English Phonology Search

A program by

Bruce Hayes
Department of Linguistics

Download the program


What the program does

This program assumes that you are a phonologist, phonetician, psycholinguist, etc., and want to find words of English that have some particular property.  For example, you could use it find words that end in a voiced obstruent, or which have a /t/ that is flanked by stressless vowels, or which have two rounded vowels in consecutive syllables, etc., etc.

You type what you want in little search windows (see interface below), click a button, and the program will search for all the words that match your description.

The most important part of the program is that it can use natural classes; that is, classes of sounds defined by some set of phonetic properties, such as "voiced obstruent" or "stressless vowel."  You can easily add whatever natural classes you want by making up little files that list the sounds in the class.

Beta version; please report bugs to me at bhayes@humnet.ucla.edu.

Download and installation

The program, regrettably, runs only in Windows.  

If you click the Download button above, you'll get a zip file, which, if unzipped, will produce a single folder with everything in it.  Put this folder wherever you like, renaming it however you like.

I've tried not to use any fancy stuff in creating this program.  As a result, there is no need to do "install" it; it should just run when you click on it.  The program itself (the thing you click on) is called EnglishPhonologySearch.exe.

Rough guide to the interface

Looking at this picture should give you the basics.  But to learn to use the program properly, please click on Help and read what it says.  It's not long.

In this picture, the interface is set up to search for words that have the vowel "IH1" (which means lax high front unrounded main stressed vowel), followed by a coronal sonorant and then a noncoronal obstruent, which is word-final.  The search returned only four words, namely bilk, ilk, milk, and silk.

English pronouncing dictionary

The program uses an edited version of part of the widely-used CMU Pronouncing Dictionary, included as the file "in.txt" in the download package.  This file is an kind of "intersection" of CMU and CELEX; specifically, all the words in CMU that have a CELEX frequency of at least 1.  The idea is to have a vocabulary representing American English pronunciation (and thus suitable for preparing experiments with American participants), with lexical frequencies high enough that the words are likely to be known to most or all of the participants.

The file also has exclusion codes, with the goal of marking words that are compounds or formed with productive suffixes.  (This is an option on the program interface.)

Both the pronunciation entries and the exclusion codes still need lots of corrections, which I would appreciate if you send to me (bhayes@humnet.ucla.edu).  For convenience it would be nice if you save up a batch of them rather than sending me repeated messages.

Other languages?

This would not be too hard, if you've got an electronic dictionary.  For how, please start the program, click on Help, then on Adapting The Program To Other Languages.

Program details

English Phonology Search was programmed by Bruce Hayes in 2011 using Microsoft Visual Basic 6.  Source code on request.

Bruce Hayes's home page