Archive for the ‘Phonetics’ Category

Shona

Sunday, August 29th, 2010
Flag of Zimbabwe

Shona, as you probably know, is a Bantu language spoken mainly in Zimbabwe, whose flag you can see to the left. There are also speakers in Zambia and other neighbouring states. Shona is famous (or should that be notorious?) for having so-called “whistle” fricatives. The voiceless one is written <sv> in the standard orthography of the language. Its voiced counterpart is written <zv>. So far, so good, though I should add that there are couple of corresponding affricates, written <tsv> and <dzv> respectively. Also <v> represents [w] rather than [v]. Shona has [v], but this is written <ṿ>. Still with me?

Both the Omniglot page and the Wikipedia entry for the language recognise that they are a bit out of their depth. One has to sympathise. When I was a postgraduate student at UCL in the 1970s I and all my fellow students failed to convince our fellow student, Alex Pongweni, a native Shona speaker, with our efforts to imitate his production of these dratted sounds.

Published descriptions all agree that these sounds involve multiple articulations, but they do not come to a consensus on what exactly is going on. One disagreement is as to whether they are double articulations, with friction noise being generated at two places simultaneously, or whether they are a primary articulation at the alveolar ridge with some sort of labial secondary articulation. The first of these ideas would lead to symbols such as [ɸ͡s] and [β͡z]. However, * and [zʷ], which is what the second idea implies, certainly would not have done for the sounds Alex Pongweni tried to get us to produce all that time ago.

* there should be s with superscript w here, but for some completely weird reason that combination refuses to display in my browser!

Underlining again

Sunday, August 22nd, 2010

A few more uses of underlining in orthographies. They are all from North American languages.

The first, a rather puzzling one, is Tuscarora, spoken in Southern Ontario in Canada, and in New York State and North Carolina in the USA. The orthography used for this language contains <eh>. This represents the vowel [ə̃]. As the letter <h> is used to represent [h] and <e> to represent [ɛ], it’s a bit difficult to see the logic behind this.

Vancouver Island
Vancouver Island

The use of underlining in Kwakiutl orthography, on the other hand, is rather neat.  Kwakiutl, also known as Kwak’wala, is spoken in the northern part of Vancouver Island and also in adjacent areas on the mainland.  It has <a> = [ə], <k> = [q], <kw> = [qʷ], <k‘> = [qʼ], <k‘w> = [qʷʼ], <x> = [χ], and <xw> = [χʷ]. It seems likely that whoever devised the orthography knew a bit about phonetics and perhaps had the IPA retraction diacritic in mind in including the underlined symbols.

Another nice thing about this language is what the speakers of it call themselves: Kwakwaka’wakw [kʷakʷakaʔwakʷ], which means “speakers of Kwakiutl”.

The final two languages use underlining to indicate nasalisation of vowels. Chickasaw, spoken mainly in Oklahoma, has <a> = [ɑ̃, ə̃], <i> = [ĩ, ɪ̃], <o> = [ə̃ʊ̃], and <u> = [ũ, ʊ̃]. Choctaw, spoken in central Mississippi, Louisiana and Tennesee, is practically the same.

Devoicing

Sunday, August 15th, 2010

In a comment on my post about Scilly place-names Glorios from South Korea asked some questions about English pronunciation. I promised to deal with them in another post. So here we go…

Glorios wrote:

my own finding is that for me when some voiced sounds come at the end of words or syllables, the sound of voiced sounds seem to get closer to the sound of its voiceless partners.

can you explain why and let me know the way to distinguish them.

Glorios is quite right, of course, but we can get a bit more specific than that. There is a process in English called devoicing. Some voiced sounds lose some or all of their vocal fold vibration in certain environments.

Question (1) Which sounds are affected? There are two different groups, which behave differently from one another. Group 1: Obstruent consonants /b d ɡ v ð z ʒ dʒ/ Group 2: approximant consonants /w r l j/.

Question (2) When do they lose their vocal fold vibration? For the Group 1 consonants to retain full voicing they must be surrounded by voiced sounds. If one of the Group 1 consonants is preceded by a voiceless sound or silence, or followed by a voiceless sound or silence, then some or all of its vocal fold vibration will be lost. Here are some examples of devoicing of Group 1 consonants with a phonemic and a phonetic transcription:

  1. this boy /ðɪs bɔɪ/ [ð̥ɪs b̥ɔɪ]
  2. my dog /maɪ dɒɡ/ [maɪ dɒɡ̊]
  3. misbehave /mɪsbɪheɪv/ [mɪsb̥ɪheɪv̥]
  4. She’s so good /ʃiːz səʊ ɡʊd/ [ʃiːz̥ səʊ ɡʊd̥]

From the above we can learn a number of things:

  • Devoicing is indicated in a phonetic transcription by a small circle below the relevant symbol, or in the case of symbols with descenders, above it.
  • Devoicing can take place at the beginnings of words, at the ends of words, and in the middle of words.
  • Devoicing of Group 1 consonants is very frequent in English.

I’m going to leave the rest of the story for another day. Watch this space!

Underlining

Friday, August 13th, 2010

On 8 March 2008 I wrote a short blog on Kāi Tahu Māori. You can see the complete post on this page. In that post I made two claims which I repeat here for convenience:

The letter k used to write KTM is unique in two respects. (1) it is the only orthographic symbol which incorporates underlining. (2) it is used to reference two different dialects of a language — in effect it means a [k] sound, which in other types of Māori is an [ŋ] sound.

(2) above I still believe to be true, but (1) definitely is not.

Having been trawling through Omniglot, I have come across a handful of languages which use a Latin-based script and which use underlining for various purposes.

For instance, a few Native American languages which indicate voiceless vowels this way. An example is Comanche which has voiceless versions of all its six vowels <a e i o u ʉ> (the last represents [ə]). So <a> = [ḁ], <e> = [e̥] etc.

More on this another time.

Pulmonic ingressive

Tuesday, August 3rd, 2010

Just in case you are not familiar with the terminology, pulmonic ingressive sounds are made with air drawn into the lungs rather than the usual outward direction of pulmonic egressive sounds. When I was teaching phonetics and dealing with the matter of airstream mechanisms I regularly skimmed over the matter of pulmonic ingressive sounds. This was just to keep things simple. Pulmonic ingressive sounds are quite interesting, however.

Usually published accounts of airstream mechanisms do what I used to do and say airily that they are only used for paralinguistic purposes like expressing pain. There is more to the matter than that.

First, there is the matter of Damin, a ritual language spoken by members of two Australian peoples, the Lardil and the Yangkaal, who live on islands in the Gulf of Carpentaria. It seems likely that the language was invented by the Lardil, though they maintain that it was created by a mythological figure in the Dreamtime. The phoneme inventory of Damin is very similar to that of Lardil, except that it contains six extra consonant segments:

  • Four nasalised clicks: ŋʘ ŋ| ŋǃ ŋǂ (Damin is the only language outside Southern Africa to have clicks)
  • A reverse click kʘ↑
  • A pulmonic ingressive voiceless lateral fricative ɬ↓

Next we have ǃXóõ, a language of Botswana, which has a stupendous array of clicks, including aspirated clicks with a voiceless ingressive nasal accompaniment at five places of articulation – ↓ŋ̊ʘʰ and the like. You can hear sound files for these and all the clicks at the UCLA Archive.

Then there is the mysterious case of Tsou, a Taiwanese language. Sometime in the 1990s a study of the speech of one speaker of this language was published in the Journal of the International Phonetic Association. It was reported that word-initial voiceless fricatives were produced with a pulmonic ingressive airstream mechanism. A subsequent study with other speakers failed to corroborate this finding.

I think the use of pulmonic ingressive speech as a paralinguistic feature is rather under-researched. However, there is some interesting information about this on Robert Eklund’s site.