Linguist Forum

Specializations => Phonetics and Phonology => Topic started by: freknu on January 09, 2014, 03:41:15 AM

Title: Norse: Phonemic or not?
Post by: freknu on January 09, 2014, 03:41:15 AM
Consider the voiced coronal phonemes /n ɽ r ʋ j/ which are represented in writing as {n l r v j}. Immediately next to voiceless consonants /n ɽ r/ are devoiced to [n̥ ɽ̊ r̥], and immediately after voiceless consonants /ʋ j/ are devoiced to [ʋ̥ j̊].

Here the voiceless coronals are allophones of their voiced counterparts near voiceless consonants — /ʋ j/ occur only after a consonant, while /n ɽ r/ can occur in any environment — and thus non-phonemic.

Initially the situation is somewhat different.

The initial biliterals {hn hl hr hv hj} represent the superficial [[n̥ ɽ̊ r̥ ʋ̥ j̊]] (I'm unsure what transcription to use, above phonetic but below phonemic) which has a free variation of [C̥~kC~çC~C]; as well as the irregular variants [ʃ(ʳ)~ɕ(ʳ)] for {hj}, [ɕɽ] for {hl}, and [ɕr] for {hr}. I'm unsure whether {hn hv} have any irregular variants.

If these biliterals are realised as voiced [C] then they remain non-phonemic, however, cross-dialectally/-idiolectally the biliterals would have to be analysed as voiceless phonemes in order to allow for the "contrastive" voiceless variants.

Should these sounds be analysed strictly as non-phonemic variants that may be pronounced contrastively, or would it make more sense to analyse them as phonemic variants that may merge with their voiced counterparts and thus losing their phonemic status.

To summarise (not including irregular variants):

n [n̥ n] ← normally; devoiced [n̥] next to voiceless

l [ɽ̊ ɽ] ← normally; devoiced [ɽ̊] next to voiceless

r [r̥ r] ← normally; devoiced [r̥] next to voiceless

v [ʋ̥ ʋ] ← after consonant; devoiced [ʋ̥] after voiceless

j [j̊ j] ← after consonant; devoiced [j̊] after voiceless

NOTE: What type is this site using? Arial? It seem kind of incompetent at properly rendering IPA :/
Title: Re: Norse: Phonemic or not?
Post by: MalFet on January 09, 2014, 08:42:06 AM
Interesting. From the outside, at least, it looks like your coronals are simply unspecified for voice and then acquire it from environment. I always enjoy looking at your datasets, but I tend to be at a loss for good answers since I don't know much about Germanic. I'm glad lx is back around, as he's probably got some insights.

Generally speaking, though, it's not a problem for an alternation to acquire contrast in one dialect but remain free in another. In fact, along with direct loan vocabularies, that's the major mechanism of phonogenesis.
Title: Re: Norse: Phonemic or not?
Post by: freknu on January 09, 2014, 07:18:10 PM
Quote from: MalFet on January 09, 2014, 08:42:06 AM
Interesting. From the outside, at least, it looks like your coronals are simply unspecified for voice and then acquire it from environment.

That's certainly a novel and interesting idea that I haven't thought about.

Quote from: MalFet on January 09, 2014, 08:42:06 AM
I always enjoy looking at your datasets, but I tend to be at a loss for good answers since I don't know much about Germanic. I'm glad lx is back around, as he's probably got some insights.

It's good to have him back! :)

Quote from: MalFet on January 09, 2014, 08:42:06 AM
Generally speaking, though, it's not a problem for an alternation to acquire contrast in one dialect but remain free in another. In fact, along with direct loan vocabularies, that's the major mechanism of phonogenesis.

Actually, the voiceless initials stem from Old Norse /hC~C̥/ from Proto-Germanic */hC he/.

*hnakkô m. → hnakki m. → hnakka m. "neck, nape"
*hliþą n. → hlið n. → hlið n. "door, gate"
*hrabnaz m. → hrafn m. → hrafn m. "raven, corvine bird"
*hernô m. → hjarni m. → hjernu f. (*hjerna m.) "brain" (reanalysed as feminine; cf. Sv. hjärna with feminine -a suffix, OSv. hiærne with masculine -e suffix)
*hwalaz m. → hvalr m. → hval m. "whale"

Since I don't have the means to collect the required data, I cannot say how many (proportionally) use the voiceless initials. However, based on my intuition and experience alone, I would say it's going to be fairly high.

The dialects where the [kC̥~çC̥] variants are used the initials are undoubtedly phonemic — contrasting with [C]; dialects where initial voicing contrast has been (partially?) lost there wouldn't seem to be any perceived contrast (non-phonemic) but do still to some degree (varies from area to area and person to person) of variation where the phonemically voiced initial can still be realised with either ambiguous voicing or as outright voiceless, but importantly, without contrast.

So it's probably a bit of both. The voiceless initials are still partially there (cross-dialectally), and influence from nearby dialects still allows free variation where voicing has been lost, which in turn might lead to (or has lead to) them (perhaps) reacquiring (partial) contrast.

Quote from: MalFet on January 09, 2014, 08:42:06 AM
In fact, along with direct loan vocabularies, that's the major mechanism of phonogenesis.

So reacquiring a (perhaps partial) contrast due to influences from nearby closely related dialects is not such a crazy idea after all?

Words do sometimes seem to flow fairly unhindered between the dialects, so even though there is strictly no contrast in the local dialect, continued exposure from the nearby dialects still maintains some "wont" or "accustomisation" to the contrast.

Even though I call them dialects, they are very close to each other.
Title: Re: Norse: Phonemic or not?
Post by: MalFet on January 09, 2014, 08:50:00 PM
Quote from: freknu on January 09, 2014, 07:18:10 PM
Quote from: MalFet on January 09, 2014, 08:42:06 AM
In fact, along with direct loan vocabularies, that's the major mechanism of phonogenesis.

So reacquiring a (perhaps partial) contrast due to influences from nearby closely related dialects is not such a crazy idea after all?

Words do sometimes seem to flow fairly unhindered between the dialects, so even though there is strictly no contrast in the local dialect, continued exposure from the nearby dialects still maintains some "wont" or "accustomisation" to the contrast.

Even though I call them dialects, they are very close to each other.

Not crazy at all! I think this kind of fieldwork should be required for any graduate degree. Of course we all know intellectually that languages are always changing, but it really takes wrestling with this kind of data to understand practically what that means.