Some Observations On the Suffix -nt- in the Indo-European Languages

The aim of this article is to briefly present and summarize the meanings of the suffix -nt- in forms attested in some of the Indo-European (IE) languages. Having those meanings classified will make it possible to reconstruct the principal primary usage of the suffix as well as to clarify its origin. I shall focus my attention mainly on two languages – Hittite and Sanskrit as being the oldest ones we have data for; I shall also briefly present the state in the Slavic language family without getting into detailed study, which itself could be the aim of another paper. During the classification, various data from other IE languages where the suffix is also attested will be adduced too, which itself will help placing the observed problem into a wider comparative background.

The nt-suffix appears in noun, adjectival and verbal formations, which I believe is the state in (late) Proto-Indo-European (PIE) as well.

I. Attached to verb roots, the suffix forms verbal adjectives, which later acquire the sense of a present participle. This formation is spread in almost all of the IE languages and it presents a stable productivity. The Hittite language, though, shows an earlier state of the matters when the meaning of the participle was not related to a specific connotations of voice. Thus in Hittite this participle expressed state. It could be said that its meaning, when formed from a transitive verb expressing a state of the object or a person affected upon, was close to the meaning of a passive participle in the latter languages: – appant- ‘seized, piyant- ‘given, ḫaššant- ‘born. Formed from intransitive verbs, the participle could express an attained state: akkant- ‘died’, pānt- ‘departed’. Generically used, the participles from transitive verbs could express both active and passive voice: šekkant- ‘knowing (spirit)’ and ‘known (person)’ (Hoffner, Melchert 2008: 339). It is also worth noting that this participle was used along with the conjugated forms of the verb ‘to be’ for forming the passive voice of transitive verbs which usually exists in active. This construction was more preferable than the medio-passive forms of those verbs, which can mean that these forms expressing voice were a more recent feature of the language (Hoffner, Melchert 2008: 304).

Passive connotations in the nt-participles are also encountered in Old Norse, though its meaning is oriented in the future (Marsh 1838: 130). Jay Jassanoff proposed that the Latin gerundive (used as a future passive participle) had also originated from an earlier nt-formation, which has been extended by an additional suffix *-in-o: secundus ‘disposed to follow, following by nature’ <*sek^wont-no- <*sek^wont-i-no-‘following-related, disposed to follow’ < *sek^wont-i- ‘act of following’ < *sek^wont- ‘following’ (Jassanoff 2006).

In the morphological plan the present active participle (PAP) was first declined as a consonant stem noun. In some of the languages attested later, though, among which are Ancient Greek, Latin, Old Bulgarian and Lithuanian, this participle shows influences from the i-stems. In the forms for feminine in Ancient Greek (N. sg. φερούσα < *b^her-ont-y-a ) and most of the forms in Old Bulgarian (Acc. sg. f berǫštǫ < PSl *ber-ont-y-a-m) and Lithuanian (Acc. sg. m dìrbantį < *dirb-ont-i-m but G. pl. mf dìrbančių < *dirb-ont-y-ōm ‘working’) the participle is declined as a yo/ya-stem, which I regard as an older i-stem additionally extended by a thematic vowel in a later period. Some of the forms in Lithuanian (s. the above) as well as the declination of the PAP in Latin (Acc. sg. mf ferentem < *fer-ent-m̥ as a non-thematic stem, but Gen. pl. mf. ferentium as a vocalic stem) could be considered as proofs for such ‘i-thematic past’ of the participle.
I will try to explain this phenomenon. Initially the present nt-participle as a deverbal noun was a consonant stem1. Such athematic consonant abstract nouns are attested in Hittite, where they are significantly more widespread than the thematic ones. In PIE the thematic stems were characteristic for the adjectives which furthermore carried the accent on the thematic vowel. I assume this occurred in the period when the grammatical agreement has already arisen as an influent grammatical feature. It is proposed (Добрев 1982: 151, Burrow 2001: 211)2 that in the earlier PIE period the defining word (being a pure stem, i.e. not extended by additional grammatical affixes) was positioned before the word defined which carried the grammatical markers.

It is very probable that initially the PAP was not extended by t-suffix. The N. sg. m. endings are traditionally viewed as having lost -t- before the N. sg. ending -s. In my opinion, it is more probable that they show a more ancient way of forming the N. sg., i.e. with lengthening3 the vowel in the suffix: cf. the AGr nomina actoris in –τωρ, -τηρ where there is no nominative ending –s which would have produced a compensatory lengthening after having been dropped, thus resulting in ει and ου instead of η and ω. The tendency of extending the n-stems with additional suffix -t- has been observed by Burrow and Benveniste (Burrow 2001: 127, 130, 144; Бенвенист 1955: 54). It could be assumed that it was a marker of an initially non-nominative, later applied to any oblique case.

In a later period, the participle, initially an abstract deverbal noun, is reconsidered as an adjective. Surely, this has happened when the adjectival class had advanced in its development and has started to influence the noun class4. As a result, the process of thematization among the substantives (and later among the participles) advanced. Thus the participles began to be perceived not only as deverbal nouns which could be used independently (like the infinitives and the supine), but also like attributives standing in this way close to the adjectives and in this manner starting to acquire their formal markers. One of the clearest examples for this marker acquisition could be observed in the PAP formed with the –i/y- suffix in most of the IE languages, cf. the formation of feminine is Sanskrit where the forms in i/ī were still widespread.

II.The formant -nt- is also used in noun and adjective formation.

1. Nouns formed with -nt- can bear the following meanings or appear in these functions:

1. Youth of humans and animals: a wide spread class in the Slavic family. Although in the oldest texts from X-XI centuries there are only several such formations (agnę, žrěbę, klyusę,kozьlę, ovьčę, osьlę, otročę Бернштейн 1974: 200), in the history of the Slavic languages this model for creating diminutive nouns and words denoting small/young living creatures has appeared very productive.
2. Diminutives in Old Irish (centat ‘capitulum’), Hittite and Luwian. Vladimir Georgiev (Georgiev 1965: 177) mentions a frequent usage of such nouns in Hittite and Luwian magic rituals. According to him, in such cases those nouns possessed a strong hypocoristic meaning, often as captatio benevolentiae.
3. Hydronyms and toponyms in the Baltic languages5 (Gāters 1955: 55): names of rivers in Lithuanian (Žuvinta, Virinta) and Latvian (Viesīte where -īt-< *-int- < *-n̥t). Jan Kretschmer also studied some Italic and Greek toponyms with a suffix -nt- reaching to the conclusion that the suffix expressed a sense of belonging, membership (Zugehörigkeit) (Kretschmer 1926).
4. Personal names are frequently attested mainly in the Slavic languages: family names in Polish (Magięta, Chwalęta), diminutive personal names (Old Polish Borzęta < Borzysław); personal names in Serbian and Croatian (Mile, Hrvoje6) and Bulgarian (Яне, Гоце, Павле). Moreover, Georgiev gives some examples from the Thracian language relicts: Δορζινθης from Δορζας (Георгиев 1969: 134).
5. Seasons in Hittite and Sanskrit - Hitt. gimmanta, Skr. hemantá- ‘winter’ (same as AGr. χειμών, G. sg. χειμῶνος, OBg. zima, Lith. žiemà < * ģ^heym- ). V. Ivanov reckons that here Hittite forms in -ant (ḫamešḫanza = ḫamešḫant-s ‘spring’ ) are used as verbal subjects (ḫamešḫanza kišari ‘spring is coming/ наступает весна’), while forms ending in -i appear instead of the old pure stem used in a nominal phrase like an adjective or an adverb7 (ḫamešḫi kišari ‘it happens at spring time /наступает весной ’) (Иванов 1965: 274).
6. Another important usage of the suffix -nt- in Hittite is what Hoffner and Melchert name split ergativity : ‘Split ergativity’ refers to a situation in which a language uses both accusative and ergative patterns of case syntax side by side, the distribution of which can be conditioned by a variety of grammatical factors. ’(Hoffner, Melchert 2008: 66 n.11). The authors lean upon the paper of E. Laroche (1962) where he establishes that whenever a neuter (i.e. inanimate) noun functions as a subject to a transitive verb, it receives an extension –anza (=ants) (sg.) and -anteš (pl.) . They regard this as purely grammatical phenomenon having nothing in common with the semantic animacy.
7. A similar to the above mentioned function is the usage of -ant in Hittite when a neuter noun is in Vocative. Although, as Hoffner and Melchert claim, we can’t search a semantic animacy, in my view it could be said, that what we have here is a  grammatical animacy, i.e. the inanimate neuter noun becomes grammatically animated and in this way it can be referred to and appear in the function of vocative case. We can also compare this ‘individualizing’ function of the -nt- suffix with the names of the seasons (s. above).
8. Collective nouns. We have some attested cases when secondary collected nouns are formed with nt-suffix in Hittite: udnē- ‘country’: udnēyant- ‘population’, tuzzi- : tuzziyant- ‘army’, ešḫar- : ešḫanant- ‘blood, a criminal deed’, wet- : wetant- ‘year’ (Бенвенист 1955: 157). E. Benveniste considers the meaning of the suffix in this case, as well as in the words for seasons and ‘year’ as abstract and collective. He compares the above mentioned wet- : wetant- ‘year’ with the French jour: journée and an: année. We can see a similar collective function in the neuter plural forms in Luwian and Tocharian, too: Toch. ākrunt ‘tears < ākär < PIE *ak’ru, ñomäntu < ñom ‘name’ (Petersen 1939: 75, Бенвенист 1955: 158).
9. Another abstract function is attested in Sanskrit where feminine nouns8 are produced from prepositions and adverbs for place with a suffix -vat< *-vn̥t9 arvāvát-, āvát- ‘proximity, tidvát- ‘hight, nivat- ‘depth, parāvat- ‘longitude, pravát- ‘hight’, saṁvát- ‘region’ (Macdonell 1910: 142).

2. Numerous cases of denominal adjectives with the nt-suffix appear in the IE languages.

1. E. Benveniste noticed that in Hittite each adjective can form a doublet with –nt- without changing its meaning (Бенвенист 1955: 157) – daššu- : daššuwant- ‘strong’. In Sanskrit the same suffix appears in the highly productive formants –vant and –mant10. Meaning of ‘”possessing X” can be assigned to both of them. Arthur Macdonell defines them as expressing the sense “possessed by something” (Macdonell 1910: 193). Suffix –ant appears in small number of adjectives: bṛhánt- ‘great’, mahánt- ‘id.’, ṛhánt- ‘small’, ruśánt- ‘brilliant’, íyant- ‘that much’, kíyant- ‘how much’. Primary adjectives formed with the suffix –vant stay close to the neuters formed with the suffix –van, not extended by –t- : ṛ́kvant- : ṛ́kvan- ‘worshiping’, yahvánt- : yahú- ‘young’. Beside these examples, Thomas Burrow (2001: 145) adduces several others from Avestian, among which ǝrǝzu- : ǝrǝzvan- : ǝrǝzvant- ‘straight’ showing how the suffix –vant- was gradually built. As mentioned before, this suffix appears in numerous secondary formations. Burrow believes that there was an initial class of neuter nouns formed with the suffix –vant, referring in the usual way to the adjectives (i.e. adjectives carry the accent on the suffix while nouns have it usually on the root, thus affecting the vowel grade; the same shift of the accent was used for forming nomina agentis). The author reconstructs the following process of derivation: áśva- ‘horse’> *áśvavar/ *áśvavan/ *áśvavant11 - ‘collectivity of horses, property in horses’ > *aśvavánt- ‘possessor of horses’ (Burrow 2001: 146). The suffix seems to have been extended additionally in a later stage. In Vedic it appears in approximately 20 derivations, nearly all of which adjectives expressing possessiveness or “connection with something” (Macdonell 1910: 146). The other highly productive suffix is –mant. It appears in small number of primary derivations such as virúkmant- ‘shining’, dyumánt- ‘brilliant, splendid’ (cf. dyumná ‘light’). There is not a strict difference in the meaning of both suffixes, but it is noticed that –mant is preferred after stems in u (a tendency, observed in Hittite, too) (Burrow 2001: 146–147). Benveniste reckons that –mant could be a variant of –vant after u-stems, so that *–u-v-ant> -u-mant (Бенвенист 1955: 145).
2. Suffix –rænt appears in Old Icelandic adjectives expressing the course of winds, currents, etc.: norrænt, austrænt, væstrænt, fjallrænt – from the North, East, West, from the mountains (Marsh 1838: 106–107).
3. Baltic languages also present diminutive adjectives formed with suffix –nt- and additionally extended by another widespread diminutive suffix: e.g. Lith. –int-elis, ė (cf. Lat. -ulus): gražùs ‘beautiful’> gražintelis (Георгиев 1969: 133).
4. Slavic languages also possessed adjectives extended by *–ǫt- <*-ont-. V. Machek studied them closely (Machek 1949a). They are presented mainly in idiomatic petrified phrases in Czech and Russian with highly expressive meaning: e.g. Russ. ведьма зла злущая ‘a very evil witch’ where злущая< *zъl- ǫt-ya-ya, OCz. běloucí < *bělъ (MCz. běloučký with additional modification), *čistъ > čist´úcí> čist´oučký. Those adjectives are usually used in phrases together with the motivating word thus forming a kind of figūra etymologica: e.g. Rus. тьма тьмущая, Cz. tma tmoucí ‘a deep, intense darkness’, Cz. div divoucí, leta letoucí ‘(during) long years’, do smrti smrt’oucí ‘until death’, etc. To this group belong also secondary adjectival formations as Rus. глазучий ‘one who sees everything’ derived from глаз ‘eye’12.

III. After some of the most important attested usages of the suffix –nt- in the IE languages have been observed, I shall discuss some matters concerning its origin and meaning.

1. Forms derived with this suffix are in close connection with the old IE heteroclite declination. Its most important morphological feature is that the belonging nouns are only neuters which an ending –r- in the Nom.–Acc. sg. (i.e. when there were no other grammatical markers following them, because the form for those cases was the pure, non-extended consonant stem) which alters with –n- in the other oblique cases. Suffixes presenting such alternation are *-mer/ -men, *-ter/ -ten, *-ser/ -sen, *-ver/ -ven and *-er/ -en. This once very productive declination has left traces in most of the IE languages – Hittite, Sanskrit, Latin, Greek, as well as in Baltic and Slavic (s. Eckert 1963 for the later). Emil Benveniste (1955: 27–46, 129–150) has studied this declination on the account of its origin and usage in the PIE word formation. The above mentioned suffixes were initially used for forming abstract deverbal nouns which were later used in some languages as infinitives. With a shift of the accent, which is probably the most ancient way of derivation, and a following shift of the root and suffix vowels later nomina agentis and adjectives were formed from the old heteroclite abstracts13. As mentioned above, all grammatical markers were added to the oblique n-stem of those nouns. In this way, when extended by the additional suffix -t-, the ending –nt/n̥t was formed.
2. In my opinion, an initial meaning of an abstract reference could be ascribed to the suffix -nt-. This reference has developed in two directions:

1. Attached to verbal roots, the suffix produced a noun which could either retain its abstract meaning and relation to the verb or turn into an adjective which still expressed referential nuances to the processing/static meaning of the primary verbal root.
2. When the root was nominal, the derived word, retaining the referential meaning of the suffix, could develop into a collective noun. Such, in my view, is the case with the Slavic diminutives in –ęt-, which were initially collective names, referring to a specific multitude of small living beings14. Other examples of referential meaning are the hydronyms and the toponyms usually derived from other names; I think that it could be assumed that those names are again collectives or names denoting a specific feature of a multitude.
3. I assume the derived adjectives with nt-suffix as secondary formations of a later period. Probably such derivation could arise when the suffix has developed one of the more strict senses mentioned above. In my opinion, it is highly probable that those adjectival formations have evolved by analogy to the collective names, derived with the suffix. In this way the suffix could form a derivational model for enforcing the meaning of the base adjective, as is the case in the forms attested in the Slavic languages shown above.

• 1. Ancient PIE deverbal abstract nouns are believed to have been non-thematic consonant stems, s. Бенвенист 1955, Burrow 2001: 120
• 2. . The authors consider Karmadhāraya compounds in Sanskrit to be relicts of such type of noun-phrase coordination.
• 3. If we use the Sanskrit grammar terms: “with vr̥ddhi of the suffix vowel”.
• 4. As I mentioned before, I believe that the transition of the ancient athematic stems to the thematic class was caused by the adjectives in order to reduce the differences between the grammatical marking of both (cf. also Burrow 2001: 208, 211). Predicative usages of the participles may, in my view, be considered as relicts of such ancient state of matters when they were functioning as nouns.
• 5. Lithuanian physiographic names, including those formed with a suffix –nt-, have been studied in detail by Jan Otrębski ( Otrębski 1949).
• 6. Their old nt-stem appears in the oblique cases, s. in examples like: video sam Mileta, prićao sam sa Hrvojetom.
• 7. See above about the attributive usage of the pure stem.
• 8. It should be mentioned here that it is already almost widely accepted that the feminine nouns descend from older collectives – e.g. Clackson 2007: 100-112.
• 9. This is the reduced grade of the suffix -vant which appears in numerous adjectival formations in Sanskrit as we will see below.
• 10. The below presented examples are according to: Burrow (2001: 144, 147) for Sanskrit and Macdonell (1910) for Vedic.
• 11. The nature of this –*r/-*n/ -*nt formation will be discussed below.
• 12. This is the meaning given by Machek. СРНГ (Филин 1970: 192), on the other hand, defines it as ‘multi-coloured, bright, striking one’s eyes.
• 13. Nouns could also be used as adjectives when posed as pure stem preceding the defined word, which on the other hand received the grammatical markers. From this point of view adjectives are a more recent phenomenon in the IE languages and have surely affected or/and caused the process of thematization of the nouns (majority of which were initially pure roots, i.e. without additional affixes).
• 14. For a more detailed study on this topic, s. Machek 1949b.