In general, the speech of our speakers H and F both showed a tonal pattern consistent with the description of Saeed (1993). However, the following several cases might be worth mentioning as dialect specific tonal characters that goes against the description in Saeed (1993). These are declension2b, declension 5 and declension 6.
Declension 2
Saeed states that pre-pause lowering does not take place in nouns with all LOW tone as in the case of the singular Declension 2b nouns. They include words like abti 'maternal uncle', durbaan 'drum', gorgor 'vulture', huudhi 'canoe', markhaati 'markhaatiyo', sambab 'lung' and suldaan 'sultan'. Namely, in these nouns the last LOW syllable should not lower before a pause.
From a theoretical point of view, this is a bit puzzling. If pre-pause lowering is a phonetic effect, such an exceptional behavior of a specific class of nouns would be unexplained. Then, we might be able to think whether the noun classes described as LOW throughout the word is actually a sequence of LOW followed by a MID or HIGH.
Interestingly, the speech of our speaker H showed a completely different pattern from such speculations. His speech clearly put a pitch accent on one of the syllables in the singular nouns, and the pattern can be classified into two categories. Namely, some nouns, he not only does not lower the tone at a pre-pausal position, but, moreover, place a succinctly high pitch on the last syllable of the word.
ex)abtí, durbáan, markhaatí, suldáan
ex) górgor, húudhi, sámbab
To generalize, we could say that long vowels get a H tone either in a falling or rising contour. And then, nouns ending in -i gets a H tone on 'i'. The H-tone accentuation on 'huudi' and 'markhaati' arise as problematic since they have two potential docking points for a H tone, yet show a different pitch accent pattern; in 'huudi' the pitch accent is on the long vowel, while in 'markhaati@ ', it is on 'i'.
It might be that in 'markhaati', there is sort of a rhythm rule that moves the accent of the second syllable 'khaa' to the initial syllable 'mar' to avoid stress clash. Then, followed by End rule-Right, the pitch accent would end up at the right position.
In any case, it seems to be obvious that both of our speakers' dialect do not follow the pattern described in Saeed with regard to the tonal pattern of the declension 2b.
Declension 5
According to Saeed (1993), Declension 5 is characterized by a final HL on the last two syllables in singular and final H in plural without any segmental change.
ex) Declension
5 in Northern Dialect
awr (sg) 'male camel'
vs awr (pl)
'male camels'
balli (sg) 'pond'
vs balli (pl)
'ponds'
Carab (sg) 'an Arab'
vs Carab (pl) 'Arabs'
However, In our speaker F's dialect, plural formation is done not by means of a simple tonal shift, but it also involves segmental addition of the suffix 'yo'. In other words, most of the examples Saeed list as belonging to the Declension 5 in Northern dialect of Somali are in fact members of other declensions such as Declension 2b or Declension 4 in their pattern of plural formation.
ex)
Plural formation in Central-Southern Dialect
awr (sg) 'male came'
vs awrar (pl)
'male camels'
balli (sg) 'pond'
vs balliyo (pl) 'ponds'
ey (sg) 'dog'
vs eyo (pl)
'dogs'
Nevertheless, our speaker F's dialect still do have a class of nouns that can be categorized as the Declension 5 according to the definition of Saeed. They are nouns indicating nationals such as the following:
ex)
Plural formation of nouns of Nationals in Central-Southern Dialect
Carab 'an Arab'
vs Carab
'Arabs'
Soomaali 'a Somali'
vs Soomaali
'Somalis'
Xabashi 'an Ethiopian'
vs Xabashi
'Ethiopians'
Click
here to hear the recording of the data for this declension class nouns
in our speaker's dialect.
Declension 6
This declension class is described to be final HL both in singular and plural according to Saeed (1993). Thus, the high tone in the singular as in hóoyo is shifted to the next syllable in plurals as in hooyóoyin to maintain the final HL pattern, as the whole length of the word gets longer in plurals due to suffixation of /-oyin/ in this declension.
However, in the dialect of our speakers, the plurals are realized with no succinct pitch accent. Speaker H's speech rather shows LOW throughout the plural form of this declension. Speaker F's speech varies: sometimes she shows the accentual pattern as described in Saeed, but other times she shows LOW as with the speech of the speaker H.
