Lith. “mirti” / Latv. “mirt” ‘to die’ and Lith. “miršti” / Latv. “mirst” ‘to forget’ in East Baltic

S aipsniai / A icles 11 YOKO YAMAZAKI S ockholm Uni e si y, Uni e si y o Zu ich Fields o esea ch: Bal ic linguis ics, his o ical linguis ics, Indo-Eu opean compa a i e linguis ics. DOI: doi.o g/10.35321/all83-01 LITH. MITI / LATV. MIRT ‘TO DIE’ AND LITH. MIŠTI / LATV. MÌRST ‘TO FORGET’ IN EAST BALTIC1 Lie . mi i / la . mi ‘mi i’ i lie . miš i / la . m s ‘pami š i’ y ų bal ų kalbose ANNOTATION The e bs Li h. mi i / La . mi ‘ o die’ and Li h. (-)miš i / La . (-)m s ‘ o o ge ’ sha e se e al ea u es in his o ical mo phology: bo h ake s a-p esen s em, in spi e o hei Indo-Eu opean cogna es in he *-ye/o- p esen s em; he oo -ao is in he middle oice in lec ion can be econs uc ed in PIE; and bo h a e also seman ically middle. Howe e , hey a e con as i e in he pas ense in Bal ic, aking di e en p e e i s ems, i.e., Li h. m ė / La . mi u(ē) and Li h. mišo / La . m su(ā). This a icle will in es iga e wha led hem o choose he di e en p e e i s ems by compa ing hei seman ic and phonological p ope ies, and will con ibu e o he econs uc ion o he en i e p ehis o y o he Bal ic p e e i sys em. In his a icle, i will be p oposed ha Li h. m ė / La . mi u(ē) is p obably descended om he olde impe ec , while i s ao is ic na u e led Li h. mišo / La . m su(ā) o inhe i he olde ao is s em, and his his o ical di e ence may be e lec ed in hei di e en p e e i s ems. 1 This a icle p esen s some o he esul s o a p ojec en i led The p ehis o y o he Bal ic p e e i sys em – diach onic changes and mo phoseman ics ( eg. N . 2018-00473), inanced by he Swedish Resea ch Council (Ve enskaps åde ). I am hank ul o gene ous inancial suppo om he council, and o D . Paul Widme , D . Ka in S übe , and D . Ma hias Jenny o Uni e si y o Zu ich o hei aluable commen s on an ea lie e sion o his pape . My co dial g a i ude also ex ends o D . Lamon An ieau o his p oo eading o he manusc ip , and o MA G e a Klimai ė o he p oo eading o he anno a ion (ano acija) in Li huanian. YOKO YAMAZAKI 12 Ac a Linguis ica Li huanica LXXXIII KEYWORDS: Bal ic e bs, aspec , mi i, miš i, ā-p e e i , ē-p e e i , his o ical mo phology, ao is , impe ec . ANOTACIJA Du eiksmažodžiai – lie . mi i / la . mi ‘mi i’ i lie . (-)miš i / la . (-)m s ‘pami š i’ – u i kele ą bend ų is o inės mo ologijos ypa ybių: abu šie eiksmažodžiai u i esamojo laiko s a- kamieną, nepaisan giminiškų indoeu opie iškų esamojo laiko *-ye/o- kamienų o mų; iš indoeu opiečių p okalbės galima ekons uo i, kad šių eiksmažodžių šakniniai ao is ai linksniuojami kaip medijai; abu šie eiksmažodžiai pasižymi medijų seman ika. Vis dėl o palyginus bal ų kalbas ma y i, kad jų p e e i ų kamienai ski iasi: lie . m ė / la . mi u (ē) i lie . mišo / la . m su (ā). Šiame s aipsnyje lyginan seman ines i onologines ypa ybes i iama, kodėl šie eiksmažodžiai įgijo ski ingus p e e i o kamienus, bei p isidedama p ie isos p o obal ų kalbos p e e i o sis emos ekons ukcijos. S aipsnyje eigiama, kad lie . m ė / la . mi u (ē) ik iausiai y a kilęs iš senojo impe ek o kamieno, o lie . mišo / la . m su (ā) dėl ao is inių seman inių ypa ybių galė ų bū i kilęs iš senojo ao is o kamieno. Šis is o inis ski umas a sispindi ski inguose jų p e e i o kamienuose. ESMINIAI ŽODŽIAI: bal ų eiksmažodžiai, eikslas, mi i, miš i, ā-p e e i o kamienas, ē-p e e i o kamienas, bal ų kalbų mo ologinė is o ija, ao is as, impe ek as. §1. The mo phological con as o he wo e bs, Li h. mi i / La . mi ‘ o die’ and Li h. (-)miš i / La . (-)m s ‘ o o ge ’, in hei p e e i o ms is he cen al opic o his pape . Excep o hei oo s uc u e (C- o Li h. mi i / La . mi and CC- o Li h. (-)miš i / La . (-)m s ), he wo e bs sha e a numbe o his o ical and mo phological ea u es. To name a ew, bo h o m s a- p esen s in Eas Bal ic, bo h ha e cogna es in Indo-Eu opean languages wi h he p esen s em in *-ye/o- and he oo -ao is in lec ing in he middle oice, and bo h a e also seman ically middle: 1) Li h. in . mi i, p es. m š a, p e . m ė; La . in . mi , p es. mi s u, p e . mi u(ē) ‘ o die’: OCS p es. 3sg. umьŕe ъ (Zo), in . m ě i, oo -ao . 2/3sg. m ě( ъ); Sln. p es. 1sg. m jèm; Sk . p es. 2sg. mid. m iyáse, p . 3sg. mam a, oo -ao . mid. 3sg. ám a (LIV2 439); YA . ma - ‘ o die’, p es. 3sg. mid. ami iiei e, 2sg. inj. mid. auua.mi iiaŋha, OPe s. ma - ‘ o die’, imp . 3sg. mid. am(a) iya ā, Manichean MPe s. my -, Zo oas ian MPe s. my - /m - / ‘ o die’, Pa ian my - ‘ o die’, Sogd. my - (bo h in Ch is ian and Buddhis ex s) ‘ o die’ (Cheung 2007: 264–265); La . mo io ; OHi . (* oo -ao . ac . >) oo -ao . mid. me a ‘disappea ed, wen los ’; om PIE *me - ‘ o die’ (LIV2 439); 2) Li h. (-)miš i, (-)miš a, (-)mišo; La . (-)m s , (-)m s u, (-)m su(ā) ‘ o o ge ’: Sk . p es. 3sg. mid. mya e, p . mamá a, oo -ao . mid. 3sg. amṭa, a-ao . mid. 3pl. áman a ‘ o o ge ’ (LIV2 440–441); S aipsniai / A icles 13 Li h. mi i / La . mi ‘ o die’ and Li h. miš i / La . m s ‘ o o ge ’ in Eas Bal ic Manichean MPe s. ’mwš-, Zo oas ian MPe s. pl’mwš- / āmōš- / ‘ o o ge ’, p es. 3sg. pl’mwšy / āmōšēd /, Pa ian ’mwš- ‘ o o ge ’, p es. 3sg. ’mwšyd; Sogd. ’’wyšcy, β ’wcy, MSogd. ’wycyh ‘ o ge ulness’ (Cheung 2007: 268–269); TochB (III) mä se ä ‘ o o ge ’; om PIE *me s- (LIV2 440–441). Consequen ly, a p esen s em in *-ye/o- (3sg. *m-yo ‘dies’ 3sg. * ms-yo ‘ o ge s’) and he oo -ao is (3sg. *m- o ‘died’, 3sg. *ms- o ‘ o go ’; p obably wi h a ze o-g ade oo and den al endings) in he middle oice can be secu ely econs uc ed o P o o-Indo-Eu opean o bo h. Fu he mo e, he pa e n sha ed by hese wo e bs is associa ed wi h he “s a i e-in ansi i e sys em” (Jasano 2003: 155, 160), cha ac e ized by he a es a ions o passi e ao is (o he ablau ing middle oo ao is ) in non-passi e use, s a i e pe ec , and in ansi i e(-inchoa i e) ye/o-p esen (and also he Bal o-Sla ic -p esen s o ē-s a i es) in he Indo-Eu opean languages. The e bal oo s associa ed wi h his sys em a e supposed o deno e en y in o a s a e. Appa en ly, he a chaic ablau ing middle oo ao is does no seem o ha e been inhe i ed by hese wo e bs in Eas Bal ic. Gi en hese esemblances, hey can be expec ed o ha e simila his o ical de elopmen om PIE o Bal ic, s ill exhibi ing he same mo phological pa e ns in he Bal ic languages. Howe e , Li h. mi i / La . mi akes he ē-p e e i , while Li h. miš i / La . m s akes he ā-p e e i . O cou se, he p esence o mé a onie ude in m š a ‘die(s)’ and he absence he eo in miš a ‘ o ge (s)’ does no appea o be a i ial di e ence. This will be discussed in §7 below. I he ia-p esen s em is he sou ce o he ē-p e e i s em, as ad oca ed by Ba on (1980) and Villanue a S ensson (2005, 2006), he e mus ha e been a eason o PB *miśya- no in oducing *-(i)yā- o i s p e e i s em. This aises he ques ion (a leas o he p esen au ho ) o wha allowed hem o c ea e di e en p e e i s em o ma ions. §2. Be o e mo ing o wa d in ou discussion, some discussion o he p ehis o y o he Bal ic p e e i sys em and o p e ious esea ch on he his o ical backg ound o he Bal ic ā-p e e i and ē-p e e i is equi ed. Al hough a ious hypo heses ha e been p esen ed o explain why hese o ma ions di e , no all o hem can be p esen ed he e due o limi a ions o space. To begin, he Bal ic languages ha e wo p e e i s ems, as is well known: he ā-p e e i and he ē-p e e i . Any gi en e b in hese languages akes one o hese wo s ems in he pas ense, which means ha he wo s ems a e in complemen a y dis ibu ion. The e is only one excep ion o his ule – he p e e i o b i ‘ o be’ in Old Li huanian. While he p e e i o ms o Li h. b i / La . bû ‘ o be’ a e bu o and biju in s anda d Li huanian and La ian, espec i ely, OLi h. 3sg. p e . bi i and OP u. be (beside bēi, bei) a e a es ed as he p e e i o ms o he e b meaning YOKO YAMAZAKI 14 Ac a Linguis ica Li huanica LXXXIII ‘ o be’. These o ms appa en ly had no s em- o ming su ix like *-ā- o *-ē-. OP u. be can be equa ed wi h OCS bě, which is he impe ec o by i ‘ o be’, bu in lec s like an ao is (c . S ang 1942: 85). OP u. be can also be compa ed wi h some hing akin o he ao is o s a i e e bs o he ype sėdjo o sėd i, sdi ‘ o be si ing’, p esen ed in S ang (1966: 381), and i s his o ical backg ound does no seem s aigh o wa d.2 Bu OLi h. bi i is p obably based on an impe ec o *bhuh2-ye/o- ‘ o become’ (→ PBS *bh(w)ī- > PB *bī), combined wi h he a hema ic p ima y ending 3sg. - i (S ang 1966: 380).3 4 La . biju has he same impe ec -based s em, bu is ex ended by he Bal ic p e e i s em *-ā- and he hia us b eake -j-, e.g., 1sg. *bī-jā-ō. Thus, i can be obse ed ha , la e han he o he e bs, he p e e i o Li h. b i accep ed he new Bal ic p e e i o ma ion ule using he su ixes *-ā- and *-ē-, whe eby he o me impe ec s em was di ec ly ollowed by he p ima y ending. This enables us o in e how he Bal ic e bs migh ha e been u nished wi h hei p e e i s ems. Mos likely, he p e e i s ems in Bal ic we e o iginally o med wi h an impe ec o ao is -based s em ollowed by he ( hema ic o a hema ic) p ima y ending. The p e e i su ixes *-ā- and *-ē- we e la e in oduced o all he e bs due o addi ional ac o s, wi h a leas one such ac o immedia ely coming o mind. The impe ec o ms wi hou he augmen o some e bs would ha e sha ed he same o m wi h he p esen o ms, i he impe ec -based p e e i o ms had s a ed aking he p ima y endings in P o o-Bal ic, as seen in OLi h. bi i, e.g., PBS p es. *Ha [H]-ye- (i), imp . *Ha [H]-ye- (→ p es. / imp . *a -ya- (i) > Li h. ã ia, ã ė o á i ‘ o plow’)5. I mus ha e become necessa y o dis inguish 2 Klingenschmi (1982: 3–5) sugges s ha OCS bě and OP u. be a e a con inua ion o he augmen ed impe ec *(h1)e-h1es-, wi h a seconda y b- aken om associa ed o ms, such as in . by i, bū on, e c. 3 ModLi h. bu o is a comple ely new o ma ion based on he in ini i e b i / pe ec ac i e pa iciple bu - wi h he Bal ic p e e i su ix *-( )ā-. 4 A di e en his o ical in e p e a ion o he o m is p oposed by Os owski (2008: 463–464), acco ding o which he inal syllable - i in bi i is no a pe sonal ending bu would his o ically be an encli ic p onoun, e.g., dsi- (duosiu au) ‘I will gi e you’ (BB Pa , 3,28). This would, howe e , lea e a ques ion, i.e., whe he bi i would be well- o med, since he p edic ed o m bi i would ha e nei he a s em- o ming owel no a pe sonal ending. As igh ly s a ed by Os owski (2008), he a hema ic p ima y ending was o he wise ne e a es ed in he Bal ic p e e i o ms. Howe e , since he hema ic p ima y endings a e used in p e e i o ms (e.g., 1sg. -au < *-ā-u, 2sg. -ai < *-ā-i o he ā-p e e i ; 1sg. -iau < *-ē-u, 2sg. -ei < *-ē-i o he ē-p e e i , also see Ba on 1980: 269), i is no un easonable o assume ha he a hema ic p ima y endings could ha e been also used in p e e i o ms. The e o e, in he discussion in his sec ion, he classic iew by S ang is main ained. 5 The i-apocope would also ha e played a ole in he p ocess. S aipsniai / A icles 15 Li h. mi i / La . mi ‘ o die’ and Li h. miš i / La . m s ‘ o o ge ’ in Eas Bal ic he p esen and p e e i o ms by he s em o ma ions, which possibly equi ed an in oduc ion o he dedica ed p e e i su ixes. §3. I seems o be a widely sha ed no ion ha he Bal ic ā-p e e i su ix mo e o less sha es he same o igin as he Sla ic ā-ao is , o he second-s em ao is in -a- o he ype 1sg. bь axъ, 2/3sg. bь a ‘ o ake’ (S ang 1942: 189; Jasano 1983: 62), o med o he ze o-g ade oo .6 The p oblem wi h his idea is p esen ed by he o igin o he Bal ic ē-p e e i , which does no seem o ha e a di ec compa andum e en in i s closes b anch, Sla ic. S ang’s (1942: 83) assessmen should p obably be main ained on he poin ha he Sla ic ě-impe ec does no coincide wi h he Bal ic ē-p e e i because he Sla ic ě-impe ec does no pa icula ly occu wi h ansi i e e bs (bu i s occu ence is ins ead phonologically condi ioned), while he Bal ic ē-p e e i clea ly occu s equen ly wi h ansi i e e bs.7 Di e en ocalisms in he oo o he Sla ic ě-impe ec and he Bal ic ē-p e e i should also no be igno ed (S ang 1942: 84). The e o e, p e ious s udies ha e le he his o ical o igin o he ē-p e e i a p oblem o be sol ed. I is ue ha he Bal ic ē-p e e i o en occu s (e.g., nešė ‘ca ied’) whe e he s-ao is occu s in Sla ic (1sg. něšъ ‘ca ied’), and hus hese wo ca ego ies mus ha e had some hing in common. The e o e, Schmals ieg (1961) en isaged he possibili y ha he ē-p e e i de eloped om he Indo- Eu opean impe ec (2sg. *h1nee-s, 3sg. *h1nee- ‘ o ca y’), which occu s in he 2nd and 3 d pe son singula o he s-ao is pa adigm in Sla ic (2/3sg. neše). On he o he hand, he e is a comple ely di e en app oach ha sees he su ix o he ē-p e e i as de eloping h ough a sound change (*iyā > ē (Ku scha 1876; Schleiche 1856: 224 .; La sson 2010: 71 .)). This can be suppo ed by he ac ha he ē-p e e i is egula o he causa i e-i e a i e e bs in in . -y i, p es. *-ā (← PB *-iya- < PIE *eye/o-) o he ype móky i, s d. móko / dial. mókia, mókė ‘ o each’, whe e he p e e i s em *-ē- can be assumed in all likelihood o ha e de eloped om he impe ec -based s em in *-iyā- (c . Pede sen 1926: 11). Mo e ecen ly, Ba on (1980) and Villanue a S ensson (2005) elabo a ed on his hypo hesis. Ba on (1980) sugges s ha *-(i)y- was in oduced o he 6 S ang (id.) dis inguished be ween wo ā-p e e i ypes: 1) descendan s o he Bal o-Sla ic ā-ao is men ioned immedia ely abo e, and 2) descendan s o he Bal o-Sla ic hema ic ao is . As men ioned in S ang (id. 383), he hema ic ao is was p obably eplaced by he ā-ao is . The e o e, i can be said ha he ā-su ix o he ā-p e e i is a mo phological con inua ion o he Bal o-Sla ic ā-ao is . 7 Pace Ko land (2005: 168) and Jasano (2017: 229, n. 107), whe e he ē-p e e i is associa ed wi h he s a i e e bs in *-ē- (e.g., ek i ‘ o low’, gūž i ‘ o lie (unde some hing wa m)’), which is descended om he o iginal ins umen al ending *-eh1 acco ding o Jasano (1978: 123–125). YOKO YAMAZAKI 16 Ac a Linguis ica Li huanica LXXXIII p e e i s em analogically om he p esen s em in *-ya- (< PIE *-ye/o-), pa icula ly in he cases o some e bs, including mi š a (← *mi ia( )). The p oposed analogical p ocess could be o mula ed as such: ea ly P o o-Bal ic p es. *mi -ya- :p e . *mi -ā- ↓ *mi ya- :*mi -yā- → *mi -ē- The -ē- su ix could ha e de eloped om *-(i)yā- h ough he pa adigm le eling igge ed by he 1s and 2nd pe son singula endings (1sg. *-yāu > -iau, 2sg. *-yāi > -ei), which would ha e esul ed in he same o ms as hose o he ē-p e e i (1sg. *-ēu > -iau, 2sg. *-ēi > -ei (Ba on 1980: 269)). Villanue a S ensson (2005) akes a mo e adical posi ion, p oposing ha o iginally P o o- Bal ic had only he ā-p e e i , and he ē-p e e i s basically de eloped om he *-iyā-s em, whe e he abo e-men ioned analogy in oduced he *-(i)y- elemen o he o iginal ā-p e e i s em o he ia-p esen s.8 This could ce ainly explain why he ē-p e e i is egula o he ia-p esen s, as desc ibed in such g amma s as Endzelīns (1923: 590) and S ang (1942: 191; 1966: 377). Applied o he case o mi i and miš i, howe e , his hypo hesis would p edic ha bo h o he e bs should ake he ē-p e e i , and would no explain why he p e e i o miš i emains an ā-p e e i . I is ue ha ha is exac ly how analogy wo ks, i.e., i does no a ec linguis ic o ms egula ly as a sound law does. In his pa icula case, howe e , pe haps a mo e sensible eason could be ound o why hese wo e bs ake di e en p e e i s ems in Bal ic. §4. A leas wo di e ences be ween he wo e bs can be obse ed. One is a di e ence in alency. Al hough bo h ha e middle- oice seman ics, whe eby he subjec s ha e li le con ol o e he e en s deno ed by hese e bs, Li h. miš i / La . m s is a ansi i e e b, as i (unde lyingly) akes an objec , whe eas Li h. mi i / La . mi p esumes no objec . Bu his ea u e would ha e na u ally led o Li h. miš i / La . m s aking an ē-p e e i a he han an ā-p e e i , and hus did no con ibu e o he e en ual ou come. The second di e ence can be seen in hei lexical aspec s. While Li h. mi i / La . mi has a dual aspec ual cha ac e , i.e., i can be ead as ei he impe ec i e o pe ec i e depending on con ex (c . Amb azas 1997: 235), his does no 8 In he cu en au ho ’s iew, i is possible ha he ē-p e e i has se e al di e en his o ical sou ces, c . Schmals ieg (1961). S aipsniai / A icles 17 Li h. mi i / La . mi ‘ o die’ and Li h. miš i / La . m s ‘ o o ge ’ in Eas Bal ic seem o be he case o Li h. miš i / La . m s . Gene ally in Li huanian, e bs wi hou any p e ixes a e g amma ically impe ec i e (c . Damb iūnas e al. 1990: 385; Amb azas 1997: 234–235), and Li h. mi i / La . mi is no excep ion o his ule, as shown in (3a, b). The La ian example in (3d) may bes demons a e he usual con as o unp e ixed and p e ixed e bs. Howe e , as in (3c), he impe ec i e o ms o Li h. mi i / La . mi in he p e e i ense can con ey pe ec i e meaning in a speci ic con ex : 3) a. Jis m š a. ‘he dies / is dying’ (Li huanian, Amb azas 1997: 235) ‘he’ ‘die’ imp . 3p es.; b. Žmónės m ė kasdien. ‘People died e e y day.’ (Li huanian, id.) ‘people’ pl. nom. ‘die’ imp . 3p e . ‘e e y day’; c. Me gai ė mi ė aka . ‘ he gi l died yes e day’ (Li huanian, id.) ‘gi l’ sg. nom. ‘die (p .)’ imp . 3p e . ‘yes e day’; d. Dažu dienu mi a un nenuomi a. ‘many’ ‘days’ gen. pl. ‘die’ imp . 3p e ‘and / bu ’ ‘die’ p . 3p e . ‘(he was) dying o many days, bu he did no die.’ (La ian, Endzelins 1948: 241). This is wha is mean by he “dual aspec ual” ea u e (Amb azas 1997: 235) o Li h. mi i / La . mi . Now, in he case o Li h. m š i / La . m s , he con as o unp e ixed (impe ec i e) and p e ixed (pe ec i e) o ms appea s o be di e en . 4) 9 a. Mi šau d ejus me us, ką jis sakė. ‘ o ge ’ imp . 1sg. ‘ wo (coll.)’ ‘yea ’ pl.acc. el-p on. acc, ‘he’ sg. nom. ‘say’ 3p e . ‘Fo wo yea s, I was ying o o ge wha he said.’ b. Užmi šau pe d ejus me us, ką jis sakė. ‘ o ge ’ 1sg. p e . p . ‘du ing’ ‘ wo (coll.)’ pl. acc. ‘yea ’ pl.acc. el-p on. acc. ‘he’ sg. nom. ‘say’ 3p e . ‘Wi hin wo yea s, I comple ely o go wha he said’. c. Rùsiškai a jaũ nemi šiù. (Li huanian, LKŽe) 9 I am indeb ed o my Li huanian in o man , MA G e a Klimai ė, o he language check on he da a in (4) and he ansla ions, and o he linguis ic ad ice on he da a. Needless o say, all emaining mis akes a e mine. YOKO YAMAZAKI 18 Ac a Linguis ica Li huanica LXXXIII ‘Russian’ ‘so’ ‘al eady’ NEG- ‘ o ge ’ imp . 1sg. u .10 ‘I will de ini ely no o ge Russian.’ d. Mi šaũ mi šaũ i užmi šaũ isa, k js sãkė. (Li huanian, LKŽe) ‘ o ge ’ 1sg. p e . imp . ‘and’ ‘ o ge ’ p . 1sg. p e . ‘all, comple ely’ el-p on. acc. ‘he’ sg. nom. ‘say’ 3p e . ‘I kep o ge ing and comple ely o go all ha he said.’ e. Ká ė ešį jaũ p àdeda apmiš i. (Li huanian, LKŽe) ‘cow’ sg. nom. ‘cal ’ sg. acc. ‘al eady’ ‘s a ’ 3p es. ‘ o ge pa ially’ p . in . ‘The cow has s a ed o o ge i s cal .’ . Ká ė ešį jaũ p àdeda užmi šin i. ‘cow’ sg. nom. ‘cal ’ sg. acc. ‘al eady’ ‘s a ’ 3p es. ‘ o ge li le by li le’ p . (i e a i e). in . ‘The cow has s a ed o o ge i s cal .’ Li h. mi š i usually akes pe ec i e (p e ixed) o ms,11 and i s impe ec i e (unp e ixed) o ms in he p e e i could be associa ed wi h a peculia si ua ion like “ac i e o ge ing”, whe e he e en “ o ge ing” can ake some ime, and he subjec can be conscious o he p ocess, e.g., when ying o o ge a auma ic o unpleasan memo y, as in (4a). This indica es ha mi š i could ake g amma ical impe ec i e o ms; howe e , hey a e seldom used. In he u u e ense, he impe ec i e o ms seem o ins ead o e a pe ec i e meaning, as in (4c). Needless o say, he pe ec i e o ms deno e he pe ec i e aspec , as in (4d), as is o be expec ed (c . Damb iūnas 1958).12 These ac s s ongly poin o he pe ec i e lexical aspec o he e b mi š i. This may be suppo ed by he ollowing: he pe ec i e in ini i es can be used wi h o begin (p adė i, im i) and o end (baig i), as in (4e), bu such cons uc ions sound be e wi h he pe ec i e in ini i e o i e a i e (lexically impe ec i ized) e bs like užmi šin i, as in 4 (c . Damb iūnas 1958: 260). The ac ha he impe ec i e o ms o Li h. mi š i / La . mi s a e a ely used may indica e, om a his o ical pe spec i e, ha i may be he case ha he old ao is s we e ein o ced in Bal ic (and Sla ic) languages by adding p e ixes, i he p oposal by Os owski (2019: 57) is o be accep ed in ou cu en 10 The nega ion ma ke ne- is ce ainly a p e ix, bu i does no a ec e bal aspec (Damb iūnas 1958: 255). 11 La . mi s is also usually used wi h a p e ix aiz- o pie-, see ME II: 634. 12 The pe ec i e o ms can also indica e a esul a i e s a e, by aking a empo al ph ase exp essing he ex ension o ime, as in (2d). S aipsniai / A icles 19 Li h. mi i / La . mi ‘ o die’ and Li h. miš i / La . m s ‘ o o ge ’ in Eas Bal ic discussion. Acco ding o Os owski, e bs such as hese ha lack impe ec i e (= unp e ixed) o ms and o which only pe ec i e o ms a e used (e.g., *žin i, pažin i ‘ o ge o know’) a e p obably based on hei p e e i o ms, which a e descended om he olde ao is s. Thei ao is ic o pe ec i e aspec ual ea u e was ein o ced in Bal o-Sla ic by adding a p e ix (i.e., o ming he pe ec i e o ms). F om a phonological poin o iew, i canno be cla i ied whe he Li h. mi š- / La . mi s- is de i ed om he o iginal p esen o ao is s em, since he oo ocalism was he ze o g ade (*ms- > *mi ś-) in bo h enses in he middle oice. Ne e heless, such a heo e ical sugges ion would speak o an old ao is as he o igin o he p e e i *mi šā. §5. As ela ed o he specula ion abo e, an obse a ion can be made o he cogna es o Li h. mi i / La . mi and Li h. mi š i / La . mi s om Indic (c . 1 and 2 in §1). I is in e es ing o no e ha bo h he ao is and p esen injunc i e o ms a e a es ed o √m, while no p esen injunc i e o √m is ound in he Ti us Sansk i da abase.13 This con as can be ound only in Indic, because no in lec ional o ms o *me s- we e inhe i ed by Sla ic, bu bo h impe ec (3pl. mь ěxǫ, Sup asliensis 398, 24) and ao is (3sg. umь ě ъ Assemanianus 11c28, m ě ъ, Assemanianus 11d2’) o ms a e a es ed wi h ega d o *me -14. This a leas implies ha bo h he impe ec and ao is ca ego ies we e p obably in use in P o o-Bal o-Sla ic, albei he inno a i e o ma ion o he Sla ic impe ec . In ac , he p esen injunc i e o √m is no ound in Rig Veda (RV) no in A ha a Veda (AV), bu when he co po a o Sama Veda Jaiminīya B āhmaa (SV JB) and Jaiminīya Upaniad B āhmaa (SV JUB) a e included, a couple o hem a e ound (e.g., 3sg. m iya a SV JB 1, 242, 34a; SV JUB 3, 8, 10a,), while ha o √m is no . Al hough he exac dis inc ion be ween ao is and impe ec in he indica i e mood has been he subjec o much deba e,15 i seems clea ha 13 All he Rig edic and A ha a edic da a quo ed he e we e aken om Ho mann (1967). The da a om B āhmaas o Sama Veda we e supplied by he da abase a ailable a Ti us (A ailable a : h p:// i us. kidg1.uni- ank u .de/da abase/ i usinx/wo dwhl.h m#dS a [Accessed 1 Ma ch 2020]). Since he aspec ual dis inc ion be ween ao is and impe ec was obscu ed in he indica i e mood bu be e p ese ed in he injunc i e mood, I added he injunc i e o ms o he compa ison. This kind o quick su ey o he da abase is insu icien o knowing exhaus i ely he a es a ions o ce ain o ms, bu o he pu pose o he cu en discussion, i su ices o seeing he gene al endency o he s a e o a es a ions. 14 C . Ai ze mülle 1977: 252, 664. 15 The ao is indica i e is usually hough o e e o he ecen pas , whe eas he impe ec e e s o he emo e o his o ical pas (Delb ück 1876; Tichy 1997); consequen ly, he impe ec is ypically used in na a i es in la e Vedic ex s (Ho mann 1967). Howe e , he e a e o he iews, as well. Some (e.g., Gonda 1962; Kipa sky 1998) assume ha hei unc ions a e mo e aspec ual han he YOKO YAMAZAKI 26 Ac a Linguis ica Li huanica LXXXIII Lie . mi i / la . mi ‘mi i’ i lie . miš i / la . m s ‘pami š i’ y ų bal ų kalbose SANTRAUKA Veiksmažodžiai – lie . mi i / la . mi ‘mi i’ i lie . (-)miš i / la . (-)m s ‘pami š- i’ – y a paliudy i y ų bal ų kalbose, o giminiški žodžiai y a gana plačiai paliudy i ki ose indoeu opiečių kalbose. Šių d iejų eiksmažodžių i giminiškų žodžių analizė odo, kad jie u i kele ą bend ų is o inės mo ologijos ypa ybių: abu šie eiksmažodžiai u i esamojo laiko s a- kamieną, nepaisan giminiškų indoeu opie iškų esamojo laiko *-ye/o- kamienų o mų; iš indoeu opiečių p okalbės galima ekons uo i, kad šių eiksmažodžių šakniniai ao is ai linksniuojami kaip medijai; abu šie eiksmažodžiai pasižymi medijų seman ika. Vis dėl o jų p e e i ų kamienai ski iasi: lie . m ė / la . mi u (ē) i lie . mišo / la . m su (ā). S aipsnyje pa eikiamas išsamus šių eiksmažodžių mo ologijos p iešis o ės, seman ikos i leksikos aspek ų y imas, įskai an lyginamuosius duomenis iš sla ų i indų kalbų. Ty imas suponuoja, kad lie . mi i / la . mi leksinis eikslas galė ų bū i impe ek inis a ba ao is i- nis, p iklausomai nuo kon eks o, o lie . (-)miš i / la . (-)m s a eju impe ek inė eikšmė y a e a; šis ski umas galėjo egzis uo i da p iešis o iniame e ape. Ki a e us, anks esnia- me y ime eigiama, kad iš ao is o kamienų kilę eiksmažodžiai dažnai ne u i g ama inio impe ek o o mų ( . y. ne u i p iešdėlių) bal ų kalbose. Taip gali bū i i lie . (-)miš i / la . (-)m s a eju, nes šis eiksmažodis e ai a ojamas be p iešdėlio. Taigi p ielaida, kad lie . (-)miš i / la . (-)m s ik iausiai y a kilęs iš ao is o kamieno, o lie . mi i / la . mi galė ų bū i kilęs iš senojo impe ek o kamieno, galimai susi o ma usio iš esamojo laiko *-ye/o- kamieno, pa i inama iek seman iniu, iek o maliuoju aspek u. Senasis impe ek- o kamienas *mi ya- ik iausiai ga o bal iško p e e i o kamieną suda ančią p iesagą *-ā-, iš ku ios kilo *mi yā- i išsi u uliojo į *mi ē- bal ų p okalbėje. Da oma iš ada, kad ski in- gos lie . mi i / la . mi ‘mi i’ i lie . (-)miš i / la . (-)m s ‘pami š i’ p e e i o kamienų o mos a spindi jų ski ingus is o inius kamienus, . y. impe ek o i ao is o kamienus, iš ku ių jie i kilo. Į eik a 2020 m. liepos 27 d. YOKO YAMAZAKI S ockholm Uni e si y Uni e si y o Zu ich Seile g aben 53, 8001 Zü ich, Swi ze land yoko.yamazak[email p o ec ed]