Printer Friendly

Livonian and Leivu: shared innovations and problems.

1. Introduction

Both Livonian and the isolated Leivu South Estonian were spoken in what is now Latvia. As far as we know the Livonian language area bordered on Livonoid or Tamian Low Latvian and the Leivu language area was surrounded by Latgalian High Latvian. After the beginning of the second Soviet occupation in Latvia (1944-1991), the traditional lifestyle on the Livonian Coast in the northern tip of the Courland peninsula was dramatically changed and the younger people had to move to cities or inland towns, where they lost their Livonian language in the Latvian environment. As a consequence, Livonian has ceased to be used in everyday communication. The Leivu linguistic island, as far as it has been studied, is usually divided into two subdialects, conservative Alamoiza (Latvian Lejasmuiza) and more innovational Saltna (Latvian Zeltini). Leivu was mostly replaced by the Latvian language already by the beginning of 1930s; the last Leivu speaker died in 1988. Despite their isolation, Livonian and Leivu have several shared or at least similar innovations. Hopefully, the present pilot study, which is based on some data from Saltna, will provoke a more thorough study of the field.

Here, the Livonian data are given orthographically, except that the East Livonian lower o has been extracted from o (instead of o there is always a in West and Ira Livonian) and the stod is indicated by an apostrophe as a substitute for the symbol ' of the Uralic Phonetic Alphabet. The Leivu data are taken unchanged from Niilus 1935. (1) The transcription system used by Valter Niilus differs from the contemporary one in rendering the three distinctive quantity patterns, hereinafter Q1, Q2, and Q3, of Estonian. Namely, this system leaves the final components of polyphthongs and the initial components of consonant clusters unmarked for duration (i.e. full-short) in the initial syllable of words of Q2. As the vowel of the 2nd syllable of words of Q1 and Q2 is mostly full-short and not half-short in Leivu, this convention must not be forgotten. (2)

2. Livo and leivu

Characteristically, Livonians and Leivus use the same stem in the name of their languages, cf. livi kel 'the Livonian language' and livu--livu kil' 'the Leivu language'. In the most part of Leivu, the former long * i has been diphthongized into ei under the Latgalian influence. The stem-final o [e] in Livonian is the result of a regular reduction of vowels except * a or * a in the 2nd syllable. The stem-final a in the Leivu name (3) may be influenced by Latvian: Livonian livo kel 'the Livonian language' is in literary (Low) Latvian libiesu valoda or livu valoda and leivu kil' 'the Leivu language' is in Latgalian leivu volida. The boldfaced u in Latvian words is the genitive plural ending and the exact meanings of the two Latvian/Latgalian phrases are 'the language of Livonians' and 'the language of Leivus'.

The origin of the name for Livonians and Leivus is unclear. SSA 2 (sv Liivi) considers it as possibly connected with the Estonian word liiv 'sand' and ascribes this etymology to Friedrich Kruse (1846 : 95) and adds a bare reference to Vaari 1959 : 196. Eduard Vaari does not mention Kruse at all but claims that in the second volume of August Wilhelm Hupel's "Topographische Nachrichten" (Hupel 1777 : 183-187) there is "a longer etymology of liivi". Actually, Hupel reports already in the first volume of "Topographische Nachrichten" (1774 : 67-68) that the derivation of the country name by Moritz Brandis (4) from the Livonian Liiw (5) 'sand', Liwa ma 'sand soil' can well be accepted. In addition he claims that "In Livonian as well as in Estonian, Liiw is a small net, but also sand. In the latter sense many persons have found with a well-founded reason a motive for giving the name to the people and country" (1774 : 69). To be more exact, Estonian liiv (GSg hivi) is 'a short triangular net for fishing in the shallow water', and liiv (GSg hiva) is 'sand'. Hupel obviously tries to explain why there is an [f] in the German name of Livonia, cf. Liefland--Livland. According to Hupel, Russians and Vends who earlier inhabited Livonia, could pronounce Liiw as Lhf (Hupel 1774 : 72). Hupel (1777 : 183) quite on the contrary insists that the neighbors (actually Estonians) of Livonians have given them the name Livi rahwas 'the Livonian people' and he is not sure whether this name is general or also Liwa rahwas is used. He thinks that i has replaced a so that Livonians could use the plural Livid because Livad from Liwa were "an entirely abnormal expression" (ein schlechterdings ungewohnlicher Ausdruck). Friedrich Kruse, in addition to a laconic presentation of the same etymology (1846: 95) thinks that Livonians seem to have received their names from Liv i.e. sand of the sand coast they populated (1846 : 167).

Undoubtedly, the beautiful sandy beach of both the Livonian Coast in Courland and of the eastern coast of the Gulf of Riga may tempt to associate this name with Estonian (including Leivu) liiv (GSg liiva), Votic and Ingrian hiva. Unlike in Estonian, in Votic, and Ingrian, 'sand' is called jou go, (PSg jogta ~ jogta a) in East and Ira Livonian, jugt (GSg jugo, PSg jugta) in West Livonian, and jug in Salats (in

Latvian Salaca, in German Salis) Livonian. In addition, wet sand in beach berm or in soil is called liedog or liedig in Courland. We have no data about any fishing tackle with a name resembling liiv in Livonian or Leivu. There is no stem in Livonian, which could serve as a source of the name of the people or country. Livonians in Courland have traditionally called their language randakel 'coastal language' and themselves randalizt 'coastal people' or kalami'ed 'fishermen'. The names livo kel and livod or livlizt may be borrowed from German. But we do not know how the Daugava and Gauja Livonians called themselves.

In the Nestor's chronicle ([TEXT NOT REPRODUCIBLE IN ASCII.] 'Tale of Bygone Years', written in Old Church Slavonic at the beginning of the 12th century), Livonians are called [TEXT NOT REPRODUCIBLE IN ASCII.]. As in Latvian Livonians are usually libiesi (singular libietis), the Old Russian term must have been borrowed from Latvian, probably from Selonian. Similarly, the Salats Livonian name Lib ma 'Livonia' and lib 'a Livonian' are borrowed from or influenced by Latvian.

In view of the possibility that the name for Lithuania, which in Lithuanian is Lietuva, and the name for Lithuanians lietuviai (singular lietuvis), cf. also Latvian lietuviesi (singular lietuvietis), dialectal (e.g. in Courland) leigi (singular leitis), may be connected with the word for 'rain', cf. Lithuanian lietus, Latvian lietus, Curonian liets, which is derived from the Balto-Slavic verbal root * lei- > li- 'to pour', it is possible that the name for Livonia and the name for Livonians may be connected either with the East Slavic stem for rainstorm, cf. Russian [TEXT NOT REPRODUCIBLE IN ASCII.] or its Slavic verbal base stem diva-, which occurs in imperfective prefixal verbs, e.g. Russian [TEXT NOT REPRODUCIBLE IN ASCII.] 'to pour out', Polish wylewac 'to overflow', Czech vylevati 'to flow out', Slovak vylievat' 'to pour out; to spill'. The stem * liva- goes back to the Balto-Slavic stem * lei-. To be true, this explanation clearly requires for Livonians other Slavic neighbors than Old Russians. On the other hand the name of Lithuania has usually been connected with the Latin word litus (< * leitos) 'coast, sea shore', which is also considered to be derived from the Indo-European root * lei- 'to pour', (Fraenkel 1960, sv lietuvis; Vasmer 1950-1958, sv [TEXT NOT REPRODUCIBLE IN ASCII.]; Karulis 1992, sv lietuviesi). This etymology of Lietuva meets the problem of location of this coast because the Grand Duchy of Lithuania subjected the coastal Baltic tribes as late as in the 13th century.

One could as well or even better derive the name for Livonians from Scandinavian. It is possible to speculate that for the ancient Vikings who in storms succeeded in bypassing the huge shallow water sandbank north of the Cape of Kolka and reached the eastern coast of what is now called the Gulf of Riga, the coast was a real Lifland 'land of life' and the friendly inhabitants of Lifland were naturally lifri and maybe also lifri 'brothers', cf. Swedish liver 'Livonians'. The probably unattested meaningful Old Norse placename Lifland became later the more obscure Liefland ~ Livland in German and Livonia in Latin. If this stem was borrowed also by the Daugava and Gauja Livonians they must have accepted it no later than in the first half of the 11th century and not before the Vikings began to use the Daugava River as a waterway to what is now Polatsk in Belarus.

In the chronicle of Henricus de Lettis, written in Latin, Livonians are called Livones ~ Lyvones. The vowel o in the second syllable can be considered an argument against Moritz Brandis's etymology of the name of Livo nians because of the similarity of the rounded vowels o and a in Livones and leivu. On the other hand, most peoples have names ending in ones, cf. Theutones 'Germans', Saxones 'Saxonians', Lettones ~ Letthones 'Lithuanians', Curones 'Curonians', Selones 'Selonians', Estones 'Estonians'.

No matter what the origin of livo and leivu is, it is noteworthy that some time after the Livonians were defeated, the label Livonian became an object of desire of non-Germans in Latvia. There exists at least one Old Latvian text with an adequate translation into German that is presented as a Livonian text. (6) Similarly, there is a doubt that Mahrz Sahrum (7) was not a Livonian although he introduced himself to a German, who had thought he was a Latvian with the following words: Herr! ich bin keine Lette, sondeen der letzte Live in der Umgegend Wenden's 'Sir, I am no Latvian, but the last Livonian in the vicinity of Wenden (= Latvian Cesis)', cf. Hagemeister 1849 : 78). Henricus de Lettis, who in 1204 probably attended the christening of Vends in Cesis, wrote that Vends where first driven away from the Venta river in Courland and later from the area of present-day Riga, and therefore Vends escaped to Letts (Latgalians). Hence, Henricus did not associate Vends with Livonians or Latvians. (8)

Leivus, like other Estonians, called themselves earlier maarahvas 'land people' (Wiedemann 1869 : 499). As Leivus were isolated from other South Estonians in their Latgalian environment and their dialect had its distinctive phonetic and lexical characteristics, they must have considered themselves different from other South Estonians. The local Germans, who also noticed this difference, probably decided that Leivus are Livonians, in German Liven. As there were no Livonians in the area, local Latgalians and Leivus accepted this decision and borrowed the corresponding German stem to denote Leivus.

3. Breaking of vowels in initial syllables

Both in Livonian and Leivu, the Proto-Finnic breaking of mid vowels involves both long and short vowels. Still, breaking has its own restrictions in each language.

3.1. Breaking of mid vowels in Livonian

In Livonian, breaking of mid vowels occurs both in syllables with plain (raising) tone and with broken tone:
* e > ie [ie]:        miez 'man', mi'ed 'men'
* o > no [uo]:        suo 'marsh', ku'odi 'direct'
* o > uo [uo] > ie:   uo > ie 'night', i'ezo 'night (IllSg)'
* e > ie [ie]:        tiera 'grain; blade', pie'zzo 'to wash'
* o > no [uo]:        suoda 'war', kuo'd 'home'

The breaking of * e, * o, and * e did not occur (a) before a syllable containing a monophthongal * i (ve'rri 'bloody (NSg)', veriz 'bloody (GSg)' from ve'r < * veri; te'b < to'b < * tobi 'epidemic', in contrast to lie 'ggi 'muddy (NSg), liegiz 'muddy (GSg) from lieges 'mud'); (b) in the diphthong * ei (leba < * leiba '(rye) bread', leibo < * leibado 'bread (PSg)', and (c) before a palatal or palatalized consonant (keja 'grindstone', reja 'rake', tedi 'works (PSg)'). Note that the vowels o and o, which are usually transcribed as and [e] and [e], are not mid vowels but high vowels, and East and ira Livonian mid vowels o and o go back to no and no, whose components fused when preceded by p, m or v, cf. padub < puodub 'it aches' and pa'ddo < puo'ddo 'to ache' (cf. also Posti 1942 : 7, 12, 129). In the later word polo 'rag' (PSg po'ddob), the mid vowel did not undergo breaking any more. Note that word initial ie and uo have received in Livonian a prothetic consonant j and v respectively, cf. West Livonian vuolda 'to be', vuo'l [vuo'l'] 'he was', and jiema [jiema] 'mother' vs. East Livonian volda, vo'l and jema ~ jiema.

Breaking of long mid vowels was also characteristic of Salats Livonian.

3.2. Breaking of mid vowels in Leivu

In Leivu, long mid vowels underwent breaking only in syllables of Q2, whereas breaking of short vowels occurs in syllables of all three quantities Q1, Q2, and Q3: (9)
* e > ie:  vie 'water (GSg)', kiele? 'tongues (NPl)'
* o > uo:  uomen 'tomorrow', uona 'lamb (GSg)', duova 'they drink'
* o > uo:  suova 'they eat', uoze 'at night'
* e > ee:  veeras 'stranger'; meedet 'measured (PastPple)'
* e > ie:  ieza 'father', nieli 'four'; viet 'water (PSg)'
* o > uo:  uoza 'part'; kuot' 'sack'

The breaking of long mid vowels * e, * o, * o' has regularly occurred in long syllables of Q2 and locally in a set of stems of Q3 with a long vowel resulting from vowel contraction after the loss of an intervocalic consonant, e.g. vie < * veten 'water (GSg)'. The breaking of * e was not a general innovation. Additionally, the quantity pattern of the resulting diphthongs differs from the diphthongs resulting from the breaking of * e, * o, and * o. It is obvious that the diphthongs under discussion need a more thorough study. The breaking of * g has also occurred in one part of North Estonian, the Central North Estonian, where the resulting diphthong is ee in Q2 and ee ~ ee in Q3, and in Hill East Votic (i.e. in Kattila and in the area south of it) where the resulting diphthong was ie.

The raising of * e, * o, and * o in Q3 and the breaking in Q2 occurred also in Kodavere East Estonian. Lauri Kettunen (1912 [1914] : 67; 1913 : 176; 1962 : 138) has suggested that such raised monophthongs in Kodavere and in South Estonian rose from ie, uo, and iii; and this has been the generally accepted view unlike the one proposed by Viitso (2003 : 177). Although breaking of long mid vowels in Q3, which rose in the course of vowel contraction triggered by the loss of intervocalic * h and weakened stops preceded by a short monophthong, cf. vie < * veten 'water (GSg)', (ma) tae < * taken 'I do', mie < * mehen 'man (GSg)', occurs only in one part of Leivu, it nevertheless makes a strong argument against the theory about the long mid vowel raising in Q3 via breaking of long mid vowels. This breaking is later than long mid vowel raising in Q3 and has taken place at the same time with breaking in Q2. Hence the diphthongs that developed from former long mid vowels in syllables of Q2 are not a residue of the once general breaking of long mid vowels, which remained untouched by the fusion of the diphthongs ie, uo, uo to i ~ i, u ~ u, u ~ u. The diphthongs are the result of breaking that took place after long mid vowels in syllables of Q3 were already raised to i ~ i, u ~ u, u ~ u, cf. mis < * mes 'man', nur < * nori 'young', u < * o 'night'. In another part of the Leivu area, the contracted long mid vowels underwent both the long mid vowel raising and the breaking of long high vowels component, cf. section 4. Hence the long mid vowel raising and the vowel contraction after the loss of intervocalic consonants occurred in reverse order in different Leivu subdialects.

General breaking of long mid monophthongs has also taken place in several North Estonian dialects, Hill and Pontizoo East Votic, Finnish, Karelian, and Lude; in Estonian East Finnish, and Karelian, it involved also the long contracted mid vowels. The area of breaking of long monophthongs is discontinuous; the southwestern pocket includes partially the Estonian western islands, namely the eastern Saaremaa, Muhu, and Kihnu. Additionally, both the long * e from Proto-Baltic * ei or * oi and the long * o from Proto-Baltic * ou have undergone similar breaking in Latvian and High (or Aukstaitian) Lithuanian (cf. Rudzite 1993 : 150-153; 154-156, 163; Zinkevicius 1978: 85-86). Low (or 2emaitian or Samogitian) Lithuanian dialects are more conservative: * ou is preserved in Northern and * o in Western Zemaitian. As the breaking of long mid vowels occurred also in Salats Livonian, once probably spoken up to the P5rnu River in Estonia, there are no known grounds to decide whether breaking in Livonian was somehow related with breaking on Estonian islands or took place under Latvian influence. In Leivu, the breaking of long mid vowels was a relatively late change and possibly influenced by Latgalian.

Breaking of short mid vowels is relatively rare. Still, in addition to Livonian and Latvian it was also characteristic of the Lutsi South Estonian pocket in the vicinity of Ludza in south-eastern Latvia. According to Mari Must and Aili Univere (2002 : 117), irregular and also individual breaking of * e, * o, and * o has been noted in five North Estonian parishes which can be considered four separate areas. I have heard such breaking also in the speech of one person from Simuna. The breaking of * e to ja (probably * e > * ie > * ia > ja) occurred also in Old Norse. In Livonian and Estonian, characteristically, the supershort i in such diphthongs cannot be identified with the palatalization of a preceding consonant as sometimes done by Sjogren and Wiedemann, who usually identified short ie and uo as e and o. In Leivu, the diphthong ie follows both a palatalized and an unpalatalized consonant. Lembit Vaba (1997 : 41) explains this by the Latvian influence. Marta Rudzite presents the diphthongs ie and uo in the list of 56 Livonoid Low Latvian monophthongs and diphthongs (1964 : 160) and ie in a similar list of 104 High Latvian monophthongs and diphthongs (1964 : 266) but she does not mention them elsewhere. In these Latvian dialects, ie and uo are probably rare local diphthongs. As the only example of ie I could find results from the shortening of a (long) diphthong, it is possible that such diphthongs in High Latvian never developed as a result of short mid vowel breaking. In Livonian, there are no examples of shortening of a long diphthong. Hence, up to now there is no known reason to substantiate breaking of short mid vowels in Livonian and Leivu by direct Latvian influence. Most probably the breaking of short mid vowels in Livonian and Leivu took place by analogy with the breaking of long mid vowels.

4. Triphthongs

As a result of breaking of the initial mid component of a former diphthong both Livonian and Leivu have triphthongs.

In Livonian stem-initial syllables, two sequences of three different vowels ieu and uoi occur in different tone and quantity patterns:
ieu   <   * eu:  lieudo [lieude] 'to find', kie'v [kie'u] 'cough',
          kie'vvo [kie'uvva] 'cough (PSg)'
ieu   <   * eu:  West Livonian lieudab [lieudab] (10) 'he finds'
uoi   <   * oi:  kuoigid [kuoigID] 'ships', kuo'ig [kuo'iG] 'ship',
          kuo'igo [kuo'ige] 'ship (PSg)'; suoimo [suoime]
          'to swear', tuoi [tuoi] 'other, second'
uoi   <   * oi:  suoimob [suoimeB] 'he swears', tuoizta [tuoista]
          'other, second (PSg)', tu'oigoz [tu'oigez] 'birch bark'

In these examples breaking has been applied to (a) diphthongs of Proto-Finnic origin (lieudo, suoimo : suoimob, tuoi : tuoizta); (b) diphthongs resulting from contraction after loss of * h (tu'oigoz), (c) diphthongs resulting from fission of (* )v and (* )vv <* h (kie'v : ki'evvoo), and (d) diphthongs resulting from fission of * g and * gg (kuoigid, kuo'ig : kuo'igo).

In Leivu the number of sequences of three different vowels is bigger:
idu < * au:  kiaus 'he walked'
uoi < * oi:  uoija 'I hold': uoitma 'to hold'; puois 'boy' : GSg
             puoizi; ruoi 'grass'
uoe < * oe:  suoe 'wolf (GSg)', tuoe 'support (GSg)'
uoa < * oa:  kuoa 'hut (GSg)'; uoas 'thistle'
iei:         miei 'honey (GSg)', viei 'water (GSg)'; miei 'man (GSg)',
             riei 'threshing house'
uou:         suouwa 'to eat (Inf)'; ruous 'sputum'

In Leivu, breaking has been applied to short initial components of (1) original diphthongs (kiaus; uoija : uoitma; puois : puoizi); (2) diphthongs resulting from contraction after loss of (a) * t (suoe, kuoa), (b) * k (tuoe), (c) * t and * h (uoas < * oas < ohas < * ohtas, cf. GSg uoxta), * (3) diphthongs in syllables of Q3 resulting from breaking of a long high vowel, which rose from raising of a contracted vowel after loss of (a) * t (miei 'honey (GSg)', viei), (b) * k (ruous), (c) * h (miei 'man (GSg)', riei, ruoi uoas), and (4) diphthongs resulting from breaking of * u < * o before * t > w in certain verb forms of Q2 (suouwa). Note however, that it has not been proven that uoe and uoa are triphthongs and not heterosyllabic sequences. Cases (3) and (4) are not general, cf. section 3.2, on the other hand they are somehow related with breaking of primary long high vowels where the initial components of the resulting diphthongs have not been subject to breaking, cf. e.g. * kounar vs. suouwa.
* i > ei:  leina 'city (IllSg)', leiva 'sand (GSg) : PSg leiva
* u > ou:  soud 'mouth (PSg)', d'ouwa 'to drink'
* u > ou:  * koudz 'nail, claw': GSg * kuudze; * kounar 'ell': GSg
           * koundre

What remains problematic here are the criteria of breaking vs. non-breaking of the initial components of diphthongs resulting from breaking of long raised or high monophthongs, cf. e.g. suouwa vs. * kounar.

Breaking of long high vowels was not common even in Saltna, cf. Ariste 1931. This was most probably influenced by Latgalian as there are parallels for breaking of * i and * u in Latgalian (Rudzite 1993 : 150, 245-246; 158-159, 246-247) and in Old Prussian (Zinkevicius 1980 : 86; Rudzite 1993 : 163).

5. Loss of intervocalic * h after a short initial syllable

In Livonian and Leivu, the Proto-Finnic * h is lost without any provable trace in stem-initial and stem-final positions." For Proto-Finnic consonant clusters on the border of the stem-initial and the following unstressed syllable in Livonian, the preconsonantal * h is lost, giving rise to stud in the initial syllable but the postconsonantal * h is lost with no trace. In Leivu, the postconsonantal * h was removed to the beginning of the cluster, after which both the earlier and the new preconsonantal * h were either assimilated with the following consonant or lost in Q2 (i.e. at the end of primary-stressed syllables which originally were followed by a closed unstressed syllable that now may be open) and preserved in syllables of Q3. The intervocalic * h is lost in both Livonian and Leivu due to either (1) substitution with another consonant or zero, or (2) substitution with stud.

5.1. Substitution of intervocalic * h with another consonant and the total loss of * h

In Livonian, there is j instead of * h in the former sequences * iha and * eha (vija 'poison', keja 'body', leja 'flesh', reja 'rake'), and v in the environment * uha (12) (piva 'holy; holiday', pivapava 'Sunday').

Unlike in Livonian, in Leivu * h is lost (13) in former sequences * iha, * eha > * iha, and * uha (vianu 'gotten angry (past participle)', l'ia'meat', ria 'rake', puab 'Sunday'). * h is also lost in former sequences {* ehi, * ehe) > * ihi (miei 'man (GSg)'), riei 'threshing house'), * ehti > * ehi (iei 'I adorn, decorate)', * ihko > * ihu (viu 'sheaf (GSg)'), * uhka > * uha (tua ~ tuad 'thousand'), * ohta > * oha (uoas 'thistle'), * ohi (ruoi 'grass'), * ohtu > * ohu (tou 'birch bark (GSg)'), * ahvu > * ahu (ruou? 'kidneys (NPI)'), * uhku > * uhu (puou 'palm, hand (GSg)'). Additionally there is an intriguing group of words with the sequence ij as the substitute for * h or * hh in environments a_e and u_e: jaije 'cool', aijer 'barren' : GSg aHtre, vaijer 'maple' : GSg vaHtre, uijerd 'auger', cf. also Voru South Estonian jahhe, aher, vaher, Finnish vaahtera 'maple', Erzya Mordvin ukstor 'maple'. Voru jahhe is a regular reflex of the underlying stem * jaheta, cf. also North Estonian NSg jahe (< * jaheta), GSg jaheda (< * jahetan). Hence, the Leivu stem jaije could be explained when proposing a change of * h to j before e. This would entail jaije to be a stem with a regular gemination of /j/: * jaheta > * jahea > * jahe > * jaje > * jajje > jajje,=jaije or * jaheta > * jahea > * jajea > * jajjea > * jajje > jajje = jaije. In this case, by analogy with Voru and North Estonian forms, we could expect in other nominative case forms only a single conso nant [dagger]j as the reflex of the weak grade cluster * ht, cf. e.g. * vahter > * vaher > * vajer > [dagger]vajor = [dagger][vajer]. Hence, a special development, an epenthesis, must have occurred in Leivu which created preconditions for the gemination of the weak grade reflex of * ht that must have been * j, cf. e.g. * vahter > * vaheter > * vaher > * vajer > * vajjer > vajjer = vaijer.

5.2. Substitution of * h with stod

Stod, laryngealization, or creaky voice is not a segment but modulation of a sonorous segment, which is produced by means of an additional effort of vocal cords that in case of especially emphatic pronunciation may be realized even as a glottal stop but most usually as a drop (or even a break) of pitch and intensity. The latter property has induced the term BROKEN TONE. In spontaneous speech, for different reasons, stud may often be hardly noticeable if at all.

In Livonian, except in cases mentioned in section 5.1, the former postvocalic * h that followed the vowel of the stem-initial syllable is, potentially, in all cases represented as stod, cf. e.g. rQ' < * raha 'money', ri' < * rihi 'threshing house', pa'zo < * pahesen 'head (IllSg)' with stod from the former intervocalic * h and no'go < * nahga 'skin'), vi'mo < * vihma 'rain', le'd < * lehti 'leaf' : GSg li'ed < lehten. In addition to stud from * h, Livonian offers even more cases of stud connected with the loss of vowels in non-initial syllables and cases of stud in Latvian borrowings.

In Leivu, * h is substituted with stud mostly in illative forms of monosyllabic vocalic stems and in stems where * h occurred between identical vowels, e.g. na'a? 'skins', pa'a 'head (IllSg)', ra'ad 'money (PSg)' (note that this way of transcribing stud in Leivu words has insisted unnecessary decisions about the position of stud in a vowel).

In Livonian, stud was first identified in 1890 by Vilhelm Thomsen. Later measurements have shown that the laryngealization in such words takes place somewhere near the midpoint of a long vowel, e.g. le'd is actually pronounced as [le'eD] or [leeD] (Kettunen 1938 : XXI). In Leivu, stod was first mentioned in writings of Valter Niilus. Although beginning with Lauri Kettunen, stod in Livonian was mostly considered a tonal feature, (14) some linguists considered it a sound, i.e. a segment (cf. German Stosslaut, Finnish katkoaanne, Estonian katkehaalik). Fanny de Sivers (1965) equated it with a glottal stop (French le coup de glotte). However, the tonal essence of stod is proved by a morphological argument: the noun stems rQ' 'money' and vQ' 'foam; wax' take in partitive and illative singular respectively the case endings -do and -zo that occur only with monosyllabic stems ending in a long monophthong or a long diphthong ie or uo (Viitso 2007 : 27, fn. 8). In Leivu, stod was considered a tonal feature (an intonation) by Eberhard Winkler (1999 : 202) although formally his example na'a? 'skins' does not prove it. Yet he is right, as also in Leivu the word ra'a 'money' takes the partitive singular ending -d that occurs only with monosyllabic stems ending in a long monophthong, cf. ra'ad.

6. Prepalatal sibilants s and z

Livonian and Saltna Leivu share their characteristic shift of sibilants from palatalized alveolar to prepalatal [s] and [z]. This shift is probably only a part of a more general shift of palatalized dental or alveolar consonants to prepalatal consonants. Palatalized alveolar sibilants have different history. Here only some comparable examples of probabilistic development of some words of Livonian and Leivu are presented:

A. Livonian
* ukti     >   * ukt'i > * ukt'si > * uksi > * uks > * uks > iks 'one'
* tauti    >   * taut'i > * tausi > * tauzi > * tauz > tauz = [tauz]
               'full', PPl tauzi [tauzi]
* viti     >   * vit'i > * visi > * vizi > * vizi > viz 'five'
* itseh    >   * it'seh > * it'se > * ize > * izi > * izi > * i'z >
               * i'zz > i'z 'self'
* ottsi-   >   * ot't'si- > * uot't'si- > vuoise 'to seek, look for' :
               3Sg vuotsuB > > East Livonian votso : votsub
* asja-    >   * asja- > * azja- > * azja > aza [aza] 'thing'

B. Leivu
* ukti     >   * ukt'i > * ukt'si > * ut't'si > uts 'one'
* tauti-   >   * taut'i > * tausi > * tauzi > tauz 'full'
* viti     >   * vit'i > * visi > * vizi > * vizi > viz 'five'
* et'seh   >   * et'se > * et'si > * esi > * ezi > * ezi > iezi 'self'
* ottsi-   >   * ot't'si- > * uot't'si- 'to seek, look for' : 3Pl
* asjan    >   * asjan > * azja > * azja > aza 'thing' (GSg)'

In both Livonian and Leivu there are words where one language has a prepalatal sibilant while the other language has not. Note that in Livonian after the vowel i, prepalatal sibilants have become unpalatalized alveolar sibilants. Characteristically, Livonian has no s and z, but Leivu has at least s, cf. uoas 'thistle'. Hence, in Leivu the shift * s > s either had already occurred when palatalization of consonants in new words was still an active process or was not fully completed. In Leivu, palatalization has been considerably more productive than in Livonian.

Fusion of s with j has parallels also in Baltic and Slavic (cf. Rudzite 1993 : 308-309; [TEXT NOT REPRODUCIBLE IN ASCII.] 1983 : 130).

7. Voiced obstruents

In both Livonian and Leivu, Proto-Finnic single obstruents (stops and sibilants) have become voiced in voiced environments, cf. Livonian viedab 'he pulls', jega 'every', pieza 'nest', andab 'he gives'; Leivu vieda 'he pulls', iega 'every', pieza 'nest', andaw 'he gives'.

According to tradition, the voiced single obstruents (stops and sibilants) appeared into different Finnic languages under the late influence of Russian or Latvian languages. In Leivu, voicing of intervocalic single obstruents is relatively late as the geminate obstruents that have risen through gemination are voiceless (rieppan 'fox', vietta 'to pull (Inf)'). In Livonian, on the contrary such geminates are voiced (rie'bbi, vie'ddo), i.e. voicing is older than gemination. Livonian voicing is also somehow connected with voicing in South Karelian, Aunus, Lude, and Veps, cf. the names of two berries in these Finnic dialects:
Livonian   South      Aunus     Lude     Veps     Gloss

buolgoz    buola      buolu     buole    bol      lingonberry
garban     garbalo    garbalo   garbal   garbol   cranberry

However, even Leivu voicing seems to be connected with voicing in South Karelian, Lude, and Veps, and maybe even in Votic (cf. Atlas Linguarum Fennicarum 2 : 483-485, 487-489):
Leivu       South Karelian   Lude     Veps     Votic      Gloss

babarn(a)   babarno          babarm   babarm   baabukka   raspberry
bauar                                                     baabukaz

For the distribution of the names for lingonberry and raspberry see Atlas Linguarum Fennicarum 2 : 475-476, 478, and 483-485, 487-489. Although the names for raspberry beginning with b in South Karelian are restricted to Ontarvi and in Lude to Kuuj5rvi, berry names beginning with a voiced obstruent show that voicing of initial stops is older than any Latvian and Russian influences.

8. Some typological conclusions

As can be seen, shared innovations or changes discussed above took place when Livonian and Leivu were already different. Therefore the changes met different preconditions and had different outputs. Formally identical changes could also meet different restrictions; e.g. the change * e > ie could not produce the sequence iei in Livonian because, unlike in Leivu, it was blocked before j, palatalized consonants, and i. When a change was induced by a language where the change was applied to a poorer vowel system, the change was not copied but generalized to the existing richer system.

doi: 10.3176/lu.2009.4.03


3Sg--3rd person singular of present indicative; GSg--genitive singular; IllSg--illative singular; Inf--infinitive; NSg--nominative singular; NPl--nominative plural; PastPple--impersonal past participle; PPl--partitive plural; PSg--partitive singular


Ariste, P. 1931, Lisateateid Koiva murrete haalikuloole.--EK, 175-179.

Atlas Linguarum Fennicarum. Itamerensuomalainen kielikartasto. Laanemeresoome keeleatlas. Ostseefinnischer Sprachatlas. [TEXT NOT REPRODUCIBLE IN ASCII.] ALFE 2, Helsinki 2007 (Suomalaisen Kirjallisuuden Seuran Toimituksia 800. Kotimaisten kielten tutkimuskeskuksen julkaisuja 118).

Fraenkel, E. 1960, Litauisches etymologisches Worterbuch. Band 1, Heidelberg-Gottingen.

Hagemeister, A. v. 1849, Mahrz Sahrum, der letzte Live in der Umgebung Wenden's.--Das Inland. Eine Wochenschrift fur Liv-, Est- und Kurlands Geschichte, Geographie, Statistik und Literatur 24 1849, 787-791.

Hupel, A. W. 1774, Topographische Nachrichten von Lief- und Ehstland. Erster Band, Riga.

--1777, Topographische Nachrichten von Lief- und Ehstland. Zweite Band, Riga.

Karulis, K. 1992, Latviesu etimologijas vardnica divos sejumos, Riga.

Kettunen, L. 1912, Lautgeschichtliche Untersuchung uber den kodaferschen Dialekt II. Vokalismus. Akademische Abhandlung, Helsinki [Reprinted as: Lauri Kettunen, Lautgeschichtliche Darstellung iiber den Vokalismus des kodaferschen Dialekts mit Berucksichtigung anderer estnischen Mundarten, Helsinki 1914 (MSFOu XXXIV)].

--1913, Lautgeschichtliche Untersuchung iiber den kodaferschen Dialekt, Helsinki (MSFOu XXXIII).

--1938, Livisches Worterbuch mit grammatischer Einleitung, Helsinki (LSFU V).

--1962, Eestin kielen aannehistoria. Kolmas painos, Helsinki (SKST 156).

Kont, K. 1954, Leivu murdetekste (Manuscript: KKI MT 154.)

Kruse, Fr. 1846, Ur-Geschichte des Esthnischen Volksstammes und der Kaiserlich Russischen Ostseeprovinzen Liv-, Esth- und Curland uberhaupt, bis zur Einfuhrung der christlichen Religion, Moskau.

Must, M., Univere, A. 2002, Pohjaeesti keskmurre. Haalikulisi ja morfoloogilisi peajooni, Tallinn (Eesti Keele Instituudi toimetised 10).

Niilus, V. 1935, Leivu murret. Haalikulooline ulevaade (Manuscript: AES 179).

--1936, Leivu (Koiva) murde h-haaliku arengust.--EK, 36-41.

Posti, L. 1942, Grundzuge der livischen Lautgeschichte, Helsinki (MSFOu LXXXV).

Rudzite, M. 1964, Latviesu dialektologija, Riga.

--1993, Latviesu valodas vesturiska fonetika, Riga.

Suhonen, S. 1989, F. J. Wiedemann ja Heikki Ojansuu leivun murteen tutkijoina.--Ferdinand Johann Wiedemannin muisto. Suomalais-virolainen Wiedemann-seminaari Helsingissa 5.-6.10.1987, Helsinki (Castrenianumin toimitteita 31), 39-52.

Vaba, L. 1997, Uurimusi lati-eesti keelesuhetest, Tallinn--Tampere.

Vasmer, M., Russisches etymologisches Worterbuch 1-4, Heidelberg 1950-1958.

Viitso, T.-R. 2003, Rise and Development of the Estonian Language.--Estonian Language, Tallinn (Linguistica Uralica. Supplementary Series / Volume 1), 130-230.

--2007, Survey of Previuos Research on Livonian Prosody.--LU XLIII, 23-28.

Vaari, E. 1959, Liivi keele uurimise ajaloost.--ESA V, 190-225.

Wiedemann, F. J. 1869, Die Ehsteninseln in den lettischen Kirchspielen Marienburg und Schwanenburg in Livland. Ein Nachtrag zu den Artikel des verstorbenen Akademiker Sjogren vom 11. Juli 184, "Zur Ethnographie Livlands".--Bulletin de l'Academie Imperiale des Sciences de St.-Petersbourg XIII, St.-Petersbourg, 497-524.

Winkler, E. 1999, Katkeintonatsiooni tekkimisest laanemeresoome keeltes.--Odagumeresoomo veeremaaq. Laanemeresoome perifeeriad, Voro (Voro Instituudi Toimbtiseq 6), 201-206.

Zinkevicius, Z. 1978, Lietuviu kalbos dialektologija, Vilnius.

--1980, Lietuviu kalbos istorine gramatika I. Ivadas. Istorine fonetika. Daiktavardziu linksiavimas, Vilnius.


(1) Seppo Suhonen (1989) has systematically compared Ferdinand Johann Wiedemann's and Heikki Ojansuu's notations from 1868 (cf. Wiedemann 1869) and 1911 respectively, giving thus an overview of somewhat older states of certain Leivu features under discussion here.

(2) In Leivu, a stressed syllable of Q1 is short, i.e. it ends in (a) a full-short or half-long (e.g. e, o) monophthong or (b) a diphthong whose first component is super-short (i, u, or u) and the second component is full-short or half-long; elsewhere the stressed syllable is long. A long stressed syllable is of Q2 if (1) its vocalic nucleus contains either (a) a full-short vowel (e.g. i, u, a), that can be preceded by a supershort i, u, or u and/or followed by a full-short monophthong and/or a consonant that is not longer than full-short, or (b) a polyphthong whose first component is not longer than half-long (e.g. e, o) and the second non-final component is not longer than half-long and the final component is not a half-long i, u, or u, and (2) its coda, if present, contains only consonants that are not longer than full-short. All other long stressed syllables are of Q3.

(3) Karl Kont (1954 : 4) has recorded from an informant from Onti the phrase livi kil'. Still it is not certain that it belonged to the local usage because as in the same text also the North-East Estonian place name Narva is mentioned, a North Estonian influence cannot be excluded.

(4) Moritz Brandis, who was born in about 1550 in Germany, was the first codifier of the Estonian Knighthood rights and wrote a chronicle of the older history of Livonia.

(5) Here Hupel's and Kruse's orthography is unchanged.

(6) This tendency goes on even in our days. Latvian linguists translate the name of the northwesternnmost Low Latvian dialect, which characteristically has several Livonian substratum features (called libiskais dialekts in Latvian and called here in section 1 as Livonoid Low Latvian) into English as Liuonian dialect, cf., the section on dialects. On the other hand, the Livonian language is often deprived of its historical Latin-based English name and called the Liu language, cf. e.g. the official English translation of the Official Language Law, section 18 (4) in the Latvian state portal

(7) Mahrz Sahrum (1799-1859), who was born in Priekuli northeast of Cesis, is nowadays introduced as the first known Latvian builder Marcis Sarums (-is and -s are Latvian nominative case endings). He built the tower of St. John's church in Cesis and restored several churches.

(8) It is possible that Vends in Cesis, like Vends in Germany, were West Slavs. Cesis, in German Wenden, is Vannu in Estonian. Additionally, there are two places in Estonia called Vannu, one in northwestern and another in eastern Estonia.

(9) Although the DOMAIN of the three distinctive quantities is the foot (Q1 and Q2 are possible only in at least a disyllabic foot, the initial syllable makes the FOCUS where all possible contrasts occur and which dictates the general structure conditions of the following syllables in the foot. Moreover, an initial syllable in any of the three quantities may occur in an up to trisyllabic foot.

(10) In East Livonian, u is always deleted after the long first component of a polyphthong. Note that in Livonian no consonant is deleted in this position.

(11) Niilus (1936: 38) reported that he had heard an informant pronouncing the intervocalic h in two words, and that another investigator had heard even more examples. As the Leivu settlement never was compact, it is possible, that the process of the loss of intervocalic * h had not yet ended for all speakers, Note that in 1868 when Wiedemann visited the Leivus, the intervocalic h was still regularly used.

(12) In Livonian, the front round vowels u and o were delabialized to i and e during the second half of the 19th century. People born in 1880's and later did not learn to use these vowels in Livonian.

(13) The sequence * eha changed to iha already before the loss of * h.

(14) Kettunen and several subsequent linguists have actually accepted the Balticists' terminology where both lexical tones and phrasal intonation are called INTONATIONS.



Tiit-Rein Viitso

University of Tartu

COPYRIGHT 2009 Estonian Academy Publishers
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2009 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Viitso, Tiit-Rein
Publication:Linguistica Uralica
Article Type:Report
Geographic Code:4EXLA
Date:Dec 1, 2009
Previous Article:Medieval Livonian county Metsepole and the historical and linguistic border between Livonians and Estonians.
Next Article:Lexical relations between Salaca Livonian and Estonian dialects.

Related Articles
Livonian orthography/[TEXT NOT REPRODUCIBLE IN ASCII].
Survey of previous research on Livonian prosody/[TEXT NOT REPRODUCIBLE IN ASCII].
Quantity in Livonian: preliminary results/[TEXT NOT REPRODUCIBLE IN ASCII].
A few more possible traces of the lost language chain of North-East Europe/[TEXT NOT REPRODUCIBLE IN ASCII].
Coasting to a different beat.
The role of duration ratios and fundamental frequency in spontaneous Livonian.
Medieval Livonian county Metsepole and the historical and linguistic border between Livonians and Estonians.
Lexical relations between Salaca Livonian and Estonian dialects.
Obscure place name elements ura and ora in South Estonian.
Heinrich von Jannau on the Livonians and Livonia language in Rosenpanter's beitrage (1828).

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |