§ 112. distributive in these forms, thus quaterni (quadrini), octōni,

etc.

1000. The Indo-European original-language seems not to have possessed a word to express 1000.

The two Aryan languages have a common word, Sanskrit sahásra (masc. neut.), Zend hazanra- (neut.).

Greek χίλιο-, Hom. in compounds χλο-, Boiot. χειλίο-, Lesb. xéo-, Dôr. xnλío-, which points to a f.f. *xeλyo-, i.e. gharya-, of obscure origin.

Latin. St. mili-, milli- (neut.), obscure.

STEMS OF THE ORDINAL NUMERALS.

§ 113. The ordinals are, except 2, superlatives, partly however with peculiarities distinguishing them from other superlatives. The forms of the Indo-European original-language for the most part cannot be restored, because the different languages frequently do not coincide in the choice of the suffix.

1-10.

1. Original-language probably pra-ma-, stem pra(before); Sanskrit pra-thamá- from pra- (as prep. 'before') +-thama-, with th for t (§ 52, 2); Greek πрw-Tо-, Dôг. πрâ-TO-, from Tрo-suffix -ta- (§ 106), and with step-formation or lengthening of stem-vowel; Latin pri-mo-, with suffix -mo(§ 107), probably from *pro-imo-, so that -imo- not -mo- has here been added, according to the analogy of other forms in *-timo-; according to Pott (Etym. Forsch. I.2 560) from *prīs-mo- (§ 77, 1, a), *prīs=prius, f.f. pra-yans, comparative of pra-, cf. pris-tino-, pri-die, for *pris-die; according to Corssen (Krit. Beitr., 433) pri- is an archaic form (attested) = prae (prep. 'before'), which is however clearly a case form, and would scarcely have admitted a superlative formation, cf. Umbr. pru-mu- pro-mo-; it would perhaps be difficult to come to a certain decision on this point.

2. Original-language (?); Sanskrit dvi-tiya-, i.e.

*dvi-tya- (§ 15, 2, b), probably therefore a further formation by § 113. means of -ya- from *dvi-ta-; Greek deú-Tepo-, a comparative (§ 105) dev- appears to be a raised-formation from *du from dva; Lat. secundo- is formed not from stem dva-, but from sec, seq (sequi); on the suffix cf. § 89, 2, n. 2.

Indo-European

original-language

3. probably tar-tya- or tra-tya-, -tya- being here also a further-formation of -ta-; Sanskrit tr-tiya, i.e. *tar-tya-, (§ 15, 2, b), as it appears to belong to the root of the stem tr-i-, i.e. tar or tra (v. sup. cardinal 3); Greek Tрí-To-, with superlative suffix -ta-, which helps to form all other ordinals in Greek except 7 and 8; Aiol. Téρ-TO-; Lat. ter-tio-, like Sanskrit.

4. Original-language probably katvar-ta-; Sanskrit katur-tha- (tha=ta), also túr-ya-, tur-iya-, for *katur-ya-, with loss of initial and suffix -ya-, not elsewhere used by itself to form superlatives; we have already noted the combination of -ya- with -ta-; here also we see the frequent phenomenon of two suffixes occurring combined,—as here -t-ya-, i.e. -ta-ya-,— either of which can exercise the same functions as the compound; Greek Téтаρ-Tо- for *TET Faρ-TO-; Latin quar-to- for *quatuor-to-, quator-to- (on these forms cf. Corss. Krit. Nachtr., p. 298, 3).

5. Original-language probably kankan-ta- or perhaps already kan-ta-; Sanskrit panka-má-, Vēd. panka-thā-, with well-known suffixes; Greek Téμπ-то-; Latin quin (c)-to-.

6. Suffix -ta- throughout, which therefore must be ascribed to the original-language with certainty, thus perhaps ksvaks-ta-; Sanskrit šaš-thá-; Greek ex-To-, probably for *E-TO-, cf. Latin sex-to.

7. Indo-European original-language sapta-ma-, or sapta-ta-, or perhaps saptan-ta-?; Sanskrit sapta-má-; Greek ¤ßdo-μo- for *¿πто-μo-, with irregular softening of mutes π into sonants B8, according to the conjecture of G. Curtius and Leo Meyer (cf. Curt. Gr. Et.3 p. 488), through the influence of

§ 113. the μ, before which o forced its way as an auxiliary vowel at a later date (cf. Old Bulg. sed-mů for *sept-mů); archaic and poët. form &ẞdó-μaтo-; Latin septi-mo-.

8. Original-language perhaps aktu-ma-; Sanskrit asta-má-; Greek öydoFo-, with the same weakening as in the case of eßdo-μo-, for *ỏктOFо-, which, as Curt. conjectures, arose from *ỏKTFO-, and whose weakening of T to yd must be ascribed to the F (cf. No. 7); on the other hand oydoFo- and Latin octauo- point to a common f.f. aktāv-a-, which is opposed to Curtius' supposition. The suffix here is therefore only -a-, which is added to the raised stem aktu- (consequently we must not assume either suffix -va- as in *par-va-, or still less -machanged into -va-).

9. Indo-European original-language doubtful whether with suffix -ma- (nava-ma-), or with -ta- (navan-ta); Sanskrit nava-má-; Greek eva-To-, evva-To-, probably shortened from *eve Fα-To-; Latin no-no- from *nou-no-, *noui-no-, probably from *noui-mo- by assimilation to the initial sound.

10. Original-language doubtful, as in the case of 9, whether daka-ma- or dakan-ta-; Sanskrit daça-má-; Greek Séκα-TO-; Latin deci-mo-.

11-19. Originally by means of two words.-Sanskrit. Here, as in other compounds, the final a of the second element of daçan-, which has lost its n, serves likewise for an adjectiveforming suffix, e.g. 11, ēkā-daçá-, from ékā-daçan-; 12, dvā-daçā-, etc. Here consequently we see suffix a also serving to form a superlative (as in Gk. Lat. *aktāv-a-, 8). Greek throughout -δέκα-το- ; 11, ἑν-δέκα-το-; 19, ἐννεα-και-δέκα-το-. Latin. 11, un-deci-mo-; 12, duo-deci-mo-, and the remaining numerals by separate words.

20-90. Originally by two words. -Sanskrit either with -tama-, e.g. 20, vìçati-tamá-; 30, tricati-tamá-; or by suffix -a- like 11-19, with loss of final -t, -ti, e.g. vîçá-, trîçá-. Greek.

To -koti-, -kovтα-, was added suffix -70-, in such a way that -KOTI- and -KOVтa- were shortened to -κOT-; hence arose *-KOT-TOand by rule ($ 68, 2) -κοσ-το-, thus 20, εἰκοστό- ; 30, τριακοστ -Tó-, etc. Latin. Suffix -tumo-, -timo-, added to the suffix *-cinti-, *-cinta-, which loses its final; or rather to an earlier *-centi-, *-centa-; thus *-cent-tumo-, and thence regularly (77, 1, b) -censumo-, -cēsimo-, and -gesimo- with c softened to g. E.g. 20, early vicensumo-, hence vicēsimo-, vigēsimo-, f.f. would therefore be some such form as *dvi-(da) kanti-tama-; 40, quadra-gensimo-, -gēsimo-, etc.

100-1000. 100. Sanskrit çata-tamá-; Greek, with suffix -σ70-, apparently through the analogy of -lσTο- (§ 106), formed from -To-, EKATO-σTÓ-; Latin, according to analogy of the tens, cent-ēsimo-, as though -ēsimo- were the suffix (from *cent-tesimo- would have arisen *censēsimo-, § 77, 1, b).

200-900. Sanskrit with çata-tamá-; Greek with -σTO(v. 100); e.g. 200, Siakoσio-σтó-; Latin with centēsimo, e.g. 200, du-centésimo-, octin-gentēsimo-, etc.

1000. Sanskrit sahasra-tamá-; Greek with -σTO-, xiλo-σTÓ-; Latin with -ēsimo-, mill-ēsimo-.

$113.

STEPHEN AUSTIN AND SONS, PRINTERS, HERTFORD,