PIE *-eh2 stems

PIE *eh2-stems Miguel Carrasquer Vidal, August 2013 In most IE languages that can provide the required information, the paradigm of the *eh2-stem nouns and adjectives can be reconstructed straightforwardly as follows: sg. nom. acc. voc. gen. dat. loc. ins. abl. pl. nom./voc. acc. gen. dat/abl. loc. ins. du. nom./acc/voc. gen./loc. dat./abl./ins. - m -a -ãs -ãi -ãi - ot < -ah2 < -ah2m < -a(h2) < -ah2as < -ah2ai < -ah2i < -ah2ah1 < -ah2ot -ãs < -ah2as - ns < -ah2ns - õm < -ah2õm os < -ah2 - su < -ah2su - bhis < -ah2bhis -a < -ah2ih1 - ou(s) < -ah2 < -ah2bhioh3 A different picture, however, emerges from Vedic, where we have: sg. nom. acc. voc. gen. dat. loc. ins. pl. nom./voc. acc. gen. dat/abl. loc. ins. du. nom./acc/voc. gen./loc. dat./abl./ins. -ai s i ,- s, - sas - n m yas -ai - yaus - bhy m The bold forms (the sg. oblique) are deviant. The usual explanation for these forms is contamination from the equally feminine *ih2-stems of the -type, but there is really no phonetical rationale within Vedic, with its completely uneroded word-final vocalism and consonantism, for such a confusion or amalgamation of forms. As far as I know, all the attempts to explain the Vedic forms with - - (GDL sg.) have tried to derive the long vowel in these forms from PIE *-eh2-1. Strangely enough, the possibility that the - - here derives from PIE *-o-, Brug a s la , has e er ee o sidered, even though it is well known that *eh2-stems are generally the thematic counterparts of *ih2-stems (i.e. athematic words make their feminines with -ih2, while thematic nouns and adjectives make their feminines with *-eh22). Another persistent notion is that the feminine *eh2-stems share a common origin with collective *-h2, as seen in the o-stem NA n. pl. *-e-h2 > *- . I believe this is wrong, and that instead the feminines in *-ih2 (athematic) and *-eh2 (thematic) both contain a suffix *-ih2, which is ultimately connected with the thematic diminutive suffix *-ikos (< *-iq-). As I argued in ‘a n law, ar ta l ngt n ng an t t at c w l, the thematic vowel was originally *-a-. This *-a-, when stressed, absorbed the *i of the suffix *-ih2, leading eventually to the familiar ending *-eh2(-) of the nom. and acc. sg. I would therefore propose to reconstruct the pre-forms of the *ih2- and eh2-stems approximately as follows: sg. N A V G D L I Ab pl. NV A G DAb L I du. NAV GL DAbI 1 ih2-stems -ix -ix-m -i(x) -íx-âs -ix-ái -íx-ai -ix-át -íx-ât eh2-stems (mobile) -a-[i]x -a-[i]x-m - -i(x) - -íx-âs - -ix-ái - -íx-ai - -ix-át - -íx-ât (immobile) -a-[i]x -a-[i]x-m -a-[i](x) -a-[i]x-âs -a-[i]x-ai -a-[i]x-ai -a-[i]x-at -a-[i]x-ât -ix-asw -ix-ms -ix-ám -ix- asw -ix-swí -ix-bhísw -a-[i]x-ásw -a-[i]x-ms - -ix-ám - -ix- asw - -ix-swí - -ix-bhísw -a-[i]x-asw -a-[i]x-ms -a-[i]x-âm -a-[i]x- asw -a-[i]x-swí -a-[i]x-bhísw -ix-ih -ix-xwáus ? -ix- axw -a-[i]x-ih - -ix-xwáus ? - -ix- axw -a-[i]x-ih -a-[i]x-xwaus ? -a-[i]x- axw And Jasa off s AHIHA-rule tries to explain the ins. sg. form -a (with short a) the same way, from underlying *-eh2-ih2-eh1, i.e. the eh2-stem suffix *-eh2- plus the -stem instrumental sg. -ih2eh1, with dissimilatory loss of the first *h2. The present analysis renders the AHIHA-rule unnecessary. 2 An exception would seem to be itself, the feminine of thematic *deiwos. Through regular sound change (and some analogy), this would have developed into something like: sg. N A V G D L I Ab pl. NV A G DAb L I du. NAV GL DAbI PIE > -ih2 -ih2-m -i -jáh2-as -jáh2-ai -jáh2- i -jh2-áh1 [-jáh2-ot] Vedic -i -- PIE + -ah2 -ah2-m -oi -o- jah2-as -o-jah2-ai -o-jah2- i -o-jh2-ah1 [ -o-jah2-ot] PIE > -ah2 -ah2-m [ -a] [ -ah2-as] [ -ah2-ai] [ -ah2-i] -ah2-ah1 [ -ah2-ot] Vedic -ai -a , -- -ih2-asw -ih2-ns -ih2-õm -ih2- ósw -ih2-sú -ih2-bhísw -n a - -ah2-asw -ah2-ns [ -o-jh2-õm] [ -o-jh2- osw] [ -o-jh2-su] [ -o-jh2-bhisw] -ah2-asw -ah2-ns -ah2-õm -ah2- osw -ah2-su -ah2-bhisw - n a - -ih2-ih1 -ih2-h3óus -ih2- óh3 -yaus - -ah2-ih1 -ah2-ih1 -o-jh2-h3ous [ -ah2-h3ous] [ -o-jh2- oh3] -ah2- oh3 -ai -ayaus - Apart from Indo-Ira ia , e ide e for the o ile eh2-stems can also be found in Slavic. The ins sg. form is - jǫ < - j N) < -ojh2ah1, as in Vedic. In the pronominal paradigm, forms with *-ojah2- and *-ojh2- are found even in the plural and dual: sg. N A G DL I pl. NA G D L I du. NA GL DI ta tǫ t ję toi t jǫ t t t j N)s t j t j N) < -ah2 < -ah2m < -ojah2as < -ojah2ai < -ojh2ah1 ty těxъ tě ъ těxъ tě t n < toisõm < toimos < toisu t < -ah2ns < -ojh2(s)õm < -ojh2mos < -ojh2su < -ojh2 tě toju tě a < toi < tojou(s) t < -oih1 < -ojh2Hou(s) < -ojh2moh3 The dual and plural oblique forms are of course identical with the o-stem forms, which contain the o-stem plural oblique morpheme *-oj-. The standard explanation is therefore to assume that the eh2-stems aquired the o-stem endings. But this obviously does not apply to the singular oblique (GDLI) forms, where the Slavic pronouns fully agree with the Vedic evidence. Additional evidence comes from Armenian. Feminine nouns in -i decline in the singular as follows: NA GD L Ab I -i -woy -w -woy, -w -eaw Some irregular -stems also show reflexes of this ending - . The noun kin Ab n . The numeral mi o e has G mioy or , DL miowm or , Ab o a has GDL n and . The ending - occurs only in feminines, in various oblique functions (GDL). The usual explanation, that it derives from masculine (o-stem) *-o-dhi (as in Greek οἴκοθι at ho e ) is phonetically implausible: intervocalic -dh- gives -r- (or -z-)3, and only an ad hoc sandhi-variant *- - could have given Arm. - . Furthermore, this does not explain why the suffix is added only to feminines. A better explanation is to derive the ending from PIE ins sg. *-ojh2ah1 > *- jj > - . The evidence from Indo-Iranian, Slavic and Armenian thus warrants the reconstruction of a PIE o ile *eh2-stem paradigm, with original Ablaut (1) *-eh2-, (2) *-o-jeh2-, and (3) *-o-jh2-. Variant (1) occurs in forms where the thematic vowel was originally stressed, (2) where the suffix *-ih2 as stressed, a d 3 here either as. Besides this o ile paradig , there also e isted a paradigm with originally fixed stress on the thematic vowel (perhaps created analogically after the thematic masculines and neuters), with only *-eh2-. References Godel, Robert, 1975, An introduction to the study of Classical Armenian, Wiesbaden. MacDonell, Arthur Anthony, 1916/1975, A Vedic Grammar for Students. Bombay. Jasanoff, Jay 2003, Hittite and the Indo-European verb, Oxford. Sihler, Andrew L., 1995, New Comparative Grammar of Greek and Latin, Oxford. Leskien, August, 1969, Handbuch der altbulgarischen (altkirchenslavischen) Sprache, 9th ed., Heidelberg. 3 See my paper on Armenian intervocalic *-dh-.