Volume 43, No. 2 - Summer 1997
Editor of this issue: Antanas Klimas
ISSN 0024-5089
Copyright © 1997 LITUANUS Foundation, Inc.



My experience with certain "scholars" has shown that it will be easier, though not much easier, to get acceptance for my definition here of "Baits" than of "Carpathians". Here "Baits" include Prussians, Latvians, and Lithuanians in the broad sense of Baits proper (including Selonians, Curonians, etc.) and Baits by extension, Dacians and Thracians, all of whom I once called "Baltoidics", that is, peoples who have spoken some Baltoidic language (see Mayer, 1992). Here "Carpathians" include not only "Albanoidics" that is, peoples whom I have previously designated as having spoken some "Albanoidic" language (see Mayer, 1992) equally designatable as "Slavoidic" (see Mayer, 1995), that is, Albanians, Illyrians, Messapians, and Slavs, but also Rumanians. Essentially, I believe that Baits in this larger sense in early times had a certain dominance over the Carpathians. This explains why the ancestors of the Rumanians so quickly and thoroughly Romanized. It was political. It was to gain Roman protection against former Baltic masters.

The linguistic prototypes of the two groups, Baltic (here, the same as "Baltoidic") and Carpathian, started as different dialects of Proto-Indo-European that ostensibly in opposition to one another innovated in different areas. These have been the even if not deliberately mutually opposing patterns of development, the continuing separate drifts of these two groups even after centuries of continuing mutual influencings of one another which show that when Baltic started innovating in phonology, Carpathian started innovating in morphology, and which show that later when the opposite happened with Baltic innovating in morphology, Carpathian started innovating in phonology.

Baltic innovated first in phonology changing initial ks-to initial sk- and, like Germanic, reestablishing uniformity now in the reflexes of syllabic resonants, a process continued into changes started by satemizing. Thus it went from syllabic resonant uniformity, Į, r, m, ų, to reflex non-uniformity, ur, ul/9m, 9n, that is, u plus liquid versus shwa plus nasal, to reestablished unity, this time, in the reflexes, that is, to ul, ur, urn, un, a pattern remaining even after the change to more common i, ir, im, in brought on by satemizing. For Baltic this change was interpreted as phonologically conservativizing. Baltic thereafter remained essentially conservative in phonology and switched its innovating to morphology.

Carpathian innovated first in morphology creating from an active verb suffix, -sk-, by reversing its components, a non-active one, -ks-r still found as the -x- in Slavic ėxati 'to ride' (versus iskati 'to seek') and the medio-passive -h- suffix in Albanian verbs like dihet 'is known'. The Carpathian switch of innovation from morphology to phonology began with the reinterpretation of suffix formation from -sk- to -ks-, wherever it was being done for grammatical purposes and was being viewed as morphological, as, in the end, essentially more phonological, that is, the phonological process involved received special emphasis.

This possibility of dual interpretation of change was not to be found in Baltic's reversal of the identical components in initial ks- to initial sk-. This could be interpreted as only phonological, not also morphological. So we start with a strange sort of mutual oppositioning between Baltic and Carpathian. Baltic reverses every initial ks- to initial sk-, a purely phonological change. Carpathian reverses not every non-initial -sk- to non-initial -ks-t a phonological change done for morphological purposes (ks- always to sk- versus -sk- sometimes to -ks-: how opposite could one get!). Once Carpathian interpretation of -sk- to -ks- concentrated on its phonological aspect, on its involving consonants, it was then open to phonological change involving vowels. But the dual interpretation involving morphology as well was bound to be continued here. I say the earliest and, most likely, last common Carpathian purely phonological change was that of semantically less heavily marked vowels to shwa, a basically allophonic change allowed by morphological minor importance. This is something, a drift into shwa, we never find in Baltic.

We find this change in Albanian and Rumanian reflected nicely in special vocabulary occurring in these languages thought by specialists to be of, possibly, Dacian or Thracian origin like buzė/buza 'lip, edge', kėpushė/căpusă 'cattle tick/sheep tick' (with shwa represented by semicircled a in Rumanian and dieresised (umlauted) e in Albanian). Where morphologically heavily marked these vowels regain their full, clear phonemic and phonetic expression as in buza 'the lip, the edge' for both languages where the final a there is viewed as the result of adding a suffix.

In Slavic, having been heavily influenced by Baltic for centuries, though this drift toward shwa is found continuing as such in individual languages, it can be pointed out continuing in Common Slavic (in spite of opposing Baltic influence) only with considerable subtlety. In Mayer (1995) I allude to the phonetic ambiguity of the back jer, ". This ambiguity is also indictable for the front jer, '. Phonemically, back jer, ", is short u while front jer, ', is short i, Allophonically, both are 9, shwa, with the front jer,', allophone palatalized. Thus nov" gost' 'a new guest' were phonetically nov9 gost,9 with ",' pronounced as shwa and palatalized shwa.

The phonetic value of marked jers, ", ', changed with time in Slavic. Before the beginning of loss of the jers, a more positional process (e.g., jers followed by vowels other than jers) in Slavic, marked front jer, ', was given its then existing phonemic value, short i, while marked back jer, ", was given its then existing phonemic value, short H, as seen in the borrowings risti 'cross' in Finnish (from kr'st"; the final -i was added later) and (from s"to) sută 'hundred' with shwa represented by semi-circled a in Rumanian. The Rumanian word sută shows that when it was borrowed, the Slavic word's initial syllable was stressed as its counterpart was in Baltic (Lith.. šimtas). Its final vowel, now o, was not. Thus the final vowel of the Slavic word for "hundred" must have been shwa dating the borrowing as from non-late Middle Common Slavic. The emphasis on the jer, ", causing it to be realized as short n, dated from the period of Slavic opposition to Baltic in the pronunciation of that word. A comparison of forms with Lithuanian representing Baltic shows the following developments; *s9mt9/šimtas (with shwa everywhere) to s"t"įšimtas (with non-final marked " standing for short u and final unmarked " standing for shwa with Slavic -m- deleted as in word final position since its syllable was marked and given quasi-word status including rules of external sandhi. Later, at the beginning period of the losing of jers, the values of jers so marked in Slavic that permanently internal n was deleted were their long phoneme counterparts, long i for front jer,', and long y (from long ū) for back jer, ", as we find in Old Russian Igor' from Scandinavian Ingvarr and lyko 'mask'/Lithuanian lunkas showing a Slavic development from Proto-Indo-European *lak- with syllabic n to *l9nk- with shwa plus n to *lunk- (with shwa plus n changed to u plus n under Baltic influence) to *l"k-r i.e., *luk-, with n deleted in a marked syllable to lyk- with u lengthened to y, the then pronunciation of back jer, ", in specially marked syllables, similar treatment was in effect with Ingvarr to *'ngor't i.e., *ingor,9, (with o for va) to igor'. In the final stages of the losing of jers, marked jers including tense jers, i.e., jers followed by j, received local phonetic values. Tense jers everywhere except Great Russian o, e for ", ' became y, i: Ukr./Rus. molodyj/molodoj, pyj (from pij)/pe; 'young, drink'. Other marked (strong) jers changed into various vowels depending on the local languages: son/den' (Rus.), sen/den (Cz), san, dan (Serb.), s"n/den (Big.) 'sleep, day' (from s"n"/d'n''). Note Rus. dosku (ace.) from Old Rus. d"sku 'board'.

Continuing patterns of drift for less heavily marked vowels in word-final position throughout the following older periods in Common Slavic seen in contrast to clearly marked ones are these:

Pre-Law of Rising Sonority: poljom (nom. -acc. sg., gen. pl.); tod(jod) novom slovom (nom. -acc.); tos novos domus, slovārjos, mangjus (man'), jis; novă leubū ('love') bogūnįl, ketūres (nom.); tom novom domum, slovārjom, mangjum, jim, novăm leubūm, bogūnjim, ketūrems, slovārjoms, mangjums (acc.); slovesom (gen. pi.) / boge, gena (voc), slovesen (loc.) = marked (with paradigm leveling out of *slevesom (gen. pi.) and *slevesen (loc.) after *slevo to slovo).

Initial Law of Rising Sonority Changes from the loss of final nasal and syllable redefinition: poljo (nom. -acc. sg., gen. pi.); novo slovo (nom. -acc.); to novo domu; leubuvm, bogūnjijm, slovarjo, mangju, ji (acc.); sloveso (gen. pi.) / slovese (loc.) - marked.

Adjustments: polj9 (nom. -acc. sg., gen. pi.); nov9 slov9 (nom. -acc.); t9 nov9 dom9, leubuv'm (to leubuv' to leubuv,9), bogūnjij'm (to bogunjij' to bogunjiį9 and later developments), slovarj9, mangj9, j9 (acc.); sloves9 (gen. pi.). Note Early Common Slavic first palatalization of the velars k, g to č, ž yielding manžį9, etc.

Middle Law of Rising Sonority Changes from loss of final obstruents and syllable redefinition: todjo/to (nom. -acc.); to novo domu, slovārjo, manžju, ji, četūre (nom); četūrem, slovārjom, manžjum (acc. pl.).

Adjustments: tod}9įt9 (later to todj9/to) (nom. -acc.); t9 nov9 dom9, slovārj9, manžjd, j9, četūr9 (nom.); četūrėm, slovārjom, manžjūm (acc. pl.).

Note Middle Common Slavic merging of short o, a to short a, later to short o, and long ā, ō to long ā and pronunciation of long ū as long y.

Late Law of Rising Sonority Changes: četyrę with slovārję, manžję by analogy with četyrę with nasalized long e in those last words also reflecting the law of syllabic synharmonization = all hard/all soft (palatalized) syllables whose initiation, I believe, was suggested by the earlier reduction of all unmarked or less heavily marked short vowel syllables to 9/,9 or/9, i.e., to shwa/palatalized shwa. Like from -ėm, -ią from -ām.

Adjustments: četyrę (nom. and acc.) by analogy with e in non-final desinences (e.g., loc. *četyres9, attested: četyrex").

Comments: Old or new heavily marked final vowels did not change to shwa. Thus voc. and loc. -e remained. The above-depicted drift of semantically or morphologically unmarked or less heavily marked vowels to shwa clarifies what has been a mystery — how singular o- and w-stem nominatives and accusatives and plural genitives all fell together in Slavic to their Late Common Slavic attestations of ",'. Otherwise, only long nasal vowels in the accusative which were indivisible in rising sonority (i.e./ unlike long ā, long ū (y) was divisible into wu, etc.), long nasal ą, ę, could permanently retain nasalization in preventing merger with the nominative.

This phonological drift of less heavily marked vowels to shwa finds striking parallels in Modern Albanian/ particularly those of unstressed root syllables versus stressed ones shown in Çabej (1976): stressed/unstressed a/ë, i/ë, o/ë, u/ë, y/ë, in gaz/gëzoj, lireë/lëroj, dromcë/dërmoj, luftë/lëftoj, mbyl/mbvllij. All this including later Slavic shortening of final longs to shwa, a pattern reflected in Albanian and Rumanian nom. sg. feminine, marks these languages as continuing Carpathian tendencies even in the case of Rumanian which is ex-Carpathian. We find none of this kind of drift, this turning of unheavily marked vowels into shwa in Baltic which emphasizes separate developments from Proto-Indo-European on for Baltic and Carpathian.

Heavy Baltic influence has only sometimes seemed to have inhibited the changing of unmarked or less heavily marked vowels to shwa. And this only minus a traceable significant Baltic phonetic substratum. Once such a substratum is involved (like an Illyrian one), then shwa tends to develop from an unmarked phoneme a as in Russian dobr9j9 (phonemically dobraja, morphophonemically either dobraja (fern.) or dobrojo (neut.) with similar tendencies in Slovene and Rhodopian Bulgarian, both with substrata having merged short o, a to short a phonemically.

Post-Indo-European Carpathian has remained most conservative morphologically when it has been most pervasively pressured by Baltic influence. When it tended to escape it, it tended to relax activity in that direction, a clear example of this involves maintenance of the neuter. It has tended strongly to disappear with no trace in Albanian and Rumanian. In Albanian, the descendant of Illyrian more than Messapic, whose ancestral speakers had escaped any heavy Baltic influence of Thracian by having moved as inaccessibly westward as possible, the neuter is not as clearly represented as it is in Slavic. In Rumanian with no form specifically neuter a similar situation exists. Its ancestral speakers escaped any heavy Baltic influence from Dacian by Romanizing, I believe, both linguistically and politically. With this the neuter vanished almost traceless in Rumanian. Only in Slavic is it clearly and sharply retained, in Slavic, which was the most heavily Balticized of the Carpathian languages, so much so that generations of specialists have insisted that Slavic surely is a sort of Baltic, even, maybe, a degenerate sort having strayed so much by innovating. Baltic proper, with the exception of Prussian which was the most heavily influenced by Germanic and its insistent maintenance of the neuter, has tended strongly to lose the neuter. Slavic, as if for spite, has seemed to insist on maintaining it. It did so in Common Slavic days under the influence of demonstratives with fluctuating boundaries. I refer to OCS demonstratives to(žde), e(že) from, originally, *tod (+ *jod), *jod (+ *ge). Specialists have specified *tod alone as the maintainer of the neuter, especially its desinence -o (nom. -acc. sg.). And they usually do not cite its combination sometimes with *jod as the real source of longer maintenance of final -d (which made possible the ultimate pronunciation of final neuter -o in the pronoun) than final -s or -m (i.e., final nasal).

We find t" for masc. nom. -acc. sg. Though there may have been parallel combinations *tos + *jos, *tom + *jom, neither kept final -s, -m yielding something like *toš' or *tomlj'. All we find is t"j' and its various reflexes. The cause of the existence of only the neuter form is the seemingly constant role in Slavic of almost exclusively the neuter modifier as adverb. This made special words of neuter *todjod, *jodge as opposed to adjectivally functioning *tod *jod, *jod ge with different boundaries. Hence the fluctuating boundaries of *tod(jod), *jod(ge). And with any consideration of later *t9 *slov9/*todj9 *slov9 with *todj9 easily analyzable as *tod plus *j9 the o marking the neuter decisively won out over then gender ambiguous 9 and spread to the substantive.

Slavic seems to have chosen to maintain a clear neuter in spite of any Baltic influence against this. But even with the greatest morphological conservatism maintained in Slavic, Carpathian generally has continued a tendency toward conservatism even despite the language exchange wrought by ancestors of speakers of Rumanian. Rumanian, despite its loss of the neuter, an innovation paralleled by that in French, Italian, Spanish, Portuguese, and the rest of Modern Romance, in opposition to French, Italian, Spanish, Portuguese, and the rest of Modern Romance has, in its way, kept oblique gen. -dat. separate from direct nom. -acc. which I consider a conservative feature via the influence of Carpathian substrate drift on Rumanian (a feature lost in Modern Bulgarian and Macedonian which were otherwise influenced).

Post-PIE Carpathian morphological conservatism has been compensated for by its phonetic innovation, especially in unmarked syllables. This has been entirely in opposition to Baltic drift whose phonological conservatism has been strongly encouraged by its continuing morphological innovation which made its speakers much more syllable conscious and thereby prevented them form viewing any syllable as unmarked and via this made particularly susceptible to phonetic leveling and related changes.

Thus, finally, Baltic has remained essentially Baltic and Carpathian Slavic has remained essentially Carpathian despite mutual influencing. Was the Russian writer Gogol' experiencing some form of Slavic original homeland déjà vu when he located the source of the problems in his story, "The Terrible Vengeance", in the Carpathian Mountains?


Çabej, Eqrem (1976). Hyrjeënė historinë e gjuhës shqipe: fonetika historike e shqipës, Katedra e gjuhës shqipe, Tirana, p. 226.
Mayer, Harvey E. (1992). "Dacian and Thracian aš Southern Baltoidic", Lituanus 38.2:24-30.
Mayer, Harvey E. (1995). "Baltoidic versus Slavoidic Drift", submitted to Lituanus.