PhonologicalCorpusTools / CorpusTools

Phonological CorpusTools
http://phonologicalcorpustools.github.io/CorpusTools/
GNU General Public License v3.0
111 stars 16 forks source link

Feature system issues #748

Closed stannam closed 3 years ago

stannam commented 3 years ago
  1. [h] is under uncategorized (as mentioned in issue #584)

  2. In SPE, [ɹ] is a mid front vowel. image

  3. compared to the list of SPE features in Mielke (2008), a number of features are missing in ipa2spe.feature.

  4. due to 3., [m] and [m̩] (syllabic m) are identical in features. (c.f., other syllabic consonants are not specified in the feature file)

stannam commented 3 years ago

list of features missing in ipa2spe.feature, but present in Mielke (2008, p. 50). The Sound Pattern of English is open-access here, but I haven't checked if they all appear in SPE itself.

[consonantal]
[syllabic]
[covered]
[delayed release of secondary closure]

[consonantal] and [syllabic] are in boldface as they seem to be critical for categorization.

Currently, we use [voc] as a determinant of vowel/consonant (instead of [consonantal]). Due to the lack of [syllabic], [m] and [m̩] are identical in features.

Do we want to add those missing features?

kchall commented 3 years ago

Hmm, [consonantal] should definitely be in there, as it's part of the 'original' set of features in chapter 7. It's defined on p. 302 as "Consonantal sounds are produced with a radical obstruction in the midsagittal region of the vocal tract; nonconsonantal sounds are produced without such an obstruction." Here's a table of their original major class features:

image

They do suggest in chapter 8 (p. 354) that [vocalic] should be replaced with [syllabic], as in the following table, to specifically handle syllabic consonants:

image

I think I'm inclined to simply have BOTH [vocalic] and [syllabic] in there, since PCT is much more about descriptive features than it is about analytical features.

Will this take care of all the above issues? What about [h]?

stannam commented 3 years ago

[h] will still be uncategorized, unless we add another column for 'glottal'.

Note: SPE classifies [h] as a glide, not fricative.

kchall commented 3 years ago

If you mean a column in the actual consonant inventory chart, that's fine, and yes, fine to put it in as a glide, as that's what SPE does (as with glottal stop). I don't want to add a [glottal] feature though. The place features for [h] (p. 307) are: [-ant, -cor, -high, +low, -back].

stannam commented 3 years ago
stannam commented 3 years ago
kchall commented 3 years ago

@stannam can you find a page reference for [ɹ] being [+lat]?

stannam commented 3 years ago

@kchall Oooops.. Sorry, I think I was not fully awake then. I meant [+continuant], and when I checked now both [l] and [ɹ] are [+cont]! The book even mentioned that '[l] [...] must be distinguished from [r] by the feature of "laterality."' (p. 318, Ch. 7) image

stannam commented 3 years ago
stannam commented 3 years ago

cf. Phonological_CorpusTools_Public/other/example_corpus_in_different_transcription_systems.xlsx

stannam commented 3 years ago