I am currently developing a program that allows the user to search for words with various properties (e.g., the property of having all voiced consonants, or the property of having a certain syllabic stress pattern). I hope that by studying patterns in English words, we can learn more about human nature and our cultural habits of thought.

For this end, I would like to hear suggestions from expert linguists about what features would make the program more useful to the scientific community. I am already thinking about letting the user define his or her own properties.

Any suggestions are welcome and most appreciated.

Look up corpus linguistics and phonetic corpora, there should be some stuff on that already.

What is the searchable body of text? The features you want to have the program look for have to be annotated or indicated in some other way in the text, of course.