Call for Papers:
Communicative Action in Humans and Machines
November 8-10, 1997, MIT, Cambridge, MA, USA
Contents
Introduction
Since at least the 50s when Austin told us how we do things with
words, it has been recognized that language performance can be
fruitfully viewed as action. There has subsequently been a range of
work reasoning about the action involved in the spoken language
communication process (speech acts), using both formal and empirical
methods. Views of communication as action have also been influential
in reasoning about machine communication in multiprocessor or
distributed systems. Moreover, many human-computer interactions have
also been described as actions similar to Austin and Searle's speech
acts. In recent years there has been an increased emphasis on theories
of action covering other aspects of the communication process,
including other modalities than speech and other aspects of dialogue
than the illocutionary acts associated with the utterance of
sentences. There has also been much subsequent work in philosophy,
logic, linguistics, and AI on the nature of actions, which can help
shed light on communicative action. We seek to bring together
researchers from a variety of perspectives on action in communication,
to discuss these issues, including the current state of the art and
assess prospects for synergy and future applications.
The symposium will focus on the following themes:
- Theories of action and agency to support representing and reasoning
about communicative action.
- Theories of communicative action including other modalities than
speech, and non-traditional levels of action.
- Empirical investigation of communicative action.
- Use of communicative action in applications.
- Relations between the communicative action of differing types of
communicators (humans, machines, and mixtures of the two).
- Relations between communicative action and other kinds of physical
and mental action (e.g., reasoning and learning).
Submission Information
Potential participants should submit each of the following:
-
Name, physical and electronic addresses, also fax number and WWW URL
if available. If several people working together (e.g. collaborating
authors) wish to attend, each should submit separately, but should
also name the others in the group.
- Bibliography entries to related papers (preferably in html and/or
bibtex format), and links to URLs related to the theme of the
symposium. These will be made publicly accessible via the symposium
WWW page: http://www.cs.umd.edu/users/traum/CA/
- Either an extended abstract of a research paper to be presented at
the symposium, or a brief statement describing why you wish to attend
and how you believe that you can contribute to the symposium (describe
your own related work and/or specific questions and issues that you
feel should be addressed in the symposium). Abstracts should be no
more than 10 pages (exclusive of references) in plain text or
postscript files (12pt).
Please send your submission via e-mail to traum@cs.umd.edu
Contributions should address one or more of the following questions:
- Are existing AI theories of (physical) action adequate for
representing and reasoning about communicative action?
- If not, how can they be adapted to serve this function, or are
different approaches required?
- What kind of representation of the communicating agents (including
"mental states" such as belief and intention) is necessary to model
the conditions and effects of communicative actions?
- How can a particular theory of action be empirically tested for
validity, utility, etc.? What methods (e.g., corpus-based, system
building, empirical investigations) can help elicit deeper and broader
models of communicative action?
- What are the compelling applications in which reasoning about
communicative actions are a requirement (e.g., interfaces,
communication analysis, network/agent management)?
- Is the "speech act" a useful intermediary concept (e.g., for
representing intentions), or is a direct "context-change" model more
appropriate to the tasks?
- Are speech act theories developed for human-human interaction adequate
or appropriate for machine-machine or human-machine communication? If
not, can they be felicitously adapted to serve as such?
- Can machine communication "simulations", communicating using explicit
speech acts in their communication protocols, provide useful insights
into the human communication process, in which speech act
interpretation is also a necessary component?
- Are theories of speech acts well suited for analyzing other
communication modalities, such as gestural communication in humans or
graphical presentation in machine interfaces?
- What kinds of dialogue actions, other than sentence-level speech acts,
occur in dialogue? How do these kinds of actions relate to traditional
speech acts?
- What is the relationship between communicative actions and rhetorical
relations, for example, in the context of generating
multisentential and multimedia presentations (e.g., is there a
"hierarchical" element of communication?).
- What is the relationship between dialogue acts and other kinds of actions
affecting mental states, such as reasoning and learning?
- When language is only part of an interaction also involving
non-linguistic domain action, what is the relationship between speech
and other action, both for communication and task performance?
April 15 | Deadline for submission of papers |
May 15 | Notification of acceptance |
August 22nd | Papers due for the working notes |
November 8-10 | Symposium |
Organizing Committee
David Traum