Volume 42, No.2 - Summer 1996
Editor of this issue: Robert A. Vitas
ISSN 0024-5089
Copyright © 1996 LITUANUS Foundation, Inc.



From a Common Baltic dialect cluster arisen from Indo-European via the following changes: initial ks-to initial sk-and the merger of palatal and ruki law reflexes to ś, ź the following further subdivisions surfaced. East Baltic versus West Baltic with West characterized by s, z versus East characterized by š, ž from Common ś, ź matched by West marked by š, ž versus East marked by s', z' from Common sj, zj. Both East and West Baltic agreed on š, ž from Common Baltic śj, źj. There also arose North Baltic versus South Baltic with South characterized by the merger of long ō with long ā versus North marked by the continuing separation of long ō from long ā, a conservatism, matched by the loss of final n from the postpositive -en 'in', a form having originated in East Baltic. All Baltic agreed on Tjē, Tje to Tē, Te (T = any consonant) and the merger of short o and a. Morphologically, East Baltic tended toward analysis versus West Baltic's favoring fusion.

Further differences involved South Baltic's conservatism against North Baltic's rising sonority innovations in long vowels and diphthongs in i and n with extremism growing in a northerly direction. Thus, North Baltic long o diphthongizes to uo. South Baltic matches this with its long ē tending toward ie, a change which we can see in Prussian accusatives, mien, tien, sien 'me, thee, oneself from *mēn, *tēn, *sēn (and, in Lithuanian, more northerly Žemaitian general ie from long ē), when Prussian long ē does not stay changed to ei (as in seyr 'heart'), a diphthong with falling sonority. Otherwise, North Baltic tends to continue the rising sonority trend which it started with the loss of final n in the postposition, -en (to -e), with the following continuing changes of diphthongs in n and i.

Lith. ei to ie matched by ai to *ia to ie; in, un. en, an to į, ų. ę. ą to ī, ū, ē, ā, etc.

Latv. ei to ie matched by ai to *ia to ie; in, un. en, an to į, ų, ę, ą to ī, ų, ie, ua to uo, etc.

These changes are found more frequently in Latvian than in Lithuanian. Forms not showing them are called "Curonianisms" sometimes in Lithuanian and almost always in Latvian.

I have called Lithuanian "East Baltic" and Prussian and Latvian "West Baltic". I now emphasize my agreement with Mažiulis (1988, 87) in his calling Curonian and Selonian "West Baltic" as well. Here, I also name Zemgalian "West Baltic". Also I now call not only Prussian, but also Curonian, Selonian, Zemgalian, Thracian, and Dacian "South Baltic" and estimate the northern boundary of South Baltic to have reached roughly midway between Riga and Saulkrasti running from west to east. Finally, I label Thracian and Dacian as East Baltic. My arguments for these decisions are the following with support drawn from Duridanov (1969).2

The South Baltic features evident in numerous examples from Prussian, Curonian, Thracian, and Dacian involving diphthongs with falling sonority, ei, en, in, un, an, ai, are the following with examples furnished for Thracian and Dacian with supporting data from the rest of Baltic if available. Dac. Preides, Prus. Preyd-azare l Latv. pride 'pine'. Thracian *sunketa, Dac. or Thrac. sunkētēnē / Latv. sūkala 'drop'. Thrac. Brentopara, OCur. Brendicke, Prus. braydis I Lith. Bríed-upis, Latv. briêdis 'deer'.

 Thrac. Mentēs, Prus. Mente (personal name).

Thrac. Seietovien(us), Prus. seyte / Lith. sietuvà, Latv. sietava, 'river deep spot'.

Dac. Stendai, Prus. San-stangen (different grade and suffix) (lake name).

Thrac. Spindeēnē / Latv. spîdêt 'shine'.

Thrac. Kuntis, Purs. Kunte (personal name).

Thrac. Strambai, Prus. strambo / Latv. strùobs (from *stran-b-) 'reed'.

Dac. Karpudaimon, Prus. Deyme (different grade) (river name).

Thrac. Gaidre(a)s / Lith. gidras 'clear', Latv. dziedrs 'azure blue'.

Dac. Spand- in Rum. spînz I Latv. spuôds 'shiny'.

Thrac. Kintis, Prus. Jo-kintis / Lith. kęsti, Latv. cìest (different grade) 'suffer'.

Dac. Grandeton, Prus. grandico / Latv. gruõdi 'beams'.

Thrac. Brinkazis / latv. Brĩkūsis (place name).

Thracian and Dacian do not drift from this pattern of falling sonority diphthongs which marks them as South Baltic.3

The same features marking Lithuanian as East Baltic are not as unambiguously portrayed, indisputable, or obvious in available Thracian and Dacian materials. Yet these following two items justify designating them as East Baltic. The Rumanian word juvete 'small fish' which Duridanov (1969, 95) implies as derived from Dacian comparing it with Lith. žuvìs, -is, (nom. pl. žùves) and Latv. dial, zuva 'fish' with initial ž- (written j-) rather than initial z- can come only from East Baltic either as an unlikely borrowing into Dacian or as an original Dacian word. This is unambiguous phonetic evidence marking Dacian as East Baltic. The item which, most likely, does the same for Thracian is morphological involving zero grade rather than full grade. It appears in the word for "pine"; written pusigon and listed by Duridanov as pusis. The West Baltic words for "pine" are either preide (appearing as preyde in Prussian or priedē in Latvian) or e- grade peusē (appearing only in Prussian). The zero grade correspondence of the second word occurs, otherwise, only in East Baltic Lithuanian where it seems to have had special significance. Lithuanians have kept place names with the South Baltic Prussian form of preidē fundamentally unchanged with falling sonority diphthong ei in the place names where they had found them (and preidē was not native to Lithuanian as shown by príedė borrowed from Latvian pride into Lithuanian) like Priedžių káimas and Preidžiai, a procedure common in names of that sort which had been given by West Baits before the arrival in Lithuania of East Balt Lithuanians. But if they had found any West Baltic e-grade forms matching Prussian peuse 'pine' as in Prussian Peusebalten, and, surely, these must have existed, Lithuanians East Balticized them into zero grade ones with š, the East Baltic reflex of the Indo-European palatal k', giving the forms Pušynės káimas, Pušinė, Pušyno káimas, Pušinėlis, and even Žem. Pušynė, Pušynytė based on Lith. pušýnas 'pine forest'. For me this is evidence enough for regarding Thracian, too, as East Baltic. With these facts in mind, Duridanov's lists of Dacian and Thracian forms should be revised to show their East Baltic character, thus, for Dacian "fish" we should find not the West Baltic shape zuv-, but the East Baltic one, žuv-. Note the following list of Dacian and Thracian forms accompanied by Lithuanian counterparts to the right and normalized to show their true East Baltic character. Thracian: bružas: brùžas, dižas, paišą: paĩšąs, prauš-, pušis: pušìs, tarš-: taršýti, ž(i)burul-: žiburýs, žilma: žilmà, keršas: kéršas, traušas, šēkas: šėkas, švit-: švitėti, žilas: žìlas, žvaka: žvãkė, ž(v)ēris: žvėrìs; Dacian: žudąs, berža: béržas, šuka: šùkė, želmas: želmuõ, -eñs; žuv-: žuvìs; Dacian or Thracian place name: Kratiškara: Kratìškiai.

The Thracian and Dacian diminutive suffix -už- is ambiguous since the correspondences are with Latvian which the following normalized examples show. Dacian: Katuža: Latv. place names Katuža-visimins, Katužs, Naruža (Rum. Nãruja): Latv. river name Naruža, Thracian: Kartuža: Latv. place name Kartuži. In this suffix (Lith -užė as in mergùžė) the Indo-European source is *-ug'j- which changed first to Common Baltic -*uźj- and then to Early Common East Baltic *-užj- I Early Common West Baltic *-uzj- and finally to Lith. and Latv. -už-. Other similar cases of ambiguity are Dacian žūras: Lith. Žiūrà, žiūrėti, Latv. žũrêt and Thracian laža, -as: laža, Lith. pã-lažas.

East Baltic Thracian and Dacian (as well as Lithuanian) show Latvian influence. This attests to Latvian's having been the most easterly of West Baltic languages. In Thracian we find the lone example of ge to ze like Latvian ge to dze in zerva 'crane' in the place name Zervae: Latv. dzrve. This is hardly native since Thracian keeps ge (3x), gi (2x), ke (5x), ki (4x) matching Dacian ge (5x), gi (5x), ke (5x), ki (1x) against ki to ci also in one lone example: Mamutzis: OLatv. Mamwtze. In the Dacian example, the suffix -uc- seems directly borrowed from Latvian. Thracian zerva with z, not dz (according to the spelling) may show an indirect borrowing via Dacian, which, remaining more northerly, surely had more sustained contact with North Baltic.

In demonstration of this, Dacian, rather than Thracian, shows initial sk- from initial ks-, the clear, unique sign of Baltism in two roots, skaudus 'painful', skuja 'pine' attested as Scaugdae (from *skaud-gdae) and Skuanes (versus initial ks- to x-in Russ. xudoj 'bad', xvoja 'pine needles') and the change tl to kl typical of North Baltic in the suffix -ukl- in Brucla: Latv. brūkle and Genukla: ELatv. dzanuklis, Lith. ganyklà, stebùklas, žénklas / Prus. ebsentliuns with tl intact paralleled buy dl in addle 'pine' / Lith. glė with dl to gl (but note Prus. kl from tl in gurcle 'throat'). Yet Thracian, through having been far away to the south, shows one more West Balticism, initial sr-to initial str- in strūna 'stream' (and also in Strymon, strambas) matched by Lith. river name Strūna with the same West Baltic feature found also in Lith. Dial, straumuõ, a feature resisted in Lith. sraunà, srūti, sravėti. This West Balticizing of initial sr- to initial str- is also found in Dacian *strunga 'milking room' corresponding with Lith. 's borrowing from Latv. strúoga matched by an East Balticized doublet srúoga!

Otherwise, we find evidence for East Baltic influence on West Baltic and North Baltic influence on South Baltic and a possible new piece of evidence of a surviving East Baltic declensional desinence in š!

I believe that East Baltic had been strongly influenced both phonologically and morphologically in very early times by languages to the east of it including Iranian and Indie and, possibly, some form of Uralic. This resulted in the following. 1. The making of joint ruki law and palatal reflexes compact: ś, ź to š. ž. 2. From sj,. zj not š, ž, but s', z'. 3. The keeping of initial sr- as in Indic and Iranian. (West Baltic changed it to str-.) 4. The adaptation of the Iranian (and Indie) form of the o-stem ablative singular, -ād (rather than -ōd) with final d dropped yielding, originally, long -ā. 5. The postpositioning of the e-grade morpheme for "in", originally, *-en, with inessive, locative meaning. These changes made East Baits more syllable-conscious than West Balts which resulted in one major difference in morphological technique between East Baltic and West Baltic. East Baltic has been comparatively analytical. West Baltic has been comparatively fusional.

East Baltic influence on West Baltic Latvian, but not Prussian which was too far away, that is, before the intrusion of East Baltic Lithuanian into originally West Baltic territory (modern Lithuania), involved the following changes in its morphology. 1. Long -ā as the o-stem ablative singular used, ultimately, for the genitive instead of the original ending, -as, maintained in Prussian. 2. Postposition e-grade -en to mean "locative".

North Baltic influence on South Baltic involved the loss of final n in the postposition -en yielding the lone example of it in Thracian given by Duridanov (1969:86, 98), midne "in the place", in the expression midne Potelense. This midne, I believe, originated as *mītinen or *mītenen. The loss of -n in the postpositive was due to North Baltic influence. Duridanov believes that the form exactly matches Latvian mītne from an earlier *mītinē, -enē. But West Baltic is morphologically fusional. This we see in Latvian where we find o-stem long ā from a plus e, i-stem long i from i plus e, etc. in the locative. East Baltic Thracian, like Lithuanian, was surely analytical with locative -e kept consistently discrete. So, though the Latvian form can be only e-stem, the Thracian form, like any conceivable Lithuanian counterpart, could not have been e-stem. It could have been, originally, a consonant stem or an o-stem, something like nom. *mītinas, -enas: loc. *mītin-e, -en-e with nom. desinence vowel -a- dropped, suppleted by loc. -e from -en. (If it had been e-stem, we would have *midneie from earlier *mīdnēie with the suffix resembling Lithuanian -ėje as in žemėje 'on earth'.)4

The East Baltic ablative-genitive suffix, long -ā, is found South Balticized to long -ō in the Dacian Decebalus per Scorilo 'Decebal, son of *Skorilus' (the true nominative form here, I believe, despite the use of Skorilo for that function by foreigners) as we find it in Lithuanian, e.g., draũgo 'of the friend'. But the word order matching that of Albanian shows a sort of early Balkanism like the conservatism of anti-Lidén law initial vr-'s reflex, br-, in Thracian briza from earlier *vriza. This keeping of initial v- followed by r is not to be found in Baltic north of the Carpathians any more than it is in Slavic, a Carpathian language, originally, (like Albanian) (see Mayer, 1996), which I have called "Northern Albanoidic" elsewhere (Mayer, 1992). But Albanian, presently a Balkan language, has also, contrary to Lidén's law, maintained initial vr- (e.g., vrasje 'murder'), the reflex of Indo-European wr-.5

If the transcription of the somewhat shaky Thracian form in a Mysian inscription, instr. (dat.) pl., cited in Duridanov (1969) as -aiś is to be understood to stand for -aiš (with ś = š and not just any s), (see Duridanov 1969, 97), then we have an expected East Baltic ruki law š in a declensional desinence, something which, I believe, was West Balticized out of Lithuanian (see Mayer, 1994b). The Thracian form labeled Mysian by Duridanov allows us to posit ruki law š as an original item in all Lithuanian desinences like *sūnùš, *sūnaũš, *sūnumíš, etc.6

We may conclude here by emphasizing the following. 1. All South Baltic languages, Thracian, Dacian, Prussian, Curonian, Selonian, and Zemgalian, are extinct. 2. Two South Baltic languages, Thracian and Dacian, were East Baltic, a fact emphasized by their very names which have reflections either only in Lithuanian like Dakai 'Dacians' tied with Lith. dàknoti 'to make disorderly', dakãnyti 'to drive away' (and are never in Prussian), or are less numerous in Latvian, i.e., more numerous in Lithuanian, as Thrakoi Thracians' with Lith. trãkas 'mad, furious' or 'open place for burning' accompanied by allomorphs in two other grades, trėkti' 'to ruin', trìkti 'to become furious' with Latvian correspondences covering fewer grades, traks 'mad', trękns 'powerful'. The meanings of the names Dakoi, Thrakoi 'Dacians, Thracians' seem to have the common element of interpretation of "frantic, unsettled" which is reminiscent of goidelic from Welsh gwyddel 'wild man' as a term meaning "Irish". 3. South Baltic territory included all of Lithuania and much of Latvia. So place names there should be seen in that context. With East Baits having left traces near Moscow, West Baits near Germany, North Baits near Scandinavia, and South Baits in the Balkans and Asia Minor, we can say that Lithuanians and fellow Baits really got around!



1 See my sister article, "North Baltic" (Mayer, 1994a) where I classify Lithuanian-Latvian agreements as "North Baltic".
2 See Mayer (1993).
3 Dacian ie from e (according to Duridanov), a possible early Balkanism found also in Albanian, might really or originally have been from long ē (as it, most likely, was in Thracian) (Note Dac. Dierna: Jatvingian Dērna, Dac. Degis: Diegis transcribed as Diēgis with Greek symbols.) If so, then the rising sonority diphthong here shows a development similar to Prussian mien, tien, sien from *mēn, *tēn, *sēn with long ē. This seems definitely confirmed by Dacian certie matched with Lithuanian kertė where Dac. final ie is matched by Lith. final long ė. Note -ien- for ēn- in Thrac. Seietovien(us) versus Spindeēnē.
4 Duridanov (1969) mentions the possible tie to OCS mĕsto 'place' of Thrac. midne, Latv. mītne, mìets 'beanpole', Lith. mitas 'fence pole'. A possible explanation of midne with an ē-stem loc. -ē would be a misspelled -ēj like Lith. ėj in žmėj.
5 See also Klimas (1967) who insisted that Baltic and Slavic be understood as separate families of Indo-European. Otherwise, unlike Duridanov, I interpret the Dacian abl. -gen. -o in Scorilo as derived not directly from long -ō, but from long patterned originally by East Baits after Iranian and Indie long here, their expected reflex of original Indo-European long ō. Dacian, in characteristic South Baltic fashion, blended long ō and long ã into one phoneme which here appears, I believe, as long ō. This Dacian form happens to resemble its Lithuanian counterpart, also long ō now, a form that exactly matches its Latvian counterpart, ã from an original long ã. Thus, even though, under South Baltic influence, Lith. original long ã here as elsewhere has, like Dac. long ã, been changed to long ō, unlike Dac.'s, Lith.'s original long ō, like Latv.'s, remains distinct via change to uo. Therefore, any depiction in Duridanov (1976, 108) of Thracian, like Dacian also South Baltic, as showing separate developments for reflexes of Indo-European long o and a must be viewed with suspicion. Was it just long ō that developed to long ū? Or did the joint reflex of both long ō and long ã, as in Prussian, another South Balatic language, mutate to long ū after labials and velars and was this long ū spread elsewhere by analogy? Finally, Duridanov (1969, 98) is mistaken in calling Dacian and Thracian more or less generalized linking vowel u in compounds like Desu-daba and Tranu-para gen. pl. formations "exactly" corresponding to those in Latvian and Lithuanian as in Desu-grava and Žąsū-gala. Aside from the North Balticism final -e for final -en in one example in Thracian midne, Thracian and Dacian stay true to the South Baltic falling sonority pattern of keeping the n after the vowel in diphthongs. Latvian u in these compounds can not with certainty be identified as a genitive plural. Nor can the corresponding Lithuanian long ū whose length can be attributed to something else, say, compensatory lengthening for loss of something other than nasalization. The following set of examples from Prussian correspondences to these compounds furnished by Duridanov (1969, 36) shows what was happening. Dac. Kapi-dava, Kapi-sturia (here with a different linking vowel, i!), Kopu-storos (with o in Kopu- not from 6th century Slavic as Duridanov says, but really, as he alternatively says, from Late Dacian, and I say, not necessarily Late, where, as in Prussian, also a South Baltic language, no reflex of o ir a either length differed from one another in the same context) Prus. Auctacops, Auctu-kape, Awctum-kape. Note the different vowels, a and u, in Prussian as well. This we find matched by Latvian examples on the same page: Kapu-mäjas, Kapa-kalns. The Prussian alternation, Auct-kape / Awctum-kape gives some clue to what must have been happening. Lithuanians copying this kind of name-giving seem to have lengthened the u to compensate for the loss of m in something like West-Balticized aukstum- (from aukštùmas). So, from, say, *Kapum-ëžeras to *Kapū-žeras the u was lengthened to compensate for the lost m. Later, to give sense to the compound, the long ū was reinterpreted as ų, a "genitive plural". Note — in Mayer (1996) I identify Thracian and Dacian as Baltic.
6 A possibly ambiguous further piece of evidence proving that Thracian was East Baltic is an item said to be from Thracian in a Middle Bulgarian text, Knišava 'a dug out place', the Lithuanian counterpart verb, knìsti (-sù /-s(i)aũ) 'to dig', has been West Balticized out of its expected ruki law relfex, š. The Bulgarian text form has preserved this East Baltic reflex if it does not represent the Slavic reflex of this law, x, altered to š via the first palatization of the velars (*knix- = -ēva to knišava) or a stem formed from sj (knis- + -java to knišava) or even a stem formed from š (from I.E. palatal k') + j (with *knišjava to knišava). Duridanov (1969) mentions this form as typical for river names and cites Nišava formed from the "ancient" name Naissos (now Niš), spelled Naissus in Russu (1967, 239-241) who designates that name as "Thraco-Dacian", no doubt, in this instance, correctly enough since its location in Upper Moesia, equidistant from Thrace and Dacia, shows that it could have been named by either Thracians or Dacians.


Duridanov, I. (196). Die thrakisch- und dakisch-baltischen Sprachbeziehungen, in Linguistique balkanique, XIII, 2, Bulgarian Academy of Sciences, Sophia.
—————— (1976). Ezik"t na trakite. Nauka i izkustvo, Sophia.
Klimas, A. (1967). "Balto-Slavic, or Baltic and Slavic? (The Relationship of Baltic and Slavic Languages)", Lituanus, 13:2.5-37.
Meyer, H. (1992). "Dacian and Thracian as Southern Baltoidic", Lituanus, 38:2.24-30.
———— (1993). "West Baltic Latvian/East Baltic Lithuanian", Lituanus, 39:3.57-62.
———— (1994a). "North Baltic", Lituanus, 40:1.5-13.
———— (1994b). "Baltic Diglossia in Lithuanian", Lituanus, 40:3.51-56.
———— (1996). "Balts and Carpathians", submitted to Lituanus.
Mažiulis, V. (1998). Prūsų kalbos etimologijos žodynas, I, A-H, Mokslas. Vilnius.
Russu, I.I. (1967). Limbą traco-dacilor, Editura Stiintifică, Bucharest.