196 |

RECENZIJOS

(‘to move’) and eiti II (‘to hold office’); iššukti I (‘to remove by turning’ etc.) and iššukti II (‘to stray off’); iššaukti I (‘to revoke’) and iššaukti II (‘to be able to express’).

The inconsistencies noted in the presentation of the semantics of words, as well as other inaccuracies, do not diminish the dictionary’s value. The fact alone that the dictionary’s compilers have already refined the semantic structure of a considerable number of words shows their determination to prove that compiling Lithuanian dictionaries on the basis of modern lexicography is not a Sisyphean task: with a conception of lexical semantics and lexicography and a vast corpus, all that is needed is time and diligence.

The innovations discussed in the first Lithuanian–Norwegian Dictionary show that merely repeating the lexicographic experience of other languages is no longer considered sufficient: this dictionary lays the foundations of a modern lexicographic conception in Lithuanian lexicography. At first glance modest, this bilingual dictionary ranks among the finest works of Lithuanian lexicography in the quality of the grammatical and semantic information it provides about words in both languages. The compilers of the Dictionary of Contemporary Lithuanian, when preparing a new edition, as well as compilers of dictionaries of Lithuanian and other foreign languages, could also draw on its theoretical lexicographic ideas.

Dalia Barauskaitė Mikkelsen Dommervænget 26 C, 1.tv. 4000 Roskilde, Danija [email protected]

Received 2002 07 03

ONDŘEJ ŠEVČÍK, BOHUMIL VYKYPĚL, EDS., GRAMMATICUS. LINGUISTIC STUDIES PRESENTED TO ADOLF ERHART ON HIS SEVENTY-FIFTH BIRTHDAY.

Brno: Masarykova universita, 2001, 234 p.

The occasion for which the Festschrift under review has appeared is sufficiently explained by its title. There is ample reason to review it in this journal, not only because the esteemed honoree, Prof. Adolf Erhart, former holder of the chair of comparative linguistics at the University of Brno, has always devoted particular attention to Baltic studies (as evidenced by his introductions to Lithuanian and to comparative Baltic linguistics; see Erhart 1956, 1984), but also because Baltic topics are exceptionally well represented in this collection.

First, let us give a brief overview of the book’s contents. Etymological topics feature prominently. Special studies in this field have been contributed by Eva Havlová (“I.e. root *al- and Germanic terms for ‘age’”, pp. 56–63), Helena Karlíková (“On the Old Church Slavonic noun spodъ”, pp. 99–104), Boris Skalka (“Old Church Slavonic setъ/seti semper provocans”, pp. 144–148), and Pavla Valčáková (“Regarding Proto-Slavic *ědъ”, pp. 201–205). Some essays devote particular attention to one subfield of the Indo-European lexicon or to the lexicon of one of the Indo-European language branches. These include studies by Václav Blažek (“Indo-European kinship terms in *-ə₂ter”, pp. 24–33) and Ilona Janyšková (“Etymology of the names for yew in the Slavic languages”, pp. 91–98). Alongside Indo-European etymology, Dravidian and Altaic etymology are also represented (Jaroslav Vacek,

Recenzijos

| 197

“Dravidian and Altaic ‘hot – fire – heat’”, pp. 175–200). Karl Horst Schmidt addresses more general questions of etymological method, showing, through the example of Kartvelian-Indo-European parallels, how important it is in etymologizing to take recurrent and universal types of semantic motivation into account (“Etymologie und semantische Parallele”, pp. 135–143). Vladimir N. Toporov’s contribution, “Dr.-ind. prathamá- k rekonstrukcii odnogo iz variantov indoevropejskoj koncepcii «pervogo»”, pp. 165–174, belongs to the borderland between comparative grammar and the comparative history of religion.

From the field of Indo-Iranian studies, the studies by Vít Bubenik (“On the Remaking of the Middle Voice in Indo-Iranian”, pp. 34–44) and Ondřej Šefčík (“Word Borders in Old Indo-Aryan”, pp. 156–164) should be noted. Slavic syntax is represented by Radoslav Večerka (“Poznámky k disjunktivnímu/alternativnímu li v staroslověnštině”, pp. 206–210) and William R. Schmalstieg (“A Common Mistranslation of a Few Lines in the Igor Tale”, pp. 131–134). The latter demonstrates, with numerous parallels, that prepositional phrases otъ nego, rendered ablatively in most translations, should rather be understood agentively in a sentence from the Igor Tale (vidě těmoju vъše svojě voě prikryty ‘he saw his entire army covered by her with darkness’). The preposition otъ serves here to reinforce the agentive genitive, a phenomenon that Schmalstieg has also discussed in detail in connection with the Baltic data (especially those of Lithuanian). From the field of synchronic syntax, Stanislav Žaža’s study of the use of verbs meaning ‘give’ in Czech and Russian light-verb constructions (“K funkci slovesa dát – дамъ v češtině a ruštině”, pp. 226–232) should be mentioned.

The contributions by Ludmila Buzássyová and Eva Pallasová do not belong to any larger thematic group. The former deals with questions of Latin paradigmatics (“Parisylabá v latinčine”, pp. 45–55), while the latter concerns the early history of Slavic linguistics, specifically the treatment of word formation by Miklosich (Jskoni bě slovo – Počátky slovanské slovotvorby“, pp. 120–135).

Helena Kurzová discusses interesting questions of Indo-European case syntax in her contribution “Zur Syntax und Semantik der indogermanischen Kasus” (pp. 106–119). The article opens with a discussion of the “basic cases”, i.e. nominative and accusative. With regard to the proto-language system, the author emphasizes above all the categorial aspects in the semantics of these cases, that is, aspects linked to the categorial features of noun phrases. Every noun, she argues, was marked above all as animate or inanimate, and the configurational functions arose only from this categorial marking: “... the cases [were] also used semantically to express the position of nominal participants in the sentence. By marking as s-nominative and m/n-accusative, animates were marked for their opposite positions in the state of affairs denoted by the sentence” (p. 107). This, she argues, gave rise to a “subsequent functional identification of the neuter’s indifferent case (‘absolutive’ without a corresponding ergative) with the nominative on the one hand and the accusative on the other” (p. 106). No one would probably dispute the relevance of categorial features, but an absolutive without a corresponding ergative is a rather far-fetched concept from a functional and typological point of view. The absolutive is in itself a configurational case concept, since it represents the unmarked member in every (transitive or intransitive) configuration. The sole argument of intransitive structures is of course unmarked, but in transitive structures normally either the

198 |

RECENZIJOS

agent or the object is formally unmarked in the sense that the unmarked case coincides with that of the intransitive subject. Since both animates and inanimates must have been able to occur in transitive and intransitive structures, one may perhaps posit the nominative as the case of the animate intransitive subject and the absolutive as the case of the inanimate object in transitive structures. In this way, we essentially obtain a system with nominative marking for animates and ergative marking for inanimates. The author perhaps dismisses a little too readily the question of the typological plausibility of such a system. Her contribution also contains ideas on constructions with participant exchange (circumdare urbem muro: circumdare murum urbi), the deviant syntactic treatment of neuters (meaning the lack of case differentiation among neuters, expressed, for example, in contrast with id gaudeo vs. gaudeo de aliqua re), the relationship between sociative and dual, double accusative, and so on. The author advances the view, probably going back to Meillet, that the proto-language had no verbal inflection in the proper sense, but rather that the cases were “... forms independently marked for the referential structure of the state of affairs” (p. 117). Among the historically attested Indo-European languages, too, the author distinguishes those with more or less “advanced sentence formalization”. But are “proto-languages” subject to laws different from those of historically attested languages? One can of course distinguish between semantically marked and purely configurational cases: for instance, the accusative as a purely configurational case may contrast with the partitive genitive (marked with respect to quantification, cf. for example Lithuanian gerti vandenį: atsigeri vandens) or the dative (marked with respect to the low degree or temporal limitation of affectedness, cf. Latvian sist suni ‘to hit a dog’: iesist sunim ‘to give a dog a blow’). In many cases, however, the daughter languages introduced new distinctions of this kind. This applies especially to such archaic languages with richly developed case systems as Baltic and Slavic. When the case system is reduced, there are also fewer opportunities for semantic differentiation within a syntactic category (subject, object, etc.), which is precisely what creates the impression of “advanced sentence formalization”. It is not very plausible that the functions of the grammatical cases (nominative, accusative, genitive, dative) could be defined as wholly asyntactic in any language, and the assumption that the proto-language differed from the historically attested languages in this respect contradicts the principle of uniformity of linguistic developmental laws, implicitly accepted since the Neogrammarians (the “uniformitarian principle”).

In his article “Přemýšlení (nejen) o litevském kondicionálu” (pp. 64–90), Tomáš Hoskovec criticizes the conventional theory of the emergence of the Lithuanian-Latvian subjunctive, while incidentally offering no explanatory proposal of his own. The author’s conclusion is as follows (for convenience, I quote from the French summary): “Dans le système panchronique du verbe lituanien, il n’y a rien sur quoi puisse s’appuyer la théorie, qui fait pourtant grande tradition en linguistique comparée, selon laquelle le conditionnel lituanien serait issu d’une soudure du supin (présumé lituanien) avec une forme ancienne (présumée indo-européenne) du verbe être” (p. 89). The existence of such a “système panchronique” is now highly questionable. Syntactic change rests on the mechanism of structural reinterpretation (reanalysis): a given construction A receives a new structural interpretation without anything changing in its substance

RECENZIJOS

| 199

changes. This reinterpretation leads to the formulation of a new grammatical rule, which, like the original one, is capable of generating the initial structure A, but is also capable of generating a structure B, for which the original structural rule is no longer sufficient. The original structural rule on which A relied is no longer recognizable to speakers. As a result, the grammar of each generation differs at least in certain subdomains from the grammar of earlier generations, and the original system can at most still be grasped by the linguist. The notion of a panchronic language system seems to contradict that of language change. The author formulates his methodological approach as follows: „sledujeme, podle jakých principů se jednotlivé indoevropské jazyky coby systémy aktualizovaly z jisté společné potence výrazových prostředků, ptáme se, jak uvnitř jednoho jazyka charakterizovat individuální nakládání s morfologickým materiálem, který má obdobu i v jiných jazycích“ (p. 72). In many respects this approach may be sound, but no one can determine where the proto-language system ends and the individual language systems begin. After all, from the perspective of the development of the language system, the beginning of the written records of the individual languages is a purely arbitrary dividing line. Moreover, there seems to be no room here for the mechanisms of reanalysis and resegmentation, so important in historical morphology and syntax.

Incidentally, Hockovec is right to emphasize that various questions concerning the reconstruction of the origin of the Lithuanian-Latvian conditional remain unresolved. For the most important elements of this reconstruction, however, Stang has proposed an exceedingly valuable explanation, and it is somewhat surprising that the author acts as though no one had yet attempted to make the formation of a conditional on the basis of the supine semantically and syntactically plausible („O co by se vůbec opřela sémantika supínového syntagmatu?“, p. 86). Some of the questions raised in the article do, however, merit serious discussion. For example, the author asks (p. 86) how it is to be explained that the original supine survived rather than being entirely absorbed into the new conditional. This question can probably best be answered as follows. According to Stang’s brilliant hypothesis, the supine, just as the infinitive did in Slavic, took the place of an original participle in clauses with coreferential subjects, while the corresponding form of the auxiliary verb būti was attached to the conjunction: cf. Lithuanian *jei-b(i) valgytų instead of *jei-b(i) valgęs and Old Russian čto-by ěstъ instead of čto-by elъ (Stang 1966: 429–431). This construction must initially have been infinitival, since the only purpose of this substitution could have been to provide final subordinate clauses with an infinitival counterpart. It is striking, then, that the Baltic conditional forms are no longer capable of infinitival use (a construction such as *Reikia dirbti, kad valgytų ‘one must work in order to eat’ is not possible in modern Lithuanian), whereas the corresponding Slavic (Russian and Polish) sentences with the infinitive are still infinitival (Russian Nado rabotat’, čtoby est’). The most plausible explanation, then, is that the supine in this construction was reinterpreted as the finite third-person form, which then became the basis of a new paradigm. Once this process had taken place, the original connection with the still-infinitival supine (despite the presumed panchronic language system) became unrecognizable, and what remains now is simply a case of homonymy that is by no means disruptive. Had this finitivization not taken place, the emergence of the new conditional would of course have had to lead to the demise of the old supine. Another objection raised by the author, namely that it is impossible to

200 |

RECENZIJOS

regard the monosyllabic sequences in bim(e), bit(e) as forms of the verb būti, is not to be taken entirely seriously, since shortening in enclitic position is self-evident. There is therefore nothing to prevent comparison with Latvian biju (cf. recently Smoczyński 1999). No one denies that the first-person singular form in -čia, -čiau requires a special explanation—not even Stang. It remains unclear whether the author attaches any significance to the Old Lithuanian attestations, since features of Old Lithuanian spelling, such as -i- in (giwen)tumbim as against modern Lithuanian (gyven)tumėme, are declared irrelevant when a “panchronic” view of the language system already suggests a structural analysis that is satisfactory from the author’s perspective.

Zigmas Zinkevičius also discusses the Lithuanian subjunctive, specifically the question, already raised above, of the origin of the mysterious first-person singular form (“Pastabos lietuvių kalbos 1. sing. cond. formų istorijai”, pp. 224–225). He cites new evidence from Lithuanian dialects indicating that -čia is an older form in contrast to the analogical innovations -čiau and, less commonly, -čiu. This contradicts the view advanced by various scholars that -čiau arose through irregular (frequency-conditioned) shortening of -tumbiau.

In his article “Argument laryngalistyczny w etymologii bałto-słowiańskiej” (pp. 149–155), Wojciech Smoczyński examines several roots ending in a laryngeal which, in Balto-Slavic, incorporated an original formant through a shift in the morpheme boundary so long ago that, at first glance, one might be inclined to think that these were inherited ablaut patterns. Thus, Lithuanian sietas ‘sieve’ (Latvian siēts) presupposes a secondarily expanded root *seh₂i-, which was separated off by a shift in the morpheme boundary from the original -ie-present *sē-jo- (root *seh₁-, cf. sē-ti). Similarly, Slavic zbjati (alongside the later zijati) can be reconciled with the present žęjo only if one assumes that *gʰeh₁- was expanded in an original -ie-present to *gʰeh₁i-, with a renewed zero grade *gʰih₁- (instead of *gʰh₁i-); this assumption also leads to the further conclusion that Lithuanian žiōti is a recent formation arising through syncope from *žiōjoti (cf. niók-oti < *nijok-oti, from nijóks ‘not one’). The laryngealist reconstruction also suggests a new interpretation for Lithuanian siuvu (alongside Old Church Slavonic šijǫ): *si̯u- is not to be explained (as Trautmann assumed) by segmentation of the preconsonantal form *sīu-, but arose regularly from *siuH-é-, with u inserted as a hiatus-filler after the loss of the laryngeal. Accordingly, only Old Church Slavonic šijǫ is to be explained by generalization of the preconsonantal form; compare the preserved *šъv- in the past passive participle šъv-enъ. The Balto-Slavic alternations ūC: uuV, iC: iiV are thus to be understood as a direct consequence of the loss of intervocalic laryngeals. To conclude his discussion of laryngeals, Smoczyński proposes definitively abandoning Trautmann’s Balto-Slavic reconstruction *sātu- as the basis for both Lithuanian sótus and Slavic syťъ. The author now compares the latter with Hittite sunnai ‘fills’ < *sunh₃-, related to Indo-European *seuh₃- ‘be full’, whose zero grade is likely attested in the verbal adjective *suh₃-tó- ‘full, satiated’.

In the first part of his article “Zwei lettonistische Bemerkungen” (pp. 211–223), Bohumil Vykypěl attempts to specify the semantic foundations of the grammaticalization in Latvian of adjective formations of the type Lithuanian ilgōkas ‘rather long’ as comparatives. In the second part, he discusses the Latvian possessive construction man ir ‘mihi est’ and its chronological relation to Lithuanian and Old Prussian tureti, or

RECENZIJOS

201

turīt(wei) ‘to have’ apart. In the author’s view, the Latvian construction is the older one, while Lithuanian and Old Prussian have undergone the frequently observed transition in possessive constructions from the type mihi est to the type habeo. Evidence for this includes the Lithuanian construction man yra, preserved as a relic in the dialects of northern Lithuania bordering Latvia, and the numerous traces of the use of Lithuanian turēti in its original meaning ‘to hold’. The author’s arguments are highly convincing, and it is becoming increasingly clear that the view expressed by various researchers (Veenker, Rūķe-Draviņa)—that Latvian man ir is an innovation that arose under Finnic influence—must be rejected. This view reflects the widespread tendency in Baltic studies to interpret every difference between Lithuanian and Latvian as an innovation on the Latvian side, without more thorough investigation. The substitution of man ir for an original turēt ‘to have’ is implausible for three reasons: (1) typologically, only the transition from ‘mihi est’ to ‘habeo’ is well attested; (2) since the process of grammaticalization is fundamentally irreversible, the elimination of an original possessive construction with turēt ‘to have’ would hardly have led to its reintroduction with the original meaning ‘to hold’; and (3) Old Latvian turēt ‘to have’, attested here and there, is probably an example of what Harris and Campbell (1995: 55) call “exploratory expressions.” A particular construction may initially be used tentatively to express some grammatical meaning, and may subsequently either become grammaticalized or remain an occasional formation. Old Latvian turēt ‘to have’ was probably an attempt at grammaticalization that did not catch on. Of course, one must also consider the possibility that turēt ‘to have’ was created in Old Latvian texts as a translation equivalent of the German possessive construction: as the author notes (p. 220), accurately rendering the German construction in Teuw neues wairak Dewes turret required the use of a personal verb, since būt + dative had already occurred once in the sentence as a modal verb.

As the overview above makes clear, this rich and valuable collected volume, for which we may congratulate both the honoree and the editors, contains much that is new and worth reading in the field of Baltic studies.

LITERATURE

ERHART, A. 1956: Litevština. Praha: Státní Pedagogické Nakladatelství.

ERHART, A. 1984. Balt­ské jazyky. Praha: Státní Pedagogické Nakladatelství.

HARRIS, A. C., CAMPBELL, L. 1995: Historical syntax in cross-linguistic perspective. Cambridge: University Press. (Cambridge Studies in Linguistics, 74.)

SMOCZYŃSKI, W. 1999: Geneza starolitewskiego conditionalis na -biau, -bei, -bi-, Acta Baltico-Slavica 24, 13–18.

STANG, Chr. S. 1966. Vergleichende Grammatik der baltischen Sprachen. Oslo–Bergen–Tromsö: Universitetsforlaget.

Axel Holvoet Lietuvių kalbos institutas Vileišio g. 6, 2055 Vilnius, Lietuva [email protected]

Gauta 2002 04 26