I am surprised but pleased to see how many echoes this article produced, and this is why I decided to add (at the end) a few specifications, until I shall take my time to write an updated and enriched version.
A New Dimension of the Linguistical Relations between Romanian and Greek
by Sorin Olteanu
The lexical relations between Romanian and Greek consist, following the main works on the history of the Romanian language, only in borrowings, grouped in three layers:
- words entered in Latin and from here inherited in Romanian, as martur "witness"< vulg.lat. *marturu < μάρτυρ; papură "reed"< vulg.lat. papura < πάπυρος; arom. stur "pillar" < vulg.lat. *stulu < στύλος, drum "way"< δρόμος "a run"a.s.o.
- words borrowed, directly or through other balkanic languages, from byzantine greek: ieftin "cheap"< εὐθηνής, a lipsi "to miss, deprive"< ἔλειψα, a mirosi "to smell"< ἔμύρωσα and – trough church slavonic – Hristos "Christ" < slav.Hristosŭ < Χρηστός, evanghelie < slav. evangélija < εὐαγγελία, and
- words received from new Greek: ighemon "leader"< ἡγεμών, grămătic < γραμματικός, drahmă < δραχμή ş.a.m.d.
In my opinion there must be a fourth and oldest layer, if the following judgement is correct:
- as a thracologist, I know that old Greek and Thracian languages were, on one hand, genetically related and that, on the other hand, as neighbours for a long time, they must have had mutual borrowings (see examples below).
- it is generally admitted that Romanian inherited a number of words from Thracian substratum
- therefore we must accept the hypothesis that words common to both Thracian and Greek could have also survived in Romanian.
Practically speaking, this means that, hypothetically again, we could find parallels between the vocabularies of modern Romanian and old Greek. But - and this is in what consists the new approach of the matter -, in order to find these common words, we must not compare their actual forms - as they are given by the Greek and Romanian dictionaries -, but their reconstructed etymons. The imponderables of this approach are the phonetic rules which governed the Thracian languages. Some of these rules are yet accepted by most thracologists (for example the satem character of Thracian), and I shall only take these ones into account.
The kinship of Greek and Thracian was denied because of their belonging to different groups following the centum-satem isogloss. In my opinion this split is not as relevant as it was thought to be. Late Latin and Romance languages have more satem characters than classical latin, there are also some modern german dialects in this situation (see the pronounciation iš for ich) etc. As this character can change in time for other languages, I don't see why it couldn't for Thracian. The lexical comparison prove that Thracian was close to both Greek and Slavic groups. Based on this fact, my theory is that Thracian belonged at first to the Greco-Macedonian group and then it received a massive influence from Balto-Slavic which determined it to change its centum into satem character. For the oldest Greek-Thracian relations we might first quote some mythical facts: Greek gods (Dionysos), heroes (Diomedes and the Troans), tribes (Abantes, some Thessalian or insular populations, as those of Thasos, Samothrace) etc. were known as having Thracian origin. Then there are, among the few words known of Thracian, some whose immediate etymon can be also found only in Greek: the Edonic name Paibes (also in compounds like Aliupaibes, Mestypaibes a.s.o) is for sure the same as gk. pais, paidós "child"; thr. toponym Salmuris - a small fortress by the Black Sea shore in the Romanian Dobroudja - is the same as the greek word halmyrís "salt water, sea"; thr. sulā "forrest" (see Scaptensula translated by Herodotus as Skaptē hýlē) is the gk. hýlē "forrest"; thr. deiza "fortress, town" corresponds entirely to the gk. teichos "wall"; thr. Arzos - a river in central Thracia with white water (today Sütlü-dere in turkish, meaning "milky river") - is the gk. argós "(shiny) white" and many others (thr.-zenis = gk.-genēs, thr. Gaidreas = gk. Phaidros, thr. zetraia = gk. chytra etc.). There are also some morpho-phonetic developments which bind these two languages. Among them let us mention the metathesis of yot with role of de-palatalization; in Greek: phainō from *phanyō, keirō from *keryō a.s.o. (it is a very large group of words); in Thracian: the later river name Oiskos (a Bulgarian tributary of the Danube, named Isker) has, at Herodotus and Thucydides, the form Oskios; the compounds with Roime- (Roimetalkas, Roimezeues...) come for sure from *Romye <ie. *rōmā "strength" and others.
I think these common inherited features are enough reasons to accept the kinship between Greek and Thracian and thus to take the premises of my hypothesis as valid. Let us now search for parallels between Romanian and old Greek lexica.
There is a Romanian word generally thought as coming from the substratum: drom. brâu "belt (also as anatomical region)", coming from an earlier *brănu, arom. brăn(u). It occurs in albanian also: brës "belt", mbrenj "to belt". Now, if this word is of Thracian origin, it must have sounded in this language something like *bren(o-) (if not even *brən(o-)), the only form which explains both Romanian and Albanian developments. This *bren(o-) supposes an indo-european root *bhren(o-), whose greek result would be *phren(o-). We know indeed such a word: phren, phrenos "mind", whose oldest meaning was in fact "diaphragm, the membrane which separates the thorax from the abdomenon", thus almost identical to that of the romanian word.
Another case: rom. a zburda "to sport" [to note how much alike is this word to its english correspondent, though the origin of the latter is thought to be different!), with unknown etymology. It has already been related to the rom.root burd- "leather recipient, abdominal sack" (in burduf, burduhan, a îmburda) - with "prothetic mobile s-" -, but it is difficult to explain their semantic relation. All its sounds are very stable in indo-european languages, so we can easily suppose a root *s-bhurdh-, undoubtly a late development of the PIE *bher- "to bear". In Greek, this form would change first to *s-phurth- and then (following Grassman's law) to *s-purth-. Such a word exists in old Greek indeed: spyrthízō means "(about animals, mainly horses) to run with jumps, or hitting their bellies with their legs" which explains also the semantic relation between rom. a zburda and rom. burd-.
Rom. beregată means "throat" and is probably a participial development of a root *bereg-. Another word, with the same consonantic structure, is borugă "pass, strait". Both words have unknown etymologies and they are probably related (see the french word gorge for the semantic compatibility). We can find almost the same relation between the greek words *pharanx,-nngos "pass" and *pharynx,-nngos "throat" from the roots *bhara(n)g- and *bharu(n)g-. If theese roots existed in late i-ean, their primary meaning must have been "passage, place to pass through". As nasal infixes were "optional" in indo-european, these two roots could also explain the Romanian words.
There are many other such pairs of words, more or less concordant, but obviously related and whose Romanian correspondents (sometimes entire families) have no etymology and can be ancient: rom. bumb, bumburez, bumbărează "small round object, coccis" < (*bhombh-) =gk. pomphos, pompholyx (see also bombylis of non-greek [thracian?] origin) "id."; rom. bală, balaur(e) "mythological monster, dragon" (<*bhalla-)=gk. phale, phalaina (but lat. ballaena with double ll)"id., whale"; rom. strugur(e) "grape(s)" (<*s-trug-)=gr. root trug- "vineyard" (in trugō, trugesis...) a.s.o.
The examples above illustrated words of (late) indo-european origin preserved both in Romanian (as substratum heritage) and in old Greek. There are yet other words common to these two languages which seem to be greek borrowings from Thracian. Let us take some examples:
There is a very old rom. root, ţur-, whose primary meaning was "tube": ţurţur(e) "icicle", ţurloi "tibia, shin bone", ţurcă"a children game, played with a small round stick", ţuţuroi "water pipe", maybe even ţurcan(ă) ("tzurkan, a kind of sheep with long and rough hear (like tubes"). This root is comparable with the greek root *sur- (syr-) with the same meaning, in syrinx,-nngos "pipe" and syristēs "flute player". But initial antevocalic s- can not be organic in Greek, because in this context it always changes to rough spirit (see hex "six" from *sex, histēmi from *sistāmi etc.; this is a very common and well known Greek phonetic law). Therefore we can suppose that this word was borrowed in Greek from another language, and its Romanian variants seem to point to the Thracian group. The same happens with another Greek word: sigalos "soft" and its family. With the same phonetical correspondence (gk. s- / rom. ţ-) its Romanian pair is ţigaie (from an older *ţigalie) "(about the sheep's hear) soft, oil-like". Both the above words belonged probably to the language of Thracian shepherds, which explains better the how and why were they borrowed in Greek.
At the end I should stress a collateral utility of this new approach. As I said at the beginning, the Thracian was also very close to Slavic, even closer - in its later age - than to Greek. So, it is to be expected that many Thracian words were also close to their Slavic relatives. There are many proves in this direction and I shell mention here only one of them but very meaningful. Dioskurides, in his "De Materia Medica" mentions the Dacian name of the plant known to science as "Verbascum (thapsus or niger)". This is diesema, beautifully explained by the great bulgarian linguist Vladimir Georgiev from the i-ean *diwes-eusmn "daylighter" (germ. Himmelbrandt). The same plant is named in russian, czech, bulgarian and polish divizma or divizna. Thracian and Slavic are the only groups having this word. This alikeness raises a very interesting question: is it possible that some Romanian words - considered now as borrowings from untypical south-Slavic words - to be in fact taken not from Slavic but from substratum lanuage(Dacian)? Let me detail. There are many Romanian words supposed to come from south-Slavic, but whose direct bulgarian or serbian correspondents either do not fit the possible etymon, or they differ from their common Slavic pattern. Could we assume a substratum origin or influence in this cases? Let us take some examples again.
Rom. gorun(e) means a kind of oak growing at the mountain and it is explained by the Romanian etymological dictionaries as a borrowing from bulgarian, where it would be derivated from bg. gora "mountain". This explanation seems to me very unlike, as long as the word does not and did not exist in bulgarian. On the other side, there was an IE root *gwalan- which meant "oak" (from where Greek got balanos). Such a form would have become in Dacian, perfectly regular, *golon-, with the known vocalic evolution of a to o in this language. Then, borrowed in the Danubian Latin, it became, perfectly regular again, gorun(e) in Romanian.
Rom. dâmb "small hill" is explained from hung. domb, where it would be a borrowing from sl. dombu "oak tree". Some linguists rejected this etymology on the legitimate ground that the semantic mutation cannot be explained. In fact, we deal here with the well known i-ean root *dhmbho-"hill" which became in Greek taphos "tomb". The Thracian phonetic result would have been *dumb(o-) - and indeed the Greek tymbos might come frome here. So there are chances that a Daco-Moesian variant of this word would be the etymon of the romanian dâmb.
Nobody doubts that rom. a zbârci "to wrinkle" comes from bulg. zbrčkvam (se) "id." and it may be so indeed. But all slavic languages (except for the bulgarian) have a m in this root instead of b: smrč. The bulgarian itself has both forms: zbrč and smrč. On the other hand, the old Greek pharkis "wrinkle" and pharkidoō "to wrinkle" witness an old balkanic root with *bh (even if without prothetic s-): *bhrk. Taking into account that Thracian did use prothetic s-, it lost aspiration and it palatalized velar consonants, the Thracian form of this root would be exactly *z-brč. Is it therefore possible that in bulgarian mixed up both old balkanic and common slavic forms? In this case why does Romanian have only the old balkanic variant, while there is no trace of *smrč?
-I never said that Thracian was a centum language. Read again to see that my opinion is that thracian varieties were, at the time of their attestation, satәm languages. But there are facts that undoubtedly show that they presented common features with old greek and old macedonian, centum languages. These features can not be explained by further influence suffered by Thracian from these languages (for ex. thr. sula can not come from gr. ὕλη, but only from its etymon *sulwa). This is why I supposed that, at their beginning, they might have belonged to a centum group (Greco-Macedo-Thracian).
-Concerning the lexical parallels between Romanian (Latin) and Greek (see the diagram below):
-the oldest has to be the one inherited from a coomon branch of indo-european languages (called here Greco-Macedo-Thracian), of centum character.
-after the split of this group into Proto-Greek and Proto-Thracian began the process of satemization of the later (maybe due to its position, closer to the Slavic (and Baltic?)). Mutual borrowings between Thracian and Greek.
-historic (Classical) age. Thraco-Dacian and Classical Greek have mutual borrowings. In the same time the same happens between Classical Latin and Classical Greek.
-Danubian age. Danubian Latin inherits the Classical borrowings from Greek, as well as borrowings from Dacian (including some words of Greek origin in Dacian).
-previous borrowings are transmitted to Proto-Romanian, then to Common Romanian and Old Romanian. At each stage we must suppose also mutual borrowings
-Modern age. Some of the old words of Greek origin survive in modern Romanian. Mutual borrowings with Modern Greek.
Across all these ages, Greek language has got words from Latin, Thracian and Romanian, which in a certain percentage were passed to the next stage of its evolution.