Read Part I http://vetiver.weblog.ro/2012/01/19/mihai-vinereanu-the-origin-of-the-romanian-language-i/#axzz1jki4yVPD and Part III http://vetiver.weblog.ro/2012/01/19/mihai-vinereanu-the-origin-of-the-romanian-language-iii/#axzz1jki4yVPD
In what follows I will discuss a few centum elements in Slavic and Baltic languages which have correspondents in Romanian found in Golab (1972) such as OCS gleznŭ “Knöchel”, Russian glezna “tibia”, Polish glozna, Slov. glezeny “ankle”, Lithuanian žlezan “ankle” from PIE *gel-, gleg- “to be or become round; something round” (IEW, 357-58); cf. Romanian gleznă “ankle”. Walde-Pokorny derives the Slavic forms from this Proto-Indo-European root. From PIE *akmo- “stone” (IEW, 18), Baltic languages have pairs of cetum/satem forms: cf. Lithuanian akmus “stone”, ašmenys “edge”, Latvian asmenas “edge, precipice”, while Slavic languages have only the centum type forms: cf. OCS kamy, Russian kamai “stone”. Cognates in centum languages: Greek ákmon “anvil”, Phrygian place-name Akmonia and perhaps Romanian ocnă “salt mine” and satem cognates: Sanskrit asman “stone”, Avestan asman “stone”.
In several other cases both Slavic and Baltic languages have centum/satem pairs. From PIE *gherdh- “to enclose”, ghordh-“fence, enclosure” (IEW, 444), we have the following forms:
1. satem type: Lithuanian žardas “a wooden construction”, Latvian zards “horse enclosure”, Old Prussian sards “id”, OCS žrŭdŭ “hen coop”, Russian žerd “id”.
2. centum type: Lithuanian gardas “enclosure for animals, fortress”, OCS gorditi “to enclosure, to build”, graditi “to build”, gradŭ “city”, Russian gorod “city”, Polish grad “city”, etc.
Furthermore, Golab shows that pre-Slavic place-names in Poland have centum forms, while other have the root ap- “water, river”; cf. Romanian apă “water, river”. The root is quite frequent in the Balkan region place-names, as well as in some Celtic regions. Thus in Gallia, we have: Geld-apa, Arn-apa, Len-apa, Ol-epa, Man-apia, Appa (cf. Holder, Alt-celtischer…), Gaulish Apa-va almost identical to Pannonian Ape-va (cf. Holder, vol. 1), in Greece: Apia, In-ōpos (river names), Api-don, Api-danos (place-names). An-apos, river-name is attested in Greece (Tucydides, 2, 82) and in Sicily as well (Titus Livius, 24, 36, 2; Tucydides, 6, 96, 3: 7, 8, 3; Dyodorus of Sicily, 15, 13, 5), Ap-sus, river in Southern Pannonia (Krahe, ZONF, 20, 1931), Sald-apa, in Dacia and many others. Therefore, there is no doubt that all these place-names, mentioned by Golab, are of Thraco-Illyrian origin. They come from Northern Thraco-Dacians who brought a great contribution to the old Slavic civilization. Regarding other (perhaps more recent) Old Slavic loan-words from other Indo-European languages, Bernstein (p. 87) thinks that OCS sluga “servant”, braga “a kind of ale”, ljutŭ “sour, cruel” are loan-words from Old Irish slog, sluag “crowd, army” Irish braich “malt”, Welsh llid “malice” < Proto-Celtic *lūdu. For sluga, there is a cognate in Lithuanian slauga “helper, servant”. I should say that in Old Church Slavonic and Romanian, the forms are absolutely identical; cf. Romanian slugă “servant”, bragă “kind of ale, malt” and iute “1. quick, energetic; 2. spicy” from an older *liute. Bernstein continues saying that these words are etymologizing well in Celtic languages, but we may say same thing about Romanian, where both slugă and iute have many derivatives. We have similar situations in the case with other loanwords in Old Church Slavonic: vino “wine” which he thinks that they are of Gothic origin or popŭ “priest”, pila “saw” vitez’iu “brave” loanwords from western Germanic dialects, but he does not specify which those Germanic dialects might be. Regarding popŭ “priest” it has a correspondent in Latin popa “a priest in charge with sacrifices in old Roman religion”. There are also Latin loanwords into Old Church Slavonic. Most of them are Christian Church terms: oltarŭ “altar”, koleda “Christmas carol ”, poganŭ “pagan”, as well kanopl’a < Latin *canapis “hemp”. All these loanwords have their correspondents in Romanian with the same meaning as in Slavic. There are reasons to believe that all these are loanwords from Old Romanian of 7-8 centuries into Old Church Slavonic after the Slavic tribes settled in the Balkan region during this period (see vin, popă, pilă, viteaz, as well as altar, colindă, păgân and cânepă).
Some of loanwords into Common Slavic are considered of Iranian origin, but many of them are, in fact, borrowed from Romanian: rajŭ, toporŭ, mogyla, vatra, while bogŭ “god” and kurŭ “rooster” have no correspondents in Romanian and they may be of Iranian origin. Bernstein never mentions the Romanian (or Albanian) correspondents, even when the relationship between Slavic and Romanian forms is more than obvious. Thus, he associates Old Slavic vatra with Avestan athaurvan “sacred fire” and Sanskrit atharvan “priest of fire cult”, but he ignores the fact that there are identical forms with the same meaning in Romanian and Albanian: cf. Romanian vatră “hearth, fire” Albanian vatrë “id”. Regarding Old Slavic sъto “hundred”, Vasmer and other Slavists believe that it is of Celtic origin, namely from Old Irish, but Common Slavic speakers were never in any contact with any Irish people. Even phonologically speaking Old Irish cet “hundred” cannot the etymon for Old Slavic sъto. It is well known the fact that PIE *ŭ turned into ъ, ь in Common Slavic. In other words Common Slavic borrowed a form *sŭta, when it still had the short Proto-Indo-European vowel *ŭ. Therefore, Romanian sută “hundred” cannot be a loanword from Old Church Slavonic sъto as all linguists believe, borrowed after the arrival of Slavs into the Balkan region during the 7th and 8th centuries AD, but Slavic sъto was borrowed from Thraco-Dacian at earlier time, sometimes before the first millennium AD (see sută).
The Relationship between Thraco-Dacian and Romanian
It was necessary to show the position of Thraco-Dacian with other Indo-European related languages. I have shown that Thraco-Dacian and Illyrian were related to Italic (especially Oscan and Umbrian) and Celtic (especially Continental Celtic). To the east they had have the Balto-Slavic group which has many loanwords from Thraco-Dacian. As I mentioned above, this language had some important phonological features in common with Italic and Celtic languages and, to a lesser extend, it shared some (other) features with the Balto-Slavic group.
To reconstruct the phonological features of Thraco-Dacian language, the real ancestor of Romanian, I have compared the Romanian lexicon and Thraco-Dacian names and glosses and with cognates from other Indo-European languages in connection with Proto-Indo-European roots as in Walde (LEW) and Walde-Pokorny (IEW). Although the German linguists did not use Romanian for their reconstructed roots, Romanian words match them very well. In the beginning, I paid a special attention to the Romanian lexical items considered already to be of Thraco-Dacian origin, later on there were taken into consideration those with uncertain or unknown origin, in order to find common phonological features. In the ‘60s of the last century the German Romanist Günter Reichenkron (1966) tried a new method in Romanian historical linguistics. He analyzed 130 Romanian words of Thraco-Dacian origin, by comparison with other Indo-European languages, using Walde-Pokorny dictionary as well. Unfortunately, this new method was rejected by Romanian linguists. However, several years later, the Romanian linguist I.I. Russu, tried the same method, but it seems that he was insufficiently familiar with the field and his unsuccessful attempt was perceived by other Romanian linguists as “inadequacy” of the method itself. In fact, this is the only method in Romanian etymology since there are thousands of words with uncertain or unknown origin or wrongly attributed origin by those using a simple comparison with Latin, Slavic or with other neighboring languages.
The Phonological features of Thraco-Dacian language
Partial reconstruction of the phonological features of Thraco-Dacian was done by G. Reichenkron, I.I. Russu, V. Georgiev, C, Poghirc and Gr. Brâncuş, but all of them are far of being complete. Most of these authors used only Thraco-Dacian glosses and names and very few Romanian words. Unlike them, I have used a large number of Romanian words. On the other hand, the phonological configuration of Proto-Indo-European is well-known today . Therefore, to reconstruct Thraco-Dacian phonology, I have started from the Proto-Indo-European sounds and followed their evolution to the sounds of modern Romanian, along with Thraco-Dacian glosses, compared with other Indo-European languages. The discovery of phonological principles governing the evolution of the sounds from Proto-Indo-European to Thraco-Dacian and to Romanian was done for the first time by the author of these lines.
The vowel sequences: A general phonological characteristic of Thraco-Dacian was that Proto-Indo-European vowel sequences turned into simple vowels. It seems that they usually turned into long vowels, as in Latin. However, in some monosyllabic words the sequence was preserved. Proto-Indo-European had six vowel sequences: *ai, *oi, *ei, *au, *ou and *eu. The first three of them were not very common, *ou was not so frequent as well, while *ei and *eu were much more frequent.
The *au sequence: In some Romanian short words, this sequence was preserved. It seems that originally in Thraco-Dacian, these words were bi-syllabic with the stress on the first syllable which preserved the sequence. In Romanian, this sequence was preserved (with the two vowels separated in hiatus ) in words with uncertain or controversial origin such as auş “old man, grandfather” from PE *aueo-, *auo- “grandfather” (Lehmann, A242), auşel “a little insectivore bird” (Regulus regulus), from PIE *auei- “bird” (IEW, 86) as well as aur “gold” considered to be of Latin origin from PIE *aus-os “to be bright, gold, down” (IEW, 86); cf. Sabin ausom “gold”, Ir. or “id”, Welsh aur “id”, Albanian ar “id”, Old Prussian ausis “id”, Old Lithuanian ausas “id”, Armenian oski “id”, Tocharian A wäs “id”. The other two have also a considerable number of cognates in various Indo-European languages (see auş, auşel, aur). In other cases (longer words), PIE *au turned into /u/, perhaps a long */ū/ in Thraco-Dacian. The verb a (se) gudura “1. to fawn (upon); 2. to be happy (about dogs)” was associated with Albanian gudulis “to tickle” which is in fact cognate with Romanian a gâdila “to tickle”. Romanian a (se) gudura is cognate with Latin guadeo, gaudere “to enjoy, to be happy”,but it cannot come from Latin *gaudulare, which not attested anywhere or have any correspondent in any of the Romance languages. Both come from PIE *gāu- “to enjoy, to be happy” (IEW, 353); cf. Greek gedéō “I enjoy, I am happy”, Dorian Greek gadéō “id”.
The other sequences have a similar evolution. The sequence *ai: it was preserved in shorter words such as coică “forested hill”; cf. Albanian kojkë “id”, Old Welsh coit “forest”, Welsh coed “id”, Old Cornish cuit “id”, Breton coed “id”, all from PIE *kaito “forest, untilled land” (IEW, 521). In this case *a > o. In late Thraco-Dacian there was a general tendency of a > o and o > u (cf. river names Mureş < Maris, Olt < Alutus, Dunăre “Danube” < Donaris).
In other cases PIE *ai > e (or i) in Thraco-Dacian and preserved as such in Romanian. The noun petec “patch, a piece of fabric” derives from PIE *baita, *paita “goatskin” (IEW, 93): cf. Albanian petk “patch, a piece of fabric”.
The sequence *ou. It turned into a simple vowel: o or u as in Romanian cocoaşă “hump”, coacăză “cranberry”, cocon “child, baby”, all from PIE *kouko-s “round” < *keu- “to bend” (IEW, 588); cf. Albanian koqë “berry, any berry”.
The sequence *ei: Romanian ţep “thorn, spike”, as well as ţeapă “1. stake, point of a stake; 2. splinter” ţepos “thorny, spiky, prickly”, a înţepa “to prick, to sting; 2. to bite (about insects)” are considered of Thraco-Dacian origin (Reichenkron, 166: Poghirc, ILR, 2, 352: Brâncuş, VALR, 124). Reichenkron shows that they come from PIE *k’eipo-, *k’oipo- “pale, stick , a sharp stone or wood” (IEW, 542) where PIE *ei > e in Thraco-Illyrian; cf. Albanian thep “sharp stone”.
Romanian a leşina “to faint (away), to swoon”, leş “corpse, carcasse”, a lihni “to starve”, all have same origin, although etymologists gave them different origin (see a leşina, leş, a lihni). All originate in PIE *leik-, *leigh- “1. weak, miserable; 2. death”. In all these examples, PIE *ei > e in Thraco-Dacian language and preserved as such in Romanian.
In other cases, PIE *ei > i as in Romanian mic “small” from PIE *mei-ko-s “small” (IEW, 711).
In words considered to be of Latin origin we encounter the same evolution, as in Romanian a zice “to say, to tell”. In most cases, in Latin, the Proto-Indo-European vowel sequences turned into long vowels. Thus, Latin dico, dicere “to show, to say” < PIE *deik- “to show, to indicate” (IEW, 188).
The sequence *eu was the most frequent in Proto-Indo-European. In Romanian it shows as o (or u), a transformation inherited from Thraco-Dacian. In some cases, in short words, it was preserved as sequence slightly different as in lăun (pronounced lă-un) “a plant that grows in stagnant water” and lăunos “dirty” from PIE *leu-, *lū- dirt, to make dirty” (IEW, 681); cf. Greek lǘma “1. dirt; 2. insult, outrage”, Albanian (Tosk dialect) lum “swamp, pond”, (Gheg dialect) ljum ‘id”, Lithuanian liūnas “swamp”.
Instead, in most cases (in longer words) PIE *eu turned into o (u). Thus, Romanian broască “frog”, (where o > oa by umlaut), originates from PIE nominal from *preu-sko of *preu- “to jump, to hop” (IEW, 845-46) with cognates in Albanian, Italian (dialectal) and Germanic languages; cf. Albanian breskë “frog”, Italian (dial.) brosca “id”, Old English frosc “frog”, Old Icelandic froskr “id”.
In other cases, it turned into u, as in Romanian ciucă (var. cucă) “ridge, peak” which was borrowed into all Balkan languages. It is frequently found in place names and personal names. It originates in PIE *keu-, *keuk- “to bend, to wind, curvature” (IEW, 589) (see ciucă). The same rule applies to words considered to be of Latin origin, such as a luci “to shine, to gleam” from Latin lucio, lucire “to shine, to gleam” (see a luci).
The Vowels: Although Proto-Indo-European had short and long vowels, the vowel quantity disappeared most probably in Late Thraco-Dacian, a phonological trait transmitted to Romanian. At a certain moment in history the quantity stopped to play a role and, short and long vowels developed in the same way.
Proto-Indo-European short and long *a: The short Proto-Indo-European vowel *a, at initial or in stressed syllable remained unchanged in Thraco-Dacian and Romanian. Romanian argea “subterranean room” was considered Thraco-Dacian since Hasdeu (Col. lui Traian, 232, 1873) from a Dacian *argilla and later in Etymologicum… is associated with Greek árgilla “subterranean house”, Old Macedonian árgella “id” and Cimmerian argill “id”). This hypothesis was adopted also by Gr. Brâncuş (VALR, 30) and I.I. Russu (Elemente, 133). All these forms come form PIE *areg- “to enclose” (IEW, 69) (see argea).
Instead, in unstressed position or at the end of the word (which is generally unstressed) PIE *a turned into ă (ə) as Hasdeu showed more than 100 years ago. For Romanian măgură “hill” he identified a PIE *mag- (Cuvente…, 288), similar to PIE *mak-, mək- (IEW, 699); cf. Greek maketa “hill” > Makedones “the ones who live on hill and mountains” (cf. IEW), Albanian magullë “id”, Neo-Greek mágoula “id” (a loanword), as well as Sardinian moγoro and Italian (Campidan dialect) moγoro “hill”. Common Slavic borrowed it from Thraco-Dacian as *magula “mount, hill” > Slavic mogyla “id” (see supra). Romanian noun vatră “hearth” (Aroumanian, Megleno-Roumanian vatră “id”) < PIE *(ŭ)āter “fire, hearth”; cf. Albanian vatrë “hearth”, where (long) vowel *a was preserved in stressed syllable. Romanian vatră was borrowed into Slavic (see supra).
The Proto-Indo-European short and long vowel *e may have a different evolution depending on the phonological environment. Sometimes in stressed position PIE*e > je in Thraco-Dacian. The iota palatalized the consonant in front of it such as t, d, k, g. The phenomenon can be seen in Thraco-Dacian names and it was transmitted to Romanian. In these names, there is an oscillation in spelling either with a full stop or with a sibilant: Sabadios/Sabazios, Dierna/Tsierna, Germizara/Zermizara. This evolution of Proto-Indo-European stops was transmitted to Romanian, not only in words of Thraco-Dacian origin, but also in those considered to be of Latin origin. Thus, Romanian miere “honey” considered to come from Latin mel “honey”, from PIE *mel-it (IEW, 723); cf. Hittite milit “honey”, Greek melì “id”, Gothic meilith “id”, Armenian melr “id”, Old Irish mil “id”, Welsh, Cornish, Breton mel “id”. In fact, Albanian mjal, mjaltë “honey” has a iotacized vowel as in Romanian. As we saw, this phonological change along with other such changes took place long time ago as in god-name Sabazios or place-names such as Tsierna or Zermizara. Other times, it remained unchanged, in lexical items considered either Thraco-Dacian origin as in a legăna “to rock, to swing, to balance” from PIE *leig-, *loig- “to jump, to tremble, to swing” (IEW, 677), lespede “slab” from PIE *lep- “stone, rock” (IEW, 678) and many others. Same with other words considered to be of Latin origin such a lega “to tie, to bind, to attach” < Latin līgāre “to tie, to bind” < PIE *leig- “to tie, to bind” (IEW, 668). In some cases the Proto-Indo-European correspondent is a vowel sequence, but in this case it makes no difference, since iotacism appear after the reduction of Proto-Indo-European vowel sequences to (long) vowels.
For Al. Cihac (2, 47) and Gustav Weigand (BA, 2, 108) Romanian ceaţă “fog” is of Slavic origin; cf. OCS kaditi “to smoke”, Russian/Ukrainian čad “smoke, steam”, but most Romanian linguists believe that it derives from Latin *caecia < caecus “blind”, a hypothesis that should be rejected. Indeed, Romanian ceaţă is cognate with the Slavic forms, but it is not of Slavic origin, since we do not have a Slavic form from which it may derive. All of them derive from PIE *ked- “to smoke, to make smoke” (IEW, 103). In other words, ceaţă derives from an older *ketia, where the iotacism of e turned *k into a č.
The Proto-Indo-European short and long vowel *u remained unchanged. The noun buză “lip, edge” > PIE *bŭ-s “lip, to kiss” (IEW, 103); cf. Albanian buzë “lip”, Old Irish bus, pus “lip”, busóc, pusóc “kiss”. The root is attested in Thraco-Dacian personal names such as Byzas, Bysos, Beuzos, as well as Illyrian Buzos, Buzetius. We have the same evolution in the noun vătui “one year-old goat” from an older *vituliu, considered to be of Thraco-Dacian origin, because that it has cognates in Albanian ftuj, vëtulë “id”, as well as viţel “calf”, vită “cow, or other domestic animal”; viţel is considered of Latin origin because it has a correspondent in Latin vitulus, while vită has no cognate in either of these languages. All these words derive from PIE *ŭet- “year”, *ŭetelo “one year-old animal” (IEW, 1175).
Proto-Indo-European short and long vowel *o: In stressed position remained unchanged as in the noun boare “breath of wind, breeze; 2. aroma” (Aroumanian boră, Megleno-Roumanian boari) from PIE *bholo- „steam, fog”(IEW, 162), where o > oa, by umlaut. From the same root derives the noun abur „steam”, where Proto-Indo-European *o turned into u in unstressed position. In some cases, it may turn into an a. For a long time, Romanian gard (Aroumanian, Megleno-Roumanian gard) “fence” was considered to be of Slavic origin, namely from OCS gradŭ “city”. Later on, Poghirc (ILR, 2, 341), I.I. Russu (Thraco-Dac…, 109; Elemente…, 159) consider it to be of Thraco-Dacian origin. Furthermore, Brâncuş (VALR, 76-77) shows that Albanian dh of gardh “fence” does not reflect a Slavic loanword. In the same time, Romanian and Albanian forms do not exhibit the metathesis of the lateral sound (r) as in the Slavic, but it remained in the same position as in Proto-Indo-European. It is quite obvious that OCS gradŭ is a loanword from one of Thraco-Illyrian dialects (see supra). All these forms originate from PIE *ghordho-s “fence” (IEW, 444). This root is wide-spread in Indo-European languages (see gard). The same phonological change can be found in Romanian mal “bank, shore”. It was considered Thraco-Dacian because it is found in Ancient glosses; cf. Malua, Dacia Maluensis alternating with Dacia Ripensis, from Latin ripa “bank” which makes clear the meaning of Maluensis. Romanian noun mal originates from PIE *molā “shore” (IEW, 721); cf. Albanian mal “hill, mountain” Latvian mala “bank, shore”, Gaullish -melos (in place names). Cognates are found in other Nostratic language families as well, such as Dravidian languages: cf. Tamil malai “hill, mountain”, Malayalam mala “mountain, hill-land”, Kannada male “mountain, forrest” (Bomhard& Kerns, 550, 1994).