1

Looking for a systematic online step-by-step process to codify English words into vowel/consontant patterns (CVC, CVCe, CVVC, etc.) and the correct sound (long-vowel, short-vowel, blend, diagraph, etc.).

6
  • Written words in ASCII, or spoken words in an audio format? "English words" means the latter, unless you specify more clearly. And which kinds of vowel and consonant are you looking for? Phones? Phonemes? Allophones? Syllables? There are a lot of things you could be asking for, some of which clearly don't exist, with others that clearly do. – jlawler Dec 15 '11 at 16:09
  • @jlawler: "Dictionary notation" to CVC, beyond that I'd say the symbols should be limited to get would be found if you used a dictionary style rendering of word sounds; if you'd like it limit it to one online dictionary, that's find. Basically I'm not even able to find an online dictionary that would let me query the 40k common words in English, so I'm attempting to make my own. Beyond that, I'm unable to find a notation standard, for example "crop" is CVC, and not CCVC. Let me know if that was of any help, since if I knew the answer to you questions, I wouldn't be asking the question. Thanks! – blunders Dec 15 '11 at 17:17
  • 2
    OK, then don't use an American dictionary. The sound symbols should be either IPA or K&K, and American dictionary publishers think Americans are too stupid to use these, so they still use Webster's 18th-Century non-phonetic system. With one exception, maybe. K&K standard notation is an English phonemic system used in Kenyon & Knott's Pronouncing Dictionary of English, published 1953 by Merriam-Webster. It gives just the phonemic transcription of most common words. – jlawler Dec 15 '11 at 19:51
  • @jlawler: Yes, I know. If you're able to create an answer that uses IPA and supports converts to "dictionary notation" that's fine. So, is what I'm looking for clear now? – blunders Dec 15 '11 at 20:12
  • 1
    The cheapest way would be to scan K&K and then do OCR. It'd take a lot of training, but unless K&K is online, which I doubt, it's the only way to get that information electronically. – jlawler Dec 15 '11 at 20:42
1

The task you are asking for is a problem encountered in text-to-speech systems and is called grapheme-phoneme-conversion. A system performing this task takes an english word and translates it to a phonetic transcription. If you search for this on Google you will find information how it is done and there are even some freely available systems for this task, I didn't find one that works online though. If you want CVC instead of a phonetic transcription, it should be a simple replacement task from there.

About your data problem: To get a list of frequent english words have a look at the 1-Grams at Google Ngrams or get an english wordlist and sort it by word frequency. If you are interested in dictionary entries, you can look up the words in Wikipedia afterwards.

All in all this is a task which can be done, but you probably won't solve it if you don't have a little experience in programming, as I doubt that there are online tools which do exactely what you want to do.

7
  • PS: if you are working with NLP-tasks more often you should have a look at bit.ly/u4lZUG. – tobigue Dec 16 '11 at 20:31
  • Thanks, main problem is not data or mapping the transformation, it's finding a standard for the transformation. I have yet to find specs for CVC pattern notation from any source. Thanks for posting an answer, since I'm aware this a none standard topic. – blunders Dec 16 '11 at 20:36
  • Thanks, I'll take a look at the Area51 NLP exchange later, though this is really more of a token transformation issue than a language processing problem in my opinion; meaning the task is a simply transcribing via a standard method from one form to another, though might be wrong. – blunders Dec 16 '11 at 20:40
  • FYI in rereading your answer, getting phonetic transcriptions is not the issue; example, Google "define: phonetic". The issue is a schema for a phonetic notation, for example Google's to CVC notation. Again, thanks for the input! – blunders Dec 16 '11 at 20:44
  • yes, but it will be much easier to derive a CVC notation from a phonetic notation than from normal text. so if you don't find a better way you might want to go around that corner. – tobigue Dec 16 '11 at 20:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.