Volume 42, No.1 - Spring 1996
Editor of this issue: Antanas Klimas
ISSN 0024-5089
Copyright © 1996 LITUANUS Foundation, Inc.

Saulius Ambrazas. Daiktavardžių darybos raida 

(The Development of Noun Formation). Vilnius: Mokslo ir enciklopedijų leidykla (1993) pp. 294.

In the foreword (pp. 3-4) the author writes that in this book modern theoretical methods of the investigation of word formation are applied to Lithuanian historical word formation. As the fundamental object of the study five classes of verbal nouns were chosen, viz. those which denote (1) action (nomina actionis), (2) result of an action (nomina acti), (3) performer of an action (nomina agentis), (4) instrument (nomina instrumenti) and (5) place where an action occurs (nomina loci).

Ambrazas (p. 11) uses as a starting point Kuryłowicz's notions concerning the historical principles of word formation. Sometimes the derivative morpheme changes, but the motivation for the derivation remains the same. For example, the suffix -ia is characteristic of earlier nouns of action derived from prefixed verbs with the root man- 'to think,' thus ìšmonia 'understanding,' núomonia 'thought,' prãmonia 'invention fiction,' sùmania 'thinking, judging,' priemonia 'understanding' (p. 31). Later, however, the suffix replaces -ia, thus išmonė and núomonė 'opinion,' etc.

Another possibility is that the motivation for a derivation changes. For example nouns in Lith. -yba, Latv. -ība, Slavic -ĭba were originally derived from nouns, thus Lith. dalýba 'division' (cf. Latv. dalība), Slavic borĭba 'struggle' were derived from the nouns dalìs 'part' and borĭ respectively, but later they came to be felt as deriving from the verbs dalýti 'to 'divide' (Latv. dalīt) and borot'sja "to struggle' respectively (p. 12).

Sometimes the process of lexicalization takes place (p. 12). Thus, for example, only by etymological analysis can we determine that Lith. gãlas 'end' is derived from the verb gélti 'to sting' and originally had the meaning 'mortal sting, prick' and is to be related to Old Prussian gallan 'death' and the verb gallintwei (which Ambrazas translates as Lith. mirti 'to die,') but which means rather 'žudyti, to kill'. Ambrazas quotes also the cognate Old English cwellan 'to kill' also incorrectly glossed as mirti. The word cwelan with one l means 'to die.' But this may be a mere misprint.

Ambrazas remarks (p. 32) that favorable conditions for the spread of the suffix result from the mixture of -iā and ė stem nouns. In Lithuanian dialects and especially in those in which long unstressed vowel endings are shortened many endings merged or at least became similar. Therefore some -iā stem nouns, particularly the stem-stressed nouns, passed into the stem category. This process was most characteristic of the Samogitian dialect, but it also took place, although not quite to such a degree, in certain Western High Lithuanian dialects of Lithuania Minor according to Ambrazas.

I myself would add that the genitive plural and the nom. -acc.- voc. dual for these classes would be the same (sharing the same ending with other classes as well). According to the index to Daukša's Postilė prepared by Czesław Kudzinowski, 1977, 527, one encounters the instrumental singular in the following orthographic forms nuomone 452, nuomonia 4331, nuomoinia 25540. Granting that the form nuomoinia is a misprint, one wonders about the difference in pronunciation between nuomone and nuomonia. If the palatalization of consonants by following front vowels goes back to the seventeenth century one can wonder about the possibility of a phonemic contrast between /e/ and /a/ in that position. This problem was discussed by Prof. Antanas Klimas and myself more than thirty years ago (see Klimas and Schmalstieg, 1962 and furthermore Klimas, 1970).

Ambrazas writes (p. 21) that both in contemporary Lithuanian and Old Lithuanian the derivatives with the suffix -imas, -ymas constitute the most numerous type of the formation of nouns of action (nomina actionis). In contemporary Lithuanian the derivatives are generally formed from the preterit stem of the verb, thus kalbėjimas 'speaking' beside 3 pret. kalbėjo "spoke,' radìmas 'finding' beside 3 pret. rãdo 'found.' Exceptions are furnished only by būstymas 'life' beside būsta (a rare present tense form of būti to be'), kvepìmas 'smell' beside kvpia 'smells,' penìmas 'feeding, nursing' beside pni 'feeds' attested in the dialect of Zietela, and from the dialect of Lazūnai tekìmas 'rising (of the sun).' From the historical point of view it is important to note the derivatives in -imas /-ymas from the present stem of unprefixed verbs in Old Lithuanian, e.g.; imimas 'taking' (cf. contemporary ėmimas), 3 pres. ìma 'takes,' 3 pret. ėmė. Cf. also from Sirvydas' dictionary duodimas 'giving,' 3 pres. dúoda 'gives,' 3 pret. dãvė. The present tense is represented in the nasal infix form skrendimas 'flying' in Sirvydas' Dictionary, cf. contemporary 3 pres. skreñda 'flies,' 3 pret. skrìdo.

Although in Latvian the derivatives with the suffix -im(a)s are not attested, it is thought that they were ousted by derivatives with the suffix -um(a)s. Neither is the suffix attested in Old Prussian, although in Slavic we encounter *pisĭmo (cf. modern Russian pis'mo 'letter' which corresponds exactly to Lith. piešìmas 'drawing,' which latter form could be derived either from the 3 pret. pišė or the 3 pres. pišia 'draws' [p. 24]). Since in other Indo-European languages such as Hittite and Albanian the nouns of action in *-imo- are formed on the present tense it seems likely that Lithuanian nouns of action in *-imo- were formed on the present tense of i-stem verbs as might be reflected in such forms as penìmas beside pni mentioned above (p. 25).

Although nouns of action with the suffix -imas have correspondences in related languages, this does not necessarily mean that the genesis of this type of formation is to be traced back to the Indo-European proto-language. In the Slavic languages alongside the group of derivatives with the suffix ĭmo there is also a group of derivatives with the unextended suffix -mo.

Two circumstances have led to the bond between the preterit stem and the suffixes in -imas /-ymas: (1) the general tendency of verbal nouns to be based on the preterit stems and (2) the possibility of confusion with the forms of the present passive participles. This is shown by the generalization of the suffix -imas in some forms of the present passive participles, e.g., in the Wolfenbüttel Postilė pagimdimus 'pagimdomus, being borne,' pagirdimas 'pagirdomas, being given to drink.' In Latvian dialects also one encounters a confusion of the suffix -ums with that of the present passive participle (p. 26).

Possibly the general tendency of verbal nouns to be based on preterit stems is a result of the fact that Baltic preterits in *-ā derive from old stative presents which came to be felt as preterits because of their original stative function, which is very close to the perfect function. This, for example, a 3 sg. pret. such as skrìdo 'flew' was originally a stative present meaning 'is flying, is in the state of flying' and came to mean 'flew' only secondarily. The original present meaning of -o- < *-ā- is retained in such verbs as žìno 'knows,' íeško 'searches for.' The preterits žìnojo, íeškojo result from the addition of the suffix -jo as a result of its later reinterpretation as a preterit marker. Once the principle of the addition of the suffix -imas to the preterit stem was established for the *-ā-stem verbs it was spread to other verbal categories.

Undoubtedly Ambrazas is right in writing that the -m- of the present passive participle is somehow connected with the nouns of action in -imas /-ymas. As his illustrious father Vytautas Ambrazas, 1979, 52, has shown, the participial forms in -mo- in the Slavic and Baltic languages were indifferent as to voice. My own view is that they were originally active intransitive and later came to be interpreted as passives of transitives in some cases. Thus there is no cause for surprise that verbal stems with some form of the -m- suffix came to be interpreted as nouns of action even with a transitive meaning.

Ambrazas writes (pp. 15-16) that a suffix may have its origin as the second element of a compound word and he gives as examples Lith. avìdė 'sheepfold' and alùdė 'alehouse,' etc. where the second element is derived from the Indo-European root *dhē- (cf. Lith. dėti 'to put'). He denies, however, (p. 31) that derivatives such as nuodžià 'nuodėmė, sin' (cf. nusidėti 'to sin'), pradžià "beginning' (cf. pra[si]dėti 'to begin') and atidžià 'dėmesys, attention' (cf. atsidėti 'to apply oneself to') reflect the suffix -djā-. I wonder if, however, -djā-could not be just another form of *dhē-. Differently from most specialists I would derive *dhē- by way of a monophthongization from an earlier *dhoi-; the zero grade ablaut would have been *dhį-, which can be observed in the formant *-djā- mentioned above. The alternation of *dhoi- vs. *dhi- can be observed in Hittite, which did not, in my view, undergo the internal Indo-European monophthongization typical of other attested languages. Cf. the present conjugation of the Hittite verb ti-ih-hi / ti-hi/ 'I put': 2 sg. ta-it-ti /taiti/, 3 sg. da-(a-)i /tai/, l pl. ti-(i-)ya-u-e-ni /tiyaweni/, 2 pl. ta-a-it-te-ni /taiteni/, 3 pl. ti-(ya)-an-zi /tiyanzi/. Cf. also the Slavic present conjugation of dĕti 'to put': 1 sg. deždo (< *de-dj-o), etc.

Iljinskij, 1902, 224, suggested that the Slavic suffix -ba be connected with the Indo-European root *bhu. Jagić, 1902, 228, however, criticized Iljinskij's note, because the latter omitted Baltic data. Ambrazas writes (p. 77) that the Lithuanian/Latvian suffix *-ībā and Slavic -(ĭ)bā developed from the Indo-European formant *-bho-/bhā- and he quotes Brugmann, 1906, 386-389, who writes that the suffix *bho- occurs partially as a primary and partially as a secondary formant in adjectives (particularly in names of colors) and nouns (particularly in names of animals). In addition Ambrazas references Bammesberger, 1973, 131-132, who writes that the suffix *-bhā could be the feminine form of *-bho- encountered in names of animals where it derives from *-bhwo- which Bammesberger traces back further to Indo-European *bhū-. Differently from others who see a reduction of *bhwo- in the suffix *-bho-, I would reconstruct a minimorpheme for the root, viz., *bhe/o- 'to be, to become' with possible suffixes -y and -w. In my view there is no need to posit a *-w- in derivative forms which do not show direct evidence of it. I believe that *-bho-without any *-w- is attested in such words as Old Indie vrśa-bhá- 'manly, mighty, vigorous, strong,' rśa-bhá- "bull,' Benveniste, 1949, 102, suggested that the suffix *-bho- seems to have denoted especially (young?) male animals with horns. Benveniste does not mention, however, Old Indie garda-bhá- or rāsa-bha- 'donkey' both of which Mayrhofer, 1956 ff., Vol. 1, 327 and Vol. 3, 57, etymologizes as 'Schreier, brayer,' derived respectively from the words gard- 'to cry out' and rās- 'to howl, to cry.' It seems to me that it would be hard to understand either garda-bhá- or rāsa-bha- as a young male animal with horns (unless, of course, French asses have horns?). I would think rather that garda-bhá- and rasa-bha could be etymologized as 'braying beings' and vrsa-bhá and rsá-bhá as 'male, sowing beings.' One could compare the English expression human being.

Szemerényi, 1967, 280, proposed a connection between Lith. jūrės 'sea' and Slavic ryba 'fish' for which he reconstructs a proto-Slavic *yūrba. With the expected Slavic metathesis this passed to * jryba which was then simplified to Slavic ryba. If Szemerényi's etymology is correct, then ryba could easily be explained as originally denoting 'sea being' or 'water being.'

Although the suffix -ba- may indeed go back to the Indo-European root *-bho-, Ambrazas (pp. 85-86) shows that there is indeed a verbal determinative -b- attested in such words as Lith. dirbti 'to work', Old English deorfan 'to work,' Old Indic (3 sg. pres.) drbhati 'braids' vs. Lith. dìrti 'to flay,' Gk. déro 'I skin, I flay' and Lith. griebti 'to seize, to catch,' Goth griepan 'to grasp,' Old Indic (3 sg. pres.) grbhayati 'grasps' beside Lith. grieti 'to skim off (cream),' Gk. khrìein 'to rub (with scented oil)'. Contemporary paliáuba 'truce' seems to be derived directly from paliáuti 'to cease,' but in 16th century texts a synonymous verb paliaubyti (3 pres. paliaubo, pret. paliaubė) was used.

Ambrazas concludes (p. 232) that the historical analysis of the categories of verbal nouns of the Baltic languages and a comparison with the data from related languages permits the assumption that already in the Indo-European proto-language there existed two syncretic categories, viz. on the one hand (A) nouns denoting (nomina actionis) and nouns denoting the result of the action (nomina acti) and on the other hand (B) nouns denoting agent (nomina agentis) and nouns denoting the instrument (nomina instrumenti). Probably the nouns denoting the result of the action result from the concretization of nouns denoting action, thus, e.g., for the noun válgymas the meaning 'process of eating' is original and the meaning 'food' is later (p. 108). Lithuanian sėmuõ in the 16th and 17th centuries could still mean 'sowing' whereas the Latin cognate semen means 'seed' (pp. 111-112).

According to Ambrazas (p. 206) at one time the paradigm for the names of agents and instruments was the same. This is proved first of all by the fact that the only suffix characteristic only of instruments is the suffix *-tlo- which is an apophonic variant of the agent suffix *-tel-. In the oldest period of the development of the Indo-European proto-language names of instruments and agents were created with the same suffix *-ter-/-tro- (and its variant *-tel-/-tlo-) and the word-forming element *-o-. Thus in ancient Greek the suffix *-ter- could be used to denote not only agents, but instruments, cf., e.g., hrais-ter 'hammer,'hrū-ter 'drawer (of a bow, arrows' and also 'strap, rein.' In addition in Homeric language one encounters names of agents with the suffix *-tro- which is usually characteristic of instruments, thus iatrós 'physician' vs. aiter 'id.' In my view the nom. sg. iatrós derives from the agentive genitive singular of iater. The word is a good example of the two different origins of the attested Indo-European nominative case, the form iater being derived from the old absolutive and the form iatrós being derived from the old sigmatic ergative (= genitive).

In Lithuanian, according to Ambrazas, some names of instruments are formed with suffixes borrowed from names of agent. Thus the suffix -ikas typical of the agent can also denote an instrument, e.g., grūdìkas can mean 'the person who pounds, crushes' or 'an instrument to mash porridge' (p. 199).

On the other hand nouns of agent are sometimes formed with suffixes which have derived from the suffix *-tlo- which developed from the suffix for names of instruments (p. 153). Thus, for example, the suffix -klė is used for nouns of agent, cf. áuklė 'nurse, nanny.' The noun arklys 'horse' is probably derived not directly from árti 'to plow,' but rather from the noun ar-kla-s 'wooden plow' where -kla- derives from the Indo-European suffix of instrument *-tlo- (p. 207). One may speculate on whether in plowing the horse is to be considered an instrument or an agent. Certainly as an animate being a horse can perform actions of its (his, her?) own will such as eating or spooking, but it seems doubtful that a horse would go out and plow a field without human direction. In any case examples such as áuklė, etc. would seem to stand almost in direct contradiction to Dressler's, 1980, 113, universal (which Ambrazas, p. 206, quotes) that agent suffixes can get productive instrumental meanings, but that the reverse is apparently not observed.

It might be surprising also because at least in inflectional morphology Russian shows a good example where the instrumental case has taken over the function of the agent, thus, we now encounter in Russian napisano mnoju (instr. sg.) 'written by me' with the instrumental case replacing an old agentive genitive, cf. Lith. parašyta màno (gen. sg.) 'written by me' (see Paulauskienė, 1979, 99).

Pages 254-276 contain a number of very interesting tables showing the productivity of various Lithuanian suffixes at various periods, in various texts and in various regions. The book contains also an index of Lithuanian derivational affixes (pp. 277-282) and a German resumé (pp. 283-294).

Although misprints seem few, Oswald Szemerényi's name occurs consistently with an extra -n- as Szemerénnyi (pp. 14, 17, 246) The Hittite word for 'to thunder' is written mistakenly as tethoye- (p. 25) with an impossible -o- instead of tethaye-. In my copy of Kronasser, 1963, 178, the -a- is poorly printed, so that it looks like an -o-, but an -o- does not ordinarily occur in a Hittite transcription. The first name of Dressier is Wolfgang and it seems odd to quote the middle initial U. rather than the initial of his first name. (pp. 206, 237).

In sum, however, Ambrazas has written an excellent and interesting study on noun formation in Lithuanian, a study which is eminently useful and one which will retain its value for a long time to come. Our congratulations to the young author.


Ambrazas, Vytautas. 1979. Lietuvių kalbos dalyvių istorinė sintaksė. Vilnius, Mokslas.
Benveniste, Emile. 1949. "Noms d'animaux en indo-européen". Bulletin de la société de linguistique de Paris 45. 102.
Bammensberger, Alfred. 1973. Abstraktbildungen in den baltischen Sprachen = Ergänzungshefte zur Zeitschrift für vergleichende Sprachforschung auf dem Gebiete der indogermanischen Sprachen 22. Göttingen, Vandenhoeck and Ruprecht.
Brugmann, Karl. 1906. Grundriss der vergleichenden Grammatik der indogermanischen Sprachen. Vol. 2, Part 1. Strassburg, Kari J. Trübner.
Dressier, Wolfgang U. 1980. "Universalien von AgensWortbildungen." Pp. 110-114 in Wege zur Universalienforschung. Festschrift Seiler, ed. by G. Brettschneider and C. Lehmann. Tübingen, Narr.
Iljinskij, Grigorij. 1902. "Zur slavischen Wortbildung." Archiv für slavische Philologie 24. 224-228.
Jagic, Vatroslav. 1902. "Zusatz." Archiv für slavische Philologie 24. 228-229.
Klimas, Antanas. 1970. "Some Attempts to Inventory Lithuanian Phonemes." Pp. 93-102 in Baltic Linguistics, ed. by Thomas F. Magner and William R. Schmalstieg. University Park, Penn State University Press.
Klimas, Antanas and William R. Schmalstieg. 1962. "A Note on the Vocalic Phonemes of Lithuanian." The Slavonic and East European Review 41 (96). 245-246.
Kronasser, Heinz. 1963. Etymologie der hethitischen Sprache. Lieferung 2. Wortbildung des Hethitischen. Wiesbaden, Otto Harrassowitz.
Kudzinowski, Czesław. 1977. "Indeks-Słownik do 'Daukšos Postilė." Poznan, Uniwersytet im. Adama Mickiewicza w Poznaniu. Seria Filologia Baltycka Nr. 2.
Mayrhofer, Manfred. 1956ff, Kurzgefasstes etymologisches Wörterbuch des Altindischen. Vols. 1-3. Heidelberg, Carl Winter.
Paulauskienė, Aldona. 1979. Gramatinės lietuvių kalbos veiksmažodžio kategorijos. Vilnius, Mokslas.
Szemerényi, Oswald. 1967. "Slavic etymology in relation to the IE background." Die Welt der Slaven 12. 267-295.

William R. Schmalstieg
The Pennsylvania State University