Linguist Forum

Specializations => Phonetics and Phonology => Topic started by: daniel.c.gallagher on July 08, 2018, 10:43:22 PM

Title: IPA k vs kk
Post by: daniel.c.gallagher on July 08, 2018, 10:43:22 PM
Hi there!

I'm new to linguistforum. As a prospective linguistics phd, I thought it'd be good to get set up somewhere I can get help when needed (coming from an engineering background, said help will probably be primarily basic stuff).

I'm practicing an entrance exam for a linguistics department and the question requires separating syllable boundaries (with a period) in the IPA transcriptions of words from the language Alabaman. I think I understand syllable boundaries generally speaking pretty well, but one word is throwing me off.

[okkʰi:tʰatkʰa:] ('sea')

First, I'm wondering what is the difference between the above and a hypothetical word [okʰi:tʰatkʰa:] (in other words, what's the difference between [kkʰ] and [kʰ])?

Second, I'm thinking the syllable boundaries would be: [ok.kʰi:.tʰat.kʰa:]. What do you think?

I appreciate your help!!
Title: Re: IPA k vs kk
Post by: Daniel on July 09, 2018, 12:01:24 AM

Languages do syllabification differently, so the only way to answer this correctly is to know the specific syllabification rules/patterns for this language.

There are some reasonable ways to guess (consonant clusters tend to split, for example) but that's just a guess.

Typically geminates (doubled consonants) are found as syllable boundaries, as you intuit. It's hard to imagine a language with phonetically geminate onsets or codas in a single syllable. (I wouldn't entirely rule out the possibility, because languages do some almost unimaginable things.) If they were to appear in those positions phonemically, they'd probably simplify phonetically. But at the same time, sometimes geminates aren't actually pronounced like two copies of the sound anyway (for example, a glottal stop followed by the second consonant), so it's hard to guess about the assumptions of the question.

The best advice I can give you is to answer the questions on any kind of exam based on the instructions and assumed knowledge for that exam. General answers may be 'wrong' in that context, and theories vary (as do notations!).
Title: Re: IPA k vs kk
Post by: panini on July 09, 2018, 09:23:48 AM
There is an article, "The Structural Phonology of Alabaman, a Muskogean Language", published in 1968 in the International Journal of American Linguistics, which "touches on" the topic, but I would say that the question hasn't actually been touched, there has simply been hand waving. You should, however, be aware that linguists have a bad habit of making stuff up. You might want to look at Montler's Alabaman dictionary at, which gives the form [okihatka] as "ocean". You should note that aspiration is not phonemic in the language, so the transcription you've been given is an "interpretation" of something else.

My reading of Rand's article indicates that the [kkʰ] transcription may be the phonetic property of being "fortis, slightly aspirated" in onset position. He actually exemplifies k-allophony with the word [okki:tatka:] "sea", without the aspiration or the syllable boundary. In other words, somebody else modified Rand's data, assuming some rules. This then raises the question of how Montler's data conflicts with Rand's. You could in fact ask Montler -- it may be disagreements in how to interpret the same phonetic output, or there could be real language differences, though Rand and Montler partially used the same informant.

It is not clear what the "question" requires of you / tells you. In my opinion, it is unreasonable and professionally outrageous to give a transcription like "[okkʰi:tʰatkʰa:]" and require a person to supply syllable boundaries. It is well known that "syllable boundaries" are not phonetic transcriptional primitives, they are a phonological device invoked to simplify certain phonological rules. That means that unless you have some reason to posit a syllable bounday, for some phonological purpose, you should not provide any syllable boundaries. Such a request is analogous to asking "What is the underlying form of [akʰatta]", when you have no basis for saying anything about underlying forms.

Since this is such an outrageous requirement, I don't actually believe that that is what you're being told to do, so maybe re-read the instructions, and provide a bit more information on the data and instructions.

Title: Re: IPA k vs kk
Post by: daniel.c.gallagher on July 21, 2018, 09:41:40 PM
Thank you both for your very helpful responses!! (and sorry for my late response; I was expecting an email when a response came in, so I'm still figuring this forum out).

Daniel, I appreciate your response very much. I suspect that reasonable guessing is all that is really wanted out of this particular question of the test. To that end, your tips are very helpful, so thank you!

Panini, thanks so much for doing some extra research on the matter! It is confusing that there would be such a discrepancy. Anyway, I don't think I've misinterpreted the directions. Just for good measure, I've included the instructions here as shown on the exam (in the original Japanese) and with my translation:

[alabaman word data list as previously included, with more examples]
(Rand, Earl (1968) The Structural Phonology of Alabaman, A Muskogean Language.
International Journal of American Linguistics, vol. 34-2, pp. 94-103. の表記を若干改変)

Q1. The following data is of Alabaman. Answer the questions below.
[Alabaman word list]
(Rand (1986)... has been slightly altered)
[Q1-1] For a,b,d,k,p, place a dot at the syllable boundary, as shown in the example.
So at least, there is an acknowledgement that the original data has been altered and might even just be fictitious, probably for the sake of exercise. Regarding the role of syllables in simplifying phonological rules, I'll say that the following questions ask to describe the rules for long vowel sounds, aspiration, and nasalization. So maybe the syllable question is just a way of priming us to think about it when we describe the other phenomena. And I'm sure the whole thing is more of a "given these examples, what kinds of linguistic patterns can you detect?" rather than "tell us the real story about Alabaman" (hence the altering of the original data).

Well thank you both again for your (very quick) answers! I will look forward to continuing to use linguistforum!
Title: Re: IPA k vs kk
Post by: Daniel on July 23, 2018, 12:45:49 AM
Often test/homework questions will alter data for convenience, even if just by selecting 'relevant' (=easy or exceptionless, etc.) data. And completely constructed data sets are also sometimes used, even with good reason: to make it impossible to look up the answer. The only advice that fits is to suggest trying to figure out WHY the question is being asked, and then answer it from that perspective.