All Slavic languages: The sensitive -g- of the genitive of masculine adjectives

  • There are three different, and probably unrelated, aspects in your question.

    (1) The change g>ǥ>(h) in the strip ranging from Upper Sorbian to Southern Russian is regular and occurs to any g. There is no good explanation to the origin of this phenomenon so far: some prefer to regard this shift as caused by the presumable Iranic substrate (Scythian and Sarmatian, being East Iranic languages, must have had ƀ, đ and ǥ as opposed to West Iranic b, d and g), though this hypothesis fails to explain why weren't Slavic b and d affected as well, plus e. g. Croatian, whose name is almost certainly of Iranic origin and thus may continue the speech of slavicized Sarmatians, lacks this development. On the other hand, Ukrainian and South Russian do have Iranic substrate (the idea supported by archeology, craniology and toponymy), so this scenario can't be ruled out.

    (2) The loss of the final vowel in the Serbo-Croatian ending is peculiar to that language and seems to be absent anywhere else. Actually, it affects the Dat. Sg. form as well: nov — novog — novom (https://en.wikipedia.org/wiki/Serbo-Croatian_grammar#Adjectives).

    (3) The special development g>h>∅>u̯>v is found only in North and Central Russian and in Cassubian and only in this ending. Its origin is unknown: there has been a thread in the Russian forum a year or two ago with more information. North Russian (derived from the speech of the tribes Slovenes and Krivichis) may have been particularly related to the dialects of other Baltic Slavs, including Cassubians, though the development -ogo>-ovo seems to be about 8 centuries younger than the migration of Baltic Slavs to the European North-East.
     
    Last edited:

    Christo Tamarin

    Senior Member
    Bulgarian
    (2) The loss of the final vowel in the Serbo-Croatian ending is peculiar to that language and seems to be absent anywhere else. Actually, it affects the Dat. Sg. form as well: nov — novog — novom (https://en.wikipedia.org/wiki/Serbo-Croatian_grammar#Adjectives).
    This can be traced in Bulgarian dialects, too. Please go to Стойков - Българска диалектология, look for "С м о л я н с к и г о в о р", and see "чẹлềкạтọк", where the final voiceless K comes from G (чẹлềкạтọк = Russian человека того).
     
    Last edited:
    This can be traced in Bulgarian dialects, too. Please go to Стойков - Българска диалектология, look for "С м о л я н с к и г о в о р", and see "чẹлềкạтọк", where the final voiceless K comes from G (чẹлềкạтọк = Russian человека того).
    That's interesting, but don't you think it may be a coincidence? If it is named after this Смолян (https://bg.wikipedia.org/wiki/Смолян), this phenomenon seems to be not contiguous in Serbo-Croatian and Bulgarian. Or is it?

    As far as I understand, standard Serbo-Croatian still has a tendency towards monosyllabic adjectival endings, so its -og belongs to different context. Of course, however, it could have been so that the shortening -ogo>-og had triggered this development in other case forms: there a probably studies dealing with the evolution of the Serbo-Croatian adjectival declension that provide evidence how it all developed in reality.
     

    Christo Tamarin

    Senior Member
    Bulgarian
    Now, it is not contiguous. Another feature, the triple articles, is not contiguous as well: it can be found in the Rhodope mountains to the east and again to the west of the Vardar river where it contributed to Standard Macedonian.

    In the past, the loss of the final vowel in the adjective endings was probably contiguous in the Bulgarian/Serbo-Croatian area. The problem is that now, most of Bulgarian dialects do not preserve such declined forms of adjectives at all - they are preserved in isolated locations only.
     
    Interestingly, judging from your link, that Smolyan dialect doesn't drop the vowel in the Dative, so we find йề си гàл’ạ хỳбạвêк Пèтрạ; ср’ồшнạх лèлинцèк Нòт’ạ vs. нàшему Ивàну, йề вѝкạм мòẹму сѝну or, with articles, чẹлềкạтọк, чẹлềкạсọк, чẹлềкạнọк vs. чẹлềкутуму, чẹлềкусуму, чẹлềкунуму. Also interesting is the distribution of vowels in the pronominal endings in these articles: preservation of the older o/e in the Genitive -тọк (<того), -нọк (<оного) and, analogically, in -сọк (<сего), vs. penetration of the contracted adjectival endings to the Dative -туму, -суму, -нуму (after *старуму<старѹ ѥмѹ).
     

    Anicetus

    Senior Member
    Croatian
    (2) The loss of the final vowel in the Serbo-Croatian ending is peculiar to that language and seems to be absent anywhere else. Actually, it affects the Dat. Sg. form as well: nov — novog — novom (https://en.wikipedia.org/wiki/Serbo-Croatian_grammar#Adjectives)

    Considering the syncretism of the dative and locative, novom could simply be the old L sg. form.

    In any case, loss of the final vowels in the those endings wasn't a phonetic change.

    Also note that the development wasn't directly *novogo > novog, but probably *novajego > *novajega > *novāga? > novōga > novōg. The forms without the final -a are in fact rather recent, Matasović claims in his Poredbenopovijesna gramatika hrvatskoga jezika (2008: 219, 229) that they're first attested in the 18th century. As Panceltic has already said, the forms with the final vowel are still entirely possible in modern BCS, and are actually much more common for some nominal pronouns (for example: koga, toga, svega). There are also dialects in which only the longer forms exist.

    On the other hand, the change *novajego > *novajega happened very early on, before first written records from the BCS area. It is usually explained as analogy to the G sg. ending -a of o-stem nouns. The innovation is shared with Slovene, where novega is still the only possible G sg. form of novi it's actually one of the isoglosses that separate the Western South Slavic (the Slovene-BCS continuum) from the Eastern South Slavic group (the Macedonian-Bulgarian continuum).

    It's very interesting that the loss of the final vowel in G sg. may have been shared by Bulgarian and BCS. Is it possible to date the change in Bulgarian?

    Matasović (2008: 219) doesn't say all that much about the evolution of the BCS adjectival declension, but he seems to suggest that -og as an ending concurrent to -oga came to be by analogy to coexisting D/L sg. endings -om and -omu, which is what syncretising the dative and locative resulted in, rather than the other way round. However, in the next paragraph he states that syncretising of the definite dative and locative sg. can be observed from the 15th century, with the original dative forms prevailing at first which is opposite from the situation today, and in my opinion could suggest there may have been a "shortening" of the endings later on. On the other hand, written records of the vernacular from that time mostly come from a rather limited geographic area, so the syncretism could have simply given different results elsewhere...
     

    franknagy

    Senior Member
    (Scythian and Sarmatian, being East Iranic languages, must have had ƀ, đ and ǥ as opposed to West Iranic b, d and g), though this hypothesis fails to explain why weren't Slavic b and d affected as well, plus e. g. Croatian, whose name is almost certainly of Iranic origin and thus may continue the speech of slavicized Sarmatians, lacks this development. On the other hand, Ukrainian and South Russian do have Iranic substrate (the idea supported by archeology, craniology and toponymy),

    I have heard about Scythians and Sarmatians only as ancestors of the Polish szlachta (noblesse) as difference from blood of serfs.
    I am surprised that these ancient tribes took part in the development of all Slavic nations.
    ....
    I know a saying that if you scrape a Russian then you will see the Tartar.
    I understand it so that Moscow was to be able to unify the surrendered small Russian principalities by learning the Tartar way of government.
     
    I have heard about Scythians and Sarmatians only as ancestors of the Polish szlachta (noblesse) as difference from blood of serfs.
    Actually, educated people of various nations in the late Middle Ages derived themselves (or part of themselves) from various prestigious sources: contemporary Lithuanians, for example, after having noticed the similarity of their language with Latin, concluded that they were direct descendants of the Romans, and from this height looked at the szlachta with understandable condescension.

    I am surprised that these ancient tribes took part in the development of all Slavic nations.
    The relationships of Slavic and Iranic (or more precisely, the adstrate or substrate influence of Iranic to Slavic) are a never ending story, discussed for more that 150 years, and limited first of all by the scarcity of Scythian and Sarmatian linguistic records. Yet, for example, the Slavic word for "god", *bagas>bogъ is usually suspected to be of Iranic origin, for phonetic (lack of lengthening of the root vowel expected after Winter's law — https://en.wikipedia.org/wiki/Winter's_law, cp. *nogos>*nōgos>nagъ "naked" [Lithuanian nuogas] with an expected development vs. bʰogos>>bogъ with an otherwise inexplicable root o) and semantic (Iranic shift *deı̯u̯os "god">"wrong god" and *bʰogos "good, blessing, welfare">"god" — https://en.wikipedia.org/wiki/Daeva) reasons. On the other hand, Ossetic, the modern descendant of Sarmatian, has a non-Iranic system of verbal prefixes conveying the perfective meaning, like in Slavic, and a similar system of using the Genitive in the Accusative meaning for animate agents (Абаев ВИ · 1965 · Скифо-европейские изоглоссы. На стыке Востока и Запада: 54–79 — https://drive.google.com/open?id=0B_7IkEzr9hyJXzBhVGFTd21oR2c).

    As to the anthropological evidence, Aleksejeva (Алексеева ТИ · 1973 · Этногенез восточных славян по данным антропологииhttps://drive.google.com/open?id=0B_7IkEzr9hyJTEZPWC1MbHd6cjg) states e. g. that (p. 273):
    На юге европейской части СССР намечается определённая линия антропологической преемственности: племена степной полосы эпохи бронзы (исключая трипольцев) — скифы лесной полосы — население Черняховской культуры — поляне.
    Учитывая антропологическое сходство украинцев со славянскими племенами Днепро-Днестровского междуречья, с одной стороны, и с полянами — с другой, можно сделать заключение о том, что в сложении физического облика украинского народа принимали участие наряду со славянскими элементами элементы дославянского, возможно, ираноязычного субстрата.

    [In the south of the European part of the USSR, there is a certain line of anthropological continuity emerging: tribes of the Steppe zone of the Bronze age (excluding Trypillians) — Scythians of the forest zone — population of the Chernyakhov culture — Polyans. Taking into account the anthropological similarity of Ukrainians with the Slavic tribes of Dnieper-Dniester interfluve, on the one hand, and with the Polyans, on the other, a conclusion can be drawn that, along with Slavic elements, those of pre-Slavic, possibly Iranic-speaking, substrate took part in the shaping of the physical outlook of the Ukrainian people].

    See also:
    https://en.wikipedia.org/wiki/Origin_hypotheses_of_the_Croats#Iranian_theory
    https://en.wikipedia.org/wiki/Serboi
    https://en.wikipedia.org/wiki/Origin_hypotheses_of_the_Serbs#Iranian_theory

    I know a saying that if you scrape a Russian then you will see the Tartar.
    I understand it so that Moscow was to be able to unify the surrendered small Russian principalities by learning the Tartar way of government.
    This proverb Grattez le russe et vous verrez le Tatare is usually ascribed to Voltaire, but people who made some research, conclude that it actually represents a later aphoristic summary of this statement from de Custine (https://en.wikipedia.org/wiki/Marquis_de_Custine):
    Celles des Russes, malgré toutes les prétentions de ces demi-sauvages, sont et resteront encore longtemps cruelles. Il n'y a guère plus d'un siècle qu'ils étaient de vrais Tatares; c'est Pierre-le-Grand qui a commencé à forcer les hommes d'introduire les femmes dans les assemblées; et sous leur élégance moderne, plusieurs de ces parvenus de la civilisation ont conservé la peau de l'ours, ils n'ont fait que la retourner, mais pour peu qu'on gratte, le poil se retrouve et se redresse.
    (http://www.gutenberg.org/cache/epub/25850/pg25850.html)

    The mentioning of Tatars here, as you see, is just a typical French witticism: the author used this word in its most general sense. That's actually part of the same tradition that proclaimed in e. g. 1914 that Europe ends on the French bank of the Rhine and has to defend herself from the Hunnic hordes (https://en.wikipedia.org/wiki/Huns#20th_century_use_in_reference_to_Germans).
     
    Last edited:
    Matasović (2008: 219) doesn't say all that much about the evolution of the BCS adjectival declension, but he seems to suggest that -og as an ending concurrent to -oga came to be by analogy to coexisting D/L sg. endings -om and -omu, which is what syncretising the dative and locative resulted in, rather than the other way round. However, in the next paragraph he states that syncretising of the definite dative and locative sg. can be observed from the 15th century, with the original dative forms prevailing at first which is opposite from the situation today, and in my opinion could suggest there may have been a "shortening" of the endings later on. On the other hand, written records of the vernacular from that time mostly come from a rather limited geographic area, so the syncretism could have simply given different results elsewhere...
    I'd also point at the preference of -im in the Dat.-Instr.-Loc. Pl., where, again, the shorter form (originally Dat. Pl.) has been chosen as opposed to -ima in the nouns (which, by the way, descend from only 2–3 earlier instances of the ī-Dual, like očima or ušima).
     
    I am surprised that these ancient tribes took part in the development of all Slavic nations.
    Well, first of all, the change g>ǥ>h occurs only in a part of Slavic (and the minor one, judging from the modern amount of speakers). Second, such changes often spread like waves, from one center, so it is enough if it had occurred in one compact area before the beginning of the Slavic migrations. In particular, South Russian differs in exhibiting the stage ǥ vs. h elsewhere, which can be regarded either as an archaism (and thus the south-eastern East Slavic as the source of this mutation) or, in contrast, as a later change induced by the Ukrainian/Belarusian influence (else, if it occurred before the Mongol and Lithuanian subjugation of the East Slavs in the 13–14th centuries, by the dialects that later became Ukrainian and Belarusian).
     
    On the other hand, the change *novajego > *novajega happened very early on, before first written records from the BCS area. It is usually explained as analogy to the G sg. ending -a of o-stem nouns. The innovation is shared with Slovene, where novega is still the only possible G sg. form of novi it's actually one of the isoglosses that separate the Western South Slavic (the Slovene-BCS continuum) from the Eastern South Slavic group (the Macedonian-Bulgarian continuum).
    Such an ending is also attested in northern East Slavic (Зализняк АА · 2004 · Древненовгородский диалект: 120 — https://drive.google.com/open?id=0B_7IkEzr9hyJUEhqQzJXT2p3ZUk). The Novgorod birch bark manuscripts (https://en.wikipedia.org/wiki/Birch_bark_manuscript#East_Slavic_texts) of the 11-12th centuries know only -аго/-ago (мьдвьнаго, ѹ котораго), which is attested for the last time around 1250–1275. -Ого/-ogo appears in 1220–1260 and becomes a norm since around 1275 (зеленого, великого, доброго). -Ога/-oga is first attested in 1271 (ѿ осмога дни) and preserves in modern northern dialects as -ова/-ova (молодова, старова), being first attested in this в-form around 1430 (головнова).
     

    Panceltic

    Senior Member
    Slovenščina
    It might be useful to mention that Slovene -ega, -emu forms are in fact recent (19th century) "artificial" creations that were introduced to the literary language during the efforts to "purify" the language. Prior to that (and even now in dialects and familiar language), the forms were -iga, -imu or even just -ga, -mu. Some dialects (particularly in Prekmurje) preserve the -oga, -omu forms.
     

    ilocas2

    Banned
    Czech
    In Czech G > H change happened in the 12th century, so I doubt that it has anything to do with Iranic tribes.
     
    In Czech G > H change happened in the 12th century, so I doubt that it has anything to do with Iranic tribes.
    As I had written, such changes spread like waves. Plus, the letter h began to be used for the outcome of g when the latter reached the glottal fricative stage: when it was still a velar fricative ǥ (https://en.m.wikipedia.org/wiki/Voiced_velar_fricative), the Latin letter was inappropriate, e. g. it isn't used for the Dutch sound (which was voiced until recently, groot, wage, and still is voiced in Flemish) and its counterparts in older Germanic languages or in Spanish (amigo).

    Update. Another, phonological, scenario of the origin of ǥ is presented e. g. in Shevelov GY · 1964 · A prehistory of Slavic: the historical phonology of Common Slavic: 593–595 (https://drive.google.com/open?id=0B_7IkEzr9hyJYUZ1ck5vdWE2Q1U), where it is explained as the attempt of the language to provide a voiced counterpart of x and to normalize the alternation k—č, g>ǥ>h—ž, x–š. Shevelov's argumentation is flawed, however, in that he regards the shift as if it were g>h, without this intermediate stage ǥ, which, again, couldn't be attested in writing and in German borrowings (in particular, the Low German g is ǥ now and was so in the past, thus the German Glesien, Gaussig etc. on p. 594 only prove that the Upper Sorbian sound wasn't h at that time). Otherwise, as with the substrate/adstrate hypothesis, this one is not without problems: in particular, I am not aware of any major attempts towards this development outside Central Slavic — both Northern and Southern Slavic idioms function perfectly without such a leveling.

    Interestingly, many Ukrainians don't realize that their г is actually h, and use g in the Latin transliteration. Moreover, the Ukrainian norm is to pronounce foreign surnames with g as h, e. g. Gordon/Гордон becomes [hordon], Angela/Ангела [anhela] etc.
     
    Last edited:

    franknagy

    Senior Member
    1. Ascherson, Neal wrote a good book about the peoples living around a Black Sea.
    2. The G has come back to the Czech and Slovak language from the Gerrman and Hungarian loanwords.
     
    Top