Nyelvtudományi Közlemények 103. kötet (2006)

Tanulmányok - Tamás Ildikó: The Lule Saami vocalism 7

+jádáv, +jadá, *jahtá (VxSgl-3), but in fact, the correct forms are: jädäv, jädä, jähtä (VxSgl-3).19 As we can see, the

ä in the second syllable appears with roundness and height adjustment as an alternant of the á, but never in the place of the original e root vowel: jähtebihtte, jähteba (VxDu2-3), jähtep, jähtebihtit (VxPll-2). In the nominal declension, the expected forms do not occur if there is a marginal e-á alternation in the background, so in the case of manne 'egg' the regular *mannáj (Sglll) is replaced by männäj. On the other hand, a-harmony does not apply if, before a Cx or Vx, the original e marginal vowel is kept: manne (SgNom), mäne (SgGen), mänev (SgAcc), etc. The domain of a-harmony, in this case, is the short foot, therefore (although there is progressive spreading) instead of suffix harmony we only talk about root harmony.20 The end of the short

foot (as seen before in several types of alternation) serves as a boundary in this case too: in longer root alternants and in suffixed roots (where the number of syllables rises to three) the a of the final syllable does not participate in the harmony: tjättä 'throat' (SgNom) - tjäddäga 19 The question might arise of whether the appearance of á in the second syllable has taken place in order to avoid homonymy. First, we must say that the paradigm containing the secondsyllable á is irregular. We might see this easier with the help of some Northern Saami examples. In Northern Saami bisyllabic verbs belonging to the (Nielsen spelling) -e or (contemporary spelling) -i root class have an irregular paradigm in the singular (in the second syllable, an á appears instead of an e.) According to Korhonen, the appearance of the irregular forms was needed in order to avoid formal identity between past and present forms (M. Korhonen 1981, 270). The present tense singular forms of the verb boahtet are: boadán, boadát, boahtá (instead of the regular -kboadin, •kboadit, *boahti), and in the past tense they are: bohten, bohtet, bodii. We can see that there is no homonymy because the strong and weak roots are used in a complementary manner in the present and past tense paradigms; furthermore, the i past tense marker was dropped intervocalically already in Proto-Saami (Sammallahti 1998: 79). It would be strange if a language, in order to avoid homonymy, changed both present and past paradigms, when only one change would be sufficient. In verbs with three syllables, having an -i root, the present and past forms would truly overlap, therefore in the present tense, before personal suffixes, there is an a instead of an i (cf. muitalit (inf.), Sgl-3 praes. muitalan, muitalat, muitala and Sgl-3 praet. muitalin, muitalit, muitalii.) In this case the two paradigms would actually overlap, so the i alternating with a in the present singular form has become the marker of present tense. In my opinion, this analogy may have spread over to bisyllabic verbs with an i root. Etymologically, Lule Saami, two-syllable, eroot verbs correspond to Northern Saami г-root verbs (new spelling). The marker of the past tense is also i, so the appearance of the ä in the second syllable is not an attempt to avoid homonymy, but to comply with ii-harmony. 20 Albeit labial harmony „means that a given suffix has two allomorphs, according to the labial/illabial opposition," and so „primarily it is typical of complexes of free and bound morphemes (suffix harmony)" (Rédei 1976: 421), in Lule Saami, this harmony type works within the suffixless root, too.