Grammatische Kernbereiche, deren Rekonstruktion und deren Relevanz für historisch belegte Sprachstufen des Baltischen und Slavischen

ACTA LINGUISTICA LITHUANICA LI (2004), 81-112 Grammatical core areas, their reconstruction and their relevance for historically documented language levels of the Baltic and Slavic languages BÖRN WIEMER Universität Konstanz The paper offers a critical account of prominent theoretical approaches to prehistorical developments of basic morphosyntax in Baltic and Slavic: voice, alignment and aspectuality. For comparative reasons the relation between possessive BEand HAVE-constructions is discussed as well. The methodological practice in reconstructing and explaining the diachronic evolution of ancient Baltic and Slavic is assessed critically, first of all because it has abided by ‘in vitro’-approaches without an account of language contact and typological evidence. It is argued that an adequate theory of grammatical change would further have to try to be more accurate with respect to the relative chronology of ‘layers’ of coding techniques and their possible intersections during long periods of diachronic development. This includes the possibility of the cooccurrence of grammatical coding. As for this issue, speakers of contact languages could trigger the functional repartition of such cooccurrent devices to such a degree that one of these devices would become predominant, fitting their own language- ’s form:function distribution. 0. INTRODUCTION This paper can be understood in a certain sense as a critical tribute to two scholars who have made a great contribution to the historical-- comparative Baltic and Slavic studies: Jurij Sergeevi¢ Stepanov and William R. Schmalstieg. For each of them, a series of more or less influential publications will be taken as an occasion to show how much diachronically oriented Baltistics suffers from the fact that chronological sequences of the change of grammatically relevant structures and analogy formations are considered too little concretely in their mutual interweaving in subsequent periods and that too little functional points of view are taken into account. In addition, typologically plausible developmental processes and the factor of language contact are often neglected, as well as the grammatical vs. lexical status of morphological distinctions. The areas that are discussed here are central to language construction, and accordingly they are important for diachronic morphology and syntax as well as for morphological and syntactic reconstructions. It is about diathesis, actionality and alignment. Discussion of the theses 82 | BJÖRN WIEMER and the approach of Stepanov (section 1) and Schmalstieg (section 2), reflections on an analogous case from the history of the Slavic language follow (section 3). Together, these three roughly designed “case studies” of historical-comparative argumentation on a typological background should lead to methodological postulates, which are formulated at the end (section 4). 1. STEPANOV: CONNECTION BETWEEN Stepanov (1976; 1977; 1978a, b; 1989) has made an impressive attempt to draw up a synthetic outline of the development of actionality and diathesis in the field of verb morphology from the historically documented facts and the reconstructable prehistoric states of the Baltic and Slavic languages. Stepanov’s theses touch on the pre-or early phases of grammaticalization, but they are based on the “in vitro” method of historical-comparative linguistics, i.e. without reference to contact linguistic factors. I will try to give a critical summary of his position on this subject. Questions that go deeper into the detail of the idea. Sound and form theory go, I will not discuss; nor will I problematize statements on the correctness of Stepanov's reconstruction in these questions. Rather, I am interested in a critical examination of his methodology and the interpretation of the facts thus provided. Stepanov starts from facts common in Indo-European studies of the stem change in the verb system of the older ide. Languages, which not least concerned the prehistoric Baltic and Slavic, as well as that it in the early-ide. Syntax no clear means to characterize transitivity (as a sentence- -semantic quantity). Furthermore, there was no special morphology which would extend the verb, respectively. Neither a reduction of the argument structure would have indicated, nor one that would have enabled a more systematic distinction between different actional types (in the sense of aspect-sensitive classes). Rather, actional forms were leximmanent fixed!. Starting from these premises, which are also generally accepted in Indo-European studies, Stepanov endeavours to uncover “layers” of changes in the verb stem that are etymologically common to Slavic and Baltic languages and that occurred in the course of the dialectal differentiation of ide. In the prehistoric Baltic and Slavic languages, the morphological correlations between the dialects led to more systematic correlations that marked transitivity, argument reduction and expansion, and actional types. The differentiation of Slavic and Baltic against each other is based on the different way in which these correlations were functionally developed. According to Stepanov, the affixing methods, which were common to the Baltic and Slavic languages in terms of type and (at least initially) also etymologically (i.e. in terms of form-substance), first formed functional clusters of actional, i.e. aspektrelevanten and diathesei In research these have occasionally been referred to as lexical aspect”. RECONSTRUCTION OF GRAMMATIC CORE AREAS | 83 related features. By ‘diathesis’ Stepanov understands the ‘relation of the action to its subject’ (1978a: 338). This corresponds to the ‘basic voice’ of Klaiman (1991), which also includes the medium of historical-- comparative linguistics. Actional features, on the other hand, are for Stepanov inherent features that determine whether a reference to a boundary (Russian: ‘npenen’) exists or not. It is therefore mainly about a distinction between terminative vs. aterminative verb stems. In Stepanov- ’s model, trunks (and their possible further affixation) can be related to three situational phases, each of which focuses on different parts of a complex situation arranged around the attainment of such a boundary. Phase (i) consists of a process that strives to reach such a limit ("pollecc, CTPEMSILUMÜCH K CBOEMY rıpeneny- "), phase (ii) corresponds to that limit itself ("HacTyrusenne rıpenena- "), phase (111) finally represents a state which necessarily follows from the attainment of that limit. This is itself aterminative, since it is open towards its end („mpouecc, OTPaHMYEHHbIH B Hayalle, UCXOASILUMI M3 paHee NOCTUTHYTOTO TIpenena Mm He OTPaHMYEHHBbIÄ KOHUOM“); See. Stepanov (1977: 145). What is important is that we are dealing here only with a conceptual distinction of actional stages, which are grouped around a state change. Although Stepanov repeatedly speaks of “aspect oppositions” or even “aspect pairs” (“BUAOBbIE IIPOTUBONOCTABJIEHHUS /TTapbI”), this is initially only a foundation of aspect-sensitive classes of verbs, but not a grammatical aspect opposition (see below). Aspects would be distributed complementarily to these situational phases” -which in the language state reconstructed by Stepanov (and also in later periods) can by no means be assumed. The root-changing affixes (infixes, suffixes or reduplications) originating from the Slavic and Baltic protoforms led in many cases to a shift in the root meaning from situational phase (ii) to situational phase (i) or (iii). Standard examples are reduplications to form the static perfect and alterations of positional verbs; cf. about from the idea. Root *séd-> aks. sed5.1.5G. AOR to ses-ti ‘to sit’ (= situational phase (ii)) against sed-e-ti ‘to sit’ (= situational phase (iii)), analogous to lit. sed-u.1.sG. PRs to ses-ti ‘to sit’ versus séd-é-ti ‘to sit’ (1989: 221£.). Stepanov now searches in the system of the ide. Verbs according to what he calls the "diachronic prototype" of the morphological marking of such a three-phase structure and sees two "favorites" for this role, namely the ide. Distinction (I) between present and aoristum perfect roots and (II) the comparison of activa tantum, media tantum and perfecta tantum (each with its own root structure); See. Stepanov (1977: 146ff. ; 1978a: 341). Functional clusters of actional and diathesis-related characteristics can be reconstructed, among others, for the proto-Slavic language, in which ablatives are still productive Notabene: the three phases of Stepanov correspond exactly to the three components from which the most complex class is constructed in an aspect theory, which treats aspects as operators on the lexeminherent actional content (as the operanda). In Breus’ ILA theory, this is the INCO class (see Breu 1998); to it belong above all so-called. holistic verbs such as russ. nanoanume /Kanonkamo ‘fill up’, 3acaonums /3acaonsme ‘cover, cover up’, cnpamamoca / npamamoca ‘hide’. On these verbs in particular cf. Paduceva/Rozina (1993). 84 | BJÖRN WIEMER must have been, so that among others Ablaut formation with stem extension of a root like *nes-(> russ. nes-ti, lit. nes-ti ‘to carry’) to *nos-i-ti a causative as well as an iterative formation caused’. This type of trunk resp. Root change later led in Slavic to the formation of iterative or “indeterminative” (and thus later formed one of the formal foundations for the ipf aspect), while in (east) Baltic it became above all a productive type of the formation of causative and curative (cf. Stepanov 1976: 420). Another example: the opposition between short and long root vowels (+ stem extension by suffix) led to the productive contrast of verbs that were interpreted punctually or in some other way perfectively (short vowel) vs. verbs that denoted a frequentative, iterative, intensified or in some other way progressively interpretable type of fact (long vowel). See for example aks. umreti & umirati ‘to die’ (elongation of the root vowel + stem extension), pogrébati = pogreti ‘to bury’; begraben’ (ė < *¢ wegena < *äin of the stem extension) as well as combined with prefixation vszbanoti ‘awake up’ & bodeti ‘be awake’ (ė < *E). Analogue “pair” is still used in modern Lithuanian; cf. for example klupti ‘stumbling’ = klüpoti ‘stumbling- ’, kišti ‘(in)stalling’ = kysoti ‘(somewhere) stuck, stuck’ (in the latter case coupled with the opposition [+ transitive- ]). In contrast to Slavic, this morphological opposition was also used in (East)Baltic languages to mark intransitives and transitives differently; cf. lit. kisti ‘to change’ = keisti ‘to change’, plisti ‘to spread’ = plėsti (as well as plėtoti) ‘to spread’ (where ė < *ė), busti ‘to wake up’ vs. = bausti originally as much as ‘to tear out of sleep’ —‘to force something’ or also ‘to abuse, beat’, today with the meaning of ‘to punish’ (cf. e.g. Stepanov 1978a: 36, 342). Furthermore, nasal infixes and the sz-suffix have become the most productive means of present tense formation primarily of intransitive verbs in Baltic languages (Stan 1943: 131ff.), while nasal infixes have been reduced in Slavic languages*. These intransitives are regularly in opposition to transitive verbs with different root formation and accordingly different temporal root formation (partly from later “layers”, e.g. with the suffix -(d)y-< *7). Cf. a comparison of typical representatives of this opposition: (la) tamp-a.3. PrRs vs. tap-o.3. PRT (1b) tap-o.3. Press vs. tap-€.3. PRT to tap-ti ‘to become’ to tap-y-ti ‘to paint’. (originally as much as ‘to stick to, to stick to’). 3 Cf. analogously in Ancient Greek. gėpo ‘I carry’ and its iterative derivative gopéw compared to poßew ‘I frighten somebody.’, which has the same suffix and the same o-ablation as causative to the medial peßoua: ‘I fear myself’ (Stepanov 1978a: 359). Cf. worrisome reflexes in the form of nasal vowels in Polish or their continuation in East Slavic as in ide. *sed-> urslav. *se-n-d-> ax. sedbzw. full. sigdzie ‘he/she sits down’ or russ. damaged ‘dass.’ (vs. preteritum poln. siadf, russ. sel ‘he sat down’). In contrast to nasal infixes for the formation of the present trunk, nasal suffixes in the form of -ng-(< *na?) acquired an extremely important role in the development of the aspect system and the actional functions in Slavic as a trunk extension that was not limited to a tempus (cf. Sadnik 1962). 4 RECONSTRUCTION OF GRAMMATICAL CORE AREAS | 85 (2a) link-st-a.3.- prs vs. link-0.3. PRT (2b) left-ia.3.prs vs. left-e.3. PRT to link-ti ‘lean; Neigung zeigen’. zu links-ti ‘to bend, to bend over’. (3a) proc-est- -a.3. PRS vs. vyk-0.3. PRT (3b) vyk-d-0.3. PRS vs. vyk-d-é.3. PRT to vyk-ti ‘to happen’. to vyk-dy-ti ‘to execute’. From these and a number of other facts, according to Stepanov, the following conclusion can be drawn: in the Baltic languages, the diathesis function predominated in the initial categorical clusters, while in the Slavic languages the actional function prevailed. So Stepanov's theory contains in general form the following statements? ! (a) The affixation types which have firmly entered into the system of Slavic and Baltic verb morphology (and are there today either fossilized or still visible to some extent or even productive) represent an identical morphological technique. (b) Most of these affixes even have a common etymological origin. (c) Both these affixes as well as the types of affixation must come from the time for which one can find intensive contacts between the preforms of the historically documented Slavic and Baltic languages (which in turn contributed significantly to the hypothesis of a “Balto-Slavic language community”- ). Stepanov himself (d) In the course of differentiaspricht hier von „proto-balto-slavisch“. tion to Slavic and Baltic, the clusters of actional and diathesis-related characteristics were gradually dissolved, i.e. as Stepanov would say, either the actional or the diathesis-related component was “suppressed” in them, whereas the other came to the fore in the further development from a functional point of view in the overall system of verb morphology and was constantly expanded or remained preserved for longer. In Baltic, the diathetic changes are better preserved (i.e. more archaic), in Slavic, however, the actional changes are so strong that they laid the foundation for the later aspect system. Cf. his own words, to which I will return later: “Ilpeo61anarılmMH TIPOTHBOIIOCTABJICEHHAMH B TJIATONBHOH JIEKCHKE COBPEMEHHOTO JIMTOBCKOTO A3bIKA SIBJISIIOTCS TTAPHbIe OIIIO3MLIHH TIPOCTOHU TJIATON ACHCTBUA WIM COCTOAHUA /Kay3aTUBHbIa WIM (DaKTHTHBHbIH TIMIarojJI Toro xe KopHsi. Hampumep: bėgti ‘Gexatp' /béginti ‘3acrTaBııaTb GexaTb (LEAHTbB MOJIOKO)’, ėsti ‘eCTh (0 XHBOTHEIX)' /ėdinti “KODMHMTb (XHBOTHbIX- )', siūti “1lmTB’ /sitdinti ‘OTNaBaTb IMT, faire coudre’, kesti ‘meHsatbca’ /kesti 'MeHsTb'. DTOMY THUITY TIPOTHBOTIOCTABJIEHMÜ OTBEYAET IIpOJIYKTHBHAS TpaMMaTHYEecKasl OIMO3HUNSI «<MSTKOC» IIEpeXOJIHOC CIIPSDKEHHE /«TBEPIOE» HeIIepexX0JIHOE CIIPSDKEHHE; T.e. TUII ia-TIPe3eHc E-TIPETePHT IIPOTUB THIIa G-TIPE3EHC O-TIPETEPUT. AHaJIOTHYHO 3TOMY TIpeOOJIaJIaIONIHMH ONMO3MIMHAMM B COBPEMEHHOM PYCCKOM A3bIKE (Cp. TAKXKE JIpyTHe CIIaBIHCKUE) ABIISIIOTCH BUJIOBbIE ITapbi TJIATOJIOB, BKJIIOYas1 IIapbi «CIIOCO60B Aeiictsust, Aktionsarten». (...) BriojiHe cornacyloTcst OCHOBHbIE OTHOLIEHHUSI B KATETOPHH TIEPeXONHOCTU /HEIIEPEXONAHOCTH H «OIPEAEIIEHHOTO» /«HEOMPENENEHHOTO» BHUNA. [lepBUYHbIE TIPOTHBOTOCTABAEHHBIE TTapbI NPE/ICTABJIAIOT OHO M TO Xe NCHCTBME C JIBYX Pa3JIMYHBIX TOYeK 3peHus (...). B stoi JIBOMCTBEHHOCTH TOYEK 3PeHNA Ha OJĮHO H TO Xe NEÜCTBHE (XOTS CAMH TOYKH 3peHHS pa3IMYHbI) COCTOHT OÖLLHOCTB CHABAIHCKON MH ÖANITMÄCKOH CHCTEM H OCHOBAHHE EIMHCTBA MX IIPOTOTHIIA «BH/Į — AnaTe3a». (...) HAIIPallImBaeTca TIPeANONOKEHHE, YTO 6ANTHÜCKHE TIPOTMBOTIOCTABIEHHA MMaTe3 H CJIABSHCKHE TIPOTUBOIIOCTABJIEHHSI BUJIOB BO3HUKAIOT U3 ONHOTO H TOTO X€ HCTOUHHKA —U3 HEKOTOPbIX TIOCTOSIHHBbIX KODpDpeJISIHŪ MEXIy BUIOM H MMATE3OH, CYIIECTBOBABIUNMX B IIPOTO-6AJITOCJIABSHCKOM HM COCTABJISBINHX TaM 3JIEMEHTbI MOCTATOYHO YCTOMYMBOM CHCTEMbI «BMI —IMaTe3an.“ (Stepanov 1978a: 335-337; blocked by the author himself, BW). 86 | 2012 As for the formation of clusters of actional and diathesis-related characteristics, Stepanov unfortunately does not specify whether the respective types of stem modification affect the actional and diathesis-related behaviour simultaneously and en gros, or whether from verb type to verb type (or from one lexical group to another) they sometimes only cause an actional modification, and other times only a change in the diathesis structure. His series of examples demonstrate both one and the other. And it is the differences between actional and diathesis-related characteristics that lead to cluster formations. Thus, for example, the st-suffix mentioned above does not occur widersprechen sich zum Teil auch, indem nicht immer klar wird, welcher Art die Affiexclusively in intransitive present tense stems, but Stepanov (1976: 409) also sees in it a carrier of an ingressive function („HaunuHaTeJIbHOcCTS“), which should actually bring a perfecting effect with it. He makes this assertion on the basis of modern Lithuanian’. However, he does not ask the question whether (a) an ingressive component could not simply be a result of interaction with the lexical actional default (see for example ding-st-a ‘disappears’) and (b) whether the current situation can easily be projected onto prehistoric (or at least earlier) Baltic. Another problem arises: how is the alleged perfecting effect to be reconciled with the view, elsewhere held by Stepanov, that presence strains in early phases primarily correlated with progressivity’? A further concern is that -stdamit is thus brought into opposition to stem extension by *e (> aks. &, lit. E), which very often only led to an actional modification, namely “progressivation”, but behaved neutrally with regard to the diathesis properties of the stem (i.e. it is proven for both transitive and intransitive stems). Stang (1943: 118, 132) counts both the *é suffixation and the suffixation by {st} to the relatively late procedures. If this is the case and *ė only led to an actional modification of the stem, it would thereby be in contradiction with the diachronic trend postulated by Stepanov; In his opinion, this goes to the side of the diathesis in the Baltic language (p.0). Intransitive verbs with nasal present tense also enter into actional opposition to *é extensions, in that the verb with *e suffix denotes situational phase (iii), the verb with nasal present tense denotes situational phase (i) or (ii) from the above three-phase model, but leaves the diathesis untouched; This would mean that *e as a marker of the terminativeness would have reached a terminative root (stem). Such “couples” postulated by Stepanov are not even so rare; cf. the following representative comparisons® (more on this in Schmid 1963: 60ff.): (4a) bu-n-d-a.3.prs vs. budo.3. PRT (4b) budi /bud-e-ja.3. Breast vs. bud-E-jo.3. PRT to bus-ti (< *bud-ti) ‘wake up’. to bud-é-ti ‘to be awake’. i It can also be found in Stang (1943: 131). ? See for example Stepanov (1989: 180, 186) himself or Kazlauskas (1968: 289). These fills demonstrate not only a difference in phase, but also a lexical distinction in terms of terminativity: verbs such as those in the (a) sentences usually presuppose a terminative course (= phase (i)) or an event (= phase (ii)), while those in the (b) sentences do not, since they focus on Stepanov’s phase (iii). RECONSTRUCTION OF GRAMMATICAL CORE AREAS | 87 (5a) Svi-n-ta.3. PRS vs. Svit-o.3. PRT (5b) svit-i /švit-ė-ja.- 3. PRS vs. Svit-&-j0.- 3. PRT to švis-ti ‘dusk, light/day become’. zu švit-ė-ti ‘to shine brightly, to shimmer’. (6a) kru-n-t-a.3. PRs vs. krut-0.3. PRT (6b) krut-a.3. Breasts vs. krut-and-- jo.3. PRT to krus-ti ‘to get in motion’. zu krut-e-ti ‘to move, shake, jiggle’. The next example is somewhat different: on the one hand, the transitivation of stems in Baltic is made to the *é-extension ide. stems connected, which in Lithuanian is supposed to have resulted in the past tense with palatal stemmaus (in the 3rd person on -€ as for example in pies-ti “draw, paint’: pies-é ‘he/she drew, painted’); This is opposed by the preteritum without stem declination (in the 3rd person on -0) (such as beg-ti ‘run’, beg-o ‘he/she ran’)’. In Slavic, the same stems were also transited in this way (cf. ex. pas-a-ti ‘write’ = pis-e-tö ‘he/she writes’). In both cases, these are with the so-called. „undetermined aspect“ („HeonpeneneHHbIH BUI“) (Stepanov 1976: 409), and thus potential precursors of an ipf. Aspects (on this question see below). Now, however, in other places transitivation is associated with perfecting, understood as the setting of an immanent limit (i.e. terminativation) (see, among others, Stepanov 1977: 145ff. himself). This is also what one would expect and is typologically consistently confirmed (cf. for example Bybee, Dahl 1989). Another example: the morphological opposition of i and a roots in later Slavic aspect pairs such as russ. ocmaeumo /ocmaenamo ‘(leaving- )’ is said to come from an opposition of causative (i-root) and iterative (or durative, a-root). Stepanov generalizes to the conclusion that there is a clustercorrelation of ‘intransitive root + indeterminate aspect + iterative’ vs. ‘transitive root + determined aspect + causative’ (Stepanov 1977: 149). Here he contradicts himself, however, see already above for the extension ide. of athematic strains. For the i-extension has been in the Baltic (and not only there; cf. Stepanov 1978a: 359f. himself) at the same time the basis for both causative and iterative formations (in the broad sense, i.e. including intensive and frequentative); cf. for example lit. ly-dyti ‘(in a form) pour’ as causative formation to /y-ti ‘pour, rain’ (see also e.g. 3) against graib-y-ti ‘(hastily) to grab’ as iterative to grieb-ti ‘to seize’. In addition, there are *i-rooted iterative formations of intransitive verbs; See e.g. svisti ‘leuchten. NTR’ = to spend ‘fechten’ (altlit.), to break ‘waten’ = to braid ‘1. emerge again and again (e.g. sun behind the clouds), 2. wading kneading- ’. From such comparisons contradictions arise, which are based on the fact that the functional clusters taken by Stepanov as a diachronic starting point dissolve in different ways and the involved features can then come together again in different ways in different periods. Thus it can happen that the causative and the iterative/intensive formed a cluster in an earlier period, to? See the quotation in footnote 10. 5. By the way, this view was already held by Stang (1943: 82f.). Schmid (1966) provides a clear analysis of the distribution of ė/o in Lithuanian past tense. Michelini, of course, expresses concern when he concludes that “morphological units of different origins have merged in the suffix -ė or Baltic preteritum” and that “in most cases (...) -ė can be traced back to *i@” (Michelini 1990: 845). 88 | BJÖRN WIEMER Later on, however, they were separated from each other and thus could come into opposition with each other. The comparison of the above-mentioned i and a stems in Slavic is, however, already a comparatively late product of development. The fact that in Lithuanian the i-stamp was used for both causative and iterative functions can be interpreted in two ways: either both types of verbs arose at an earlier stage when these functions were still “paired” and then represented different “dissolutions” of the cluster; or they arose in different periods when causativity and iterativeness did not (anymore) form clusters. Although Stepanov partly puts the morphological oppositions he discussed in detail in a rough chronological order (see the list in the appendix), the periods in which they were respectively productive apparently intertwined, so that it resulted in an often very opaque interweaving of different layers and the functions associated with them. An open problem is therefore the chronological sequence of the different morphological opposition types and the length of their overlap in real time." A related problem, not addressed by Stepanov, lies in the question of analogous compensations, which obscure a pure picture of the consistent sequence of the layers in the trunk changes. In order to avoid a real battle of materials, I will refrain from an example discussion on this topic. A careful reading of Stepanov- 's works could provide the corresponding cases, and today's Lithuanian shows this particularly in the nasal-infected verbs in the present and in the development of the *i stems (i.e. -(d)in and -(d)y-) a very colourful picture, since these affixes induce both actional and diathesis-- related modifications. Furthermore, some remarks on Stepanov- ’s conceptual apparatus and his approach based on it would be appropriate. Here it should be emphasized that Stepanov makes no statements about the degree of grammaticalization of the respective affixing methods (and this was not actually his intention). However, it can be indirectly inferred that their productivity and the regularity of their formation in the respective functional area (diathesis or actionality) could be the decisive factor for a gradation of the degree of grammaticalization, if one wanted to determine one. More importantly, however, is the fact that Stepanov aspect (sun) and lexical modifications such as the action types (cmoco6bI NeicTBus), which include an actional (and therefore aspect-relevant) cause change of meaning, relatively undifferentiatedly ranks under a global category of actionality (acrıeKTyanbHOCTB- ). Accordingly, he understands his three-phase model as a schematic representation of aspect types, although it actually per se represents only a portion of aspect-sensitive LEXICS (see above). Similarly, Stepanov considers the genus verbi and the passive (combined as ‘saror’) in one breath with changes in the lexical diathesis (i.e. the argument structure) under the one category of ‘3ar1oroBocTp’. This is legitimate, provided that, as "especially in Stepanov (1977) it is not clear how his synchronous classification of the transitivity types and the actional three-phase model relates to the sequence of the morphological oppositions. RECONSTRUCTION OF GRAMMATIC CORE AREAS | 89 Stepanov states that it is primarily about demonstrating the systematic conceptual relationships between grammatical categories such as aspect or passive and the functional (conceptual) areas that motivate them. However, a joint consideration becomes precarious if the status difference of the categorical oppositions involved (more grammatical vs. more lexical) is blurred or even declared irrelevant. Stepanov is basically aware of this fact, at least in the area of diathesis. On the one hand, he points out that an affixation that explicitly marks transitive stems (i.e. derives from intransitive or diathetically labile stems) also leads to the derivative and the derived stem (verb) receiving different lexical explanations (cf. among others Stepanov 1977: 140). He is also aware of the fact that one and the same technique of affixing can have a purely syntactic function (marking of the passive), another time a lexical derivative function (e.g. formation of decausative verbs); such as the already common example of the Russian postfix {sja} (Stepanov 1978a: 339). And finally, he shows at the same place that the opposition of active vs. passive and an opposition of lexical diathesis distinctions can have very different, sometimes even incompatible consequences for the interpretation of the intended fact. Thus, the active sentence (7a) can be contrasted with the denotatively equivalent (i.e. synonymous) passive sentence (7b): (7a) Jlabopanm pacmanausaem gewecmeo ‘The laboratory assistant melts the substance’. (7b) Bewecmeo pacmanausaemca (nabopanmonm) “The substance is melted (by the laboratory assistant)’. Neither of the two statements necessarily implies that the processed substance actually melts. But this is exactly what is indicated (asserted) when the transitive verb pacmanausamo ‘to melt, to bring to melt’ is compared to its intransitive lexical equivalent maamo ‘to melt.ıntr’; See: (8) Bewecmeo maem. “The substance melts’, i.e. with respect to the terminative component implies (7b) not necessary (8). Conversely, (8) does not imply (7b) when it comes to the agent presupposition: only (7b) presupposes an external causator (it is precisely this component that makes 7b a denotative periphrase of 7a), (8) on the other hand excludes it. The same applies in principle to Lithuanian, where this opposition is still expressed very regularly (though no longer productively) by the relation between simple and extended stems; See. Stove heats ‘Der Kamin wird heiß’ = The owner heats the stove ‘Der Hausherr heizt den Kamin anjein’. Paraphrasing Stepanov, one can conclude that in the case of a lexical diathesis-opposition of the kind that occurs within the framework of the so-called "translational" dialectology, the "translational" dialectology is not the same as the "translational" dialectology. ‘inchoative—causative alterna- 96 | The 100 So what is the thesis, which Schmalstieg is one of the advocates of? If there was an active:inactive system, it must have preceded an ergative system”. Furthermore, the distinction between active vs. inactive classes is used as the basis for the emergence of both a genus opposition according to ‘genus commune’ vs. ‘genus neutrum’ and for the dichotomy of aorist/present vs. present. Perfektstimmen?® considered. If the basis for these already extremely old (about 2000 BC) morphological distinctions is to have represented the postulated active: inactive system, then one recognizes from this alone how far back in time the latter must have existed. Furthermore, the active class is said to have formed the basis for the ergative and this in turn for the nominative and, complementarily, the inactive class the basis for the ‘casus indefinitus’ (= absolute) and this in turn for the accusative (see below). According to this hypothesis, the result would be the following (cf. Schmidt 1980: 103; 1983: 9f.); However, see Dixon (1994: 185f., note 2) for a critique of this sequence. Abfolge der Alignment-Systeme: Aktiv:Inaktiv > Erg:Abs (‘indefinitus’) > Nom:Akk We want to concentrate on the discussion of ergative syntax or supposed remains of an ide. Inherited ergative system. I would like to emphasize that ‘ergative form’ 0.4. cannot simply mean a frequent characterization of the agent in an oblique case or a PP”. Such an equation has also sometimes led to misunderstandings in the Slavistic literature, when some authors speak of “ergative syntax (construction, etc.)”. Confusion arises because it is not taken into account that a morphosyntactically marked expression of the highest-ranking argument does not per se constitute an ergative scheme. On the one hand, this argument must be opposed in the sentence to a complementary argument (which would correspond to the absolute), and on the other hand, in contrast to the passive, the oblique marking of the highest-ranking argument must be qualified as the morphosyntactically unmarked pattern. I will come back to that in a moment. For the Baltic languages, especially Lithuanian, ergative syntax in connection with Sets by type (9) Here Zmoni-y lives-t-a | lives-m-a here Mensch.Gen. PL leben.pTC. NEW ‘Here people have (probably) lived / live (probably) people’. (10) Sen-y misk-ai myle-t-a alte.GEn.prL Wald.nom. PL love.PTC. NEUT “The old people loved the forests” has been said. In such sentences it is striking that (a) the pradicative participle with no N See the very critical review of Stepanov (1989) by Krasuchin (1990, especially pp. 73 and 78) and more recently also by Comrie (2001: 24-27). = On these and their connection with the medium, see in particular Stepanov (1989: 24ff.); See section 1. * So it understands e.g. Rosen (1979: 165). RECONSTRUCTION OF GRAMMATICAL CORE AREAS | 97 NP congruent, although a nominative NP can stand (see e.g. 10), and that (b) the respective highest rank (possibly only) argument of the participle deriving verb is expressed in the genitive. There are at least two hypotheses for the origin of this construction, one of which is based on the assumption of a prehistoric ergative syntax. The main advocate of this hypothesis is now Schmalstieg (1982; 1988; 1993; 1994), and I will therefore call it briefly ‘ergative’ or ‘Schmalstieg hypothesis’?. Participles and the 3. Press. Sgt. That's an idea. Mediumaorists (e.g. in Vedic and Greek)”. Schmalstieg erkennt die etymologische Verwandtschaft zwischen dem *fo-Suffix des tApart from this assumption, he assumes that in the early Ide. there was no congruence between sentence members, that there was no transitivity opposition and that accordingly stems were indifferent (or diffuse) with regard to the diathesis. Therefore, sentences like (10) should also be seen in diachronic analogy to other nominal sentences (Schmalstieg 1982: 132f.); vgl. etwa (11) We-are-healthy-to-eat Fleisch.NoM.- sG. F healthy. 100% new. INF ‘Eating meat is healthy’. This would, however, correspond to a neutral alignment, since neither with regard to the feature [+ active] nor with regard to the distinction [+ transitive] a clear pattern of the choice of case or the designation of congruence can be recognized (see Plank 1995: 1185). After all, both existed — even according to Schmalstieg’s reasoning, not at the beginning. Now, with the emergence of the transitivity distinction, which also manifested itself in the form of case markings, in intransitive sentence patterns the original ‘indefinite case’ was first refunctioned to the absolute, then to the nominative, while in transitive predicates a special case was used for the highest-ranking argument. In this way, an ergative pattern had first emerged. The former ergative case then changed to the present genitive- ?” (Schmalstieg 1982: 121). Now, although the development of an ergative pattern from a neutral Alignment typologically well documented due to transitive sentence types (Plank 1995: 1194). To the former ide. However, ergative is more likely to have been reinterpreted to the nominative by expansion to arguments of intransitive predicates; these ® Other supporters of this thesis are Marvan (1973) and Palmaitis (1977). * This assumption is based on Hirt's studies of ide. Syntax; See. Schmalstieg (1988: 36f.). 27 Schmalstieg (1982: 121) assumes here reconstructed sentences like *pater bhr-tö vir-os = ‘father carried by the man’ with pater as the original indefinite case, bhrtö as a verbal noun and viros as an ergative case (later genitive). Unfortunately, he does not indicate where this reconstruction comes from and how it is secured by historical-comparative data. His claim that the genitive in sentences like lit. The room. NOM. SG pribego vandens.- GEn.sG ‘The room ran full of water’, Zemé.Nom.sG primirko lietaus.GEN.sG ‘The earth sucked itself full of rain’ go back to an ergative (Schmalstieg 1984). Today, the genitive is used here to indicate a certain amount of a substance in two-digit, but intransitive verbs (see the examples just cited). Such an “ergative addition” to intransitive propositions would logically only be conceivable for the time when there was no reasonably clear transitivity opposition (see also Stepanov’s hypotheses in section 1). 98 | The New York Times Development finds parallels in Caucasian languages (cf. Schmidt 1980: 103f.). Thus, the question arises whether sentences of the type (9-10), which are the object of Schmalstieg's explanation, are really ergative alignments. More on that immediately. The Schmalstieg hypothesis, however, assumes even more. Thus, in the originally assumed juxtaposition of the sentence members, there must have been a model in which the nominative (or, for Schmalstieg, the absolute) would have stood next to a non-congruent ¢- participle (or verbal noun) of intransitive stems; this would correspond to Lithuanian of a structure like (12) *tev-as gule-t-a”® Father. NOM lie.PTC. NEUT = “Father laid down” = ‘The father lay down’. Furthermore, this hypothesis sets because of the for the early Ide. The assumed lack of subject-object distinction also presupposes that forms like pasnig-t-a ‘snowed’, palyt-a ‘rained’, i.e. of zero-digit verbs (pasnigti ‘snow’, palyti ‘rain’), belonged to the original participles (Schmalstieg 1982: 119), but not that these were only later created by lexical expansion of a morphological marking pattern. In connection with this, the non-congruent sentence pattern of (10) diachron before a congruent (corresponding to the passive) pattern should also (10°) Sen-y misk-ai myle-t-i alte.GEen.pL Wald.nom.pL. M love.PTC.NOM.PL. M “The forests were loved by the ancients” have existed (Schmalstieg 1982: 119f.). However, Schmalstieg does not attempt a credible reconstruction that would support this assumption. (neither do Matthews 1955 or Ambrazas, who in this cover point with narrow staircase; zu Ambrazas s.u.). Apart from that, it would be for his Hypothesis a decisive question, in which (prehistoric) chronological relation is the postuled by him Kakhethwandel and the origin of the *fo-Forman (see Ambrazas 1994: 9). Schmalstieg does not respond to this either. About the *to-formans we can say with some certainty that it has contoured itself in the Baltic (and Slavic) relatively late, possibly even only with the beginning of writing (in the Baltic, therefore, only in the last 300-400 years), according to the diathesis, i.e. a clear determination on one of the arguments of the derivative verb has been made?. Die ® This illustrates intransitive stems with an inactive argument. However, since ergative syntax does not distinguish between active and inactive predicates, sentences like *Tevas.NOM.SG would hypothetically also have to. M eita. NEUT = “father gone’ have passed (cf. the discussion in Ambrazas 1994: 8). ® The process of contouring consists in the resolution of an underdetermined (= diffuse) prototype; this can be represented by a ‘gram’ (in the sense of Bybee, Dahl 1989). Before the resolution (contouring), there is a logical ‘and/or’ relationship between the variants of the prototype, which can only be decided (or left open) under the influence of the context. A typical case of a diffuse (i.e. only contourable by the context) RECONSTRUCTION OF GRAMMATICAL CORE AREAS | 99 morphological cases ide. Languages, on the other hand, are of much older origin. From a typological point of view, the role of morphological cases in the transition from an ergative to an accusative system can be described as the extension of the initial ergative case (used only to mark the highest-ranking argument of transitive predicates) to arguments of intransitive predicates. By such an expansion, the ergative ceases to be an ergative and becomes what is commonly referred to as the ‘nominative’. Therefore, there is above all a reversal of the markedness ratios in the alignment.” And that is exactly what does not happen in the above case (9-10): the genitive remains an oblique case, the sentence pattern in (9-10) remains “non-basic” against the background of an otherwise accusative syntax?. It should also be noted that Schmalstieg's reconstruction prima facie refers only to non-congruent constructions with t-participles. If one wants to explain why m-participles also occur in incongruous sentence patterns of the type (10-11), one must necessarily assume that after the pattern of *to a very powerful analogy has taken place”, and one would have to relate this analogy to other relevant processes in the diachronic chronology of the alignment systems and involved grammatical phenomena. Similarly, the incongruous use of the today gender-neutral t-participles of intransitive verbs (as in example 9) would have to be explained by analogy. After all, Schmalstiegs thesis is based on the fact that an ergative case first arose only in transitive sentence patterns. But again: even if this case had been used by analogy to mark the only argument of intransitive predicates, such an analogy would have led to a marking pattern which would not be classified as ergative alignment but as accusative alignment”. See (9-10): there is the only argument of the intransitive verb gyventi. The morpheme is in Lithuanian the pronoun kas ‘who and/or what’. On the concept of contouring cf. Lehmann (1996; 1999: 234), Wiemer (2001: 47ff.). For the diathesis orientation of this and comparable deverbal formants cf. Wiemer (2004: 288-293) with the relevant references. Vgl. Plank (1995: 1186f., 1991), der in einem Wechsel des Alignment u.a. „consequences of revaluations of non-basic constructions as basic and vice versa“ sieht, sowie auch Dixon (1994: 182-206, passim). Only as a curiosity, the following should be noted: Schmalstieg (1982: 122) speaks, contrary to his own hypothesis —which makes the ergative become the genitive—- and the view of other proponents of the thesis, of an ergative past of the Ide. according to which not the genitive, but the nominative should be considered as the successor of the ergative (this would correspond to the typeologically proven sequence of development just described). Schmalstieg considers this assumption to be “completely unnecessary” (in Schmalstieg 1993: 41, however, he finally takes it into consideration again!). His explanation then ends in an accusative system —without the “intermediate stage” of the ergative system. In the light of this kind of argumentation, one cannot help but get the impression that we are dealing with incoherent ad hoc attempts. X ws The same applies to Slavic n-participles, provided that their incongruent sentence patterns can be linked to structures as in (10-11). Therefore, Stepanov (1978b: 348) pointed out that the complementary morphphonological distribution of nund t-suffixes ultimately speaks against a hypothesis such as that of Schmalstiegs. This is precisely what distinguishes the Lithuanian case from the Indo-Iranian one, which is often cited for comparison purposes (cf. Christen 1998: 58). For the basic marking patterns, see also Comrie (1975: 114f.). Accordingly, this would stand out from e.g. (9-10) resulting case patterns close to an antiergative system, as one finds it for example in modern Finnish. 100 | 100 Years ‘live’ (žmonių) in the same case as the highest-ranking argument of the transitive verb mylėti ‘love’ (senų); Syntactically this corresponds to an accusative pattern, in which of course the morphological marking relationship would be “swap”. Neither this, however, nor an idea. Pre-stage with juxtaposition of the type *tev-as.- NOM.SG. I'm d-ir-b-a /gule-t-a. NEUT can be credibly reconstructed in some way (Ambrazas 1990: 213). Finally, the Schmalstieg hypothesis implies that the present passive and the subject-impersonal in Lithuanian (with the gender-neutral ta/ma participles) developed from an ergative system. However, this would be the only case in the typology so far in which a passive would have arisen from an ergative system. Otherwise, only those cases are known in real (not reconstructed) cases, in which a passive verb has been converted to an ergative marker (see Estival/Myhill 1988 and #1142 from the Konstanz “Universals Archive”). Stepanov (1978a: 355ff.; 1978b: 348ff.) had already taken a position that was only partly contrary to Schmalstiegs assumptions, in that he did not discuss the emergence of the sentence types (9-10) under the assumption of an ergative case marking, but implicitly linked them to an active:inactive alignment, which in his view also included the precursors of Baltic and Slavic (he speaks of „proto-Balto-Germanic“). Slavic") from the early Ide. should have taken over. Accordingly, the oblique marking of the argument in (9) should be regarded as original because an animated speaker appeared in an “untypical” role for him, namely inactive; Accordingly, it was also marked morphologically and syntactically (in the form of lack of congruence). On the other hand, Stepanov sees in sentences of the type (9) a connection to the ide, which has disappeared in Baltic and Slavic languages. Perfect. This is generally considered as a state category (and is therefore associated with the medium or formed with this form:function cluster). After the demise of the perfect in the Baltic and Slavic languages, an empty systemic space had, in a sense, emerged in these languages, which, according to Stepanov, had been filled by infinite forms of the verb and oblique argument marking’- *. The formation of an analytic passive was thus diachronically secondary to the “perfect” (or rather: static) semantics of the Set types (9). Against this thesis it is to be objected, that it just like that of Schmalstieg raises questions about the verifiability of the chronological sequence: in which temporal Relation are the function as perfect, the function in the passive and the function in the (inferential) Evidentialis®? Moreover, Stepanov does not raise the question of whether the sentence type (9) really corresponded to a tripod or rather to an ‘experiential’ (which of the evi- ¥ hereby he also establishes a direct relation to intransitive sentences with nominal predicate and oblique argument marking in modern Russian and Lithuanian; so above all with the dative-experiencer (cf. Russian Mne xonodno, lit. Man šalta “I am cold”) or an external dative-possessor with the AKK of the lowest rank argument as in russ. (dial) Mr. DAT Hoey. AKK 6026n0 wörtl. ‘My leg. Akk schmerzlich' or lit. Man.par headache. AKK wörtl. “My head hurts’. Stepanov overlooks, however, that many of these sentence patterns have apparently only formed themselves relatively late, namely in historical time (cf. in general in Galkina-Fedoruk 1958). ® Especially for this function, sentences of the type (9-10) are well documented in modern Lithuanian. RECONSTRUCTION OF GRAMMATIC CORE AREAS | 101 dential function nearestunde). Furthermore, his assumptions (just like the Schmalstiegs) are based on hypotheses about a change in alignment, which he bases on a purely system-immanent reconstruction, in aufgrund der typologischen Gegen-Evidenz die Beweislast überlassen (s.0.). Schließlich which possible influences through language contacts are completely excluded. Thus, in Schmalstieg's explanations, it comes to completely unnecessary ad hoc interpretations, which are not covered by facts that would be relevant for the reconstruction. So, for example, when he ad libitum sentences of the type (13) Ant Sakel-iy ant Zal-iyjy vainik-us kabin-t-a on Ast.GEN. PL on green.GEn.- pL Kranz.akk.- pL. M hang up.PTc. NEW ‘On the branches on the green are (there) crowns. Akk aufgehängt’, in which the lowest rank argument is marked as an accusative object. Schmalstieg (1982: 121f., 129) speculates here that the verb kabinti ‘hang up’ was only later (in the course of the transitivity opposition) clearly reinterpreted as transitive and that thereby the ¢-participle kabinta (now gender-neutral) formed by it was also understood as transitive (i.e., oriented to the highest-ranking argument). Such an “explanation” ignores the fact that in the diachronic development of participles with -formans the orientation to the lowest rank argument has prevailed (see footnote 29). If in individual cases they are oriented to the highest-ranking argument in the case of transitive basic verbs and can otherwise be formed by intransitive and even zero-digit verbs (and thus be used “actively”), this could be seen to a greater extent as the result of a far-reaching expansion of this deverbal suffix to other than transitive stems and to a lesser extent as the result of a syntactic reanalysis (such as in the Polish impersonal with the no/to participle). Whether Polish influence can be assumed in the case of Lithuanian participles on -fa (as well as on -ma) of transitive verbs with object-rection is still open”. Apart from this, it is much more plausible to assume that the expansion of the deverbal t-formant did not take place from transitive to intransitive verb stems, but rather from intransitive to transitive verb stems —i.e., exactly in the opposite direction of that which would correspond to the Schmalstieg hypothesis. Ambrazas (1990: 200-214; 1994; 2004) put forward a few arguments against the Schmalstieg hypothesis to justify the latter assumption. He also assumes that the early ide. Sentence construction was primarily characterized by the juxtaposition of nominal sentence members with only weakly pronounced word class differentiation. However, the genitive in sentences like (9-10) does not refer to a ergative case back; this has very likely never existed. Rather, the genitive in sentences of the type (9-10) is to be regarded as the continuation of a very old genetivus subiecti, as it is also in other older ide. Verbalnomina * On the diachronic development of the so-called “languages of the world” „impersonal“ passives in the Baltic-Slavic language area cf. Wiemer (in print 2). 102 | The New York Times (nomina actionis) can be observed- ?’. According to this reconstruction, the present-- day gender-neutral fa-participles go back to *o-rooted verbal nouns, in which first arguments were in the nominative, which did not congrue with the participle on -ta (< *to). Such a structure is to this day in the sentence pattern of the kind (14) rug-iai see-m-a | see-t-a Roggen.noM. PL säen.PTC. NEW ‘Der Roggen wird / ist gesät’ reflected. To Verbalnomina with toder m-Suffix the “carrier” of the respective property could be indicated in the genitive (as genetivus subiecti), and that both in transitive stems (valstiecio.GEN- .SG.M sėjama /seta originally ‘of the peasant’s sown’, i.e. sown by the peasant’) and in intransitive stems (valstiecio.GEN- .SG.M gulima /guleta originally ‘of the peasant’s lying/lying thing’, i.e. ‘of the peasant is/was lying’)®. This reconstruction has the advantage that (a) it is much more transparent and (b) it does not distinguish between £fund m-participles or between transitive and intransitive stems. In contrast to the Schmalstieg hypothesis, this approach does not fall into an explanatory constraint that requires additional hypothetical assumptions about thorough analog alignments. This greater clarity is linked to the fact that the sentence structure in (910) can be attributed to an extremely long diachronic continuity, in which the syntactic status of the genitive NP would at most have changed (and this too might only have happened in the last few centuries). Finally, Ambrazas’ approach also allows to integrate the thesis of the possessive origin of the genitive-marked highest-ranking argument (see footnote 37). Due to the extremely speculative character of the ergative hypothesis®, Schmalstieg's decisions, which are usually made ad hoc (cf. above all Schmalstieg 1988: 30ff.) and which cannot generally be solidly supported by facts, as well as the inconsistencies set out above and the fact that Schmalstieg's approach would have to lead to a passive arising from an ergative system (which would run exactly against the unambiguous picture of typologically evidenced cases), the explanatory approach of Ambrazas would clearly have to be preferred. 7” This point of view is only formulated in such a clear form in Ambrazas (2004). Earlier, he also seemed to have been not averse to Holvoet’s thesis that the genitive in these sentence types was derived from a possessive, initially only adnominal case, which had later been reinterpretated as an adjunct. Stepanov also assumed that the genitive in (9) was to be interpreted first possessively. He saw this, however, as a further development of the early Baltic and early Slavic “perfect substitution” (cf. 1978b: 347f.). ® It is noteworthy in this context that in Lithuanian dialects the genitive is even more frequently used for taund ma participles of intransitive roots than for such participles of transitive roots (Ambrazas 1994: 9). On other objections to the thesis of the "ergative past" of the Ide. with further references cf. Comrie (2001: 24f.) and Matasovi¢ (2000: 330f.). RECONSTRUCTION OF GRAMMATIC CORNER AREAS | 103 3. UNIQUE CASE IN SLAVIC: ‘ESSE’ AND ‘HABERE’ IN POSSESSIVE CONSTRUCTIONS Until recently, other researchers in addition to Schmalstieg have been building their reconstructions on hardly durable ground. Sometimes even goats become a supposedly archaic, still from the proto-Ide. derived morphosyntaksis instead of innovations, which may be in individual ide. Languages have only been established through language contact (including with non-identical languages). A very clear case is e.g. Bauer (2000: 10, 151), who cites the Russian possessive construction of the type u + GEN (y mens knuea ‘I have a book’ etc.) as an example of traits that are “inherited from the protolanguage”. The question arises whether the chronology has not been turned upside down in some way. Since this possessive construction can also be placed in a close connection with the question of alignment (see section 2), it is worthwhile to go into the methodological problem a little more closely. So let us take a position on the thesis that the possessive construction u + GEN (+ EssE, with the possessive in the nominative) in Russian represents a Finno-Ugric substrate. This thesis has been discussed and actually accepted by Veenker (1967: 117-119), while Panzer (1982: 330f.) and Vasilev (1973) have cast doubt on it. Panzer’s skepticism, however, is de facto based solely on the (somewhat dogmatic) assumption that Slavic languages are actually “BE-languages” and not “HAvE-languages” (following Isacenko 1974) and that it is therefore more likely that Russian merely continues an inherited state of Proto-Slavic; if the other Slavic languages use a HABERE-construction for the possessive expression (e.g. imeti and its equivalents), this is due to a supporting effect of Greek (or, by way of mediation, of e.g.), Latin or later West European languages. If Panzer is probably right about the supporting effect, this would presuppose that the HABERE verb is a Slavic common property from Proto-Slavic. Also the material of the etymological dictionaries speaks relatively clearly for such an assumption (cf. Vasmer 1953: 479 and Walde 1930: 124). Panzer's skepticism about Finno-Ugric influence is therefore not particularly well-founded, and it is neither refuted nor confirmed. Rather, it should be considered that both types of possessive constructions have existed in parallel in prehistoric times and the role of a Finno-Ugric substrate may have been to support a repartition of these competing constructions in favour of the adhesive (with ESSE). This is precisely the assumption supported by the investigation of Safarewiczowa (1964, in particular p. 17, 55f.). It must therefore be noted that Veenker's important remark that the adessive esse-construction only occurs in Russian, i.e. in the north-eastern periphery of the Slavic-speaking area, as a predominant possessive construction (with all its functional developments) over the HABERE construction, against the evidence of Vasilev that the adessive construction is already evidenced in early documented stages of various Slavic, among them also South Slavic, languages, and namely in addition to the possessive function in the narrow sense also in partitive function (with inalienablia), the function of possession and that of relative affiliation. Vasilev Page 104 | (1973: 363) also rightly points out that, on the basis of knowledge of the settlement movement of the direct ancestors of the East Slavs, one must assume that direct contact between these and Finno-Ugric tribes did not occur until the 5th century. Jh. AD could have begun. Since the construction ‘u + GEN + EssE’ already appears in the earliest preserved East Slavic texts (cf. Safarewiczowa 1964), almost 500 years can therefore be assigned as the “virulent period” for its establishment. This, one should think, would be sufficient for a calquation under Finno-Ugric substrate conditions, for which there have been ample prerequisites in Russian since that time (cf. to this a.o. Veenker 1967: 17–21). This would be another reason, Panzers and Vasilevs To dampen skepticism. The beginning of the decay of Proto-Slavic (about 500 AD) coincides more or less with the beginning of the Slavic settlement movement to the northeast. Observations that in the historically documented period ‘u + GEN + ESSE’ occurred practically in the entire Slavia are not very meaningful. Vasilev himself points out (in addition to Safarewiczowa 1964 for common Slavic and older Russian) the fluctuation in the frequency of use of this construction and the HABERE construction. The fact that it has remained in the norm in the long term only in Russian is therefore most plausibly attributable to the continued substrate influence of Finno-Ugric speakers. The increased use of this construction in possessive (and semantically related) functions is also in recent times easily explained by contact with idioms, which already used it exclusively or predominantly in this way. This is shown by the significantly more frequent occurrence of u + GEN in possessive function in northern ‘polszczyzna kresowa’ (vs. standard Polish) as well as in Lithuanian colloquial language and those Lithuanian dialects which have been in close contact with East Slavic varieties. In general, these considerations are supported by a comparison with the Baltic. Also in this a HABERE verb can be counted to the hereditary vocabulary: lit. tureti is to be understood as a general Baltic ablatio to the root *fuer (cf. the etymological relationship with tverti ‘(to) encompass’; Fraenkel 1965: 1142f. ; Pokorny 1959: 1101). In Latvian there is no longer a HABERE verb; the verb turet, which is lit. furėti, has acquired the meaning of ‘keeping’ and is mainly used in this concrete meaning. It can therefore be assumed that in the Baltic languages a HABEREverb also participated in the formation of basic sentence types and thus determined the alignment (as this is still the case in Lithuanian). Furthermore, it is to be assumed that Latvian has had much more extensive linguistic contacts with Baltic-Finnish languages and is in many respects much more strongly influenced by a corresponding substrate than Lithuanian. The repression of the HABERE verb in Latvian could thus be considered analogous to an identical process in Russian —and this makes both area-linguistically as well as with regard to the known settlement-historical facts more than Sinn. For the time being, we can only say that this question still needs to be examined in detail. The previous discussion of the EssE-HABERE problem alone, so to speak “in a bird’s eye view”, shows the following: if one wants to speak of “archaic syntax”, one should first clarify how the competition relationship of the constructions in question should be addressed. RECONSTRUCTION OF GRAMMATICAL CORE AREAS | 105 their morphosyntactic constructions (marking methods, etc.) and how their chronological relationship to each other was shaped. It may well prove that a coding pattern that appears archaic from today’s point of view has not existed continuously and unchallenged since the time of a protolanguage (however to be used), but could be temporarily pushed back, only to be activated again later as a primary procedure. As a trigger for such a revival comes then above all language contact (especially substrate influence) into question, so that one should not stop at system-immanent explanation attempts. 4. A PAIR OF CONCLUDING POSTULATES In the above discussion of case examples of central morphosyntactic coding methods and their supposed diachronic changes, hardly anything can be proven, but at best made plausible. However, the plausibility of explanatory approaches will increase to the extent that arguments can be drawn from areas that go beyond internal system reconstruction. Internal system argumentation can be considered as a necessary condition of the explanation, but it is not a sufficient condition, especially when there are alternative approaches. Nothing else should be demonstrated in this post. The examination of the arguments underlying the theoretical approaches of Stepanov and Schmalstieg should show, representatively for many other cases, how necessary the comparison of the historical-- comparative reconstruction with typological facts and with contact linguistic findings appears and what can remain of this reconstruction if one does not endeavour for such a comparison. So here to the idea. Prehistory of the historically documented Baltic (as well as Slavic) no competent judgments should be made, since in relation to it a number of conclusions and postulates can be formulated: Speculative assumptions about disruptive changes in alignment are anything but convincing, they lead to internal inconsistencies and for the most part they even contradict typological comparison data. In any case, they lead the adherents of the theory of an ergative syntax into explanatory distress. Especially in the case of primarily syntactic problems, the consideration of area-linguistic factors (language contact, substrate formation) furthers the explanation, making it much more plausible and simpler. In addition, it allows an assessment of the postulated developments under the conditions of real time and real communication as well as under the assumption of an already existing variation of functionally equivalent means. Finally, it allows a better appreciation of what is archaic and what is innovative. In this regard, a precise distinction should be made between form (etymological “substance”) and functions: it is apparently precisely the latter that can be altered by language contact. In particular, latent or previously in the foreground functions can be (re)activated by voice contact. 112 | The New York Times Bisang, W., Himmelmann, N. P., Wiemer, B., eds., What makes Grammaticalization? A Look from its Fringes and its Components. Berlin-New York: Mouton de Gruyter, 271-331. —im Druck 1: O razgranicenii grammaticheskih i lexicheskih protivopostavleniij v glagol''nom slovoobrazovanii, or: What can aspectologists learn from the examples of sha-verbs? Lehmann, V., ed., Semantika i struktura slavjanskogo vida IV. Munich: Sagner. —im Druck 2: On the diachronic and areal background of constructions called ‘impersonal’ in Polish, East Slavic and Baltic. Manninen, S., Hiietam, K., Kaiser, E., Vihman, V., eds., Passives and Impersonals in European Languages. Amsterdam, Philadelphia: Benjamins. Björn Wiemer Gauta 2004 11 19 Universität Konstanz Department of Linguistics Universitätsstrasse 10, 78457 Konstanz, Germany [email protected]