- Loans between Slavic and the Finnic and Baltic languages, while offering much more fragmentary information, appear to bear this assumption out. See particularly Valentin Kiparsky's “The earliest contacts of the Russians with the Finns and Baits” in Oxford Slavonic Papers III and the same author's “Chronologie des relations slavo-baltes et slavo-finnoises” in Revue des Etudes Slaves XXIV (1948), and K. Būga's “Die Vorgeschichte der Aistischen Stämme im Lichte der Ortsnamenforschung” in the Streitberg-Feslgabe (Leipzig, 1924).
- My interpretation of the Indo-European phonemic system is based, particularly insofar as the velar series are concerned, mainly on W. Lehmann, Proto-Indo-European phonology, Austin, 1952. As it represents a late stage of the variety of Indo-European from which Slavic developed, laryngeals and reduced vowels are absent, having merged with other phonemes by this time. The much disputed question as to whether an exclusively Balto-Slavic proto-dialect ever existed or whether there was merely a period of parallel Baltic and Slavic development characterized by geographic contiguity and consequent close contacts between Baltic and Slavic speakers is not touched upon this paper. Certainly the late Indo-European dialect sketched in 1.1 is also the ancestor of Baltic. Almost all the developments outlined in 1.2 are also shared by Baltic. Kury-lowicz has placed the consonantal j-changes mentioned in 1.3 in the period of Balto- Slavic common development (see his L'apophonie en Indo-Européen, Wroclaw, 1956, pages 235–236).
- There is reason to suppose that in the variety of Indo-European underlying Baltic and Slavic the length contrast was neutralized in diphthongs (vowel plus resonant) but that the former length contrast left its reflex in that long vowel plus resonant became a diphthong with acute accent, whereas short vowel plus resonant became a diphthong with circumflex accent. The phonemic status of accent remained the same into and through the Common Slavic period we are here considering, though the position of the accent underwent a number of complex and, in their details, not clearly reconstructed shifts. Since the position and nature of the Common Slavic accent is not pertinent to the argument of this paper, no further mention will be made of it.
- For an exposition of the phonemic development of Indo-European s and kˆ, ĝ, ĝh in Slavic, see André Martinet, Economie des changements phonétiques (Berne, 1955), pp. 240–242.
- I offer the following hypothesis regarding the phonological development of the reflexes of the velars before j and front vowels. In their initial stage, č and ž, resulting from consonantal j-changes and the so-called first palatalization, were allophones of k and g ([k g] before back vowels, something like [č' ž'] before front vowels). The passage of ě after š ž č ž to a would have been sufficient to establish č and ž as phonemes, but this probably occurred later than the passage of åj to ě (or i in certain positions), which established č and ž as phonemes, for now [k] and [g] (probably fronted varieties of these consonants) again appeared before the front vowels ě and i. Presumably the new allophones of k and g before these front vowels were at first highly fronted and then later assibilated, becoming [ć ž] eventually. A corresponding allophone of x eventually became [ś] and then coalesced with s in East and South Slavic dialects and with š in West Slavic dialects. In various parts of the Slavic domain, c and z (from the allophones [k' > ć] and [ǵ > ž] attain phonemic status in various ways during or immediately after the period when the Slavs began to migrate. In the West Slavic area they first coalesced with the reflexes of t' and d', which also originally appeared only before front vowels, but later, as the result of the backing of certain front vowels or analogical substitution of back vowel for front vowel in certain paradigms came to stand before both front and back vowels. In a large part of the South Slavic area (the dialects from which Slovenian, Serbo-Croatian, and Macedonian developed) the reflexes of t' and d' took the place of c and z as allophones of k and g before frpnt vowels, since the former were phonetically closer to k and g, being then probably [t' d'] or [k' g'] (cf. the reflexes of t' and d' in modern Macedonian, namely k' g'). The reflexes of t' and d' continued to function as allophones of k and g until the coalescence of y with i in South Slavic and of ŭ with ǐ in western South Slavic caused k and g to appear before front vowels again. In this connection, note
- Trubetzkoy's statement in “Die altkirchenslavishe Vertretung der urslavishen *tj, *dj,” Zeitschrift für slauische Philologie XIII (1936), pp. 88–97: “Im Urserbokroatishen müssen die aus *tj, *dj entstandenen palatalen Verschlusslaute ebenfalls als kombinatorische Varianten der Velaren k g gewertet werden. Zu selbständigen Phonemen werden sie erst nach Aufheben der Gegensätze u-ü, y-i, a-ä.”
- The relative chronology of the so-called second and third palatalizations has been much discussed. Recently F. V. Mareš in his article “Vznik slovanského fonologického systému a jeho vývoj do konce období slovanské jazykové jednoty” in Slavia XXV (1956), see especially pages 464–468, has renewed the argument that the so-called third palatalization preceded the second. Whatever the merits of Mareš's arguments (some of his reasoning seems quite compelling, but other parts weak or even self-contradictory) regarding the priority of the third palatalization, he is surely right in asserting that the first stage of the processes traditionally termed second and third palatalizations was an allophonic fronting of the velars (Mareš holds, correctly I believe, that when the velars were fronted in the third palatalization, back vowels following them were replaced by the corresponding front vowels, the fronted velars thus remaining allophones not contrasting with the unfronted velars before back vowels), that the results of both palatalizations were at this stage already phonemically and phonetically identical and that the final stage (assibilation or affrication) was therefore undergone simultaneously by the fronted velars resulting from both second and third palatalizations.
- For a discussion of the phonemic analysis of Old Church Slavic, which in its oldest stages is almost identical with the stage here designated Common Slavic (the only difference being that Common Slavic t' and d' become št and žd in Old Church Slavic and that an additional phoneme j˛ must be added), see Horace G. Lunt, “On Old Church Slavonic phonemes: the Codex Zographensis” in Word VIII (1952), 311–328 (see particularly the discussion of the phonemic status of j on pages 318–321) and N. S. Trubetzkoy, Altkirchenslavische Grammatik, Sitzungsberichte der Österreichischen Akademie der Wissenschaften, Philosophisch-historische Klasse, vol. CCXXVIII: 4, particularly the discussion of nasal phonemes on pages 36 and 81–82.
- The treatment of the sequence vowel plus liquid plus consonant varies in the various parts of the Slavic domain. The tendency toward open syllables was not carried through completely in all parts of the domain. In this connection, see Roman Jakobson “On Slavic diphthongs ending in a liquid” in Word, VIII (1952), 306–310.
- Accounts of the evolution from Indo-European to Common Slavic differing somewhat from the foregoing may be found in: N. van Wijk, “K istorii fonologičeskoj sistemy v obščeslavjanskom jazyke” in Slavia, XIX, 293–313; F. V. Mareš, “Vznik slovanského fonologického systému a jeho vývoj do konce obdobi slovanské jazykové jednoty” in Slavia, XXV, 443–495; and J. M. Kořinek, Od indoeuropského prajazyka k praslovartˇine (Bratislava, 1948).
- Some investigators (e.g. van Wijk in the article cited under footnote 5) assume a phoneme ě2 < åj which caused the second palatalization of preceding velars and then fell together with ě1. Such an assumption would permit overlapping of the first and second palatalizations. However, not only is this solution unnecessarily complicated and phonetically unlikely, but it will not take care of cases of second palatalization in loanwords caused by front vowels other than ě2 e.g. cǐrky (cf. Russian cerkóv´) ‘church’ < Greek kirjakí or from a Germanic word derived from the Greek etymon.
- This delineation of the linguistic situation in the Balkans prior to the Slavic settlement there is based primarily on Konstantin von Jireček's Die Romanen in den Städten Dalmatiens während des Mittelalters (published as nos. 48 and 49 in the Denkschriften der kaiserlichen Akademie der Wissenschaften, Vienna, 1902 and 1904) I, 13. I have deviated from this line in including Scupi and at least part of the Haemus in the Latin area (following Skok in “Zum Balkanlatein IV”, Zeitschrift für romanische Philologie, LIV, 179).
- Concerning the location of the area in which contact between proto-Rumanian and proto-Albanian could have taken place, see Henrik Barić, Lingvističke Studije (Djela Naučnog Društva NR Bosne i Hercegovine I, Sarajevo, 1954), particularly p. 25 et seq. N. van Wijk offers evidence implying that the area along the Serbo-Bulgarian border was long occupied by a Romanic population in his “Taalkundige en historiese gegevens betreffende de oudste betrekkingen tussen Serven en Bulgaren” (Medeelingen der kortinklijke akademie van wetenschappen, deel 55, series A, Amsterdam, 1923). A discussion of the various theories offered concerning the area or areas in which proto-Rumanian was spoken is contained in M. Friedwagner, “Ueber die Sprache und Heimat der Rumänian in ihrer Frühzeit”, Zeitschrift für romanische Philologie LIV (1934), 641–675.
- E. Kranzmayer in his “Frühromanische Mundarten zwischen Donau and Adria” (Zeitschrift für Namenforschung XV) makes some daring hypotheses regarding possible Romanic isoglosses in the Pannonian and inner Balkan area on the basis of the few German and Slavic place-names in this area derived from Latin names.
- Skok holds (and brings impressive, though not completely conclusive, evidence to support his thesis, in “Considérations générales sur le plus ancien istro-roman” in Romanica Helvetica XX [1943], 472–485) that Istro-Romanic was originally a Romanic dialect of the Eastern type, closely allied to Dalmato-Romanic, which was later overlaid and replaced by a dialect of the Rhaeto-Romanic (West Romanic) type. Skok's evidence consists largely of the Slavic place-names in Istria, of which such as Capris > Slovenian Koper ‘Capo d'Istria’, Petena ˜ Pelina > SCr. Pićan ‘Pedena’, which are apparently derived from VL etyma with unvoiced intervocalic stops (but note that a number of apparently equally old Istrian place-names in Slavic show voicing of intervocalic stops: Piquentum > Buzet, ‘Pinguente’, see 3.3 below for development of this item) and of apparent survivals in present-day Istro-Romanic dialects. R. L. Politzer in “A note on the North Italian voicing of intervocalic stops”, Word XI (1955), 416–419, adduces evidence from eighth-century documents that voicing of intervocalic stops had not yet taken place at that period in some areas of Northern Italy, particularly in northeastern Italy. If Western Romanic voicing of intervocalic stops spread from a western center (Gaul?) and had not yet covered all northern Italy by the eight century, this would support the thesis that voicing of intervocalic stops had not yet taken place in Istro-Romanic when the Slavs began to settle in Istria (seventh century).
- The Vegliote affrication of k and g before i is very probably a late development, as it takes place also before the reflexes of VL u (which passed through a stage ü). It is surely late if Skok is correct in deriving the place-names on Veglia Maknel from *machinile, Krknul from *circīnālis, and Kakorajne from *cicerīna see his “Studi toponomastici suir isola di Veglia” in Archivio glottologico italiano XXV, 319–355. See also Carlo Tagliavini, Le origini delle lingue romane (Bologna, 1948), p. 265 (cited hereafter as Tagliavini, Le origini).
- See Tagliavini, Le origini, pp. 257–258.
- In most of the VL area, vowels under stress were phonetically (nondistinctively) long in open syllables, short in closed syllables. Furthermore, the stressed vowels in antepenultimate syllables were shorter than those in penultimate syllables. It is probable that even short stressed vowels were phonetically slightly longer than unstressed vowels, which were uniformly short. See chapter II of Harald Weinrich's Phonologische Studien zur romanischen Sprachgeschichte (Münster, 1958); also C. Tagliavini, Le origini, pp. 143 and 145 and C. H. Grandgent, Introduction to Vulgar Latin (Boston, 1907), pp. 76–77. I am indebted to André Martinet for the information regarding difference of vowel length between the penult and antepenult in Vulgar Latin.
- The Proto-Rumanian vowel system also underlay the Vulgar Latin loans in Albanian. The evidence for Dalmato-Romanic, as reflected in Vegliote, is not entirely clear. A. Colombis in “Elementi veglioti nell'isola di Cherso-Ossero,” Archivum Romanicum XXI (1937), 243–267, postulates, on the basis of Dalmato-Romanic loans in the Venetian and Serbo-Croatian dialects of Cherso, the same vowel system for Proto- Dalmato-Romanic as for Proto-Rumanian.
- A good part of the place-name etymologies adduced as examples in this paper are drawn from the various publications of the late Yugoslav linguist Petar Skok, particularly his “Studije iz ilirske toponomastike” which appeared in the Glasnik zemaljskoga muzeja u Bosni i Hercegovini, vols. XXIX, XXXI, and XXXII (Sarajevo, 1917, 1919, and 1920), his “Zum Balkanlatein” in the Zeitschrift für romanische Philologie, vols. XLVI, XLVIII, L, and LIV, and his Slaoenstvo i Romanstvo na jadranskim otocima (two volumes, Zagreb, 1950, hereafter cited as Slavenstvo i Romanstvo), as far as the identification of the modern word with its ancient etymon is concerned. On the details of the evolution from the Vulgar Latin etymon to the modern South Slavic form I differ from Skok in a great many cases.
- Latin intervocalic voiceless stops are voiced in the Friulan area, as in all of the Western Romanic domain, the exact time when this occurred being uncertain (see Politzer's note cited under footnote 12). Prothetic v is regular before initial back vowels. m instead of n is a hypercorrection (m # > n # in Slovenian dialects on the western periphery).
- Ulpiana > *lyptanü > Lipljan (a town in Kosovo-Metohija) is puzzling. Here we have metathesis of the initial liquid, as is expected, but the reflex of the unstressed u is that of proto-Slavic long rather than short u (the expected development is that of Ulcinium > Lacin, cf. footnote 29 below). The labial context may have preserved a closer articulation of u in Ulpiana, thus accounting for the exceptional treatment (personal communication of André Martinet). The treatment Ulpiana > Lipljan parallels the development *ūrba > ryba ‘fish’ tentatively proposed by Roman Jakobson in footnote 1 to his article “Slavic Diphthongs Ending in a Liquid” in Word VIII (1952), 306. Transmission through or contamination by aboriginal (i.e. non-Romanic) Balkan dialects may possibly account for some of the discrepancies noted.
- Final g rather than z, which would be expected as a result of the second palatalization has been explained (by Fran Ramovš in “O prvotnih substitucijah za balk.-lat. k, g pred e, i,” Južrtoslovenski Filolog VI, 165) as due to the borrowing of the place name as a locative and the replacement of z by g in the other cases by analogy with the regular morphophonemic alternation g˜z (later z) in Common Slavic. Cf. Old Church Slavic bogŭ ‘God’, o bozě ‘about God’.
- To these examples of Latin stressed i (VL ė) > Slavic icaron;, all of which are taken from South Slavic, there might be added the following two loans which are widespread in the Slavic domain: piperem 'pepper'> *pǐpǐrǐ > SCr. papar, Czech pepř, Polish pieprz; missa ‘mass’ > SCr. dialect and Slovenian maša, Czech mše, Slovak mša, Polish msza. Unstressed VL e > Slavic ǐ is also exemplified by piperem > *pǐpǐrǐ.
- Some investigators (most recently Mareš in the work cited under footnote 4) have held that the changes in Slavic traditionally labeled “metathesis of liquids” involved, not only in East Slavic, but in the whole Slavic domain, the existence at a certain stage of an anaptyctic vowel between the liquid and the following consonant. A certain amount of support for this theory is lent by the names Srakane (two small islands west of Cherso; in Italian, Canidole,) if this name is, as appears possible, derived from Saracenus ‘Saracen’ (Arab corsairs were active in the upper Adriatic at various times in the ninth century and c. 841 a Saracen fleet inflicted a severe defeat upon the Venetians in a sea battle off Susak/Sansego), and Mljet (from Latin Melila), if the second vowel in the ancient name had not been already lost in Vulgar Latin by syncope (Constantine Porphyrogenitus in the tenth century has Méleta; modern Italian, derived from the Venetian name, is Méleda), and by Aspaláthos > VL *spelet- (with the alternation of eand a found frequently in names of this area; cf. Spalatum˜Spaletum in Thomas Archidiaconus, Delmati˜Dalmati, Serdica˜Sardica, Méleta in Constantine Porphyrogenitus ˜ Malata in the anonymous geographer of Ravenna) > *splětŭ; > Split (Spalato) and possibly by Curicum > *kürǐkü > Krk (Veglia). Lorn < Almus (in north-western Bulgaria) does not give the expected South Slavic long vowel reflex and deserves further investigation.
- V. Kiparsky in Die gemeinslawischen Lehnwörter aus dent Germanischen (Annales Academiae Scientiarum Fennicae, Series B, Tomus XXXII, Helsinki, 1934) and F. Mareš in the article cited under footnote 4 assume a development åw [ō] > u, for which Slavic borrowings from and into Germanic, Baltic, and Finnic appear to argue, e.g. Germanic plōg- > OChSl. plugŭ. Kiparsky on p. 285 of the above work says: “Got. ō und urgerm. o haben stets slav. u, niemals y, gegeben, weil das heutige slav. u noch in urruss. Zeit (etwa um 900) denselben Lautwert wie das germ, ō gehabt hat (die Ostseefinnen, die sowohl ū als ō hatten, wählten zur Wiedergabe der urruss. u ihr ō… und ebenso taten die Letten, wenn sie slav. duma durch duoma < dōma wiedergaben).” The very numerous Slavic borrowings from Balkan Latin on the other hand contradict this hypothesis, for if a phoneme ō had been present in the Slavic system when the Slavs entered the Balkans, they would not have borrowed VL ⊙ as ū > y > i (Salona > Solin). In general, the borrowings between Slavic and Latin, Germanic, and Finnic corroborate each other's evidence very nicely. On this one point they seemingly do not, and it deserves further investigation. With regard to the loans from Germanic, I venture the following very tentative hypothesis: It is not at all uncommon for long e and o to become diphthongized, acquiring a high semivocalic offglide in many or all positions (cf. modern English, some dialects of German, and some varieties of colloquial Hungarian). It is quite possible that in those varieties of Germanic, from which Slavic borrowed, long vowels were diphthongal in nature. This would agree nicely with the position of those who regard the long vowels of early Germanic dialects as complex nuclei (see Robert P. Stockwell, “The Phonology of Old English”, Studies in Linguistics XIII, 1–2, particularly pp. 16–17). Thus, the Slavs might have equated a Germanic long o having semivocalic offglide with their own åw, which at this stage might well have been [oy]. I am not acquainted with the details of Finnic and Baltic evolution. However, it is evident from the examples quoted by Kiparsky that his “urrussische Zeit” refers to a period after Common Slavic and after monophthongization, since he cites Russian words as giving Finnic ō where the u in the Russian word is not from proto-Slavic åw. Thus, Veps kōma < kumŭ < Latin commetter. It is, of course, quite probable that after monophthongization Slavic u was relatively open, so that Baits, Finns, Greeks, and others would equate Slavic u with their own relatively closer o when they borrowed from Slavic, just as the South Slavs after monophthongization (e.g. at the stage when Duklja < Doclea and Solun <Saloniki were borrowed) equated Greek and Romanic o with their u.
- Josip Klemenc in Ptujski grad v kasni antiki (Academia scientiarum et artium slovenica, opera 4, Ljubljana 1950), pp. 76–77, believes the Slavs settled the area near Poetovio shortly after the Langobards abandoned Pannonia to the Avars (569), probably about 580.
- The Freising Monuments, three short passages written in an archaic Slovenian, apparently by German missionary priests around 1000 A.D., indicate that at the time of their composition ĭ and ŭ had fallen in a weak position and coalesced into one phoneme (probably a schwa-like vowel) in strong position. In two of the three monuments y and i had also merged (into i), while in the other the distinction was preserved after labial consonants with merger taking place in other environments. See Fran Ramovš and Milko Kos, Brižinski Spomeniki (Ljubljana, 1937), particularly pp. 7 and 12, and also Fran RamovS, “Drobnosti iz slovenske gramatike,” in Slavia I, pp. 27–37. Ramovš believes that the monuments represent copies of originals composed as much as one or even two centuries previously. They would thus represent a stage of the language considerably earlier than 1000 A.D.
- The explanation of the change of y to i in *rymˇ, *kryžǐ as due to initial borrowing into a Slavic dialect in which i and y coalesced was, as far as I know, first suggested by Meillet in Le Slave Commun (Paris, 1924), p. 102. Vaillant in his Grammaire comparée des langues slaves (Lyon, 1958), I, 120, attributes the y in *rymŭ, *solynŭ, etc. to the Rumanian change Latin o > u before nasal consonant, but this assumption is mistaken, for we have the same borrowing of -ona as -ynŭ in Romanic areas where o definitely did not pass to u before nasal consonants (e.g. in Friuli Cormones, VL *kormón- > Slovenian Krmin ‘Cormons’, Glemona > Slovenian Humin ‘Gemona’), which latter fact indicates that Romanic o was borrowed in the South Slavic area as Slavic ŭ > y > i without an intervening stage of Romanic u.
- As stated above (see footnote 14), there is considerable doubt whether Rumanian had affricated its velars before front vowels by the time the Slavs reached the Balkans. Latin loans in Albanian show no affrication (Latin k g before front vowels > k' g' in Albanian, just as do the reflexes of the Indo-European velars in the same position. See G. Mayer “Die lateinischen Elemente im Albanesischen” in Gröber's Grundriss der romanischen Philologie (Strassburg, 1904–06), and “Albanesische Studien III” in volume CXXV of the Sitzungsberichte der philosophisch-historischen Klasse der kaiserlichen Akademie der Wissenschaften (Vienna, 1892). Kiparsky's (in the work cited under footnote 23) assumption that “romanische Assibilierung” is responsible for the changes k > c and g > z in Balkan place names is mistaken (particularly for the Adriatic coastal area for which we have incontravertible evidence that no affrication of velars before front vowels took place).
- With due respect for Skok's excellent and voluminous philological and etymological researches it must be stated that his position on a number of details of the evolution of Slavic and Romanic is quite untenable. Skok was taken to task on some of these points by Anton Grad in “Della palatalizzazione di k latino intervocalico nel dialetto veneziano,” Vitalia Dialettale IX (1939), 230–239, and by Fran Ramovš in the article in Južnoslovenski Filolog VI cited under footnote 20. Regarding the treatment of velar stops before front vowels in loans from Vulgar Latin, Skok held (see p. 386 of “Zum Balkanlatein II” in Zeitschrift für romanische Philologie XLVIII) that velar before front vowel in items borrowed contemporaneously might undergo the first, second, or third Slavic palatalizations or remain unchanged. Ramovš correctly points out that affrication in those items which Skok interprets as examples of first palatalization actually results from other sources and that the items which show unchanged velars before front vowel were borrowed from Dalmato-Romanic after the palatalizations had ceased to be operative in Slavic and after further changes had made it possible for a velar to occur before front vowel. Regarding the items which show the reflexes of the second and third palatalizations RamovS states that they cannot be the result of the Slavic palatalizations since these had already long since taken place. Rather, these items, according to Ramovš, represent cases where the Slavs in borrowing had substituted their phonemes c and z for VL positional allophones (before front vowel) [k'] and [g']. Skok, replying in Južnoslovenski Filolog VIII (“Slavenska palatalizacija sa romanističkog gledišta”), accepted this interpretation. I do not. Ramovi offers no evidence whatever that the Slavic palatalizations were operative at a “prehistoric” period long before the Slavic invasions of the Balkans, but simply states it as an accepted and proven fact. More important, Ramovš's explanation is phonetically and phonemically unlikely. As stated in footnote 4 above, by the time the original velars before front vowels had become [ć] and [ž] they were no longer allophones of k and g, but their place as allophones of k and g before front vowel was taken by the reflexes of proto-Slavic tj and dj, which were phonetically nearer k and g, being then probably [t' d'] or [k' g']. Had the sound changes we call second and third palatalization really been long completed as Ramovš maintained, then VL [k' g'] before front vowel would have been borrowed as the (then) Slavic allophone of k and g before front vowel, which would have been the reflexes of proto-Slavic tj and dj, and loans made in this period would show ć and zacute; in modern Serbo-Croatian. This is not the case; early VL loans in SCr. show second-third palatalization reflexes; ergo, second and third palatalization was operative when the linguistic ancestors of present-day Serbo-Croatian speakers settled in the Balkans.
- Skok in Slavenstvo i Romanstvo, p. 35, mentions the occurrence in Cassiodorus without giving any further citation.
- Skok's assumption (in Slavenstvo i Romanstvo, p. 69) that the present dialect form of the name, Caska, is a result of cakavism (cakavism in Serbo-Croatian dialectology signifies the merger of the phonemes š ž č with &s z c in many areas of Dalmatia under the influence of Venetian, which shows a similar coalescence) is mistaken. Rather the original form of the name must have had c, and the form Čaška which sometimes occurs in the standard language represents a hypercorrection. Cf. also the hypercorrect form Čres, sometimes used for Cres.
- The name of this town, ancient Ulcinium (Latin) or Olkinion (Greek), now in Albanian speaking territory (though within the political boundaries of Yugoslavia), although it apparently had a Romanic speaking population until late in the middle ages, shows many variants in Serbo-Croatian. Ulclnj is the present official form and the following variants are attested: Ocin, Ocinj, Ucin, Dulcinj, Ulćin, Lǐcinǐ (this latter writing is found in medieval Serbian documents from the thirteenth century on and probably represents a' pronunciation /lacin/). Lacin is probably the native Serbo-Croatian development of this word and the remaining forms show varying degrees of contamination by foreign forms such as Albanian Ulqin/ulk'in/ and Italian Dulcigno. The native form of the name would then have resulted from metathesis of an original ŭl- (cf. Ulpiana > Lipljan treated above). In any event, important for the present argument is the fact that all the Serbo-Croatian variants of the name show k > c (excepting only Ulćin, which is based on Albanian/ulk'in/in all probability).
- Skok points out that this word has a modern Serbo-Croatian reflex ceta ‘ancient coin’ in the dialects of the Sisak (> Siscia) area. Siscia was an important Roman mint and Roman coins are still frequently dug up in this region.
- The voicing of the initial consonant in the Slavic reflex is difficult to explain, but the development of the remainder of the word is quite normal.
- The Gailtal lies north of Friulan speaking territory, thus outside the area where we know VL velars were preserved without affrication before front vowels. However, the German form Gail seems to indicate that this name was borrowed before the affrication of velars had reached this area (as was Kelsbach < Celeusus in Bavaria). Perhaps the Slavs encountered German speakers (who arrived in the western Austrian Alps shortly before the Slavs began to settle the Balkans) in this area and borrowed the name from them, rather than from Latin. The German form Gil (presumably with a long vowel) is attested three' times at least in documents from the early twelfth century. Ramovš's derivation (in the article cited under footnote 25) of both the Slavic and German forms of the name from a Vallis Julia attested in Paulus Diaconus is unconvincing. The Friulan name Zeje seems to be a borrowing or reborrowing from the Slavic. Kranzmayer and Ramovš cite OHG Gila without an asterisk, but do not state where it is attested.
- The significance of these Austrian place-names in dating the Slavic third palatalization was, to my knowledge, first pointed out by Primus Lessiak and has later been discussed by a number of scholars, including R. Ekblom in his “Germanische und slavische Palatisierung in Ortsnamen des östlichen Alpengebietes,” Språlkvetenskapliga Säliskapets i Uppsala Förhandlingar 1949–1951, pp. 37–52. The Greek examples were listed by Max Vasmer in his work Die Slaven in Griechenland (Abhandlungen der preussischen Akademie der Wissenschaften, phil.-hist. Klasse nr. 12, Berlin, 1941).
- The vowel in Cres represents the reflex of Common Slavic ě (the local dialect is ekavski; neighboring ikavski areas call the island Cris) as is usual in South Slavic metathesized forms where the initial element of the diphthong had been e. With regard to my statement that Cres represents the second and not the first palatalization, it should be noted that the ča-dialects of the Quarnero region and neighboring Istria do not show the change čr > cr which occurs in the standard language. It is true that the town Cres and some other localities on the island have “cakavian” dialects (i.e. all č's are replaced by c; see footnote 28), but the name of the island is everywhere Cres (or Cris in ikavski dialects), in all localities, including those with non-cakavian dialects (which ought therefore to show čr-, had the original form of the name had that cluster rather than cr) on the island and in neighboring regions. The form Čres which appears on a few recent maps is nowhere attested in local speech and evidently (like Čaška mentioned in footnote 28) represents a hypercorrection made on account of local cakavism (see Skok's Slavenstoo i Romanstvo, particularly pp. 35–36 and note 8 on p. 43; also Hamm, Hraste, and Guberina's listing of localities on the island which show cakavism in Hrvatski Dijalektološki Zbornik I, 45).
- Bulgarian and Macedonian distinguish the reflexes of ŭ and ǐ (where these reflexes are not zero) and allophonic or phonemic palatalization is retained to the present day in many varieties of Bulgarian, particularly in the East. The falling together of y and i in the eastern branch of South Slavic must be placed at a later date than that in the western varieties, since OChSl., based on a Macedonian variety of Slavic speech, distinguishes i and y, at least in the earliest monuments.
- Choose new content alerts to be informed about new research of interest to you
- Easy remote access to your institution's subscriptions on any device, from any location
- Save your searches and schedule alerts to send you new results
- Export your search results into a .csv file to support your research