The Indo-European genitive of agent with finite verbs and participles

АСТА LINGUISTICA LITHUANICA ХУІІ (2002), 41-54 The Indo-European genitive of agent with finite verbs and participles ‘ WILLIAM R. SCHMALSTIEG The Pennsylvania State University Examples of the use of the finite verb and the genitive with agent function are given from Old Indic, Old Persian, Tocharian, Greek and Slavic. It is suggested furthermore that the genitive object of certain verbal classes, e.g., Indo-European verbs with the suffix *-emay originally have functioned as an agent and that the etymological subject was the patient. In this paper I should like to give some examples of the genitive of agent (or perhaps better, genitive of source) with finite verbs in some Indo-European languages other than Baltic'. The etymological possessive interpretation is always possible for passive participle constructions, but it would seem to me to be unnecessary to have one explanation for the finite verb constructions and a second explanation for the passive participle constructions, although I suppose that in principle the constructions with finite verbs could be explained as being analogical to those with the participles. First of all I would quote the example of the use of the genitive of agent with a finite verb from Old Persian (Cardona 1970: 2; Kent 1953: 133): (1) avaiy uvjiya arika aha NOM. PL. MASC. NOM. PL. NOM. PL. MASC. 3 PL. IMPERF. the Elamites unfaithful were u-tasam Auramazda naiy ауайіуа CONIJ.-GEN. PL. MASC. NOM. SG. MASC. NEG. 3 SG. PSV. IMPERF. and by them Ahuramazda not was revered “The Elamites were unfaithful and Ahuramazda was not revered by them.’ Note the Old Indic example with a finite verb (Jamison 1979: 134): I should like to thank herewith Prof. Vytautas Ambrazas and Prof. Algirdas Sabaliauskas for reading and commenting on earlier versions of this paper. I should also like to thank Prof. Giedrius Subacius for transcribing the passage from Rhesa’s Bible in the Chicago Newberry Library. Although I have not mentioned him directly in this paper, I should like to express my thanks to Prof. Axel Holvoet, with whose conclusions I frequently disagree, but whose articles have been helpful to me in revising and formulating my own thoughts. His 2001 article quoted in the references was the immediate stimulus for this paper. Haudry (1977: 409) writes that the subjective genitive and the genitive of belonging go back to the same origin as the genitive of agent, viz. the genitive of source, which when used with the passive participles denotes the author of the action and is opposed to the instrumental which denotes the element in contact, the object possessed or the instrument of the process. 42 | WILLIAM R. SCHMALSTIEG (2) idam tyat pātram indrapānam, indrasya NOM. SG. NEUT. МОМ. SG. NOM. SG. СЕМ. SG. this that cup Indra's drink by Indra priyam amrtam apāyi NOM. SG. NEUT. NOM.SG.NEUT. 3RD SG. PSV. AOR. dear immortal was drunk "This very cup, Indra's drink, dear, immortal, was drunk by Indra.” A Tocharian B example of the genitive of agent is given below (Krause, Thomas 1960: 83): (3) srukor aišaumyepi olypo [riļtoy(tā)r NOM.SG. СЕМ. SG. ADV. 3RD SG. MIDDLE OPTATIVE death by wise (man) very is to be sought ‘eher diirfte der Tod von einem Weisen gesucht werden’ ‘death ought rather to be sought by a wise man.’ Now it is well known that in the history of languages in the course of time the preposition comes to support the simple case ending. Typically language teachers tell their students that such and such a preposition requires such and such a case or cases, but in fact historically it is usually just the reverse. The preposition (as some kind of etymological adverb) supports an original case. The replacement of Lat. ire Romam by ire ad Котат ‘to go to Rome’ is well known, but there аге many other examples also. For example, Krys’ko (1997: 121-122) writes that for Russian during the process of the formation of the direct object category the contrast between object and circumstantial forms is strengthened as a result of which the accusative of spatial object which combines the lexically locative and the functional object categories, і.е., marked both by circumstantiality and object quality, is gradually replaced by the preposition čerez ‘through’ plus case form. Interestingly enough Кгуѕ Ко predicts that the simple accusative of object of verbs of motion will gradually be replaced by the preposition plus object construction as it has already happened with certain constructions, e.g., perelezt’ Cerez zabor ‘to climb over the fence’, pereletet’ čerez more ‘to fly across the sea’. In Greek the preposition 6x0 ‘by’ has come to support the genitive case in the sense of agent. Thus: (4) тухүхосшол опо тоу YEYEVHUEVGY... тотоу хоттүүор=їу 1 5С. PSV. АОК. PREP. СЕМ. РІ. СЕМ. PL. СЕМ. SG. INF. I am forced by the happenings... thisone to accuse ‘I am forced by what has happened to accuse this man.” (Goodwin, Gulick 1958: 261.) Paul Andersen (1991: 123) showing how in Greek the introduction of an agent can change the interpretation of an active intransitive construction to a passive construc- THE INDO-EUROPEAN GENITIVE OF АСЕМТ | 43 tion gives the following examples (Herodotus 3, 15, 4): (5) apa TAVPOV TLV amedave пархурўџх АСС. SG. СЕМ. SG. NOM. SG. МАЅС. PAST ACT. PART. 3 SG. AOR. ADV. blood of bull having drunk died on the spot ‘Having drunk the bull’s blood he died on the spot.” (Herodotus 7, 213, 2) (6) amedave Ono "А 9туад=о адрос Тотугуѓоо 3 SG. АОК. PREP. СЕМ. SG. СЕМ. SG. СЕМ. SG. МАЅС. died by Athenades man Trachinian ‘He was killed by a Trachinian (called) Athenades.” With regard to Old Church Slavic Vaillant wrote (1964: 189) that the agent of a passive verb may be rendered by the instrumental or ољ plus the genitive. Vaillant (1964: 195) gives the example from Matthew 3.6: (7) kroštaaxo se otv перо 3 PL. IMPERF. REFL.PTCL. PREP. СЕМ. SG. were baptized by him “They were baptized by him.’ The addition of ot» merely strengthens the original genitive of agent meaning. Above I have given evidence for the use of an agentive genitive with a finite verb in Old Persian, Old Indic, Tocharian, Greek and Slavic. It seems to me then that an IndoEuropean origin for the construction is quite likely. Adhering to the notion that unmotivated syntax has its origin in motivated semantics, I propose that the unmotivated genitive case which is required by certain verbs is the reflection of an earlier agentive meaning. From the synchronic point of view one would say that such and such a verb requires a certain case, but there is no motivation for that. In many constructions, then, one can understand the subject as originally being the object (or patient) of the action or state and the genitive object as the agent or the source of the action or state. When the Indo-European language shifted from ergative typology to nominative-accusative typology, the original patient came to be felt as a subject and the analysis of the old ergative type sentences was changed. Thus, for example, certain Old English constructions are traditionally called ‘impersonal’, although Visser (1963: 20) considers this term a misnomer. According to Visser (1963: 23): “This construction is not infrequently accompanied by a complement in the form of a noun or pronoun in the genitive... or by a preposition (for; of, at, to etc.) + noun... Since it is convenient to have a name to refer to the complement іп the form of a (pro)noun in the genitive it will be called CAUSATIVE OBJECT, as in the majority of cases it expresses the cause or the occasion of the action or state denoted by the verb it qualifies.” Visser's term ‘causative object’ is very appropriate since it fits with the notion that the genitive is an agent or a source producing the action or state. Example: 44 | WILLIAM К. SCHMALSTIEG (8) him sceamode Oces mannes DAT. SG. 3 SG. PRET. GEN. SG. MASC. СЕМ. SG. to him it shamed of that man ‘he was ashamed of that man.’ This ‘impersonal construction’ with a ‘causative object’ was probably preceded by a personal construction also with a ‘causative object’. In Old English a dative object replaced an original nominative subject. Compare the Gothic example with a cognate verb (Luke 9.26): (9) saei... skamaip sik meina NOM. SG. 3 PRES. REFL. СЕМ. SG. whoever shames himself ofme ‘whoever is ashamed of me’ (Streitberg 1919: 131) According to Erhart (1980: 182) Indo-Iranian fourth class verbs (those in -ya) were originally stative, but some had developed into passives already in the Indo-Iranian period. Let us compare the following sentence from Old Indic (Macdonell 1916: 319): (10) аппаѕуа trpyati GEN. SG. 3RD SG. PRES. with food he refreshes himself “He refreshes himself with some food.” According to the text-book interpretation the verb trpgoverns the genitive case, but note that this verb belongs to the Indic fourth class and that one could reverse the syntactic interpretation assuming a genitive agent (or in Visser's terms, a causative object) and a nominative patient. One could then understand the sentence as: “the food refreshes him’. Consider now the following Latin example: (11) pudet теё tui 3 SG. PRES. ACC. SG. GEN. SG. shames me you “I am ashamed of you.' (Bennett 1945: 140) This also could be understood as: ‘you shame me’. Compare the Old English example (8) and Gothic example (9) above. Hermann (1926: 286-287) connects the Latin verbs in *-ėwith the Lithuanian stative verbs in -ėwhich freguently reguire a genitive object, cf. the Lithuanian sentence from Basanavičius: 12 0 ii tū manēs гаі bjauriesi J е. р J] CONIJ. CONIJ. NOM. SG. GEN. SG. ADV. 2 SG. PRES. but if you by me so are repelled ‘but if you are so repelled by me.’ THE INDO-EUROPEAN GENITIVE OF AGENT | 45 The verb bjaurétis “to be repelled’, is an -6stem verb (<*-ē-). This could be understood as: ‘but if I repel you so’. Other Lithuanian examples include: (13) Vaikas svečių drovisi (inf. drovėtis) NOM. SG. СЕМ. РІ. 3 РКЕЅ. child by guests is made shy (DŽ 1993: 13) “The child is made shy by the guests, i.e., the child is shy in the presence of the guests.” This could also be reversed and understood as “The guests make the child shy’. (14) Мапёѕ visi bédis (inf. bodčtis) GEN. SG. NOM. PL. 3 PRES. by me everyone is annoyed (LKŽ, 966) ‘Everyone is annoyed by me’ which could be understood as ‘I annoy everyone”. The verbs in examples (9) through (14) above all belong to the same Indo-European class of stative verbs and all here take genitive objects, or in Visser's term *causative objects’. See also Ambrazas (2001a: 29-34) for many other examples and somewhat similar conclusions. In my view an analysis of the genitive of agent cannot be separated from an analysis of the instrumental with which it competes in certain syntactic collocations*. Note the following Lithuanian sentence (Paulauskienė 1979: 99): (15) žėmė ėsti sniegū (sniėgo) nuklota NOM. SG. 3 PRES. INSTR. SG. GEN. SG. NOM. SG. FEM.-T-PART. earth is withsnow (by snow) covered “The earth is covered with (by) snow.’ In this case the use of the instrumental sniegt would denote that snow is considered the instrument, whereas use of the genitive sniégo would denote that snow is the agent 2 Now one may apply a similar dichotomy to the *-о stem nouns. Then the nominative singular in *-s (probably an old ergative) may be considered the source of the action as opposed to the nominative singular in *N (= *-m or *-n, probably an old instrumental) which may be considered the instrument. The reason for the assignment of animate nouns to the *-s morphological category is that they are more likely to function as the source of the action. This may explain then why such Latin nouns as gladium ‘sword’, scutum ‘shield’, caelum ‘sky’ have older alternate nominatives singular gladius, scutus, caelus. Cf. also Gk. Quyév ‘plow’ beside Cuyog. Originally the choice of the nominative singular depended on the speaker’s perception of the function of the subject, viz., either as source (with the ending *-os) or instrument (with the ending *-oN). One is, of course, familiar with the phenomenon of gender change in the adjective, viz. Lat. bon-us, bon-um, bon-a, etc., but for the noun I presuppose different original functions for the nominative singular forms *-05, *-0N, *-a, although at the moment I would have no explanation for the difference between *-oN and *-ā, for the latter of which Haudry (1977: 449) suggests an original instrumental function. He compares the form of the *-ā stem instrumental function and nominative singular in the Latin sentences (1) miles sagitta hostem vulnerat ‘the soldier wounds the enemy with an arrow’ and (2) militis sagitta hostem vulnerat ‘the soldier’s arrow wounds the enemy’. This would explain why the same root may appear in related languages as masculine or neuter * о stems or (mostly feminine) *-ā stems. 46 | WILLIAM R. SCHMALSTIEG or the source of the action. It should be noted that the agentive use of the genitive is not restricted to animate beings. Other examples guoted by Paulauskienė (1979: 99) include: (16) Daiktinis ағ veiksminis simbolis būvo ūžstelbtas NOM. SG. CON. NOM. SG. NOM. SG. 3 PRET. NOM. SG. -TPART. material or active symbol was choked off ir užtėmdytas Zodiniu apvalkalu (žodinio apvalkalo) CONI. NOM. SG. -TPART. INST. SG. INST. SG. GEN. SG. GEN. SG. and eclipsed verbal covering (verbal covering) “The material or active symbol was choked off and eclipsed by the verbal covering.’ (17) Bendruomenés istorijos kūrimas уға СЕМ. SG. СЕМ. SG. NOM. SG. 3 PRES. of the society of the history establishment is nejžvelgiama paslaptim (neįžvelgiamos paslapties) pridengtas... INST. SG. INST. SG. GEN. SG. GEN. SG. NOM. SG. MASC. -ТPART. with an impenetrable secret (by an impenetrable secret) hidden ‘The establishment of the history of the community is hidden with (by) an impenetrable secret.’ One must say that here the author and the instrument (or perhaps source and path) are not clearly differentiated, although in most cases the Lithuanian language would differentiate quite clearly (Paulauskienė 1979: 99). Still even with verbs in -ėti one notes the competition between the genitive and the instrumental, cf. the following example with (14) above: 18 Kai jisai sirgo, visi iuo bodėjosi J rg Ј J СОМЈ. NOM. SG. 3 PRET. МОМ. РІ. INSTR.SG. 3 PRET. When he was sick everybody with him was annoyed “When he was sick, everyone was annoyed with him.” (LKŽ, 967) In example (3) above I have given a Tocharian example with a genitive of agent. But in addition to the genitive in Tocharian the instrumental (for inanimate) and the perlative (for animate) cases may function as instrument and agent cases also. Note the following example from Tocharian A (Krause, Thomas 1960: 84; Lane 1947: 49): (19) mā poryo tskāmsamtār mā wdryo ADV. INSTR. SG. 3 PL. MIDDLE PRES. ADV. INSTR. not by fire are burned not by water THE INDO-EUROPEAN GENITIVE ОЕ АСЕМТ | 47 sikamtār mā lancsa pārtsi уйіейс 3 PL. MIDDLE РКЕЅ. АРУ. PERLATIVE РІ. INF. 3 PL. PRES. washed away not by kings to take can be “They are not burned by fire, not washed away by water, cannot be carried (away) by kings...” I have noted above in Greek examples (4) and (6) the genitive of agent governed by the preposition оло. The Greek dative functions as instrumental, since Greek has not inherited the instrumental from Proto-Indo-European. Thus similar to the Lithuanian competition of genitive and instrumental is the Greek competition of the genitive with the dative. Note the examples from Schwyzer (1966: 526): (20) &daun оло үрс! ROŠOAEOC Alaxidao 3 SG. PSV. AOR. PREP. DAT. PL. GEN. SG. GEN.SG. was slain by hands of swift-footed Aeacus ‘He was slain by the hands of swift-footed Aeacus.’ (Шаа П, 860) (21) о оло доор! TOALV пёр9о! DAT. SG. PREP. DAT. SG. ACC. SG. MIDDLE INFINITIVE thy by spear city to be laid waste “city to be laid waste by thy ѕреаг (Шаа XVI, 708) As with the genitive case the preposition ото merely comes to strengthen the original instrumental. The fact that ото is used with both the genitive (with agentive meaning) and the dative (with instrumental meaning) shows that the original difference in meaning was determined by the case and not the preposition. The use of the instrumental with the passive (as contrasted with the genitive as in sentence [2] above) is attested in Old Indic (Jamison 1979: 131): (22) adbhir gobhir mrjyate INSTR. PL. INSTR. PL. З PRES. with water with cows is wiped ‘He is wiped with water (and) cows (= milk).’ (23) sd myjyamanah kavibhih... NOM. SG. MASC. МОМ. SG. MASC. PSV. PART. INSTR. PL. he (= Soma), being wiped by the kavis... According to Haudry (1977: 414) the use of the instrumental as an agent in Old Indic is a result of the extension of the subject of origin to the subject of instrument. Haudry writes that this results from the fact that fsayo hinvanti ‘the seers incite’ and the innovation matayo hinvanti ‘the thoughts incite’ are parallel, so that the parallel to matibhir hitd ‘incited by the thoughts’ would be Fsibhir hitd ‘incited by the seers’. 48 | WILLIAM К. SCHMALSTIEG In Russian the competition between the genitive of agent and the instrumental is evident in this example from the Igor Tale, cf. the Jakobson reconstruction (1966: 170): (24) Poskėpani sabljami Каіепуті šelomi NOM. PL. MASC. PAST PSV. PART. INSTR.PL. INSTR. PL. NOM. PL. cleft with sabers tempered helmets Ovarbskii ots tebe, Jars ture Vbsevolode! NOM. PL. MASC. PREP. GEN. SG. УОС. 56. УОС. 56. УОС. SG. Ауаг Бу уои Fierce Bull Vsevolod! ‘Cleft with tempered sabers are [their] Avar helmets — by you, Fierce Bull Vsevolod" In modern Russian, of course, this distinction has been lost and correspondingly such modern Russian phrases as (25) and (26): (25) napisano mnoju NOM. SG. NEUT. PAST PSV. PART. INSTR. SG. ‘written by me’ (26) napisano karandasom NOM. SG. NEUT. PAST PSV. PART. INSTR. SG. ‘written with a pencil’ would be rendered in Lithuanian by (27) and (28) respectively, see Paulauskiené (1979: 99). (27) parasyta mano NOM. SG. NEUT.-T-PART. GEN. SG. ‘written by me’ (28) parasyta pieštukū NOM. SG. NEUT. -TPART. INSTR. SG. ‘written with a pencil’ In modern Russian we see a development parallel to that of Old Indic, viz. the replacement of (ofa plus) the genitive by the instrumental. The competition between the genitive and instrumental is also encountered in the Old Indic sentences quoted from Patanjali by Wackernagel and Debrunner (1954: 582): (29) idam aheh srptam ADV. GEN. SG. NOM.-ACC. SG. NEUT. -TPART. here serpent crawled THE INDO-EUROPEAN GENITIVE OF AGENT | 49 (30) ihāhinā (= iha + ahinā) srptam ADV. INSTR.SG. NOM.-ACC.SG.NEUT. -T-PART. here serpent crawled "The serpent crawled here.” In modern Armenian the participle in -c is used with a genitive of agent which apparently does not go back to Indo-European times and must be considered an innovation, differently from the use of the genitive with the participle in -/ which is most likely ancient. Note the modern Armenian example (Fairbanks and Stevick 1975: 227): (31) im barekami grac girk'2 hetak'okir ё POSS. PRON. СЕМ. SG. PASTPART. NOM. SG. NOM.SG. 3 PRES. my by friend written book interesting 15 “The book written by my friend is interesting.’ Such a sentence could be translated into Lithuanian with almost exactly the same syntactic structure: (32) mano bi¢iulio parašyta knyga yra įdomi POSS. PRON. СЕМ. SG. РАЅТРАКТ. NOM. SG. З PRES. NOM. SG. my by friend written book is interesting With the participle in — we also encounter the instrumental usage in Armenian (Tumanjan 1962: 58): (33) angin k'arov šarac ADI. INSTR. SG. РАЅТ PART. valuable withstone studded ‘studded with valuable jewels’ Cf. Lith. (34) papuostas brangakmeniais NOM. SG. INSTR. PL. ‘adorned with jewels’ One argument for the possessive origin of the genitive of agent in Baltic is the accepted use of the possessive pronouns Lith. mano, Latv. mana ‘my, by me’ and Lith. tavo, Latv. tava ‘your (sg.), by you (sg.)” as agents, rather than the standard genitives Lith. manęs, Latv. manis ‘my, by me’ and Lith. tavēs, Latv. tevis. There are, of course, exceptions. Endzelins (1923: 384) notes the vacillation between Latv. dziésma mana padziēdāta ‘the song sung by me’ and later in the same place manis jaūna padziéddta ‘a me juvenco cantata, sung by те, a young (fellow)’.