South Estonian written standard and actual spoken language: variation of the past participle markers/[TEXT NOT REPRODUCIBLE IN ASCII].
From the 16th to the early 20th century there were two literary languages in Estonia called the Tartu literary language (in South Estonia) and the Tallinn literary language (in North Estonia). During the second half of the 19th century the South Estonian literary language started a gradual decline because of the socio-political changes in the society, and the North Estonian literary language gained prominence. At the same time in Southeast Estonia, the old local variety, Voru, partly retained its status as a local spoken language.
Even nowadays, the sharpest linguistic differences are found between North and South Estonian. Furthermore, unlike in other areas of Estonia the old local dialects of South Estonian are not levelling to Standard Estonian but to the Common South Estonian language where the most typical South Estonian markers are used. (1) Since the end of the 1980s a standard Voru language is being developed. It is mostly based on the eastern sub-dialects, i.e. Vastseliina and Polva.
Due to many discussions about the written standard of Voru the spoken variant of the language has become quite a popular research topic among Estonian linguists. Previous studies dealing with the morpho-phonological issues of Voru have focussed on the spoken materials mostly collected during the 1990s (e.g. see Iva 2002a; 2002b : 84-92; Mets 2000; 2004a; 2004b : 657-669, 719). All these researches have revealed that there exists significant variability within the spoken language.
The main purpose of this article is to find out whether the older variant of spoken Voru is more rule-based, and thereby more similar to the Voru standard, or coincides with the spoken language usage of the 1990s. For that the variation of the past participle markers is analysed with regard to number agreement and word structure (the number of syllables and the word quantity degree). The topic is intriguing because various markers are used in different sources. For example, on the maps of Estonian dialects (see Saareste 1938 : map 27; 1941 : map 48; 1955 : 34-35, 39, 45, 54, 64; Toomse 1998 : 44, 46, 94, 109, 131) one can see different formatives for marking the past participles. These maps are based on interviews rather than on recorded speech. Yet, later recordings made in the 1990s show that there are even more markers for the past participles (for example, see Mets 2000; 2004a; 2004b : 660-662; Iva 2002a : 96-97). Notwithstanding, the standard of Voru allows only one marker per form. -nU[??] (2) is the standardised form for the active past participle, though according to the history of language the active past participle was marked by two markers: -nU[??] denoted the singular and -nU[??] the plural. (3) Today this old system does not exist any more. There are two markers for the passive past participle: -t/-D (4) marks the singular and -tU[??]/-DU[??] the plural (VL 1996 : 14; VES 2002 : 581). One of the hypotheses is that the older variant of spoken Voru could include more historical markers (i.e. -nU[??]) than the spoken Voru of the 1990s. Another hypothesis is that the old system, according to which the singular and plural forms are marked by different markers, is not valid either in the recordings made between the 1960s and 1980s or in the recordings of the 1990s. The main focus of the paper lies on Voru; Setu is used as an additional material for comparison. (5) The analysis is based on three corpuses: recordings made between the 1960s and 1980s, (6) recordings made in Vastseliina during the 1990s and recordings made in Polva in 1999.
2. Materials and method
In March, 2007 the corpus of the Estonian dialects (7) contains five sub-dialects (8) of Voru: Hargla, Urvaste, Polva, Rapina and Vastseliina, and two sub-dialects of Setu: North Setu and West Setu. Among these sub-dialects the morphologically tagged ones are Hargla (4), (9) Vastseliina (7), Polva (1), Rapina (1), West Setu (1) and North Setu (5). For the present article only the morphologically tagged texts have been taken into consideration; untagged texts have been left out.
In order to find participles in the texts two search engines available in the corpus (search.php and search2.php) are used. One of them (search.php) shows a form in its context; the other (search2.php) shows how many times the form is used. In the corpus the words are tagged so that one can see the word class (e.g. V = verb) and its morphological status (e.g. nud = the active past participle).
The recordings made in Vastseliina and Polva during the 1990s are searched manually through.
The database of the recordings made between the 1960s and 1980s consists of 467 active past participles (377 from Voru and 90 from Setu) and 163 passive past participles (100 from Voru and 63 from Setu). The second database, the recordings in Vastseliina during the 1990s, contains 882 active past participles. Finally, the third database, the recordings in Polva in 1999, consists of 1259 active past participle and 259 passive past participle markers. Some of the markers are borrowings from Standard Estonian, but these have been excluded from the analysis. Likewise, in the database there occur some participles (both active and passive) which are adjectives by their grammatical status. As there are only a few of them presented, the number and case agreement in adjectives is not analysed here.
The first part of the following analysis is based on the recordings made between the 1960s and 1980s. The overview of the markers in all six subdialects is given. The second part of the analysis focuses on three subdialects that are better presented in the corpus: Hargla, Vastseliina and North Setu. Only verb forms are analysed. For the active past participle the main purpose is to determine the usage of -nU and -nU[??], whereby all markers ending with thae glottal stop (-nU[??], -U[??], -nO[??], -O[??]) are counted as one, and all markers without the glottal stop (-nU, -U, -nO, -O) are coded together. The analysis of the passive past participle investigates the variation of -t and -tU([??]), whereby -t and -D are counted as one and -tU, -DU, -tU[??] and -DU[??] are coded together. The intra-sub-dialectal and intersubdialectal differences are analysed. The results will be compared to the data on the dialect maps and also to the spoken Voru of the 1990s.
3.The active past participle
3.1. General analysis
In the Voru written standard the active past participle marker is -nU[??]. In spoken language the choice of markers is much more complicated. Below the overview of the active past participles is given.
-nU (e.g. vetnu '(have) taken') is the most common South Estonian suffix for marking the active past participle. In the corpus, the marker -nU is most often used in almost all the sub-dialects under investigation, except in Rapina where its phonological counterpart -nO (e.g. nanno '(have) seen') is typical. -nO is a later development where [sup.*]U > O in non-initial syllables. The variants with O are also used in Polva and Setu, i.e. in the northeastern parts of Voru.
In Vastseliina and West Setu, the marker -nU[??] (e.g. lannu[??] '(have) gone') is more frequent than the marker -nU. (1)0 The variant which ends with the glottal stop is the older variant of -nU, and Vastseliina is said to be one of the sub-dialects which has preserved many unique traits. Historically -nU[??] was used to mark the singular and -nU stood for the plural. Today this old morphological distinction has disappeared and the occurrence of the glottal stop at the end of the marker has altered into a stable structured phonological variation.
In olema ('to be') and tulema ('to come') subtypes the n in the beginning of the marker is assimilated to l (i.e. olnu > ollu '(have) been') above all in Hargla. A few more examples can be found in Vastseliina, North Setu and Rapina. The recordings of the 1990s show the same assimilation in Polva (Mets 2000 : 23; 2004b : 660). Likewise, the maps of Estonian dialects (Saareste 1955 : 34-35) demonstrate that the n is assimilated to l in all six sub-dialects: North and West Setu, Vastseliina, Hargla, Polva and Rapina.
On A. Saareste's map the most frequent marker of the active past participle is -nU[??] (see Saareste 1938 : map 27; 1941 : map 48; 1955 : 34-35, 39, 45, 54, 64). On the other hand, the present findings show that in almost all South Estonian sub-dialects (except Vastseliina and West Setu) the non-glottal variant -nU is used more often than the glottalised -nU[??] (for the relationship of -nU and -nU[??] see Map 1). This means that already in the 1960s the marker -nU was more frequent than -nU[??]. The same result appeared from the analysis of the data from later periods (see Mets 2000; 2004a; 2004b : 657-669; Iva 2002a : 96-97) where it could be seen that the occurrence of the glottal stop depends on such phonological factors as the word quantity degree and the number of syllables. Map 1 demonstrates the frequency of the occurrence of the markers with the glottal stop (here -nU[??], -U[??], -nO[??] and -O[??] on the one hand and -nU, -U and -nO on the other hand have been coded together).
In the plural, the third person can be marked by the suffix -vA([??]) (e.g. joznuva[??] '(they have) run'). The corpus of Estonian dialects shows that -nUvA([??]) has spread in northern Voru (Polva and Rapina), but is not used in the southern areas. According to the tape recordings of the 1990s it appears that the suffix -nUvA([??]) has spread to the areas of Vastseliina (Mets 2004a : 23; 2004b : 660).
Additionally, the corpus contains some active past participles that are adjectives or nouns by their grammatical status. In Voru, the adjectives when followed by the noun agree in number and case (Keem 1997 : 47), e.g.
(1) ja kinela-ze sis mu k aDu-nu-le lella-le and speak-3PL then my.GEN loose-PRTCPL-ALL paternal.uncle-ALL tu-D asja that-PART thing.PART 'and then they speak it to my dead paternal uncle'
(2) eGa siz i-s ole-k'k'i oste-ttu-i-zi surely then NEG-IMPRF be-CLTC buy-PRTCPL-PL-PART nuppe button.PL.PART 'surely then there were no bought buttons'
Because of their rare occurrence these have not been analysed in the present article.
To sum up, in spoken Voru and Setu there are many more markers for the active past participle than the written norm allows. The standardised marker -nU[??] is mostly used in Setu and Vastseliina. (11) The rest of the subdialects incline towards the non-glottal marker -nU. When compared to the results of the 1990s Vastseliina and Polva the picture is almost the same - the marker -nU is more frequent than -nU[??], even in Vastseliina. The present tendencies show that the spoken language is moving towards a simpler and more common variant which uses the more common markers of South Estonian. For the active past participle this common marker is definitely -nU. It is normal that sounds that do not distinguish any grammatical meaning (12) (e.g. the glottal stop at the end of the marker of the active past participle) are not always pronounced. Likewise, the previous discussion reveals that the active past participle markers found in the corpus of Estonian dialects are more similar to present-day spoken South Estonian than to the older dialect norms.
3.2. The singular and plural in three South Estonian sub-dialects
The written standard of Voru does not make a distinction between the singular and plural forms of the active past participle. Historically, the marker -nU[??] marks the singular and the marker -nU stands for the plural. For the singular the historical development is [sup.*]olnut > olnu[??], and for the plural it is [sup.*]olnu[delta]et > olnu. Earlier studies on Vastseliina and Polva in the 1990s (Mets 2000 : 43-46; 2004a : 106-110) and on the speech of the Voru educated people in 1999 (Iva 2002a : 100) show that this distinction does not exist in the everyday language usage either. Similar result emerges from the analysis of the corpus of Estonian dialects where both markers, -nU[??] and -nU, are used both in the singular and plural (see Table 1).
Still, there are some differences between the sub-dialects. In Hargla, -nU is frequent both in the singular and plural. Vastseliina offers a slightly different picture: the usage of -nU and -nU[??] is similar in the singular and plural. North Setu differs from the rest - the formative -nU is used more in the plural and -nU[??] is used more in the singular. The analysis of the 1990s Vastseliina reveals that the glottalised formative is used less, -nU dominates in both numbers. In the 1990s material the elderly and middle-aged informants had a tendency to mark the plural with -nU[??], but the younger informants used the glottal stop less frequently (Mets 2004a : 109-110). The 1990s material collected from Polva shows the same tendency.
These results show that in the database recorded between the 1960s and 1980s the old differentiation between the singular and the plural has survived more in the most eastern sub-dialects than in the western subdialects. An irregular usage of the singular and plural markers is apparent in Hargla whereas Vastseliina forms a transition area where both markers are used equally in both numbers. In North Setu, the old paradigm of the singular and plural is more regular than in the other two sub-dialects. It can be concluded that the more peripheral the sub-dialect the more it has preserved its older traits. On the other hand, the data collected during the 1990s reveal some changes--in both numbers -nU is used more than -nU[??]. Whereby, it can be suggested that this old distinction between the singular and plural markers has lost its original denotation.
Additionally, it has been claimed that the occurrence of glottal stop depends on word structure (see Juva, Kasak, Help 1995 : 10; Keem, Kasi 2002 : 32; Iva 2002a : 96-97; 2003 : 79-80, 82; 2005 : 124-125, 131; Mets 2004a : 80-83; 2004b : 665-666): glottal stop is most likely to be pronounced at the end of the disyllabic Q2 words. In disyllabic Q3 and longer forms its occurrence declines. The data analysed for the present article demonstrated the same tendency. The only exception is Hargla where disyllabic Q2 forms do not favour the marker with the glottal stop. The overall tendency seems to be that in the western sub-dialects the glottal stop occurs more rarely than in the eastern sub-dialects. Other studies have indicated on it as well (e.g. Nigol 1994 : 73; Keem, Kasi 2002 : 33).
Likewise, dialectologists (Keem, Kasi 2002 : 20, 50) have pointed out that in Voru dialect in the case of disyllabic forms and in some Q1 verbs (e.g. tul ema 'to come', teGema 'to do', naGema 'to see', minema 'to go') the number is marked by the quantity: Q2 stands for the singular (e.g. ta om nannu[??] 'he has seen') and plural forms are marked by Q3 (na omma nannu[??] 'they have seen'). Though, it is especially characteristic of Hargla, but this kind of distinction can be found even elsewhere in South Estonia (e.g. in Vastseliina and Setu). According to that the variation of -nU[??] and -nU is analysed with regard to the singular and plural. The study reveals that both quantities appear in both numbers. Secondly, in both numbers the occurrence of -nU in Q2 and Q3 words is higher than the occurrence of -nU[??]. The analysis of the 1990s data shows the same result.
According to the previous discussion the overall conclusion could be that in the case of disyllabic words the quantity degree does not seem to be so relevant for distinguishing between the singular and plural. Likewise, the occurrence of the markers ending with the glottal stop is more influenced by the word structure than by the number.
The next part of the analysis deals with the passive past participles.
4. Passive past participle
4.1. General analysis The Voru written standard has two markers for the passive past participle: -D/-t for the singular and -DU([??])/-tU([??]) for the plural. In spoken language the choice of markers is much more complicated.
On the basis of the corpus of Estonian dialects the formative -t (unehtet '(have been) forgotten') dominates in Vastseliina and North Setu. On the other hand, the marker -tU (kirotteDu '(have been) written') is more prevailing in Hargla. In West Setu, the usage of -t and -tU is more or less equal. (13) As compared to other sub-dialects, in Vastseliina the marker with the glottal stop, -tU[??], is preferred (kenelDu[??] '(have been) spoken'). The formatives containing an O (tetto '(have been) done') are more typical in North Setu and Rapina; other sub-dialects do not have such markers. In the plural, the third person can be marked with -tU(v)vA (loettuva '(have been) read'), such markers occur in Polva and Rapina, i.e. in the north-eastern sub-dialects. In the corpus there are six performative forms (verb saama ('to get') + -tUs), all of them end with the regular marker -tU s (sai tettus '(it) got done'). These, like the adjectives, have been excluded from the analysis.
According to A. Saareste the only formative for marking the passive past participle in Hargla is the formative -t. The rest of the sub-dialects (Vastseliina, Polva, Rapina, North Setu and West Setu) have -t, -tU and -DU[??] as possible variants (Saareste 1955 : 69). Unlike A. Saareste, M. Toomse shows that in Hargla the formative -tU is also possible. According to his maps there are two possible markers (-t or -tU) for the passive past participle in Vastseliina, Polva and Rapina. In the areas of Setu the only formative is -t (Toomse 1998 : 46, 94, 109, 131). Additionally, on the recordings of the 1990s there are three different formative variants used in spoken Voru: -t, -tU and -tU[??], the most common one being -tU (Mets 2000 : 81-87; Iva 2002a : 97-99, 101-103, 121). The occurrence frequencies of marker -t vs. -tU([??]) are shown on Map 2 which presents the percentages of -t.
In sum, the corpus of Estonian dialects and the 1990s data collected from Polva demonstrate that both -t and -tU([??]) are used in spoken Voru and Setu. The only difference from the written standard of Voru is the fact that both markers are used both in the singular and plural. There may be several reasons for this. One possible reason is that the paradigms of singular and plural have lost their original denotation (for further analysis see 4.2). Another possible reason might be an analogy with the active past participle: if the most common marker for the active past participle is -nU, language users might regard -tU as the most typical marker for the passive past participle. The latter argument is favoured by the notion that similar forms are formed according to the same principles. This is connected to the explanation that the present-day South Estonian is at some point strongly influenced by Common Estonian. It might be that local people, who use two varieties, tend to form South Estonian forms according to Common Estonian rules; i.e. in some grammatical forms (e.g. the passive past participle) they draw parallels with Common Estonian. Notwithstanding these different interpretations, it is evident that the actual spoken usage of the past participle markers is different from the Voru written standard.
4.2. Singular and plural in three South Estonian sub-dialects
Historically, -t marked only the singular and -tU[??] only the plural. The analysis of the recordings made in the 1990s show that like in the case of the active past participle some of the historical rules are no longer valid for the passive past participle either (see Mets 2000 : 93-95; Iva 2002a : 97-99, 101-103, 121). Similar tendencies emerge in the analysis of the recordings made between the 1960s and 1980s: historical paradigms are not relevant in the actual spoken language of the time, i.e. plural markers are used in the singular, and vice versa (see Table 2). Though, these results differ for each sub-dialect. In Hargla, the most common formative is the plural marker -tU([??]) which is used to mark both the singular and the plural. Other dialectologists (Keem, Kasi 2002 : 20) state that in Hargla the marker -t is used for the singular and -tU[??] marks the plural. Still, the present database proves that the distinction between the singular and plural is not so rule-based. In Vastseliina, the singular is mostly marked by -t and the plural by -tU([??]). Likewise, the same tendency can be seen in North Setu.
These results show that the old system to distinguish between the singular and plural has been confused already in the 1960s. Moreover, the data collected at the end of the 1990s in Polva exhibits strong tendencies to use -tU([??]) in both numbers. Marker -tU seems to become general in spoken Voru. This generalisation is typical for the northern and western parts of Voru dialect area--northern parts are influenced by neighbouring sub-dialects of Tartu and western parts by Mulgi (Toomse 1998 : 46, 94, 109, 131). In Tartu the passive past participle marker is -tU (Keem 1970 : 40), and in the eastern sub-dialects of Mulgi (including Helme which is very close to Hargla) the passive past participle is marked by -tU, though in general the marker -t is common for the other sub-dialects in Mulgi (see Tanning 1961 : 48).
Just like in the case of the active past participle, the word structure is relevant for the variation of the passive past participle markers. The analysis of the 1960s-1980s material reveals that the usage of the marker -t decreases in longer forms - all monosyllabic words end with -t, di- and trisyllabic forms favour -tU([??]). There occur no differences between the three sub-dialects. Divergent results emerged from the analysis of the 1990s material - -tU dominates in all forms, -t is used rarely and mostly in longer forms (Mets 2000 : 97-100).
The above presented analysis shows that spoken South Estonian differs from the written norm of Voru. The differences can be seen already in the material collected between the 1960s and 1980s. In regard to the past participles the generalisation of the markers -nU and -tU seems to be an ongoing process as the 1990s data indicates. Whereby, the occurrence of the glottal stop at the end of the markers is regulated by the phonological conditions.
Another interesting observation concerning the active past participle emerges from the analysis of the singular and plural forms. Already in the 1960s the old singular-plural paradigm has not preserved, and both markers are used in both numbers. Likewise, the recordings of the 1990s do not exhibit a completely rule-based paradigm, but show an inconsistent language usage.
Similarly, an irregularity in the usage of the singular and plural markers can be seen in the case of the passive past participle, where in the plural both singular and plural markers are used.
To conclude, the above-analysed data from the 1960s to the 1980s affirm that the old paradigms, which invalidity emerged from the 1990s materials, were not very well preserved already earlier. The Voru written standard does not correspond to the colloquial language of neither the period between the 1960s and the 1980s nor the 1990s.
ALL--allative; CLTC--clitic; GEN--genitive; IMPRF--imperfect; NEG - negation; PART--partitive; PL--plural; pl--plural; PRTCPL--participle; sg--singular.
Har--Hargla; Kan--Kanepi; Krl--Karula; NSe--North Setu; Plv--Polva; Rou--Rouge; Rap--Rapina; Urv--Urvaste; Vas--Vastseliina; WSe--West Setu.
VES 2002--Voro-eesti synaraamat, Tarto-Voro 2002 (Voro Instituudi toimundusoq 12); VL 1996--Vorokiilne lugomik. II parandot ja tavvendet trukk, Voru 1996.
[TEXT NOT REPRODUCIBLE IN ASCII].
Iva, S. 2003, Voru korisulghaalik.--Lounaeesti haalikud II, Tartu (Tartu Ulikooli eesti keele oppetooli toimetised 27), 68-91.
--2005, Glottal Stop in Voru South Estonian.--LU XLI, 123-133.
Iva, T. 2002a, Haritlaste voru keel. Magistritoo, Tartu (Manuscript).
--2002b, Eesti uhiskeele mojutused voru keeles.--Vaikeisi kiili kokkoputmisoq, Voro (Voro Instituudi toimondusoq 14), 84-92.
Juva, S. 2005, Voro keele astmovaeldusost.--Piirikultuuriq ja keeleq, Voro (Voro Instituudi Toimondusoq 17), 103-112, 185.
Juva, S., Kasak, E., Help, T. 1995, Till'okono tiijuht' Voru-Setu kirakeele manoq.--Vorukiilne lugomik, Voru.
Keem, H. 1970, Tartu murde tekstid, Tallinn (Eesti murded III).
--1997, Voru keel, Tallinn.
Keem, H., Kasi, I. 2002, Voru murde tekstid, Tallinn (Eesti murded VI).
Lindstrom, L. 2004, Lounaeesti keelematerjalid eesti murrete korpuses.--Tartu Ulikooli Louna-Eesti keele- ja kultuuriuuringute keskuse aastaraamat III, Tartu, 85-94.
Lindstrom, L., Bakhoff, L., Kalvik, M-L., Klaus, A., Laanemets, R., Mets, M., Niit, E., Pajusalu, K., Teras, P., Uiboaed, K., Veismann, A., Velsker, E. 2006, Sonaliigituse kusimusi eesti murrete korpuse pohjal.--Keele ehe, Tartu (Tartu Ulikooli eesti keele oppetooli toimetised 30), 154-167.
Lindstrom, L., Pajusalu, K. 2003, Corpus of Estonian Dialects and the Estonian Vowel System.--LU XXXIX, 241-257.
Mets, M. 2000, nud- ja tud-partitsiip Polva murrakus. Bakalaureusetoo, Tartu (Manuscript).
--2004a, Voru konekeel: nud-partitsiibi tunnuse varieerumine Vastseliina murrakus. Magistritoo, Tartu (Manuscript).
--2004b, Voru konekeel: nud-kesksona varieerumine.--KK, 657-669, 719.
--2005, Voru ja setu konekeele mineviku kesksonade tunnused: kas tegelik keelekasutus vastab voru kirjakeele normile?--Tartu Ulikooli Louna-Eesti keele- ja kultuuriuuringute keskuse aastaraamat IV, Tartu, 65-77.
Nigol, S. 1994, Hargla murraku konsonantism, Tallinn.
Pajusalu, K. 2003, Estonian Dialects.--Estonian Language, Tallinn (Linguistica Uralica. Supplementary Series/Volume 1), 231-272.
Pajusalu, K., Velsker, E., O r g, E. 1999, On Recent Changes in South Estonian. Dynamics in the Formation of the Inessive.--International Journal of the Sociology of Language 139, 87-103.
Pajusalu, K., Hennoste, T., Niit, E., Pall, P., Viikberg, J. 2002, Eesti murded ja kohanimed, Tallinn.
Saareste, A. 1938, Eesti murdeatlas. I vihik, Tartu.
--1941, Eesti murdeatlas. II vihik, Tartu.
--1955, Petit Atlas des parlers estoniens. Vaike eesti murdeatlas, Uppsala.
Tanning, S. 1961, Mulgi murdetekstid, Tallinn (Eesti murded I).
Toomse, M. 1998, Louna-Eesti murded 1-30. Kaardid, Turku (Turun yliopiston suomalaisen ja yleisen kielitieteen laitoksen julkaisuja 56).
* This research was supported by the Estonian Science Foundation grant no. ETF5968. The author is grateful to Karl Pajusalu and Sulev Iva for their valuable comments.
(1) For example, see Pajusalu, Velsker, Org 1999; Iva 2002a; 2002b; Mets 2004a; 2004b.
(2) The markers are given in the Finno-Ugric transcription and the vowel phonemes that may be the subjects to vowel harmony are marked by capital letters: -nU[??] = -nuq/-nuq (q marks the glottal stop in the Voru written standard), -nU = -nu/-nu, -t/-D = -t/-d and -tU[??]/-DU[??] = -tuq/-tuq/-duq/-duq.
(3) Though, in general the glottal stop marks the plural in Voru, e.g. kala 'fish' and kala[??] 'fishes'. For more, see Keem 1997 : 5-6.
(4) In Voru the consonants in non-initial syllables may be the subjects to grade alternation: veet '(have been) taken' (strong grade) and veeDu '(have been) taken' (weak grade). For more, see Juva 2005 : 103-112, 185.
(5) In nowadays South Estonia only the Voru variety has written standard, Setu is still being standardised.
(6) The tapes and texts are preserved in the corpus of Estonian dialects at the University of Tartu and the Institute of the Estonian Language.
(7) About the corpus, see for example Lindstrom 2004; Lindstrom, Pajusalu 2003; or Lindstrom, Bakhoff, Kalvik, Klaus, Laanemets, Mets, Niit, Pajusalu, Teras, Uiboaed, Veismann, Velsker 2006.
(8) In Estonian dialectology the dialects are divided into sub-dialects according to the borders of the historical parishes (see Pajusalu, Hennoste, Niit, Pall, Viikberg 2002 : 46; Pajusalu 2003 : 231).
(9) The number marks the number of tagged texts in the corpus.
(10) As this concerns West Setu the result may be inconclusive because only one text was analysed, i.e. only one idiolect was under observation. On the other hand, Setu is known to have preserved many old South Estonian traits.
(11) The written standard of Voru is based on the eastern sub-dialects of Voru where Vastseliina is central. On the other hand the history of language has also been taken into consideration.
(12) The distinction between singular and plural by markers -nU[??] and -nU has disappeared in Voru as the following analysis will show.
(13) One must not forget that here only one idiolect is being analysed because no more tagged texts are available in the corpus for West Setu.
University of Tartu
Phone: +372 5040379
Table 1 Active past participle markers in the singular and plural 1960s-1980s 1990s Hargla Vastseliina North Setu Vastseliina Polva -nU -nU? -nU -nU? -nU -nU? -nU -nU? -nU -nU? sg 98 20 58 64 31 26 530 165 785 323 % 83 17 48 52 54 46 76 24 71 29 pl 25 2 15 16 16 6 136 51 98 53 % 93 7 48 52 73 27 72 28 65 35