Does anyone know of any sources to check the relative frequency of various consonant places of articulation in word-initial position in English (or any other language)?
In other words, what percentage of word-initial consonants in English are coronal, labial, dorsal, etc.?
Any corpus which includes phonological annotations would work for this. For instance, the CMU corpus (http://www.speech.cs.cmu.edu/cgi-bin/cmudict)
You'll also have to consider whether you're interested in the type or token frequencies.
Thanks for this, Stephen. I think my query involves both type and token frequency. I am essentially looking for some documentation of a generalization that is not really controversial: namely that coronals are the most common initial consonants in English. I am 99.99 percent sure that this is true, especially considering the fact that coronals incorporate θ, ð, s, z, l, n, t, d, ʃ and tʃ.
We checked this for a subset of fricatives in one of our papers (http://www.tandfonline.com/doi/full/10.1080/23273798.2016.1151058) and that was indeed the case, although we didn't check the rest. But yeah, I think some of the coronal underspecification literature has looked into this.
For type frequency this will be straightfoward to do from the CMU corpus. For token you'd have to combine those findings with some other corpus (like celex, mcword, etc) that has token frequencies in it.
Hi Geoff! If you want the type frequency, then CUBE (always the first place to go to for a rough idea) gives 39821 word-initial coronals out of a total of 85576 word-initial consonants in a dictionary of 103208 entries. Not an absolute majority, so you will need to dig a bit more. Or shall I?
(Edit: OK, just had a second look and it's 29270 for the labials. Settled. For token frequency, we'd have to do what Stephen says.)
http://seas3.elte.hu/cube/
You may also want to check the CELEX database for Dutch, English and German (Baayen et al., 1995). The English part is based on a 17 million word corpus (mostly - if not exclusively - based on written texts). You can chose between type (lemma) or token (word form) corpus and then you just count - with a simple script - all entries that start with a coronal consonant in English. CELEX comes with four different transcription systems all of which are ASCII coded.
If you were able to find a database of English words in IPA and import that into SIL's Phonology Assistant program (https://www.sil.org/resources/software_fonts/phonology-assistant, its free) you could easily answer your question and also look at it from multiple angles.
If you can't find a database in IPA, you could import the CMU corpus above, but you would have to define the phonological features of each of the graphemes and digraphemes. If would be a little bit of work, but not too much. You could then easily compare your results against a token frequency list such as found at (http://www.wordfrequency.info/)
- Badges
- Science topic
- Similar topics
- Linguistics
- Philology
- Language
- English