Notes on the structure and meaning of some Slavic nominal compounds.
Usually, the topic is given a short overview at best, which is also the case in the fine new book on Slavic Nominal Word-Formation by Ranko Matasovic (2014: 183-189). In the following, a number of objections are raised with regard to both the proposed classification and to the interpretation of some compounds mentioned there.
In Slavic, as well as in the other branches of the family, nominal compounds are a large class of formations inherited from Indo-European. This heritage has undergone further developments, as usual. Perhaps due to the succinct form of his chapter, Matasovic (henceforth: M; numbers refer to pages of his book) has proposed to range Slavic compounds into three types only, viz. A Copulative Compounds, B Determinative compounds, and C Exocentric compounds.
Class A "Copulative compounds", called dvandva "pair" in Comparative Grammar (following the terminology introduced by Indian grammarian Pa[??]ini, ca. 500 BC) is without problems, but one should have added that though usually only two members occur, compounds with three and even more members are possible in IE, cf. Fr (drapeau) bleu-blanc-rouge, Skt hasty-asva-ratha-"(caused by) elephants, horses and chariots", garga-vatsa-vaja[??] "(descendants) of Garga, Vatsa and Vaja". (2) Such compounds with more than two members are not found in Slavic as far as I know. (3)
In discussing type B "Determinative compounds" (Skt tatpuru[??]a "his man" = "servant"), there is, unfortunately, a serious oversight in M's definition (183): it is, of course, the first member which determines the second, and not vice versa. The subdivisions proposed are too loosely formulated, I am afraid: 1. "two nominal elements", where "nominal" seems to mean substan tives only; 2. "a nominal and an adjectival element", with both orders possible; 3. "two adjectives"; 4. "a verbal and a nominal element". In fact, both adjec tives and substantives are nominals; this word class includes all kinds of verbal nouns as well (including root nouns and participles). Furthermore, semantics are helpful for a clearer subdivision. This is why one usually distinguishes between attributive, casual, adverbial relations of the determining (first) element to the determined (second) one. M's definition of type B.4 ("combine a verbal and a nominal element" [my italics]) seems to be open to misunderstanding. The two examples cited from OE and Gr both contain two nominal elements: OE [TEXT NOT REPRODUCIBLE IN ASCII]-lida (4) "sea-farer, sailor", Gr [TEXT NOT REPRODUCIBLE IN ASCII] "cattle-thief". There are no "verbal elements" here other than the verbal root which served as base for the derivation of the agent noun in OE, and as a simple root noun in Gr (resulting in a Governing compound, on which type see below).
Tatpuru[??]as with an attributive first member fall into two sub-types: those with adjective as first member (called karmadharaya (5) by Indian grammarians and in Comparative Grammar), e.g. G Rot-wein, E bluebird, etc., and those with a substantive or pronoun as first member (tatpuru[??]a proper). In compounds of substantive + substantive, the first member has to be understood in most cases as standing a casual relation to the second, i.e. if paraphrased, the first word of the phrase might stand in any possible case expressing all sorts of logical relations. We may speak of a comparative subtype if the meaning of the second member is modified by comparison to the first. Numerous compounds with an adjective as second member belong to this latter group, cf. G blut-rot, E blood-red and the like. An adverbial function of the first member may be found in cpds. of the type Gr [TEXT NOT REPRODUCIBLE IN ASCII] G Immer-grun.
Under C. "Exocentric compounds", two quite different groups are taken together, the bahuvrihi type and the Governing compounds. The latter, however, are not "exocentric" at all, and have to be discussed separately. The term 'exocentric' suits only those compounds whose referent is not directly expressed by one of the lexemes used in the compound. As a rule, all these compounds, whether their second member is an adjective or a substantive, are adjectives. They may, of course, be used secondarily as substantives at all times in suitable context. The term bahuvrihi itself is a transparent and illuminating example of the type: although formally it could be a karmadharaya compound meaning "much rice", its usage as an adjective shows that the formation refers to something which is "characterized by much rice" i.e. "fertile land". (6) G Rotkappchen, a fairy tale figure, is a girl "characterized by her little red hat" (not "a small red cap"!). (7) In Slavic, many bahuvrihi compounds with a numeral in first position are known, e.g. tri-glav[??] (theonym and oronym: a god or peak "with three heads"), (8) [TEXT NOT REPRODUCIBLE IN ASCII] "trident" (an instrument with "three teeth", clearly a loan translation from Greek). (9)
The last main type to be discussed here are Governing Compounds (G term: Verbale Rektionskomposita). (10) Here, one element is a verbal noun and the other designates the object of the action expressed. Both orders of elements are old and should be assumed for (at least, late) IE. In scholarly literature, one often meets with the opinion that the 'verbal' element is an imperative. It looks indeed like an imperative in many cases, especially in Modern Romance languages (e.g. Fr nettoye-pipe "pipe cleaner", It stuzzica-denti "toothpick"); but this is popular etymology, not linguistic classification. In Slavic, the type is mostly (11) found in proper names (see below). In the older languages, this type quite often occurs in poetic texts, and the 'verbal' element can hardly be understood as an imperative, cf. Gr [TEXT NOT REPRODUCIBLE IN ASCII] "carrying the house" = "snail" and [TEXT NOT REPRODUCIBLE IN ASCII] "rejoicing humans", besides, with 'reverted' members, [TEXT NOT REPRODUCIBLE IN ASCII] "cattle hitting" = "whip" and Skt visva-vid- "knowing everything". (12)
II. Selected Slavic compounds
As stated above, I am unable, unfortunately, to agree with my old friend R. Matasovic's interpretation of a number of compounds. The following few objections and counter-proposals should be understood as part of a friendly discussion, and not as punches in a brawl. In all such disagreements, we should not forget that, after all, linguistics is not mathematics, i.e. absolute truth is unreachable.
1. The theonym *[TEXT NOT REPRODUCIBLE IN ASCII], [TEXT NOT REPRODUCIBLE IN ASCII] Daz[??]bog[??], OPol Dadzibog (M 185) is obviously a compound of some form from the common verb dati "to give", which originally had a reduplicating present still visible in Ru, Bulg (and some SCr dialects) plural forms, and bog[??]; but the latter word does not mean "god" here at all (it may well have been re-interpreted as such later, of course, after Greek names like [TEXT NOT REPRODUCIBLE IN ASCII], [TEXT NOT REPRODUCIBLE IN ASCII]). As known since long, and shown repeatedly, a.o. by me (cf. lastly Zimmer 2012), the original meaning of Slavic bog[??] is "share, part, portion", presumably in the booty or the food (also at rit ual meals) to be divided among those entitled. The older meaning is still trans parent in a few other words not cited in M's sketch: u-bog[??] "poor" and bog-at[??] "rich", *[TEXT NOT REPRODUCIBLE IN ASCII] in *[TEXT NOT REPRODUCIBLE IN ASCII] "property" (cf. Cz zbozi "property, riches", Po zboze "corn", Ukr zbizze "id.", etc. (13)). How [TEXT NOT REPRODUCIBLE IN ASCII] could have acquired the meaning "god" among Slavs is an old problem discussed many times. I hope to have shown a convincing way for understanding the semantic development in which just this latter *s[??]-bog[??] is central. It is identical with an Indo-Iranian compound attested in Skt su-bhaga-, Av hu-ba[??]a- "well characterized with share(s)", later "rich, happy", etc. (for details, see Zimmer 2012). (14) It should be noted in passing that the widely used element IE *h1su- (Skt su-, Av hu, Sl s[??]-, etc.) is an adverb, not an adjective (therefore to be translated literally by "well"; M (188) keeps to the adjectival "good"), cf. *[TEXT NOT REPRODUCIBLE IN ASCII], OCS [TEXT NOT REPRODUCIBLE IN ASCII], etc., "healthy", lit. "well equipped with firmness or strength". (15)
2. Are [TEXT NOT REPRODUCIBLE IN ASCII], Ru Vladimir purely Slavic formations (M 185)? Alternatively, one might think of an Old Norse model such as ODan Valdemar (cf. also Volomar, a prince of Minsk). In OIc, -valdr occurs as second member in proper nouns only. This is the common Germanic type, cf. Marcomannic Cat-valda, OFranc Chario-valda, etc.
3. Croat kazi-prst "index finger" (M 185) looks like a translation of G Zeige-finger. The hypothesis of an independent Sl cpd. would be stronger if OCS or ORu attestations could support it.
4. *[TEXT NOT REPRODUCIBLE IN ASCII] "lord" is traditionally regarded as corresponding to L hospes "host". If true, the first element is *[TEXT NOT REPRODUCIBLE IN ASCII] "guest", and the second is IE *poti- "lord, master". But M (186) rightly remarks that Sl -d- would remain unexplained by that etymology. This invites us to reconsider Vaillant's explication of *gospod? as a loan from Germanic *[TEXT NOT REPRODUCIBLE IN ASCII]- (or [TEXT NOT REPRODUCIBLE IN ASCII]). Unfortunately, this is not attested in any Germanic language (but of course, a possible word, perhaps an archaism lost early). Slavic would have substituted the foreign sounds by familiar ones (-f- by -p-, -a- by -o-, -[TEXT NOT REPRODUCIBLE IN ASCII]- by -d-). (16) By the way, the well-known L hospes < *ghosti-potis is not without qualms either. (17) Instead of posing a second member *poti- "master", one has also proposed to find a form of the root *pet- "to seek, make for" so that the PIE *ghosti-potis would have been "looking for/after guests", vel sim. (18) This makes good sense with regard to the great importance of hospitality in PIE society. (19)
5. The plant name *[TEXT NOT REPRODUCIBLE IN ASCII] "dog's tongue" is traditionally explained (M 186) as a compound of "cow" and "tongue" but this is at least not immediately transparent. How should the word be analysed? OCS [TEXT NOT REPRODUCIBLE IN ASCII] "cattle" and the other pertaining attestations are derivatives of the PIE word for "cow", but details are unclear (cf. M 81-82). The second member also is far from clear; why should *[TEXT NOT REPRODUCIBLE IN ASCII] have lost its -k-[??] Is it really a loan translation from Gr [TEXT NOT REPRODUCIBLE IN ASCII]? (20)
6. The term for "threshing-floor",*gum[??]no-, is regarded by M (186) as "atypical, as the nominal element is not the object of the verbal root, but its (locative) adjunct" (186). The roots of the two constituents are probably gu- "cow" and*[TEXT NOT REPRODUCIBLE IN ASCII]- "to trample, step", but the composition type has not been recognized. It simply is a bahuvrihi "characterized by the trampling of cattle", viz. a special place or spot, the threshing-floor, as already Vaillant and others have seen.
7. The inherited status (M 187) of the taboo word*medv-ed[??] "bear" < "honey eating" = Skt. madhv-ad- (epithet of birds) may be supported by a similar compound in Greek: [TEXT NOT REPRODUCIBLE IN ASCII] "bee" < *[TEXT NOT REPRODUCIBLE IN ASCII] "honey-licker" (by haplology; cf. the Skt poetic term madhu-lih- "id.").
8. Many Slavic compounds are clearly loan translations or calques. (21) Some have been mentioned before. All the words given by M in the last para of p. 187 are also such formations, e.g. OCS bogo-rodica < Gr [TEXT NOT REPRODUCIBLE IN ASCII], Croat pismo-nosa < G Brief-trager, Ru miro-ljubie < G Friedens-liebe. Even in a brief chapter on compounding, the possible model(s) of every such formation could be helpful. In Slavic, as well as in Germanic and other languages, the need to translate (esp. the Bible and other Christian literature) into the vernacular is responsible for the majority of nominal compounds. But inherited formations, often coming from the common IE poetic tradition, are the basis, and should accordingly be highlighted - and, as far as the textual sources permit, clearly distinguished from the newer coinages.
Forssman, Bernhard (1989). Vedisch athitipati-, lateinisch hospes. In: Jasanoff, Jay et al., eds., Mir Curad. Festschrift Calvert Watkins. Innsbruck: IBS, 115-126.
Ernout, Alfred, Antoine Meillet (1937). Dictionnaire etymologique de la langue latine. Paris: Klincksieck.
Heidermanns, Frank (2002) Nominal composition in Sabellic and Proto-Italic. Transactions of the Philological Society 100: 185-202.
Loma, Alexandr (2009). Krieg und Frieden - vom Indogermanischen zum Griechischen. In: Luhr, Rosemarie, Sabine Ziegler, eds. Protolanguage and Prehistory. Wiesbaden: Reichert.
Matasovic, Ranko (2014). Slavic Nominal Word-Formation. Proto-Indo-European and Historical Developments. Heidelberg: Universitatsverlag Winter.
Pinault, Georges-Jean (1998). Le nom indo-iranien de l'hote. In: Meid, Wolfgang, ed. Sprache und Kultur der Indogermanen. Innsbruck: IBS, 451-477.
Pohl, Heinz Dieter (1973). Verbale Rektionskomposita mit regierendem Vorderglied im Slavischen und anderen indogermanischen Sprachen. Wiener Slavistisches Jahrbuch 19, 190-202.
Pohl, Heinz Dieter (1974). Die slavischen zusammengesetzten Personennamen. Osterreichische Namenforschung, 33-40 [repr. in Namenkundliche Informationen, Beiheft 17, Leipzig 1993, 7-20].
Pohl, Heinz Dieter (1977). Die Nominalkomposition im Alt- und Gemeinslavischen. Ein Beitrag zur slavischen, indogermanischen und allgemeinen Wortbildung. Klagenfurt: Klagenfurter Sprachwissenschaftliche Gesellschaft.
Speyer, Jacobus S. (1896). Vedische und Sanskrit-Syntax. Stra[sz]burg: Trubner.
Vaillant, Andre (1950-1977). Grammaire comparee des langues slaves. 5 vols. Paris: Klincksieck.
Vasmer, Max (1953-1958). Russisches Etymologisches Worterbuch. 3 vols. Heidelberg: Winter, repr. 1979.
Wodtko, Dagmar, Brita Irslinger, Carolin Schneider (2008). Nomina im indogermanischen Lexikon. Heidelberg: Universitatsverlag Winter.
Zimmer, Stefan (1984). Iranisch baga- - ein Gottesname? Munchener Studien zur Sprachwissenschaft 43, 187-215.
id.: (1992). Die umgekehrten Bahuvrihi-Komposita im Kymrischen und Indogermanisches. In: Beekes, Robert et al., eds. Rekonstruktion und relative Chronologie. Innsbruck: IBS, 421-435.
id.: (1994). Griech. [TEXT NOT REPRODUCIBLE IN ASCII] (-) usw.: Ablautstufe und Wortart. Munchener Studien zur Sprachwissenschaft 55, (recte 1995), 157-171.
id.: (2000). Studies in Welsh Word-Formation. Dublin: DIAS.
id.: (2012). Wie kommt slav. bog[??] zur Bedeutung "Gott"? In: Podtergera, Irina ed. Schnittpunkt Slavistik. Ost und West im wissenschaftlichen Dialog. Festgabe fur Helmut Keipert zum 70. Geburtstag. Gottingen, vol. 3: V&R unipress, 133-141.
id.: (2016). Three Indo-European Values (Halle Conference Papers, ed. H. Bichlmeier et al., in press for 2016).
(1) The only attempt to such a comprehensive description of Slavic nominal composition I am aware of is the one by Pohl (1977). See also his two other studies (Pohl 1973, 1974). There are, of course, many publications on nominal compounds in various Slavic languages.
(2) The longest known of such enumerative compounds in the older layers of Skt literature has 22 members, see Speyer (VSS[section]106).
(3) Cf. also the similar [TEXT NOT REPRODUCIBLE IN ASCII]- ("repeated"), i.e. iterative compounds (e.g. L quis-quis "whoever", Skt dive-dive "day for day, daily", eka-eka- "one after the other, every single"). They exist(ed), albeit sporadically, as adverbs in Slavic, e.g. Ru gde-gde (archaic) "somewhere, here and there".
(4) Sic. M prints sae-lidan.
(5) A term of unknown literal meaning.
(6) In Vedic Sanskrit (as already in PIE) the mobile accent distinguished between Tatpuru[??]a and Bahuvrihi : the first type is accented on the second member, the second on the first. Classical example are Skt raja-putra- "son of a king" vs. raja-putra- "whose sons are kings" (cf. RV 2,27,7, said of the goddess Aditi), and Greek [TEXT NOT REPRODUCIBLE IN ASCII] "feeding wild animals" (said of a man or a place) vs. [TEXT NOT REPRODUCIBLE IN ASCII] "feeding on wild animals" (said of a drake), both used by Euripides. In all other languages, having lost or replaced the free IE accent, only syntax and semantics can decide. They are quite often further characterized by the addition of an adjectival suffix in order to secure proper understanding. - Note that in poetic language, widely influencing the giving of proper names, a 'reverted' or 'inverted' subtype of bahuvrihi is found in many languages (not in Slavic), see Zimmer (1992, 2000).
(7) The G non-diminutive formation Rotkappe of practically the same meaning is the name of an edible mushroom.
(8) Also as a family name in G, mostly spelled Triglaff.
(9) M. cites a OCS tr[??]zab? 'trident' (189), but calls it erroneously a "determinative compound". This type is fundamentally different from ino-c[??]d[??] "only child" (cited in the following line), a clear karmadharaya compound: "a child which is single/alone", whereas a bahuvrihi like tr[??]zab[??] is not "a tooth which is three" but something which is 'characterized by three teeth'.
(10) Cf. Pohl (1973).
(11) But not exclusively, cf. [TEXT NOT REPRODUCIBLE IN ASCII] "evildoer" (transl. Gr [TEXT NOT REPRODUCIBLE IN ASCII] and also [TEXT NOT REPRODUCIBLE IN ASCII], so not necessarily a calque); voje-voda "(military) leader", perhaps a calque of OHG heri-zogo (or of a corresponding form from another Old Germanic dialect), or neto-pyr[??] 'bat', lit. 'flying by night' (the first member seems to have been modified for taboo reasons) and medv-ed[??] 'bear', lit. 'honey-eating'.
(12) Note the accent of the Skt word which excludes an interpretation as bahuvrihi.
(13) More in Vasmer.
(14) By the way, "iran. baga- 'Verteiler'", as cited by Pohl (1977: 39) does not exist. In Iranian, baga- either means "share, portion" (perhaps in a few difficult Avestan passages) or "god" (in Avestan, Old Persian, and still in classical Kurdish poetry, as Sebastian Heine kindly tells me).
(15) See the detailed argumentation in Zimmer 1995.
(16) See also Pohl (1977: 36-38). Loma (2009) gives a review of alternative explanations.
(17) It is hardly possible to say whether the (three?) terms are inherited from PIE or are parallel formations. For semantics, remember that L hostis is not always the "enemy", as Classical literature may suggest. Its basic meaning is "foreigner", perhaps from "standing alone" (Heidermanns 2002: 190-191). See also Wodtko (2008: 173), with references.
(18) The first is the standard etymology, the second is to be found in Ernout-Meillet (1932: 441). For the pros and cons of both, see the considerations offered by Forssman (1989).
(19) Cf. the explanation of Vedic athiti- "guest" as "(who is) standing at the side (of the pater familias)" by Pinault (1998), another semantic variant to the two possible senses of hospes as suggested by the etymologies just cited. For the institution, cf. also Zimmer (2016, in press).
(20) For the meaning, cf. the L and G plant name Cynoglossium, Hundszunge. On the one hand, it is hard to assume that the Slavs should have 'translated' a Greek name referring to a cow's tongue literally when they thought of a dog's tongue. On the other, if the word were a loan translation from L, why did they use a word for cow as first member?
(21) Discussion of these terms (definition, difference, etc.) is avoided here.
|Printer friendly Cite/link Email Feedback|
|Date:||Jul 1, 2016|
|Previous Article:||Relative Constructions in European Non-Standard Varieties.|
|Next Article:||Stereotypes as an interdisciplinary construct: Implications for applied linguistics research/Stereotip kao interdisciplinarni konstrukt: implikacije...|