converbs in translation : the role of aktionsart in the interpretation and translation of russian converb constructions into english and norwegian

My ongoing investigation of the role of Aktionsart in the interpretation and translation of Russian sentences with converb constructions (cf. (Krave in progress)) suggests that the lexical aspect of a particular converb construction in the source data (Russian) may trigger a specific interpretation restricting the range of possible translation patterns found in the target data (i.e. English and Norwegian). For example, it was found that coordinate clauses (e.g. VP coordination – by means of the conjunction ‘og’ (‘and’) linking the main event and the event expressed by the converb) occur most frequently in the Norwegian translations of the ‘achievement’ type of converbs (including the semelfactive lexical type). This pattern is contrasted with the preference for the Norwegian construction with ‘etter’ (‘after’) followed by either a DP or a Perfect Infinitive construction – ‘å ha’ (‘to have’) + past participle – found in translations of the delimitative Aktionsart of the converb construction [1] introduct ion Corpus research shows a relatively high frequency of converbs in Russian of the semelfactive and the delimitative Aktionsart (compared to other Aktionsarten). This article investigates Norwegian and English translations of all instances of converbs of these two semantically distinct types in our Russian-Norwegian-English corpus (the RuN corpus)1. First of all, the aim of this detailed analysis is to provide convincing evidence that the differences in the lexical semantics of converbs of these two types do affect the translator’s interpretation of the converb [1] Examples from the corpus are presented in the following way: on the first line is the original Russian sentence (source text), then the Norwegian or/and the English authorized translation (target text). [156] maria filiouchkina krave sentences in the source text and restrict the choice of the corresponding constructions in the target text. The second aim of this analysis is to compare the Norwegian and English translations of the same Russian constructions. Differences between the target languages can shed new light on language-specific features in English and Norwegian. The article is structured as follows: after a brief data description in section [2], section [3] compares the Norwegian translations of the semelfactive [3.1] vs. the delimitative [3.2] type of Russian converbs, while section [4] compares the two types based on the English target data. Section [5] discusses similarities and differences between the English and Norwegian translation patterns. In section [6] we discuss the differences in the lexical aspect of delimitative vs. semelfactive converbs that underlie the different translation patterns. Summary and conclusions are presented in section [7]. [2] the data Following the classification of Russian verbs in (Zaliznjak & Šmelev 2000), the semelfactive lexical type of converbs (expressing a single occurrence of events) is marked in Russian by the suffix -nu-. A search for converbs with this formalmarking in combinationwith the converb suffix -v (PF form) in the Russian-NorwegianEnglish parallel corpus (RuN) returned 164 Russian sentences and their corresponding translations in Norwegian and English. Examples of the converbs found are: • vzgljanuv – ‘having glanced’ – 42 occurrences; • vzdohnuv – ‘having sighed’ – 10; • vspyhnuv – ‘having flushed’ – 9; • vzdrognuv – ‘having shuddered’ – 5; • povernuv – ‘having turned’ – 4; • zagnuv – ‘having bent / crooked (a finger)’ – 4; • kivnuv – ‘having nodded’ – 3; • skinuv – ‘having thrown down’ – 3; • zapahnuv – ‘having wrapped around’ – 3; • mahnuv (rukoj) – ‘having waved (with one’s hand)’ – 3; • hlopnuv – ‘having slammed’ – 2 etc. OSLa volume 2(1), 2010 converbs in translation [157] Due to the specific semantics of semelfactive converbs, which denote a single occurrence of a momentaneous event, it is often difficult to determine whether the event expressed by the converb construction temporally overlaps with the main clause event or precedes it. It is assumed here that the meaning of the converb suffix -v is ‘anteriority’. However, events expressed by converbs of the semelfactive type are often interpreted as partially simultaneous with the main clause event: e.g. in the sentence Vzdrognuv, on skazal ... – ‘Shuddering (PF), he said ... ’ – the event of shuddering may be interpreted as preceding the event of saying or as simultaneous with it. This vagueness in interpretation is arguably due to the fact that prototypical semelfactives do not have well-defined “target” states. The delimitative Aktionsart (also known in Russian aspectology as the pofective) is marked by the prefix poand denotes events of limited duration – e.g. events that last ‘for some time’. For example, poguljat’ means ‘to walk for some time’. Importantly, applied to converb constructions, themeaning of ‘po-fectivity’ (or delimitativity) has two components: (i) the precedence relation (coming from the meaning of the converb suffix -v): i.e. the converb event is ended before the start of the main clause event; and (ii) the event goes on for some (explicitly specified or contextually implied) time interval (the meaning of po-). Although in general, converbsmarked by poare numerous in the data, not all these cases can be attributed to this specific semantic interpretation (e.g. položiv – ‘having put’ is clearly not a pofective). 49 sentences with converbs interpreted as a delimitative have been found in the corpus with their corresponding translations into English and Norwegian. Some examples of such converbs are: • pomolčav – ‘having been silent for some time’ – 14 occurrences; • podumav – ‘having thought for some time’ – 9; • pogovoriv – ‘having talked for some time’ – 7; • pogljadev – ‘having looked at something for some time’ – 5; • postojav – ‘having stood for some time’ – 2; • pobyvav – ‘having been (somewhere) for some time’ – 2 etc. [3] analys i s of the norwegian translat ions [3.1] Norwegian translations of Russian sentences with semelfactive converbs The Norwegian translation patterns for 164 Russian sentences containing 69 different semelfactive converbs (lexemes) are here listed in the order of decreasing frequency: OSLa volume 2(1), 2010 [158] maria filiouchkina krave (i) Coordinate Past clauses with ‘og’ (‘and’): converb event + og / comma + main event (65 tokens, hence approx. 40 %): (1) Vzgljanuv iz okna, ona uvidala ego koljasku. (Lev Tolstoj, “Anna Karenina”) Hun så ut av vinduet og fikk øye på kalesjen hans. Glancing out of the window, she saw his carriage. (ii) Coordinate Past clauses: main event + og / comma + converb event (50 tokens, hence approx. 30%): (2) – Ja ne vovremja, kažetsja, sliškom rano, – skazal on, ogljanuv pustuju gostinuju. (Lev Tolstoj, “Anna Karenina”) “Jeg er ikke presis, ser det ut til, det er visst altfor tidlig,” sa han og så seg om i det tomme rommet. “It’s not time yet; I think I’m too early,” he said glancing round the empty drawing room. (iii) Past (main event) + med (‘with’) + nominal expression (converb event) /med + nominal expression (converb event) + Past (main event) – 9 tokens, approx. 5%: (3) ... govorila v osnovnom Lena, Sergej sidel v uglu, zakinuv nogu na nogu, i kuril ... (Viktor Pelevin, “Generation P”) ... hvor det hovedsakelig var Lena som førte ordet, mens Sergej satt med korslagte bein i et hjørne og røykte ... ... Lena did most of the talking; Sergei sat in the corner with his legs crossed, smoking ... (iv) Relative clauses in the Past tense: Past (main event) + som (relative pronoun) + Past (converb event) – 5 tokens, approx. 3%: (4) – Da, ja pišu vtoruju čast’ “Dvuh načal”, – skazal Goleniščev, vspyhnuv ot udovol’stvija pri etom voprose ... (Lev Tolstoj, “Anna Karenina”) “Ja, jeg holder på med annen del av “De to prinsipper”,” sa Golenisjtsjev som ble fyr og flamme i glede over dette spørsmålet ... “Yes, I’m writing the second part of the Two Elements,” said Golenishtchev, coloring with pleasure at the question ... (v) Coordinate Past Perfect (converb event) og / comma +main event in the Past form (5 tokens, approx. 3%): OSLa volume 2(1), 2010 converbs in translation [159] (5) Šurik, vsporhnuv k nemu na grud’, obnjal ego za šeju. (Ljudmila Ulitskaja, “Medea and her children”) Sjurik hadde hoppet opp og hang nå om halsen på ham. Shurik flew up to his chest and hugged him around the neck. (vi) Temporal adverbial clauses introduced by da (‘when’) in the Past (converb event), following or preceding thematrix clause in the Past tense – 5 tokens, approx. 3%: (6) Vzgljanuv v ego starčeskie milye glaza, Levin ponjal daže čto-to ješčo novoe v svojom sčastje. (Lev Tolstoj, “Anna Karenina”) Da han så inn i de kjære gammelmannsøynene hans, gikk endog en ny side ved lykken hans opp for Levin. Looking into his kindly old eyes, Levin realized even something new in his happiness. (vii) Temporal adverbial clauses introduced by da (‘when’) with Past Perfect (converb event) preceding or following a simple Past form (main event) – 5 tokens, approx. 3%: (7) – Mne žalko, čto ja rasstroil vaše ženskoe tsarstvo, – skazal on, nedovol’no ogljanuv vseh ... (Lev Tolstoj, “Anna Karenina”) “Så synd at jeg har brakt uro inn i kvinnenes verden,” sa han da han hadde sett seg rundt på alle sammen ... “I’m sorry I’ve broken in on your feminine parliament,” he said, looking round on every one discontentedly ... (viii) Temporal adverbial clauses introduced by mens (‘while’) in the Past (converb event) following or preceding the matrix caluse in the Past tense – 4 tokens, approx. 3%: (8) – A, da! – skazal on na to, čto Vronskij byl u Tverskih, i, blesnuv svoimi čornymi glazami, vzjalsja za levyj us... (Lev Tolstoj, “AnnaKarenina”) “Å, ja!” sa han til det at Vronskij hadde vært hos Tverskojs, og mens det glimtet i de sorte øynene, grep han fatt i den venstre barten ... “Ah! yes,” he said, to the announcement that Vronsky had been at the Tverskoys’; and his black eyes shining, he plucked at his lef

OSLa volume 2 (1), 2010 converbs in translation [157] Due to the specific semantics of semelfactive converbs, which denote a single occurrence of a momentaneous event, it is often difficult to determine whether the event expressed by the converb construction temporally overlaps with the main clause event or precedes it.It is assumed here that the meaning of the converb suffix -v is 'anteriority'.However, events expressed by converbs of the semelfactive type are often interpreted as partially simultaneous with the main clause event: e.g. in the sentence Vzdrognuv, on skazal … -'Shuddering (PF), he said … ' -the event of shuddering may be interpreted as preceding the event of saying or as simultaneous with it.This vagueness in interpretation is arguably due to the fact that prototypical semelfactives do not have well-defined "target" states.
The delimitative Aktionsart (also known in Russian aspectology as the pofective) is marked by the prefix po-and denotes events of limited duration -e.g.events that last 'for some time'.For example, poguljat' means 'to walk for some time'.Importantly, applied to converb constructions, the meaning of 'po-fectivity' (or delimitativity) has two components: (i) the precedence relation (coming from the meaning of the converb suffix -v): i.e. the converb event is ended before the start of the main clause event; and (ii) the event goes on for some (explicitly specified or contextually implied) time interval (the meaning of po-).
Although in general, converbs marked by po-are numerous in the data, not all these cases can be attributed to this specific semantic interpretation (e.g.položiv -'having put' is clearly not a pofective).49 sentences with converbs interpreted as a delimitative have been found in the corpus with their corresponding translations into English and Norwegian.Some examples of such converbs are: • pomolčav -'having been silent for some time' -14 occurrences; • podumav -'having thought for some time' -9; • pogovoriv -'having talked for some time' -7; • pogljadev -'having looked at something for some time' -5; • postojav -'having stood for some time' -2; • pobyvav -'having been (somewhere) for some time' -2 etc.
[3] a n a l y s i s o f t h e n o r w e g i a n t r a n s l at i o n s [3.1] Norwegian translations of Russian sentences with semelfactive converbs The Norwegian translation patterns for 164 Russian sentences containing 69 different semelfactive converbs (lexemes) are here listed in the order of decreasing frequency: [158] maria filiouchkina krave (i) Coordinate Past clauses with 'og' ('and'): converb event + og / comma + main event (65 tokens, hence approx.40 %): (1) Vzgljanuv iz okna, ona uvidala ego koljasku.
Glancing out of the window, she saw his carriage.
(ii) Coordinate Past clauses: main event + og / comma + converb event (50 tokens, hence approx.30%): (2) verb precedes the verb denoting the converb event (approx.30%) 3 -cf.( 1) and (2).As mentioned in section [2], Russian converbs of the semelfactive type are somewhat vague with respect to the temporal relation between the two eventsi.e. both relations of temporal overlap and consequentiality are possible, and it is often difficult to determine precisely whether the event described by the converb temporally overlaps or precedes the main event (as opposed to the delimitative type of converbs which always convey the precedence relation).In (1) repeated below, the event of 'looking out of the window' is from a grammatical point of view naturally understood as preceding the event of 'seeing'.(1) Vzgljanuv iz okna, ona uvidala ego koljasku.
Glancing out of the window, she saw his carriage.
However, lexically the converb and the matrix denote events which are clearly related to each other, hence there is also a flavour of overlap in the interpretation.
The coordinate construction used in the Norwegian translation seems to reflect this temporal ambiguity.In (2) repeated below, the converb is in the final position but the event expressed by it is still most naturally interpreted as preceding the main event (the 'saying' event).At the same time, it is possible to infer pragmatically that the converb event is co-temporal with the main event (the saying event happens while the event of looking around is taking place).
The comitative med ('with') followed by a DP (which is often followed by a PP, or preceded or followed by a past participle form, cf.(3) repeated below 4 ) is another alternative construction used in the Norwegian translations of the Russian [3] Note, however, that the order of events (converb event and main event) with respect to each other in the Norwegian target text does not always match their order in the source text (Russian).[4] Such "small clauses" often denote sets of states according to (Fabricius-Hansen & Haug forthcoming).Syntactically, clauses such as "med korslagte bein" (with crossed legs), where the comitative preposition is followed by a DP, are distinct from clauses of the type "with his legs crossed", where "with" is followed by a small clause (SC): a non-finite construction with an overt argument and a predicate (cf.(Fabricius-Hansen & Haug forthcoming)).
(3) … govorila v osnovnom Lena, Sergej sidel v uglu, zakinuv nogu na nogu, i kuril … (Viktor Pelevin, "Generation P") … hvor det hovedsakelig var Lena som førte ordet, mens Sergej satt med korslagte bein i et hjørne og røykte … … Lena did most of the talking; Sergei sat in the corner with his legs crossed, smoking … In the definition of semelfactives most researchers (cf.(Smith 1991)) assume that these verbs denote events which do not bring about a change, i.e. a semelfactive VP does not have a well-defined target state 5 (contrary to normal accomplishments and achievements).Example (3) seems to go against this pattern.The verb with a semelfactive suffix (-nu-) clearly denotes, in this particular construction, events which have a "target state" of the legs' being crossed.It is precisely this state, and not the preceding event /activity of crossing the legs itself, which is referred to by the construction chosen by the translators.Hence, the complex situation is interpreted as simultaneity although the "semelfactive event" in a narrow sense may still precede the matrix.
Other semelfactive converb VPs translated into Norwegian with the comitative construction (hence implying a "target state") in our data include: sognuv svoju dlinnuju spinu ('med den lange ryggen bøyet' -'his long back bent'), sognuv koleno ('med bøyet kne' -'with one knee raised'), povernuv nabok golovu ('med hodet på skakke' -'her head on one side').Note that all these examples involve transitive verbs that denote a change of position of some body parts and are usually attached to matrix verbs that have a stative reading and typically occur in the IPF form in Russian (e.g.stojal (IPF) -stood, sidel (IPF) -sat, etc.).Within the group of target state predicates, (Grønn 2004, 232) distinguishes between permanent vs. reversible target states: e.g. the target state of the book being read to shreads is permanent (irreversible), while the target state of the window being open is reversible (i.e. the window can be closed again).Noteworthy, all semelfactive converbs that are interpreted as involving a "target state" are similar to predicates which denote a reversible target state (a change in the position of some body part is usually followed by placing this part back to its initial position) and can be used with temporal adverbial modifiers ('for some time'): e.g.sognuv spinu na sekundu -'having bent the back for a second').However, most semelfactive converbs in our data do not have target states (but only consequent states) and cannot be modi- [5] The term "target state" is used here as in (Parsons 1990, 235) and (Grønn 2004) and refers to a particular, semantically visible state which is part of the lexical meaning of certain telic VPs.It should not be confused with the term resultant state (Parsons 1990, 235) -also called the consequent state (cf.(Grønn 2004, 232)) -i.e. a state that holds for any event e after its culmination.
Relative clauses with som ('which') in the Past tense form are also found in the corpus data (3%) -cf.(4) repeated below.Combined with main clauses in the Past tense, such constructions are underspecified with respect to the temporal relation between the events, but are compatible with the 'overlap' relation.
(4) -Da, ja pišu vtoruju čast' "Dvuh načal", -skazal Goleniščev, vspyhnuv ot udovol'stvija pri etom voprose … (Lev Tolstoj, "Anna Karenina") "Ja, jeg holder på med annen del av "De to prinsipper"," sa Golenisjtsjev som ble fyr og flamme i glede over dette spørsmålet … "Yes, I'm writing the second part of the Two Elements," said Golenishtchev, coloring with pleasure at the question … Further, we find constructions which explicitly encode the 'anteriority' reading of the converb event with respect to the main event: e.g.coordinate constructions with the Past Perfect form of the verb denoting the converb event and the simple Past form of the main verb (3%) -cf.( 5); and two subordinate constructions with da ('when') attested in 6% of all translations -i.e.constructions with da followed by the simple Past form of the verb denoting the converb event -( 6), and constructions with da followed by the Past Perfect form of the verb denoting the converb event and the simple Past form of the main verb -cf.(7).Constructions in the Past tense introduced by idet ('as') are found in 2% of the data -cf.(9).These are quite similar to subordinate clauses in the Past tense linked by the conjunction mens ('while'), which make the overlap relation between the converb event and the main event explicit (3%) -as in (8).The Russian converb construction in these examples is truly underspecified with respect to the temporal reference and the translator is forced to choose a construction with a more specific temporal relation than in the source text.It is in principle possible to use the 'etter' ('after') construction (referring to anteriority) in Norwegian, but our data shows such constructions are dispreferred in translations of the semelfactive type of converbs.
Other constructions include depictive adjectives (2%) -( 10) and juxtaposed sentences in the Past tense (1%) -(11).Eight other constructions have also been attested in the data, but with only one occurrence and thus will be considered peripheral compared to the patterns described above.
The 'semelfactivity' reading in the Norwegian sentences with finite verbs in the Past tense may often be reflected in the use of a DP with an indefinite (singular) article -e.g.kastet et blikk ('cast a (one) glance'), through the combination of a verb with a particle -e.g.'hadde hoppet opp' ('had jumped up') -as in (5).(Lev Tolstoj, "Anna Karenina") "Ja, hva var det så vi snakket om?" sa Levin efter et øyeblikks taushet.
"Yes, what were we talking about?"Levin said, after a pause.
"How can new conditions be found?" said Sviazhsky.Having eaten some junket and lighted a cigarette, he came back to the discussion.
The analysis of the data shows that Norwegian translations of delimitative converbs exhibit preference for constructions which explicitly encode the temporal relation of 'anteriority' -i.e. the converb event temporally precedes the event expressed by the main predicate.Thus we find constructions with the temporal preposition etter ('after') in about 38% of all translations.These include the following four types: (i) 'etter + DP' following the main predicate in the Past tense (where the DP includes information about the temporal span of the event described -e.g.etter et øyeblikks taushet -'after a moment's silence') -cf.( 12); (ii) 'etter å ha + past participle' following the main verb in the Past tense -cf.( 13); (iii) 'etter at + Past Perfect' preceding (both semantically, i.e. temporally preceding, and in the surface structure, i.e. linearly preceding) the main clause in the Past tense -( 14); and (iv) 'etter å ha + past participle' preceding the main verb -(15).
In total, 23% of all the Norwegian constructions involve temporal adverbial clauses with the temporal conjunction 'da' / 'når' ('when').The following three types of sentences are attested in this group: (i) the 'da' clause in the Past tense followed or preceded by the main clause in the Past tense -cf.( 16); (ii) the 'da' clause in the Past Perfect tense followed or preceded by the main clause in the simple Past -(17); and (iii) the 'da' clause in the Present Perfect tense followed by the main clause in the Present tense -cf.( 18).
Both the 'etter' constructions and the 'da' clauses contribute to the 'anteriority' reading of the converb event with respect to the main event, which is an important part of the semantics of delimitative converbs -i.e. the contribution of the suffix -v (cf.section [2]).
21% of all translations of pofective converb sentences involve coordinate constructions with 'og' ('and') or a comma (if there are more than two clauses in the sentence) in the simple Past tense.Most typically, the verb denoting the converb event precedes the main verb ( 19), but the reversed pattern is also attested -cf.( 9).Note that a DP expressing repetition/duration is added in the Norwegian translation of sentence ( 19): e.g.'tygget noen tak' ('chewed a few times').
In ( 19), the Norwegian sentence has several clauses in the Past.Such coordinate clauses are usually interpreted in terms of narrative progression -i.e.events temporally follow one after another.In (20), on the other hand, only two verbs are linked by 'og' ('and'), which allows for the temporal relation of overlap between the two events (cf.also the English translation with the comitative construction which is also compatible with this temporal interpretation).Note, how- [168] maria filiouchkina krave ever, that in (20), pogljadev 'having looked' can be interpreted both as a true pofective (e.g.'having looked at the sun for some time'), but also as a semelfactive ('having looked once').In case of the semelfactive reading of this converb, the relation of overlap seems justified (cf.section [3.1]).Thus the use of coordination in the Norwegian translation can be explained by the translator's interpretation of the converb event as a semelfactive event.
Other translation patterns are less frequent.Pofective converbs translated as temporal adverbs occur in 6% of all sentences but only as a translation pattern of one particular converb -pomolčav ('after being silent for some time').The temporal adverbs used include: senere ('later'), til slutt ('eventually'), litt efter ('a little later').In this case, the lexical meaning of the converb (i.e.'being silent') is made implicit -cf.( 21).4% of all translations involve relative clauses introduced by som ('which') in the Past Perfect form preceded by the main clause in the Past formas in ( 22).Such constructions are also compatible with the interpretation of the converb event as temporally preceding the main clause event.

[3.3] Semelfactive vs. delimitative converbs in the Norwegian translations
The contrastive analysis presented in [3.1] and [3.2] shows that there is a clear correlation between the semantics of converbs and their translation equivalents in Norwegian.Figure 1 on the facing page shows differences in the percentage distribution of different constructions found in the Norwegian translations of semelfactive vs. delimitative converbs.
Converb constructions with the semelfactive Aktionsart are predominantly translated into Norwegian by means of coordinate clauses linked by og ('and') or a comma (70% of all translations).Converb constructions with the delimitative Aktionsart, on the other hand, show a preference for constructions with the temporal preposition etter ('after'), in particular, the 'etter å ha + past participle' construction and the 'etter + (temporal) DP' construction (38%).Note that this construction is almost completely avoided as a translation of the semelfactive type of converb constructions (only one instance was attested).Another preferred construction in translations of delimitative converbs involves subordinate da ('when') clauses (23%).Coordinate clauses in the Past tense are only found in 21% of all translations of sentences with delimitative converbs (vs.70% for semelfactive converbs), and, as argued in [3.2], the use of coordination with pofective converbs could sometimes be due to the 'semelfactive' interpretation of some of them (e.g.pogljadev 'having glanced').
The predominant use of coordinate clauses in translations of semelfactive converbs reflects the underspecified temporal relation of the events expressed by such converbs with respect to the matrix clause (both 'overlap' and 'anteriority' relations are possible).On the other hand, the explicit temporal constructions with etter 'after' and conjunctions da / når 'when' in translations of delimitative OSLa volume 2( 1), 2010 converbs in translation [169] figure 1: Russian semelfactive and delimitative converbs and their translations in Norwegian converbs convey the 'anteriority' relation.This is the expected pattern for converbs due to the anteriority meaning of the morpheme -v, and indeed the only possible interpretation of the lexical type of delimitatives.In addition, in translations of semelfactives, we find lexical means of expressing 'semelfactivity' (i.e. the use of verbs whose semantics refers to a unique occurrence of a momentaneous event), combinations of verbs with verbal particles etc., whereas in translations of delimitative converbs, we observe explicit reference to the temporal duration of events -e.g.en stund 'a moment', etter et øyeblikks taushet 'after a moment's silence', etc.
In the next section, we will consider English translations of the same lexical types of converbs (semelfactive vs. delimitative) to see if also these data provide evidence for the correlation between the lexical semantics of converbs and their interpretation as converbs.
[4] a n a l y s i s o f t h e e n g l i s h t r a n s l at i o n s construction and the 'after + the -ing adjunct' combination -as shown in ( 36) and (37).Note that in constructions with DPs, the DP often includes information about the duration of the event -e.g.we find expressions such as: after a little conversation, after an instant's thought, after a short pause, etc).
The English converb (the -ing participial adjunct) is relatively frequent, occurring in 22% of all translations -cf.examples ( 38) and (39) for converbs in the post-verbal and pre-verbal positions, respectively.Although the -ing form by itself does not provide information about the duration of the described event explicitly, there are many instances in the data where the -ing form is combined with a DP which provides such information -e.g.thinking a moment -as in (38).
Further, we find that the 'having + past participle' construction (the Perfect counterpart of the -ing adjunct) occurs in 6% of all translations of the pofective converbs -cf.example (40).
The 'having + past participle' construction is compatible with the 'anteriority' reading of the converb event.Coordination with and (or by means of a comma) is less frequent (4%) -as in (41).In addition, we find one occurrence of the comitative 'with + DP' construction and the instrumental 'by + -ing' construction.Note that absolute constructions are not found in translations of pofectives (as opposed to semelfactives), as they are not used for the coding of successive events (cf.(Kortmann 1995, 220)).Figure 2 convincingly shows that the semelfactive type of converbs is predominantly translated into English by the -ing adjunct (68% of all data), while VP coordination ('and' + Past) and comitative 'with' constructions are the other two frequent patterns observed in this group.In contrast, the delimitative type of converbs in the Russian source data corresponds to constructions involving the preposition after (64%).The -ing converb is used in 22% of all translations of this type of sentences.Thus the English translation data provides further evidence in favor of our initial hypothesis concerning the role of lexical semantics in the interpretation and translation patterns of converb constructions.
[5] c o m p a r i s o n o f t h e e n g l i s h a n d t h e n o r w e g i a n t r a n s l at i o n s The main difference between the English and the Norwegian translations is the wide use of the -ing participial adjunct in English (which corresponds to the Russian converb construction) as opposed to Norwegian.Norwegian translations, on the other hand, are represented by a greater variety of constructions (11 main patterns vs. 6-8 in English).
OSLa volume 2 (1), 2010 [176] maria filiouchkina krave figure 2: Russian semelfactive and delimitative converbs and their translations in English The majority of semelfactive converbs are translated into Norwegian by 'og' coordination in the Past tense (70%) (cf. the predominant use of -ing clauses in English -68%).The Norwegian 'og' conjunction allows for a temporal overlap relation between the main event and the event expressed by the converb, which is often a natural interpretation for semelfactive converbs in Russian 10 .In English, on the other hand, the 'overlap' relation is mostly conveyed by the -ing form, while coordination with 'and' is employed for temporal succession.Apart from the correlation between the -ing form and 'og' coordination, there are parallels (in translations of the same semelfactive converbs) in the use of the comitative construction ('med' / 'with' + DP) in Norwegian (5% of all data) and English (11%) -cf.our example below repeated from section [3.1]: ( In all the three examples, the anteriority relation (≺) between the converb event and the matrix event is due to the semantics of the converb suffix (-v).However, unlike the accomplishment converb in example (42a), semelfactive and delimitative converbs lack the semantically visible target state and cannot be modified by temporal adverbial phrases ("for X time").Moreover, delimitatives require a specification (overtly, i.e. lexically, or covertly, i.e. contextually) of some temporal span denoting the duration of the event (e.g."2 minuty"), which is not possible with instantaneous events expressed by semelfactives.These differences can be illustrated in the following way (τ = temporal trace/span of the event, OVL = temporal overlap): Obviously, the minimal duration of semelfactive events does not provide the speaker with an interval or state s1, which is salient or relevant enough to be designated by temporal adverbials.However, as pointed out in section [3.1], certain semelfactive converbs (denoting a change in the position of some body part) are interpreted as having a target state s2 which overlaps with the matrix event: (43) d.Zakinuv nogu na nogu (na 2 minuty), on kuril.'With his legs crossed, he sat there smoking.' (in Russian: √ for 2 minutes) Thus, in this case, we get a relation between the events that is similar to that we observed for the accomplishment converbs (cf.point (a.) above): (d.) Semelfactive converbs (type 2): e1 (crossing the legs) ≺ e3 (smoking); s2 is the target state of the legs being crossed; and s2 OVL e3.Comparing two semantically distinct converbs from the point of view of their translation alternatives in English and Norwegian, we find that the lexical aspect of converbs affects the interpretation of sentences and the choice of particular translation patterns: e.g.semelfactive converb constructions are mostly translated by coordinate constructions in Norwegian and the -ing converb in English, while delimitative converbs correlate with temporal prepositions (etter in Norwegian and after in English) as well as temporal adverbial clauses in Norwegian and the Past Perfect counterpart of the -ing converb (having + past participle) in English.
The majority of semelfactive converbs in the source data correspond to constructions which by themselves leave the semantics of semelfactivity implicit: coordination is used in the majority of the Norwegian translations and the ingparticipial adjunct is employed in most of the English translations.In this case, the 'semelfactivity' reading is expressed by other lexical means: e.g. the semantics of the verb itself (e.g.glance means to look quickly at something once), certain combinations of verbs and verbal particles producing a semelfactive interpretation (e.g.hoppet opp 'jumped up'), or verb phrases with singular DPs of a certain lexical class (e.g.kaste et blikk -'throw a glance'), etc.
'Pofectivity' (with regard to converbs) encodes two semantic components: (i) that the converb event is ended before the matrix event (the anteriority meaning of -v); and (ii) that the event goes on for some time.The first component is clearly realized in the translation data: in the Norwegian data we find 38% of the OSLa volume 2( 1), 2010 converbs in translation [179] constructions with the temporal preposition etter 'after' and 23% with subordinate da ('when') clauses; while in English, 64% of all constructions involve combinations with the preposition after.As for the second component, it is reflected in the Norwegian construction 'etter å ha + past participle' which due to the presence of the Perfect Infinitive makes the event's boundaries more salient, but also in combinations of etter with DPs that provide temporal information (e.g.etter et øyeblikks taushet 'after a moment's silence').In cases of coordinate verbs, we note the presence of DPs containing information about the duration (or quantity) of the events (e.g.tygget noen tak 'chewed a few times').
To conclude, the analysis of translation data presented in this article shows that the lexical semantics of converb constructions (semelfactive vs. delimitative) in the source language (Russian) affects their interpretation and the range of possible translation alternatives in the target languages (English and Norwegian).This claim is also supported by the fact that we find similarities in the translation patterns observed in the two target languages.
The study presented in this article thus provides further evidence for the importance of distinguishing carefully between lexical classes of verbs (here: semelfactives and pofectives) in the study of grammatical verbal categories (here: converbs and their temporal/aspectual interpretation).The semantic properties of the two lexical groups in question also set them clearly apart from other perfective converbs, both of the achievement and accomplishment type.This last point is neatly confirmed by the translation data.r e f e r e n c e s Norwegian translations of Russian sentences with delimitative converbsIn this section, we look at different Norwegian constructions found in translations of 48 Russian sentences containing converbs of the delimitative Aktionsart -the so-called pofectives.The following constructions in the Norwegian target text are attested (in decreasing order):[A.]Four construction types involving the temporal preposition 'etter' ('after') -18 tokens, hence about 38% of all sentences:(i) Past (main event) + etter / efter + DP -7 tokens:(12) Da, tak o čjom že my govorili?-pomolčav, skazal Levin.

[4. 1 ]
English translations of Russian sentences with semelfactive converbs164 Russian converb sentences with semelfactive converbs marked by -nuv (69 different lexemes) correspond to the following constructions attested in the En-

[ 4 . 3 ]
Semelfactive vs. delimitative converbs in the English translations Figure 2 on the next page is an illustration of the different distribution of various constructions attested in the English translations of sentences with semelfactive and delimitative converbs.

[ 7 ]
s u m m a r y a n d c o n c l u s i o n s 14)Levin rasskazal … i, pogovoriv o politike, rasskazal pro svojo znakomstvo s Metrovym i pojezdku v zasedanie.(Lev Tolstoj, "Anna Karenina") Levin fortalte … og efter at de hadde pratet politikk en stund, fortalte han om sitt bekjentskap med Metrov og om besøket på møtet.Levin told him … and after talking a little about politics, he told him of his interview with Metrov, and the learned society's meeting.skački.(Lev Tolstoj, "Anna Karenina") Aleksej Aleksandrovitsj rakk ikke hjem igjen før klokken fem … og efter å ha spist sammen med forretningsføreren, innbød han ham til å vaere med ut til landstedet og siden til veddeløpene.Alexey Alexandrovitch only just managed to be back by five o'clock … and after dining with his secretary, he invited him to drive with him to his country villa and to the races.
We need to explain the nature of the differences in the lexical aspect of delimitatives vs. semelfactives which triggers the different temporal relations between the converb event and the matrix event attested in the translation data.Let us consider the following constructed examples of sentences with (a) an accomplishment converb with a reversible target state, (b) a pofective converb, and (c) a semelfactive converb -and their compatibility with temporal adverbial phrases: … hvor det hovedsakelig var Lena som førte ordet, mens Sergej satt med korslagte bein i et hjørne og røykte …[10]In general, though, the temporal reference of PF converbs is 'anteriority', encoded in the semantics of the suffix -v.