Finite state and Constraint Grammar based Text-to-Speech processing
View the project on GitHub giellalt/speech-sme
This is a document about word stress in North Sámi
Our end product will be expected to “read” everything from whole paragraphs to words in isolation, so we need to assign stress to all polysyllabic words. Here are the main rules.
This means that in trisyllabic stems, the third syllable will be stressless, when being the ultimate syllable, and stressed when being the penultimate.
" = main stress
'= (weak) secondary stress
( )= stress group
("mui.ta.lus)
("mui.ta.)('lu.sa)
("mui.ta.)('lu.saš)
("mui.ta.)('lu.sa.ža)
("mui.ta.)('lu.sa.)('žiid.da)
("mui.ta.)('lu.sa.)('žiid.da.met)
There are some exceptions to this assignment, the most important ones being compounds and words with consecutive stress.
Word boundaries are also stress group boundaries in compounds. Main stress falls on the first syllable. Secondary stress falls on the first syllable of the second (and third) element of the compound:
"mán.na#'bii.gá
"ov.da#'skuv.la
"boah.ku.han#'plá.na
"seas.tin#'báŋ.ko#'vahk.ku
In compounds with two stress groups, the only possible pattern is main stress followed by weak secondary stress. In compounds with three or more stressgroups, there are two possible patterns, main stress can be followed by two weak secondary stresses (pattern 3 1 1), or followed by strong secondary stress, which is followed by weak secondary stress (pattern 3 2 1):
("gula)+('hallan)#('váhnen): 3 1 1
("váhnen)#('gula)+('hallan): 3 2 1
3 1 1 is the default pattern. 3 2 1 is what you get if the first element of the compound is a disyllabic word followed by two stress groups. These patterns are the same for compounds of 3 disyllabics. If the first two elements make a constituent to which the final element is added, we get the 3 1 1 pattern. If the final two elements make a constituent to which the first element is added, we get the 3 2 1 pattern:
goike+biergo#haddi: 3 1 1
goike#čielge+suotna: 3 2 1
In compounds of three words, the middle element is sometimes shortened. When it is, it does not bear stress, but constitutes an unstressed syllable in a trisyllabic stress group:
("sámegiel)('girji)
("sotnabeai)('skuvla)
Disyllabic suffixes constitute disyllabic feet, such as
-goahtit
and -vuohta
(also haga, according to Sammallahti 1977, but it is usually written separately).
("vázzi)('goahtit)
("boradiš)('goahtit)
("ráhkis)('vuohta)
("oskkáldas)('vuohta)
There are some words with consecutive stress, on the first and second syllables. These are words in which the first stress group consists of only one syllable. The first syllable has main stress and the second has secondary stress:
("mu.)('sihk.ka)
("sma.)('đáh.kis)
("oam.)('beal.li)
("oam.)('be.le.)('ža.gat)
("má.)('il.bmi)
These might just have to be listed. Otherwise, many of them can be identified on the basis of the vowels and consonants of the second and third syllables. In a word with ordinary stress patterning, the second vowel position is the latus, and diphthongs cannot occur there. If a diphthong is in second position, the word has two consecutive stresses. The next consonant position after latus would be the consonant margin, and only some consonants can occur there. Two possible rules:
I. If there is a diphthong in the second syllable, the first syllable constitutes a stress group of its own and bears the main stress, while the second syllable bears secondary stress and begins a new stress group.
II. If there are consonants between the second and third syllable, which cannot occur in the consonant margin, such as /lbm/, then the first syllable is a stress group of its own, and the second syllable bears secondary stress and starts a new stress group.
The importance of this goes well beyond the stress assignment itself. It will also affect the correct identification of the consonant centre, for example in the word (“vil.)(‘bealli), which is treated by our current phonological rules as if the ‘lb’ is the centre, and a schwa is inserted between the consonants: vilᵊpealli.
Some monosyllabic suffixes and enclitic particles are not included in the preceding stress group, even though they are written together with the word:
beatnagago?: ("beat.na.ga) go
Márehiiges: ("má.re.hii) kes
bođiigoson Máret: ("bo.đii)ko son ("Má.ret)
hálddahuslaš: ("háld.da.hus) laš
(In inflected form, it makes a group of its own: ("háld.da.hus)('lač.čat)
)
mánáidisguin: ("má.náj.dis.)guin
beatnagiiguin: ("beat.na.gii.)guin
Some words have main stress on the second syllable. These must be listed.
vih.("guht.ta)
ieš.("guh.te.nai)
ieš.("guh.te.ge) Compare to ("ieš.)('lá.gan)
(o.ba)("nas.sii.ge)
vuo."hon