Here are some references to get you going, in pseudo-BiBTeX format.

A good survey paper is:

author="C. D. Paice", title="Constructing Literature Abstracts by Computer: Techniques and Prospects", journal="Information Processing and Management", year="1990", volume="26", number="1", pages="171--186"

This will suggest many other references. For a more theoretical discussion, see:

author="Sparck Jones, K.", title="What might be in a Summary?", booktitle="Proceedings of the German Information Retrieval Conference", year="1993"

Particular summarising methods:

word-frequency methods:

author="H. P. Luhn", title="The Automatic Creation of Literature Abstracts", editor="Schultz", booktitle="H. P. Luhn: Pioneer of Information Science", publisher="Spartan", year="1968"

author="S. Williams and K. Preston", title="Managing the Information Overload", journal="Physics in Business", publisher="Institute of Physics", year="1994"

(this system -- BT's NetSumm -- can be tried at

author="E. F. Skorochod'ko", title="Adaptive Method of Automatic Abstracting and Indexing", booktitle="Information Processing 71", year="1971", pages="1179-1182"

@techreport author="M. Benbrahim and K. Ahmad", title="Computer-aided Lexical Cohesion Analysis and Text Abridgement", series="Knowledge Processing", number="18", institution="University of Surrey", year="1994"

c(l)ue phrase methods:

author="J. E. Rush and R. Salvador and A. Zamora", title="Automatic Abstracting and Indexing. {II}. {P}roduction of Indicative Abstracts by Application of Contextual Inference and Syntactic Coherence Criteria", journal="Journal of the American Society for Information Science", year="1971", month="July", pages="260--274"

author="C. D. Paice", title="The Automatic Generation of Literature Abstracts: an Approach based on the Identification of Self-indicating Phrases", booktitle="Information Retrieval Research", editor="R. N. Oddy and S. E. Robertson and C. J. van Rijsbergen and P. W. Williams", year="1981", publisher="Butterworths", pages="172--191"

methods which use domain-knowledge:

author="G. F. DeJong", title="An overview of the FRUMP system", editor="Lehnert and Ringle", booktitle="Strategies for Natural Language Processing", publisher="Erlbaum", address="Hillsdale HJ", year="1982"

@techreport author="J. I. Tait", title="Automatic Summarizing of {E}nglish Texts", number="47", note="PhD thesis", institution="University of Cambridge Computer Laboratory", year="1983"

As far as I am aware, there are very few comparative studies. Here is one:

@techreport author="P. Gladwin and S. Pulman and Sparck Jones, K.", title="Shallow Processing and Automatic Summarising: a First Study", number="223", institution="University of Cambridge Computer Laboratory", year="1991"

You may also be interested in theories of how people summarise text as they read it, in which case take a look at:

author="T. A. van Dijk and W. Kintsch", title="Strategies of Discourse Comprehension", publisher="Academic Press", address="New York", year="1983"

author="W. Kintsch and T. A. van Dijk", title="Toward a Model of Text Comprehension and Production", journal="Psychologial Review", year="1978", volume="85", number="5", pages="363--394"

For information about how professional abstracters work, there is lots of good work by Liddy, for example:

author="E. D. Liddy", title="The Discourse-level Structure of Empirical Abstracts: an Exploratory Study", journal="Information Processing and Management", year="1991", volume="27", number="1", pages="55--81"

author="E. D. Liddy and S. Bonzi and J. Katze and E. Oddy", title="A Study of Discourse Anaphora in Scientific Abstracts", journal="Journal of the American Society for Information Science", year="1987", volume="38", number="4", pages="255--261"

Richard Tucker []