SIGMORPHON: Implemented Systems

This page is intended to give useful starting points if you want to find out more about implementations and toolkits. This information has been freely plagiarised from a range of sources, the page is still under construction, and the information is woefully out-of-date. If you would like to contribute to this page, please let us know.

AMAR: A Computational Model of Autosegmental Phonology

Author:

Daniel Albro (albro@humnet.ucla.edu)

Description:

Version:

Platform:

Anything with a C++ compiler (though it currently does not compile with newer versions of C++).

Development status:

Price:

Documentation:

A thesis.

For further information:

Author:

Philip Brassett (c/o geoffrey.hunt@sil.org)

Description:

Provides a portable acoustic phonetics laboratory, with spectrograms and pitch tracking

Version:

DOS: version 2.1, 1995
Windows: (WinCECIL) version 2.2, 1997 (Will be superceded by Speech Analyzer and some related programs)

Platform:

DOS: uses SIL CECIL Hardware Interface, a 12 volt rechargeable battery operated box that plugs into the parallel port, with a loudspeaker & throughput connection to a printer.
Windows: uses a Windows sound card or a built-in Windows sound system.

Development status:

mature system

Price:

Free (web distribution), printed manuals must be purchased

Documentation:

Three manuals: Interpreting CECIL, how to use CECIL for field linguistics; CECIL Reference Manual, hardware and software installation, configuration and use; recording techniques; problems and recovery; hardware servicing. Technical Support Manual for the CECIL Interface (CI500) (not included in the standard CECIL package), Hardware information for the maintenance technician; assistance for the programmer developing speech software to be used with the CECIL hardware.

For further information:

gopher://gopher.sil.org/11/gopher_root/computing/software/linguistics/speech

Authors:

András Kornai and Gábor Tóth (kornai@bbn.com)

Description:

A web program for reconstructing the accent marks often missing from Hungarian electronic texts. Our program, based on a statistical analysis of a large corpus of properly accentuated text, makes very little use of symbol-manipulation techniques, but performs reasonably well (1.58% error) because of the breadth of data it stores about the relationship between the unaccented and the accented forms.

Version:

Platform:

Web

Development status:

Price:

Free (Gnu public license)

Documentation:

http://www.cs.rice.edu/~andras/ekezes.html

Author:

David Bevan (david.bevan@sil.org)

Description:

FindPhone is a DOS program that you can use to analyse phonetic data. With FindPhone, you can generate and test hypotheses in order to discover the sound patterns that are the basis for understanding the phonology of a language.

Version:

6.0, 1995

Platform:

IBM PC or compatible running DOS 2.11 or higher, 512K of free RAM and a CGA, EGA, VGA or Hercules graphics display.

Development status:

mature system

Price:

US$38 plus shipping (includes a 226 page manual). Note that the software itself is free, but the manual is almost essential

Documentation:

FindPhone User's Guide: Phonological Analysis for the Field Linguist 234pp spiral bound, plus online context-sensitive help.

For further information:

http://www.sil.org/computing/findphone/findphone.html

Author:

Bart de Boer (bartb@arti.vub.ac.be)

Description:

Age2.exe is a windows 95 based application for investigating the role of self-organisation in the formation of vowel systems in a population of agents. A population of artificial agents that each are able to produce and perceive sounds in a human like way develops a realistic system of speech sounds.

Version:

2.1, 1997

Platform:

IBM PC or compatible running Windows '95.

Development status:

research system

Price:

Freely available, as long as the author is contacted first if any of the system's results are to be published.

Documentation:

rudimentary documentation can be obtained by clicking help/about.

For further information:

VUB AI-lab origins of language group

Author:

Frank Brandon (dec.)

Description:

Fonol is a programming language that simulates phonological rules of the sort described in Chomsky and Halle's Sound Pattern of English and Schane's Generative Phonology. It also incorporates the input and output filters (conditions) which came into common use at about the same time. These filters are currently restricted to the identification of a pattern and two actions: preventing the successful application of a specific rule or blocking (starring) the entire derivation.

Version:

4.2.1, December 1991

Platform:

DOS

Development status:

Not in active development

Price:

Documentation:

For further information:

http://128.2.242.152/afs/cs/project/ai-repository/ai/areas/nlp/morph/fonol/0.html

Author:

Mike Maxwell (mike.maxwell@sil.org)

Description:

An implementation of classical generative phonology and morphology (in the form of a recogniser and a generator), incorporating some more recent innovations from lexical phonology and realisational morphology.

Version:

Platform:

Prolog (for a textual interface) and SIL LinguaLinks (MS Windows) for a graphical interface

Development status:

In active development

Price:

Documentation:

An article in SIGPHON 1: Maxwell (1994)

For further information:

Author:

Steven Bird (steven@cogsci.ed.ac.uk)

Description:

A web-based lexical database tool for quantitative phonological research, incorporating speech files and powerful search facilities tailored to the needs of the phonologist. The current version is set up to work for a particular group of languages (Grassfields Bantu).

Version:

0.9, November 1997

Platform:

Any web browser supporting tables and forms, sound card

Development status:

beta testing

Price:

Free (Gnu public license)

Documentation:

Online interactive tutorial

For further information:

http://www.cogsci.ed.ac.uk/hyperlex/

Authors:

Arthur Dirksen and John Coleman (John.Coleman@phonetics.oxford.ac.uk)

Description:

Version:

1995

Platform:

Windows, 486PC, standard 16-bit sound card

Development status:

Price:

Documentation:

For further information:

http://ouplsun.phon.ox.ac.uk/~public/IPOX/ipox.htm

Author:

Sheila Williams (williams@psyvax.psy.utexas.edu)

Description:

Version:

Platform:

Development status:

Price:

Documentation:

PhD Thesis

For further information:

Author:

Richard Sproat (rws@research.att.com)

Description:

Lextools is a package of tools for creating weighted finite-state transducers from high-level linguistic descriptions. These descriptions include:

• Regular expressions or strings
• Lists of regular expressions or strings
• Context dependent rewrite rules
• Context free rewrite rules
• Various more specialized tools for building particular kinds of grammars -- for instance grammars that specify between the mapping between digit strings and the number-names for those strings.

Version:

3.0

Platform:

Unix

Development status:

Price:

Free for non-commercial use

Documentation:

Web pages, downloadable man pages

For further information:

http://www.research.att.com/sw/tools/lextools/

Author:

Avery Andrews (Avery.Andrews@anu.edu.au)

Description:

OTW attempts to generate phonological structures in accord with some of the principles of Optimality Theory, implementing some of the analyses described in Prince & Smolensky (1993/1995) Optimality Theory. Areas covered are (a) `Galilean' Syllable Theory (ch 6); (b) Basic Syllable Structure Theory (ch 8); (c) Constraint Interaction in Lardil (ch 7) (partial; long vowels and vowel-sequence reduction are not handled).

Version:

1.1, August 1995

Platform:

Windows 3.1

Development status:

Price:

Documentation:

Paper: Andrews (1994)

For further information:

http://www.anu.edu.au/linguistics/software/otw.html

Author:

Markus Walther (walther@ling.uni-duesseldorf.de)

Description:

OT SIMPLE is a software tool for modelling analyses in the Optimality Theory framework (Prince & Smolensky 1993) on a computer. It reuses the standard UNIX tools 'sed' and 'sort' for implementing constraints CON and optimization EVAL, whereas the structural grammar for GEN is interpreted by a separate component written in BinProlog. CON and GEN are fully user-definable.

Version:

1.0

Platform:

Unix with BinProlog

Development status:

usable for research and pedagogical purposes

Price:

Free (Gnu public license)

Documentation:

Paper: Walther (1996)

For further information:

http://www.phil-fak.uni-duesseldorf.de/~walther/otsimple.html

Author:

Evan Antworth (evan.antworth@sil.org)

Description:

PC-KIMMO is an implementation of a program dubbed KIMMO after its inventor Kimmo Koskenniemi. The program is designed to generate and/or recognize words using a two-level model of word structure in which a word is represented as a correspondence between its lexical level form and its surface level form.

Version:

2, 1995

Platform:

DOS, Macintosh, UNIX

Development status:

mature system

Price:

software free when obtained via the Internet

Documentation:

A book: Antworth (1990), $24 plus shipping (includes software)
Online documentation at http://www.sil.org/pckimmo

For further information:

http://www.sil.org/pckimmo

Supplementary software:

Englex, a PC-KIMMO morphological grammar/lexicon for English, is available at http://www.sil.org/pckimmo/v2/englex.html.
Englex documentation is at http://www.sil.org/pckimmo/v2/doc/englex.html.

Author:

Lee Hartman (lhartman@siu.edu)

Description:

Phono is a Windows software tool for developing and testing models of regular historical sound change.

Version:

4.1, 2003

Platform:

Windows

Development status:

Functioning, but not under active development

Price:

Free (web distribution)

Documentation:

User's manual in "read.me" file, online help screens

For further information:

http://mypage.siu.edu/lhartman

Authors:

Paul Boersma and David Weenink (paul.boersma@hum.uva.nl)

Description:

a system for doing phonetics (and some phonology) by computer. Comprehensive speech analysis, synthesis, and manipulation package, including general numerical and statistical stuff, built on a general-purpose GUI shell for handling objects and producing publication-quality graphics. Contains several learning algorithms (neural nets, principles & parameters, discrete and stochastic OT, adaptive resonance).

Version:

4.1, May 2003

Platforms:

Macintosh, Windows, Linux, SGI, Solaris, HPUX.

Development status:

mature status, though actively extended.

Price:

free.

Documentation:

1000 pages as of May 2003, browsable on-line in the program, on hardcopy, and at http://www.praat.org; Pitch-extraction algorithm, articulatory synthesis, and the learning of stochastic OT grammars are described in papers available from http://fon.hum.uva.nl/paul/.

For further information:

for up-to-date info, visit http://www.praat.org/

Author:

John Wimbish & Karen Buseman (Karen.Buseman@sil.org)

Description:

Shoebox is a database program oriented to the needs of a field linguist's dictionary. It builds a lexicon either directly or through interlinearizing text. (It can be used for other types of data also.)

Version:

3.02, January 1997

Platform:

Microsoft windows, Macintosh

Development status:

mature system

Price:

approx US$20 plus shipping

Documentation:

For further information:

http://www.sil.org/computing/shoebox.html

Author:

Terry Gibbs (Terry.Gibbs@sil.org)

Description:

Version:

1.0, 1997

Platform:

Windows (requires a sound card)

Development status:

Price:

Documentation:

For further information:

http://www.jaars.org/icts/softdev.htm

Author:

Michael Hammond (hammond@u.arizona.edu)

Description:

1. A Perl program with a web interface for OT-style parsing English and French words into syllables.
2. A Prolog program with a web interface for OT-style parsing of syllables, allowing user control over constraint ranking.

Version:

Platform:

Web (also 1. anything running Perl; 2. anything running Prolog)

Development status:

1. Not under active development.
2. Continuing to be developed.

Price:

Free

Documentation:

Two papers: Hammond (1994), Hammond (1997)

For further information:

http://www.u.arizona.edu/~hammond/

Authors:

William Fisher (william.fisher@nist.gov)

Description:

A C implementation of Daniel Kahn's theory of English syllable structure

Version:

1.1, August 1996

Platform:

Anything with a C compiler

Development status:

Price:

Free

Documentation:

For further information:

ftp://jaguar.ncsl.nist.gov/pub/tsylb2-1.1.tar.Z

Authors:

Xerox PARC (webmaster@xrce.xerox.com)

Description:

The Xerox Finite-State Compiler allows you to enter a regular expression in a text window and compile it into a finite-state network. Depending on the expression you type, the result is either a simple automaton or a transducer.

Version:

Platform:

Web, Unix

Development status:

commercial system

Price:

Documentation:

Web pages

For further information:

http://www.rxrc.xerox.com/research/mltt/fst/home.html

Authors:

John Local, John Coleman, Richard Ogden and Steve Harlow (lang4@york.ac.uk)

Description:

A speech generation system incorporating real phonology...

Version:

Platform:

Development status:

Price:

Documentation:

For further information:

http://www.york.ac.uk/~lang4/Yorktalk.html