Tips on Writing Abstracts for Conferences
Handout from workshop for the Linguistics Club
Caroline Wiltshire, Asst. Professor, Program in Linguistics

Before writing:
Attend conferences whenever possible
Find conference announcements:
Bulletin Board in Anderson Hall
Linguist List
Have a conference in mind, and research it:
who will be reading the abstracts? faculty? students?
     specialists? generalists?
what theories can you assume knowledge of?
any theoretical biases?
what kinds of papers are usually presented at this conference?
how long are the presentations? will the papers be published?

When Writing:

     Start Early! You want to be able to show your abstract to a critical reader and revise it.

     Have well-defined, worthwhile objectives:
          specific point your paper will prove (that can be accomplished within the time)
          more general point to hint at; the issues addressed should be topical/important

     Have an argument: sketch it in the abstract provide relevant data in an organized way

     Discuss alternatives: include arguments for preferring the given analysis over others
                                    demonstrate familiarity with the relevant literature
                                    relate your work to recent literature so that you don't seem isolated/naive

     Be civil, even generous in your criticisms of others' works

     Pay attention to format:      obey the format guidelines; don't go over the length limit.
                                              make it easy on the eye for the poor abstract readers
                                              make it easy on the brain for the poor abstract readers: be clear, concise,
                                                     do not load with abbreviations, give (only) enough background for
                                                      comprehensibility, do not make leaps in argumentation
                                              do not keep your conclusions a secret! Reveal them early.

      Make the abstract look as if the paper were already written
            be as specific about your conclusions as possible
            avoid the future tense (e.g., this paper will show )
            do fill up allotted length; too short looks worse than too long

Before submitting:

        Have someone else read it! Beg for criticism.
        Abstracts will be rejected if they: do not present or clearly specify the data
                                                        contain greap leaps in reasoning or gross flaws in argumentation
                                                        obviously misconceive the issues they deal with or the proposals they criticize
                                                        express undue hostility towards the authors of works they criticize
                                                        deal only with data that have already been beaten to death in the literature
                                                        cite only unpublished works by unknowns
                                                        seem to say nothing beyond what has already been said
                                                        promise to resolve several fundamental questions of linguistics in 20 minutes

Typical abstract:
First paragraph:
state topic and background
state specific problem or question to be discussed
specific conclusion
Second paragraph:
lay out the problem in detail
present data
Third paragraph:
lay out the solution in detail
show how it handles the problem
explain what is new/interesting about it
Fourth paragraph:
consider alternatives (past or present)
show that they can't handle the data, or
show that they are inferior to your proposal in some way
Fifth paragraph:
conclusions, implications, and
more general connections with the field

What constitutes a point? (may vary, depending on conference)
argue to extend a theory to accomodate your data
discuss the predictions such an extension makes or the implications of accepting it
two theories can both explain some old data; your data will decide between them
a recently proposed theoretical device provides an insightful analysis for old data
a cross-linguistic typological generalization can be explained by a theory
a theoretical idea from one domain can be applied insightfully to a new domain or to phenomena that appear to be quite different
describe and analyze a phenomenon that no one has ever noticed before present data that has not been gathered before (that bears on a theory's formulation)

Further advice and examples are available in the LSA bulletin No. 158, December 1997
(available from me as a loan for a day or two for people at UF).

Back to my homepage: Caroline R. Wiltshire

Also see