DAG SVEDSTEDT

THEME

POSTPONEMENT

IN R U S S I A N

0. Sometimes sentences are found in Russian in which an object is not in its usual linear position but is placed further towards the end of the sentence. In the present paper this phenomenon will be called p o s t p o n e m e n t (PP for short). Both thematic and rhematic objects 1 may be postponed: (1) A~tmunbt~t, n p e O A a e a ~ xoOumb h a m eo~enbKuMu u p a 3 y mbtMU, mo~b~:o n e Ooeoaapuea~." (Hp. I, 39) /1/10/12 2 (2) otta n a y u u m mam/eaamb u e e o , xoTa... (Flo.~., 197) In (1) the object is thematic, and the utterance is said to have t h e m e p o s t p o n e m e n t (TPP). In (2), on the other hand, the object is the rheme of the utterance, and in this case we speak about r h e m e p o s t p o n e m e n t (RPP). In this paper sentences with TPP will be the main subject of study. Our aim is to find an explanation for what may cause this type of PP, to describe its communicative function, and to discuss it from a normative point of view. The source material mainly consists of literary texts, principally postwar fiction. In addition use has been made of recorded material of spoken language (below designated (Sp.)), consisting of stories retold by a number of Soviet Sixth Form students. In the great majority of sentences with TPP the postponed object is a personal pronoun. This calls for an introductory discussion of what may cause the positioning of these items in different types of sentences. 0.1. In Russian, as in many other languages, objective personal pronouns (O) mostly occur as clitics. In utterances with non-inverted word order consisting of S (one-word subject), P (predicate in the form of a finite verb) and O, there is a clear tendency to link O enclitically to a prosodically marked (stressed) S. Some examples: (3) 6yc)mo u n e Ba6e.ab e e o nucaJt, ... (Hall. II, 118) (4) Toct
16

DAG SVEDSTEDT

(1974b, 172) observes, typical of rheme components. S as rheme or forming part of a rheme is always stressed. (5) H a c m o / t e .aeatcaau .~6.aoKu. Ma.aertbtcuft Maab,-tUt¢ o,.tenb x o m e a ~ e c r n b ux. M a M a e M y ctca3aaa : ... (Sp.) (6) Ocrnaabrtbte Pbta~ceeo m e p n e m b n e MOeaU - on u x m e p p o p u 3 u p o a a a . (AKc., 326) In these utterances S is a stressed primary theme. In (5) the stress is due to the fact that the theme is new. 3 In (6) the change of primary theme vis-/~-vis that of the preceding utterance causes S to be prosodically marked. 4 If, on the other hand, S is not stressed, O tends to be linked enclitically to P instead: (7) B a p a a p a Oaa~ce noa~caae~a eeo. (~p., 120) (8) Ya~c a uctcaa m e 6 a , ucKaa... (F~., 338) In (7) and (8) we must assume that P is the only prosodically marked component, in (7) because of the rheme-emphasising particle, in (8) because of the emphatic repetition of P. In utterances with the above-mentioned syntactic structure the positioning of O is thus primarily dictated by prosodic factors (in so far that O is enclitically linked to the first stressed component), while syntactic-hierachic factors (the regens - rectum order) and communicative factors (the theme rheme order) play a less important part. 0.2. Also in utterances with a more complex syntactic structure than the ones described above it seems that prosodic factors are of great importance for the positioning of O. Such utterances are usually divided into rhythmic segments which are formed by a prosodic marking of the last element of the segment. Bogustawski (1977, 200) calls such a marking a n t i c a d e n c e . It is characterised by a rising tone and its function is, among other things, to signal a continuation. Each component to the left of the one carrying c a d e n c e (final intonation) can in principle have anticadence. In the great majority of utterances with non-inverted word order P takes anticadence, and O is enclitically linked: (9) C a u t a n p o m ~ n y : t M n e m o a c m y t o mempac)b. (TeH~p., 183) (10) B y , y ~ a r e y C m e n K u neOoeOeuubtft r y c o r x a e 6 a . O n 6 p o c a e m eeo Mampocy. (Co., 82) (11) A u a m o a u f t Bacu:tbeeu~t n o s e a MeUa e ceoft K a 6 u n e m . (Herp., 95) (12) £I noO~tuuuaca eft c yOoso:tbcmsue~u. (F~., 109) It is highly probable that P will carry anticadence in (9)-(12). Utterances with the above linear-dynamic structure are very common in Russian. Adamec (1966, 66) points out that the structures (9)-(11), i.e. with S as primary theme, P + final object as rheme, and a highly "individualised"

THEME

POSTPONEMENT

IN R U S S I A N

17

object of type O as interrhematic theme, are among the most common in written language. In our literary material a summing up of the great number of utterances with O having non-inverted word order and one or more "final components" (F) to the right of P, none of which is an infinitive, shows that O is placed immediately after P in about 84 ~ of the utterances. Not a single instance of TPP has thus been found in the literary material. 5 If O is postponed it always constitutes the rheme of the utterance: (13) Cna~taJta on nocaOuA na me~eey e e , a n o m o m nomoe a3ne3mb u Mne. (F~., 48) (14) HomoM o n OaJl 9 m u c m u x u n a ~ t u ctca3asl, ~tmo6~,t ~tbt St, Iy,U~U UX. (Coqbp., 349) 0.3. In utterances of the same type as those in 0.2 but having a compound predicate (P + subjective infinitive governing O) there is a very strong tendency to place O immediately after the infinitive. Some examples: (15) KoeOa Clbponos tcou~tun eosopumb, 60ftttbt c m a ~ u ~ a O a s a m b e M y s o n p o c b t . (HnK., 164) (16) ,el xo~ty noOapumb sam s a m n o p m p e m . (AKC., 226) (17) O n a noo6euta:ta npunnrnb nac s 6~uatcaftutue Onu. (Coqbp., 147) In utterances of this type O is placed directly after the governing sentence component even more often than in the utterances under 0.2, viz. in about 87'Yo of the utterances. Once more the explanation may be found on the prosodic level: (18) M e p e c b e s cxaamu:t 6yJnaatcKy c)poaccaulu•u pytcaMu. O n n p o ~tuma:t e e m y r a atce, y cmo:ta, (l-Ion., 225) (19) 14 o n h a m a n ~ t u m a m b u x m y r a atce, 3a cmonoM, (KopuI., 157) Because of its relatively limited length (18) can be spoken without rhythmic segmentation, i.e. without anticadence. In (19), on the other hand, it is highly probable that segmentation will take place, and that the infinitive (I) will carry anticadence. In other words, the higher frequency of IOF compared to POF may be explained by the fact that I is relatively more often stressed than P. 1. In the utterances which have been described by way of introduction and which are very heterogenous as regards both syntactic and communicative structure, we want to claim that O has a clear tendency to be enclitic. However, as has already been stated, this tendency towards encliticism does not result in TPP in written language. Such utterances, in which TPP occasionally does occur in written language, can be grouped into two main types:

18

DAG SVEDSTEDT

(a) M o n o s u b j e c t i v e s e n t e n c e s , which are characterized by O being the underlying subject of the sentence, e.g. : (20) PucKo~anno, d a u m p y O H o 6btAO c n y c K a m b c a e M y no y3Kof, ~pymoft ~ecmnut4e. (Fpan., 269) (b) H e t e r o s u b j e c t i v e s e n t e n c e s 6, e.g.: (21) on lieMeO~enno o c m a n o a u n eeo u a lie~moeux, no pe3tcux c n o a a x OaR 3 a M e m u m b

eMy,

u m o nepeO h u m crnoa.~ nopyuutc, a ne

Opyeoft tcaKoft o06utlep. (For. II, 55)

In these sentences O is the underlying subject of the embedded infinitive clause. When discussing the underlying structure of the heterosubjective sentence we will call this subject S~. It is not coreferent with the subject of the finite clause, Si. Monosubjective sentences with TPP have a very low frequency and occur mostly in colloquially coloured language, e.g. : (22) I£a~ 3liamb, ~tmo cmaJtocb 6bt c 6eOm,lm C m e n a n o m C a x a p o s u ~teM? IloOu, neeo Oo6poeo, n p u u t ~ o c b

6bt U C n b t m a m b

e M y na

c m a p o c m u Jtem maKoft liecJtbtxanubfft cpaM, u6o, (Coa., 94) In heterosubjective sentences TPP occurs more frequently and is found in all kinds of style. (23) is taken from a scientific text, while (24) is from a literary text: (23) K o n m e t ¢ c m moaw.ern O a m b n o n a r n b l i a M , ttrno pe~tb uOem o c)e~cmsuu, ~at¢ 06 06bt~tnOM csoftcmae. (Pacc., 70) /5/11/7/ (24) F ~ a O a eMy a nul4o, a cMoey MU~tUKOft, eJte 3a.~emnoft monKoft a4UMUKOfi noKa3amb e m y , u m o a lie mate-too n p o c m , ~tmo .~elt.~t He m a K - m o n p o c m o yliu3umb, O a m b n o H a m b Ma3na,

a ~ya~c,~ulia,

~tmo ~to~ e u z u m

e M y , ~mo a lie pa3-

9rno

moatce

aKm

mya~cecmaa, a lia neeo ~ l i e quxarnb. (AKc., 213)

In Gogol TPP is found in both mono- and heterosubjective sentences to a much higher degree than in modem fiction. Whether the phenomenon is a stylistic characteristic typical of Gogol or whether it occurred more generally in the 19th century is a question that must be left open because of the lack of more extensive material from that time. 1.1. In almost monoproper (25) (26)

the introduction we mentioned that postponed thematic objects without exception are personal pronouns. This holds true both in and heterosubjective sentences. Only single instances of TPP of nouns have been found, e.g. : 3 n a u u m , ~toa~cuo y f t m u HaM c K y n p u a u b t ~ t e . ~ s ~ec. (Flp. I, 284) O n uecrostbro pa~ ocmauaaitusaJtca ua :tecmnutte, raK 6bt a~ceJtaa co6parnbca c c)yxoM u noOy~tamb o morn, KaKuM 6~a o6pa3om ,gamb nouyscrnaoaamb I 1 1 u : t n e p y eeo Oep3ocmb. (For. II, 61)

THEME POSTPONEMENT IN RUSSIAN

19

In other words T P P seems to presuppose that the postponed component carries a very high degree of "givenness". It is thus avoided if the component is not given enough to be expressed anaphorically, i.e. when reintroduction of the "full" N P is necessary. 1.2. Another typical feature of TPP is that the postponed component is not in final position (cf. ex. quoted above). If O is in final position we are dealing with RPP instead: v (27) H a c m o ~ e a o n e p a t t u o n n o f t aeaw, a~ Mo~wOoft ~teaoaet¢ u c yOueaen u e M c M o m p e : t Ha c a n u m a p t c y , K o m o p a ~ peMH~MU npua~3b~aa.,m e e o p y K u u n o e u t¢ cmo:ty. KoeOa a~ce o n a npunoOna:ta p y 6 a u t t c y , mo mym

npuut:tocb

yOuaumbca

A4He, u 6 o a yauOe:t y Heeo

Ha atcuaome p y 6 e t l . (Fa~., 225)

(28)

B :temmoto pacnymu~y

O o c m a a u m b u x MO:rtCHO 6blAO mOAbKO Ha

ca~to:teme. B m a f t e e u cpeOu 6o:tom agpoOpoMoe He 6bt:to. AnOpeft uMe:t n p a K m u ~ y nopy~tu~

no

npbt:~qCKaM C n a p a u o o m o ~ t ,

Oocmaaumb

3aaomOe~oM

t c a p m b t e M y . (Fp., 88)

1.3. One more feature characteristic of utterances with T P P is that I carries anticadence 8, e.g. : (29) KoeOa OeO c m a ~ c m a p , cbm u n e e e c m K a n e cma:tu caatcamb e e o 3a c m o ~ , a O a a a ~ u o 6 e O a m b e M y 3a neu~o~. (Sp.) /3/8/12/ In their rhythmic structure the utterances under discussion are most often organised so that prosodically marked and unmarked components follow each other regularly, a feature they have in common with the utterances described under 0.2 and 0.3. In (9)-(12) we have: P with anticadence clitic w o r d - F with cadence; in (15)-(17) and in the utterances with TPP: unmarked P (verb or adverb) - I with anticadence - clitic word - F with cadence. Thus the utterances with T P P follow a very common and so more or less automatised rhythmic pattern. This may be a contributing factor to the fact that T P P occurs much more frequently in spontaneous oral production than in written production. However, it is true that T P P is not at all a common phenomenon either in spoken or in written language. This is probably primarily due to syntactic factors (see below for a more detailed discussion). Another reason may be the fact that the communicative structure of mono- and heterosubjective sentences far from always permits the prosodic pattern described above. Some illustrative examples: (30) H o e m o a~ce a p e M ~ ef~ tte Hpasu:tocb, ~tmo e o c m b ma:lo e c m , u oHa ynpautusa:ta

eeo

xo~tern. (Fl.riaT., 115)

ecmb

6 0 J l b U t e U o6u:~ca.~acb, ~tmo oH n e

20

DAG SVEDSTEDT (31)

M b t earn ~an:tauupoea:tu noesOKy n a A c y a n u e f l y l c c o p .

[...]

B nepebte Osa n o c e u 4 e n u a IOceo5 9c-Cu6af~ n e n p e O : t a e a : t

Mne

(Coqbp., 104) In (30) P is rhematic. I signifies something given in the preceding context and is thus context-dependent (in terms of Czech FSP terminology9). This means that the action or state signified by the verb in principle can be seen as predictable, in any case to a higher degree than a context-independent component. A context-dependent component is prosodically marked only in special cases, and I will not be marked here. In (31) the whole infinitive phrase is context-dependent, and a prosodic marking of I is not possible. Since the utterances above are characterised by a stressed P and an unstressed I, there are no premises for an "automatised" TPP. Not a single instance of TPP has been found in utterances with rhematic P and contextdependent I. (32) M b t c:tuu4~om nepeepyaccenbt ,~ya~cumu n p e O c m a e : t e n u ~ m u , ~¢omonocemumb

Acyan.

pbte m e u t a t o m

Ham

euOemb

mup makuta,

Ka~coft on e c m b .

(Fpan., 38) (33)

B o ~ b u t e f t atce ~tacrnbto n a m a m b m e u t a e m rune, m n o e o e o ~ n e x o ~ y nomHumb, aocnomunanua

geutamm

auc)emb

rune aeulu ma-

(FpaH., 289) /3/12/8/ In (32), too, P is rhematic and stressed. I is here the first element in a rhematic segment and will probably not be prosodically marked. Very few instances of TPP have been found in utterances of this type. In (33), on the other hand, P is context-dependent and unstressed whereas I is rhematic and takes anticadence. In other words, the conditions favourable for an automatised TPP are present, and it is, as has been pointed out earlier, under these very conditions that the phenomenon generally occurs. (34) B m o a~ce s p e m a C e p n u n u n n e Oe~aJt suOy, ~tmo n e n o n u m a e m , Kumu, KaKue o n u e c m b .

tcat¢ 9 m u dtoc)u 6 o a m e a , 1oak 3 m o .onu m o e d u no6eatcamb, p a c m e p a r n b c a , 6 p o c u m b opyatcue. H a o 6 o p o m , on O a e a ~ n o ~ t y s c m s o a a m t , u r n , ~ m o n o n u m a e m 9too.

(35)

(CHM., 123)

/0/8/15/

O n 3 n a ~ : ona u mate 3 a s m p a atce s e p n e m c a n a 9 m y p a 6 o m y , n o xome:t

bamb

eft

nouyscmsosamb,

~tmo 9 m o

aaa~cno ne

(CnM., 394) In (34) both P and I are rhematic, as in (32). However, P is here desemanticised (vide 1.52). It is semantically integrated with the governed verb to a higher degree than P in (32). Under these conditions it gets no prosodic marking while I gets anticadence. In utterances with desemanticised P the rhythmic pattern favourable for TPP is thus present. In (35) d a t ' is semantically reduced in the same way as d a v a l in (34). But in the sentence structure there are no less than three verbs, which mom,~:o c):la nee, n o u O~a n e e o .

T H E M E P O S T P O N E M E N T IN R U S S I A N

21

probably results in the prosodic marking of dat' despite its desemanticised status. This marking is then not communicative but purely rhythmic. If the marking takes place, the prosodic pattern favourable for TPP is not present. 1.4. In spontaneous oral production TPP, as stated above, is prosodically conditioned to a large extent. Two tendencies co-operating on the prosodic level seem to be the ones that mainly cause the phenomenon: (a) The tendency to link O enclitically to a stressed component. If P is unstressed O tends to be moved further right and to be linked enclitically to the following stressed component, the infinitive (a tendency which is strongly counteracted by syntactic factors, i.e. that O through the move will follow after a verb to which it is not subordinate). The fact that postponed objects nearly always are personal pronouns seems to justify the view that TPP may be seen primarily as a form of clitic movement, not to the left, as in the utterances described in 0.1, but to the right. (b) The dominating tendency in spoken language is to organise speech prosodically so that marked and unmarked components follow each other regularly. TPP gives rise to an unstressed component between two prosodically marked components, I and F. 1.5. In written language the intonation is not fixed. Word order is then the primary means of indicating how the sentence is meant to be spoken. The fact that O has a strong tendency to encliticism makes it possible to use it as an a c c e n t u a t o r , ~° i.e. to place it after a component which ought to be stressed if the sentence is spoken. In a few cases the position of the accentuator organises the utterance in the rhythmic (non-communicative) respect. Thus for instance its position in (35) will probably result in the prosodic marking of P, despite the fact that the verb is desemanticised. In her study of Russian spoken language (1975, 195) Lapteva states the following: "B xo me apeM~ npnHunn p~IVMHqeCKO~ opFaHH3allHI4 BblCKa3blBaHHfl MO~eT BOHTH B CTO.UKHOBeHHe C HeO6XO~2HMOCTblO no,~qepKHyTOrO npori3HeceHria HHqbOpMaTHBHO 3HaqHMblX

3~eMeHTOB. I-Io HamnM na6~Im1IennaM, B 3TOM cny,~ae O6bX~HO no6e>r~aeT • pI4TM". However, in the great majority of cases the accentuator organises the sentence in both the rhythmic and the communicative respect. In sentences with TPP it indicates that I is context-independent and/or that P (verb or adverb) is context-dependent, semantically reduced or, to use Firbas' terminology, otherwise has a lower degree of communicative dynamism than I. 1.51. The communicative function of TPP is very clearly seen in monosubjective sentences. Some illustrative examples:

22

DAG

(36)

SVEDSTEDT

- H e uMeto n p a s a s c)annom c : t u t a e c o a e m o s a m b , 6blmb, 6 d a e o p a 3 y M n e e

6btAO 6bt nepeexamb

no, Moatcem sam

n a anne

(13ona., 208) In (36) the position of the accentuator shows that I is context-independent. Thus it is possible to deduce that there has been no earlier mention of a move (to altemative places) in the preceding context. The placing of O further to the left would make I ambiguous in the communicative respect: it could then be interpreted either as context-dependent or contextindependent. (37) - Ax, epycmno • n e e ~ a O e m b n a aac, - (OK., 238) (38) Cmparmo u a t c y m t ¢ o n o p o f t e . a a O e r n b M n e n a nee. ( I 0 . Ka3aKOB) (KOBT. 99) In the context-free utterance (37) the linear position of the accentuator causes a prosodic marking of the adverb, while I is probably not marked since it is interpreted as context-dependent. The situation will, in other words, be interpreted so that the speaker looks at the listener before he speaks, and in his view the listener is aware of this. In (38), on the other hand, both the semantic structure of the sentence and the position of the accentuator show that I is context-independent.1 (39) - ecJtu 6 ~ m o 3ha3, ~:aK xo~tem ona, KaK n y Y , c n o e f t c b t e p a m b H a m a u t y . . . . (Hall. I, 290) (40) Ox, KaK x o m e . a o c b n o e o a o p u r n b e f t c 6a6aMu... ((I)p., 62) In (39) both the context and the accentuator contribute to the interpretation of the adverb as rheme, while the infinitive phrase is seen as contextdependent. In (40), on the other hand, the position of the accentuator results in I being interpreted as context-independent. (41) TaK n a m o ~ t u o c m a s u J t u e e o o 6 e c)aMbt u o m n p a s u J t u c b tcaatcc)aa apMuu.

s csoto c m o p o n y 6 y n m o s a m b eopoc). 3 m o n p e O n p u a m u e y c)a.lt o c t ,

(For. I, 272) This utterance informs us about the time required to complete an operation. The successful completion of the operation is a presupposition. The positioning of O further to the left would make the immediate comprehension of the communicative structure of the sentence more difficult, because the presupposition, the theme, would not then be as evident: (41a) 7)TO npe~npnaTne u ~ t y O a ~ o c t , n p o u 3 a e c m u c He60~bUXnM npou3aecmu

u . ~ c n e 6 o m , tuu.~ a no.a,.taca.

B no.rlqaca.

(41b)

~TO npe~xnprtaTrm y O a J t o c b u m n p o u 3 s e c m u c He60.ru,m r l i B noa~taca. In (41a), and even more in (41b), there is a risk that the utterance would get the wrong intonation.

THEME

POSTPONEMENT

IN

RUSSIAN

23

When dealing with the function of personal pronouns on the communicative level Kovtunova (1976, 86) states the following: I'(OMMyHHIgaTHBHaf He3HaqllMOCTb J'IHqHblX MeCTOHMeHHHOTKpblBaeT H.rlft HI,IX BO3MO)KHOCTb 6 o n e e CBO6OHHO n e p e a a n r a T b c u a n p e a n o x e u n n . 9 T a OTHOCrITeabHaf CBO6OHa nO3aOnfeT nCnOnb3OBaTb nX a qbyHrUrm, He OXHOCfmeficf r CHHTaKcnqeCrOMy a~rl arTyanbHOMy qneHennro n p e m a o x e H n f : nriqHue MeCTonMennu q a c x o n r p a r o T pnxMoo6pa3ytonty0o ports.

By this she naturally means that personal pronouns may have various linear positions in the sentence without this fact causing any change in their syntactic or communicative (thematic) status. However, as shown above, the choice of position for these items may be an effective means of expressing the communicative status of other sentence components. 1.52. In heterosubjective sentences with TPP the verbs dat' (davat') and zastavit' (zastavljat') are most frequent as P. However, both these verbs are polysemic, and it is only when they have a particular semantic content that TPP occurs. Like many infinitive-governing verbs dat' carries the fundamental semantic feature CAUS, 'to cause', in all its senses. This semantic feature appears to be the predominating meaning of dat' when I is a verb of notion or perception, e.g. znat', pokuvstvovat', zametit'. (42) on He saMemu.a, nan abexa~ a cpeOuny o6tuupnoeo ceAa co •noy, cecmao• us6 u y:lm4. CKopo, oOnaKo a~ce, OaA s a m e m u m b e M y a m o npenop aOo~mbtft mo a~ot¢, npouseeOennblft 6peeen~tamoto Mocmoeoto . . . . (For. I, 161)

(43)

BceM OaMaM coseptuenno ne nonpaauaocb matcoe o6xoa~cOenue ttu~tu~coaa. OOHa u~ nux Hapouno npou~:ta a4uMo neeo, qmo6bt Oamb e M y a m o s a ~ t e m u m b . . . . (For. I, 245)

Since CAUS, a feature which is normally only a subcomponent in the semantic content of the causative verbs, here predominates the meaning of the verb, one may say that dat'-cause is semantically reduced, desemanticised. The verb is then semantically integrated with the governed verb to a high degree. In connexion with certain infinitives it can, together with the infinitive, be replaced by a causative verb, e.g. dat' znat' ~ soob~bit'. The semantic integration can also be seen in the fact that dat'-cause always has the same communicative status as I in the utterances found in the material. Either both the verbs are thematic, as in (42), or rhematic,. as in (43). Dat'-cause is the most frequent P in the utterances where TPP has been found. In the following examples dat' carries a different meaning: (44) Pa3y~tm,te, JtacKoable penu aceeOa Oeflcmaosa.au na Cepeatcy ycnoKoumemmo. On Oa.a eft y~tt, l m b .uo.~otca. (HaH. I, 160)

ce6a u ebmu./t tcpya~cKy

24

DAG SVEDSTEDT (45)

Cbm He OaA e~uy O o e o s o p u m b ,

nonepxny~c~ Obl•O•,

3a-

Moma~ eoJtosoft ompuqame~bnO. (13on~., 74)

(46)

H ace t~anucaAu 9~aa~uJluu. [...] cl)eOe C l c o a o p o O t c u n y Mbt

n e Oa~u n a n u c a m b o6a~tu~uto. (From a Russian reader) In addition to CAUS the feature of 'permission', to cause by allowing, is included in the meaning of the verb. In this semantic content there is moreover an implication referring to the underlying infinitive clause, viz. the modality of 'volition': S t wants to perform the action expressed by I. Dat'-permit is not semantically integrated with the governed verb in the same way as dat'-cause. It often occurs as rheme, while I is part of the theme, as in (46). The utterances with dat'-permit thus usually have a different prosodic pattem from those with dat'-cause: either both P and I are prosodically marked, which probably will be the case in (45), or then only P is, as in (46). Not a single example of TPP has been found in the material with dat'-permit. The most common meaning of dat', 'to give', can also be found when it governs an infinitive. This meaning is present in (29) above. Like dat'cause, dat'-give seems to be semantically reduced since it can, in connexion with certain infinitives, be replaced by a causative verb, e.g. dat' poest' p o k o r m i t ' , xz In utterances with dat'-give TPP has been found only in the spoken material. Like dat' (davat'), zastavit' (zastavljat') also has the fundamental feature CAUS. 13 (47) u maK npo ce6~ OyMa~ : "COeJtamb tcana:t - 9too 3nahum noft~arnb cuAy eoObt u 3 a c m a a u m t , p a 6 o m a m b ee na ~teaoaeKa." (Hp. II, 263) /3/11/9/ (48) ~ mo~t,Ko OymaA ceeoOna, ~tmo cOe~amt, tcana~ - 9mo ~navum no(t~tamb cuJty O00bt U 3 a c m a a u m t , e e p a 6 o m a m t ,

na ~teao-

aetca . . . . (lip. II, 265)

If this semantic content is the predominating one, as in the examples above, the verb is desemanticised. In connexion with certain infinitives zastavit'-eause may also, like dat'-cause, be replaced by verbs with similar meaning, where CAUS is built in, e.g. zastavit' smejat'sja ~ rassmegit'. The verb has a different meaning in the following examples: (49) Bee MomtaAu. Homo~t eopl4t,t 3 a c m a a u J t u n a c s c m a m b u

(50)

Oo~eo o6x~ont,laa~u caou~au meepOt,tMu :taOonaMu, Oo~a~cno 6t,tmt, pa3~,tclcuaaR cnpamannoe noO oOea~cOo~ opyatcue. (Flaycx., 383) - Pant, me ~me pa~peutaAu, - npu6aawta IIIypo~ma. [...] - ,q 3acrnaaAJo m e 6 a c)amt, Mne pa3peutenue, Oy~a:ta IIIypa.

(HI4K., 256)

THEME POSTPONEMENT IN RUSSIAN (51)

- A aom a ~acmaaato

sac

aey, camb

25

n a 3eMae c)o 3 a s m p a . . . .

(CrIM., 76) In addition to CAUS there is also the feature of 'force', causing by forcing. In a similar way to d a t ' - p e r m i t , z a s t a v i t ' - f o r c e adds the modality of 'volition' to the underlying infinitive clause. The meaning of the verb implies that S j d o e s n o t w a n t to perform the action expressed by I. This action may be included in the rheme of the utterance as in (49), but it often has thematic status. Either it occurs in the preceding context, as in (50), or can be derived from it. In the isolated utterance (51) I is probably seen as context-dependent. In other words, the most probable interpretation of the context is that the persons are already lying on the ground. Since both z a s t a v i t ' - c a u s e and z a s t a v i t ' - f o r c e have similar selection restrictions, sentences with this verb may be semantically ambiguous. In this case the positioning of the accentuator O may contribute to making one or the other interpretation more likely. The following examples are interesting from this point of view: (52) - H a n o p a M b t , Y x a n o a ! CablUtUUtb, npu~eabt! - ne o 6 p a u l a ~ sttuM a u u a n a elu6ucoaa, •pu•ny:t

K y ~ n e t l o a u M e n o a e n u o noOyMa:t,

n m o x o m e : t u Moe nputca3amb Y x a n o a y - uMe:t n a 9 m o n p a 6 o Ctt~mb nanopaMb~, m o ecrnb a.~acrnb~o KoManOupa s3aoOa 3 a cmasumb

SbtCKO~tUmb e e o cef~uac noO 60M6ea~cKoft K o p y O u a M

U3 cnacume:tbnof~ 3eM:lU, cam o c m a s a a c b a poaur~e, n o n e M o e

(Bona., 146) /0/12/11/ The text informs us about the power of the platoon commander to make Uchanov expose himself to the dangerous situation of leaving the trench while he, the commander, remains there. The interpretation z a s t a v i t ' - f o r c e would mean that Uchanov did not want to put himself in this dangerous situation, i.e. an implicit denial of his bravery. This interpretation is hardly in accordance with the aim of the text - nothing in the macrocontext supports it. In this case it is thus z a s t a v i t ' - c a u s e . The position of the accentuator also supports the latter interpretation. TPP has been found in utterances with clear cases of z a s t a v i t ' - c a u s e '4, albeit in solitary instances, whereas not a single case of TPP has been found in utterances with clear cases of z a s t a v i t ' - f o r c e . ' 5 (53) O m m y O a mt, t y u t : t u ; o m c y m c m s u e 3eM:tu, m e c n o m a ~ a c m a s u e t u 9 m o e o npu~:a3amb.

ttac y f t m u , n o 6 o a b u t y t o n a c m o a u l y m P o c c u m Mt, t y n e c a u c co6oft u a e p u m s ,no, ,~mo 3Oecb ee "'cOeaaem". (~aa., 95) In (53) we have a clear case of z a s t a v i t ' - f o r c e . The utterance informs us not only that certain circumstances made some people leave their native country but also that this happened against their wish. However, Si is here

26

DAG SVEDSTEDT

inanimate, which normally favours the interpretation zastavit'-cause.l 6 In order to make the semantic content of P completely unambiguous the writer has chosen to indicate that the verb should be prosodically marked by having it in italics. Because of this the verb will constitute the rheme of the utterance. As the desemanticised verb zastavit'-cause does not alone appear as rheme, we thus here have the verb zastavit'-force, and the fact that Sj is unwilling to leave is emphasised, which certainly is very important in the context. (53) is an example of a very unusual phenomenon: a prosodic marking may not only have a communicative but also a purely semantic function. The accentuator O here has its normal position after P, which is always the case when I is context-dependent (cf. 1.3. above). (54) - ~ ue ~uato, eoeopum, ~mo •ne Oe~amb. 3 a c m a e u n , e o s o p u m , n o b n u c a m b M e n a nanyto-mo 95anbuausyto 6yMaey, 6pocuA narnuaOqamt, py6~eft accuenaquaMu ; ... (For. I, 264) Here we have TPP, which results in the prosodic marking of I. At the same time the sentence structure contains a component which, like O, functions as a kind of accentuator, viz. govorit. The linear position of this component causes even P to be prosodically marked. This is one reason why the translation 'he forced me to' would seem to be more adequate than 'he got me to'. As was stated above dat' (davat') and zastavit' (zastavljat') are the verbs that most frequently occur as P in the heterosubjective sentences with TPP. Yet it is only when the verbs are semantically reduced that TPP occurs, i.e. with dat'-cause, dat'-give and zastavit'-cause. The explanation to this can be found on both the prosodic and the semantic level: these verbs are unstressed more often than other infinitive-goveming verbs, including dat'-permit and zastavit'-force. (If prosodic marking occurs, it is solely rhythmic, cf. (35).) As desemanticised verbs they are to a high degree semantically integrated with the govemed verb. This creates a tendency for contact position between the verbs, with TPP as a result.l 7 1.6. Single instances have been found in Gogol in which TPP may be purely syntactically motivated, at least when seen from a modem reader's point of view: (55) Kan npou3ae/tucb nepabte nonynnu, uumamestb y>tce auOeJt; nan noftOem OeJto Oa~ee, nanue 6yOym yOa~tu u neyba~tu eepoto, nan n p u b e m c n p a 3 p e t u u m b u n p e o O o ~ e m b e M y 6oJtee mpyOnbte npennrncmaun, [...] rno yauOum nomoM. (For. I, 346) Here we have two coordinated infinitives. However, the final component of the subclause would seem to refer only to the latter of them. Because of TPP, with the resulting prosodic marking of the preceding infinitive, our

THEME POSTPONEMENT

IN R U S S I A N

27

attention is drawn away from the paratactic relation between the two infinitives, and thus also from the not very happy referential relation between the words razre~it' and prepjatstvija. 2. In the sentences discussed so far, the different positions of O have not involved any differences in the interpretation of its communicative status. In all cases O has been thematic and clitic. Its linear position has helped in interpreting the communicative status of other sentence components. What, then, are the effects of PP on the communicative level, if the postponed object is an NP which is more unspecified in the communicative respect than personal pronouns, e.g. an NP which, in contrast to O, may be rhematic together with the governing verb? (56) CeJtuoSan n o , t o e Oee~lonKe e 3 : t e 3 m b na ~o3Jtbt. In (56) the object of P has its usual linear position. Without a known context it may be interpreted as context-dependent (the girl), or as contextindependent (a girl), and in principle also as constituting rheme alone (the girl, and not the boy). The ambiguity of the object in the communicative respect is due to the fact that in heterosubjective sentences it is quite normal for both thematic and rhematic objects to follow immediately after P. Examples of thematic objects in the form of O have already been given above. Below a couple of rhematic objects: (57) Oc)naatcObt n n o n p o c u : t a ~ a O e M u t c a C M u p n o a a n p u n n m b Menn. (Fpan., 229) (58) flea, oce:t u Aucutla oxomu:mcb a.~tecme. Y n u x 6bt:ta 6o~butan Oo6btua. flea s e n e A o c : t y c)eJtumb O06bl~ty. Oce:t pasOe:tu:t noposny na mpu ~tacmu. flea paccepOu:tcn, cr,eA oc:ta u se:te:t n u c u q e O e : t u m b . (From a Russian reader)

Thus RPP is not normally resorted to in heterosubjective sentences, if the rhematic status of the object is clearly shown by the semantic structure of the sentence or by the context. ~8 Only regarding extended rhematic objects RPP is the normal case: (59) Caem Mecm4a o3apnA tcoMnamy, 3 a c m a s J t n n s b t c m y n a m b u3 me~nbtx

y e J t o s e e ec)e x o ~ c m ,

eOe e u n c o e y J o p y ~ y ,

eOe

o c m a a J t e n n y t o n a c m y ~ e O p a n u p o s K y , eOe n a n m a A o n b t u n e ~ t u u l e n n m e c a n o e u . (For. II, 146)

(56) can be compared with the following sentence, where PP occurs: (60) CeJtuo6an n o , t o e s s n e 3 r n b O e s u o n t c e na KOSnbt. (For. I, 83) Differently from (56) the object is here most likely to be interpreted as context-dependent.19 It thus gets a lower degree of communicative dynamism if it is moved to the right, which undoubtedly is a rare phenomenon. Possibly the interpretation of the postponed object as context-dependent is

28

DAG SVEDSTEDT

due to analogy with the utterances in 0.2 and 0.3 : as was pointed out here an interrhematic theme between P or I and F is very common. A couple of examples in which the object is another NP category: (61) On 3acmaaua a c e x a e a ~ c a r n b n a ~eMae. (62) Fostoc, Komopbt~t on n p o u 3 n e c ace 9 m o , 3at<:uo~taa a ce6e ~tmo-mo nompaca~oulee,

maK

~tmo 3 a c m a a u : t

actcpuKnymb

acex

s

(For. I, 287) In (61) the object is probably seen as constituting rheme alone. The utterance then gives the information that the persons he made lie on the ground included everyone. In (62), on the other hand, the object is unambiguously thematic, and the aspect of totality recedes into the background. The utterance informs us rather that with his voice he made those present exclaim simultaneously. oOno s p e • a :

...

3. From a normative point of view the most controversial form of PP seems to be TPP in heterosubjective sentences, and we shall therefore confine ourselves to discussing only this phenomenon. 3.1. The answer to the question how TPP is to be seen from a normative point of view will depend on the meaning given to the word 'norm'. In linguistic literature this term is sometimes used to express the linguist's view of the linguistic ideal, in other words what o u g h t to be said or written. Such a view, however well argued it may be, is of course always subjective. In her books about word order in Russian Sirotinina (1965, 1974a) gives her view of TPP in heterosubjective sentences, stating that it is not in accordance with the "grammatical norm" and is thus unacceptable (o~ibo~en). One of her examples of a sentence with unacceptable word order (1966, 160) is as follows: (63) X a p a t c m e p u 3 y a o m ~ a H a a ~ a B:tacoaa, Fopblcuf~ d a e m n o n ~ m b ham

cyutecmaytoulue

yc:toaua

ceMeftnoeo

6btma

a pa6ouux

ceMb~x.

'Norm' is also used to denote that a linguistic occurrence is productive, i.e. the norm defines what is said or written. The statement that a certain manner of expression belongs to the literary norm is usually made with a reference to authoritative literary sources. The norm does not then, in principle, express the linguist's standpoint and is, in this respect, objective. If the term is used in this sense, TPP must, against the material presented above, be seen as a linear variety falling within the bounds of the literary norm. In (47) and (48) above Prigvin makes use of two norm varieties, possibly to avoid repetition.

THEME POSTPONEMENT

IN R U S S I A N

29

TPP is thus a possibility within the language system (langue). But if we consider to what extent this variety is actually realised in practice (parole), it is obvious that it is utilised to a very limited extent. Often TPP involves so many complications in the syntactic and the communicative respect (syntactic homonymy perhaps resulting in ambiguity) that it will not be used in written (premeditated) production. Stylistic factors (if in the manner of Gvozdev (1965, 394) one of the most important functions of stylistics is seen to be the fight against syntactic homonymy) or "kul'tura re6i" (if by this is meant the utilisation of the optimal linguistic means for a given linguistic situation, in the manner of Ickovi~ (1970, 9)) strongly counteract the applicability of TPP. 3.2. In his evaluation of linguistic performance (mostly in the written form) the language teacher acts as a normator. He must take a stand as to whether the use of a less common norm variety is acceptable in the given situation. What, then, is the attitude of such a normator to TPP? In what circumstances can it be accepted, and how high is his level of tolerance under different conditions? In order to get an answer to this question a questionnaire was sent to 26 informants, all of whom were Soviet citizens and had Russian as their mother tongue and who were working as professors and lecturers in Russian at Finnish universities and language institutes in 1979. Twenty-three of the persons returned the questionnaire, a couple of them not having answered all questions. Part of the questionnaire consisted of a collection of 36 heterosubjective sentences with TPP of O. Some of these were taken from the literary material described above, others were chosen from the same material and originally contained no TPP, but in which we postponed O (designated by /--*/ below). Some of them, finally, were chosen from the material of spoken language mentioned in the introduction. The informants were told about this but were not informed as to which sentences belonged to which category. The task given to the informants was expressed roughly as follows: " I f in your capacity as a teacher of the mother tongue and thus normator you had to evaluate the word order in the sentences below as examples of written production, taking into account their communicative function and considering the excerpts out of a stylistic point of view, in which of them would you find: (a) that the word order is unacceptable and consequently in need of correcting (nedopustimye predlo~enija), (b) that the word order is acceptable and thus not in need of correcting

(dopustimye predlo~enija),

30

DAG SVEDSTEDT

(c) that the word order in no way strikes you as odd (vpolne dopustimye

predlo~enija). Some of the sentences presented above were included in the questionnaire. In this case the number of informants that chose the alternatives (a), (b) and (c) has been given in slants (/ /). In this paper only the most salient results of the inquiry will be dealt with. By way of introduction we could mention the fact that the informants generally found the task very difficult. The following excerpt is representative of the letters that frequently were enclosed with the replies to the questionnaires: "qTO KacaeTc~ co~epxxaHrm Moero '~toMatuHero 3a2IaHri~', TO ~OJDKeHnpH3HaTbCfl, tlTO OHO Kpa~He Tpyllaoe. HeT HHKaKOFOCOMHeHH~! B TOM, HTO B 3TOE O6o/acxn HeT He TOJIbKO CJ~O)KttBtne~cn HOpMbI, HO ~a~e CKOJ~bKO-Hn6y~b OTtteT.rlHBO BbIpa)reHHO~ Ten~eHtlriH. AHKeTy a c T a p a h c a - HaCKO.rlbKO :)TO BO3MO)KHO -- cHaqaJ~a 3aHOYlHHTb coBeptueHr[O HHTyHTHBHO; HOTOM n o n b i T a n c a n p o a H a o ] n 3 n p o B a T b CO6CTBeHHbIe petueHHa n CaM yBI4~e.rl, HTO OHH Henoc.rle~oBaTe.rlbHbi."

In our evaluation of the replies the main point of consideration is the ratio between the replies according to alternative (a) on the one hand, and those according to alternatives (b) and (c) on the other. It seems very difficult to analyse the motives for choosing one or the other of alternatives (b) and (c). In several of the replies according to (b) it is made clear that the informants accept TPP, even though they would not consider using it in the current sentences themselves. The choice of (c) can then be interpreted to mean that the informant would himself consider using T P P in the current cases. The attitude of the informants to T P P varies considerably. The informant with the most negative attitude rejects 24 sentences and accepts 12, while the one with the most positive attitude rejects 3 and accepts 33. This would seem to reflect a divergent attitude on behalf of the two persons as to their role as normators/informants. The former sees as his main task to establish what ought to be written, while the latter engages in analysing the sentences from the point of view of what may be written. The average is found nearer the positive limit value: 11 sentences are rejected while 25 are accepted. Only 6 of the 36 sentences are judged in a similar way by all informants, in so far that 2 are rejected by everybody while 4 are accepted. 3.21. In the 4 sentences which are accepted by all, P is a desemanticised verb. One of our constructed sentences turned out to be the one appreciated most:

THEME POSTPONEMENT IN RUSSIAN (64)

31

C n y c m a HecKom, t¢o :tem ~ o n a m b s c m p e m u ~ w a e M o c K a e c B. O n O a :~ n o ~tu m a m ~, ~t n e Katcue-mo coou pacctca3bt, - ortu 6t,t:tu n:toxue, o tleM Yt eMy cl¢.a3aA, rio oH ace-maRu He nocAyuta.~ca Mena u socope

crna~ npo~eccuona:lbHbt~t

c/ty~taflnoM 3apa6omtce...

a~cypna:mcmoM,

(~az., 52) / ~ /

na

/0/7/16/

In general there is a clear inclination to accept TPP when P is, or may be interpreted as desemanticised: ...aan no~nvaxb Mae... ...~aaaa no~yBcxBosaxb HM... ...3aCTaBHTb B~CKOHHTbero... ...aasa~n o6eaaTb eMy... ...3aCTaBHTb pa6OTaTb ee...

(64) (34) (52) (29) (47)

/0/ 7/161 /0/ 8115/ /0/12/!1/ 13/ 8t12/ /3/11/ 9I

In the following sentence there are three verbs following one after the other: (65) ,~eO CU:tbHO cocmapu:tc~, u MyY,c c afcenoft c m a : l u c ) a a a m b o 6 e c ) a m b e m y 3a ne~Koft. (Sp.) /6/13/4/ A comparison with the similar sentence (29) and with (23) shows that the attitude of the informants to TPP in these conditions is more negative: ...aaBa~n o6eaaxb eMy... ...cxa:In aaBaxb o6e~axb eMy... ...MO~eT ~axl, non~xb HAM...

(29) (65) (23)

/3/ 8/12/ /6/13/ 4/ /5/11/ 7/

The reason for the somewhat more negative attitude is probably the sentence rhythm which is caused by TPP. Vide the discussion about this in connection with (35). 3.22. As mentioned earlier, a characteristic feature of TPP postponed component is not in final position. In fact, only one a postponed thematic object in this position has been found in material : (66) - A 30ect,, s amour Oome, - npoOo:ly, cazta H a p a o c o a b a

is that the example of the literary Hempoana,

- u a m o e e o y~enutca, nuonepa, cu:~oft naOedu tcpecm, cu:toft 3acmaa:tatom mO:lUmbc~.,. say, cno,

cu:~oft

[-...] BeOb saM, ~:atc ~ nonumalo, He c y m b

Au 3 a c m a a u A u

MoAumbca

pe6enKa

u:tu

(TeHap., 130) Here the communicative structure of the sentence is completely unambiguous. Both P and the infinitive phrase are context-dependent. The postponement cannot possibly be seen as a case of RPP. We constructed an analogous example with O: y6eOu:m e e o e 9moeg.

32

DAG SVEDSTEDT (67)

ToJtbKo noO y m p o KaMennbtft con eAOMUa eeo. KoeOa concern paccseJto, soKpye Oepeea, noO KornopbtM

on cnaa, on ysuOea

ne.aKue, Kpya~cesnbte cAeC)bt JtUCbUX Man, u Meatc n u n u auOneAca na cneey npoOoAeosambtft caeOoK aoJto~tuauteeoca xsocma. TaK aom Krno n e O a s a a c n a m b e m y : (Floa., 2 3 3 ) / ~ /

/1/10/12/ As can be seen there is a very positive attitude towards this sentence. TPP to final position is thus accepted when it does not result in any ambiguity as regards communicative structure. (68) - 51 ne xo~y eemb 9moeo cyna: - cKa~aJta Oeao~tKa, ~coeOa n a M a O a 3 a o 6 e O a m b eft. (Sp.)/---,/ /10/11/2/ The attitude to (68) is a lot more negative. The communicative function of O is here unclear: the sentence may in this limited context be interpreted as having RPP. The fact that so many informants still accept the sentence is surprising, but may perhaps find its explanation in the following circumstances: Firstly, the informants have been instructed to see the sentence as having TPP 2° and they are thus aware of its prosody. Consequently, they have the same information as a listener in analysing its communicative structure, and do not then have to take the position of a reader. There is a risk that they will not discover that the communicative structure of the sentence is ambiguous, when only the context is taken into consideration. Secondly, we cannot exclude the possibility that one or a couple of informants have not been thinking of the instructions when evaluating the sentence, but have interpreted it as having RPP. 3.23. In heterosubjective sentences Sj is the performer of the action denoted by I, while Si does not participate in the action. However, sentences with the verb p o m o S ( p o m o g a t ' ) as P are a special case in this respect: (69) l~pam nomoe ~me nocmpoumb 0o•. In this sentence both Si and St take part in the. performing of the action denoted by I. A decision as to which of these is the main performer of the action is primarily made on pragmatic grounds, and is in most cases not possible to make, as is the case in (69). However, consider the following sentences: M a M a nonoeAa Bane 3asa3amb mnyp~u (ha 6omunKax). (70a) M a M a 3aaa3a~a Bane umyptcu. (70b) B a n a n o n o e n a n e npueomoeumb o6eO. (71a) B a n a npueomoauJt m a n e o6eO. (71b) In both (70a) and (70b) Si would on pragmatic grounds be seen as the main performer of the action denoted by I. In other words, these sentences have a similar semantic content and could conceivably be used in the same

THEME POSTPONEMENT IN RUSSIAN

33

situation (even though the meaning is different in so far that in (70a) Sj must be assumed to take at least some part in the action, which in (70b) is excluded altogether). (71a) and (71b), on the other hand, are very different in meaning and would certainly not be used in the same situation, unless jokingly. The questionnaires show that the informants to a large extent accept TPP in sentences of the type shown in (70a). As regards their meaning these sentences ressemble sentences of the type shown in (70b), in which the object is subordinate to another verb: (72) Ho-Moe•y, malcue 3u~tbt 6bt:tu :ty~tutu~t epeMenem eoOa a EamyMe. Hya~cno 6btao mOAbl¢O Oocmamb nenpomo•aeMbtft n~au4 u npo~mbte 6ymcbt. Hupt¢ n o M o e O o 6 b t m b H u e 9mo OoateOesoe crtapaa~ceHue.

(Hayc-r., 453)/---,/ (73)

/1/8/11/

Hec~coAbt¢oOnef~, tcomopble Mbt ablHyatcOenno npoae:tu a Totcuo, moAbKO Oo6asu~u K MOUM sne,am:teuuaM, 6oaee noAnbtM U tuupotcuM, omuacmu 3pume/Ibllbte ene~tam:tenua, om~tacmu 8nenam:tenua om necKoabtcux 6eceO u scmpe,~, t¢omopbte n o M o e ycmpoumb na~t npeOcmasume:tb A H H e £InoHuu Muxau:t Eo6uMoa, Mo:toOoft ma:taHmausbff~ a~cypnaaucrn. (Coqbp., 249)

/4/6/10/ (74)

To6apuul

~e~mop,

no~uoeume

y6eOumb

rune

moaapuufa

J/utam~oaa, ~mo cmuxu - 9too Oyxoanoe npoRa~eHue 4e:to6e~a, ma~: cKa3amb, at,tpa:~cenue eeo dytuu... (Temlp., 432) /5/9/9/ If a transformation of type (70a) ~ (70b) is performed here (disregarding that it is not felicitous from a stylistic point of view), it would not result in any radical change of meaning. In other words, it may be claimed that P is semantically reduced in these sentences. The situation is quite different in the following sentence: (75) 51 n/tatca:ta HaO nucbMamu, ~tyacmay~, ~tmo 6t,tmt, "u3aeonnoft O6ueypoft" - He s Moux cu:tax. Hulcmo He n o . ~ o e a : t o m a e ~ t a m b rune Ha ~moeo,~uc.aenm,te nuct,Ma. Ce~pemapuau y meua ne 6blaO. (An~., 2 7 8 ) / ~ /

/9/7/5/ In the same kind of transformation this sentence gets an entirely different meaning. Under these circumstances the informants find it more difficult to accept TPP: the syntactic juncture arising between O and I "offends" much more than the juncture in (72)-(74). 3.24. As shown above the sentences containing pomo?' (pomogat') as P are not quite clear as to the role of Si and that of Sj in the performing of the action expressed by I. However, here O can hardly be interpreted

34

DAG SVEDSTEDT

as being subordinate to I. This would imply that Sj is unspecified, an interpretation that seems far-fetched not only in the above sentences, but generally in heterosubjective sentences with this finite verb. In the case of other finite verbs such an interpretation is quite possible, if there is syntactic homonymy in the sentence. If pragmatic factors do not eliminate one of the alternative possibilities of interpretation/the sentence will be ambiguous. (76) ToeOa o m e ~ a e ~ e : t u • n p u n e c r n u e e n u r . (Sp.) (77) ToeOa o m e q e e J t e ~ n p u n e c r n u u M eenuK. (Sp.) The interpretation of the syntactic structure of these context-free sentences will probably depend on the linear position of O. In (76) O is seen as subordinate to P and thus as being Sj. In (77), on the other hand, St is interpreted as being unspecified, and O is seen as being an object in the embedded infinitive clause. In other words, (77) is probably not interpreted as having TPP. (78) O m e ~ aceeOa eoaopu:t c b m o a b a M , ~tmo a~cumb nya~cno e .~upe u Opya~c6e. OOnaa~cbbt on e e ~ e ~ ee:te:t CAoMamb n o u e c~oMa~u.

(79)

npunecmu

cbmoeba~t

eenu~c, u

UM 9 m o m aenutc. CblHOSb~q ~oMaetu, ~oMoAu,

(Sp.)

/9/8/6/

Hy, na 9mo cbmoaba nepeeJtany~ucb u ym,t6ny~ucb. Toeba ome~ npeO/toatcun

cTtoMamb

uM

nyttoK cmpeJt. CblHOabYt HO o m -

(Sp.) /11/10/2/ In (78) and (79) the sentences are ambiguous. Only through the subsequent context is the ambiguity eliminated. This would seem to be the factor that causes the comparatively negative attitude of the informants. Even so their attitude to (78) is more positive than their attitude to (79). This may be explained by the fact that P is context-dependent in (78), which makes TPP more justified in the communicative/prosodic respect. (80) O m e ~ a c e e b a e o a o p u n caou~t c b m o e b a M , ,¢mo6bt o n u a~cu~u d e ~ b n o c m u AeeKo n e p e ~ o ~ t a ~ u ace c m p e m , L

Opya~cno u n e c c o p w l u c b , n o o n u n e c.aytuaJtuch eeo. Oc)naatcObt on n o n p o c u / l

npunecmu

u x aenut<. Koec)a o n u e e o npunec.au,

(Sp.) /19/1/1/ The sentence above is also ambiguous. There is a syntactic juncture between O and F, whereby O is interpreted as a possessive pronoun. As can be seen, the informants' attitude to the sentence is very negative. on cKa3aa u M : " C ~ o ~ t a f t m e e e o / "

3.25. In the cases so far presented I has neither been context-dependent nor desemanticised, and thus TPP has been justified from a communicative point of view. In the following sentences, however, this is not the case: (81) B p o c b m e - K a e e o Htotu~ce. H o M o : ~ c e m :tu on e~ ebtnJlblmb? H a s p a O .au. A p e J t u e u a eft e o s o p u : t a : " H e epex!"

H H t o u a c a aepuJta. P e n u e u a - m o

npaecmsenno~.

(Ten~p., 470)

npeJuo6odeftcmayft -

noMoeaJta

cmamb

eft

/7/8/8/

THEME

(82)

POSTPONEMENT

IN R U S S I A N

35

Y m p o M so emopnut¢ •bt cnoea no~eonu:tu s Ou:mnnuncKoe noco:tbcmeo. Ce•pemapb noc:ta npon:ta HaM no me:teo6ony me:teepaM~ty AmaOy IOsona, o tcomopof~ on n p o c u ~ o6oa~cOamb n a c eute cymtcu. ToeOa M~,tpetuu~u noauOambca c 06uAunnunctcuM noc~oM. Ho~eonu~u. (Coqbp., 247) /12/5/6/

In (81) a prosodic marking of the desemanticised infinitive cannot be justified on communicative grounds. The sentence rhythm caused by TPP does not reflect the communicative structure of the sentence. In spite of this the informants' attitude to (81) is relatively positive. In (82) the semantic structure of the sentence and the surrounding context indicate that I is context-dependent (which is in fact the case if the macrocontext is taken into account) and so should not be prosodically marked. The attitude of the informants to this sentence is more negative than their attitude to the previous one, probably because of an additional complication in the form of syntactic homonymy. 3.26. Finally one of the two sentences rejected by all the informants: (83) flea cna3, u •aAenbtca,q MbtUttCa no6eatca~a no neMy. Jlee npocny~wa u nofi~taA ee. ToeOa MblUtKa n o n p o c u : l a

omnycmumb

e eo. (Sp.) /23/0/0/ What mainly distinguishes this sentence from those presented earlier is the fact that here I is a transitive verb, the direct object of which has been deleted. It is true that the object may be implied in the given context (Togda m y g k a poprosila ego otpustit'), but such a manner of expression would differ from what is expected. Through the expected complementation of I in the form of a direct object O is spontaneously interpreted as being subordinate to I, and in the given context the sentence is then incomprehensible. This would seem to be the main reason why the sentence is rejected by all the informants. It has been found in the spoken material, and is obviously a case of pure lapse (ego pro ee). 3.3. To summarise our findings briefly: The replies to the questionnaires reveal the informants' uncertainty as to how they as normators should react to TPP in heterosubjective sentences. Only if TPP is semantically motivated, i.e. if P is (or can be interpreted as being) desemanticised and thus semantically integrated with the governed verb, is there a relatively wide unanimity that TPP ought to be accepted. But even in these sentences the acceptance is given with some reluctance: many of the informants do not want to consider the sentences totally acceptable (vpolne dopustimye).

As soon as TPP has some form of negative side-effects on different levels, the informants become somewhat less inclined to accept the phenomenon:

36

DAG SVEDSTEDT

if TPP affects the sentence rhythm in a negative way (e.g. (23) and (65)); if such syntactic homonymy occurs which results in uncertainty about the syntactic structure of the sentence (e.g. (78) and (79)); if TPP does not seem to be justified from a communicative function point of view (e.g. (81) and (82)) or if through PP the communicative function of O becomes unclear (e.g. (68)). But at the same time it must be pointed out that the majority of the informants yet accept TPP under these conditions. This is to be understood in the sense that this majority considers the occurrence as a word order variety belonging to the literary norm, and as such it has to be accepted (albeit somewhat reluctantly), even when it has negative effects in various respects. TPP is totally (or practically totally) unacceptable only in sentences with syntactic homonymy, in which the homonymy results in so strong a syntactic juncture between O and some component other than P, that the sentence gets a different syntactic interpretation and meaning, which is not possible in the context (e.g. (80) and (83)). 4. TPP occurs mainly in spontaneous spoken language and arises by analogy with sentences with a less complex structure. In the great majority of cases these simpler sentences have the following characteristics. On the syntactic/linear level: simple or compound predicate-object governed by the finite verb (in compoundlpredicates by the infinitive)-final component; on the communicative/linear level: rheme (theme)-theme-rheme, and on the prosodic/linear level: stressed-unstressed-stressed component. Sentences with TPP have the same characteristics in the communicative/linear and prosodic/linear respect, but differ from simple sentences in the syntactic/ linear respect. In other words, TPP is an analogy on the communicative/ prosodic level. TPP presupposes a high degree of givenness of the postponed component, and therefore this component is nearly always a personal pronoun. The phenomenon may thus primarily be seen as a form of clitic movement. It is curious in so far that it occurs to the right instead of to the left. In written language TPP occurs fairly seldom. In this form of linguistic production (always premeditated, whether it is written language or stylised spoken language) syntactic/linear considerations predominate. Our investigation shows that in the cases where TPP does occur, the choice of this word order is caused by communicative considerations, i.e. it is used in order to make the immediate comprehension of the communicative structure of the sentence easier for the reader. The postponed component acts as an accentuator and organises the sentence in the communicative/prosodic respect. It indicates that the infinitive has a higher degree of communicative

THEME

POSTPONEMENT

IN RUSSIAN

37

dynamism than the component preceding the infinitive, and that the infinitive is the component to be stressed, if the sentence is spoken. As a phenomenon in the communicative system TPP is undoubtedly an odd one. In most types of sentences the object gets a higher degree of communicative dynamism if it is placed further to the right than immediately after the governing verb. In heterosubjective sentences with TPP, however, it gets a lower degree. The question how TPP in heterosubjective sentences is to be considered from the normative point of view is controversial. The normators/informants do not consider these sentences wrong in the sense that they need correcting (provided that TPP does not have any serious negative consequences), and yet many of them do not want to accept them as being entirely correct.

Svenska Handelshfgskolan, Helsingfors

NOTES 1 In the present study the description applies to the syntactic and the communicative level at the same time. The term s e n t e n c e is used primarily in the sense of syntactic unit. The term u t t e r a n c e is used primarily in the sense of communicative unit having certain prosodic characteristics. If the unit represented by these terms is looked upon in both the syntactic and the communicative respect, the terms are used interchangeably. 2 The figures in slants may be ignored for the time being. See below under 3.2. 3 For a more detailed discussion, vide Svedstedt (1979, 196). 4 For a more detailed discussion, vide Svedstedt (1976, 87 if). 5 Utterances of this type with T P P of O have only been found in the spoken material, e.g. Odna~dy ochotniki pojmali l'va i p r i v j a z a l i verevkoj ego k derevu. It is obvious that the tendency to link O enclitically to a c o m p o n e n t with anticadence is what causes utterances of this type. 6 Kubik (1971, 72) calls heterosubjective sentences "konstrukcii s geterosub"ektnym infinitivom". 7 Only one exception has been found in the material. See below under 3.22. s Only single instances of I with cadence occur, e.g. U~ vy p o z v o l ' t e , va~e prevoschod i t e l ' s t v o , p o c e l o v a t ' mne ego. (Gog. I, 248) 9 By c o n t e x t - d e p e n d e n t is meant both given and thematic. A c o n t e x t - i n d e p e n d e n t c o m p o n e n t m a y be either thematic or rhematic. 10 The term occurs in Pumpjanskij (1974, 73), who uses it in the sense of 'rheme-signalling word'. 11 Here the infinitive phrase is the theme of the utterance according to Kovtunova, who gives this example (1976, 99). In other words I is here context-independent and thematic. ~2 On the other hand dat'-give as well as dat'-permit m a y alone constitute rheme. Cf. the semantically a m b i g u o u s utterance Ona ne dali~ emu est" (Est" ona emu ne dala), which can be interpreted as either 'She did not give him to eat' or 'She did not let him eat'. ~3 The fact that dat' (davat') and zastavit" (zastavljat') in one of their meanings mainly contain the semantic feature of C A U S does not imply that they are freely interchangeable. Their main difference in use seems to be that with the former verb the activity of St comes into the foreground, while with the latter it is the activity of Si. la Here both pragmatic and semantic factors influence the interpretation. Zastavit'-cause must be the meaning of the verb when it is obvious that S t wants to perform the action

38

DAG SVEDSTEDT

expressed by I, e.g. ~to neo[idannoe dosti~enie zastavilo ego zabyt" vse trevogi (Pol., 161), or if the modality of 'volition' has no relevance to the action, e.g. Znakomye zapachi benzina, masla, radostnoe ogkug~enie poleta zastavili ego poterjat' soznanie. (Pol., 71). Inanimate Si or Sj also influence the interpretation in the same direction (cf. (47) and (48)). is Here pragmatic factors play a part. In (49) both S~ and Sj are animate. The relation between the two is hostile, and zastavit'-force must be the meaning of the verb. Communicative factors also influence the interpretation. In (50) P is the rheme of the utterance. The interpretation zastavit'-force is then the only possible one, since zastavit'-cause does not alone occur as rheme. 16 E.g. Vdrug kto-to zastavilo ego obernut'sja (Ok., 330). Cf. also examples in Note 14. 17 Cf. the French verbs with similar meanings, in which semantic integration is explicitly expressed in the structure: laisser faire qc ~ qn (vis-h-vis permettre ~ qn de faire qc) and faire faire qc a qn (vis-h-vis forcer qn de faire qc). The fact that TPP in sentences with desemanticised P occurs fairly frequently in the only representative of the 19th century in the material, Gogol, may be an indication of contemporary influence from French. When P is desemanticised in French, P and I will be in contact position in the same way as when TPP occurs in Russian. 18 Kovtunova (1976, 77) makes the same observation when discussing context-dependent components: "Jazyk izbegaet naru~enija ierarchii sintaksi~eskogo ~lenenija, esli v 6tom net ostroj neobchodimosti, t.e. esli prinadle~nost' k teme zavisimoj_ slovoformy, obozna~aju~ej dannoe, dostatoi~no ~etko podskazyvaetsja kontekstom". 19 When discussing word order in Czech, Biij, (1977) states in accordance with Sgall (1973): "All complements to the right of the verb, except for the first one, which, without a known context, is ambiguous from the FSP point of view (cf. (56), D.S.), are context-independent". In (60), however, the postponed object is undoubtedly seen as context-dependent. 20 In the instructions it was explicitly pointed out that O is not stressed but rather slurred ("proglatyvaetsja") in all the utterances.

BIBLIOGRAPHY Adamec, P. : 1966, Porjadok slovv sovremennom russkomjazyke ( = Rozpravy (~eskoslovensk6 akademie v~d, l~ada spole~ensk~'ch v~d, ro~:nlk 76, se~it 15), Praha. Bil~', M. : 1977, Is there any neutral word-order? (Based and tested on Czech material), unpublished paper read at the 8th Nordic Meeting of Slavists, Askov, Denmark. Bogusiawski, A. : 1977, Problems of the thematic-thematic structure of sentences, Pafistwowe Wydawnictwo Naukowe, Warszawa. Gvozdev, A. N. : 1965, O~erki po stilistike russkogojazyka, Prosve~,enie, Moskva. Ickovi~, V. A. : 1970, ' N o r m a i ee kodifikacija', Aktual'nye problemy kul'tury reki, Nauka, Moskva, pp. 9-39. Kovtunova, I. I.: 1976, Sovremennyj russkij jazyk. Porjadok slov i aktual'noe klenenie predlo~enija, P r o s v e ~ n i e , Moskva. Kubik, M. : 1971, Transformacionnyj sintaksis russkogo jazyka, Stb,tni pedagogick~ nakladatelstvl, Praha. Lapteva, O. A. : 1976, Russki] razgovornyj sintaksis, Nauka, Moskva. Pumpjanskij, A. L. : 1974, Informacionnaja rol" porjadka slovv nauknoj i technikeskoj literature, Nauka, Moskva. Sgall, P. : 1973, 'Kontextov~ zapojeni a otfizkovfi metoda', Slovo a slovesnost 34, 202-211. Sirotinina, O. B.: 1965, Porjadok slov v russkom jazyke, Izd. Saratovskogo universiteta, Saratov. Sirotinina, O. B.: 1974a, Sovremennaja razgovornaja re~" i e e osobennosti, Prosve~enie, Moskva. Sirotinina, O. B. : 1974b, 'Ob aktual'nom i~lenenii v razgovornoj re~i', Papers on functional sentence perspective (ed. by F. Dane~), Prague, pp. 171-174. Svedstedt, D. : 1976, Position of objective personal pronouns. A study of word order in modern Russian ( = Stockholm Slavic Studies 9). Svedstedt, D. : 1979, 'On word order options in Russian', Scando-Slavica 25, 193-208.

THEME

POSTPONEMENT

IN RUSSIAN

39

The literary examples have been taken from the following sources: AKC.

Ann. BOH~.

Fan. Fn.

For. I For. II FpaH. Fp. Koptu. Herp. nilK.

= = = = = =

OK.

=

Has. I Hall. II HaycT. HflaT. liO.rl. lip. I lip. II Pacc.

= = = = = = = =

=

CHM.

Con. Co. Coqbp. TeHap. TIIT. q~aa. q~p.

= = = = = = =

ArceHoa, B. H., )Ka~b, umo aac ne 6btno c ua~u. M o c z s a 1969. AanmayeBa, C., To~b~co oc)un coO. Hbro-flopK 1969. l~oruIapeB, IO. B., Fopn,tu(~ cnee. MocKaa 1974. Fanntlrnfi, A. B., 'Pa3Mumnenila o6 o~nofi n3 npoqbecca~'. Hepaoe ucnblmauue. MocrBa 1974. F ~ a a r o a , B. cP., Bonbuu~a. M o c r a a 1971. Foroab, H. B., Mepma~ae Oyuau. MocKaa 1961. Foron~, H. B., Hemep6ypecrue noaecmu. Moczaa 1965. FpaHnH, ~ . A., HeoaJcuc)annoe ympo. 51ennHrpaa 1970. FpetuHOS, M. H., Bo~ute6u~af~ Ko/toOett. Moczaa 1974. Koptuynoa, E. A., Onepatfun 'Xame~eou'. MocKaa 1971. Hezpacoa, B. H., B a~cu~nu u a nucbMax. M o c r a a 1971. HnrnTIIH, H. H., Ceaepnan aepopa. 5IenilHrpaa 1971. Ozynx~aBa, B. HI., F~omoK cao6o&,t. M o c r a a 1971. liaHoaa, B. qb, Hosecmu. JleHilHrpaa 1969. HaHoaa, B. qb., 3aMemKu .~umepamopa. 5Ienilnrpa~ 1972. FlayCTOBCrii~, K. F., tI36paunoe. M o c r a a 1961. HnaTOHOa, A. H., CMepmu uem. r M o c r a a 1970. HoneBo~, B. H., llo~ecmb o uacmo.,~uteM ,~eno6eKe. M o c r a a 1964. HpiltUanH, M. M., K~mOoaan co:tmla. MocKaa 1970. HpmuanH, M. M., OcyOapeaa Oopoea. FIeTpo3aBo~cK 1970. PaccynoBa, O. H., Ynompe6:teuue auboa a pyccroM nz~atce. M o c r a a 1968. CIIMOHOB, K. M., .,aKua~,te u Mepmabte. MocrBa 1965. CoaoBbea, B. A., KoeOa yxoOam eo&,t, lieTpo3aaoacr 1972. ConoyxnH, B. A., Kyuyturuu c~au. M o c r a a 1969. CodpponoB, A. B., ,lIopoeu. M o c r a a 1972. TeH~paroa, B. ~ . , ttpe3nbt,~aftuoe. M o c r a a 1972. TilTOa, B., 'BceM cMepXaM na3~o', llepaoe ucn~mauue. M o c r a a 1974. dPa~eeB, A. A., Hosecm~, uau~eft muocmu. Moczaa 1968. OpOnOB, 2I. A., Ho:te:~,caeacrue neo&,t. MoczBa 1972.