CC BY
88
I. Yakubovich
Indo-European *ma 'to grow'
Резюме. В настоящей статье обосновывается реконструкция нового праиндоевропейского глагольного корня *ma 'расти'. Реконструкция произведена на основания сравнения трех групп лексем, рассматривавшихся как этимологически несвязанные в предшествующей научной литературе. К рефлексам реконструированного глагола относятся восточноиранские глагольные основы со значением 'быть, становиться' (напр. хотано-сакское ката- и тумшукско-сакское ramy-), хеттское mai- / miya'расти', латинское matUrus 'зрелый' и русское матёрый.
Ключевые слова: индоевропеистика, этимология, семантическая реконструкция, иранистика, анатолистика, согдийский, хеттский, латинский, старославянский.
The purpose of this paper is to argue for the existence of a previously unrecognized Proto-Indo-European verbal root by comparing three groups of forms that were commonly regarded as etymologically unrelated.1 The reflexes of Indo-European *ma 'to grow' are preserved as finite verbal forms in Iranian (*ma 'to become, be') and Anatolian (Hitt. mai- / miya-hi 'to grow'), and as adjectival derivatives in Slavic and Italic (Russian матёрый and Latin maturus 'full-grown').2
1 This paper benefited from the discussion with Dzhoj I. Edelman (Moscow), H. Craig Melchert (UCLA) and Nicholas Sims-Williams (Cambridge, UK), as well as the interaction with the audience of the 21st UCLA Indo-European Conference (2009), where its content was first presented. I am also grateful to Pavel Basharin (Moscow), Benjamin Fortson (Ann Arbor) and Alexei Kassian (Moscow) for their bibliographic help and to Douglas Hitch (Whitehorse, Canada), who took pains to improve the style of this contribution. All the shortcomings are mine alone.
2 This paper is written in the non-laryngealist framework. Consequently, I use the reconstructions *ma and not **meh2 'to grow', *me and not **mehi 'to measure' etc., unless I cite the works belonging to the followers of the traditional Laryngeal Theory. I would like, however, to stress the fact that the choice of notation does not impact
1. Iranian. The starting point of my analysis is the East Iranian verbal root reconstructed as *maH 'to be, become' in Cheung 2007: 257. The reflexes of this root cited in Cheung's dictionary include Khotanese hama- (*fra+ma-) 'to become, be', Sogdian suppletive past stems m't and wm't (*wi+ma-) 'to be, remain', Yazghulami mi-/mad 'to be', Wakhi umuy-/ umut (*a+ma-) 'to be', used in modal forms, and Yidgha farms (*fra+ma-), glossed by Morgenstierne through Persian basad, the present subjunctive form of budan 'to be'. To this one can add a few more forms. Hitch (1992: 93 with fn. 6) plausibly compares Khotanese hama- with Tumshuqese ramy- 'to be, become', both presumably derived from the same prefixed variant (*fra+ma-).3 I further suggest that Sogd. ''mt(')yc 'true' with its derivatives (Gharib 1995, #171-2, 174-6) is ultimately derived from the past stem of *a+ma- with the same semantic change that one can observe in English sooth 'true, truth' < IE *(b)sont- 'being'. The same root etymology can be suggested for Sogd.''mnk 'truth' < Ir. *a-ma-na-ka- (Gharib 1995, #164-5).4
The etymological dictionaries available to date mention a variety of unsustainable hypotheses regarding the Indo-European origin of Ir. *ma 'to be, become' or its reflexes in individual daughter languages. Bailey (1979: 483a) attempts to
my etymological conclusions. Those, who prefer the laryngealist reconstructions, will have no difficulties performing the appropriate mechanical substitutions.
3 Hitch (ibid.) suggests including Avestan *fra+m(a)y- 'to turn into' into the same family of cognates. However, pace Bartholomae (1904: 1166), Av. framita- may mean not 'turned into' but 'appeared' and represent an intransitive match to Av. a-maii- 'to show'. If so, it can be linked to the Indo-European verbal root *ma 'to show, give sign' (Rix 2001: 425; meh2). This root has other likely derivatives in Iranian, notably Manichaean Middle Persian nm'y- 'to show'. Regrettably, the Iranian reflexes of IE *ma 'to show, give sign' are commonly lumped together with those of IE *me 'to measure' (e.g. in Rix 2001, 424-25 and Cheung 2007: 254-57).
4 N. Sims-Williams (pers. comm.) informs me that Sogd. ''mtk, cited in Cheung 2007: 257 as a derivative of ''m- 'to come to' < *a+ma-, is apparently extracted from the outdated glossary Reichelt 1931: 46. This word in how recognized as meaning 'ready' and should be treated together with the other spellings of the same lexeme, which is ultimately derived from IE *me 'to measure' (Cheung 2007: 255-56).
reconstruct Khotanese hama- as *ham+i-, deriving it from Ir. *ai 'to go' and citing Persian sudan 'to become' < Middle Persian sudan 'to go' as a typological parallel. Steblin-Kamenskij (Сгеблин-Каменский 1999: 455) favorably cites the
suggestion of Szemerenyi (1971: 517), who connects Sogd. (w)m't and its cognates with Indo-European *men 'to remain' (Rix 2001: 437). As a less likely alternative, he mentions Sokolova's opinion that the same Iranian root is to be reconstructed as *mai/mi 'to be' (Соколова 1973: 23). Cheung 2007 does not offer a new Indo-European etymology for the root under discussion, but mentions the same suggestions by Sokolova and Szemerenyi, indicating that the last one is "morphologically difficult". At the same time, he adduces a possible parallel to the semantic development postulated by Szemerenyi, namely Tocharian А/B mask- < *mn-ske- 'to be (located)', which is derived from the same IE *men 'to remain'. It is appropriate to address these suggestions before embarking upon what is considered here as the correct etymology of Ir. *ma 'to become'.
Bailey's hypothesis is contradicted by the data collected in the same lexical entry. The attempt to deny the connection between the synonymous Saka verbs, Khot. hama- and Tumsh. ramy- 'to become, be', which are both amenable to the reconstruction of *ma with the same historical prefix fra-, appears to be strictly ad hoc.5 The genetic relationship between Khot. hama- and Wakhi bzmbzy-/bzmbzt, which Bailey accepts, is incompatible with his reconstruction hama- < *ham+i-, since Wakhi ы represents a regular outcome of Ir. *a (Сгеблин-Каменский 1999: 56-7). The semantically vague comparison between Sogdian wm't 'was' and Parthian wm'y-/wm'd 'to experience, endure', which Bailey (1979: 87a) uses in order to separate the Sogdian form from Khot. hama-, was shown to be non-compelling already in Szemerenyi 1971.
5 Douglas Hitch (pers. comm.) indicates that although much is not understood about Tumshuqese phonology, Tumsh. ramy-, consistently written with the Fremdzeichen ra, could plausibly represent /rami-/.
The etymology involving IE *men- 'to remain' is largely based upon the Sogdian data. Already Benveniste (1929: 13) noticed that the present stem m(')yn- 'remain, be' < Ir. *manaya-belonged to the same paradigm as the past stem (w)m't in Sogdian. This observation can be confirmed through the analysis of the paraphrase in the Sogdian Marriage Contract (Yakubovich 2006). The clause Nov. 3 rev. 2-4 in the main text of the agreement rtnms kS ZNH cttyH ZNH m'ny w'n'kw fi’ty ZY 'M ’wttkyn pr'yw wSwH L' my-n'tk'm 'And if it comes to mind to Chat that she will not remain with Ot-tegin...' is rephrased in Ot-tegin's guarantee letter as Nov. 4 obv. 15-16 rtms 'cw 'm'y cttyH S'm'k rpr'yw1 L' nm't m't 'And if it is not agreeable for Chat to remain with me.', where the infinitive m't is regularly derived from the past stem. Benveniste, however, did not claim that these two paradigmatically related forms are also genetically related. Szemerenyi formulated the diachronic hypothesis much later: "This usage points the way to the correct understanding of wimat: -ma- is not the stem
*may- but the nil grade of man- 'wait, stay'. The proper verbal adjective would, of course, have been *mata-, since *man- is an anit-root. But this would have been identical with PPP of *man-'think'. No wonder that a way out was sought by following such models as zan-/ zata- 'born'" (Szemerenyi 1971: 517).
Szemerenyi's etymology, ingenious as it is at the first glance, remains an artificial construct. Given the widespread character of *ma- in East Iranian, one would expect that the paradigmatic association of this alleged secondary stem with the reflexes of IE *men- 'to remain' would also be common. Yet, Khotanese, at least, shows the clear formal and semantic contrast between hama- 'to become, be' (Emmerick 1968: 152) and man- 'to remain' (Emmerick 1968: 109), both of which are endowed with full paradigms, while the Wakhi and Yazghu-lami languages display full paradigms of the reflexes of *ma-. The unique merger of the two genetically unrelated verbs into a single paradigm in Sogdian apparently went hand in hand with the unique semantic change. Sogdian m't can mean 'to remain', as in the example cited above, while Sogdian m(')yn-can mean 'to be', as in SCE 255-7 xwnx ZKZY fiynptw mynt 'PZY 'ny' mrtxm'k w'n'kw ywct 'PZYn ptxwy'yn'k pwt 'PZYn 'wzy'n wn'nt 'one who is a priest so as to teach other people to
act as butchers and to slaughter animals.' (MacKenzie 1970). No parallel semantic shift can be observed elsewhere in Iranian. Neither can one claim that the proportional analogy based on *zan-/ zata- is limited to the Sogdian language. In Sogdian, this verb had been itself complexly remodeled in to yield the pair of zan-/ zand- 'to give birth, to be born' and azay-/ azit- 'to be born' and therefore could not trigger the remodeling of other verbs. To this one can add that the preterit form w-m't is considerably more frequent than the simplex m't, but the matching present stem **w-m(')yn- is not attested in Sogdian.
The only argument in favor of Sokolova's reconstruction *mai/mi is the presence of i/y in the present stems of Tumshuqese, Wakhi and Yazghulami forms. There are, however, no reasons to think that this element historically belongs to the root and not to the suffix. On the other hand, the past stem forms, Sogd. (w)m't and Wakhi umut, plead for reconstructing the Iranian past participle *mata-. It is probably true that the Iranian high vowels yielded Yazghulami a in some verbal past stems, perhaps due to the exceptional generalization of the historical feminine participle with a-umlaut, as in Yazgh. vad 'was' < Ir. *buta (f.) ^ge^bMaH 1987: 393).6 Accordingly, one could venture the reconstruction Yazgh. mad < Ir. *mita (f.). Nevertheless, since Yazgh. mad is the normal outcome of Ir. *mata-, one should give preference to this reconstruction, which does not assume an exceptional pattern of analogical leveling.7 The present stem of Ir. *ma 'to
6 Alternatively, the vocalization of Yazgh. vad instead of the expected *vod is due to the analogical influence of *mad ^ge^bMaH 1987: 3589). With regard to the etymology of mad, Edelman (ibid.) follows Szemerenyi's idea, deriving this stem from the Indo-Iranian *man 'to remain'. She does not, however, explain the formation of the corresponding present stem mi- in Yazghulami.
7 Even if one accepts *m(a)i- as the Pre-Yazghulami reconstruction of the verb 'to be', it would be possible to analyze this root as a result of language-internal restructuring of Ir. *ma. Compare the Yazghulami transitive stems yay- / yed 'futuere' < Ir. *ga 'to go in', pay- / payd 'to herd' < Ir. *pa 'to herd, protect', zdnay- / zdned 'to wash' < Ir. *sna 'to swim' etc. All these stems display the reanalysis of the suffixal -y- as part of the root.
become' is to be reconstructed as *ma-ya- (Wakhi umuy- < *a-ma-ya-). Its likely dialectal restructuring to *m(i)-ya- (Yazgh. mi-, Tumsh. ramy- < *fra-m(i)-ya-) could be based on the productive Indo-Iranian pattern of intransitive present stems in *-ya- appended to the zero grade of the root.
2. Anatolian. The Hittite verbal stem mai- / miya-hi 'to grow' (act.), 'to be born' (med.) is well attested both directly, through finite forms, and indirectly, through numerous nominal derivatives, e.g. LU mayas 'adult person (?)', miyatar 'growth, abundance', and amiyant- 'small' (Kloekhorst 2008: 172, 54041). The derived adjective L^mayant- 'grown-up, adult' is recorded in both Hittite and Palaic texts (Kloekhorst 2008: 542), while the historical participle miyanti- is attested in the Iron Age Luvian inscription of KARATEPE with the meaning 'long' (Tischler 1990: 91). In addition, Luvian features the abstract noun mashahit- 'growth, prosperity', attested in cuneiform transmission, and the denominative verb mashani-'to make grow', known from the Iron Age Luvian inscription SULTANHAN (Tischler 1990: 91, Kloekhorst 2008: 541). The last two forms are presumably derived from masha- 'growth', which in turn represents a verbal noun endowed with the productive suffix -sha- (Starke 1979). This noun either reflects a contraction of the earlier maya-sha- or is derived directly from
*ma, the Pre-Luvian shape of the root 'to grow'.
Although this cluster of Anatolian cognates is much better known to Indo-Europeanists than the East Iranian forms discussed in the previous section, it also lacks a commonly accepted Indo-European etymology. The range of opinions on the origins of Hitt. mai- / miya-hi is presented in Tischler 1990: 92-94. The most popular hypothesis, explicated in Oettinger 1979: 471, connects it with Hitt. mm- 'soft' and Lat. mltis 'soft, ripe'. This would imply that the palatal glide was the organic part of the root, which can be accordingly reconstructed as meiH (Rix 2001: 428). Kloekhorst (2008: 541) justly criticizes this etymology as semantically rather vague, upholding instead the connection between Hitt. mai- / miya-hi and Old Irish mar, mar, Middle Welsh mawr 'large', with the comparatives Gothic mais, Old High German mero 'more'. While Kloekhorst's alternative is superior from the semantic
viewpoint, his reconstruction *meh2 (=ma) 'big, much' runs into serious formal problems. The Celtic and Germanic forms cited in his dictionary were taken with good reason as belonging to the family of IE *me-ro- ~ *mo-ro- 'great' (Pokorny 1959: 704, cf. Matasovic 2009: 258).8 The semantically unconvincing
attempts to link the Hittite root with IE. *me 'to measure', *ma-'good, timely', or *mei 'to move' have enjoyed less following among scholars in the recent years.
I propose that Hitt. mai- / miya-hi 'to grow' is cognate with Ir. *ma 'to become'. The natural semantic change 'to grow'>'to become' can be illustrated through the English phrase to grow strong / tired / stale / irritated / anxious / more beautiful etc. Such forms as Gk. фиш0а1 'to grow' or Russ. быльё 'grass' demonstrate that the polysemy 'to grow ~ to become' can be reconstructed for the Indo-European verb *bhu (Rix 2001: 98101; *bhuehi). The formal side of the proposed comparison is likewise unproblematic once we accept the analysis of mai- / miya-hi as a reflex of the Early Indo-European (Indo-Anatolian) i-presents reconstructed in Jasanoff 2003: 91-127.9 This morphological category was productive in Anatolian, where it could extend itself even to the verbs that have no traces of it in "classical" Indo-European, notably Hitt. dai- / tiya-hi 'to put' ~ IE. *dhe 'id.' (cf. Jasanoff 2003: 98). But in the case of Hitt. mai- / miya-hi it remains likely that this stem actually continues an
8 It is true that the adjectives cited in Kloekhorst 2008: 541 are ambiguous with respect to their vocalism, but their reconstruction with the second laryngeal or *a would force one to ignore their obvious connection with the second component of Greek eyxeCTL-pwpog 'famous with spear' and related forms, and especially with the personal names in *-mero-, such as Old High German Walde-mar or Old Russian Bxadu-Mmpv 'famous through power'. At the same time, the traditional reconstruction *mero— *moro- 'great' is also somewhat problematic, because the ablaut in a thematic adjective is not expected.
9 While Jasanoff (2003: 94) hesitated to reconstruct the i-present for this particular verbal root, this was solely due to the lack of etymological comparanda at his disposal. The proposed etymology strengthens Jasanoff's overall argument for the reconstruction of i-presents in Anatolian. In his notation, Hitt. mai- / miya-hi can be reconstructed as *moh2-i- / *mh2-i-
Indo-Anatolian i-present. We have seen in Section 1 that all the East Iranian languages forming the present stems from ma 'to become' contain traces of the *-ya- suffix, which is the expected outcome of the Indo-Anatolian the present marker *-i- in Indo-Iranian.
The proposed binary comparison between Iranian and Anatolian is not conducive to determining the original vocalism of the reconstructed root. In Iranian, long non-high vowels naturally all fall together, while Hitt. mai features the root in the historical apophonic o-grade, where the difference between the morphophonemes {a}, {e}, and {5} is likewise neutralized.10 Consequently, it could be possible in principle to assume the reconstruction *me 'to grow', thus reconciling the new proposal with the alleged nominal cognates mai- / miya-hi offered in Kloekhorst 2008. I will argue, however, in the following section that there is a better match to the Iranian and Anatolian verbs, which supports the reconstruction of their root as *ma.
3. Latin and Slavic. The latest etymological dictionary of the Latin language connects Lat. maturus 'ripe, full-grown' with Lat. Matuta 'the goddess of Dawn', and matutinus 'of early morning' (de Vaan 2008: 367). The lexemes that are commonly adduced as further cognates of this group include Lat. mane 'early, in the morning', manus 'good', and Celt. *mati-'good' yielding Old Irish maith, Middle Welsh mad, and Middle Breton mat (Pokorny 1949: 693). Heiner Eichner rounded up this etymological nest by adducing Hitt. mehur 'time' and claimed that all the nominal forms cited above are derived from IE *meh2 'to be timely', the finite forms of which have all been lost (Eichner 1973, see especially p. 65).
This construct, in my opinion, is prone to deconstruction. The concept 'to be timely' is not frequently expressed through
10 For this phonetic environment, my observation can be also reformulated in laryngealist terms as the neutralization of opposition between *ohi, *oh2, and *oh3. According to "Pinault's Law", all the three laryngeals are expected to disappear before *y already in the proto-language.
a special finite verb, definitely not in the Indo-European language family. Furthermore, the comparison between Lat. manus 'good' and mane 'early, in the morning' does not require deep semantic reconstruction: the same meaning shift is still transparent in modern French de bonne heure 'early in the morning'. In a similar fashion one also can account for the semantic difference between Celt. *mati- and Lat. Matuta, matutinus. All these lexemes appear to contain an Italo-Celtic root with the basic meaning 'good' and of uncertain origin.
The etymology of Hitt. mehur 'time' represents a notorious point of contention among Indo-Europeanists and Hittitolo-gists, but the difference in root vocalism complicates its comparison with the other lexemes adduced in Eichner 1973.11
I believe that Hitt. mehur is not to be disconnected from Hitt. me(y)an- 'range (of the year), extent', therefore I accept the derivation of both nouns from IE *me 'to measure'.12 A different explanation of mehur is offered in Kloekhorst 2008: 568, but it is likewise incompatible with Eichner's proposal. Note also that mehur is not mentioned among the cognates of Celt. mati- 'good' in the latest Celtic etymological dictionary (Matasovic 2009: 259-60).
We are left with Lat. maturus 'ripe, full-grown'. It was sporadically compared in the past with Hitt. mai- / miya-hi 'to grow' (references in Tischler 1990: 93), but invariably in association with the Italo-Celtic words for 'good' and 'early', which did not add credence to these suggestions. Once we attach the Hittite verb to Lat. maturus alone, the semantic
11 Eichner himself justifies his comparison by reconstructing the acrostatic ablaut *meh2-ur / *meh2-un-os and assuming the special development eh2 > e, as opposed to eh2 > a. The validity of the last hypothesis, known as "Eichner's Law" represents another major point of disagreement among the Indo-Europeanists. But even if one accepts "Eichner's Law", it is clear that the reconstruction of a synchronically unattested morphological alternation weakens this particular etymology.
12 In my opinion, [h] in Hitt. mehur represents not a special reflex of the "first laryngeal", but rather a hiatus-breaker (similarly Cowgill 1979: 27-8 m.8 and Lehrman 1998: 214, fn. 13). This etymology requires further discussion.
problems are all eliminated. The secondary influence of the Latin words for 'early' as a folk etymology can, however, be seen in the second meaning of maturus 'early, precocious' (cf. the semantic discussion in Ernout and Meillet 1967: 391a) and perhaps in the basic meaning of the adverb mature 'at the proper time, early, quickly', if not derived through the metaphor "when time is ripe".
This does not, however, solve morphological problems. The Latin suffix -turus is the synchronic marker of future active participles. If Lat. maturus 'ripe, full-grown' had been formed from *ma 'to grow' according to the standard pattern, one would expect it to mean "one that has to grow", i.e. 'young'. One way to get around this difficulty would be to assume that maturus is older than the other Latin adjectives provided with the same suffix, and may even have served as their formal model. This was, in fact, the scenario entertained in Fortson 2007: 88, where maturus is analyzed as the *-ro-adjective derived from *matu, the instrumental case of the verbal noun *matu-. Under my reconstruction of *ma 'to grow',
*matu- could be translated as 'growth, stature', while maturus would mean "provided with stature", i.e. 'grown up'.
Yet comparative data plead for a different solution. The Slavic cluster of cognates that includes Russian матёрый, Ukranian материй 'full-grown', Old Church Slavic матерь-ство 'np£CTp£Lov', Church Slavic матеръ, маторъ 'old', Slovenian mater 'elderly', Bulgarian матор 'firm, full-grown, old' etc. can be reconstructed as *matorU. The suffixal e-vocalism in some Slavic forms may reflect a contamination with the word for 'mother', an obvious folk etymology for this group, which is nonetheless reflected in some serious etymological works. One cannot agree more with Dercksen (2008: 304), who indicates that Lat. maturus 'full-grown, ripe' is the best candidate for an external lexical cognate of Slavic adjectives, even though their suffixes do not exactly match. The suffix of Slav. *ma-torb, unlike that of Lat. ma-turus, is synchronically unproductive and therefore must be regarded as more ancient. Presumably, the u vocalism was introduced
into the Latin adjective under the double pressure of the common suffix -turus and words such as Matuta, matutinus that were falsely perceived as its cognates.13
Given that Slav. *ko-ton 'which of the two' corresponds to Gk. noT£po^ and Ved. kataras 'id.' (the last one without the effect of Brugmann's Law), the suffix of Slav. *ma-ton can likewise be identified with the IE *-tero- signifying a choice between two alternatives. Thus the original dichotomy reflected in Slavic *ma-ton and Italic *ma-tero- could be "more grown", as opposed to "less grown, young", but the polar semantic nuance might undergo bleaching already in Proto-Indo-European. It is not quite clear whether this polar adjective could be derived directly from the verbal stem or the derivation went through a verbal noun. Whatever the derivation, it should have been similar to that of Gk. ^£qt£qo^ 'better, best', which is related in some way to ^£Q£iv 'to carry' (for discussion, see Chantraine 1980: 1188-89).
The comparison of the Iranian and Anatolian verbal roots with *ma-tero- 'full-grown' precludes their comparison with
*mero-/moro- 'large, great' for phonetic reasons and vice versa. Two considerations support the first alternative. First, the semantics of *ma-tero- speaks for its deverbal character, while an adjective with the meaning 'large' is more likely to be primary. Second, the reflexes of *mero-/moro- 'large, great' are associated with the contiguous area stretching from the northwest to the southeast of Europe and may well represent a regional term spreading by way of language contact. By contrast, a lexeme confined to Slavic and Italic is expected to reflect an Indo-European linguistic archaism. On the basis of these two arguments, I prefer to keep *mero-/moro- out of the picture and reconstruct the verbal root treated in the present paper as *ma.
13 This explanation need not undermine Fortson's hypothesis about the historical derivation of participles in *-turos from instrumental nouns. In fact, Fortson (2007: 89) discusses in some detail the structure of futurus that was more likely to serve as a formal model for the other future participles. Assuming that the meaning of Pre-Latin *fu- < IE *bhu was 'to become' rather than 'to be', futu-rus can be analyzed as "imbued with becoming, imminent", hence 'future'.
4. Concluding remarks. The reconstructed morpheme *ma was not the only exponent of the meaning 'to grow' in Indo-European. The most well-represented verbal root belonging to this semantic field was *xaug / *xweks 'to grow, increase' (Rix 2001: 274, 288-89; *h2eug / *h2ueks). Among its derivatives are Tocharian A/B ok-/auk-, Lithuanian augti 'to grow', Latin augere 'to increase', Avestan waxs-, Attic Greek au££o0ai, and German wachsen 'to grow'. Its reflexes, however, do not seem to be attested in Anatolian, and definitely not as finite verbal forms. Another common Indo-European verbal root that had 'to grow' among its basic meanings was *bhu (Rix 2001: 98-101; *bhueh2). This root is preserved in the Luvian adverb puwa 'formerly', which is the lexicalized form of bhu in its derived meaning 'to become, be' (Ivanov 2001). There is also about a dozen of other verbal roots having the meaning 'to grow' in individual Indo-European families, although their semantics need not be projected into Proto-Indo-European.
This embarrassment of the choice helps to explain why the reflexes of *ma 'to grow' are limited to four language families. Already in Indo-Anatolian this ancient root must have been in competition with *bhu, which it succeeded in marginalizing in Anatolian. In Late Indo-European, however, it was itself marginalized through the rise of *xaug / *xweks and language-specific verbs meaning 'to grow'. In Indo-Iranian it lingered on in the derived meaning 'to become' (again in competition with *bhu), while in the rest of the family only its nominal reflex
*ma-tero- 'full-grown' was sporadically preserved.
Those who are willing to venture beyond Indo-European and inquire into the possible Nostratic connections of *ma 'to grow' are reminded of Dravidian *ma(j)- 'large, great', which is adduced as an external cognate of Hitt. mai- / miya-hi in Kassian 2009: 164.
Литература
Соколова В. С. 1973. Генетические отношения язгулямского языка и шугнанской языковой группы. Ленинград: Наука. Стеблин-Каменский И. М. 1999. Этимологический словарь ваханского языка. СПб.: Петербургское Востоковедение.
Эдельман Д. И. "Язгулямский язык". 1987. Основы иранского языкознания. Новоиранские языки: восточная группа. Под ред. В. С. Расторгуевой. Москва: Наука. Стр. 348-407.
Bailey, Harold W. 1979. Dictionary of Khotan Saka. Cambridge University Press.
Bartholomae, Christian. 1904. Altiranisches Worterbuch. Strasburg. Reprint 1961. Berlin: Walter de Gruyter.
Benveniste, Emile. 1929. Essai de grammaire sogdienne. Deuxieme partie: morphologie, syntaxe et glossaire. Paris, P. Geuthner.
Chantraine, Pierre. (1968-)1980. Dictionnaire etymologique de la langue grec. Paris: Klincksieck.
Cheung, Johnny. 2007. Etymological Dictionary of the Iranian Verb. Leiden: Brill.
Cowgill, Warren. 1979. "Anatolian hi-conjugation and Indo-European Perfect: Installment II". Hethitisch und Indogerma-nisch. Ed. E. Neu and W. Meid. Innsbruck: Institut fur Sprach-wissenschaft der Universitat Innsbruck.
Dercksen, Rick. 2008. Etymological Dictionary of the Slavic Inherited Lexicon. Leiden: Brill.
Eichner, Heiner. 1973. "Die Etymologie von heth. mehur". Munchener Studien zur Sprach-wissenschaft 31: 53-107.
Emmerick, Ronald E. 1968. Saka Grammatical Studies. Oxford University Press.
Ernout, Alfred and Antoine Meillet. 1967. Dictionnaire etymologique de la langue latine. Fouth edition. Paris: Klincksieck.
Fortson, Benjamin W. IV. 2007. "The Origin of the Latin Future Active Participle". Verba Docenti (Fs. Jasanoff). Ann-Arbor: Beech Stave. Pp. 83-95.
Gharib, Badr-uz-zaman. 1995. Sogdian dictionary. Tehran: Farhangan.
Hitch, Douglas. 1992. "Tumshuqese resth- (resth-) : Khotanese *hist-'send'". Journal of Turkish Studies 16 [= Richard Nelson Frye Festschrift I. Essays presented to Richard Nelson Frye on his Seventieth Birthday by his Colleagues and Students]. Pp. 91-95.
Ivanov, Vyacheslav V. 2001. "Indo-European *bhuH in Luvian and the Prehistory of Past and Perfect". Proceedings of the Twelth Annual UCLA Indo European Conference. Ed. M. Huld et al. Washington: Institute for the Study of Man. P. 81-106.
Jasanoff, Jay. 2003. Hittite and the Indo-European verb. Oxford University Press.
Kassian, Alexei S. 2009. "Anatolian Lexical Isolates and their External Nostratic Cognates". Аспекты Компаративистики 4. Orientalia et Classica XXVIII. Moscow: Russian State University for the Humanities. P. 152-86.
Lehrman, Alexander A. 1998. Indo-Hittite Redux. Moscow: Paleograph.
Kloekhorst, Alvin. 2008. Etymological Dictionary of the Hittite Inherited Lexicon. Leiden: Brill.
MacKenzie, David. N. 1970 The "Sutra of the causes and effects of actions" in Sogdian. Oxford University Press.
Matasovic, Ranko. 2009. Etymological Dictionary of Proto-Celtic. Leiden: Brill.
Oettinger, Norbert. 1979. Die Stammbildung des Hethitischen Verbums. Nurnberg: Karl.
Pokorny, Julius. 1959. Indogermanisches etymologisches Worterbuch. Bern: Francke.
Reichelt, Hans. 1931. Die sogdischen Handschriften-reste des Britischen Museums II. Heidelberg: Winter.
Rix, Helmuth. 2001. Lexicon der indogermanischen Verben: Die Wurzeln und ihre Primarstammbildungen (2nd ed.). In cooperation with M. Kummel, Th. Zehnder, R. Ipp, and B. Schirmer. Wiesbaden: Reichert.
Starke, Frank. 1979. "Zu den hethitischen und luwischen Verbalabstrakta auf -sha-". Zeitschrift fur vergleichende Sprachforschung 93: 247-61.
Szemerenyi, Oswald. 1971. "Iranica IV". Orbis 19: 500-519.
Tischler, Johann. 1990. Hethitisches Etymologisches Glossar. Fascicle 5-6 (L-N). With contributions by G. Neumann and E. Neu. Innsbruck: Institut fur Sprachwissenschaft der Universitat Innsbruck.
de Vaan, Michiel. 2008. Etymological Dictionary of Latin. Leiden: Brill.
Yakubovich, Ilya S. 2006. "Marriage Sogdian Style". Iranistik in Europa - gestern, heute, morgen. Akten der Tagung Graz, 11.15.02.2002. Ed. B. Fragner et al. Vienna: Osterreichische Akademie der Wissenschaften. P. 307-344.
Summary
The purpose of this paper is to argue for the existence of a new Proto-Indo-European verbal root by comparing three groups of forms that were commonly regarded as etymologically unrelated. The reflexes of Indo-European *ma 'to grow' are preserved as finite verbal forms in Iranian (*ma 'to become, be') and Anatolian (Hitt. mai- / miya-hi 'to grow'), and as adjectival derivatives in Slavic and Italic (Russian матёрый and Latin matrnus 'full-grown').
It is concluded that the reconstructed root belonged to the most archaic stratum of the Indo-European lexicon. Already in Indo-Hittite (Early Indo-European) this ancient root must have been in competition with *ЪНй 'to grow, become', which it succeeded in marginalizing in Anatolian. In Late Indo-European, however, it was itself marginalized through the rise of *(x)aug / *(x)weks 'to grow, increase' and language-specific verbs meaning 'to grow'. In Indo-Iranian it lingered on in the derived meaning 'to become' (again in competition with *ЪНй), while in the rest of the family only its
nominal reflex *ma-tero- 'full-grown' was sporadically preserved. The last form represents an etymological polar adjective stressing the opposition with young animals/plants.
A number of ancient contaminations and folk etymologies complicate the analysis of the reflexes of IE *ma 'to grow'. In Sogdian, for example, it became united in a single paradigm with the reflexes of IE *men 'to remain', which contributed to the change in its semantics. It Latin, *ma-tero- became synchronically associated with mane 'early in the morning' etc. and acquired the second meaning 'early, precocious', while the change in suffix vocalism is due to the influence of future active participles in -turus. In Slavic, the inherited Mamop'b underwent a dialectal change into Mamepv by association with the word for 'mother'.
