob- 
lems, are; there is still much to do -- and much disagreement about how -- to solve 
these problems. 
References: 
\[Appelt85\] Appelt, D. 1985 Planning English Referring Expressions. Artificial Intelli- 
gence 26: 1-33. 
\[Groszeta183\] Grosz, B.J., Joshi, A.K., and Weinstein, S. 1983 Providing a Unified 
Account of Definite Noun Phrases in Discourse. Proc. 21st Annual Meeting of 
the Association for Computational Linguistics. Cambridge MAr 44-50 
\[Crosz&Sidner86\] Grosz, B.J. and Sidner, C.L. 1986 Attention, Intentions, and the 
Structure of Discourse. Computational Linguistics 12(3): 175-204. 
\[Hobbs79\] Hobbs, J. 1979 Coherence and Coreference. Cognitive Science 3(1): 67-82. 
\[Kautz&Allen86\] Kautz, H. and Allen, J. 1985 Generalized Plan Recognition. Proc. 
Fifth National Conference on Artificial Intelligence. Philadelphia PAr 32-37 
\[Mann&Thompson83\] Mann, W. and Thompson, S. 1983. Relational Propositions in 
Discourse. Tech. Report RR-83-115. Information Sciences Institute, Marina del 
Rey, CA 
\[McKeown85\] McKeown, K. 1985 Discourse Strategies for Generating Natural- 
langauge Text. Artificial Intelligence 27:1-42 
\[Pollack86\] Pollack, M. A Model of Plan Inference that Distinguishes Between the 
Beliefs of Actors and Observers. Pr0c. 24th Annual Meeting of the Association 
for Computational Linguistics. New York, N.Y.: 207-214 
\[Sidner83\] Sidner, C. Focusing in the Comprehension of Definite Anaphora. in M. 
Brady and R. Berwick (eds.), Computational Models of Discourse MIT Press, 
Cambridge MAr 331-371. 
\[Webber83\] Webber, B.L. 1983 So What Can We Talk About Now? in M. Brady and 
R. Berwick (eds.), Computational Models of Discourse MIT Press, Cambridge 
MAr 331-371. 
87 
