Östen Dahl, Maria Koptjevskaja-Tamm, eds. “The Circum-Baltic Languages: Typology and Contact”

ACTA LINGUISTICA LITHUANICA XLIX (2003), 131 - 152 RECENZIJOS Reviews OSTEN DAHL AND MARIA KOPTJEVSKAJA-TAMM, EDS. THE CIRCUM-BALTIC LANGUAGES: TYPOLOGY AND CONTACT. VOLUME I: PAST AND PRESENT. Amsterdam & Philadelphia: John Benjamins, pp. xx + 360 + (separately paginated) 22 index pp. (Studies in Language Companion Series, Vol. 54.) OSTEN DAHL AND MARIA KOPTJEVSKAJA-TAMM, EDS. THE CIRCUM-BALTIC LANGUAGES: TYPOLOGY AND CONTACT. VOLUME 2: GRAMMAR AND TYPOLOGY. Amsterdam & Philadelphia: John Benjamins, pp. xx + 400 + (separately paginated) 21 index pp. (Studies in Language Companion Series, Vol. 55.) According to the Introduction the first volume of the two-volume series: “... surveys important sub-groups in the present-day Circum-Baltic languages, placing them in their geographical, his- torical and societal setting and discussing specific contact situations’ (p. xvii). The languages surveyed in part I include Lithuanian, Latvian, southeastern Baltic Russian dialects, Swedish dialects around the Baltic Sea, and the Finnic languages. The second volume “... focuses on grammatical phenomena in the Circum-Baltic languages, relating them to the larger typological perspective’. The introduction, the name and subject indices are printed in both volumes, but in volume 2 the last page of the subject index has been omitted. Laimute Balode and Axel Holvoet’s articles about Latvian (pp. 3-40) and Lithuanian (pp. 41-79) are well-organized and concise descriptions of modern Latvian and Lithuanian and their dialects, although there is relatively little about language influence, except, perhaps for the mention of Livonian influence on Tamian Latvian (p. 27). One apparent oversight that could cause confusion is the transcription of Lithuanian [~mjz:ta] ‘throw(s)’ where the [j] seems to contrast with the simple palatalization encountered in [~mz:tat &}/[~m z:tat] ‘you throw.” (p. 5). Valeriy Cekmonas’ article “Russian varieties in the southeastern Baltic area: Urban Russian of the 19th century” (pp. 81-99) discusses the Russian used in Vilnius, Riga and Tallinn. In Vilnius the editors, authors and correctors tried to emulate the language of St. Petersburg and Moscow. Particularly interesting are the mistakes in Russian made by Lithuanians in various petitions, e.g., ibo is¢ezli mezdu nami litovskie knizki ‘Lithuanian books have disappeared among us’, where meždu nami is a calque for Lith. tarp mūsų ‘among us’ (p. 89). In Riga, which has a I should like to thank herewith Martti Nyman of the University of Turku for advice on some of the Finnish syntactic structures discussed in this review. I am particularly grateful to Virginia Vasiliauskiené of the University of Kaunas for her help with numerous aspects of this review. Neither of these is, of course, responsible for any errors which I may have made. 132 | RECENZIJOS longer history of Russian population, one of the phonological characteristics is the hard pro- nunciation of ž, Čč, š, šč, e.g., ščy ‘cabbage soup’. A morphological feature is the occurrence of plurals of the type gldzy ‘eyes’, rogi ‘horns’, xi ‘ears’ instead of the standard Russian glazd, roga, usi (pp. 92-93). Tallinn was part of the Russian empire since 1710, but after this time Russian was limited pretty much to the military garrison. According to Cekmonas’ article “Russian varieties in the southeastern Baltic area: Rural dialects” (pp. 101-136) the Russian language in this area derives primarily from the Old Believers seeking religious freedom (p. 101). Cekmonas lists a number of phonetic, morphological and syntactic features characteris- tic of these dialects. One of the most interesting syntactic features is the occurrence of struc- tures of the type v (1) men'a tavaru vz'ata ‘1 have taken some goods’ (p. 118). Čekmonas writes: “The preposition v (u) is used with the instrumental (sic!) plural of nouns to express the meaning of the Standard Russian u kogo ‘whose, who has/have’: v kupcax vsegda deneg mnoga ‘the merchants always have much money’ (p. 120). Certainly here the word ‘instru- mental’ should be replaced by ‘prepositional’ or ‘locative’. Čekmonas reports that there are no phonetic traits in the Old Believer dialects which could be connected with foreign influence. One morphological trait which might be so identified is the change of neuter to feminine, e.g., m’ésta ‘place’, at'ėc zanimal xarošuju m’estu ‘smb’s father occupied a good post’ (p. 125). Another morphological feature common to the Russian dialects of East Lithuania may be the result of Lithuanian influence, viz., the occurrence of the 3rd singular verb with a plural sub- ject, cf., e.g., ani vaz’'m’6t ‘they will take’ (p. 126). Many Lithuanian vocabulary items have penetrated the central Lithuanian Russian dialects, e.g., uskur ‘a married man who lives in his wife’s house’ (< Lith. uzkurys). The article is marred by numerous editorial infelicities, e.g., on p. 85 within the course of two consecutive lines we encounter croun (> crown), Ydish (> Yiddish) and the definition reach people (> rich people). Other mistakes include Kazlauskas > Kulikauskas (p. 89), pleophiny > pleophony (p. 112), bye > buy, by > buy (p. 117), innter > inner (p. 118), guagmire > quagmire (p. 127). The articles “Swedish dialects around the Baltic Sea” (pp. 137-177) by Anne-Charlotte Rendahl and “The Finnic languages” (pp. 179-212) by Johanna Laakso are solid and interesting, but say little about language contact. Osten Dahl's article “The origin of the Scandinavian languages” (pp. 215-235) concludes that: “Germanic-speaking groups arrived to the very westernmost corner of the Baltic (present-day Germany and Denmark) somewhat before the beginning of our era. A little later they expanded eastwards as far as Uppland on the north side and the Vistula estuary on the south side of the Baltic. During the ensuing half millennium, the languages of the different Germanic groups became differentiated, exactly how much we do not know” (p. 231). Lars-Gunnar Larsson’s article “Baltic influence on Finnic languages” (pp. 237-253) begins with many examples of Baltic loan-words in Finnish in the fields of animal husbandry, e.g., Fin. vuohi ‘goat’, cf. Lith. ožys, body parts, e.g., Fin. kaula ‘neck’, cf. Lith. kaklas, family relationships, e.g., Fin. heimo ‘tribe’, cf. Lith. šeima ‘family’, etc. (pp. 239-240). The only possible phonetic influence of Baltic on Proto-Finnic would be the change of the sequence *#i to si, i.e.. the palatalization of *t by a following front vowel, cf. Fin. silta ‘bridge’ with Lith. tiltas. However, as Larsson correctly remarks (p. 243), the Proto-Baltic palatalization of consonants by following front vowels is rather unlikely. Larsson discusses two similar syntactic phenomena, the use of the partitive genitive both as subject and as object and the agent participle in Finnish. An example RECENZIJOS | 133 of the first phenomenon is Fin. Metsdssa on susia (part. pl.) ‘In the forest there are wolves’ = Lith. Miške vilkų yra (p. 245). An example of the second phenomenon is Fin. Isdn (gen. sg.) ostama auto ‘the car bought by father’ = Lith. Tévo (gen. sg.) perkamas automobilis (p. 247). It seems to me, however, that there is a difference in force between the two participial constructions which the English translation (correct for both phrases) does not catch. The Finnish construction expresses a completed action and the Lithuanian an action in progress, more like ‘the car which is being bought by father’. Thus Finnish /sčin ostama auto can be paraphrased as Auto, jonka (acc.) isd osti / on ostanut ‘Car that father bought/has bought’. In this case, the Finnish construction certainly expresses a completed action. On the other hand the corresponding Lithuanian construction cannot be paraphrased by automobilis, kurį (acc.) tėvas pirko | nusipirko, since perkamas has a durative meaning. Larsson also points out that Mari -m- participles differ from Finnish in that they can be formed from intransitive verbs (p. 248), thus Eastern Mari: memnan (gen.) tolmo korno ‘the road that we have come’, lit. ‘the by us come road’, but similar constructions are possible in Lithuanian also, e.g., Cia kelelis pésciy (gen. pl.) einamas (nom. sg. masc. pres. psv. part.) ‘here the road is traveled on by the pedestrians’ and Cia mano (gen. sg.) eitas (nom. sg. masc. past psv. part.) kelias ‘Here is the path that I have gone on’ (LKZ, 1069). Although the correspondences between Finno-Ugric and Baltic are not exact it seems a remarkable coincidence to me that Finno-Ugric has both a participle in -m- and a participle in -t- which have somewhat (but not exactly!) similar functions to the Indo-European participles with -m- and -t-. Collinder (1964: 48, 57) would see this and many other features as evidence for a common inheritance. One might compare also the curious Hungarian parallel. Kéroly (1972: 119) writes: ‘the participle with the formant -t/-tt can be found in subjective genitive construction... az [demonstrative] én [pronoun] vdlasztottam [participle] ‘one chosen by me’. (The grammatical definitions in brackets are mine — WRS.) This seems to me to be very similar to Lith. šis mano parinktas (daiktas)... ‘the (thing) chosen by me...', although influence of Baltic on Hungarian seems unlikely if not impossible, at least within historic times. Even the fact that the Baltic and Finno-Ugric constructions are somewhat different in meaning would seem to argue for a common source, rather than outside influence, where more exact correspondences might be expected. Perhaps instead of trying to interpret Finnic syntax in terms of Baltic, one should try to interpret Indo-European syntax in terms of Finno-Ugric. It seems to me that the syntactic development of both Indo-European and Finno-Ugric needs deeper analysis before wide-ranging conclusions can be drawn about Baltic influence on Finnish syntax. Stefan M. Pugh’s article “The rale of language contact in the formation of Karelian, past and present” (pp. 257-270) describes a language which seems to have been almost overwhelmed by Russian. From the information supplied in Pugh’s article, it seems to me that Karelian faces a very uncertain future. The loss of a language is certainly a matter of regret for human cultural history, but such losses are on the other hand a boon for those American syntacticians who base their argumentation almost exclusively on their native English. If there are fewer languages, then obviously there will be fewer uncomfortable counter-examples to the language universals established by such scholars. Indeed, if English becomes the only world language, then language universals could indeed be based on English. For me this consideration calls into question also the significance of determining the number of languages exhibiting a specific typological feature. 134 | RECENZIJOS The existence of a language depends on the existence of a population speaking that language, not the nature of the language. Perhaps if the war-like Indo-Europeans had not invaded Europe, that continent would contain many more ergative languages like Basgue. After reading Eva Agnes Csat0’s article “Syntactic code-copying in Karaim” (pp. 271-283) I wondered about the future of that language also, cf. such a sentence as I uže bu fotograf turat kolo Bas yanin ‘And this photographer is already standing near Basia’ (p. 275) in which the Karaim words are bu ‘this’, turat ‘is standing’ and the name Basia, which is here in the (Karaim) genitive case as required by the Polish preposition kolo. The author argues that in Karaim °...the use of postpositions is disharmonic with the other syntactic properties of Karaim’ and that “The path of syntactic change in Karaim is, however, not postposition = preposition, but rather noun-like postposition = particle like postpositions’ (p. 276). Since I don’t know Karaim morphophonemics sometimes the reasoning is hard for me to follow. Thus the preposition referred to as bila and glossed as ‘with’ apparently appears without the final syllable in sequence: m’en’imb’a yanasa I: GEN. with next to ‘next to me’ In the author's view this “...illustrates a possible development in the grammaticalisation of postpositions as case endings. Compare this Karaim construction to the use of the Polish preposition przy ‘by, at’, which governs the instrumental (sic!) case’ (p. 281). I could not find any source documenting the use of Polish przy with the instrumental case and in fact under the heading pri Vondrak (1928: 312) writes: “Als Prap. hat es im Slav. nur den Lok. bei sich...”. Probably the translator of this article just wrote ‘instrumental’ for ‘locative’ or ‘prepositional’, curiously enough making the same mistake as in Cekmonas’ article mentioned above. Still such editorial oversights are unfortunate particularly for persons unfamiliar with the material. Neil G. Jacobs’ article “Yiddish in the Baltic region” (pp. 285-311) is a well-organized and competent (although necessarily brief) outline of the main features of that language in the Circum- Baltic territory. In his section on Lithuanian influences in Northeastern Yiddish Jacobs relies primarily on the works of Lemkhen (Lemchenas), thereby making available in English (probably for the first time) some of the results of that outstanding specialist’s study of Lithuanian impact on that variety of Yiddish. The bulk of Lithuanian influence on Yiddish is lexical and Lithuanian vocabulary borrowed into Yiddish conforms to the phonology of Yiddish. For example, the distinction between sibilants and shibilants, characteristic of Lithuanian (and many other languages) is lost and there is rather some kind of intermediate sound. Lemkhen writes that in the word borrowed from Lith. speigas ‘big frost’ he cannot say whether it is špejg- or spejg-. On this phenomenon see Martinet (1951: 92). In his article “The North Russian Romani dialect: Interference and Code Switching” (pp. 313-337) Aleksandr R. Rusakov writes that this dialect “...is a very typical example of strong linguistic interference” (p. 313). In this dialect under Russian influence a new verbal system is arising, in which the old native opposition of imperfect vs. aorist disappears, the old aorist serving as a general past, and the old imperfect serving as a rarely used form with iterative meaning. Almost every verb may use borrowed prefixes on the Russian model (p. 315), e.g., te ot-des ‘to give back’ (Russ. ot-dat’), te vy-des ‘to give out’ (Russ. vy-dat’), te roz-des ‘to distribute’ (Russ. roz-dat’). According to the author one encounters in the dialect “...the extensive use of RECENZIJOS | 135 grammatically unadapted Russian lexical elements' (p. 323). An example is: Da nat, me prosto na dumind'om, so me tut date vstrechu which is supplied with the somewhat unidiomatic English translation ‘No, I didn’t just think that I meet you here’ (p. 323). Unfortunately this is not supplied with any grammatical commentary. Someone with a knowledge of Russian can pretty well figure this out, but I doubt that a person who knows no Russian would have any idea of the extent of the code-switching. The author concludes that the speakers of the Romani language themselves feel a need to preserve it “...as a special, secret language, which cannot be understood by outsiders’ (p. 334). Therefore, one assumes, that differently from Karelian and Karaim, also discussed in this volume, the North Russian Romani dialect might not die out. Valeriy Cekmonas’ article “On some Circum-Baltic features of the Pskov-Novgorod (Northwestern Central Russian) dialect” (pp. 339-359) considers several phonological phenomena from the point of view of the substratum theory. According to the author the phenomenon of cokan’e (the merger of *¢ with *c) “...was one of the most characteristic features of the old Pskov... and old Novgorod dialects...” (p. 341). An examination of the function and the distribution of ¢ and ¢ of the various substratum Finnic languages of the Circum-Baltic area suggests that they might stimulate the coalescence of c and ¢ into ¢ in Russian dialects. Furthermore there are two other phenomena which must be taken into consideration. The first is the merger of the soft *s and *z and soft *š and *ž into *s”, *z" (an intermediate sound between sibilant and shibilant) and the second is the free (and sporadic) alternation of hard *s and *z with (hard) *š and *ž (p. 346). A similar phenomenon is Lithuanian §lekiavimas as described by Girdenis and Pabréza (1978: 127-129). The lack of a contrast between /s/ and /š/ in Finnic is an ancient feature, and the Russian phenomena in question could be explained in this way. Another phenomenon which might be ascribed to Finnic influence is the unmotivated change of voiced to voiceless consonant and vice-versa, e.g., buza < puza ‘belly’, vybucit’ (glazd) < vypucit’ ‘to open one’s eyes wide’ (p. 351). At first Finno-Ugric populations could not pronounce voiced consonants, but later after mastering the voiced consonants they could produce hypercorrect forms with voiced consonants instead of voiceless ones. In the first article of volume 2 “Impersonals and passives in Baltic and Finnic” (pp. 363- 389) Axel Holvoet points out possible areal links between the passive and impersonal constructions of Latvian and Lithuanian on the one hand and Finnic on the other hand. In many languages the impersonal constructions are etymologically derived from passive constructions, e.g., Pol. Zburzono šcianę “They pulled down a wall’, ‘A wall was pulled down’ as opposed to Sciana zostata zburzona (przez robotnikéw) “The/A wall was pulled down (by the workmen)’ (p. 365). For the purposes of his paper Holvoet finds it useful to generalize some differences between the passive and the impersonal. In his view “The passive: (a) promotes the original object of an active construction to subject; the passive verb form must agree with this subject in number at least; if the passive form is periphrastic and contains a participle, then this participle will agree with the subject in case and gender as well (provided we are dealing with a language where these forms of agreement exist); (b) must not necessarily contain an agent phrase, but may do so if necessary; if no agent phrase occurs, the sentence conveys no information about the kind of agent involved”. “The impersonal: (a) does not promote the original object to subject (the agreement features mentioned for passives will therefore not apply), and (b) does not allow an agent phrase to be added, but always applies to human agency” (p. 366). 136 | RECENZIJOS Examples of the Finnish passive are: Naapuri kutsuttiin illalliselle. NOM.SG PRET. PASS. ALLATIVE “The neighbour was invited for dinner.’ Minut kutsuttiin illalliselle. ACC PRET.PASS ALLATIVE ‘I was invited for dinner’ Naapuri ei kutsuttu illalliselle. PRTV. SG NEG PASS.PART ALLATIVE ‘The neighbour was not invited for dinner’ (p. 367). I must confess that I was confused by the last example, since I would have expected that the partitive would be different from the nominative. Since my knowledge of Finnish is minimal I asked Ilse Lehiste who suggested to me that this might be a typographical error for naapuria. Interestingly enough, Holvoet writes (p. 368) that there does not seem to be a tendency in Finnic to integrate the agentless passive into a passive paradigm, but rather a tendency to integrate passive forms into the active paradigm. I think that this might be a common tendency, one that Gotab (1975: 29) has termed ‘activization’ for Polish (cf., e.g., the examples Holvoet has quoted above). Golab compares also postclassical Latin constructions such as legitur librum (acc.) ‘a book is read’ (cf. Plautus epityra estur insanum ‘olive spreads are eaten madly’), and Pol. czyta si¢ ksigzke “a book is being read’. In this connection I should like to quote a further example from the Iliad (XIII, 597): TO d" žptixeTo NOM. [ACC.?] SG. NEUT CONJ 3 SG. MIDDLE AOR. the and dragged petAtvov žyyos NOM. [ACC.?] SG. NEUT. NOM. [ACC.?] SG. NEUT. ashen spear Schwyzer (1966: 237) writes that the sentence could be translated in several ways, (a) “der eschene Speer schleifte nach’ or (b) ‘wurde nachgeschleift’ or even (c) ‘er schleifte den e. Sp. hinter sich nach’. If one accepts translation (a) the verb is simply intransitive, and if one accepts translation (b) then the verb is passive. But if one accepts the translation (c) then the verb has been “activized' and made transitive. In my view the sequence of possible translations follows exactly the historical development. The middle voice was originally intransitive and then with the appearance of an active counterpart in the preterit the morphological middle voice could be interpreted as passive. Finally a possible ‘activized’ usage of the old passive was introduced. In his discussion of the Latvian agent phrases Holvoet quotes the examples: manis celta maja and mana celta maja “a house built by me’ (p. 371) noting that the second example is now obsolete. He writes further that originally “...mana celta maja meant ‘my house, which I have built’, and subsequently came to mean ‘the house built by me’ (the agent and the possessor not RECENZIJOS | 137 being necessarily identical any more)” (p. 372). The syntax is, of course, similar to that of the Lithuanian example: Namas yra mano pastatytas ‘The house was built by me’. Holvoet writes (pp- 377-378): “The genitive mano used here is the same as in mano tévas ‘my father’, but differs from the genitive manęs used in nuo manęs ‘from me’ and laukia manęs “is waiting for me’ [...] This suggests that in Lithuanian as well, the agentive genitive was originally an adnominal possessive genitive...” However, the use of manęs in agent usage is known also in Lithuanian (Schmalstieg 2002a: 50-51; Zulys 1969: 170). It must be kept in mind that syntactic constructions can be lost as well as gained in the history of individual languages, cf., e.g., the English agentive of, discussed below. The theory of the possessive origin for the genitive of agent is simple and attractive and has been ably defended by Holvoet elsewhere, e.g., (2001 with literature). The theory certainly works well for the Baltic passive which is only encountered in participial constructions. But in a recent article in this journal (2002a: 41-43) I have given examples from Old Persian, Old Indic, Tokharian, Greek and Old Church Slavic in which the genitive is used with agentive meaning in finite verbal constructions as well as in participial constructions. Therefore, if one were to assume the possessive origin of the agentive genitive in participial constructions in these languages, one would then have to construct a separate theory for the agentive genitive in the finite constructions (perhaps presupposing a transfer of the meaning derived in participial constructions to finite constructions). This seems less probable to me than Haudry’s (1977: 409) notion that the subjective genitive and the genitive of belonging go back to the same origin, viz. the genitive of source, which when used with the passive participles denotes the author of the action. Traugott (1972: 127-128) notes that in English the preposition of became the most popular agentive until ca. 1600, cf. Chaucer: ...that is a lord to be biloved of his citezeins and of his peple. Although the use is recessive, one can still find in Shakespeare: ‘I have been told so of many’. She writes: “...this use of of is an extension of its use expressing source or origin...” (p. 127); “In O(ld) E(nglish) of was used mainly to signal ‘out of, originating from’, or part-to-whole relationship. Probably strongly supported by the influence of French de, of spread from the partitive use to possessive constructions and to many others that had involved the genitive marker” (p. 128). Although of is no longer used to express agent in a passive construction (at least in any variety of contemporary English that I am familiar with), Traugott’s explanation provides a good example of how the agentive and possessive meanings can be derived from a common source without assuming that the former meaning arose from the latter. I would also add that the assumption of an ergative origin for the agentive genitive would explain the identity of the nominative and genitive singular of Hitt. an-tu-uh-ša-aš ‘man’, Goth. harjis ‘army’, hairdeis’ ‘shepherd’, Vedic veh ‘bird’. In my view the sigmatic nominative was originally characteristic of only the etymological *-0 or *-jo stem nouns, the ending *-os having its origin in an agentive form of an older consonant stem noun, see Schmalstieg (2000: 386; 2001). Long ago Hirt (1928: 102) recognized that there was a morphological relationship between the Greek 3rd sg. middle ending -fo (and Old Indic -ta) and the indefinite case of the Indo- European participles in *-to. Note the following parallel: et TiS éTLudto Oo CONJ. NOM.SG.MASC. 3 SG. PSV. IMPERF. PREP. if anyone was honored by 138 | RECENZIJOS Tob dnuov GEN. SG. GEN. SG. the people (Goodwin and Gulick 1958: 261). In Greek the agentive genitive has been reinforced by the preposition ūno ‘by’. The actants would be in the same respective cases (i.e., the subject in the nominative case and the agent in the genitive case) in a possible Lithuanian syntactic counterpart: Jéigu kas būvo tautos CONJ. NOM. SG. MASC. ~~ 3 PRET. GEN. SG. if anyone was by people pagerb-ta, NEUT. PAST PSV. PART. honored Tradition says that the Gk. ending -fo is the 3rd sg. middle aorist / imperfect ending, whereas the phonetically corresponding Lithuanian ending -fa is the neuter sg. past passive participle ending, but the resemblance is striking. Could the Indo-European middle verbal ending *-to derive from an earlier participle? If so, could the purported possessive origin of the agentive be saved by tracing it back to Indo-European times? The incorporation of etymological participles into the verbal paradigm has parallels elsewhere, cf., e.g., the Slavic preterit in -/-, which is of participial origin. Horn (1898: 148) writes that the two Pahlevi preterit constructions am kart ‘von mir (ward) gethan’ and man kart ‘mein Gethanes’ merged and gave the modern Farsi preterit sg. 1 man kardam ‘1 did’, etc. Could the Iranian development merely be a repetition of what happened in Indo-European times? Interestingly enough modern Farsi man ‘I’ derives from the Old Persian (gen.) mana ‘mine, of me’, and may now be used even as the subject of intransitive verbs, cf., e.g., man doktor am “I am a doctor’ (Schmalstieg 1995: 19-20). Although in general language specific explanations are more likely than those going into the distant Indo-European past, the occurrence of common syntactic phenomena in other Indo-European languages certainly suggests a common origin. I doubt that one would want to explain, for example, the subject use of the nominative case merely within the framework of the history of Baltic alone, without reference to Greek, Latin, Indo-Iranian, etc. Similarly the nominative object of the infinitive has its origins in Indo-European times, although the scope of its use may have been influenced by West Finnic (see below). Holvoet (pp. 379-386) discusses the problem of indefinite, generalized zero subjects of verbs. The use of the plural verb in such circumstances is quite common typologically and occurs both in Latvian and Finnish, e.g., Latv. Manu dzivokli kratijusi, mani mekléjusi ‘I was told that my flat had been ransacked and that I had been looked for’ (with plural participles) and Finnish Sielld kuuluvat (3rd pl. pres.) tienaavan (act. part.) hyvin ‘It is said that one earns well there’ (pp. 381- 382). On the other hand the use of the 3rd sg. verb with no subject is common in Finnic, cf. Finnish Puheesta-ni (elative plus possessive) voi (3 sg. pres.) kuula (inf.), ett olen (1 sg. pres.) ulkomaalainen (nom. sg.) ‘From my speech you can hear that I am a foreigner’ (p. 382). In Baltic languages, of course, the difference between 3rd person singular and plural verbal agreement can only be determined in those forms with participles. For Latvian Holvoet gives us the example: RECENZIJOS | 139 Ja pavasari dzird dzeguzi kūkojam un ja nav édis un naudas nav klat, tad tai gada slikti klajas “Tt in the spring one hears a cuckoo crying and one hasn't eaten and one has no money on one's person, then that year will be a bad one' (p. 383). The generalized form of the participle ėdis ‘eaten’ here is in the masculine singular (showing a neutralization of gender). Stressing the similarities between Latvian and Finnic Holvoet writes: ‘On the whole, however, it is Latvian that seems to have adapted its system of impersonal and passive constructions to a Finnic model’ (p. 387). A minor editorial slip in the preceding article is that the Lithuanian sentences (40) and (41) on p. 308 are erroneously labeled Latvian. In his article “On the development of the nominative object in East Baltic” (pp. 391-412) Ambrazas writes that there are two opinions concerning the origin of the nominative with the infinitive. According to the traditional opinion “...the nominative is regarded as the former subject and the infinitive is treated as a reflection of the purposive dative of the actional nominal” (p. 391). According to Timberlake’s (1974: 220) view the nominative used with the infinitive “...arose as a syntactic borrowing from some West Finnic language(s)” (p. 392). Ambrazas first notes the dative origin of the Baltic infinitive and then notes that in some constructions the infinitive can be replaced by some other nominal, e.g., Rugiai (mums) liko séti / séjai “The rye remained (for us) to sow / for sowing’ (p. 393). A similar Latvian example is: Pienins ėst / ésanai nederéja “The milk wasn’t fit for consumption’ (p. 394). The purposive meaning is perhaps less clear in the example: Jam teko / patiko | rūpėjo laukas arti “It fell to him / He was pleased / concerned to plough the field’ (p. 394). The nominative with the infinitive is observed with verbs of perception, cf. the expressions Lith. Kas girdėti “What is to be heard?’; Latv. Kas Jauns dzirdet “What news is to be heard?’ (p. 395). The verb ‘to be” may function with the nominative with the infinitive to express necessity, e.g., (man) (yra) buvo | bus namai statyti “(for me) it is / was / will be necessary to build a house’. In Latvian this is apparently rare as is evidenced by a single example from Endzelins (1951: 783): Kungam ėst tei meizite “The gentlemen have to eat this bread’ (p. 397). It is possible that the Latvian debitive has its origin in such constructions (Holvoet 1993: 152): *man ir maize ja ést literally ‘to me is bread which (is) to eat’, i.e., “I have bread for eating’. The normative grammars, of course, demand a nominative case for the subject of such a construction, thus, e.g., Lini bijusi Jjakalté “The flax was in need of drying’. In a letter dated 18 February 1984 the now sadly deceased Latvian linguist, Valdis Zeps wrote to me: “The business about no accusative after a debitive is stuff and nonsense, promulgated by the Academy Grammar, although to my knowledge, nobody has told the Emperor about the new clothes’. Zeps even thought to be possible the sentence: Man ir jalasa gramatu (acc.) ‘I have to read the book’, see Schmalstieg (2002b: 153). In that we have to deal with the replacement of a historical nominative by an innovating accusative now felt as an object, the phenomenon is somewhat similar to the phenomenon of ‘activization’ discussed above. The example is also mentioned in the article by Baiba Metuzale-Kangere and Kersti Boiko, see below. There are also examples consisting of an adjective and a link verb, e.g., Lith. (Man) medus (yra) | buvo / bus gardus valgyti ‘Honey is / was / will be delicious (for me) to eat’ (p. 398). With neuter adjectives one encounters in some local dialects of East Lithuania: šaltinio vanduo sveikas / sveika gerti ‘Spring water is healthy to drink’. Ambrazas writes (p. 399): ‘Kernel structures medus gardu ["honey is sweet’]; vanduo sveika ‘water is healthy’] represent the pattern of an 140 | RECENZIJOS ancient nominal sentence with the pure stem-form in the predicate”. While I agree with this statement completely, I am suspicious of the example from Old Prussian: Erains boūsei poklusman ‘Jederman sey unterthan, everyone must be obedient’. Possibly Abel Will thought that every adjective (and perhaps pronoun and noun) after a verb should have a final nasal consonant. This is somewhat similar to Petit’s idea (2001: 185) that Abel Will established a rule for himself that the accusative should have a final nasal even when the form in question was of the neuter gender. Thus I would reconstruct a neuter *poklusma. Of all the northern Indo-European languages (Baltic, Slavic and Germanic) only Old Prussian gives unequivocal evidence of an *-o stem neuter singular ending -an (< *-on). Ambrazas notes that “...the original structure of the nominative with the infinitive in East Baltic and the subject function of the nominative in the constructions [to be] of a relic character” (p- 399). This is proved by the participial constructions which agree with the noun in case and number, e.g., (p. 394): Jam (buvo) likes / ritpéjes / tekes laukas arti “The field (evidently) remained / concerned / fell to him to plough’. I agree with Ambrazas that such constructions are of a reliquary character, as may be shown by cognate constructions in other Indo-European languages, but I am not certain that the internal evidence of Baltic participial constructions is necessarily proof of this. It seems to me that an opponent of the theory might argue that the nominative participles are merely the result of assimilation to a prevailing pattern according to which, e.g., an adjective would have to agree in case, number and gender with the noun it modifies, e.g., Jam teko naujas laukas arti ‘It fell to him to plough the new field’. Ambrazas writes that the reinterpretation of infinitive constructions was encouraged by the generalized meaning of neuter adjectives (p. 402). Examples are: Pienas (NOM. SG.) saldu (NEUT.) / saldus (MASC. NOM. SG.) [gerti | “The milk is sweet [to drink]’ and Saldu [gerti] pienas (NOM. SG.) / pieną (ACC. SG.) ‘It is sweet to drink milk’. It is important, according to Ambrazas, to note that the word order is changed here, i.e., the nominative is shifted from first place to the position next to the infinitive. In Latvian where neuter adjectives have been replaced by adverbs of state in similar constructions there may be vacillation between the nominative and accusative: Aka/Aku nav viegli rakt “The well is not easy to dig’. Interestingly enough Ambrazas does not find the traditional explanation of the nominative with the infinitive and Timberlake's explanation to be mutually exclusive. They merely apply to different epochs. The original construction with the nominative was reanalyzed and became an impersonal construction. According to Ambrazas: ‘The nominative turned into the grammatical object of the infinitive and spread in its new function under the influence of the West Finnic nominative object rule’ (p. 408). My own view of syntactic change is quite similar. The seeds of the change exist already in the language but the influence of the foreign language may be a kind of catalytic agent aiding and abetting the spread of the construction in question. Bernhard Wilchli concentrates on those languages most affected by the Baltic and Finnic contact in his article “Lexical evidence for the parallel development of the Latvian and Livonian verb particles” (pp. 413-441). Both verb particles and preverbs can express accomplished activity / telicity. In Estonian only verb particles are used, whereas in Livonian, both verb particles and preverbs are used, although the former predominate. In Latvian both are used also, but preverbs predominate and in Lithuanian one encounters even more use of preverbs as opposed to verb particles (p. 420). Thus going from Livonian to Latvian to Lithuanian one encounters decreasing use of verb particles and increasing use of preverbs. Typically in Estonian the notion ‘open’ is RECENZIJOS | 141 expressed by the verbal particle lahti and in Livonian by valdin, the instructive plural of the noun valda ‘power, free will; permission, authorization; arbitrariness; administrative district’, a word related to Est. vald ‘power’ and Finnish valta, a borrowing from Germanic, cf. Old Norse vald. In Latvian the notion ‘open’ may be expressed by at-vérts, but the word vala, the locative singular of vala ‘freedom, leisure, free time; arbitrariness; power; permission; will’ is more common (pp. 420-421). In Lithuanian the usual expression is atidarytas, atvertas, atdaras, atviras and only in a band of dialects along the northern border with Latvia does one encounter the cognate (with the Latvian) expression valio(j) with this meaning (fn. 12, p. 437). Wilchli concludes (p. 434): “The lexical material of the verb particles shows that it would be somewhat too easy to conclude that the Latvian verb particles are loans from Finnic. The situation is more complex and more thrilling (sic!) than that’. It seems to me that American linguists can only applaud a researcher who finds anything linguistic to be ‘thrilling’. For those readers unfamiliar with contemporary American English usage I quote the definition of ‘thrilling’ from Webster's (p. 2383): ‘1a obs : PIERCING b: penetrating with cold: inducing shivering and shaking... 2: causing an instantaneous surge of emotion: producing tremulous excitement: deeply moving... 3: THROBBING, VIBRATING’. Perhaps the second meaning is most common in every-day American English. Research in a vein similar to that of the previous article is Helle Metslang’s article “On the developments of the Estonian aspect: The verbal particle dra” (pp. 443-479). Most interesting is the collection of sentences (pp. 445-446) which illustrate the different usages of dra ‘away, off" from directional to pure perfective: Ta saatis kiilalise (GEN.) dra ‘He / she saw the guest off’ (total object, perfective); Ta saatis kiilalis-t (partitive) dra ‘He / she was seeing the guest off’ (partial object, imperfective); (as an answer to a question): Ei, eile siinnitas Mari ta lopuks onnelikult dra ‘No, luckily, yesterday at last Mari brought the child into the world’ (pure perfective meaning). Metslang establishes five categories of transitive verbs with respect to their relationship to the particle dra (pp. 446-453). In category (1) the simple imperfective verb is used with a partial object, e.g., Ma tundsin presidenti “I knew the president’ whereas with a total object the particle is used, e.g., Ma tundsin presidendi dra ‘1 recognized the president’. In category (2) “...the particle is optional, it emphasizes and doubles the perfective meaning of a sentence expressed already by the total object” (p. 449), e.g., Ta koristas tuba (with partial object) ‘He / She was cleaning the room’ vs. Ta koristas toa dra ‘He / She tidied up the room’. In category (3) “...either a directional adverbial or perfective particle is obligatory when a perfective interpretation is intended. dra has both aspectual and directional meaning and the same characteristics as in the previous groups” (p. 450). Here dra is used according to the situation, e. g., Ta veeretas vaadi oue/dra ‘He rolled the barrel into the yard / off, away’. Such examples are opposed to perfective transition verbs which must be used with a total object, e.g., Ta pani raamatu lauale / dra “He / she put the book on the table / away’. Category (4) includes verbs that usually are incompatible with the particle dra, but which may take it occasionally, in which case the particle has a pure perfective meaning. The perfective particle in a sentence such as Ta siinnitas lapse dra emphasizes perfectivity and expresses “... a special information structure, where all the semantic content (e.g., the expected birth) of the sentence belongs to the theme and its realization (the perfective meaning alone) to the rheme” (p. 451). Category (5) includes “Verbs that are totally incompatible with [the] perfective particle and are used only with the partial object..., e.g. vajama ‘to need’...” (p. 452). 142 | RECENZIJOS The development of dra within the history of the Estonian language provides a mirror of the various foreign influences to which Estonian was subjected. At the very beginning of the creation of the literary language dra was used as a particle denoting perfectivity (p. 475) under the influence of German syntactic structures. During the period of independence from 1918 until 1940 an attempt was made to free the language from German influence and there was a corresponding decrease in the use of the verbal particles. Later during the Russian rule the use of aspect markers with verbs became common. In the 1990s period of democratization the use of dra to mark perfectivity in informal newspaper texts has become more common thereby reflecting colloquial style. The purpose of the article “Case systems and syntax in Latvian and Estonian” (pp. 481-497) by Baiba Metuzale-Kangere and Kersti Boiko is to provide a contrastive analysis of the functions of the various cases in these two languages. The authors first study subject-object relations which in Latvian are quite straight-forward and traditional with a nominative subject, accusative direct object and a dative indirect object. In Estonian on the other hand the situation is much more complex, where “...subjects and objects can be in the three so-called basic cases” (p. 485). Sentences involving subject-object relationships are dubbed type 1 sentences. Next the subject-object relationships involving an oblique case are studied (type 2 sentences). Traditionally a distinction is made between the ‘logical subject’ and the ‘grammatical subject’ and there is a class of sentences in Latvian that: “...have the logical subject in the dative and the logical object in the nominative” (p. 488). This applies to such debitive sentences as: Man (DAT.) ir jalasa gramata (NOM.) ‘I have to read the book’. The authors write, however, that “...it is not at all so very clear that that the nominative is an object at all nor is the dative an unequivocal subject” (p.488). One also encounters the colloquial: Man ir jalasa gramatu (Acc.). According to the authors: “This actually indicates a shift in perception, i.e. that the ‘logical object in the nominative’ is becoming a regular object in the accusative as it has always been for personal pronouns. How far this grammaticalisation has progressed and what its limitations are is not being researched since the sentence is regarded as ungrammatical, not as a sign of an ongoing process of grammaticalisation™ (p. 490). But, of course, this is exactly how language change takes place, when older norms are replaced by innovating norms. What is the rationale for having one set of rules for the nouns and another set of rules for the pronouns? The Estonian construction: Mulle (ALL.) meeldib siitia “I enjoy eating’ is compared with Latv. Man (DAT.) patik ėst ‘id.’ revealing a close similarity between the Estonian allative and the Latvian dative (p. 491). Existential and equational sentences are called type 3 sentences. An Estonian example is: Peenral kasvab lilli “There are flowers growing in the flower bed’ which might be compared with the synonymous Latv. Dobé aug pukes. The authors write: “The productivity of Latvian sentences of Type 2 may be explained as a generalisation of the model of the possessive sentence in the absence of the verb ‘to have’ (p. 493). But several of their examples of Latvian type 2 sentences have syntactic counterparts in Lithuanian which does have the verb ‘to have’, viz. turėti. Cf. Latv. Man (DAT.) sap galva (NOM.) “I have a headache’ (p. 490) = Lith. Man skauda galva; Man (DAT.) patik ėst ‘1 like to eat’ = Lith. Man patinka valgyti; Man (DAT.) nomira tevs (kad man bija pieci gadi) ‘My father died (when I was five years old)’ = Lith. Man numirė tėvas (kai man buvo penkeri metai). (According to my colleague, Virginija Vasiliauskienė, this latter sentence would now have an archaic or dialectal RECENZIJOS | 143 flavor. She also writes that the seguence Man skauda galva would more often nowadays be replaced with Man skauda galvą (Acc.). (I would note here another case of the reinterpretation of the subject as an object.) One could, of course, object that the meaning ‘to have’ for Lith. turėti is recent as opposed to Latv. turėt “to hold’ (although the Old Prussian cognate turit also means ‘to have’). But, of course, verbs with the meaning ‘to have’ appear to be innovations in all the Indo-European languages in which they are encountered and the argument must be moved from a specifically Baltic level to an Indo-European level and a much wider research plan must be established for studying the occurrence of the dative of possession vs. the verb ‘to have’. The existence of a verb ‘to have’ and the dative of possession are not necessarily mutually exclusive. Baldi (2002: 23) notes that the dativus possessivus is one of the oldest uses of the dative in Proto-Indo-European and in Latin, of course, the dativus possessivus is well known, but according to Baldi (2002: 25): “...there is no time in the attested history of Latin that the meaning ‘have’ is not present in Lat. habeo, being found in some of the earliest authors such as Livius Andronicus..., Ennius and Naevius... It is also well represented in the inscriptional corpus...” Baldi notes further (2002: 28): “Mihi est is information-oriented; the construction foregrounds the possessum, and backgrounds the possessor... Habeo on the other hand, is subject-prominent and person-differentiated...” Baldi and Cuzzolin (to appear) have identified at least nine different roots which have given the meaning ‘to have’ in various Indo-European languages. Simon Christen’s article “Genitive positions in Baltic and Finnic languages” (pp. 499-520) examines the structural relevance of different genitive positions in these languages. The author concentrates on Finnish, Estonian, Latvian and Lithuanian with occasional references to other minor Finnic languages, Livonian, Veps, etc. Typically the possessive genitive precedes the item possessed in Finnish, Estonian, Latvian and Lithuanian, cf. Estonian tiidruk-u koer ‘the girl's dog’ (p. 499) Similarly the genitive of material and the agent genitive precede the head noun, cf., e.g., Lithuanian aukso žiedas ‘ring of gold’, Ciurlion-io paveikslas. The author writes: “Action nominals do not only combine with agents, but also with patients” (p. 502). In this case, too, such constructions are more readily formed in Finnic than in the Baltic languages. This is certainly true with respect to Lithuanian where nominalizations frequently seem awkward to native speakers. In Soviet time when it became necessary to translate the then current definition of capitalism, viz., aKcnayamauus 4eno6eKa veaosexom ‘exploitation of man by man’ into Lithuanian, at first the expression žmogaus (objective genitive) išnaudojimas žmogumi (instrumental agent) was used, but Jonas Kruopas (1963: 24; 1998: 410) had the courage to note that the instrumental of agent was not good Lithuanian usage, so he suggested rather žmogaus (GEN.) išnaudojimas žmogaus (GEN.) (in order to keep to three words as in Russian, although žmogaus išnaudojimas would have apparently been sufficient [personal letter dated 2001/11/09 from V. Ambrazas]). My colleague, Virginija Vasiliauskienė writes (personal communication) that in September of 2003 she questioned a group of fourteen students about the meaning of the sentence Zmogaus išnaudojimas žmogaus and ten thought that it meant that it was the exploitation of one man by another. They were, however, unanimous in their opinion that the first word of the phrase denoted the exploiter and the last word the one exploited, viz. the subject (agent) comes first and the object (patient) comes last. She writes also that such a phrase as Jono kankinimas Petro ‘John’s torture of Peter’ ("Peter's torture of John?’) is, indeed, ambiguous as to who is the torturer and 144 | RECENZIJOS who is being tortured, but by analogy with the sentence Jonas kankina Petrą the first word of the phrase would probably be understood as the subject (agent) and the last word as the object (patient). In addition to the interesting work of Virginija Vasiliauskienė on the history of Lithuanian word order, it seems to me that more might be done with surveys and guestionnaires on the status of contemporary Lithuanian word order. With regard to the genitive placement Christen puts Lithuanian in a special category noting that here the word order is freer than in the other languages examined in this article (p. 517). I think that the author is correct and that the norm in Lithuanian is relatively recent and could hardly have been affected by Finnic influence. V. Vasiliauskienė has studied in detail the history of preposed vs. postposed modifiers in a number of articles, e.g., (1997; 1998; 2001; 2003). She writes (2003: 102): “...statistical data prove that at the end of the 19th century the preposed genitive attribute began to oust the postposed genitive attribute which had dominated since the appearance of the first written Lithuanian documents“. This topic is discussed again below in the context of the final article of the volume. In her article “A piece of the cake’ and ‘a cup of tea’: Partitive and pseudo-partitive nominal constructions in the Circum-Baltic languages” (pp. 523-568) Maria Koptjevskaja-Tamm distinguishes between the partitive construction represented by a piece of the cake as opposed to the pseudo-partitive represented by a cup of tea where we are really talking about the amount of some substance. Thus, e.g., the Finnish partitive construction pala téstd hyvista kakusta (elative) ‘a bit of this good cake’ is distinguished from the pseudo-partitive sikki perunoita (partitive plural) ‘a sack of potatoes’. In English a true partitive construction would be a cup of that good tea, whereas a pseudo-partitive construction would be a cup of tea. The definition is: “partitive nominal constructions involve a presupposed set of items referred to by one of the nominals (‘that good tea’...); and the quantifier indicates a subset which is selected from it; in a pseudo- partitive nominal construction the same word merely quantifies over the kind of entity (‘tea’...) indicated by the other nominal” (p. 527). The author writes further that the starting point for the grammaticalisation of the partitive construction is when: “...predicates of separation (such as ‘take away’, ‘cut’, ‘remove’ etc.) combine with two dependents, one referring to a part and the other referring to the whole from which it comes” (p. 535). Later comes reanalysis of the original construction such that “... the ‘part’ and the ‘whole’ are... reanalyzed as making up one constituent instead of being two different dependents of the same predicate” (p. 535). Pseudo-partitive constructions arise when the difference between ‘set’ and ‘kind’ becomes ambiguous. She writes that the history of partitive and pseudo-partitive constructions in Finnish shows several processes often found in grammaticalization: at some stage “...the original separative meaning can gradually get lost [...] a new marker with the separative meaning can start expanding to partitive-like uses” (p. 536). I think that her analysis of the grammaticalisation process is completely credible and that this could be applied to the internal history of Indo-European where the ablative case has arisen as a new marker of separation, whereas the genitive case was the original marker of separation. Therefore the traditional partitive genitive terminology for Indo-European is completely justified. See Schmalstieg (1995: 36-37) and compare the quotations from Haudry (1977: 409) and Traugott (1972: 127-128) mentioned earlier in this review. The author writes, however, (p. 559): “...there is no historical evidence whatsoever for a separative/directional origin of the genitive in a number of Indo-European languages where it has partitive and pseudo-partitive uses”. RECENZIJOS | 145 Evidence, of course, is just what one can get an audience to accept as evidence. But in a paragraph entitled Le gėnitif partitif comme gėnitif d'origine Haudry (1977: 62) writes: “Le gėnitif d'appartenance a un ensemble est lui aussi un gėnitif d'origine: on le voit clairement dans un emploi adverbal comme Y.9.13 yat [CONJ.| hė [3RD GEN./DAT. SG. PRON.] tūm [2ND SG. PRON.] us.zayanha [2ND SG. MIDDLE PRET.) ... nmanahe [GEN. SG.) pourusaspahe [GEN. SG.] ‘quand tu lui naquis ... dans la maison de P".” (In the previous example I have supplied the bracketed morphological description. The *-o0 stem genitive seems to imply that at least potentially a contrastive ablative would have been possible also, see Sims-Williams [1998: 141].) The ancient syntactic structure encountered with the Indo-European verb ‘to fill’ (root *pl-) seems to imply a genitive denoting either ‘source’ or ‘part’. Note Old Indic sémasya (gen. sg.) jatharam prnati "he fills his stomach with soma’ (Meillet 1964: 345); Gothic ...gup lubainais Sulljai izwis allaizos fahedais (GEN. SG.)... *...may the God of hope fill you with all joy... (Streitberg 1919: 249 [Romans 15:13]) = Russ. (http://www.russianbible.net/) 502 ace nadexcow da ucnoa- Hum 6ac écskou padocmu (GEN. SG.) = Gk. 6 dė ede thc ėinidos manpomoat Huds namą yopds ... (GEN. SG.) = Lith. ... vilties Dievas jus... tepripildo visokių džiaugsmų... (GEN. PL.) (Rubšys 1998: 1781), but with the innovating instrumental notice the Protestant Bible Society translation (NT 1972: 216) ...Dievas ... tepripildo jus visokiu džiaugsmu... (INSTR. SG.). Similarly in the Vulgate we encounter Deus...spei repleat vos omni gaudio (ABL. SG.), where the innovating ablative has replaced the old genitive, still attested, e.g., in such sentences as: amphoram puram impleto aquae purae (GEN. SG.) ‘fill a clean jar with clean water’; convivium vicinorum (GEN. PL.) cotidie compleo ‘1 fill up my company with neighbors everyday’ (Woodcock 1959: 55; Hofmann and Szantyr 1972: 82). It has long been known that originally probably only the Indo-European *-o stem nouns showed a morphological difference between the ablative and the genitive. With other stems the genitive form in -es, -os and -s also had ablative meaning (Brugmann/Delbriick 1911: 163-164). The analysis of Hittite and Tokharian did not change the picture very much. Hittite has two ablative endings for the *-o stem, viz., -az (possibly derived from the adverbial suffix *-tos and then spread to other stems) and a second one homonymous with the gen. sg. -aš, this latter certainly the original ablative undifferentiated from the genitive (Kronasser 1955: 101-103; Sturtevant 1951: 88-89). In Later Avestan also the - of the etymological *-o stem ablative spread to other stems (Reichelt 1909: 168; Sims-Williams 1998: 142). The Tokharian A and B ablatives (different from each other) are apparently derived within that language family itself, see van Windekens (1979: 254-256). One notices that the Indo-European *-o stem nom. sg. ending *-os is fairly easy to etymologize, cf., e.g., Lat. lup-us ‘wolf’, Gk. Adxoz, Old Indic vrk-ah, Lith. vilk-as, Slavic vivk-», Hitt. an-tu- uh-ša-aš ‘man’, etc. In my view the original identity of the nom. and gen. sg. (=ergative sg.) *-os was for the most part not tolerated in those Indo-European languages other than Hittite and led to the creation of new genitive forms, Lat. lup-i, Gk. *Adx-ov < *hdx-oto, Old Indic vrk-asya, Lith. vilk-o, Slavic vlrk-a. The Baltic and Slavic genitives are said to be cognate with the ablative represented in Latin lup-o(d) and Old Indic vrk-at. There is, of course, a problem with the equation of Lith. vilk-o and Lat. lup-o(d), because a Lith. 6 under ordinary circumstances derives from Indo-European *4 and only by the assumption that unstressed Indo-European *6 passes to Baltic *a could one make it to correspond to the Lat. final -o (Mažiulis 1970: 23). Mažiulis (1970: 106) affirms also that the Balts and the Slavs, like the Greeks, never had a paradigmatic 146 | RECENZIJOS ablative. Shields (2001) gives a brief review of the history of the problem and suggests an origin for this ending. It is the common opinion that in Greek the ablative has been lost and that in the *-o stems the genitive has taken over the functions of the Indo-European ablative. Schwyzer (1966: 90) suggests that the form roixw ‘from the house, at one’s own expense’ represents a relic of the Indo-European ablative. I suggest just the contrary, viz., that the new *-o stem ablative (*-o[d/t] or *-a[d/t]?) never took hold in Greek as it did in Latin, Old Indic, Baltic and Slavic. Under the heading Gen. der Abstammung Schwyzer (1966: 124) quotes the Iliad (21, 109): mateo [GEN. SG.] 8 iw ayadolo [GEN. SG.) literally “I am of a good father’. The elementary Greek grammar by Goodwin and Gulick (1958: 236) notes that poets used the genitive of separation with verbs of motion, e.g., Odd umno:o [GEN. SG.] xatT1fADoucv ‘we descended from Olympus’ (Iliad 20, 125). I would suggest then that the Indo-European genitive did at least originally have separative meaning and that the author’s developmental scenario (apparently contrary to her assertion) is valid for Indo- European as well as for Finno-Ugric. I have suggested (Schmalstieg 2000; 2001) that the Indo-European syntactic nominative singular had three morphological origins, a zero ending for the etymological absolutive case (originally the subject of intransitive or antipassive constructions), an animate marker *-s (originally the agent of ergative constructions) and an inanimate marker *-m (originally the instrument in constructions similar to the ergative constructions). Hittite shows the retention of the *-os as -a$ in the nominative, genitive and even partially in the ablative function (although, as mentioned, a new ablative in -az is encroaching on the older -aš). The reinterpretation of the etymological ergative ending *-os as a nominative ending led to a complete restructuring of Indo-European *-0 stem morphology such that in some of the Indo-European languages the old genitive ending *-os was modified or substituted with another ending and in the case of Latin and Indo-Iranian a new marker of source (the ablative) was created for the older functions. This same marker of source (*-a[t/d]?, *-o[t/d]?) served the Baltic and Slavic *-o stem nouns with genitive as well as ablative meaning. The theory is developed in detail in Schmalstieg (2000). The ending *-s which originally had genitive-ablative-ergative meaning was retained as such in all but the *-o stems. In the *-o stems *-s came to have the additional nominative singular meaning and was also spread to other stems. That this spread of -s is an ongoing process can be shown by the creation within historical times of such new nominatives as Latv. akmens ‘stone’ (beside Lith. akmuo) and modern Gk. matepag ‘father’ beside classical mathe. This notion explains why the morpheme *-s is encountered in the genitive, ablative and nominative singular function in the ancient Indo-European languages. In the restructuring of the *-o stems a new genitive singular was created in all the Indo-European languages but Hittite and a new ablative was created in some of the Indo-European languages. In his article “Nonverbal predication in the Circum-Baltic languages” (pp. 569-590) Leon Stassen notes that adjectival and nominal predicates may be either in the nominative case or some oblique case, the difference depending upon the Relative Time Stability. He writes (p. 569): “... encoding in the nominative is used to refer to situations which are relative ‘time- stable’ and thus unlikely to change over time, whereas the oblique encoding emphasizes the ‘fleeting’ or temporary nature of the situation”. The author confirms the existence of this semantic contrast in a number of different languages from the well-known Spanish distinction between ser and estar to examples in such disparate languages as Tamil, Kannada, Telugu, Avar, Chechen and RECENZIJOS | 147 Archi. He concludes that “...the double encoding of nonverbal predicates, which takes the form of a nominative-oblique case opposition, is an areal feature of the Circum-Baltic area...” and that “... this double encoding is in essence a non-Indo-European characteristic, which has been ‘pushed aside’ by Indo-European expansion, but which may, in some border areas, continue to exert its influence on Indo-European languages as well” (p. 588). Thomas Stolz’ article “On Circum-Baltic instrumentals and comitatives: To and fro coherence” (pp- 391-612) begins with the statement that “Recent investigations ... have revealed that syncretism of comitative and instrumental categories is not as common among the languages of the world as formerly postulated”. According to the study of 323 languages completed by Stolz’ research group slightly less than 25% of their sample languages were like German and are characterized by comitative-instrumental syncretism (p. 593). Those languages which display comitative-instrumental syncretism are labeled ‘coherent’ languages and those which distinguish comitatives from instrumentals are called ‘incoherent’ languages. Those languages which mix coherent and incoherent morphemes are called ‘mixed’ languages. (p. 593). Thus, for example, Swedish is coherent, Lithuanian is mixed and Finnish is incoherent. The author writes: “None of the Slavonic languages of the Circum-Baltic area are strictly speaking coherent, though other members of the same family have become coherent, some of them, e.g. Slovene and Sorbian, most probably because of German influence....” (p. 601). One could possibly add to this list Serbo-Croatian where also, contrary to school book prescriptions, s is sometimes used to denote ‘with, by means of’. Some speakers say: Idém s vlūkom (vozom) ‘I'm going by train’ instead of the prescribed form without the preposition (Magner 1972: 255). Although the author is correct with regard to Russian, in Srz4 (IV 9) we encounter definition no. 5 of ¢: “Ynotpebnsiercs npu 0603Ha4eHHH MpejiMėTA, JULIA, MTOCPEICTBOM HJIH C MOMOIIbIO KOTOpOro ocyuectsasercs aeiicrsue” ‘It is used to denote an object [or] a person by means of which or with the help of which an action is performed’. Examples include: omnpasums naxem co ces3nbiM ‘to send a package with (by?) a messenger’; pacecmampueams ¢ aynoi ‘to examine with a magnifying glass’; Mbimb ¢ mbaom ‘to wash with soap’; —-Zfla kak nocetiams nucemo-mo? — C noymou nado, — omeexan Hava Heanoeuus (““ And how should one send the letter?” “One must [send it] by [with the?] mail”, answered II'ja Ivanovic’); 5 yesxcaro ... c noezdom 6 namb namrao- uams ‘1 am leaving by train at five-fifteen’. The line between the instrumental of instrumentality (opyodutinocmes) and the instrumental of means (cpedcmeo) is hazy in Russian and I have predicted (Schmalstieg 1966: 179) that one day Russian will become a ‘mixed’ or ‘coherent’ language. I based my speculation on Kurytowicz’s (1960:131-150) notion that bipartite morphemes tend to replace monopartitite morphemes. The author concludes that despite the lack of historical evidence it is probably the case that “...the progress of coherence in the Circum-Baltic region was enhanced principally by the partial Germanicization in northern Scandinavia (Sami) and in the former state of the German Knights (Latvian and Estonian)” (p. 609). The volume concludes with a comprehensive article “The Circum-Baltic languages: An areal- typological approach” (pp. 615-750) by Maria Koptjevskaja-Tamm and Bernhard Walchli. According to the authors: “The goal of this concluding paper is to show that the Circum-Baltic languages as a whole form an interesting linguistic landscape of their own among the languages of the world, in general, and the European languages, in particular” (p. 615). This article begins with a rather interesting study of the historical background of the area. I 148 | RECENZIJOS was surprised to read, however: “Old Prussian and Lithuanian catechisms began to be printed in the duchy of Prussia at the end of the 15th century...” (p. 620). I assume the authors had in mind the first Old Prussian and Lithuanian translations of the Lutheran catechisms (First Old Prussian Catechism 1545; Mazvydas 1547), so even if this is a simple misprint for “16th century” it is still wrong. Martin Luther was born in 1483, died in 1546 and the usual date given for the publication of his Smaller Catechism is 1529. I know of no catechism by the teenage or younger Martin Luther dating from the end of the 15th century. Next the authors proceed to the examination of earlier studies of language contacts in the Circum-Baltic area. First the authors take up the vexed question of Sprachbiinde and write: “An obsession with Sprachbiinde is, however, not a necessary precondition for finding the linguistic situation the CB area a thrilling (sic/) object of inquiry” (p.624). As mentioned above it is always good to find emotional attachment to one’s area of research. The authors then distinguish those areal studies which merely (p. 626) catalogue similarities between languages and those which try to explain the similarities. They also find that the frequency of occurrence of an observed feature is also an important aspect of their study. It is also very difficult to determine whether the loss of an inherited structure is just a natural development of the language or influenced by a neighboring language. In the rest of their article they“... hope to combine the achievements of areal linguistics in the CB area and general linguistic typology” (p. 629). First they discuss Pluralia tantum which were originally more common in the Indo-European languages than in Finno-Ugric. Sometimes the pluralia tantum may be borrowed along with their etyma, e.g., Finnish rartaat ‘cart’ < Baltic (Lith. ratai, Latv. rati) which is literally the plural of the word for ‘wheel’, Finnish ratas ‘wheel’ < Baltic (Lith. ratas, Latv. rats). Plurality spread from this word “...to other words denoting carts in Finnish such as vaunut and kdrryt” (p. 633). In the authors’ view: “...the general tendency for or against pluralia tantum is inherited, but it may change in the development of languages” (p. 633). The authors conclude their discussion of suprasegmental phonology with the statement that there are at least three different areas in the languages of the Circum-Baltic region displaying the following features: (1) the opposition of tone contours in long syllable cores [found only in the Baltic languages]; (2) ‘overlength’, i.e., the existence of three degrees of length exhibited by Estonian, etc.; (3) word-tone the origin of which is some way connected with the original number of syllables in a word and the presence or absence of an original secondary stress encountered in many Scandinavian dialects (p. 645). The initial stress of many of the Circum-Baltic languages and the polytonicity may be related (p. 646). A very interesting portion of this article is devoted to the morphological cases, particularly the use of the genitive and partitive in the Baltic and Finnic languages, although the authors don’t always favor mutual influence. For example, they question the common opinion that the developed system of secondary locative cases in the Baltic languages is of Finnic origin, e.g., because: “The secondary local cases of Eastern Baltic represent grammaticalizations of nouns with their postpositions. Similar developments are found also in Umbrian, Tokharian and Ossete in Indo- European languages” (p. 672). One might argue for Finno-Ugric influence in Tokharian (van Windekens 1979: 166), but hardly in Umbrian or Ossete. With respect to the expression of possession the authors write that the locative type“... is the option that is invariably chosen for those Indo-European languages that lack ‘have’-verbs (e.g. RECENZIJOS | 149 Celtic, Indic, Anatolian)” (p. 676). But the ‘locative’ type of expression does not necessarily preclude the existence of ‘have’ verbs, as must be obvious to the authors from their discussion of Russian umems ‘to have’. Baldi has clearly demonstrated this with regard to Latin (see above) and I have mentioned it with respect to Lithuanian. Although in Hittite (an Anatolian language) the usual means of expression of possession is with the dative, Hittite does use a verb ‘have’ (interestingly enough also as an auxiliary verb with perfect tense meaning). Thus, e.g., from the Hittite laws (Hoffner 1997: 152): tdk-ku (conj.) LU-a$ (nom. sg.) MUNUS-an (acc. sg.) har-zi (3rd sg. pres.)... ‘If a man has a wife...” In languages with both means of expression of possession the function of the verb ‘to have’ is different from that of the dative of possession. The authors rightly refine the notion of possession in the ‘be’ languages (p. 677). After reviewing the comitative and instrumental, the comparative, the passives, desubjectives and zero-subject constructions the authors proceed to sentence syntax. Here they discuss adjective agreement, loss of gender in Low Latvian dialects. and the syntax of numeral constructions. On pp. 702-703 the authors present a table illustrating the internal syntax of numeral constructions in Finnic, Baltic and Slavic showing the possibility of either government strategy ‘case on the nominal governed by numerals’ or agreement strategy ‘case on the nominal determined by the function of the whole phrase’. Among the footnotes on p. 704 we find the statement: “In Old Church Slavonic, lower numerals decline, higher numerals are indeclinable”. I don’t understand why the authors write that higher numerals are indeclinable. Diels (1932: 218) gives the example sv desetijp (INSTR. SG.) tyspsty (GEN. PL.) ‘with ten thousand’ in which the word for ‘thousand’ in the genitive plural is governed by the word for ‘ten’, which itself is in the instrumental case as required by the preposition sz. Diels gives also the example s» džvėma desetvma tyspstama ‘with two thousand’ in which all the numerals are in the instrumental dual as governed by the preposition sz. The only larger Old Church Slavic numeral might be considered tema which for the most part translates pvg:az and denotes an indefinite extremely large number. The word appears to be fully declinable according to the entries in SJAST (pp. 545-546). One is left to wonder how the church fathers might have reacted to gigabytes. The authors find that (p. 704) “...a large portion of the European flexible SVO languages is found in the CB area. These include Slavic, Baltic, Finnic, as well as Northern and Eastern (at least Inari) Sami”. The authors then offer a table with the title: “From SOV to SVO in Europe” (p. 705). This runs from SOV (Hittite, Nenets) to highly flexible SOV (Latin, Old Greek) to split SOV/SVO (Hungarian) to highly flexible SVO (Russian, Lithuanian, Komi, Veps, etc.) to flexible SVO (Finnish, Estonian) to V2 (Swedish, German) to SVO (English). (Hittite, of course, although Indo-European, is an Anatolian language at one time presumably spoken in Anatolia, at least that is where the majority of the texts have been found.) Still the table is interesting in classifying the word order types encountered mostly in Europe. The authors write further: “...Baltic and Finnic appear to be the only VO-language families in Europe in which all nominal modifiers, except for relative clauses and prepositional phrases precede the head“ (p. 707). The authors appear to dispute Vasiliauskiené’s claim that in Lithuanian there was originally some freedom of word order in the noun phrase. They write: “Vasiliauskiené’s statistics are problematic for areal comparative purposes since we do not know which percentage is covered by genitives in (pseudo-) partitive functions, such as ‘a pound of apples’ (these always follow their heads)” (p. 708). 150 | RECENZIJOS Vasiliauskienė replies, however, (personal communication) that in the first part of Sirvydas' Punktai sakymu 16 partitives were found. Of these 13 (81,3 %) were NG, and three (18,7 96) were GN and in general in the Punktay sakymu 81 per cent of the genitive nouns are used postpositionally (in all, 1291 seguences with the genitive were found and investigated). In Eastern Lithuanian dialect texts collected by A. Baranowski 62 examples with the partitive were investigated of which 44 were NG (71 %) and 18 were GN (29 96). Vasiliauskienė writes further that she has just completed the collection of material from the first three issues of Aušra (the first Lithuanian magazine, 1883), where from several hundred sequences with the genitive she encountered only 13 with the partitive meaning and nine of these were NG and four GN. In the majority of the 16-19th century texts we encounter the predominance of postposition. The contemporary GN rule is much more strict than the relatively free word order of the early texts. It should also be emphasized that in sequences with the partitive there was much more freedom of placement than there is in contemporary Lithuanian where the NG order predominates. In summary the articles by the individual authors are well-written and interesting (although perhaps not always ‘thrilling’, at least from the American interpretation of that adjective). Among the fairly frequent misprints are New Mexvico (p. ii) > New Mexico, Winfried Lehmann (p. I[ndex]-4) > Winfred Lehmann, Allan Timberlake (p. I[ndex]-7) > Alan Timberlake, etc. Although the editors and authors might have done a better job of proof-reading and paid more attention to detail, these volumes present a wealth of new ideas and frequently a reorganization of well-known facts. The editors are to be congratulated on compiling path-breaking tomes for the study of both the languages of the Baltic area and typological linguistics. ABBREVIATIONS LKZ — Lietuvių kalbos žodynas 1-20, ed. by J. BALCIKONIS, J. KRUOPAS, K. ULVYDAS, V. VITKAUSKAS, 1941-2002. NT 1972: Bible Society translation of the New Testament: Naujasis Testamentas, No place of publication given. SRIA — Slovar’ russkogo jazyka IV, ed. by S. G. BARXUDAROV et al., Moscow: Gosudarstvennoe izdatel’stvo inostrannyx i nacional’nyx slovarej, 1961. SIAST — Slovnik jazyka staroslovénského, ed. by ZOE HAUPTOVA et al., fasc. 44, Prague: Academia, Nakladatelstvi Ceskoslovenské Akademie véd, 1992. WEBSTER'’S — Webster's Third New International Dictionary, ed. by PHILIP BABCOCK GOVE et al., Springfield, Massachusetts: G. & C. Merriam Company, 1966. LITERATURE BaALpi, P. 2002: Where does Latin habeo come from? In: Cavoto, E, ed., Festschrift for Alexis Manaster-Ramer, Munich: Lincom Europa, 23-32. BALDiI, P, CuzzoLiN, P, to appear: Etymological, geographical and typological dimensions of “have” verbs in Indo-European languages. In: Salvi, G., ed., Festschrift for Joseph Herman, Budapest. BRUGMANN, K., DELBRUCK, B. 1911: Grundriss der vergleichenden Grammatik der indogermanischen Sprachen. Vol. 11/2. 2nd ed., Strassburg: Karl J. Triibner. COLLINDER, B. 1964: Sprachverwandischaft und Wahrscheinlichkeit, Uppsala. (Acta Universitatis Upsaliensis, Studia Uralica et Altaica Upsaliensia.) DIELS, P. 1932: Altkirchenslavische Grammatik, Vol. I, Heidelberg: Carl Winter. GIRDENIS, A., PABREZA, J. 1978: Nauji Slekiavimo stebėjimai Žagarės apylinkėse. Baltistica 7, 127-129. RECENZIJOS | 151 GoktaAB, Z. 1975: Endocentricity and endocentrization of verbal predicates: Illustrated with Latin and Slavic material. General Linguistics 15, 1—35. Goobpwin, W. W., GuLick, C. B. 1958: Greek grammar, Waltham, Mass., Toronto, London: Blaisdell and Co. HAUDRY, J. 1977: L'emploi des cas en védique, Lyon: Editions CHermes. Hirt, H. 1928: Indogermanische Grammatik, 4. Doppelung, Zusammensetzung, Verbum, Heidelberg: Carl Winter. HOFFNER, H. A. 1997: The laws of the Hittites: A critical edition, Leiden, New York, Cologne: Brill. HOFMANN, J. B., SZANTYR, A. 1972: Lateinische Syntax und Stilistik, Munich: C. H. Beck. HOLVOET, A. 1993: On the nominative object in Latvian, with particular reference to the debitive. Linguistica Baltica 2, 151-161. HoLVOET, A. 2001: Once more the Baltic genitive of agent. Baltistica 35, 45-58. HORN, P. 1898: Neupersische Schriftsprache. In: Geiger, W., Kuhn, E., ed. Grundriss der iranischen Philologie, Vol. 1, part 2, Strassburg: Karl Triibner, 1-198. KAROLY, S. 1972: The grammatical system of Hungarian. In Benkd, L., Samu, 1., eds., The Hungarian Language, The Hague and Paris: Mouton and Co, 85-170. KRONASSER, H. 1955: Vergleichende Laut- und Formenlehre des Hethitischen, Heidelberg: Carl Winter. KRruoras, J. 1963: Dabartinės lietuvių kalbos vystymosi tyrinėjimas. Lietuvių kalbotyros klausimai 6, 21-30; reprinted as pp. 407-416 in Rinktiniai raštai, compiled by S. Keinys, K. Morkūnas and A. Sabaliauskas, Vilnius: Mokslo ir enciklopedijų leidybos institutas, 1998. KuUrYLOwICz, J. 1960: Esguisses linguistiques. Wroctaw, Krakow: Ossolineum. MAGNER, TH. E. 1972: Introduction to the Croatian and Serbian Language, State College, Pa.: Singidunum. MARTINET, ANDRE. 1951: Concerning some Slavic and Aryan reflexes of IE s. Word 7, 91-95. Mažiuris, V. 1970: Baltų ir kitų indoeuropiečių kalbų santykiai (deklinacija), Vilnius: Mintis. MEILLET, A. 1964: Introduction a l'étude comparative des langues indo-européennes, University, Alabama. (Reprint of the 1937 ed.) PETIT, D. 2001: Pronoms neutres et rėgle SVO en vieux prussien. Baltistica 36,173-188. REICHELT, H. 1909: Awestisches Elementarbuch, Heidelberg: Carl Winter. RUBSYS, A., transl. 1998: Senasis ir Naujasis Testamentas. Šventasis Raštas, Vilnius: Katalikų pasaulis. SCHMALSTIEG, W. R. 1966: The preposition s + the instrumental. Slavic and East European Journal 10, 178-180. SCHMALSTIEG, W. R. 1995: A Student Guide to the Genitive of Agent in the Indo-European Languages, Washington, D.C.: Institute for the Study of Man. (Journal of Indo-European Studies, Monograph Number Fourteen.) SCHMALSTIEG, W. R. 2000: Lithuanian and Indo-European Parallels, Journal of Indo-European Studies 28: 385-398. SCHMALSTIEG, W. R. 2001: La triplice origine del nominativo singolare indoeuropeo. Res Balticae 7, 23-32. SCHMALSTIEG, W. R. 2002a: The Indo-European genitive of agent with finite verbs and participles. Acta Linguistica Lithuanica 47, 41-54. SCHMALSTIEG, W. R. 2002b: Valdim Jurim Zepam 70. Baltu Filologija 11 (1), 151-154. SCHWYZER, E. 1966: Griechische Grammatik. Vol. 2. Syntax und syntaktische Stilistik. Vervollstandigt und herausgegeben von Albert Debrunner, Munich: C. H. Beck. SHIELDS, K. 2001: On the origin of the Baltic and Slavic o-stem genitive singular suffix *-ad. Baltistica 36, 165-171. Sims-WiLLiams, N. 1998: The Iranian languages. In: Giacalone Ramat, A., Ramat, P, eds. The Indo-European Languages, London and New York: Routledge, 125-153. STREITBERG, W. 1919: Die gotische Bibel, Heidelberg: Carl Winter. STURTEVANT, E. H. 1951: A comparative grammar of the Hittite language, New Haven and London: Yale University Press. 152 | RECENZIJOS TIMBERLAKE, A. 1974: The nominative object in Slavic, Baltic and West Finnic, Munich: Otto Sagner. (Slavistische Beitrage, 82.) TRAUGOTT, E. C. 1972: A history of English syntax. A transformational approach to the history of English sentence structure. New York, etc.: Holt, Rinehart and Winston, Inc. VASILIAUSKIENĖ, V. 1997: Daiktavardžio ir įvardžio genityvo vieta senojoje lietuvių kalboje. Lietuvių kalbotyros klausimai 38, 104-121. VASILIAUSKIENE, V. 1998: Atributiniy junginių modeliai Mikalojaus Daukšos Postiléje ir Antano Baranausko tarmių tekstuose. Baltistica 33 (2), 215-232. VASILIAUSKIENE, V. 2001: Atributin¢ frazė su derinamuoju komponentu Kiprijono Lukausko Pamoksluose. Acta Linguistica Lithuanica 45, 151-178. VASILIAUSKIENE, V. 2003: Postposition in the Attribute Phrase of Old Lithuanian. Zeitschrift fiir Literaturwissenschaft und Linguistik 33 (129), 96-108. VONDRAK, W. 1928: Vergleichende Slavische Grammatik, Gottingen: Vandenhoeck & Ruprecht. VAN WINDEKENS, A. J. 1979: Le tokharien confronté avec les autres langues indo-européennes, Vol. II, 1. La morphologie nominale, Louvain, Centre International de Dialectologie Générale. Woopcock, E. C. 1959: A new Latin syntax, Cambridge, Mass.: Harvard University Press. Žurys, V. 1969: Vadinamyjy nekaitomuyjy įvardžių vieta lietuvių kalbos gramatinéje sistemoje. Baltistica 5 (2), 167-177. William R. Schmalstieg Gauta 2003 10 28 814 Cornwall Road State College, Pa. 16803 email: [email protected]