Though there are problems with the reconstruction of a common form for the number one (see below), the following cardinal numbers one to ten are traditionally reconstructed for later Proto-Indo-European (for additional information, cf. Blaek 1999:141324; Meillet 1964:409413; Beekes 1995:212213; Szemernyi 1960; GamkrelidzeIvanov 1995.I:740 744; Sihler 1995:404433; Adrados 1975.II:871877; AdradosBernabMendoza 1995 1998.III:127131):
The numbers in Anatolian are, for the most part, not known inasmuch as they are written ideographically (cf. Luraghi 1997:27). The number seven occurs in Hittite in the ordinal (dat.) i-ip-ta-mi-ya seventh (cf. Sanskrit saptam- seventh; Latin septimus seventh) (cf. Sturtevant 1951:30, 44, 60, 63, 77, and 87; Kronasser 1956:152; Benveniste 1962:83). The number three is also represented in Hittite in (adv.) te-ri-ya-an-na for the third time, and the military title te-ri-ya-al-la, tar-ri-ya-na-al-li third-in-command, officer of the third rank (cf. Kronasser 1956:151; Benveniste 1962:82; Blaek 1999:186187), apparently to be read *tri- three (cf. Benveniste 1962:86), while two is found in Hittite in the military title du-ya-na-al-li second-in-command, officer of the second rank, the compound ta-a-i--ga-a, da-a-i--ga-a, ta-a--ga-a two years old (da-/ta- two + i--ga-a yearling), da-a-an, ta-a-an a second time; second, and (nom. sg. c.) da-ma-a-(i-)i second, other (cf. Sturtevant 1951:34, 58, 61, 67, and 110; Kronasser 1956:151; Benveniste 1962:81), and in Hieroglyphic Luwian tu-wa/i-zi two (cf. Laroche 1960:206; Meriggi 1962:136; Blaek 1999:164). All three of these forms 2 agree with what is found in the non-Anatolian Indo-European daughter languages. The forms in the Anatolian languages for the number four, however, differ from those that are found elsewhere: Proto-Anatolian *meyu- four > Hittite (nom. pl.) mi-e-(ya-)wa-a, (acc. pl.) mi-e-- u, (gen. pl.) mi-i--wa[-a] four, Luwian mauwa- four (instr. pl. ma-a-u-wa-a-ti) (cf. Benveniste 1962:81; Laroche 1959:70; Blaek 1999:201202). Two basic stems may be reconstructed for the number one: *Hoy- and *sem- (cf. Sihler 1995:404407; Fortson 2004:131). The underlying meaning of the first stem appears to have been single, alone, while that of the second stem appears to have been together (with) (cf. Szemernyi 1996:222; Blaek 1999:155). The first stem only occurs with various suffixes: (1) *Hoy-no- (cf. Latin nus one [Old Latin oinos]; Old Irish en, in one; Gothic ains one; Old English n one; Old High German ein one; Old Church Slavic in one it is also found in Greek ovq, oivo roll of one [in dice]); (2) *Hoy-wo- (cf. Avestan ava- one; Old Persian aiva- one it is also found in Greek oio alone, lone, lonely [Cyprian oi+o]); (3) *Hoy- ko- or *Hoy-ko- (cf. Sanskrit ka- one; Mitanni [Proto-Indic] aika- one). The second stem is found in Greek: Attic (nom. sg. m.) ci one, Doric one, Cretan cv (< *cv < *c < *sems) one; Attic (f.) iu (< *o-iu) one. It is also found in Armenian mi one. To complicate matters, the various forms of the ordinal found in the daughter languages are based upon yet another Proto-Indo-European stem: *per(H)-/*p3(H)- first (> *p3H-wo-, *p3H- mo-, *prey-mo-, *prey-wo-, *proH-to-, *proH-mo-, etc. [for details, cf. Blaek 1999:141 162; see also Szemernyi 1996:228; Sihler 1995:427428]). There was a variant form *tw-i- (traditional reconstruction *dw-i-) two in Proto-Indo- European that was used in compounds (cf. GamkrelidzeIvanov 1995.I:742) and in the adverbial form *tw-i-s twice (cf. Latin bis twice [Old Latin duis]; Sanskrit d(u)v twice; Avestan bi twice; Greek oi twice; Middle High German zwir twice). The regular form for the number two is traditionally reconstructed as a dual *duw/*dw (Szemernyis reconstruction), though the dual forms may have arisen in the early prehistory of the individual daughter languages themselves (cf. Sihler 1995:408). This view is quite attractive, and I would reconstruct *t(u)w-o- as a plural (originally indeclinable) and not as a dual at the Proto-Indo- European level (the plural is still found, for example, in forms such as Greek [nom. pl.] oto, [nom.-acc. pl.] ouoiv). Attempts to come up with an etymology within Indo-European itself for this number have met with little success (cf. Blaek 1999:175179). That the core form was *t(u)w- (cf. Blaek 1999:178; Villar 1991:136154; ErnoutMeillet 1979:187188) is shown by the fact that the thematic vowel *-o- could be added directly to the that form, on the one hand, to yield the form traditionally reconstructed for the independent word for the number two, while, when used in compounds or to express twice, the extension *-i- could be added directly to the core form instead. Thus, we get *t(u)w-o- ~ *t(u)w-i- two. There are several forms in Hittite that point to an alternative form for two in Proto- Indo-European these are: the compound ta-a-i--ga-a, da-a-i--ga-a, ta-a--ga-a two years old (da-/ta- two + i--ga-a yearling), da-a-an, ta-a-an a second time; second, and (nom. sg. c.) da-ma-a-(i-)i second, other. These forms point to a Proto-Indo-European *te- /*to- (earlier *te-/*ta-) two (cf. Sturtevant 1951:61 [Sturtevant reconstructs Proto-Indo- Hittite *do- two]; Benveniste 1962:7886 [Benveniste brings in data from non-Anatolian Indo-European daughter languages to support his views]). There is absolutely no way to reconcile *te-/*to- with *t(u)w-o/i- phonologically so that they can be convincingly combined 3 in a single reconstruction (AdradosBernabMendoza 19951998.III:138 note the problems involved and discuss proposed solutions). Consequently, two competing forms must be reconstructed for the number two in Proto-Indo-European. If the Proto-Indo-European number ten were originally a compound meaning two hands, that is, *te- two + *ki(t)- hand, as some have suggested (cf. Szemernyi 1960:69 and 1996:224, fn. 16; Markey 1984:284285; Justus 1988:533; GamkrelidzeIvanov 1995.I:747; AdradosBernabMendoza 1995 1998.III: 131; but rejected by Blaek 1999:295296), it would provide additional evidence for reconstructing two separate forms for the number two. This situation raises the question as to why there should be two alternative forms for the number two in Proto-Indo-European. A possible answer is that *te-/*to- may have been the native form, while *t(u)w-o/i- may have been a borrowing. Given the geographical location of the Indo-European homeland in the vicinity of the Black Sea near speakers of early Northwest Caucasian languages, these languages might have been a possible source for the *t(u)w-o/i- form. Indeed, there is a striking resemblance between Proto-Indo-European *t(u)w-o/i- two and similar forms for this number in Northwest Caucasian: Proto-Northwest Caucasian *tqo- two > Proto-Circassian *t# two, Proto-Ubykh *tq (> * tqa) twice, Proto-Abkhaz- Abaza *t two (cf. Colarusso 1992:45). Kuipers (1975:19) reconstructs Proto-Circassian *Tq(a) two (> Bedux t(a)/t(a)w, -t(a) two [twice]; Kabardian -ta only in mzamta more than once, repeatedly, literally, not-once-not-twice). Colarusso (1992:45) derives the Proto-Indo-European form for the number two from *t#, which he claims first became *t# and then *t(u)w-o- [traditional *d(u)w-o-]. Colarusso (1992) documents many other similarities between Proto-Indo-European and Northwest Caucasian. These similarities lead Colarusso to think about possible genetic relationship. I prefer to see the similarities to be due to the fact that the Indo-Europeans occupied territory north of and between the Black and Caspian Seas that was originally inhabited by speakers of early Northwest Caucasian languages. We can further speculate that *t(u)w-o/i- two eventually replaced the native Proto-Indo-European word for two, which survived only in relic forms and in the word for the number ten (*te- ki(t)). The Proto-Indo-European word for the number three is completely straightforward and can be reconstructed *tr-ey-/*tr-i-. Sanskrit (nom.-acc.) tisr and related forms in other Indo- European daughter languages are dissimilated from *tri-sr- (cf. Sihler 1995:410). The word for the number four is traditionally reconstructed *ketwores (so Szemernyi; Brugmann reconstructs *qetor-). The most convincing etymology is that offered by Burrow (1973:259) (see also Beekes 1987:219):
4. This number is formed on the basis of a root ket which seems originally to have meant something like angle (cf. Lat. triquetrus triangular), whence square and from that four. In the masc. and neut. (catvZras, catvZri, Lat. quattuor, etc.) the stem is formed by means of the suffix -var, with adjectival accent and v0ddhi in the nominative. In the other cases (acc. catras, etc.) the suffix has the weak form according to the general rule. A neuter noun *ctvar, or its IE prototype, is presupposed by the thematic extension catvara- square, crossroads. Elsewhere the simple r-suffix may appear (Gk. Dor. tctopc, Lat. quarter), or the elements of the suffix may be reversed (Av. aru-).
In accordance with Burrows views, the form *ket-wor- square may be reconstructed for later Proto-Indo-European. It was preserved in Sanskrit in the thematic derivative catvar-m 4 quadrangular place, square, crossroads (cf. Mayrhofer 19561980.I:371). It was this form that served as the basis for the number four found in the non-Anatolian daughter languages: (nom. pl.) *ket-wgr. Curiously, the suffix *-wor- is replaced by *-sor- in the feminine (cf. Sanskrit ctasra). Thus, the root was *ket-, to which different suffixes could be added. It is intriguing to speculate that *ket-wor- may have replaced an earlier form for four, which is preserved in Anatolian. On the other hand, some have suggested that the original form for the number four was *Hokto- and that eight was simply the dual of this stem, whose underlying meaning was two fours (cf. GamkrelidzeIvanov 1995.I:747; Burrow 1973:260). This suggestion finds support in Kartvelian (cf. Blaek 1999:268). The number four is reconstructed as *otxo- in Proto-Kartvelian, and this is generally taken to be a loan from Proto- Indo-European (cf. Klimov 1998:145146; FhnrichSardshweladse 1995:269; Gamkrelidze Ivanov 1995.I:775 [GamkrelidzeIvanov reconstruct Proto-Kartvelian *(o)t(o)-]). I favor this explanation and consider *Hokto- to be the original form of the number four in Early Proto-Indo-European. It was replaced by *meyu- in Anatolian, while, in the Proto-Indo- European antecedent of the non-Anatolian daughter languages it was replaced by *ket-wor-. It only survives in the later Proto-Indo-European form for the number eight, *HoktoH(w), a dual formation originally meaning two fours. No doubt, this replaced an earlier form for the number eight, which, unfortunately, can no longer be recovered. One final comment may be made here: in Etruscan, there is a number hu. Its exact meaning is uncertain it could be six, or it could be four (cf. Cristofani 1991:77; Blaek 1999:235; BonfanteBonfante 2002:9495). If it is six, then the number a is four. On the other hand, if it is four, then the number a is six. Without going into the whole question here of whether Etruscan and Proto-Indo-European are ultimately genetically related, we can say that hu more closely resembles Proto-Indo-European *Hokto- four, while a more closely resembles Proto-Indo-European *s(w)eks six (Szemernyis reconstruction). As noted by Blaek (1999:211 and 235) and Briquel (1994:329), support for considering the meaning of hu to be four comes from the identification of hu in the Pre-Greek name 'Yttqviu for the city Tetrapolis (Tctpuaoti, composed of tctpu- four and aoti city) in Attica. This may provide another piece of evidence in support of considering *Hokto- to have been the original form for the number four in Proto-Indo-European. The number five was *penke (Brugmann *peqe) in Late Proto-Indo-European. It is usually identified with words for fist and finger: (1) Proto-Indo-European *pk-sti- fist > Proto-Germanic *fustiz > West Germanic *fsti- > *fsti- > Old English fst fist; Old Frisian fest fist; Middle Low German fst fist (Dutch vuist); Old High German fst fist (New High German Faust) (cf. Mann 19841987:968 *p$stis [*pqstis ?] fist; Onions 1966:358; KlugeMitzka 1967:187; KlugeSeebold 1989:205); Serbian Church Slavic pst fist; (2) Proto-Indo-European *penk-r- finger > Proto-Germanic *figraz finger > Gothic figgrs finger; Old Icelandic fingr finger; Old English finger finger; Old Frisian finger finger; Old Saxon fingar finger; Old High German fingar finger (New High German Finger) (cf. Feist 1939:150; Lehmann 1986:114; De Vries 1977:120; KlugeMitzka 1967:198; KlugeSeebold 1989:215). Though not without problems from a phonological point of view, the above comparisons can hardly be questioned. Ultimately, all of these forms may indeed go back to a verbal stem *penk- to take in hand, to handle, as suggested by Horowitz (cited by Blaek 1999:228), though it should be mentioned that this putative verb stem is not attested in 5 any of the daughter languages. Blaek (1999:229) notes that the meanings fist, etc. are primary. Several different reconstructions are possible for the Proto-Indo-European word for the number six: *seks, *sweks, *kseks, *ksweks, *weks (for details, cf. Blaek 1999:234 242; see also Sihler 1995:413). This number was also borrowed by Kartvelian: Proto-Kartvelian *ekw- six (cf. Klimov 1998:48 *eksw-; FhnrichSardshweladse 1995:125126 *eksw-; Schmidt 1962:107 *ekw-/*eku; GamkrelidzeIvanov 1995.I:775 *ekw-). Sihler (1995:413) takes *weks (he writes *we#s) to be the original form and considers the initial *s- to be a secondary development (imported from the number seven) (Szemernyi 1996:222 and Beekes 1995:213 express the same view). Thus, in accordance with Sihlers views, the earliest form of the Proto-Indo-European number six should be reconstructed as *weks. As Sihler notes, when *s- was merely added to *weks, the result was *sweks, but when it replaced the initial consonant, the result was *seks. The Iranian forms pointing to original *ksweks (cf. Avestan xva six) appear to be due to developments specific to Iranian and should not be projected back into Proto-Indo-European (cf. Sihler 1995:413). The Proto-Indo-European word for the number seven, *septi (Brugmann *septi), is sometimes considered to be a loan from Semitic (cf. Blaek 1999:256257; Gamkrelidze Ivanov 1995.I:747). That this number is ancient in Indo-European is clear from the fact that it is found in Hittite. We have already discussed the number eight. For nine, Proto-Indo-European most likely had *new (cf. Szemernyi 1996:223). Other possible reconstructions are *newi and *Hnew/i (cf. Brugmann 1904:365 *ne, *en; Blaek 1999:283 *Hnewi; Meier-Brgger 2003:234 *hn; Watkins 1998:67 *hnw; Haudry 1979:68 *nwi/; Burrow 1973:260; Sihler 1995:415 *Hnw; GamkrelidzeIvanov 1995.I:744 *neu(e)n; Buck 1933:230 [Buck takes Greek cvvcu to be a blend of *cv+u and *vc+u]; Rix 1992:172 *n). As noted above, the Proto-Indo-European number ten must originally have been a compound meaning two hands, that is, *te- two + *ki(t)- hand. The Proto-Indo-European word for the number hundred is traditionally reconstructed as *(d)$itm it is usually considered to be a derivative of *de$i(t) ten and meant something like ten tens (cf. Szemernyi 1996:226; Watkins 1998:67; Meier-Brgger 2003:235; Beekes 1995:213; GamkrelidzeIvanov 1995.I:744). Though there was probably no common Proto-Indo-European word for thousand, the form *geslo- served as the basis for the Indo-Iranian, Greek, and Latin terms (cf. Szemernyi 1996:227; Beekes 1995:216; Meier-Brgger 2003:235; Meillet 1964:414; Brugmann 1904:368). According to Bengtson (1987:260261), this form is to be derived from Proto-Indo-European *ges- hand (he writes *hes-) plus a suffix *-lo-. We may now summarize our findings. The numbers one to ten may be reconstructed as shown below for the earliest stage of Proto-Indo-European. However, inasmuch as Anatolian corroboration is lacking for several of the numbers, the following reconstructions must be considered provisional:
1 *Hay- (later *Hoy-), *sem-, *per(H)-/*p3(H)- 2 (earliest form) *te/a-; (later also) *t(u)w-a- (still later *t(u)w-o-), *t(u)w-i-; though originally a plural form, this was later reinterpreted as a dual. 6 3 *tr-ey-/*tr-i- 4 *Hok-ta- (perhaps with original, non-apophonic -o- in the first syllable) (later *Hok-to-); replaced by *meyu- four in the Anatolian languages and by *ket-wor- four-sided, square in the non-Anatolian daughter languages. 5 *penke 6 *weks (later also *sweks ~ *seks) 7 *septi 8 Original unknown; replaced by (dual) *HoktoH(w) two fours = eight in the non-Anatolian daughter languages. 9 *new 10 *te-ki(t) (original meaning two hands).
References
Adrados, Francisco R[odrguez] 1975 Lingstica indoeuropea [Indo-European Linguistics]. 2 vols. Madrid: Gredos. Adrados, Francisco R., Alberto Bernab, and Julia Mendoza 19951998 Manual de lingstica indoeuropea [Handbook of Indo-European Linguistics]. 3 vols. Madrid: Ediciones Clsicas. Beekes, Robert S. P. 1987 The Word for Four in Proto-Indo-European, Journal of Indo-European Studies 15.1/2:215219. 1995 Comparative Indo-European Linguistics: An Introduction. Amsterdam and Philadelphia, PA: John Benjamins. Bengtson, John D. 1987 Notes on Indo-European 10, 100, and 1000, Diachronica IV.1/2: 257262. Benveniste, mile 1962 Hittite et indo-europen [Hittite and Indo-European]. Paris: Adrien- Maisonneuve. Blaek, Vclav 1999 Numerals. Comparative-Etymological Analysis and their Implications. Brno: Masarykova Univerzita v Brn. Bonfante, Giuliano, and Larissa Bonfante 1983 The Etruscan Language: An Introduction. New York, NY: New York University Press. [2002] [Revised edition. Manchester: Manchester University Press.] Briquel, Dominique 1994 trusque et indo-europen [Etruscan and Indo-European], in: Franoise Bader (ed.), Langues indo-europennes [Indo-European Languages]. Paris: CNRS ditions, pp. 319330. Brugmann, Karl 7 1904 Kurze vergleichende Grammatik der indogermanischen Sprachen. Auf Grund des fnfbndigen Grundri der vergleichenden Grammatik der indogermanischen Sprachen von K. Brugmann und B. Delbrck [Concise Comparative Grammar of the Indo-European Languages. Based upon the Five Volume Elements of the Comparative Grammar of the Indo-Euro- pean Languages by K. Brugmann and B. Delbrck]. Berlin: Walter de Gruyter. Reprinted 1970. Buck, Carl Darling 1933 Comparative Grammar of Greek and Latin. 10th impression 1966. Chicago, IL, and London: University of Chicago Press. Burrow, Thomas 1973 The Sanskrit Language. 3rd edition. London: Faber & Faber. Colarusso, John 1992 Phyletic Links between Proto-Indo-European and Proto-Northwest Caucasian, in: Howard I. Aronson (ed.), The Non-Slavic Languages of the USSR: Linguistic Studies (Second Series). Chicago, IL: Chicago Linguistic Society, pp. 1954. Reprinted in Mother Tongue 21:820 (1994). Cristofani, Mauro 1991 Introduzione allo studio delletrusco [Introduction to the Study of Etruscan]. Leo S. Olschki Editore. De Vries, Jan 1977 Altnordisches etymologisches Wrterbuch [Old Norse Etymological Dictionary]. Reprint of 2nd edition. Leiden: E. J. Brill. Ernout, Alfred, and Antoine Meillet 1979 Dictionnaire tymologique de la langue latine: Histoire des mots [Etymological Dictionary of the Latin Language: History of Words]. 4th edition. Paris: Klincksieck. Fhnrich, Heinz, and Surab Sardshweladse 1995 Etymologisches Wrterbuch der Kartwel-Sprachen [Etymological Dic- tionary of the Kartvelian Languages]. Leiden: E. J. Brill. Feist, Sigmund 1939 Vergleichendes Wrterbuch der gotischen Sprache [Comparative Dictionary of the Gothic Language]. 3rd edition. Leiden: E. J. Brill. Fortson, Benjamin W., IV 2004 Indo-European Language and Culture: An Introduction. Oxford: Blackwell. Gamkrelidze, Thomas V., and Vjaeslav V. Ivanov 1995 Indo-European and the Indo-Europeans: A Reconstruction and Historical Typological Analysis of a Protolanguage and a Proto-Culture. 2 vols. English translation by Johanna Nichols. Berlin, New York, NY, and Amsterdam: Mouton de Gruyter. Justus, Carol F. 8 1988 Indo-European Numerals and Numeral Systems, in: Yol L. Arbeitman (ed.), A Linguistic Happening in Memory of Benjamin Schwartz: Studies in Anatolian, Italic, and Other Indo-European Languages. Louvain: Peeters, pp. 521541. Klimov, G[eorgij] V. 1998 Etymological Dictionary of the Kartvelian Languages. Berlin and New York, NY: Mouton de Gruyter. Kluge, Friedrich, and Walther Mitzka 1967 Etymologisches Wrterbuch der deutschen Sprache [Etymological Dictionary of the German Language]. 20th edition. Berlin and New York, NY: Walter de Gruyter. Kluge, Friedrich, and Elmar Seebold 1989 Etymologisches Wrterbuch der deutschen Sprache [Etymological Dictionary of the German Language]. 22nd edition. Berlin and New York, NY: Walter de Gruyter. Kronasser, Heinz 1956 Vergleichende Laut- und Formenlehre des Hethitischen [Comparative Phonology and Morphology of Hittite]. Heidelberg: Carl Winter. Kuipers, A[ert] H. 1975 A Dictionary of Proto-Circassian Roots. Lisse: Peter de Ridder Press. Laroche, Emmanuel 1959 Dictionnaire de la langue louvite [Dictionary of the Luwian Language]. Paris: Adrien-Maisonneuve. 1960 Les hiroglyphes hittites [The Hittite Hieroglyphs]. Vol. I. Paris: Centre International de la Recherche Scientifique. Lehmann, Winfred P. 1986 A Gothic Etymological Dictionary. Leiden: E. J. Brill. Luraghi, Silvia 1997 Hittite. Munich: Lincom Europa. 1998 The Anatolian Languages, in: Anna Giacalone Ramat and Paolo Ramat (eds.), The Indo-European Languages. London and New York, NY: Routledge, pp. 169196. Mann, Stuart E. 19841987 An Indo-European Comparative Dictionary. Hamburg: Helmut Buske Verlag. Markey, Thomas L. 1984 The Grammaticalization and Institutionalization of Indo-European Hand, Journal of Indo-European Studies 12.3/4:261292. Mayrhofer, Manfred 19561980 Kurzegefates etymologisches Wrterbuch des Altindischen [Short Etymological Dictionary of Old Indic]. 4 vols. Heidelberg: Carl Winter. Meier-Brgger, Michael 2003 Indo-European Linguistics. Berlin: Walter de Gruyter. Meillet, Antoine 9 1964 Introduction ltude comparative des langues indo-europennes [Introduction to the Comparative Study of the Indo-European Languages]. University, AL: University of Alabama Press. Reprint of 8th edition (1937). Meriggi, Piero 1962 Hieroglyphisch-Hethitisch Glossar [Hieroglyphic Hittite Glossary]. 2nd edition. Wiesbaden: Otto Harrassowitz. Onions, C. T. (ed.) 1966 The Oxford Dictionary of English Etymology. Oxford: Clarendon Press. Rix, Helmut 1992 Historische Grammatik des Griechischen: Laut- und Formenlehre [Historical Grammar of Greek: Phonology and Morphology]. 2nd edition. Darmstadt: Wissenschaftliche Buchgesellschaft. Schmidt, Karl Horst 1962 Studien zur Rekonstruktion des Lautstandes der sdkaukasischen Grundsprache [Studies on the Reconstruction of the Sound Structure of the South Caucasian Parent Language]. Wiesbaden: Kommissionsverlag Franz Steiner GMBH. Sihler, Andrew L. 1995 New Comparative Grammar of Greek and Latin. New York, NY, and Oxford: Oxford University Press. Sturtevant, Edgar H. 1933 A Comparative Grammar of the Hittite Language. Baltimore, MD: Linguistic Society of America. [1951] [Revised edition. Vol. I. New Haven, CT: Yale University Press.] Szemernyi, Oswald 1960 Studies in the Indo-European System of Numerals. Heidelberg: Carl Winter. 1996 Introduction to Indo-European Linguistics. Translated from the 4th edition (1990) of Einfhrung in die vergleichende Sprachwissenschaft [Introduction to Comparative Linguistics], with additional notes and references. Oxford: Clarendon Press. Villar, Francisco 1991 The Numeral two and its Number Marking, in: Perspectives on Indo- European Language, Culture, and Religion: Studies in Honor of Edgar C. Polom. McLean, VA: Institute for the Study of Man, vol. 1, pp. 136 154. Watkins, Calvert 1998 Proto-Indo-European: Comparison and Reconstruction, in: Anna Giacalone Ramat and Paolo Ramat (eds.), The Indo-European Languages. London and New York, NY: Routledge, pp. 2573.