Metathesis-OT

[sieledorothy]

Hey Phoneticians! I am studying Metathesis among Kipsigis speakers of English (Kipsigis is a language spoken in East Africa) within OT framework. My supervisor has asked me why study that language (Kipsigis speakers of English) in particular. Truth is, metathesis is evident in almost all languages. And I picked the aforementoned Kipsigis with no linguistic reason to it. Kipsigis exhibits metathesis just like any other language would. Now how do I explain this. Thanks in advance.

[sieledorothy]

djr33, any help?

[panini]

The simplest explanation is that Kipsigis does not have metathesis. If you think it does, you can present the evidence and we can evaluate it. Furthermore, English also does not have metathesis, so it isn't the case that "metathesis is evident in almost all languages". Icelandic and Leti do, so you could pick another language.

In standard Nigerian English, the verb "ask" is pronounced [ɑks], which seems to involve historical metathesis. However, this is found only on that one word, and not e.g. in "mask, task, bask...". It is possible that there is some such example in the speech of Kipsigis speakers, though I would be surprised if there were metathesis that is specific to just Kipsigis or Kalenjin speakers.

The problem with looking at patterns of English to Target Language loanword adaptation is that it you don't know what the input to the speaker is. Not only is there a difference between North American and British English (and many dialects of English in the UK), but there are also now Kenyan versions of English, so that the input may well be Kenyan English, which is not at all well studied. An example is that Kenyan English for chui is [leopɑd], where the diphthong is not derived from any speech input -- except, now, the fact that that is how it is pronounced in Kenyan English.

Ideally you would address this via a controlled experiment, where speakers of some language are presented with an audio sample of an unknown but phonotatically possible word of English, and they are supposed to pronounce it back.

[Copernicus]

Can you provide some examples of the data you are looking at? 

Common English examples abound, but they are usually lexicalized conventions rather than regular phonological processes--"purty"/"pretty", "comfterable"/"comfortable", "aks"/"ask", "perscription"/"prescription", "integral"/"intergal", "nucular"/"nuclear", "asterisk"/"asteriks", "calvary"/"cavalry", "introduce"/"interduce", etc.  Of course, "ask" and "wasp" were originally "aks" and "waps", so there can even be modern dialects where the metathesis was never lexicalized. 

[freknu]

Of course, "ask" and "wasp" were originally "aks" and "waps", so there can even be modern dialects where the metathesis was never lexicalized.

"Wasp" does indeed come from a PG *wabsō (PIE. *wob^hseh₂), but "ask" comes from *aiskōnam (PIE. *ais-); so "original" metathesis is true only for "wasp", whereas for "ask" it only becomes true later when it came to shift between ask/aks.

[Daniel]

You have no data. You can't argue anything. (If you get data, you can, but you can't try to figure out what you're arguing before you have the data.)

Also, metathesis is a relatively sporadic ('irregular' across words) kind of sound change. There are very few instances (vowel+R clusters in a few languages maybe English?) where this occurs, but this is a historical change, not something you'd be modelling with OT (at least not in its most common functions). And if it's sporadic, it wouldn't fit OT well anyway.

Panini, as Copernicus said, aks is more common in Old English, I believe, and it has switched back and forth a lot. Whether constant instability (and free variation?) is metathesis or not depends on your technical definition, but it's an interesting case. Doesn't seem specifically relevant to this context though.

English is not a relevant test case because it doesn't have any sort of active phonological metathesis either.

So, you can decide between these options:
1. Find some great data.
2. Pick a new language.
3. Pick a new question.

Something somewhat relevant might be epenthesis. It's a little more regular than metathesis (though still can be sporadic). For example:
pumpkin and hampster
Which one actually has a P in it?
Or scents vs sense -- they sound the same.

You could try to look at some of those effects. But having a VERY specific question like this with no data is a good way to get null results.

The best thing to do is observe an interesting puzzle in the language and try to solve it, if possible.


Edit: just saw your reply. Aks/ask is complicated

[sieledorothy]

Panini
Thanks for the insight. Perhaps I should rephrase my research questions and have one investigating whether metathesis does actually exist in the said language (the English spoken by the Kipsigis). Then, adopt the controlled experiment, presenting respondents with specific inputs where they read aloud. Another question my supervisor asked was how will you know which words are likely to be metathesized.