LINGUIST List 13.1521

Fri May 24 2002

Sum: Tones in Latvian

Editor for this issue: Marie Klopfenstein <marielinguistlist.org>

Directory

1. Jardar Eggesb� Abrahamsen, tones in Latvian

On the 13th of May I submitted this request to Linguist:

 Does anyone know an autosegmental analysis of the Latvian "stiepta
 intonacija" (words like bra:lis) and "kritos'a intonacija" (words
 like sva:rki)?

 The first group is reported to have a H tone on the first syllable and
 a L tone on the second, whereas the second group has a fall (HL)
 already on the first syllable.

(In this quote, as in the following quotes, I have rendered all 8 bit
characters into 7 bit characters, as 8 bit characters seem to have a
problem surviving the editing of the Linguist List.)

Noone replied, but thanks to AltaVista and the search criterion "Latvian
phonology" I got in contact with Dr. Aleksandra Steinbergs, associate
professor at the Memorial University of Newfoundland.

>From my e-mail to her:

 Then I remembered some Latvian pitch contours that I had seen in a book
 from the thirties (Ekblom, R. 1933. _Die lettischen Akzentarten. Eine
 experimentalphonetische Untersuchung._ Uppsala: Almqvist & Wiksells
 boktryckeri). As I understand these contours, the "stiepta intonacija"
 (bra:lis) has a H tone on the first syllable, and a L tone on the
 second. The "kritos'a intonacija" (sva:rki) har a fall HL already on
 the first syllable.

 Of course, in the good, old phonetic tradition of the thirties the
 contours are not analysed phonologically. What I would like to know is
 whether there exist any modern (i.e. autosegmental) analysis of the
 Latvian word-tones. Without knowing too much about Latvian phonology I
 can see two possible analyses:

 1) Both patterns display a prominence tone H on the first mora of
 the first syllable and a boundary tone L to the right. In words like
 (or accent phrases beginning in words like) "br�lis" there is in
 addition a lexical H on the second mora.

 2) In words (accent phrases) like "bra:lis" there is a H on the first
 two moras (or on the first syllable), in words like "sva:rki" there
 is a H only on the first mora.

>From Dr. Steinbergs' reply:

 Latvian has several tonal dialects. The "standard" tonal dialect has
 three tones: the Level tone "stiepta intonacija", the Falling tone
 "kritos'a intonacija" , and the Broken tone "lauzta intonacija".
 ...
 The usual descriptions of this system describe (a) the Level tone as
 being high throughout (or high with a slight gradual decrease in pitch
 throughout the syllable). The (b) Falling tone is normally described as
 having an extremely brief rise followed by a long fall. The (c) Broken
 tone is normally described as being a rising tone followed by a falling
 tone with either a glottal interruption in the middle or some degree of
 creaky voice during the production of the syllable.

 To illustrate the contrast, I've provided three words in this dialect.
 In the orthography are all three spelled "loks", but they contain a
 centering diphthong something like [ua]. The contrasts are: [luaks]
 "green onion" (with Level tone), [luaks] "arch, bow" (with Falling
 tone), and [luaks] "window" (with Broken tone).

 If I had to do an autosegmental analysis of this three-tone dialect, I
 would suggest that (1) "green onion" (Level tone) has a H on the first
 mora, "bow, arch" (Falling tone) has a H on the first mora and a L on
 the second mora, and that "window" (Broken tone) has L on the first
 mora but both a H and L on the second mora. Another analysis (2) would
 be same as the above for "green onion" and "bow, arch", but "window"
 would have a L on the first mora, a L on the second mora (and the
 double Low calls forth the creaky voice) , and there would be some sort
 of (dissimilatory?) restructuring whereby a H pops up between them.

Jardar