The Tocharian Adjectives in B -tse A -ts*

HANNES A. FELLNER
University of Vienna
0. Nouns in Tocharian are traditionally classified according to their inflectional
patterns, a system established in Tocharisches Elementarbuch (TEB = Krause and
Thomas 1960). Tocharian adjectives are divided by TEB (144–57) into four clas-
ses according to their nominative and oblique masculine plural endings in To-
charian B: class I with nom. pl. m. -i, obl. pl. m. -eṃ; class II with nom. pl. m. -ñ,
obl. pl. m. -(nä)ṃ; class III with nom. pl. m. -ñc, obl. pl. m. -ntäṃ; class IV with
nom. pl. m. -ṣ, obl. pl. m. -ṣäṃ.
0.1. Class I adjectives are subdivided by TEB (144–50) according to the presence
or absence of palatalization of their stem-final consonants and the formation of
the feminine plural in Tocharian B. This class goes back to various PIE thematic
adjectives.
Class Iaα has no paradigmatic alternation of the stem-final consonant and a
nominative plural feminine in -ana. Its prominent members, the productive rela-
tional adjectives in B-ṣṣe A-ṣi (e.g., Boraṣṣe Aorṣi ‘wooden’ : BAor ‘wood’), and
B
-ññe A-ñi (e.g., Bostaññe ‘domestic’ : Bost ‘house’), exhibit gemination of the
stem-final consonant in Tocharian B.
Class Ibα also has no alternation of the stem-final consonant and has a nomi-
native plural feminine in -ona. The gerundives in B-lle A-l (e.g., Bpralle Apräl ‘to
be carried’ : BApär- ‘carry’) with gemination of stem-final consonant in Tocharian
B and a subset of the adjectives in B-re A-r (e.g., Bratre Artär ‘red’) belong to this
class.1
Class Ibβ shows paradigmatic alternation of the stem-final consonant and has
a nominative plural feminine in -ona. Palatalization is found in the masculine
oblique, the entire masculine plural, and the feminine singular. The so-called
privatives in B-tte A-t (e.g., Betaṅkätte Aatäṅkät ‘unhindered’ : BAtäṅk- ‘hinder’)
with gemination of the stem-final consonant and the ordinal numbers in B-te A-t
(e.g., Btrite Atrit ‘third’) belong to this class.

* I am indebted to Jay Jasanoff, Jeremy Rau, Melanie Malzahn, Craig Melchert, Brent Vine, and
Laura Grestenberger for helpful discussion and comments on earlier drafts of this paper. The
usual disclaimers apply.
1 Tocharian B has another subset of these adjectives in B-re, which inflects according to adjec-
tive class II: nom. sg. -e, obl. sg. -eṃ, nom. pl. -eñ, -e(nä)ṃ.

Stephanie W. Jamison, H. Craig Melchert, and Brent Vine (eds.). 2014.

Proceedings of the 25th Annual UCLA Indo-European Conference. Bremen: Hempen. 49–56.
50 Hannes A. Fellner

0.2. The Tocharian adjectives in B-tse /-tˢe/ A-ts /-tˢ/ belong to class Iaβ (TEB:147–
8) and are characterized by nom. pl. m. -i, obl. pl. m. -eṃ, alternation between
palatalization and non-palatalization of the stem-final consonant, and nom. pl. f.
-ana in Tocharian B. The palatalization of the stem-final consonant appears in the
masculine oblique and the entire masculine plural in Tocharian B. Tocharian ad-
jectives in B-tse A-ts inflect in the following way:2
TB TA
m. f. m. f.
nom. sg. -tse -tsa -ts -tsi
obl. sg. -ce -tsa -tsäṃ -tsāṃ
nom. pl. -ci -tsana -tse -tsāṃ
obl. pl. -ceṃ -tsana -tses -tsāṃ3

0.3. The adjectives in B-tse A-ts consist of synchronic primary (cf. Bktsaitstse ‘old’
A
ktsets ‘perfect, accomplished’, Bwartse ‘broad, wide’ Awärts ‘id.’) and secondary
formations. As secondary formations, these adjectives are productive in Tochari-
an B and can be derived from every word class, but are only occasionally found
in Tocharian A: Bkramartse [Akrāmärts] ‘heavy, weighty’ ← Bkrāmär ‘weight,
heaviness’; Borkamotse ‘dark’ ← Borkamo ‘darkness’, adj. ‘dark’; Baletse ‘for-
eign, strange’ ← Ballek ‘other’; Boktatse Aoktats ‘eightfold’ ← Bokt Aokät ‘eight’.
1. From a synchronic point of view, there are two interesting observations con-
cerning these adjectives.
1.1. First, tse-adjectives in Tocharian B show variation between non-geminated
and geminated forms of the stem consonant throughout the paradigm.
The geminated forms themselves vary between the spellings 〈tsts〉 and 〈tts〉,
which represent /tˢː/, in the nominative singular masculine and the entire feminine
paradigm.
This difference in the treatment of the stem consonant can be exemplified by
looking at the attestation of one of the most prolific adjectives of the class in
question, namely, orotse ‘great, big’:4 nom. sg. m. orotse (e.g., THT 17a3), orots-

2 For the present study only nominative and oblique forms are relevant.
3 In the nominative/oblique feminine plural forms in -tsaṃ are occasionally attested. The alter-
nation of -aṃ ~ -āṃ in the feminine plural endings in Tocharian A corresponds to the one
found in Tocharian B -ona ~ -ana (see Fellner forthcoming a and Fellner 2014).
4 Since it has no bearing on the present discussion the variation orotse ~ wrotse (for which see
Malzahn 2013) can be ignored.
 The Tocharian Adjectives in B -tse A -ts 51

tse (e.g., THT 94a2), orottse (e.g., THT 338b5); obl. sg. m. orocce (e.g., THT
81b3), oroce (e.g., THT 333a1); nom. pl. m. orocci (e.g., THT 521b4), oroci
(e.g., PK AS 13I b7); obl. pl. m. orocceṃ (e.g., PK AS 7I b1), oroceṃ (e.g., THT
1192a6); nom. sg. f. orottsa (e.g., B85b5), orotsa (e.g., IOL Toch 739a2); obl. sg.
f. orotstsai (e.g., THT 305a3), orottsai (e.g., PK AS 15B a4 ); nom. sg. f. orots-
tsana (e.g., PK AS 7B a4), orotsana (e.g., PK AS 17G b6); obl. sg. f. orotstsana
(e.g., PK AS 4B a3), orotsana (e.g., THT 290 3).
1.2. Second, in Tocharian B the stem consonant of the adjectives in B-tse appears
as -c(c)- in the oblique masculine singular and the whole masculine plural, just
like in the ordinals and the privatives which also belong to class I:
B
tse-adjectives ordinals privatives
nom. sg. m. -ts(ts)e -te -tte
obl. sg. m. -c(c)e -ce -cce
nom. pl. m. -c(c)i -ci -cci
obl. pl. m. -c(c)eṃ -ceṃ -cceṃ

However, the stem allomorph -c(c)- is not found in the feminine singular:
B
tse-adjectives ordinals privatives
nom. sg. f. -ts(ts)a -ca -cca
obl. sg. f. -ts(ts)ai -cai -ccai

Moreover, the stem allomorph -c(c)- is not attested in Tocharian A at all (cf.
above).

2. Both of these synchronic problems of the tse-adjectives in Tocharian B, the

gemination and palatalization of the stem consonant, can be explained by analogy
with other class I adjectives.
2.1. The gemination must be due to analogical influence of adjectives that regu-
larly show gemination throughout the paradigm, namely the adjectives in B-ṣṣe
and B-ññe. This analogy is likely to have spread from cases where it was possible
to derive an adjective in B-tse and an adjective in B-ṣṣe or B-ññe from the same
base, e.g., kenaṣṣe ‘earthly’, kenätstse ‘id.’ ← keṃ ‘earth’; pyapyaiññe ‘flowery’,
pyapyaitstse ‘id.’ ← pyāpyo ‘flower’.
2.2. The palatalization in the tse-adjectives in Tocharian B must be analogical to
adjectives also with a stem-final dental, i.e., privatives, ordinals, and the pro-
nouns. These regularly alternated between unpalatalized (nom. sg. m.) and pala-
talized (obl. sg. m., nom. and obl. pl. m.) stem consonants in their paradigm. The
following proportion can be assumed:
52 Hannes A. Fellner

nom. sg. m. -t(t)e : obl. sg. m. -c(c)e ::

nom. sg. m. (ku)se : obl. sg. m. (ku)ce ::
nom. sg. m. -ts(ts)e : obl. sg. m. x, where x was solved as -c(c)e5

3. Regarding the prehistory of the suffix B-tse A-ts, Meillet (1911:146) suggests
that it reflects PIE *-ti̯ o-. This view is usually repeated in the more recent litera-
ture (e.g., Hilmarsson 1986:260, Adams 2013:749 and passim). There are no
problems from a phonological point of view: the sequence PIE *-ti̯ - and/or PT
*-ty- gives -ts- in both Tocharian languages, cf. Bepetsa ‘fiancée’6 < PT *-pætyā
< *-t-ih2 (cf. Ved. sapátnī- ‘rival; bride’); Blāntsa Alānts ‘queen’ < PT *-ntyā <
*-nt-ih2 (feminine of Bwalo, obl. Blānt; Awäl, obl. Alānt ‘king’). However, the pre-
history of PIE *-ti̯ o- as reflected in Tocharian has not yet been discussed.
3.1. Dunkel (1983; cf. also Balles 1997 [2000]:161–2 and Schulze 1907) discuss-
es a type of formation in *-ti̯ o- by which relational adjectives were originally de-
rived from local adverbs. Some of these are inherited, e.g., Hitt. appezziya- ‘rear’,
Ved. ápatya- ‘descendant’ (cf. also the Greek adverb ὀπίσσω ‘backwards’) from
*op- ‘away, back, behind’; Ved. nítya- ‘native’, Go. niþjis, ‘relative’ from *ni-
‘in’; Ved. santᵢya- ‘benign’, OHG samfti ‘gentle’ from *som- ‘together’; cf. also
Lat. vitium ‘fault, defect’ from *u̯ i- ‘apart’ and Go. auþja ‘bleak’ from *au̯ -
‘away’. The general derivational mechanism is productive in a number of
branches: Hitt. šarāzziya- ‘upper’ (cf. Lyc. hrzze/i- ‘id.’) ← šarā ‘up(wards),
above’, Hitt. ḫantezziya- ‘first’ ← ḫant- ‘forehead, front’; Ved. níṣṭya- ‘foreign’
← nís ‘out’, amā́ tya- ‘cohabitant’ ← amā́ ‘at home’; Gk. εἴσω ‘into’ ← ἐν ‘in’,
ἔπισσαι ‘after-born daughters’ (Hsch.) ← ἐπί ‘on, upon’.
It is safe to say that the continuant of PIE *-ti̯ o- in the Tocharian adjectives in
B
-tse A-ts is exactly this inherited type. This is shown by Bsnaits(ts)e ‘destitute’
(cf. Ved. sánutya- ‘being at a distance’) derived from Bsnai Asne ‘without’ (cf.
Lat. sine ‘without’‚ OIr. sain ‘separate, different’). The old pattern of deriving
relational adjectives from local adverbs enjoyed a mild productivity in Tocharian:

5 The speakers of Tocharian B were able to adopt -c(c)- as the palatalization product of -ts(ts)-
due to the lack of a paradigmatic alternation non-palatalized vs. palatalized ts anywhere else in
their synchronic grammar. In Tocharian A, alternation of ts with its regular palatalization
product ś is found in verbal paradigms (cf. pret. ptcp. tsmo vs. prs. III 3. pl. śamantär [: tsäm-
‘grow’], prs. I 3. pl. tsipiñc vs. imperf. 3. pl. śepär [: tsip- ‘dance’]).
6 Pinault (2009 [2010]:36) shows that this is the correct interpretation of the word previously
understood as †Bpetso ‘husband’.
 The Tocharian Adjectives in B -tse A -ts 53

B
aiwoläts(ts)e ‘directed to’ ← Baiwol ‘towards’, Beṣṃkets(ts)e ‘lasting’ ← Beṣṃke
‘up to, until’, Bwräntsaits(ts)e ‘directed to’ ← Bwrantsai ‘against, opposite’.
3.2. In the prehistory of Tocharian, *-ti̯ o- must have spread from its original locus
to become productive as an all-purpose adjective marker. This scenario can be
supported by a parallel development in the Luvo-Lycian branch of Anatolian. As
Melchert (2014:262) points out (following Hajnal 1994:151–2 and Gusmani
1961), the suffix *-ti̯ o- > Luv. -zza/i-, Lyc. -zze/i- (e.g., Lyc. hrzze/i- ‘upper’, cf.
Hitt. šarāzziya- ‘id.’ ← šarā ‘up(wards), above’) became productive. In Lycian,
formations going back to *-ti̯ o- are made from place names (e.g., Atãneze ‘of
Athens’ ← Atãna- ‘Athens’, neleze- ‘of the agora’ ← nele- ‘agora’) and also
make denominal derivatives in Luvian (e.g., wašḫazza/i- ‘holy’ ← wašḫa- subst.
‘the sacred’, urazza/i- ‘great’ ← ura/i- adj. ‘great’).7
4.1. Hajnal (1994:151–2; cf. also Laroche 1979:98–100) shows that adjectives in
-ze/i- are related to names of professions and agent nouns in -za- in Lycian. Ex-
amples include kumaza- ‘priest’ (cf. HLuv. kumaza-, which probably also means
‘priest’; Melchert 2014:261), wasaza- kind of priest, maraza- ‘judge’, xddaza-
‘slave’, zxxaza- ‘fighter’, and probably also HLuv. wara/ilaza- ‘?’, which is a
substantive referring to an occupation according to Melchert (2014:261). Follow-
ing Hajnal (1994:151–2) and Melchert (2014:262), the following derivational
scheme can be envisaged: (CLuv. kumma- subst. ‘the sacred’) → Lyc. *kumeze-
‘sacred’ → kumaza- ‘priest’ (cf. kumezi(je)- adj. ‘sacred’, subst. ‘sacred pre-
cinct’; kumez(e)i- ‘to sacrifice’). Melchert (2014) traces the substantives in -za-
and other substantives in -a in Lycian back to masculine endocentric individualiz-
ing *-eh2- in the sense of Nussbaum 2014:303–6 (cf. also Fellner and Gresten-
berger forthcoming).
4.2. Interestingly, there are also names of professions and agent nouns in Tochar-
ian that have the same shape as the adjectives in B-tse A-ts (< *-ti̯ o-), including
B
käryorts(ts)e ‘trader, merchant’ (: Bkaryor ‘trade, commerce, business’), Bwer-
piśkats(ts)e ‘gardener’ (: Bwerpiśke ‘garden’), Byāmäts(ts)e ‘doer’ (: yām- ‘do’),
B
rinäts(ts)e ‘renouncer’ (: Bri-n- ‘leave, give up’), Aamokäts ‘artist, craftsman’ (:
A
amok ‘art’), Atspokäts ‘taster’ (: Atspok- ‘enjoy [food])’. The use of formations
in B-tse A-ts to derive substantives originally goes back to nominalizations of
B
tse/Ats-adjectives, as evidenced by derivational chains such as: BAamok subst.
‘art’ → Bamokä(ts)tse adj. ‘artistic’ → subst. Aamokäts ‘artist’, Bkaryor Akuryar

7 On the use of adjectives in -zza- to mark the superlative in Luvian alone see Yakubovich 2013.
54 Hannes A. Fellner

subst. ‘trade, commerce, business’ → Bkäryorts(ts)e adj. ‘pertaining to trade,

commerce, business’ → Bkäryorts(ts)e subst. ‘trader, merchant’, Bñuwe adj. ‘new’
→ Bñwets(ts)e adj. ‘new’ → Bñwets(ts)e subst. ‘novice’. The names of profes-
sions and agent nouns in Tocharian B-tse A-ts are in principle parallel to the Lyci-
an nouns in -za-,8 except that they are not formed with the individualizing suffix
*-eh2-.9 This fact has so far remained largely unnoticed.
5. In conclusion, the synchronic problems of the inflection of the Btse-adjectives,
the gemination and the palatalizaion of the stem consonant, can be explained by
analogy to other class I adjectives. The adjectives in B-tse A-ts go back to PIE
*-ti̯ o-, an inherited formation that originally was used to form relational adjec-
tives from local adverbs that is found across different IE languages. There are
remarkable parallels between Tocharian and Luvo-Lycian concerning the devel-
opment of PIE *-ti̯ o- that can be summarized as follows: (1) both branches inher-
ited the derivational mechanism of deriving local adjectives in *-ti̯ o-; both
branches extended the use of the continuants of *-ti̯ o- (2) (probably) first to place
names and then (3) to other nouns; (4) both branches started to use the continuant
of the *ti̯ o-suffix for marking names of professions/agent nouns:
(1) Lyc. hrzze/i- ‘upper’
(cf. Hitt. šarāzziya- ← šarā ‘up(wards), above’)

8 The evidence of Luvo-Lycian and Tocharian taken together might suggest that the Lycian
substantives in -za- are originally (re)characterizations of ze-substantives to specifically mark
them as substantival in contrast to adjectival -ze-. While some of the ethnica in Lyc. -ze-
(which are usually adjectives) leave room for both adjectival and substantival interpretation,
there is only one clear case of a substantive in -ze- attested in Lycian so far. This is the collec-
tive prñneze/i- ‘household’ (cf. CLuv. parna- ‘house’), which is a direct nominalization of a
ze-adjective (Melchert 2014:262); an individual member of the household is expressed by the
i(je)-derivative prñnezi(je)-.
9 In Tocharian A the names of professions/agent nouns in -ts follow class I inflection, but there
are also attestations where they followed the a-inflection of the so-called agents (Baknātsa
A
āknats subst. ‘fool’, adj. ‘ignorant’; formations in B-ntsa like Bwapāntsa ‘weaver’ (: Bwāpā-
‘weave’); formations in B-nta A-nt like Bkauṣenta Akoṣant (: Bkau- Ako- ‘kill’); formations in B-
uca like Bkärstauca (: kärstā- ‘cut’); verbal governing compounds in B-a like Byolo-rita ‘seek-
ing evil’ (: Britā- ‘seek, long for’) and Aṣotre-lyāk ‘seeing signs’ (: Aläkā- ‘see’)), e.g., nom. pl.
tspoktsāñ. If these cases are not just influenced analogically by the class of agents, they also
reflect old *-ti̯ ōn- with the continuant of PIE individualizing *-ō/on- (cf. Fellner forthcoming
b). This would provide another parallel to Lycian, where next to the continuant of *-eh2- in
tewinaza- we find the continuant of *-ti̯ ō/on- in a similar function in the personal name Tewi-
nezẽi, both “the tewineze- one” (Melchert 2014:262).
 The Tocharian Adjectives in B -tse A -ts 55

Toch. B snaits(ts)e ‘destitute’ ← Bsnai Asne ‘without’

(cf. Ved. sánutya- ‘being at a distance’ : OIr. sain ‘separate’ etc.)
(2) Lyc. neleze/i- ‘of the agora’ ← nele- ‘agora’
Toch. B kenätstse ‘earthly’ ← keṃ ‘earth’
(3) Luv. urazza/i- ‘great’
Toch. B orots(ts)e ‘great, big’10
(4) HLuv. kumaza/i-, Lyc. kumaza/i- ‘priest’ ← CLuv. kumma- ‘the sacred’
Toch. A amokäts ‘artist’ ← amok ‘art’
Toch. B ñwets(ts)e ‘novice’ ← ñuwe ‘new’

References

Adams, Douglas Q. 2013. A Dictionary of Tocharian B2. Amsterdam and New York:
Rodopi.
Balles, Irene. 1997 [2000]. Reduktionserscheinungen in langen Wortformen als Ursprung
morphologischer Doppelformen im Urindogermanischen: Die Suffixformen *-i̯ o-
und *-ii̯ o-. Die Sprache 39.140–66.
Dunkel, George E. 1983. πρόσσω καὶ ὀπίσσω. Zeitschrift für vergleichende Sprach-
forschung 96.66–87.
Fellner, Hannes A. 2014. PIE Feminine *-eh2 in Tocharian. In: Neri and Schuhmann
2014, 7–21.
———. Forthcoming a. Tocharisch und das indogermanische Femininum. In: Norbert
Oettinger (ed.), Das Nomen im Indogermanischen. Morphologie, Substantiv versus
Adjektiv, Kollektivum. Akten der Arbeitstagung der Indogermanischen Gesellschaft
vom 14. bis 16. September 2011 in Erlangen. Wiesbaden: Reichert.
———. Forthcoming b. Tocharian Special Agents: The nt-participles. Tocharian and
Indo-European Studies 15.
Fellner, Hannes A., and Laura Grestenberger. forthcoming. The Greek and Latin Verbal
Governing Compounds in *-ā- and Their Prehistory. In: Bjarne Simmelkjær Sand-
gaard Hansen, Benedicte Nielsen Whitehead, and Birgit Anette Olsen (eds.), Etymol-
ogy and the European Lexicon. Akten der 14. Fachtagung der Indogermanischen
Gesellschaft vom 17. bis 22. September in Kopenhagen. Wiesbaden: Reichert.
Gusmani, Roberto. 1961. Il suffisso -tjo- di aggettivi “locali” e la sua diffusione nelle
lingue indoeuropee. Annali dell’Istituto Orientale di Napoli, Sezione linguistica
3.41–58.

10 Luv. urazza- and Toch. B orots(ts)e were already equated by Van Windekens (1960; cf. also
Melchert ap. Adams 2013:128), but the details remain to be worked out.
56 Hannes A. Fellner

Hajnal, Ivo. 1994. Die lykischen a-Stämme: Zum Werdegang einer Nominalklasse. In:
Jens E. Rasmussen (ed.), In honorem Holger Pedersen. Kolloquium der Indogerman-
ischen Gesellschaft vom 26. bis 28. März 1993 in Kopenhagen, 135–71. Wiesbaden:
Reichert.
Hilmarsson, Jörundur. 1986. Studies in Tocharian Phonology, Morphology and Etymolo-
gy with Special Emphasis on the o-vocalism. Ph.D. diss., Leiden University.
Laroche, Emmanuel. 1979. L’inscription lycienne. In: Emanuel Laroche, André Dupont-
Sommer, Manfred Mayrhofer, and Henri Metzger (eds.), Fouilles de Xanthos VI: La
stèle trilingue du Létôon, 49–127. Paris: Klincksieck.
Malzahn, Melanie. 2013. Tocharische Dichtersprache. Paper presented at Sprache und
Metrik in Diachronie und Synchronie, 2 September, at Ludwig-Maximilians-Univer-
sität, Munich.
Meillet, Antoine. 1911. Remarques linguistiques. Journal Asiatique 18.144–50.
Melchert, H. Craig. 2014. PIE *-eh2 as an ‘Individualizing’ Suffix. In: Neri and Schuh-
mann 2014, 257–72.
Neri, Sergio and Roland Schuhmann (eds.). 2014. Studies on the Collective and Feminine
in Indo-European from a Diachronic and Typological Perspective. Leiden and Bos-
ton: Brill.
Nussbaum, Alan J. 2014. Feminine, Abstract, Collective, Neuter Plural: Some Remarks
on Each. In: Neri and Schuhmann 2014, 273–306.
Pinault, Georges-Jean. 2009 [2010]. Review of Tremblay 2003. Kratylos 54.24–36.
Schulze, Wilhelm. 1907. Ahd. suagur. Zeitschrift für vergleichende Sprachforschung
40.400–18.
TEB = Wolfgang Krause and Werner Thomas. 1960. Tocharisches Elementarbuch I:
Grammatik. Heidelberg: Winter.
Tremblay, Xavier. 2003. La déclinaison des noms de parenté indo-européens en -ter-.
Innsbruck: Institut für Sprachen und Literaturen der Universität Innsbruck.
Van Windekens, Albert J. 1960. Hieroglyphenhethitisch ura- und tocharisch B orotse
‘groß’. Zeitschrift für vergleichende Sprachforschung 76.180–1.
Yakubovich, Ilya. 2013. The Degree of Comparison in Luwian. Indogermanische Forsch-
ungen 118.155–68.