Lietuvių kalbos bendraties konstrukcijų raida

LIETUVIŲ KALBOTYROS KLAUSIMAI XXXIII (1995) GRAMATIKA IR LEKSIKOLOGIJA VYTAUTAS AMBRAZAS THE DEVELOPMENT OF INFINITIVE CONSTRUCTIONS IN LITHUANIAN 1. THE ORIGIN AND ORIGINAL MEANING OF THE INFINITIVE 1.1.1. In contemporary Lithuanian, the infinitive is one of the three principal forms of the verb. Its use is also closely associated with the verb: it governs cases in the manner of a verb and forms close syntactic combinations with finite verb forms that function as auxiliaries (turiu skubėti, pradėjo važiuoti, ėmė tempti and so on). Forms of this kind, called infinitives, are found in almost all modern Indo-European languages. A comparison of them reveals striking differences in form, alongside considerable similarity in their verbal use. A historical examination of infinitives shows that they did not originally belong to the verbal system, but were incorporated into it in the various Indo-European languages at different times and to different extents. The infinitives of all related languages have stems of verbal nouns, and some (especially in earlier periods of development) have retained case endings characteristic of nouns. Both the stems themselves and the case endings vary greatly: of the 14 stems distinguished by R. Jeffers (1975:135), infinitives with only the tu and ti stems are attested in many Indo-European language groups, while those with other stems occur in only 1–3 groups. However, infinitives formed from the same stem sometimes have different case endings and therefore do not derive from the same original form. Examples include infinitives with the tu stem and the endings: Old Prussian -ton, -twei, Sanskrit -tos, and so on. In the oldest attested Indo-European languages (Hittite and Old Indic), infinitives are sometimes difficult to distinguish from verbal nouns—names of actions: the same forms are used as infinitives in some contexts, while elsewhere they retain their nominal function.1 1.1.2. Lithuanian infinitive forms in -ti and -tie < *-tei likewise developed from verbal nouns—names of actions (abstract nouns)—with the ti stem, such as mirtis, prapultis, būtis. Verbal abstracts of this stem were once highly productive and widely used in the eastern Baltic languages. After action nouns with the suffix -imas/ -ymas, now predominant, took their place, the old abstract forms with the ti stem became frozen and eventually turned into infinitives. Baltic infinitives in -i (Lithuanian nešti, nešti, nešti-s, Latvian nest, nestis, Old Prussian billīt “to speak, to talk”) and in -ie < *-ei (Lithuanian neštie, neštie-s, Latvian nestiē-s) correspond to the singular dative of nouns with the i stem (-i and *-ei may be ablaut variants of the same dative form). This 1 Pasak V. Ivanovo (1963:152–153), hetitų kalbos veiksmažodinių daiktavardžių vn. datyvo formų su -anna virtimą infinityvais galima pastebėti lyginant istorinius šaltinius. 74 suggests that forms such as mirtį/ mirtie were once used as datives of verbal nouns (meaning “for death”), and that this usage provided the basis for their development into infinitives. Many other related languages likewise tend to form infinitives from the datives of action nouns. For example, of the 700 infinitives found in the Rigveda, as many as 600 correspond to the dative of verbal nouns (Sgall 1958; Szemerényi 1970:298). 1.1.3. Formally, Baltic infinitives in -ti/ *-tei can also be derived from locatives of the ti stem (Otrębski 1956:243; Stang 1966:447; Hamp 1985:199—202); a locative origin has also been proposed for the Slavic infinitive in -ti < *-tei (Stang ib. 447; Vaillant 1966: 127).2 The dative and locative interpretations of these forms are not fundamentally opposed, since in the earliest periods of Indo-European development, datives in -i/ *-ei may have coincided with forms used adverbially, which were later incorporated into the case paradigm and interpreted within it as a separate case—the locative (Toporov 1961: 275 ff., 285; Mažiulis 1970: 80, 128—129, 248 ff.). F. Trávníček’s (1956: 167 ff.) account of the origin of infinitives from Indo-European adverbs can be interpreted in the same way. However, this formal relationship between dative and locative was apparently inherited from a period before infinitives arose. In the use of infinitives in both Lithuanian and the other Baltic languages, it is difficult to discern traces of locative meaning. 1.1.4. In Prussian, alongside the infinitive in -t (dat “to give,” gērbt “to say”), an infinitive in -twei is widely used, e.g. dātwei “to give,” poutwei “to drink,” westwei “to lead.” It developed from the singular dative of action nouns with the tu stem, which had the ending -ei, that is, *-tu-ei. Infinitives of precisely the same origin in -tave also existed in Sanskrit (cf. dhātave “to give,” attave “to eat”). However, in the eastern Baltic languages there are no infinitives derived from the dative of nouns with the tu stem, while the supine (purpose form) in -tų < *-tum derives from another case of the tu stem—the singular accusative. Thus, the Baltic languages agree in the case form of the action noun that gave rise to the infinitive, but differ in part in the stem. This shows that the datives of verbal nouns were incorporated into the verbal system and became infinitives not in Proto-Baltic, but during the independent development of these languages. It is precisely for this reason that Baltic data help us better understand the formation of the infinitive, since the connection with the original use of verbal nouns has not yet been entirely effaced here. Other Indo-European languages, too, apparently inherited not common invariant infinitive forms, but common conditions for their formation. These conditions are syntactic and must be sought in the use of infinitives. Internal reconstruction is useful here. Applying it to the syntactic system of Lithuanian, we attempt to distinguish within it the ancient types of infinitive usage and to determine the structure of the constructions from which the infinitive developed as a distinct, invariable verbal form. 2The linguistic literature also discusses the origin of the infinitive in -ti from the locative of the consonant-stem -t (Brugmann 1911: 424; Endzelins 1951: 918; Hamp 1985: 201—202; Zinkevičius 1981: 164). This view is implausible, if only because Baltic languages have no action nouns with a t stem at all. 75 1.2.1. In contemporary Lithuanian, the infinitive is most commonly associated with finite verbs expressing: a) the beginning or end of an action (I begin / start working, set off running, stopped walking, finished writing), b) ability, obligation, desire, skill, or effort (I can walk, we have to return, intends to study, tried to escape, I want to leave), c) an excess or insufficiency of an action (gave too much to drink, used too little salt), and so on. Here the infinitive specifies the content of the main verb, serves as its necessary complement, and the subject of the actions expressed by the verb and the infinitive is the same. Infinitives also form close combinations with verbs expressing an order, encouragement, or request; a prohibition or hindrance; permission, invitation, and so on, but here the action of the infinitive is attributed to a different subject (ordered [someone] to be quiet, asked [someone] not to sing, forbid [someone] to sleep, prevent [someone] from reading, allowed [someone] to go for a walk, invited [someone] to come in). In Lithuanian, as in many other modern Indo-European languages, constructions of these types are the most widespread and form the principal layer of infinitive usage. Here the infinitive is closest to finite verb forms, and traces of its earlier usage are difficult to discern. 1.2.2. The infinitive’s old—nominal—function is more clearly evident in constructions in which it occurs with full-meaning verbs (a) or verb phrases with a necessary complement (b), is not determined by their valency, and denotes the purpose or intended use of an action or of the thing under discussion: a) We recently moved here to live LnkV; The women sat down to spin ŽemR I 15; And that man remained at that manor to serve as a lord BsV 300³; b) They stationed servants to guard [it] BsP II 48; The goat found hay to eat Lt III 116; Laid a plank to cross [it] Pvn; After Gryta had finished spinning, she sorted the yarns to weave DnR 125; You cut a stick to lean on PP 206. Jau Jonas Jablonskis (1935: 30) atkreipė dėmesį į tai, kad tokia tikslą žyminti bendratis iš prigimties (“pagal kiltį”) yra veiksmažodinių daiktavardžių naudininkas. Pagal savo funkciją ji atitinka daiktavardžių naudininko formas ir gali būti vartojama su jomis pagrečiui: orei (= armiui, arti) žagrių neturime, plg. dar: sukosi lizdelį gyvenimui/ gyventi; ieškojo vietos nakvynei/ nakvoti; atnešė bulvių valgymui/ valgyti; lova bus reikalinga poilsiui/ pailsėti; sūnus gabus mokslui/ mokytis. Tikslo bendraties ir naudininko vartosenos paralelumą dažnai matome tiek tarmėse, tiek rašomojoje kalboje, plg.: Varnų ben porelę sau pietums nusisauja DnR 241 / Kad Dočys porelę varnų kept nusisauja 243. The use of the dative in these constructions is grounded in the meaning of purpose (goal), long reconstructed for the Indo-European dative, which combines freely with verbs (Brugmann 1911: 555 ff.). This meaning was especially characteristic of verbal abstracts. Although the Lithuanian infinitive has already lost its direct connection with declension, in the cases mentioned it has preserved unchanged to this day the original meaning of the dative of purpose. 3Sources of examples are indicated by the same abbreviations as those used in the large “Dictionary of the Lithuanian Language” and the three-volume “Grammar of the Lithuanian Language” (V., 1965—1976, abbreviated: LKŽ I—III). 4 Daugiau pavyzdžių su tikslo reikšmės infinityvu žr. Jablonskis 1935: 30; 1957: 475—476; Fraenkel 1928: 129—132; Balkevičius 1963: 226—227; LKG III 170—171, 260. 76 2. DATIVE WITH THE INFINITIVE (DATIVUS CUM INFINITIVO) 2.0. Once incorporated into the system of verb forms, the Lithuanian infinitive acquired verbal government of cases. Infinitives of transitive verbs govern the accusative, and more rarely the genitive, much like corresponding forms of other verbs; cf.: repaid the debt—ordered [someone] to repay the debt, wanted to repay the debt, sent an order to repay the debt; waited for the parents—asked [someone] to wait for the parents, sent a request to wait for the parents, and so on. However, in cases where the infinitive is freely associated with the sentence nucleus and has the aforementioned meaning of purpose or goal, the object of the action it denotes is expressed not by the accusative but by the dative; cf.: there was not enough money to repay the debt. This is the so-called dative with infinitive construction (dativus cum infinitivo, hereafter DI). In terms of the relation between the infinitive and the dative, this construction seems to stand apart, to “fall out” of the system of syntactic relations prevailing in the modern language. From the perspective of internal reconstruction, this may be considered grounds for the assumption that this is a syntactic archaism, based on an earlier state of syntactic relations. It is therefore useful to discuss this construction in greater detail. To establish the original structure of DI, we must set aside superficially similar utterances in which a finite verb form forms one of the close units of the type mentioned in 1.2.1 with an infinitive, while the dative serves as a complement of the verb, e.g.: ordered the boy to drive, hinders / helps people to work, advised his son to get medical treatment, forbade others to hunt, and so on. Hereafter, DI will refer to constructions in which the dative is not determined by the verb’s valency and, together with the infinitive, denotes purpose or intended use. plowed (the field) for sowing rye 2.1.1. According to their relation to other sentence constituents, three groups of DI constructions can be distinguished: a) DI is associated with a verb (predicate): Don’t go about advising others Šim; Please, sir, drink up, drink to put the worm to death Jabl; He fired into the air to frighten people Myk - Put; He can earn no more than enough to feed himself Šln. b) DI is associated with a verb and a noun or pronoun governed by it: I brought poles to build a fence Vb; I will give you a silk handkerchief to wipe away your tears tts (Jabl); He took a young girl to look after the children Simon. c) DI is associated only with a noun serving as the object or subject of the sentence: They completely took away the storeroom for putting things away Simon; That one, used for chopping firewood, has become completely bluntSrv. Constructions of group (b) are the most widespread and predominant. On the other hand, constructions of group (c) occur only occasionally. Even from a synchronic perspective, they can be regarded as derived from group (b), where the relation of DI is felt not only with the noun but also with the main verb. This relation generally remains even when the predicate is an expressed or implied form of the verb ‘to be’, e.g.: What a good spit this is for roasting a lamb BsP II 91. J. Jablonskis (1957: 598) perceptively observed that a nominal component is implied here: suitable, intended, or the like: Not every mouth is suitable (good, meant) for slurping porridge. 77 This is also confirmed by the questions asked in folk tales about the concepts denoted by nouns associated with DI, e.g.: A stork comes with a large crop. — What is that crop for [i.e., what is it intended or needed for]? — For putting money in. — What is the money for? — For buying meadows. — What are the meadows for? — For grazing cows LTRn 396 (for a fuller discussion, see Toporov 1963: 151). Examples from group (a) show that a connection with a noun is not necessary: DI here depends directly on the main verb. Constructions in this group can also be understood as answers to the questions For what? For what purpose?, cf.: What are you plowing here for, Juozas? — To sow flax (JablRR I 598). A clear connection and continuum between groups (a) and (b) are easy to observe: if we add to the verb a noun denoting an object (cf.: Bend the glass, young sir, to kill the worm) or, by paraphrasing, express the concept of an action with a phrase (cf.: He fired a shot into the air to frighten people), we obtain constructions of group (b); cf. also: At the manor, the bell rings for milking the cows; in Žem and Bokštas, the bells rang to see the young girl off tts (JablRR I 599). Thus, groups (a) and (b) are variants of the same type: DI is associated with the sentence core—the verb, or the verb together with a case phrase dependent on it. 2.1.2. Although DI is often used as an equivalent of a purpose subordinate clause, it cannot, by any means, replace every purpose subordinate clause, nor vice versa. In colloquial speech, DI is especially common in cases where a single dative expressing purpose or intended use (without an infinitive) is also possible. Cf. the comparisons given by Jablonskis: I brought home several pounds of meat for dinner / to cook for dinner; People are hauling bricks for the church / to build the church; A cross was erected in memory of his work / to remember his work JablRR I 596 (see also Valiulyte LKG II 34 — 35). Like a noun in the dative, a single infinitive is also linked to the sentence core: He laid down a board to cross Pvn; You cut a stick to lean on PP 206 (see 1.2.2). Even in the same sentence, purpose or intended use can sometimes be expressed by a noun in the dative, by an infinitive, or by both together. This parallelism in usage is observable in constructions of both group (a) and group (b), cf.: a) At first I plowed only for barley / to sow / to sow barley. b) Father brought home firewood for the cottage / to burn / to burn in the cottage. The manor gates are wide for horses / to enter / for horses to enter. All this shows that DI consists of two adjuncts that expand and depend on the sentence core: the dative of purpose (intended use), and the infinitive, which likewise expresses purpose. As already mentioned, the infinitive itself developed from the dative of a verbal noun—the name of an action. Thus, the relations between the construction’s constituents discussed above make it possible to discern the following original structure of DI5. 5 Abbreviations: N — nomen (nominal), V — verbum (verb); NV — nomen verbale (verbal noun), nom — nominativus, dat — dativus. The arrow → denotes a dependency relation, ↔ a mutual association. 78 For example, the close relationship between the constituents in the constructions pirkau arklių laukams arti ‘I bought horses to plough the fields’ can be represented as follows: Thus, DI directly reflects the origin of the infinitive in the dative of verbal nouns. In its original structure, DI appears to be a dependent, double dative of purpose (goal) (dativus duplex) attached to the sentence nucleus. 2.1.3. This conception of the original structure of DI essentially accords with the views expressed by A.Potebnia (1958: 387 ff.), F.Kuršaitis (1876: 406), J.Jablonskis (1935), and, later, P.Trost (1958: 273) and I.Marvan (1960: 10). Jablonskis characterized DI especially aptly, saying that here “to express purpose we use two datives, one of an ordinary substantive and the other, now called an infinitive. The substantive indicates for what (30) thing, for what time, or for what purpose something is needed or done, while the infinitive added to such a dative explains and specifies even better the purpose for which some thing is suitable or used” (Jablonskis 1935: 30—31). Incidentally, F.Kuršaitis wrote only about the connection between DI and nouns, while H.Engertas (1937: 109—111, 1938: 45—56), developing his idea further, concluded that the construction was purely attributive, answered the question “What is it?” and had no connection with the main verb. Linguists who later examined DI from the perspective of language culture likewise focused primarily on its semantic connection with the noun functioning in the sentence (Balčikonis 1978: 267—269; 392—393; Paulauskienė 1970: 49 ff.; Piročkinas 1986: 188—190; similarly, Haudry 1968; 1977: 104 ff., from the perspective of other Indo-European languages). However, the examples in group (a), already discussed, and the derivative nature of attributive constructions make it impossible to link DI historically only with a noun functioning in the sentence. Since DI expands the sentence nucleus, in which the verb plays the most important role, E.Fraenkel’s claim remains irrefutable: the construction originated in sentences in which DI essentially (“im Grunde”) depends on the main verb (1925:60). 2.1.4. In the DI construction itself, considered in isolation, the dative of a noun or pronoun most often denotes the object of the infinitive’s action. Some scholars who have explained the origin of DI and corresponding constructions in related languages have proposed that an accusative originally occurred here instead of the dative, governed by a verbal noun expressing purpose (since verbal nouns in ancient monuments of Indic and other languages generally retain the case government characteristic of verbs). The accusative here was supposedly replaced by the dative under the influence of the main verb (Brugmann 1916: 917 ff.) or of the dative of the action noun (attraction), at a time when the latter had not yet become an invariable infinitive (Schleicher 1856: 311; Delbrück 1897: 470; Miklosich 1874: 79 844). Recasting the notion of attraction as agreement, W.Boeder (1980) similarly explains the relation between the constituents of the construction. This view of the origin of DI is contradicted not only by the relationship between the construction’s constituents and the sentence nucleus, already discussed, but also by the variety of semantic relations between them. In Lithuanian, the dative of a noun here can denote not only the object of the infinitive’s action, but also its subject: I poured milk for the kittens to lap Ktk (cf.: so that the kittens might lap it); You have a daughter for the dragon to swallow LB 190 (cf.: so that the dragon might swallow her); Twenty nesting boxes were put up in the forest for the tits to nest in (cf.: so that the tits might nest) LKG III 173. Such constructions also occur in old linguistic monuments from the very beginning of the written record, cf.: lawonus kitus ing wandeni mete szuwims walgiti Mž 427 (= so that the fish might eat). Sometimes the dative also denotes an adverbial circumstance of the infinitive’s action, e.g.: She took yarn to spin at night (cf.: so that she could spin at night). Here the connection of each DI constituent with the sentence nucleus is obvious (cf.: took it for the night / took it to spin). 2.1.5. Although the constituents of the construction underlying both DI types were dependent on the sentence nucleus, the nominal constituent (not the action noun) apparently occupied the central position between them from the very beginning. E. Fraenkel (1925: 60; 1928: 131) regarded the infinitive (by origin, an action noun) as a complement added to the dative of a noun (“epengetischer Zusatz”). This kind of relationship between syntactic units, in which one supplements the other, while both depend equally on a third, is called appositional. Units linked in this way can each independently substitute for the other, but one of them is usually primary in meaning, while the other is supplementary, explanatory, and most often follows the primary one (Seiler 1960: 198; Valeckienė 1984: 111). That the dative of a noun was semantically primary in the DI construction is indicated by its predominant preposition both in Lithuanian and in the ancient texts of other Indo-European languages (cf. Sgall 1958: 202; Kurylowicz 1964: 162). I went to plough the field 2.2.1. The lexical restrictions on the construction’s use and the dependence of its structure on the lexical meaning of the verb also indicate the primary connection between DI constituents and the main verb. When the main verb denotes movement, a change in an object’s state, or a change (or absence of change—as with the verb ‘remain’), the goal is expressed not by the dative but by the genitive with an infinitive, for example: I will go to the hill to reap yellow wheat VD I 290 (Alv); The others ran off to fetch the horses Žem; The women sat down to spin flax Vrn; The girl stayed home to guard (but: I left the dog to guard the house), see LKG III 169. As in constructions with the dative, the goal can also be expressed by just one of the constituents (a single genitive or a single infinitive). The old man came back for lunch/ to have lunch/ to eat The supine, characteristic of the dialects of Eastern Lithuania, has the same syntactic relationships as the infinitive in the corresponding constructions; cf.: The old man came back for lunch/ to have lunch (supine)/ to eat (supine) The supine, like the infinitive, is nominal in origin: it developed from the accusative of action nouns in the fossilized *tu-stem. Actions of this stem 80 abstract nouns (abstracts) played an important role in the word-formation system of Proto-Baltic, and only later did stem abstracts with ti come to predominate markedly in the East Baltic languages (S.Ambrazas 1987: 92 — 93). In sentences with verbs of motion, the accusative of tau-stem abstracts initially expressed the goal of motion, a meaning attested in other Indo-European languages as well in earlier periods of their development; cf.: Skt. sa gamad indro... vasūnām dātum RV V 36.1. “come, Indra... to give gifts (lit.: giving)” (Jeffers 1976: 2), Lat. eunt spectatum “they went to watch (lit.: watching),” and so on. However, tu-stem abstracts became unproductive very long ago; their accusative became fixed and turned into an uninflected form—the supine. The supine attached to the genitive expressing purpose, which in turn depended on the main verb, apparently already as an uninflected form, no longer carrying case meaning. Later, in many constructions with verbs of motion or change of state, the infinitive displaced the supine. J.Palionis (1967: 199 — 200; 1974) established that even in the 16th and 17th centuries the supine (the “seeking” form) was used in many Lithuanian dialects, where it was later replaced by the infinitive. Since ti-stem verbal nouns assumed a predominant position among action nominals later, their dative retained its case meaning for longer. As we have seen, the infinitive still retains the meaning of the dative in many purpose constructions today. Therefore, the original state of syntactic relations has been better preserved in dative constructions with the infinitive.6 Change in syntactic relations 2.3.1. Although the primary relation of the DI components to the sentence nucleus is still perceptible in many cases, the construction itself can already be used as an independent syntactic unit expressing the purpose (aim) of an action. It also occurs in sentences in which a dative alone (without an infinitive) would be inappropriate or would have a completely different meaning, e.g.: To summon friends outdoors, only before the evening meal; Have you seen what a good spit this is for roasting elk BsP II 91; The others promised to let the children dig the pit, too Jabl. In such cases, the dative tends to be understood as the object of the infinitive’s action. Only occasionally, however, does this also affect the grammatical form of the construction: the dative is replaced by the usual case of the direct object—the accusative, e.g.: I lit a little fire to cook breakfast Sch 158; Grant me, dear God, a clear day to sow rue VD I 373. Such expressions abound in records of the dialects of Lithuania Minor and in Donelaitis’s writings, e.g.: They want to weave themselves bast-shoe sandals DnR 201 (see also Jablonskis 1957: 598). The spread of the accusative into the dative’s place is already observable in 16th–17th-century texts, e.g.: He appointed those bishops to govern the church of God SP I 226₃₁; For I am too old to take a husband BrB Ruth 1:12 (das ich einen man neme LB). 6Studies of corresponding constructions in the Vedas have likewise noted that the oldest usage is preserved there by those infinitive forms that separated from verbal nouns latest (Sgall 1958: 209). 81 In the standard language, such constructions are still corrected as incorrect (cf. KPP 78), but usage is already variable in some places, and the norm for case usage has been discussed more than once (Balčikonis 1978: 267—269; 392—393; Pupkis 1980: 144; Šukys 1984: 81—83; Piročkinas 1986: 188—189). 2.3.2. As syntactic relations were regrouped, the accusative also began to penetrate constructions dependent on verbs of motion and change of state. Alongside the genitive, it is often used in the Western Aukštaitian and Samogitian dialects; it is attested in written sources from the 17th century, and in folklore from the earliest recordings, for example: And three young men rode up to water their brown horses StnD 31; Mother sent me to sweep the manor house Klv (VD I 333); I go to tether the cow/cows Dkš; Leaving his legs behind, he ran to fetch an axe BsP I 35; The boy sent down from heaven is running to see S1G I 53 (see also LKG III 169; Piročkinas 1972: 133). However, in Donelaitis’s writings (unlike the cases mentioned in 2.3.1), the genitive of purpose is well preserved here; cf.: I made my way to the edge of the forest in the dark to steal trees DnR 201 and elsewhere. Jonas Jablonskis (1935: 36) yra atkreipęs dėmesį į tai, kad kilmininkas su siekiniu vartojamas tada, kai jis ir vienas galėtų reikšti veiksmo tikslą (t.y. kai tarp veiksmažodžio ir kilmininko yra tiesioginis ryšys — V.A.). Kai vienas kilmininkas dėl sakinio prasmės negalimas, siekinys valdo galininką (kaip ir bendratis ką tik cituotuose pavyzdžiuose), pvz.: Puolė visi bučiuotų rankas seneliui; Stankunas nubėgo sudėtų pinigus advokatui. Pasak Arnoldo Piročkino (1972: 132 — 133), toks skirtumas daromas ir dabar kai kuriose šnektose (apie Veliuoną, Jurbarką). 2.3.3. A later stage in the development of the DI is attested in Latvian. Constructions in which the dative marks the subject of the infinitive’s action are well preserved there and correspond in structure to Lithuanian ones; cf.: I would not take a tall wife for the wind to sway ME IV 548, gloss: “I would not take a tall person for the breeze to sway” (see also Ozols 1967: 150). As in Lithuanian, purpose is also expressed here by a single noun in the dative, for example: got up for work “got up for work”; what I will leave for remembrance “what I will leave as a memory” (Bergmane 1959: 379; 1962: 31), and so on. That the purposive infinitive in Latvian has also retained the dative meaning of an action noun is evident from K. Mülenbach’s comparisons: There is no soup even to smell (= for smelling); “There is no soup even to smell (= for smelling); It was beautiful to look at (= for looking) “It was beautiful to look at (= for looking)” and so forth (Mülenbachs 1928: 193). On the other hand, a noun or pronoun associated with the infinitive and denoting the object of the infinitive’s action already takes the accusative, not the dative, in Latvian, for example: I will give my little horse the rose garden to plough “I will give my horse the rose garden to plough” (Mülenbachs 1928: 193). The accusative is used here instead of the dative already in 17th-century writings, for example: God made... the luminaries, the greater one to rule the day and the smaller luminary to rule the night G1B I Moz 1,16, gloss: “God made... the luminaries, the greater one to rule the day and the smaller luminary to rule the night”; he took a linen skirt to deceive them Judit 16,8, gloss: “he took (i.e. put on) a linen skirt to deceive him.” Thus, the regrouping of syntactic relations in Latvian has already changed the con- 82 the grammatical form of the construction. 2.3.4. In Latvian, the accusative became established in place of the genitive also in constructions with verbs of motion and change of state, e.g.: Nu es iešu sētņā kājas, rokas pūtināt BW 6936, gloss: ”Now I will go into the yard to rest my legs and arms”; Tur teceju slotas griezt BW 1271, gloss: ”I ran there to make brooms” (Ozols 1967: 151); Visi spiedās uz durvīm saņemt negaidītos viesus MLLVG II 20, gloss: ”Everyone pressed toward the door to welcome the unexpected guests”; Saimnieks sūtīja puiši krūmus izplaut (Mülenbachs 1928: 196), gloss: ”The master sent the lad to cut down the bushes.” The accusative had already become established here in written sources of the 17th century, cf.: tas to sutia uz sawu cematu cūkas ganīt EEV(ASpr I 6210), gloss: ”that one sent that one to their own field to herd pigs”; ESUS egaa namma... maize est ib I 194, gloss: ”Jesus went into the house to eat bread.” However, dialects that have preserved the supine have also preserved the older genitive with it, e.g.: Iēsim guōvu slāuktul! Valmiera (Endzelins 1951: 562); Tēvs aizgāja bīšu kāptu ME I 301 ”Tėvas išėjo bičių koptų”. Correspondences in Other Languages 2.4.1. In other related languages, the development of infinitive constructions has progressed considerably further. In some of them, infinitive constructions can be found only in the earliest periods of their development. In Old Church Slavonic and Old Czech, remnants of infinitive constructions are found which researchers regard as syntactic archaisms; their structure and meaning correspond to Lithuanian constructions. In these, the dative can denote either the subject of the infinitive’s action, e.g. Old Church Slavonic sětvoriť mi chizinu sěsti mi vъ neï Supr 204,2, gloss: ”he will make me a little hut for me to sit in it”, or its object, e.g. Old Czech kúpichu pole pútnikóm hřesti Hrad., ”they bought a field for burying travelers”7 The dative’s connection with the core of the sentence (i.e., the verb) in Old Czech is indicated by the fact that the dative is retained here even when the infinitive has already been replaced by a prepositional construction, cf.: kúpichu pole na pohřeb pútnikóm, gloss: ”they bought a field for the burial of travelers” (Trávníček 1956: 170, 174). As in Lithuanian, a dative of purpose with a noun has also been used here since ancient times. 2.4.2. The corresponding constructions in Old Indian are particularly important for establishing the original state of infinitive constructions. In the Vedas, their structure fully coincides with that which can be reconstructed for Old Lithuanian by means of internal reconstruction. The infinitive there still has paradigmatic dative forms of verbal nouns (with -tavāi/ tave, -adhyai, -áye, -e), and the construction itself has clearly retained the character of a double dative. The first dative (of a noun or pronoun) can likewise denote either the subject of the action (a) or its object (b), cf.: a) átho indrāya pātave sunú sómam RV 1,28,6, gloss: ”thus, for Indra to drink, press the soma- 7 Plačiau apie slavų kalbų DI žr. Miklosich 1874: 857 tt., Потебня 1958: 389; Los 1922: 172—173; Gebauer 1929: 380, 579—580; Vondrák 1928: 265, 408; Trávníček 1956: 170 tt. 83 ma”; b) pība... vṛtrāya hāntave śaviṣṭha RV 4,13,3, gloss: ”drink... for Vrtra’s slaying (to slay), most powerful one”8. The relationship of this construction to the dative of individual nouns denoting purpose is similar to that in the Baltic and Slavic languages. Nevertheless, already in the Vedas the double dative (or DI) is of a relic nature (Sgall ib. 209). Owing to the regrouping of the syntactic relations between the constituents of the construction, constructions with an accusative governed by the same “infinitives” are used more often instead of the double dative (DI). 2.4.3. Hittite cuneiform texts also contain clear parallels to Baltic DI. Between the goal-denoting dative with an infinitive and the main verb, we find relations characteristic of the structure of the Baltic construction; cf.: nankān ANA ERĪNMĒŠ ŠA KUR (UGUTI) (ni)—ninku̯anzi u̯eriat Hatt. 11 IV 4, gloss: ”and he sent him to gather the country’s (upper) troops” (Ose 1944: 40, Hahn 1953: 112; 1966: 390). The Hittite infinitive in -u̯anzi (nininḳu̯anzi “to gather”), from the archaic stem u̯ant-, here has the same dative—“locative” ending -i as the Lithuanian infinitive in -ti. In Hittite, the dative often also denotes the purpose (goal) of an action when it occurs alone, without an infinitive. 2.4.4. Thus, the Baltic languages’ DI corresponds to the aforementioned constructions in related languages in its original structure as well as in its meaning and similar direction of subsequent development. The infinitive in this construction has retained the meaning of purpose, which the dative of nominal forms has had since the earliest times. Comparison with other languages confirms that the syntactic basis of DI is a double dative of purpose (goal), whose second constituent was a verbal noun. The appositional relation between the two members of this construction is an ancient type of syntactic relation, characteristic of the Indo-European languages in the early periods of their development. Double case constructions are also based on these relations. The double dative with verbal nouns—names of actions—was apparently used to express purpose or function long before infinitives arose, and performed in the sentence the role later taken over by infinitival constructions and subordinate purpose clauses. In this respect, the double dative with verbal nouns is close to double cases with verbal adjectives (later, participles). Absolute constructions and the accusative and genitive with a participle developed from such double cases. This confirms the important place of verbal nominal forms in the sentence structure of the Proto-Indo-European languages. 3. NOMINATIVE WITH INFINITIVE (NOMINATIVUS CUM INFINITIVO) the field remained nearby 3.1.1. The purpose-denoting infinitive has long been used with verbs of state- 8 Žr. Delbrück 1888: 413; Sgall 1958: 200 tt.; Haudry 1977: 106—122 su lit. F. Renou (1937: 28) manė esant galima ”infinityvą” hāntave ir pan. kildinti iš veiksmažodinio būvardžio datyvo, tačiau tam prieštarauja tiek kai kurių sen. indų (Haudry 1977: 122, 436), tiek baltų, slavų ir kitų kalbų infinityvų morfologinės ypatybės. 84 verbs, e.g.: to happen, to remain, to matter, to be necessary, to pay off, to come about, to occur and so on. In such cases, the thing whose state is described by the verb is expressed in the nominative. The subject experiencing the state may be either unexpressed (a) or expressed in the dative (b): a) Bepaliks tas sienelis sugrėbt Lpl; Darbas nelengvas teko dirbt Gg; Rūpėjo tas šienas daužyt Krp; Apsimoka linai sėt? Rd; Kitąsyk pasitaiko ir toks vyras gaut Mrj; b) Man kas belieka daryt? Ds; Man kliūn(a) visi darbai nudirbti Vvr; O čia jam rūpi pinigai skaityti BsP III 69; Nebepatiko ožkelei sodrį žolelė pešti Žem (LKŽ XIII 286). As early as 1893, Jonas Jablonskis astutely observed that in sentences of this type the nominative functions as the subject, while “the infinitive here only explains or supplements the predicate... For example, let us take the sentence ‘I already feel concerned about the oats’; if we explained the predicate ‘feel concerned’ with the infinitive ‘to reap’, we would then get the complete nominative-with-infinitive construction: ‘I am already concerned about reaping the oats’ (1935: 32). This can be said of examples in both group (a) and group (b): the nominative forms the core of these constructions, while the infinitive can be omitted without disrupting the grammatical structure: the little wall was left behind, the work fell to [someone], flax pays off, and so forth. The purposive meaning characteristic of the dative of verbal nouns and of the infinitive that developed from it is somewhat more faded here than in the DI constructions discussed, but it can readily be traced by comparing expressions with and without the infinitive. The supplementary role of the infinitive is even more pronounced in sentences where the dative marking the subject cannot be used at all, e.g.: Soon the rye will begin to be cut Mžk; The rye has already had time to be cut Žrm (LKŽ XIII 387). Thus, the construction model can be written as follows: (I) had to / got to / cared about / got tired of... plowing the stubble field Such constructions are attested from the very beginning of written records; cf.: it is fitting for one to return and to give honor, glory, and praise WP 33. 3.1.2. The relations between the constituents of the construction discussed above are now no longer felt in any way in all sentences containing the aforementioned verbs of state. Having long since lost its connection with the nominal paradigm, the infinitive began to attach itself to the finite verb form, forming a close construction, also used in sentences without a nominative, cf.: I had to drive; The child only cares about playing; Morta got tired of sitting, and so on. The object of the infinitive of transitive verbs came to be expressed by the accusative or genitive, e.g.: Atsitiko jam pro kapines praeinant duobkasius sutikt BsV 92; Bene rūpės darbą dirbti arba vargą vargti? Žem R I 173; Meškai apkrūvėjo klausyti lakštingalos Krž (LKŽ V 834). In the modern standard language, the restructured pattern of relations in constructions with a transitive infinitive has now become predominant: (I) happened to/ had to/ cared to/ got tired of... ploughing the stubble fields The restructuring of the internal relations within the construction in Lithuanian is not yet complete—it is an ongoing process. In Eastern and Southern Aukštaitian dialects, the nominative is widely used with the infinitive. However, it is often understood here as an object case dependent on the infinitive, e.g.: Sekas pagaut ešerys Mrk; Šitoks geras rūbas neapsimoka parduot Trgn; Teko man iškast senas šulnys Grnk. 85 Here the nominative can no longer be considered the subject of the sentence, cf.: *sekas ešėrus, *teko šulnys. medus gardau/ gardus valgyti 3.2.1. Similar to the aforementioned verbs of state, the infinitive is also associated with a predicate whose nominal part is an adjective. Such adjectives may have an inflected masculine or feminine form (a). Most often, however, the infinitive is associated with the neuter form of adjectives, which denotes the state of a person or thing (b). In both cases, the subject of that state (as in the sentences discussed in 3.1) may be expressed by the dative: a) Standing water is tasty to drink Jnš; A rabbit is unpleasant for me to slaughter Jnš; It is expensive to keep cockerels Akm; This basket is pleasant to carry Vdk; The ground was very hard to plough Als; b) Dry bread is not tasty for me to eat Ut; It is costly to attain knowledge Šr; It is important for me to get horses Ukm; It is pleasant to mow hay Ssk. Both types of constructions are attested in texts from the 16th century onward, cf.: a) And these troubles are very hard for a person to bearBrP II 105; b) and these pains are not easy to endureDP 214₂₃ (Palionis 1967: 153). The infinitive’s original function is particularly evident in constructions of type (a). In these, the infinitive, which complements a nominal predicate, has the purposive meaning characteristic of the dative (cf. 1.2.2). Clear equivalents of the construction with the same semantic structure are also found in related languages, cf.: Latv. saulīte silta sildīties BW 32 35 “the sun is warm for basking in” (Endzelīns 1951: 989); OCS eretici... blědnъ z sutu viděti Kozma 3 “heretics... are pale to look at” (Дьячкова 1982: 133); O. Czech chléb tvrd jesti “bread is hard to eat” (Gebauer 1929: 358); O. Russian voda Jordana sladkъ pitъ “the water of the Jordan is sweet to drink” (Спринчак 1960: 179)⁹. Just as examples of type (a) correspond to sentences without an infinitive, in which the subject agrees with the nominal part of the predicate, so examples of type (b) correspond to sentences whose predicate consists of uninflected neuter forms (pienas skanus gerti: pienas skanus = “milk is tasty to drink: milk is tasty”; pienas skanu gerti: pienas skanu “milk is tasty to drink: milk is tasty”). Sentences such as Man pienas/ arbata skanu Kd; Svetimas stogas nemiela Ut are relics of ancient nominal sentences; they were inherited from distant stages in the development of Proto-Indo-European, when pure nominal stems associated with the subject served as nominal predicates (for more detail, author 1979: 200 — 202). Lithuanian neuter adjectives ending in -a < *-o, -u: miela, skanu (author 1990: 201 — 202 with lit.) have also preserved the pure stem. The infinitive of dative origin became attached to the nominal predicate with these neuter forms just as it did to predicates with agreeing masculine and feminine forms; this is clearly shown by examples containing the same lexemes, cf.: Nespartūs miežiai pjaut Mlt/ Runkeliai nespartu kast Mrj; Medus valgyt gardus/ gardu Mlt. It is worth noting that type (b) constructions with an infinitive are most widely used in the dialects of eastern and southern Lithuania, where the corresponding nominal sentences with neuter forms have been well preserved. 9 Panašiai ir kitose kalbų grupėse: s.indų (Sgall 1958: 215), germanų (Behaghel 1924: 342), graikų (La Roche 1893: 65), tochary (Krause — Thomas 1960: 79). 86 Taking into account the origin of the infinitive in the dative of a verbal noun (an action noun), we can schematically represent the original structure of the nominative-plus-infinitive constructions discussed in 3.1 and 3.2.1 as follows: E.g.: honey is pleasant/tasty/delicious (for me) to eat (<“for eating”) This model of syntactic relations fully corresponds to the one reconstructed on the basis of dative-plus-infinitive constructions (cf. 2.1). 3.2.2. With the disappearance of the neuter gender of nouns in Lithuanian, neuter adjective forms became invariable predicates generalized in the sense of a state. The infinitive, in turn, having become detached from the declensional paradigm, began to attach closely to these neuter forms, complementing their meaning and forming with them close state-denoting phrases: pleasant to eat, difficult to work, good to live, sad to travel, and so on. Their establishment was facilitated by the fact that neuter forms corresponding to bare stems had been used in the sense of a state in Indo-European languages since the earliest times. Once phrases of neuter forms with an infinitive had become established and widespread with an impersonal meaning, (b)-type constructions with a nominative began to be understood as impersonal, and the nominative form as dependent on the infinitive. Thus, sentences such as Guest pleasant/to see (“for seeing”) were reinterpreted as Pleasant to see/guest (the guest) through a reorganization (reintegration) of syntactic relations. The nominative came to be regarded as dependent on a transitive infinitive and was replaced by the usual case of a direct object—the accusative. Thus, the original infinitive-construction model identified in 3.2.1 (abbreviated Inf) was reorganized as follows: E.g.: (for me) it is pleasant to eat honey/honey honey needs to be eaten 3.3. Some neuter adjectives, functioning as nominal parts of predicates, themselves came closer to verbs and provided the basis for their inflected forms. For example, the old neuter adjective forms gaila, verta became members of the conjugational paradigms, with present-tense meaning, of the impersonal verbs galėti, vertėti derived from them (cf. skauda: skaudėti). When they are associated with a transitive infinitive, the nominative denotes its object, e.g.: it is a pity to dig up those young potatoes Rdv.; Or 10The abbreviation Pred here denotes a verbal or nominal predicate. Other abbreviations are as in 2.1. 87 worthwhile to keep a horse? Kp (because the meaning of these forms is clearly impersonal, sentences *bulbės (nom.) gaila, *arklys verta, ‘cabbages are regrettable, a horse is worthwhile’, apparently are not used anywhere). Similarly, from the noun reiką/ reikė used as the nominal part of the predicate, the impersonal verb reikėti, -ia, ėjo ‘need’ developed (Fraenkel 1925: 37–38). The relatively recent origin of this verb is indicated by the corresponding nominal constructions used in old texts and dialects, reikā yra/ buvo ‘there is/was a need’ and the like (cf. LKŽ XI 375; Fraenkel LEW 714 with lit.); they must have been associated with the infinitive much as reikalas yra/ buvo...., ‘there is/was a matter....’, cf.: Argi jam reikalas buvo tę arkliai brukti? ‘Was it really necessary for him to drive those horses?’ Lp (LKŽ XI 377). According to Jablonskis (1935: 32), sentences such as these show the old use of the nominative: Žinau, kas žagrei reik; žinomas daiktas, kas reik, tas reik. ‘I know what the plough needs; it is a known thing: what is needed is needed.’ Thus, in ancient times there were apparently also expressions such as *reikā yra/ Ø /buvo/ bus karvė pirkti (cf. metas karvė pirkti) ‘there is/was/will be a need to buy a cow’ (cf. ‘it is time to buy a cow’), alongside expressions with inflected adjectives, e.g.: man reikalingas (t.y. reikia) vaikas paimt ‘I need to take the child’ Žž. The verb reikėti, developed from reikā, is now used very widely with the infinitive. In such cases, the nominative is understood as the object of the infinitival action, while the experiencer of the state may be expressed in the dative: Reikia uždegt žiburys ‘A lamp needs to be lit’ Skp; Ir spiečius reikia mokėt pagaut ‘And one must know how to catch a swarm’ Ktk; Anam reikš nešti vanduo ‘He will have to carry water’ Mžk; Vaikas reikia ažmig(d)yt ‘The child needs to be put to sleep’ Švnč; Reikia muš(t) ji per pilvą ‘One should hit it on the belly’ Rdš; Kaip ana reikia pasiekti? Dgl; Reikia viskas sakyt ‘How is one to reach her? Everything must be said’ Nmn. early/ time to harvest rye 3.4. As with neuter adjectives, the infinitive is also combined with adverbs expressing a state (a) or with nominative forms of nouns (b), which developed from fossilized nominal sentences and have a meaning close to that of adverbs: a) Tenai gerai vaikai augint ‘It is good to raise children there’ Kpr; Taip smagiai tie medžiai skaldyt ‘It is so pleasant to split those trees’ Pln; Da(r) anksti rugiai pjaut ‘It is still early to harvest rye’ Mlt; Nebebuvo kada ir pūdymas art ‘There was no longer time even to plough the fallow’ Dgl; Mun ir pienas toli nešt ‘It is a long way for me to carry the milk’ Ob; Gan jam liet kraujas ‘He has shed enough blood’ JnšM; b) Jau laikas karvė melžt ‘It is already time to milk the cow’ Dg; Metas poteriai kalbėt ‘It is time to say the prayers’ Dbk; Ten būtų juokai sakyt ‘It would be a joke to say that’ Jnš; Karvės laidyt vargas vasaros laike ‘Turning cows out to pasture is a hardship in summer’ Šr. Corresponding constructions are also formed with noun phrases involving nouns darbas, dalykas, daiktas ‘work, matter, thing’ and the like, combined with adjectives (pronouns): Grybai rinkt smulkus darbs ‘Gathering mushrooms is a small task’ Brž; Degutas degt tai paprastas dalykas ‘Burning tar is a simple matter’ Sv; Karvė išlaikyt yr nebrangus daikts ‘Keeping a cow is not an expensive undertaking’ Rs; Linai rauti nebokia naujiena ‘Pulling flax is nothing new’ Grg; Suverpt per žiemą keturi pūdai tai nieko nėr ‘Spinning four poods over the winter is nothing’ Mlt. The nominative forms of concrete nouns in such cases also have an object meaning and are usually understood as directly dependent on the infinitive. Two infinitival phrases with an object in the nominative may also be juxtaposed in one sentence, for example: Namai statyt tai ne pyragas krimst ‘Building a house is no piece of cake’ Šk. Establishment of the Accusative Form 3.5. According to data collected for the Atlas of Lithuanian Dialects, the infinitival constructions of the types discussed, with a nominative, have been well preserved by the eastern and southern Aukštaitians and the western Aukštaitians south of the Jonava–Čekiškė–Šakiai–Sintautai line. 88 In other dialects, the accusative is used instead of the nominative, and it also predominates in the standard language. The boundary between the use of the nominative and the accusative is not sharp; variation and transitional cases occur on both sides of it. The nominative is most prone to replacement by the accusative in constructions with the verb reikėti ‘need’ (3.3.) and adverbs (3.4.), in which the nominative cannot be understood as the subject of the sentence. In the Samogitian dialect, the nominative has survived here in a few “islands” surrounded by the accusative: around Židikai, Viekšniai, Luokė, Gargždai, Papilė, and Raseiniai; but even here the accusative is usually used alongside it (especially in constructions with reikėti).¹¹ Such sporadic dispersion is characteristic of syntactic archaisms: as newer constructions spread, older ones are not displaced all at once and usually survive in certain pockets. Constructions with the nominative have also been well preserved in Lithuanian-language enclaves surrounded by Belarusian (in Gervėčiai, Ramaškoniai, Rodūnė, Pelesa, Varanavas, Lazūnai, Zietela, etc.). Thus, the establishment of the accusative in the dialects is still an ongoing process. From the rather sparse examples in sixteenth-century texts, it can be seen that the western Aukštaitians of Lithuania Minor also used the nominative with the infinitive. The survival of the object nominative in infinitival constructions was favored by the fact that the nominative with this meaning had long been used in constructions with neuter forms of passive participles, for example: potatoes are being dug up/dug up, the rope is being twisted, the door is opened and left open; often also with a genitive agent: his rye is being sown/has been sown, forests of old were loved and the like.¹² It is noteworthy that the nominative with an object meaning is used with the infinitive and with neuter forms of passive participles in the same eastern and southern Aukštaitian dialects.¹³ Constructions with neuter forms of passive participles, too, are now often understood as impersonal, and therefore a nominative that served as the subject is sometimes replaced by an accusative functioning as an object; compare: With a hook, honey is climbed down for from the tree hollows Lp (aut. 1979: 206—207). In the same dialects, the object nominative is also used with present and past adverbial participles; compare: When work is being done, the winds are no concern Mžk; When pulling flax, one needs to dress warmly Brž; Where can a horse be put up here now Ut; If only a sweater were knitted Mžš; It would be good to have a strong worker hired Kp. The object nominative may have spread here by analogy with the aforementioned constructions involving passive participles and the infinitive (adverbial participles, like the infinitive, developed from dative forms of the -i stem; cf. aut. 1962: 17 ff.; 1979: 133—134). Constructions with the verb būti 3.6. A separate group consists of nominative–infinitive constructions with the verb būti ‘be’, whose zero form denotes the present tense. They are of two ty- 11 Remiamasi atsakymais į ”Lietuvių kalbos atlaso medžiagos rinkimo programos” (V., 1956) klausimus Nr. 115 b ir c. 12 These constructions are participial variants of such nominal sentences as honey is tasty; for more on them, see aut. 1979: 200—204; 1990: 200—214. 13 For the distribution area of constructions with neuter forms of passive participles, see aut. 1977: 36—38. 89 pù. In constructions of the first type, the infinitive of verbs of sensation or perception (most often — matyti, regėti, girdėti ‘see, behold, hear’) is linked to a nominative marking a possible object of sensation or perception, e.g.: (man) yra / Ø / buvo / bus namai matyti ‘for me, the house is / Ø / was / will be visible’. In constructions of the second type, the infinitive of various verbs is linked to a nominative marking the object of the action that needs to be performed, e.g.: (man) yra / Ø / buvo / bus namai statyti ‘for me, the house is / Ø / was / will be to build’. In both cases, the infinitive has a passive meaning. (man) yra / Ø / buvo / bus namai matyti ‘for me, the house is / Ø / was / will be visible’ 3.6.1. This construction is now used mainly in the dialects of Central and Western Lithuania, but fixed expressions with the pronoun kas (kas girdėti ‘what is heard’; kas nauja girdėti ‘what is new to hear’, etc.) occur in many other areas as well. This indicates that constructions of this type were more widespread in the past. The present-tense form of the verb būti ‘to be’ is sometimes expressed in the Samogitian dialect; elsewhere it is most often zero. Jau dugnas matyt Jnv; Per langą gerai matyt ežers Pln; Jau regėt ugnis Lzd; Varpai girdėti Prk; Užsnīgę griovai ir keleliā, net vieškelis menkai pažint Simon (LKŽ XIII 258). Antai baltoji obelìs y(r) matyti Skd; Jau trobos matyti yra Pgg; Tas akmuo buvo iš tolo matyt Klp; Vakar griaustinis buvo girdėtis Erž; Už to miško ir Papilė bus matyti Pp. Only occasionally is the experiencer marked by a dative, e.g.: Ta troba man buvo iš tolo matyt Klp. In written language, the construction is attested from the sixteenth century, e.g.: du daiktu mumis neperweszlibu regėtessi WP 72. In this construction the infinitive plays a role similar to that of the present passive participle, and is sometimes used alongside it even in the same dialect, e.g.: Jau vadakstis matyties // Nuo Šimkaus kapelių karvidės matomos Pkl; cf. also: Sodnas didelis y(r) matoms Skd. In the Eastern Aukštaitian dialect, a reflexive form of the verb often occurs instead of the infinitive, e.g.: Visas sklypelis matos kaip in delno Skdt. That the nominative has the function of a subject here is evident from agreement with a participial construction, e.g.: Petruko balsas buvęs girdėti Rs (for more detail, see Sirtautas 1971). With a negative infinitive, the genitive occurs instead of the nominative: žmogaus nematyti Plng; laumių neb’girdėtis BsP II 70. Unlike the cases discussed in 3.1–3.5, however, the nominative here is never replaced by the accusative. The construction is generally used to describe what is available to be perceived about a particular thing, cf.: jau bokštas matyti — jau galima matuti bokštą. ‘the tower is already visible — one can already see the tower’. Thus, here too the infinitive implies a purpose meaning, which is inherently characteristic of the dative. 3.6.2. The infinitive has the same meaning in corresponding Latvian constructions, most often used with negation, cf.: No tās reizes vilku ne redzēt, ne dzirdēt LP VII 878 = nebija vilku ne redzēšanai, ne dzirdēšanai (Mūlenbachs 1928: 195). “From that time on, wolves have been neither seen nor heard (= There were no wolves either to be seen or to be heard).” Constructions of exactly the same structure and meaning as those in Lithuanian are attested in old Slavic texts, cf. OCS: i glasa umlčě. i ne by slyšati Supr 570,18 “and the voi 90 sas nutilo, ir nebuvo girdėti” (Ходова 1980: 223)14, s.rus.: bė v to vremja vidėti... pečalb gorbkaja Mosk. Letop. 290 ”buvo tuo metu matyti sielvartas kartus” (Сабенина 1978: 404); jų yra išlikusių gudų ir pietinėse rusų kalbos tarmėse, pvz. derevnja vidatь ”kaimas matyti (Спринчак 1960: 179).15 Infinitivo pirminę paskirties reikšmę tyrinėtojai įžiūri ir atitinkamose Rigvedos konstrukcijose, pvz.: táva spārhé várna á samídṣi śríyah RV 2,1,126 ”tavo pageidaujamoje spalvoje kaip tik matyti grožis” (Sgall 1958: 221; Disterheft 1980: 46). (for me) there are / Ø / were / will be houses to build 3.7.1. In constructions of this type, the lexical expression of the transitive infinitive is not specified, while the verb to be has a modal nuance of necessity. Two groups of constructions can be distinguished. a) The infinitive is associated only with the nominative and a possible (optional) temporal specification. The present-tense form of the verb to be is usually zero-marked, while the past or future form is expressed. Today the hay is to be raked, and tomorrow the rye is to be cut Dglš; It is time to cook dinner again Trg; Tomorrow the horses are to be grazed Pšl; Everything has been threshed with the cows PnmP. So the family had to be hired Žd; And the rye had to be cut Brš; At first such buttons had to be picked Šil; The work had to be done like that Glv. The hay will have to be cut, the fallow ploughed Slnt; Then the hay will have to be raked Kl; When the potatoes are planted, the flax will have to be weeded Rdn; Then, one of these days, the manure will have to be hauled and the pens cleaned Mžk. b) In addition to the nominative with the infinitive and a possible (optional) temporal specification, the dative case also expresses the subject of the infinitive’s action: Tomorrow we have to graze the calves Pnd; Mother has to raise the children Vvr; For you, this has to be mended, the horse fed, and then the rye stacked—it is no joke Jnš; I alone had all the work to do Žd; Why did he have to torment these women? Šrv. Tomorrow you will have to weed the sugar beets Rk; I will have everything put in order Kltn; Will both my parents have to card the wool? Sk. Both groups of constructions are attested from the beginning of written records, for example: a) So too is wine to be drunk, for it is holy Mažv 34,17; b) They will take the strict judgment to God; accounts will be given to the Lord God WP 85 (MLLG V 238). Data collected for the Atlas of the Lithuanian Language show that these constructions are also used in those dialects where the verb to need is used, often also with the impersonal verb- 14 However, in the same syntagma, the object of seeing or hearing is here expressed by an accusative dependent on the infinitive (as also in the vellum manuscripts of most Slavic languages), cf.: bystъ viděti junosǫ krasna stoješsta předъ nímъ Supr 187, 1—4 “a handsome young man was to be seen standing before him” (Ходова 1980: 223). 15 Some consider the corresponding Czech construction, for example: hora je viděti “the mountain is visible” (for a fuller discussion, see Porák 1967: 82 ff.), which V. Kiparsky (1969: 146) recognizes as an ancient inheritance, to be a borrowing from German (Reiter 1953: 175; Timberlake 1974: 124 ff.) or a later innovation (Havránek 1968: 175—176; Dunn 1982: 523—524). However, even in German the infinitive with zu, in the cases under discussion associated with the nominative, has retained its ancient purposive meaning (cf. Brugmann 1916: 924); thus, apparently, the construction has the same original source. 91 … agree with nouns in gender as adjectives; the accusative is already used (Trg, Žgč, Škn, Kltn, Pln, Plng, Vvr, Dr, Drb, Lnk, Kl, Krp, Kv, Skdv, Nv, etc.). They differ from the type discussed in 3.6 only in the meaning of necessity, which is probably conditioned not by the verb to be itself, but by its connection with the infinitive of dative origin expressing purpose (“the flax is/was/will be intended for breaking”)¹⁶. This purposive meaning characteristic of the dative can also be traced in cases where no nuance of necessity is felt at all, e.g.: You have only crows to chase Pg; The young have only beer to drink Grz, and the like (see also LKG III 616–617). All this shows that the nominative here, too, had the function of a subject. Thus, the same original model of syntactic relations can be reconstructed for both types of constructions discussed in 3.6 and 3.7 (the verb to be is denoted here by the abbreviation Vₛᵤᵦₛₜ). This model coincides with the one established on the basis of constructions with the verbs remain, suit, be due and the like, as well as with neuter forms of adjectives (see 3.2.1), and differs only in the expression of the predicate. 3.7.2. Constructions of this structure are also attested in Latvian dialects, cf.: kaņgam ēst tei meizīte (Nereta) “for the master to eat that little loaf” (Endzelīns 1951: 554). The nominative is also used here with the relatively recent debitive forms, whose basis was provided by the infinitive, cf.: man jāēd sausa maize “I have to eat dry bread” (ibid 972 – 973). Most often, however, the nominative here has been replaced by the accusative dependent on the infinitive, e.g.: man bija tā šo lietu iesākt “I had to begin this matter in this way” (Mülenbachs 1928: 173 – 174); Tev bij ņemt niedres meitu BW 15142 “you had to take the reed-maiden”; Kā būs man purvu brist 15307 “how shall I wade through the mud?” (Ozols 1961: 96–97; 1967: 182). The accusative had already become widespread in old Latvian written sources, e.g.: Man bij grahmatu rakstīht Stender Gr §165 “I had to write a book”; No ta buhs tew to sinnahr G1B NT Praef. “From that you will know it.” Old Russian expressions with the nominative have been widely and repeatedly studied, cf.: Такова правда оузети Русину ou Rizě Smol. gr. 1229 “such is the right of a Rusin to take [it] in Riga” (i.e., a Rusin ought to have such a right in Riga; see already Потебня 1958: 405: 407); more recently, J. Stepanov (1984: 132–136) has convincingly demonstrated their historical connection with Lithuanian and Latvian constructions. The corresponding constructions in Old Indic are important: there, the modality of a verb meaning “to be” is usually further emphasized by an optative or imperative form, the nominative clearly functions as subject, and the infinitive expresses purpose in the dative, e.g.: ápo bhavantu pitáye RV 10,9,4 16 Plg. Brugmann 1916: 925. Pasak R.Mrazeko (1963: 113 — 114), datyvo paskirties reikšmė buvusi perkelta į veiksmažodį ”būti” ir sen. bažnytinės slavų kalbos konstrukcijose: мне бѣаше страдати ”мне выпало, принадлежало страдание → страдати”. Kiek kitaip aiškina J.Stepanovas (1984: 133) laikydamas reikiamybės reikšmę viena seniausių, iš indoeuropiečių kalbų bendrijos paveldėtų veiksmažodžio ”būti” reikšmių. 92 “May the waters be for drinking (for consumption)” (Sgall 1958: 221); similarly in Hittite (Disterheft 1980: 165–166). It is worth noting that in other related languages as well (Latin, Greek, Germanic), in earlier periods of their development, constructions with forms of nominal origin were used with modal meanings such as possibility and necessity, and the like; these were later replaced by newer modal verbs. In the Slavic languages, modal verbs and adverbs (predicatives) expressing these meanings are also relatively recent, becoming established chiefly during the period of written tradition. By contrast, the sentences discussed here, with an infinitive of dative origin, represent a very archaic means of expressing modality, inherited from the Indo-European linguistic community. 3.7.3. The subsequent development of the construction apparently proceeded in the same direction as in the other cases already discussed. The infinitive, having lost its connection with the inflectional paradigm and become an invariable verb form, became the grammatical center of the sentence—the predicate, with a modal meaning of possibility or necessity (based on the purposive meaning of the dative). The construction was reinterpreted as impersonal, and the nominative became linked to the infinitive as the object of the action it denotes. This gave rise to the following model of syntactic relations in impersonal infinitive sentences: Since the nominative is not a case characteristic of an object, it is entirely understandable that the accusative came to be used in its place: Tomorrow I have hay to rake, you have horses to pasture, while with an infinitive of a verb that governs the genitive, the genitive is used: What more could we need, and so on. The accusative in place of the nominative is attested in written sources from the 16th century, e.g.: Then it was for you to give my money to the money-changers BrB Mat 25,27; all Christians to the end of the world will have troubles to see BrP II 78,2 (see Palionis 1972: 126). The development of infinitive constructions expressing necessity and possibility with the verb “to be” into impersonal constructions, and the replacement of the nominative by an accusative governed by the infinitive, have long been noted not only in Latvian (3.7.2), but also in other related languages: Slavic, Germanic, Greek, and Latin (Brugmann 1916: 925–926; Потебня 1958: 391–398). Nevertheless, in Lithuanian the nominative has been preserved here much better than in constructions with reikėti and other verbs—not only in dialects (see 3.7.1), but also in the standard language (LKG III 616–617). The nominative has been especially well preserved in the expressions under discussion with the pronoun kas, attested since the 16th century (cf. kas buvo daryti BrB II Moz 36,7) and still used today in various parts of the Lithuanian dialect area, even in those where in other cases we find only the accusative, e.g.: What can we say now Lkm; There is nothing to put on the horse Lbv; There is something to sell Psv; Will there be anyone to pull it? Šil; What is there to do? Ss; There is nothing to do Pg; There is nothing to say Vlkv, Ss. 93 Such sporadic distribution also indicates the relic nature of these constructions. (for me) there is / Ø / was / will be to go 3.8.1. Infinitives of intransitive verbs with the verb būti also form impersonal constructions expressing necessity and other modal meanings. Nowadays they are most often used with a dative marking the subject of the infinitival action and with the zero form of the verb būti: For you, plague, to rot! BsP I 264; The living to live, the dead to rest! KrėvPP III 14; Jonas to go to the forest today Gr; For me, neither to shout nor to swim Skr (see also LKG III 36—37; 616—617). In dialects and folklore, constructions with an expressed (especially past- and future-tense) copula also occur. They are used both with a dative subject (a) and without one (b): a) Who is there to feed, if not to go Pvn; Perhaps it is God’s will for the cuckoo to call in the grove NS 270; My dear daughter, my young one, you were to go to the ploughman BsO 215; If he is to recover, he will recover, but if he is to die, he will die Tj (LKŽ, I 1216); b) And there was much to remember here Ktč; The little brothers had arrived yesterday evening LTRn 90; When I dismounted from my horse, there was no one to call me; when I went into the parlor, there was no one to seat me JV 683. Such constructions are attested from the 16th century, especially in Bretkūnas’s writings, e.g.: a) For he had to remain in his own city of Walnistes BrB IV Moz 35,28; then he will have to die I Moz 44, 31; We would have perished if God had not helped BrP II 53. b) Would it really not be necessary to fight for the sake of such women? BrB Mal 10,2 (for more detail, see Palionis 1972). 3.8.2. Corresponding constructions with a zero or expressed form of the verb būti are also used in Latvian, e.g.: bāliņam karā iet Spr. 271 “for the little brother to go to war”; brist bij man, vaj nebrist BW 9416 “whether I was to wade or not” (Ozols 1961: 96 — 97); Tev nebūs zagt! ME I 359 (see also Endzelīns 1951: 994). They are attested from the earliest written records; cf.: Woj tu sinnaji / kurzą Laikā tew bij peedzimt G1B, Hiob 38, 21 “Did you know at what time you were to be born”; kapehz buhs man... eet Ps 42,10 “why I will have to go.”17 In Slavic languages, finite forms of “to be” with the infinitive form impersonal constructions from the earliest written records; cf. Old Church Slavonic byti oumrěti ništumu Ostr Luk 16, 22 “the poor man was to die”; Old Russian mi budet nyňě poěati Ip. 130 “I will have to go now” (Потебня 1958: 392—393); Old Czech tu še nám (jest) s císařem sníti Dal 81 “here we (are) to sleep with the emperor” (Pořak 1967: 21). In other Indo-European languages, corresponding constructions are likewise attested with modal meanings of necessity, possibility, etc., e.g.: Greek ἀλλ’ οὐκ ἔστι Διὶ Κρονίων μάχεσθαι φ 193 “but it is not possible (= 17 This is the so-called “analytic debitive” (older than the morphologized form with jā-); examples from old Latvian grammars can be found in Ozols 1965: 92, 202, 252. 94 “to fight with Zeus Kronion”; Latin: per nares floribus frui est Tertull. Coron. 40 “to enjoy flowers through the nostrils (= by smelling them) is (= necessary)”; Old High German za petōnne ist Hymn. 17,1 “to pray is (= necessary)” (Brugmann 1916: 925–926). 3.8.3. The data cited here and in 3.7.1 confirm the view expressed earlier that infinitival constructions with an expressed form of the verb būti in Lithuanian and Latvian are a shared ancient inheritance (Palionis 1972: 128). There is no basis for deriving them from constructions without a copula. Although an increase in the productivity of the copula at certain stages of development has been observed in some Slavic languages (e.g., Czech; see Porák 1967: 18, 39 ff., 122), constructions without a copula are also attested in East Baltic and Slavic languages from the beginning of written records. In earlier times, they apparently functioned as variants of the same constructions with a zero present-tense form; later (as constructions with būti in the past and future tenses were disappearing), they were generalized as a separate type of infinitival sentence (Lith. Tau važiuot!; Latv. Kurp mums bēgt? Russ.: вам платить “you have to pay,” etc.). 3.8.4. Constructions with the infinitive of intransitive verbs are just as impersonal (i.e., lacking a syntactic subject position) as those with the infinitive of transitive verbs governing the accusative, cf.: (tomorrow) I have to mow hay and (tomorrow) I have to cut barley. The latter constructions probably began to be formed only after the infinitive had become detached from the nominal declension paradigm (cf. 3.7.3). In many cases, the opposition between transitive and intransitive verbs is neither strict nor primary in any case. Thus, from a historical perspective, the constructions I have to mow hay (there is, there was...) and I have to cut barley (there is, there was...) can be treated alike and assigned to the period in the development when, owing to a regrouping of syntactic relations, sentences with būti and an infinitive came to be interpreted as impersonal (cf. Brugmann 1916: 925–926). The circumstance that the original case meaning of the form that became an infinitive is difficult to explain and the structure of the construction difficult to motivate semantically makes it problematic to regard the construction I have to mow hay as an original, paradigmatically autonomous nominal sentence (as in Večerka 1973: 303; Степанов 1984: 132–133). Syntactically impersonal constructions with an infinitive could apparently have arisen only after the infinitive itself had become a distinct verbal form. 3.8.5. Thus, the original pattern of infinitival sentences with an expressed or zero form of the verb būti probably had a nominative subject: the houses are (Ø / were / will be) visible/to be built. The structure of this construction is motivated by the same dative meaning of the verbal noun that is reconstructed for the constructions discussed earlier with the verbs likti, tikti and the like, as well as with neuter forms of adjectives (as in cases 1–2 discussed above). In many cases, forms of a governed case dependent on the infinitive (most often the accusative) later became established in place of the nominative subject. However, in Baltic and some Slavic languages the nominative was partly retained in constructions used impersonally, where it acquired an object meaning. The survival of the so-called “object nominative” here was apparently helped not only by internal factors, but also by certain external conditions, which merit separate discussion. 95 The problem of the “object nominative” 3.9.1. The use of the object nominative in impersonal sentences is attested across a fairly large, contiguous area of dialects in contact with one another. Besides Lithuanian, Latvian, and the northern dialects of Russian, as well as a large triangle roughly between Pskov, Smolensk, and Vologda, this area also includes the West Finnic languages (Estonian and Finnish). In the latter, the so-called accusative II, which coincides with the nominative, is regularly used instead of the usual accusative I to denote the object in impersonal sentences with infinitives and converbs, as well as in impersonal imperative sentences (cf. Finnish hänen tuli rakentaa sauna “he had to build a sauna”). V. Kiparsky (1960; 1969) discussed the affinity of these constructions with the East Baltic and Russian “object nominative”; he suggested that contact with West Finnic languages may have helped preserve the inherited nominative-plus-infinitive constructions in neighboring East Baltic and Russian languages.18 B. Larin explained the relationship between these constructions as follows: the “object nominative” is supposedly a remnant of a non-accusative sentence structure, inherited from a long-extinct “littoral” language—the common substrate of East Baltic, North Slavic, and West Finnic languages (Ларин 1963: for criticism of this view, see Jacobsson 1964: 75; Veenker 1967: 126; Kiparsky 1969: 145). The hypothesis that the construction was directly borrowed from Finnic languages was set out in a special study by A. Timberlake (1974). In his view, Lithuanian and Latvian, as well as the northern Russian dialects, borrowed from West Finnic languages a syntactic rule according to which the object of the action denoted by impersonal sentences is expressed not by the accusative but by the nominative. Following the Finnic model, this expression of the object was realized in sentences lacking a subject position, and thus there was no conflict between the nominative form characteristic of a subject and the object function (Timberlake 1974: 2–3, 152). Timberlake considers constructions in Old Indian, Germanic, and other languages in which the object of an infinitive is expressed by the nominative in personal (i.e., sentences with a subject position) sentences to be an entirely different type, incomparable with the “object nominative” in East Baltic and northern Russian. He likewise draws a sharp distinction from cases of the houses are (were) visible and their Indo-European equivalents. The author failed to notice how closely the latter resemble constructions of the type the houses are (were) to be built (cf. 3.7.1–3.7.3), as well as the parallel use in Lithuanian and Latvian of sentences with agreeing and non-agreeing adjective forms (cf. honey is tasty / it is tasty to eat 3.2.1 ff.), and their paradigmatic relationship to personal sentences (i.e., with a subject position) of the type honey is tasty / it is tasty. The borrowing hypothesis rests on the conviction that the object nominative in Lithuanian and Latvian is synchronically motivated and regular in impersonal sentences (133 ff., 154). In fact, as already noted, it is being displaced here by synchronically motivated accusative counterparts, and its distribution across Lithuanian dialects has features characteristic of syntactic relics. 3.9.2. In Latvian dialects in direct contact with the Estonian area, the nominative with an infinitive is used much less frequently than in northern, eastern, and southern Lithuanian dialects. According to Latvian dialectological 18 Some researchers consider the West Finnic construction itself to have developed from a subject nominative with a purposive infinitive; see Hakulinen 1960: 246. 96 the data collected for the atlas19, at the Latvian–Estonian dialect boundary, the nominative of the object with an infinitive is attested in only a few places: in the very northernmost corner—in the Ydaus and Ypiki dialects and in Cirgali. From the northern part of the inland dialect akmenį grūti kustinat “the stone is difficult (i.e. hard) to move” type of constructions have also been recorded in the vicinity of Birzuļi, Bauni, Jaunburtnieki, Kuscen, Katraci and Dunte, in dialects separated from one another in which the accusative predominates20. Constructions of the type mentioned are somewhat more frequent in the western part of Kurzeme (in the Pope, Venta, Užava, Talsi-Èdole, Saldus and Dobele dialects); sporadically—as in the Samogitian dialect of Lithuanian—they also occur elsewhere in Kurzeme (in the Laidze, Lībagi, Kandava, Engure, Irlava and Ranki dialects); moreover, utterances of the vajag / gribas malka skaldit “one needs/wants to chop wood” type have also been recorded here. In the Semigallian dialects of the inland dialect, constructions of both types are likewise attested only in isolated pockets (in Ciecere, Jēkabnieki and Jaunsvirlauka); most are found in places where the Latvian-speaking area borders northern Lithuanian dialects that use the corresponding construction (south of Rēzekne, as well as in the Ezere, Bēne, Yslyce and Kaplava dialects). Over much broader stretches of the Semigallian inland dialect (around Dobele, Jelgava, Bauska, etc.), the nominative is used with debitive forms (jāiet pļaut siens “one must go to mow hay,” and so on). On the other hand, across a large area of the Upper Latvian dialect, the nominative with an infinitive is attested in only a few “islands”: in the Graši-Kraukļi area, around Sērene, in Dignāja and Vārkava, in the vicinity of Kaunata-Pilda, and in the Kraslava and Aknīste dialects along the Lithuanian–Latvian border (Ančītis 1977: 264—265); incidentally, here too the nominative occurs more often with the debitive than with an infinitive lacking jā-21 The scattered distribution of pockets where the construction occurs suggests that it was formerly widespread across a large part of the Latvian-speaking area, but has now declined much more extensively than in Lithuanian. This also supports the view that the nominative in this construction is a shared syntactic archaism. The Latvian nominative with an infinitive differs from Lithuanian in several innovative features as well. Except for rare cases with inflected adjectives (of the sun warm warmed type, see 3.2.1.), infinitive constructions here are impersonal, and the nominative has the function of a direct object. Since the neuter gender of adjectives has disappeared in Latvian and in infinitive constructions has been replaced by adverbs (easily “easily,” with difficulty “with difficulty,” need “it is necessary,” free “it is possible”, etc.; cf. not easy to carry a basket “not easily (=not easy) to carry a basket”), the connection with personal sentences, still perceptible in Lithuanian (work is difficult, etc.), has been severed here. Thus, the nominative form in Latvian has already lost its syntactic motivation; the nominative has become a genuine object case. This was apparently the main reason why, to a greater extent, it was replaced in Latvian by the usual object case. 19 They are kept in Riga, in the Department of Dialectology of the Latvian Language Institute. The author is grateful to the Institute’s management for the opportunity to use them. 20 Here and below, the discussion draws on responses to questions 270, 272, 276, 277 and 280 in the data-collection questionnaire for the Latvian Language Dialect Atlas (Latviešu valodas dialektoloģijas atlanta vāksanas programma. Rīga, 1954, lpp. 90—93). 21 Latvių kalbos tarmių pavadinimai pateikiami pagal M.Rudzīte 1964: 408—409. 97 — the accusative. It has been better preserved only in constructions with the new debitive form, which occupies a distinctive place in the Latvian language system. 3.9.3. The hypothesis already mentioned by P. Kiparsky—that the conserving influence of the West Finnic languages best explains why the nominative with an infinitive, long lost in other related languages, has survived to this day in East Baltic and neighboring Russian dialects—offers the best explanation. Contact with the West Finnic languages apparently helped preserve this syntactic archaism and provided a model for assigning it a new function in the same way—as the nominative of the object in an impersonal sentence. The conditions for this were especially favorable after the sixth century, when East Baltic tribes crossed the Daugava and penetrated far into lands inhabited by West Finnic peoples. Over several centuries, under the influence of the Finnic substrate, the dialects of present-day Latvia diverged from Lithuanian. Since the infinitive in -ti/-tei had apparently already separated from the nominal inflectional paradigm by then, changes in syntactic relations within the construction may have begun, giving rise to the beginnings of the “nominative of the object.” It was precisely in the seventh and eighth centuries that East Slavic dialects became established across a large territory between Kyiv and Novgorod. Intensive Finnic and Baltic Slavization took place in the area of today’s northern Russian dialects. Thus, contact with the West Finnic languages may have affected the Russian construction inherited from Proto-Slavic in much the same way as it did the Baltic construction. It is noteworthy that in the oldest northern Russian written texts, the nominative forms the same kinds of constructions with an infinitive as in the East Baltic languages. Later, it also began to be used with finite verb forms (чаша питья наливают, literally: “they pour a cup of drink”), and the construction itself took a distinctive path of development associated with processes characteristic of the Slavic languages: syncretism between nominative and accusative forms, and the grammaticalization of the animate–inanimate distinction (Timberlake 1974: 104–116; Журавлев 1984). In the western Russian dialects (within the Pskov–Smolensk–Vologda triangle), the nominative with an infinitive was apparently also sustained by a Baltic substrate. The role of Lithuanian is indicated by the fact that in the fourteenth–seventeenth centuries the construction is attested in documents exhibiting Belarusian features, written in the territory of the Grand Duchy of Lithuania—in Polotsk, Trakai and Vilnius (Dunn 1982: 517). It is still frequently used in Belarusian dialects that have recently become established in the Lithuanian-speaking area (in Pelesa, Voranava, etc.; see Судник 1982: 246–247). In contrast, in the southern East Slavic dialects (in the territory of present-day southern Russia and Ukraine), the construction with the nominative apparently disappeared before the beginning of literacy, and seventeenth-century texts from this area contain only remnants of it (in some places reactivated by the prestige of the Moscow chancery language). 3.9.4. The history of the nominative with an infinitive is not an exception. After the fifth–sixth centuries, intensive contact with the West Finnic languages also helped preserve other archaic constructions in the East Baltic languages: the nominative with neuter forms (linai minima, durys atidaryta), as well as predicative participles expressing a generalized modal meaning of indirect mood (author 1977; 1979: 213–218; 1985; 1990: 228–235). Contacts with the West Finnic peoples in much earlier times helped preserve in the Baltic languages ancient case constructions with postpositions, which developed into postpositional locative cases (author 1985: 191–193). In all the cases mentioned, we see inherited Indo-European sentence structures that the East Baltic peoples retained and reinterpreted according to West Finnic languages 98 characteristic patterns. Some researchers tend to trace the origins of the nominative-with-infinitive construction back to an even more distant past. It is thought to have been inherited from a time when the Indo-European accusative sentence structure had not yet formed and the object of transitive verbs (like the patient of verbs of state) was expressed by the same absolutive form (Palmaitis 1977: 116—119; similarly Schmalstieg 1987: 145—152).22 This assumption is possible, but not necessary, since the development of constructions in Baltic and Slavic languages can be explained on the basis of the accusative sentence structure and the primary subject function characteristic of the nominative. 4. DIRECTIONS OF FURTHER DEVELOPMENT 4.0. Constructions with an infinitive that has retained its original purposive meaning represent the oldest fragment of Indo-European sentence structure. The aforementioned constructions with the supine—the second infinitive in Lithuanian—belong to the same stratum. On the basis of the place of these constructions in the Lithuanian syntactic system and by comparison with related languages, we conclude that the infinitive originated in the use of the dative of verbal nouns, and in certain cases also the accusative, to expand and semantically supplement the verbal core of the sentence. Having become detached from the declensional paradigm of verbal nouns and turned into an invariable verb form, the infinitive in many cases lost its original case meaning and began to attach more closely to various finite verb forms; its relations with other sentence constituents changed. Viewed from the perspective of the primary usage patterns that can be reconstructed, the diversity of present-day infinitive constructions allows us to discern the main directions and stages of their further development. I can/ begin/ ask to work 4.1.1. Once infinitive constructions with the verbs reikėti, gailėti, vertėti ‘to be necessary, to feel sorry, to be worthwhile’, etc., which developed from nominal forms (reikà, gaila, verta, see 3.3), had become established, a basis arose for linking the infinitive with other verbs of modal meaning as well, e.g.: I have to/ must go (by analogy with reikia < reikà eiti ‘it is necessary to go’), can work (by analogy with galima, verta dirbti ‘it is possible/worthwhile to work’), wanted to go, and so on. Similarly, the infinitive came to be used in combinations with verbs of more independent meaning, such as prašyti, liepti, drausti, trukdyti, kviesti ‘to ask, order, forbid, hinder, invite’, etc. (see 1.2.1). Infinitives of various forms were also linked to verbs of corresponding meaning in other Indo-European languages at different periods in their development. This parallelism in usage apparently reflects not the inherited character of such constructions, but a typologically common, perhaps even universal, tendency to use in this way invariable forms naming actions that have lost their former connection with declensional paradigms. Corresponding combinations tend to develop not only in Indo-European languages but also in languages of other families wherever invariable forms (usually also of nominal origin) develop. 22Some have also regarded corresponding Finnic constructions as a relic of a pre-nominative sentence structure; see Wickman 1955: 14, Schlachter 1968: 286. 99 verb forms. 4.1.2. In constructions with phase verbs (pradėti, imti, baigti, liautis, nustoti, etc. ‘to begin, start, finish, cease, stop’), the infinitive in many cases replaced the active participles used with verbs denoting the end of an action, which had long had a function close to that of a predicative attribute, e.g.: stopped shouting, ceased drinking, cf.: fell silent/quieted down while singing, stopped while walking, returned from mushroom-picking (for more detail, see aut. 1978: 33—37; 1979: 112—114; 1990: 133—134). 4.1.3. In constructions with verbs expressing desire, request, aspiration, or invitation, the infinitive could replace the supine, which developed from the accusative and with which verbs of these meanings are still sometimes linked in dialects, e.g.: What more do you want to have Tj; We ask you to sit Ds; Mother scolds, calls (someone) to spin BM 410 (see also LKG II 419—420). Such constructions have counterparts in Latvian, cf.: wants to eat, wants to drink... little belly BW 2892 (Endzelīns 1951: 996), literally: “to eat would want, to drink would want the little belly.” The accusative infinitive (supine) is also linked with verbs of modal meanings ‘to want; to aspire’, ‘to be able’ in the Vedas, cf.: gātúm icchanta RV I, 80, 6d “(he) wants to go” (Disterheft 1980: 125); bhū́yo vā dātúm arhasi V, 79, 10, literally: “more or to give (you) are able” (Macdonell 1977: 336; see also Sgall 1958: 231—232; Jeffers 1976: 2—3); researchers recognize that dative infinitives in constructions with verbs of these meanings are undoubtedly more recent, having become established by analogy (Benveniste 1935: 99; Sgall 1958: 234). It should be noted that the supine is attested from the earliest written sources also with the verbs likti, leisti ‘to remain, to let’, e.g.: Idant passilikcza Nammūsa... mattytu graszū sluszbū WIEschpaties/ ir atlankytu Basznitzia jo PsD 27,4 (Palionis 1974: 146), cf.: attako numirėlio saugotų Kp (Jablonskis 1957: 478); Leisk maną Szmones, man slusizū BrB II Moz 9,123; similarly with the verb of related meaning duoti ‘to give’: aš tau duosiu gertų ‘I will give you to drink’ KGr (cf. Latvian ēstu deva BW 2809,5 “gave (something) to eat,” Endzelīns 1951: 996), although here the dative infinitive has predominated from the beginning of the written tradition: Ghis duos liti ant Nessibijancziu BrB Ps 10,6, and so on. Given the replacement of the supine by the infinitive with verbs of motion, observable in the period of the written tradition, it is highly likely that in some constructions with verbs of the meanings discussed in this section, too, the infinitive became established in place of the supine used earlier. Independent infinitive 4.2.1. Independent infinitive usage in imperative, interrogative, or optative sentences, with various modal nuances, is based on constructions with reikėti, norėti, galėti and the other verbs mentioned above, e.g.: Run! Children, eat! Are you going to tell? Where to go now? How to live? Now just rejoice! and so on. Similar to its use in interrogative sentences, the infinitive is also used in subordinate clauses of complex sentences with corresponding conjunctions, e.g.: I don’t know how to feed myself BsP 23The accusative of purpose can be confirmed here by the use of the corresponding verbal noun; cf.: Hey, maiden, dear maiden, will you let me stay the night? JD 761. 100 I 65; I don’t see where to go IV 3, and in fossilized (idiomatic) expressions, e.g.: What more is there to say Kv et al. (for more detail, see LKG II 411—415; III 616—618). The connection with the infinitive, which depends on the modal verbs discussed in 4.1.1 and 4.1.3, is shown by both syntactic synonymy and the use of the dative with a subject meaning; cf.: Everyone get up! Should I return? Now all we can do is live and live! I don’t know where I should go, and so on. Used without a finite verb, the purposive can also form imperative sentences, e.g.: Children, say your prayers! Rk; Hurry up and go to bed! (see Jablonskis 1957: 329; LKG II 419). Infinitive usage with an exhortative meaning in place of first-person forms of finite verbs also developed from constructions with modal verbs, e.g.: It’s already late, let’s go (= we’ll go, let’s go) home Kb; But don’t forget to bring the mare back in the evening Ign; I got up from the shade and now let’s go mow Vrn (LKT 389). In many cases, the infinitive used in conditional constructions also carries an exhortative or optative nuance, e.g.: If we work, then work we shall (LKG II 414—415). The use of infinitives with imperative and exhortative meanings is characteristic of many related languages and is attested from their earliest monuments. However, we could consider it inherited from the Indo-European linguistic community (as Disterheft 1980: 191 does) only by assuming that the infinitive itself, as a separate invariable verb form, is of Indo-European origin. This is precluded not only by the variation in infinitive forms, but also by their clear historical connection with the use of case forms of verbal nouns, especially the dative. The imperative and optative use of infinitives should be explained not by genetic inheritance, but by a typological parallel in syntactic development. Having become detached from the paradigm of verbal nouns in various language groups, infinitives became invariable verb forms, resembling those bare verb stems without inflection that had expressed commands or exhortation since ancient times. Formed at different times and under different circumstances in different languages, they came within the sphere of a general typological tendency. 4.2.2. Just as modal constructions align with the imperative infinitive, so constructions with phase verbs (especially verbs denoting the beginning of an action) align with the infinitive when it stands in expressive or emotional speech in place of finite verb forms, e.g.: He at that devil—beat him, tear him apart (cf. started beating him, tearing him apart), while that devil begged him BsP IV 48; That man jumped out of the sleigh and ran, swinging his arms II 321; The fox runs! The dogs chase her! II 321. In Lithuanian, such use of the infinitive in place of finite verbs is stylistically marked, restricted to cases expressing an action that begins suddenly and is intense, and has clearly developed from constructions with verbs pradėti, imti, šokti and so on. Some related languages have gone further in this respect, extending the use of the infinitive in place of finite verbs to past tenses (cf. Latin infinitivus historicus; see Hofmann — Szantyr 1972: 367—368 with references). 4.2.3. Infinitive constructions with a zero form of the verb būti, incorporated into the sentence structure, laid the foundation for some modal uses of the infinitive 101 in the meaning of words. The infinitives most commonly used as parenthetical modal words are matyti, regėti(s), girdėti, sakyti, e.g.: It seems, he was not at home Slnt; There, apparently, is a familiar person walking Jrb; Jonas, it seems, has not come back today Kb; It is said that mushrooms are already sprouting there Vlk; A good road, they say, is a highway Dr (LKŽ XII 34) (see also LKG II 417–418; LKŽ VII 917). 4.2.4. When used alongside finite verb forms from the same root, the infinitive is sometimes used to emphasize the action they denote, as a kind of emphatic word, e.g.: To promise, it seems, everyone did promise ŽemR I 48. With verbs in the negative form, it is equivalent in function to a frequentative, e.g.: Not one... even tried to come to the blacksmith BsP I 84 (see also LKG II 417). Secondary nominalization of the infinitive 4.3.1. Having lost its case form, the infinitive became an indeclinable verb form, but was not incorporated into the verbal systems of tense, person, and voice categories. It became a general designation of an action, creating the conditions for its use in certain cases in positions characteristic of nouns, as object or subject. 4.3.2. With verbs denoting the transporting, making, seeking, giving, receiving, having, or taking of something, the infinitive of some verbs can function as a direct object and replace the accusative or genitive of a cognate noun, e.g.: Did you bring something to drink? Skdv; Katrė brought in something to wash with ŽemR I 89; I won’t swallow it; just give me something to smoke BsP III 51; Did you give the cows something to eat? Er; Look for something to eat at home Kp; Do you have a lot to cut? Jnšk; Take something to read ŽemR I 291 (see LKG II 403; III 166–167, 468–469); similarly already in old writings, cf. give me to drink SP I 58₁₇. On the other hand, such uses of the infinitive are connected with constructions in which the infinitive occurs with the case form of a concrete noun and has its usual purpose meaning, cf.: The guests... brought that beggar something to eat and drink BsP I 76 and The old woman immediately brought him pancakes to eat I 149. 4.3.3. In sentences whose predicate is a nominal, an adverb, or another infinitive, infinitives sometimes function as the subject and occupy the nominative position, e.g.: To live is a great gift rš; To go back around is a long way ŽemR I 45; To plow is not to harrow. 4.3.4. The cases discussed in this section seem to complete the developmental cycle of verbal nouns: action names, having lost their case meaning and become indeclinable forms, once again begin to be used in positions characteristic of nominals. However, the infinitive forms themselves remain unchanged and do not acquire nominal inflections. In general, in Lithuanian we see only the very beginnings of the secondary nominalization of the infinitive, limited to a small number of lexically and structurally defined cases. In many related languages, the development of infinitives in this respect has advanced further: prepositions and demonstrative or relative pronouns that turn into articles are added to secondarily nominalized infinitives, characteristic of an analytic type of syntactic structure. For example, already in Ancient Greek li- 102 In the Baltic languages, the infinitive with the article *to* is rendered by an action noun. This process has advanced particularly far in the Germanic languages; it is enough to recall German, where every infinitive that takes an article becomes a verbal noun incorporated into the system of prepositional inflection. CONCLUSIONS Internal reconstruction and comparison with related languages show that the basis for the Lithuanian infinitive as a separate verb form was the use of the dative of verbal nouns—action nouns—with the meaning of purpose and aim, inherited from the period of the Indo-European language community. The Lithuanian data discussed here make it possible to discern several stages in the formation and development of infinitive constructions. I. The first and oldest stage is represented by usage types in which the infinitive has retained its purposive meaning, or in which that meaning can be reconstructed on the basis of the construction’s structure and its place in the language system: (1) A verb, on its own (a, c) or with an object in the appropriate case (b), is linked to an infinitive expressing purpose and having the meaning of the dative of a verbal noun: a) ariau sėti (“for sowing”); b) ariau dirvą sėti; c) dirva liko sėti. (2) A verb, on its own (a, c) or with an object in the appropriate case (b), is linked to two forms denoting purpose and related in meaning: the dative of a concrete noun and an infinitive having the meaning of the dative of a verbal noun: a) ariau rugiams sėti (i.e., “for the rye, for sowing it”); b) ariau dirvą rugiams sėti; c) The field was left to be sown with rye. (3) A nominal predicate—the neuter form of an adjective, corresponding to an ancient bare stem (a), or the nominative of an adjective inflected for gender (b)—is linked to an infinitive expressing purpose: a) medus gardu valgyti (“for eating”); b) Honey is delicious to eat. (4) The same type of nominal predicate as in (3) is linked to two forms denoting purpose: the dative of a concrete noun and an infinitive with the meaning of a dative: a) honey is delicious for children to eat (i.e. “for children, for them to eat”); b) honey is delicious for children to eat. (5) A finite form of the verb būti ‘to be’ (expressed or zero), occurring with a nominative subject, is combined with an infinitive (a) or with an infinitive and a dative (b): a) the field is (Ø, was, will be) visible/near; b) the field is (Ø, was, will be) visible/near to you. II. As the dative of verbal nouns in -ti/*-tei became distinct from the nominal declension paradigm and drew closer to the verbal system, syntac- 103 tic relations within many constructions changed. On the basis of the primary usage types, the following models of infinitive constructions became established and widespread: (6) A finite form of the verb būti ‘to be’ (expressed or zero), combined with an infinitive and a dative denoting the subject of a modal state (a) or a time/place reference (b): a) he is (Ø, was, will be) to die; b) tomorrow is (Ø, was, will be) to die. The basis is type (5). (7) A noun combined with an infinitive denoting purpose and a dative: soil for sowing rye. The basis is type (2). (8) Constructions with an infinitive and an object case dependent on it, involving syntactically impersonal neuter adjective forms (a), impersonal verbs (b), and adverbs or nouns similar to them (c): a) (for children) it is delicious to eat honey (honey-NOM); b) (for the children) it is necessary to eat honey (honey); c) (for the children) it is early/time to eat honey (honey). The basis is the types in (3 a) and (4 a). III. The following constructions should apparently be assigned to a later stage of development: (9) Personal constructions with an infinitive dependent on modal (a), phase (b), and certain other verbs (c), which may in turn govern object cases: a) I must, can, want to eat (honey); b) began, finished, ceased eating (honey); c) I ask, invite, decided... to eat (honey). The basis is the types in (1) and (8 b), and perhaps also constructions with the supine. IV. The development of infinitive constructions was further shaped by the following syntactic processes: (10) Verbalization of the infinitive (its use in the position of a finite verb form) through the transfer of the construction’s modal (a) or phase (b) meaning to the infinitive form itself: a) Children, eat! Go home; b) The hare run, the fox chase. The basis is types (8 b), (9), and constructions with the supine (eikit valgytų, and so on); (11) Secondary nominalization of the infinitive in the subject (a) or object (b) position: a) to eat is a pleasure; to live is to work; b) brought something to drink; gave something to smoke; The basis is types (8 a), (9 c); (12) Nominalization of the infinitive-governing verb: desire, request, invitation to eat. The basis is type (9 a, c). The order in which stages II–IV are presented is provisional, since some usage cases, for example those with the imperative infinitive (10) or with personal verbs of requesting, inviting, or phase verbs (9), may also have arisen in earlier periods under the influence of archaic supine or other constructions. 104 The relationship among the types of infinitive use indicated here and their approximate direction of development can be represented schematically as follows: (5) ⟶ (6); (2) ⟶ (7); (3 a), (4 a) ⟶ (8); (1), (8 b) ⟶ (9 a); (8 b), (9 a, c) ⟶ (10 a); (9 b) ⟶ (10 b); (8 a), (9 c) ⟶ (11); (9 a, c) ⟶ (12). The processes of development of infinitive constructions discussed above have clear parallels in other Indo-European (and not only Indo-European) languages (with the exception of modern Balkan languages, where infinitives disappeared owing to circumstances specific to that area). This points to a typologically general direction of development and a tendency to use and assign meanings in a similar way, within the syntactic structure of language, to indeclinable forms that evolved from the case forms of deverbal nouns. A characteristic feature of Lithuanian is that, alongside newer types of infinitive use, the original, ancient constructions from which the infinitive form itself once evolved have been well preserved. In this respect, Lithuanian differs markedly from other modern languages and provides important evidence for the reconstruction of the Indo-European syntactic system. ABBREVIATIONS ASpr — Old Latvian Language Monuments in Facsimile Prints. Edited by A. Günther. Vols. 1—2. Heidelberg, 1929. BrB — The Bible, that is, the Entire Holy Scripture, translated into Lithuanian by Jan Bretkūnas... Königsberg (1579—1590) (cited from a manuscript photocopy). BrP — Postilla, that is, a Brief and Simple Explanation of the Gospels... By Jan Bretkūnas... Königsberg... 1591. BW — Barons Kr. and Wissendorffs H. Latvian Folk Songs. Vols. 1—6. Jelgava and Petrograd, 1894—1915. DnR — K.Donelaitis. Raštai. V., 1957. EEv — Gospels... Translated from Latin into the Latvian language. By R. P. Georgium Elger... Vilnius... 1671 (= ASpr I 22—242). GlB — The Holy Bible, or the Holy Word of God... Riga... 1685—1689 (E. Glück’s Bible translation). Jab1RR — J. Jablonskis. Selected Writings. Compiled by J. Palionis. Vols. 1—2. Vilnius, 1957—1959. KGr — Grammatik der littauischen Sprache von Dr. Friedrich Kurschat. Halle a.S., 1876. KPP — Kalbos praktikos patarimai. Sud. A.Pupkis. V., 1985. LKG — Grammar of the Lithuanian Language. Edited by K. Ulvydas. Vols. 1—3, Vilnius, 1965—1976. LKŽ — Dictionary of the Lithuanian Language. Vols. 1—15. Vilnius, 1941—1991. ME — K. Mühlenbachs, J. Endzelīns. Dictionary of the Latvian Language. Vols. 1—4. Riga, 1923—1932. MLLG — Proceedings of the Lithuanian Literary Society. Vols. 1—6. Heidelberg, 1883—1912. MLLVG — Grammar of the Modern Latvian Literary Language. Vols. 1—2. Riga, 1959—1962. RV — Rigveda Saṃhitā. Hrsg. F.M.Müller, 1887. SIG — Hymns for the Use of Catholics. By Salomo- 105 ne Mozerka Słowoczyński... Wilnae... 1646 (Photographic edition: S.M.Slavocinskis. Giesmės. Edited by J.Lebedys. V., 1958). Supr — Супрасльская рукопись. Изд. С.Северьянов. СПб., 1904. VD I — Lithuanian Folk Songbook. Wedding Songs. Vol. 1. V., 1983. Other source abbreviations are the same as in LKŽ and LKG. LITERATŪRA 1. Ambrazas S. Veiksmažodžių abstraktų darybos tipų istorinė kaita. — Lietuvių kalbotyros klausimai. T. 26, 1987, p. 88—95. 2. Ambrazas V. Absoliutinis naudininkas XVI—XVII a. lietuvių kalbos paminkluose. — Lietuvių kalbotyros klausimai. T. 5, 1962, p. 3—146. 3. Ambrazas V. Netiesioginės nuosakos paplitimas ir kilmės problema. — Lietuvių kalbotyros klausimai. T. 17, 1977, p. 7—54. 4. Ambrazas V. Lietuvių kalbos vardininko su dalyviu (nominativus cum participio) struktūra ir raida. — Lietuvių kalbotyros klausimai. T. 18, 1978, p. 7—67. 5. Ambrazas V. Lietuvių kalbos dalyvių istorinė sintaksė. V., 1979. 6. Ambrazas V. Die indogermanische Grundlage des Dativus und Nominativus cum infinitivo im Baltischen. — Indogermanische Forschungen. Bd. 92, 1987, S. 203—219. 7. Ambrazas V. Сравнительный синтаксис причастий балтийских языков / Vergleichende Syntax der baltischen Partizipien. V., 1990. 8. Ancītis K. Aknīstes izlokšne. Rīga, 1977. 9. Balčikonis J. Rinktiniai raštai. T.1. V., 1978. 10. Balkevičius J. Dabartinės lietuvių kalbos sintaksė. V., 1963. 11. Benveniste E. Les infinitifs avestiques. Paris, 1935. 12. Bergmane A. Nenoteiksme (Infinitīvs). — Mūsdienu latviešu literārās valodas gramatika. S.1. Rīga, 1959, lpp. 630—632. 13. Behaghel O. Deutsche Syntax. Bd 2. Heidelberg, 1924. 14. Boeder W. Zur Rekonstruktion von Infinitivkonstruktionen im Indogermanischen. — Linguistic Reconstruction and Indo-European Syntax, ed. P.Ramat e.a. Amsterdam, 1980, S.207—224. 15. Brugmann K. Grundriss der vergleichenden Grammatik der indogermanischen Sprachen. Bd.2., T.1. Strassburg, 1906; Bd. 2., T.2. Strassburg, 1911. 16. Delbrück B. Altindische Syntax. Halle, 1888. 17. Delbrück B. Vergleichende Syntax der indogermanischen Sprachen. Bd. 2. Strassburg, 1897. 18. Disterheft D. The Syntactic Development of the Infinitive in Indo-European. Columbus. Ohio, 1980. 19. Dunn J.A. The Nominative and Infinitive Construction in the Slavonic Languages. — Slavonic and East European Review. V. 60, N. 4, 1982, p. 500—527. 20. Endzelīns J. Latviešu valodas gramatika. Rīgā, 1951. 21. Engert H. Syntaktische Studien. — Archivum Philologicum. T. 6, 1937, p. 108—122. 22. Engert H. Syntaktische Studien. — Archivum Philologicum. T. 7, 1938, p. 45—56. 23. Fraenkel E. Zur baltoslavischen Grammatik. KZ, Bd. 53, 1925, Hf. 1, S. 36—65. 24. Fraenkel E. Syntax der litauischen Kasus, K. 1928. 25. Gebauer J. Historická mluvnice jazyka českého, d. 4. Skladba. Praha, 1929. 26. Hahn A. Some Hittite-Sanskrit Parallels. — Language. Vol. 29, 1953, p. 242—255. 27. Hahn A. Verbal Nouns and Adjectives in Some Ancient Languages. Language. Vol. 42, 1966, p. 378—398. 28. Hamp E.P. Baltic Infinitives and Verbal Nouns. — Lituanistikos instituto 1981 m. suvažiavimo darbai. Chicago. III, 1985.p. 199—207. 106 29. Haudry J. Les emplois doubles du datif et la fonction du datif indoeuropéen. — Bulletin de la Société du Linguistique. T. 63, 1968, 1, p. 141—159. 30. Haudry J. L’emploi des cas en védique. Lyon, 1977. 31. Havránek B. Staroruská osobní konstrukce ryba loviti a její obdoba v severoruských nářečích. — Bulletin ústavu ruského jaz. a lit., 12, s. 169—176. 32. Hofmann J.B. — Szantyr A. Lateinische Syntax und Stilistik. München, 1972. 33. Jablonskis J. Jablonskio raštai. T. 4/ Red. J.Balčikonis. Kaunas, 1935. 34. Jablonskis J. Rinktiniai raštai. T.1/ Sud. J.Palionis. V., 1957. 35. Jacobsson G. Zur Frage vom Nominativ als Kasus des direkten Objekts im Slavischen. — Lingua viget, in hon. V.Kiparsky. Helsinki, 1964, S. 71—82. 36. Jeffers R. Remarks on Indo-European Infinitives. — Language, Vol. 51, N.1, 1975, p. 133—148. 37. Jeffers R. Syntactic Change and Syntactic Reconstruction. — Current Progress in Historical Linguistics/ Ed. Christie W. Amsterdam, 1976, p. 155—157, 163—166. 38. Kiparsky V. Über das Nominativobjekt des Infinitivs. — Zeitschrift für slavische Philologie, Bd. 28, 1960, S. 333—342. 39. Kiparsky V. Das Nominativobjekt des Infinitivs im Slavischen, Baltischen und Ostseefinnischen. — Baltistica. T. 5(2), 1969, p. 141—148. 40. Krause W.-Thomas W. Tocharisches Elementarbuch. Bd.1. Grammatik. Heidelberg, 1960. 41. Kuryłowicz J. The Inflectional Categories of Indo-European. Heidelberg, 1964. 42. LaRoche J. Homerische Untersuchungen, Bd. 2, T.1. Der Infinitiv bei Homer. Leipzig, 1893. 43. Loś J. Gramatyka starosłowiańska. Lwów, Warszawa, Kraków, 1922. 44. Macdonell A.A. A Vedic Grammar for Students. London, 1955. 45. Marvan J. Verbum infinitum в балтийских языках и его отношение к славянскому союзу. — Slavia Pragensia, 1960, №2, s. 7—15. 46. Mažiulis V. Baltų ir kitų indoeuropiečių kalbų santykiai (deklinacija). V., 1970. 47. Miklosich F. Vergleichende Grammatik der slavischen Sprachen. Bd. 4. Wien, 1874. 48. Mülenbachs K. Sintakse jeb teikuma mācība. — Latviešu gramatika J. Endzelīna un K. Mülenbacha sarakstīta. Rīgā, 1928, lpp. 112—281. 49. Ose F. Supinum und Infinitiv im Hethitischen. MVAeG, Bd. 47, T.1. Leipzig, 1944. 50. Trebski J. Gramatyka języka litewskiego. T.3. Warszawa, 1956. 51. Ozols A. Latviešu tautasdziesmu valoda. Rīgā, 1961. 52. Ozols A. Veclatviešu rakstu valoda. Rīgā, 1965. 53. Ozols A. Raksti valodniecībā. Rīgā, 1967. 54. Palionis J. Lietuvių literatūrinė kalba XVI—XVII a. V., 1967. 55. Palionis J. Über einen Typ des Infinitivsatzes im Litauischen. — Baltistica, 1 priedas, 1972, p. 125—130. 56. Palionis J. Supino vartojimas XVI—XVII a. lietuvių rašto paminkluose. Baltistica. T. 10 (2), 1974, p. 143—148. 57. Palmaitis L. Dėl baltų kalbų nenominatyvinės praeities. Baltistica, 2 priedas, 1977, p. 114—123. 58. Paulauskienė A. Paskirties arba tikslo naudininkas su bendratimi. — Kalbos kultūros teorija ir praktika. Lietuvių kalbos sekcijos sąsiuvinis N 5 (18), 1970, p. 49—53. 59. Piročkinas A. J.Jablonskis apie vardininką su bendratimi. — Kalbotyra. T. 22(1), 1972, p. 61—70. 60. Piročkinas A. Jono Jablonskio kalbos taisymai. K., 1986. 61. Pořak J. Vývoj infinitivních vět v češtině. — Acta Univ. Carolinae, Philol. 16. Praha, 1967. 107 62. Pupkis A. Kalbos kultūros pagrindai. V., 1980. 63. Reiter N. Die deutschen Lehnübersetzungen im Tschechischen. Berlin, Wiesbaden, 1953. 64. Renou L. Monographies sanscrites, t. 2. Le suffixe -tu et la constitution des infinitives en -toh, -tave, -tum. Paris, 1937. 65. Rudzīte M. Latviešu dialektoloģija. Rīgā, 1964. 66. Schlachter W. Arbeiten zur strukturbezogenen Grammatik. München, 1968. 67. Schleicher A. Litauische Grammatik. Prag, 1856. 68. Schmalstieg W.R. A Lithuanian Historical Syntax. Columbus, Ohio, 1988. 69. Seiler H. Relativsatz, Attribut und Apposition. Wiesbaden, 1960. 70. Šagall P. Die Infinitive im Rygeda. — Acta Univ. Carolinae, Philol. 2, 1958, S. 137—268. 71. Sirtautas V. Konstrukcijų buvo matyti, girdėti... struktūra. — Kalbotyra. T. 22(1), 1971, p. 71—79. 72. Stang Chr.S. Vergleichende Grammatik der baltischen Sprachen. Oslo, Bergen, Tromsø, 1966. 73. Szemerényi O. Einführung in die vergleichende Sprachwissenschaft. Darmstadt, 1970. 74. Šukys J. Linksnių ir prielinksnių vartojimas. K., 1984. 75. Timberlake A. The Nominative Object in Slavic, Baltic, and West Finnic. Slavistische Beiträge, 82. München, 1974. 76. Trávníček F. Historická mluvnice česká, 3. Skladba. Praha, 1956. 77. Trost P. Infinitiv v litevštině. — Studie ze slovanské jazykovědy. Praha, 1958, p. 271—274. 78. Vaillant A. Grammaire comparée des langues slaves. T. 2. Lyon-Paris, 1966. 79. Valeckienė A. Lietuvių kalbos gramatinė sistema. Giminės kategorija. V., 1984. 80. Večerka R. K vývoji modálních konstrukcí v slovanských jazycích. Otázky slovanské syntaxe, r. 3. Brno, 1973, s. 303—306. 81. Veenker W. Die Frage des finnugrischen Substrats in der russischen Sprache. Bloomington, 1967. 82. Vondrák W. Vergleichende slavische Grammatik, Bd.2. Göttingen, 1928. 83. Wickman B. The Form of the Object in the Uralic Languages. Uppsala, 1955. 84. Zinkevičius Z. Lietuvių kalbos istorinė gramatika. T.2. V., 1981. 85. Деянова М. Функционален развой на инфинитива в сърбохърватски език. София, 1982. 86. Журавлев В.К. Nominativus cum infinitivo с точки зрения морфологической нейтрализации. — Baltistica. Т. 20(2), 1984, p.119—125. 87. Иванов Вяч.Вс. К типологии инфинитива в балканских языках. — Славянское и балканское языкознание. Москва, 1976, с. 216—230. 88. Ларин Б.А. Об одной славяно-балто-финской изоглоссе. — Lietuvių kalbotyros klausimai. Т. 6, 1963, p. 87—107. 89. Потебня А.А. Из записок по русской грамматике. Т. 1—2. Москва, 1958. 90. Сабенина А.М. Конструкция с именительным падежом прямого объекта. — Историческая грамматика русского языка. Синтаксис. Простое предложение. Ред. В.И.Борковский. Москва, 1978, с. 403—416. 91. Спринчак Я.А. Очерк русского исторического синтаксиса (простое предложение). Киев, 1960. 92. Степанов Ю.С. Оборот земля пахать и его индоевропейские параллели. — Известия АН СССР, сер.лит. и яз. Т. 62(2), 1984, с. 128—143. 93. Топоров В.Н. Из области балто-славянских фольклорных связей. — Lietuvių kalbotyros klausimai. Т. 6, 1963, p. 149—176. 94. Ходова К.И. Простое предложение в старославянском языке. Москва, 1980. 108 THE SYNTACTIC DEVELOPMENT OF THE INFINITIVE IN LITHUANIAN Summary Differences between infinitive stems in Lithuanian-Latvian and Prussian alongside with the genetic identity of dative singular flexion permit to assume that the infinitive was isolated from the paradigm of verbal abstracts and included into the verbal system after the split of Common Baltic. Because of a comparatively late differentiation of the infinitive and the conservative nature of the syntactic system, Lithuanian has preserved the ancient use of pre-infinitival forms better than the majority of other contemporary Indo-European languages. By means of internal reconstruction and comparative analysis the main stages in the syntactic development of the infinitive are distinguished and presented in the form of patterns and schemes of syntactic relations. The oldest stage is represented by patterns in which the dative of ti- stem verbal abstract with the ending -i/*ei is a predicate of an imbedded purpose clause. Its use with another (nominal) dative of purpose or destination gave rise to dative with the infinitive. In many utterances still in use in contemporary Lithuanian dialects, both members of this construction have retained the original dative meaning, relation to the main verb and possibility of separate use, e.g. ariau lauką rugiams sėti ”I ploughed the field to sow rye” (lit.: ”for rye for sow”), cf.: ariau rugiams ”I ploughed for rye” and ariau sėti ”I ploughed for sowing”. The nominative with the infinitive after verbs like tekti ”fall to”, likti ”remain, be left”, rūpėti ”concern” and predicative adjectives (especially in neutral form) is based on the same ancient use of the verbal abstract dative expressing purpose, e.g. laukas liko arti ”field remained to plough (for ploughing)”; medus gardu/gardus valgyti ”honey is tasty to eat (for eating)”. As a result of internal re-analysis, the subject nominative has been reinterpreted as object and replaced by the accusative in many clauses turned syntactically impersonal: gardu/liko/reikia valgyti medų ”it is nice/remains/necessary to eat honey”. Contacts with West Finnic languages (especially after the 5th c. AD) can be regarded as a favourable condition for preserving the object nominative with the infinitive in some Lithuanian, Latvian and North-West Russian dialects. The subject nominative in its original function is retained after the verb būti (often in zero form) in clauses denoting content of perception (namai Ø matyti ”the house is to be seen”) and necessity (namai bus statyti ”the house is (lit.: will be) to build”. The subsequent syntactic development of the infinitive already separated from the verbal noun (as an uninflected verbal form) is reflected in personal clauses with modal, phasal and other verbs: jis nori/turi/prašo/leidžia grįžti ”he wants/must/begs/allows to return”, jie pradėjo/baigė statyti ”they began/finished building (lit.: to build)”. The existence of such patterns in many relative languages (sometimes regarded as inherited from protolanguage) is considered here to be a result of typologically determined tendencies in the use of impersonal (uninflected) verbal forms. The transfer of modal and phasal meaning to the infinitive itself has determined its use as an equivalent of imperative and some other finite verb forms in emphatic speech, e.g. Keltis! ”stand up!”; Sūnus bėgt, tėvas vytis ”The son took to his heels (lit.: to run), the father after him (lit.: to pursue)”. Secondary nominalization of the infinitive in the accusative or nominative position, attested in many related languages, is still on its rudimentary stage in Lithuanian. The syntactic development of the infinitive in Lithuanian has been influenced by its interrelations with supine. The supine (originally the accusative singular of tu- stem verbal nouns) lost its case meaning earlier and was replaced by the infinitive in many clauses with verbs denoting motion and similar processes. 109