*téri-/tri-´
‘three’
- Language
- Proto-Anatolian
- Grammar
- num
- ID
- 595
1. Individual Anatolian Languages
1.1 Hieroglyphic Luwian
/tarri-/, tara/i-
‘three’1.1.1Transmission
This lexeme is attested throughout the HLuw. period from the 12th century through to the late 8th century BCE. Being a lower number word, it is solidly attested, yet suffers a drawback faced by most number words in HLuw. and other Anatolian languages: most attestations of this lexeme are written as numbers rather than spelt-out numerals, i.e. they do not provide any point of reference as to its phonological and morphological form. Occasionally, the number attestations are followed by one or two syllabic signs – usually in later attestations – that provide an overt inflection, thereby hinting at the linguistic form, but it is far more common for the numbers to appear as is, i.e. as bare number signs.
[AH.B.]
1.1.2Forms
| nom./acc. pl. c. | /tarrinzi/ | tara/i-zi/a | |
|
3-zi-i |
|||
| “3”-zi[…?] | MARAŞ 14, §9 (Maraş) | ||
| bare number sign | ? | 3 | |
| KARAHÖYÜK (ELBİSTAN), §16 (Malatya) | |||
| KULULU 1, §6 (Tabal) | |||
| KULULU lead strip 2, §1.5 | |||
| KULULU lead strip 2, §3.17 | |||
| KULULU lead strip 3 (+ frag. 2), reg. 1 | |||
| MARAŞ 11, §8 (Maraş) | |||
| TOPADA, §22 (Tabal) |
1.1.3Graphic Features
1.1.3.1Phonological Interpretation
There are only three spelt-out attestation of this numeral, two of which are cardinals from the TOPADA inscription (Tabal), and the third is a derived adverb in KARKAMIŠ A6. The HLuw. evidence is hence very thin and the signs used leave much room for interpretation about a possible anaptyptic vowel between the initial t and r, i.e. reading /tarri-/ rather than /tri-/. All suggestions put forward so far regarding the phonological interpretation draw heavily on CLuw. and Hittite (cf. Melchert 1994a:58; EDHIL 873; Yates 2015a:154f.) and do not provide conclusive evidence for the HLuw. forms.
[AH.B.]
1.1.4Meaning
The meaning of this lexeme is reasonably clear even in cases where the context is damaged.
Regarding syntax, it should be mentioned that the lower numerals ‘two’, ‘three’, and ‘four’ typically occur with plural head nouns in HLuw., whereas head nouns following the numerals from ‘five’ onwards show singular inflection (Bauer 2014a:77-83).
[AH.B.]
1.1.5Stem
As far as can be ascertained, the plural endings evident in TOPADA, the ASSUR letter a, and in MARAŞ 14 attach directly to the root of the numeral when it appears as a cardinal number.
It remains unclear whether or not the root shows a different ablaut in the adverb compared to the cardinal because the sign used to write the form is the same one in each case. In fact, the sign itself is a ligature that consists of three strokes indicating the number ‘3’ combined with sign *383, a the small downward stroke on the side commonly used to indicate ra/i, i.e. the sign tara/i arose out of this very number word by way of the acrophonic principle to be used with a syllabic value more widely across the lexicon later on.
[AH.B.]
1.1.6Derivatives
1.2 Lycian A
*trija, *tri
‘three’1.2.1Compounds
1.2.2Derivatives
- terihe
- trijere-
- trppeme/i-
- trijẽteze/i- (PN) (Neumann 2007a:374 with further literature)
1.3 Lycian B
trija
‘three’1.3.1Transmission
This lexeme is found once in the poetical inscription engraved on the lion sarcophagus of Antiphellos dated on stylistic grounds to the beginning of the 4th century BCE (Zahle 1979a:318, Mühlbauer 2007a:91).
[D.S.]
1.3.2Forms
| indecl. | trija | TL 55.3 (Antiphellos) |
1.3.3Meaning
In his study of Lycian numerals, Bugge (1899a:60) recognizes the word for ‘three’ in Lycian B trija (thus also Schürr 2005a:152, hesitant DLL:131). Because of the false segmentation of the following word by the Lycian mason, the sequence da of da:te, was taken together with the word for ‘three’ (Neumann 2007a:373). However, it is clear that da{:}te is the verb of the sentence meaning ‘they gave’ and trija is the direct object, i.e. ‘they gave three’. There is a genitive plural in the sentence, which might depend on the number three, i.e. qirzẽ ‘?’. Contra Shevoroshkin 2014a:236, 254, who takes the word trija as referring to the Storm-god. This is refuted by the mismatch with the verb date (for *dãte), which must be plural on account of its lenited conjugation, i.e. da-(di) (see under lemma).
[D.S.]
1.3.4Stem
A priori, it is not certain whether the number for ‘three’ is declinable or not. If it is, the stem would be tri(je)- just like kbi(je)- ‘two, another’ and, thus, trija would be neuter plural. However, considering greater numbers like sñta ‘ten’, which are indeclinable and end in -a, trija could be indeclinable as well. This is even more likely in this case, because trija does not agree with anything neuter plural in the context.
[D.S.]
1.3.5Derivatives
2. Reconstruction
2.1Reconstruction ‘three’
Eichner 1992a:68f. recognizes two genuine ablaut forms in the word for ‘three’ in the Anatolian languages; cf. Hitt. teriyaš ‘third’, CLuw. *tarri- (in derivatives), Lyc. terihe ‘third’ vs. Lyc. tri-. Thus, he reconstructs two ablaut forms of the root, i.e. PAnat. *téri-/*tri-´ (followed by Kloekorst 2014a:65), but remains uncertain regarding the exact morphological process hiding behind it. What is clear is that Hitt. teriyaš (< *térii̯os) and Lyc. A terihe (< *tériso) go back to the exact same form, i.e. to the genitive singular of the number three. Čop’s Law, which can be very observed in Luwian, shows that the r-consonant was in an open syllable, which means that the glide was syllabified as *-ii̯-, i.e. *térii̯- > CLuw. *tarri- and not **téri̯-. Yates 2015a:154 (following Melchert 2013e:140 n. 4) argues for an early epenthetic vowel, i.e. *tərí, followed by an attraction of the accent, which would have made the epenthetic vowel into a full accented vowel, i.e. *téri. To account for the Lycian forms in tri-, he suggests that they reflect pre-forms in *teri-, which would have undergone a late syncope. However, what is not credible in his account is why terihe would have preserved the vowel -e-, but trija would not. Therefore, we follow the explanation with the two ablaut grades.
[D.S.]
3. Proto-Indo-European
3.1Semantic Reconstruction
PIE *trei̯-/*tri- denotes the cardinal number ‘3’.
3.1.1Root etymology
The most secure way to deal with such a basic lexeme is to recognize it as a form without root etymology, cf. Winter 1992a:17, LIPP II:804.
For various proposals, cf. the overview presented by Blažek 1999a:193–195, who prefers a connection with the root *ter- ‘to protrude (vel sim.)’ (similarly EIEC:400). Thus, the numeral ‘3’ would be named after the ‘third finger’ because the middle finger is ‘the protruding one’.
3.2Morphological Reconstruction
The Core IE numeral ‘3’ is inflected like an ordinary proterokinetic plural i-stem: Core IE nom. m./f. *tréi̯-es, nom./acc. n. *trí-h2, loc. *tri-sú (cf. PLuw. *trisu ‘thrice’), etc., cf. Blažek 1999a:193.Beside the mere i-stem, Indo-Iranian and Celtic attest a probably old compound (or univerbation?) consisting of the zero-grade (compound) form *tri- and the ancient word for ‘woman’, PIE *h1(e)s-or- ‘woman’, cf. Transmission above. While it is likely that this formation is reconstructable for PIE, due to its archaic look, it is irrelevant for the reconstruction of the Anatolian numeral.Thus, two ablaut grades, viz. *tréi̯- and *tri- have to be reconstructed for the stem. This situation would contradict the assumption that the stem PAnat. *téri- was original. Therefore, it has to be assumed that PAnat. *téri- originates in zero-grade *tri- by developing an epenthetic vowel between *t- and *-r- and a Lindeman variant resulting in *tǝrii̯-V-, which subsequently was changed to *térii̯- by accent retraction. Lyc. B trija ‘3’ fits the reconstructed nom./acc. n. form *trí-h2 (cf. Ved. trī) with the observed development of the glide in *tri- > *t(ǝ)rii̯- (cf. also Lat. tria, Goth. þrija). For the further developments in Anatolian, cf. the PAnat. reconstruction above.For the Hittite developments, cf. Eichner 1992a:66f.
3.2.1Root etymology
As mentioned above, it is preferred here to take PIE *trei̯-/*tri- ‘3’ as unanalyzable.
Blažek 1999a:193–195 opts for interpreting the stem as delocatival while connecting it with the root *ter- ‘to protrude (vel sim.)’, hence loc. *tr-í ‘on the protruder (= middle, third finger)’ → *trei̯-/*tri- ‘3’.
[A.O.]
4. Transmission
PIIr. m./n. *trái̯-/*tri-´ (pl.) ‘3’, cf. nom. m. Ved. tráyas, YAv. θrāiiō, θraiias(⸗ca), acc. Ved. trī́n, YAv. θriš, nom./acc. n. Ved. trī́, YAv. θrī, instr. Ved. m./n. tribhíṣ, YAv. dat. (n.) θribiiō, dat./abl. Ved. tribhyás, loc. Ved. triṣú, gen. Ved. trīṇā́m (hapax, RV.10.185.1), YAv. θraiiąm
PIIr. f. *t(r)i-sr- (pl.) ‘3’, cf. Ved. nom./acc. tisrás, YAv. tišrō, Ved. instr. tisṛ́bhiṣ, dat./abl. tisṛ́bhyas, gen. tisṝṇā́m, YAv. tišrąm, Ved. loc. tisṛ́ṣu
EWAia I:675–678, Grassmann 1996a:555f., Emmerick 1992a:166f., Bartholomae 1904:801f., Emmerick 1992b:293–295
PGk. m./f./n. *tréi̯-/*tri- (pl.) ‘3’, cf. nom. m./f. Att.-Ion. τρεῖς, Dor. (Thera) τρησ, Cret. (Gortyn) τρεεσ (IC IV.72 col. IX.I l. 48), nom./acc. n. τρία, τρα (SEG 26:675, 200–190 BCE), acc. m./f. Ion.-Att. τρεῖς, OAtt.(+) τρῖς,Att.-Ion. acc. τρεῖς is built after the nominative.Cret. (Gortyn) τριινσ (IC IV.81, 480–450 BCE etc.), gen. τριῶν, dat. τρισί, Aeol. τρίσσι, Myc. ti-ri-si etc.
GEW II:921f., GEW III:185, DELG:1131, EDG:1502, Eichner 1982b:221–226, Waanders 1992a:371, LSJ:1812
Arm. nom. erek‘ ‘3’, acc. eris, gen./dat. eric‘ (secondary), instr. eriw-k‘ (Bible+)The anlauting e- is a prothetic vowel, cf. EDAIL:262.
EDAIL:261f., 753 (with lit.), Eichner 1982b:209–215, Winter 1992b:349
PIt. m./f./n. *trei̯-/*tri- ‘3’, cf. nom. m./f. Lat. trēs, Osc. (f.) trís (hapax, ST Cp 24), acc. m./f. Lat. trīs (Lex XII+), Osc. tris (hapax, ST Cp 35), Umbr. trif (hapax, IgT Ib.24), tref, tre, trif, treif (hapax, IgT VIa.22), nom./acc. n. Lat. tria, Umbr. (acc.) triia (hapax, IgT IV.2), triiu(⸗per), trio(⸗per) (hapax, IgT VIIa.51), gen. Lat. trium, Osc. trium, dat./abl. Lat. tribus, Umbr. tris (bis, IgT III.18)
WH II:702f., EM:701f., EDLIL:628, Untermann 2000a:767f., Coleman 1992a:393, OLD²:2172
PCelt. m./n. *trī- ‘3’, cf. nom. m. OIr. trí, MW tri, Corn. try, acc. m. (?)Celtib. (Botorrita) tirís /tri:s/, OIr. trí h, nom./acc. n. OIr. tríL , gen. Ogham m./n. TRIA, OIr. treN, dat. OIr. m. trib etc.
PCelt. f. *ti-s(o)r-, cf. nom. Gaul. tiđres, OIr. téoir, (with secondary *te-) MW, Bret. teir, Corn. ter, tyr, acc. OIr. téora h, dat. OIr. teoraib
EDPC:390, LEIA-T:139f., eDIL s.v. trí, Eichner 1982b:226f., Greene 1992a:507f., 539
PGerm. m./f./n. *þrī-/*þri(j)- ‘3’, cf. nom. m. Run. Dan. þrīʀ, ON þrír, OHG drî, drîe, OE þrī etc., acc. m./f. Goth. þrins, ON þré etc., nom./acc. f. NWGerm. (400 CE) þrijoz, Run. Nor. þrijoʀ, ON þrjár, OHG drîo, thriio, driio, OE þrīo, þrēo etc., nom./acc. n. Goth. þrija, CrimGoth. tria, ON þrjú, OFr. thriu, OHG driu, OE þrīo, þrēo etc., gen. Goth. þrije,The ending -e in Gothic is secondary after m. a-stems, cf. EWAhd II:772.45f.ON þriggja, OHG thriio, drîo etc., dat. Goth. þrim, ON þrim(r), OHG drim, drin etc.
EDPG:546f., GED:365f., de Vries 1977a:622, EWAhd II:770.58–775.57, Kluge & Seebold 2011a:215, Eichner 1982b:228–234, 244–255, 261–292, Ross & Berns 1992a:575–579
BSl. m./f./n. *trij-/*tri- ‘3’, cf. nom. m./f. Lith. trу̃s, Latv. trîs, OPr. *tris (in place name Triskaym), OCS(+) trije, Russ.(+) tri, acc. m./f. Lith. tris, Latv. trîs, nom./acc. n. Lith. *trý (in trý-lika ‘13’), OCS tri, gen. OPr. *treon (in the place name Treonkaymynweysigis ‘trium villarum pratum’), loc. Lith. dial. trisù, OCS trĭxŭ
EDBIL:471, ALEW II:1125f., LEW II:1125, EDSIL:499, Eichner 1982b:216–220, 242f., 256–260, Comrie 1992a:737–740
Alb. nom. m./n. tre ‘3’, nom. f. trī,For details, cf. Blažek 1999a:188.gen./dat. m. treve, gen./dat. f. trīve
AED:463, Demiraj 1997a:387f., Matzinger 2006a:114, Hamp 1992a:906f.
PToch. m. *treyä ‘3’, cf. m. TochA tre, TochB trai, treyThese forms continue virtual *troi̯es which is – contra EWAhd II:774.36–49 – not the regular form of the nominative, but regarded as either influenced by *ku̯etu̯ores ‘4’ (cf. DTB²:340), a contamination with a secondary o-grade formation (cf. Winter 1992a:105), a secondary extension of PToch. *træ (< *träyä < *trei̯es, cf. Pinault 2008a:554) or a continuation of a collective formation (Blažek 1999a:192).
PToch. f. *täryā, cf. f. TochA tri, TochB tarya
DTB²:339f., Adams 2015:165f., Eichner 1982b:235–241, Winter 1992c:104–106