Welcome to the Gsearch homepage. Gsearch is a tool designed to facilitate the selection of sentences from text corpora by syntactic criteria, even where these corpora contain no prior syntactic markup. The Gsearch developers can be contacted at gsearch-dev@informatics.ed.ac.uk.

On this page, you can find:

• The Gsearch sources and documentation;
• The Gsearch mailing list;
• Electronic versions of published research using Gsearch;
• A repository of Gsearch resources.

Latest News:

Apr 28, 2005

Gsearch has a minor update: version 2.07 contains a bug fix for the BNC filter. This is now the default version for download. Also available is a patch for version 2.06w (instructions are in the file).

Oct 10, 2001

Gsearch has a minor update: version 2.06w is aware of Windows-style line endings, for easier operation under Cygwin. This is now the default version for download. Also available is a patch for version 2.06 (instructions are in the file).

Sep 30, 2001

Steffan Corley has experimented with a port of Gsearch to Windows. He found that Gsearch compiled and ran with no modification under Cygwin. This is very encouraging, though we would like to hear from other Windows user to see there are any liminations to Gsearch functionality under Cygwin.

Sep 14, 2001

Gsearch 2.06 is now available for download. Among the new features of this version are:

• entirely rewritten user documentation, covering all of the enhancements in version 2 and above;
• greatly simplified install process, based on GNU automake/autoconf, and improved portability;
• numerous bugfixes and enhancements (including off-the-shelf ability to read gzipped corpora).

Gsearch 2.07 has been tested on Solaris (5.5.1, 5.6, 5.7, 5.8), Linux (RedHat 6.1, Mandrake 7.2), and MacOS X.

Please note that we do not intend to support versions of Gsearch prior to 2.07 in the future.

Sources and Documentation

Gsearch is designed to run on UNIX-like systems. To date, it has been tested on Solaris, Linux, MacOS X, and Cygwin under Windows.

You can obtain the full Gsearch distribution here (gsearch-2.07.tar.gz). The distribution contains the main Gsearch system, the user manual, a paper describing the Gsearch system, and several example grammars, filters, and pseudofield scripts.

You can also obtain the manual separately here (gzipped postscript).

Mailing List

If you download Gsearch, you are strongly encouraged to join the Gsearch users' mailing list. This is a low-traffic list used to discuss Gsearch usage and to announce new releases of Gsearch and associated software.

You can subscribe to the mailing list using this page.

Publications

Gsearch is described in

• Corley, S., Corley, M., Keller, F., Crocker, M., & Trewin, S. (2001). Finding Syntactic Structure in Unparsed Corpora: The Gsearch Corpus Query System. Computers and the Humanities, 35, 81-94.

Note: this paper is included in the Gsearch distribution.

Other Publications which make use of Gsearch:

• Lapata, M. (1999). Acquiring lexical generalizations from corpora: A case study for diathesis alternations. In Proceedings of the 37th annual meeting of the Association for Computational Linguistics (pp397-404). College Park, MD.
• Lapata, M. & Brew, C. (1999). Using subcategorization to resolve verb class ambiguity. In P. Fung & J. Zhou (eds.), Proceedings of Joint SIGDAT Conference on Empirical Methods in Natural Language Processing and Very Large Corpora (pp266-274). College Park, MD.
• Lapata, M., Keller, F., & Schulte im Walde, S. (2001). Verb frame frequency as a predictor of verb bias. Journal of Psycholinguistic Research, 30, 419-435.
• Corley, M. & Haywood, S. (2000). Parsing modifiers: The case of bare-NP adverbs. In Proceedings of the 21st annual meeting of the Cognitive Science Society (pp126-131). Vancouver.
• Zamparelli, R. (1998). A Theory of Kinds, Partitives, and OF/Z Possessives. In C. Wilder and A. Alexiadou (Eds.), Predicates and Movement in the Determiner Phrase. Amsterdam: John Benjamins.

Gsearch Resources

Most publically available resources are currently distributed with Gsearch. If you have additional filters, grammars, or pseudofield scripts (or code enhancements) please contribute them to the Gsearch project by mailing gsearch-dev@informatics.ed.ac.uk.

Currently available:

BNCditrans.tar.gz

Grammar used with BNC to count NP-NP vs. NP-PP alternations for ditransitive verbs in English. Requires WN_anim.pl (distributed with Gsearch).

An improved raw.pl filter

The raw.pl filter (distributed with Gsearch) provides simple-minded segmentation of raw text files. With help from Garance Paris, this new version is locale-sensitive (much improving the segmentation of languages other than English).

The development of Gsearch was funded in part by the Economic and Social Research Council, and in part by collaborative ARC project 1024 of the British Council and the Deutscher Akademischer Austauschdienst.