Lat. "atrium", Gr. "ándro-" and terms from I.-E. *h(n)d(n)-(ero-)

133 LAT. A¯TRIUM, GR. ANDROAND TERMS FROM I.-E. *HNDH-(ERO-) LAT. A¯TRIUM, GR. ANDROAND TERMS FROM I.-E. *HNDH-(ERO-) Adolfo Zavaroni Lat. aˉtrium y aˉter parecen derivar de una base que expresa ‘oscuridad’ como un cambio semántico de ‘caverna, parte inferior e interior’, mientras Etr. *atrdenota ‘enterrar’<‘bajar’. Se puede plantear la hipótesis de que estos términos derivan de una raíz *HnD- ‘debajo, en, abajo’, donde D representa una oclusiva dental variable, cuya variación se debería a la presencia de la nasalización, como se ve por otras oclusivas iniciales nasalizadas *HmB-, *HnG-. Esta formulación incluye la forma tradicional *°ndh- ‘abajo, bajo’ y explica algunas variaciones tales como Lat. atry Umbr. *adr-o- ‘oscuro, negro’ or Gk. ajnqro<Éajndro- ‘hombre’, si se asume que en griego también -como en Lat. homo etc.- el significado primario de la palabra ‘hombre’ era ‘el que vive abajo, sobre la tierra, la parte más baja del cosmos’. De ahí, la raíz IE *HnD- ‘bajo, debajo, en’ experimentó un desarrollo semántico y produjo términos que significaban 1. ‘tierra, suelo’, 2. ‘hueco, cueva, caverna, hoyo’, 3. ‘hombre’, 4. ‘carbón’, 5. ‘oscuridad’, además de adverbios y verbos que expresan una dirección hacia abajo y hacia el interior. Lat. aˉtrium and aˉter seem to derive from a basis expressing ‘darkness’ as a semantic shift from ‘cavern, lower and inner part’, while Etr. *atrdenotes ‘to bury’<‘to let down, to lower’. One may hypothesize that these terms derive from a root *HnD- ‘under, in, down’, where D represents a variable dental stop, whose variation would be due to the presence of the nasalization, as we may see for the other initial nasalized stops *HmB-, *HnG-. This formulation includes the traditional form *°ndh- ‘down, low’ and accounts for some variations such as Lat. atrand Umbr. *adr-o- ‘dark, black’ or Gk. ajnqro<Éajndro- ‘man’, if we assume that in Greek too – as well as in Lat. homo etc. – the primary meaning of the word for ‘man’ was ‘he who lives down, on the earth, the lower part of the kosmos’. Hence, HABIS 38 (2007) 133-146 134 ADOLFO ZAVARONI the IE. root *HnD- ‘down, under, in’ underwent a semantic development and yielded terms meaning 1. ‘earth, ground’, 2. ‘hollow, cave, cavern, pit’, 3. ‘man’, 4. ‘coal’, 5. ‘darkness, obscurity’, beside adverbs and verbs expressing a downward and inward direction. 1. Latin glossators1 hand on that Lat. aˉtrium was a word of Etruscan origin. Ernout & Meillet consider this tradition as probable and add that ‘si l’aˉtrium n’est pas étrusque, ce serait un souvenir de l’ancienne maison où la fumée du foyer s’échappait par une ouverture ménagée dans le toit ; il y aurait ici soit un dérivé d’un ancien nom du ‘feu’ soit un dérivé de aˉter’2. As Untermann has already noted3, the thesis4 that the original sense of aˉter is ‘blackened by fire’ in virtue of a link with words meaning ‘fire’ (Avest. aˉtarš, genit. a q roˉ ‘fire’, OIr. áith ‘stove’, Slav. words as Serb. vatra ‘fire’) is untenable. On the other hand, it does not explain the variation d/t we find by comparing Umbr. adroand Lat. atro- (cfr. Osc. Aadiriís, Aadíriis also generally considered equivalent to Lat. Atrius)5. According to our suggestion, Lat. aˉtrium does not derive from a vague name of the fire, but from an Indo-European root, perhaps by means of an Etruscan intermediaton in pronounciation (/d/>/t/). In Lat. coloroˉ, occultus (from *kelof cella: cfr. also kaluvbh and kaluvptw) we have the semantic shift ‘covered, hidden’>‘dark’. In Lat. caecus the original sense ‘dug, hollow’ changes not only into ‘blind’, but also into ‘covered, hidden’ and ‘dark, tenebrous, without light’6. An analogous semantic shift is present in OInd. andha 1. adj. ‘blind, secret, hidden’; 2. noun ‘obscurity’, andhu ‘well, pit’: the direction of the shift is, as we will see, from ‘well, pit’ to ‘obscurity’. We attribute the root of OInd. andha, andhu to Lat. aˉter, aˉtr- ‘obscure, black’, although such a connection at first sight does not seem to be possible. In order to account for this connection, we may start from the Etruscan words atrane and atrs´. Etruscan is not an Indo-European language, though, in our opinion was a lingua franca which borrowed many Indo-European roots before the adoption of writing, that is before the seventh century B.C. We suggest that Etruscan can supply indications for defining the Indo-European root in question, although interpretation of Etruscan words commonly raises a justified scepticism. 1 Varro, L.L., 5.161; Fest. p. 12; Serv. Aen. 1.726. 2 Ernout-Meillet, s. v. aˉtrium. 3 Untermann (2000, 55). 4 This thesis is considered probable by Schrijver (1991, 54). 5 See bibliography in Untermann (2000, 54-55). 6 A. Zavaroni, “Imprestiti indoeuropei con Lautverschiebung in etrusco”, IF 111 (2006) 149-181. 135 LAT. A¯TRIUM, GR. ANDROAND TERMS FROM I.-E. *HNDH-(ERO-) In the bilingual inscription of the golden A-plate of Pyrgi the sequence nac atranes zilacal seleitala ‘after the burying of the bright Rector (= Jupiter)’ corresponds to the Phoenician sequence BYM QBR ’LM ‘in the day the divinity is buried’7. It is not clear if atranes is a nominative in -es (cfr. the nominatives fleres ‘bronze statue’<‘flatilis’, Artumes, Par q anapaes) or the genitive of atrane which appears as a cognomen (fem. atrania and atrani<*atranei) in some onomastic formulae and pottery seals. On pottery seals atrane (nom.) or atranes (gen.) or atranesi (abl.) are trademarks. In the Pyrgi plate atranes, which follows nac, could also be a direct case (acc.) instead of an oblique one (gen.), since in another passage of the plate nac has the function of a preposition governing the accusative ci avil ‘three years’. Formally, atrane appears to be a present participle like uslane, u esane, acilune, ezine, nun q ene etc.8: therefore atranes, as a nominative, could correspond to a Latin noun ending with -ntia. As we may infer from comparison with the Phoenician text, atranes means roughly ‘burial’, in virtue of the semantic passage we will try to define in the next paragraphs. At the end of an inscription9 on a sarcophagus from Tarquinia, one reads suq i q atr´srce: as the locative su q i q certainly means ‘in the sepulchre’, while atr´srce is a preterite in -ce, we may interpret atr´srce as ‘(he) was interred, laid’10. In some burial inscriptions of Vulci one finds the formula hels atr´s. In our opinion, hels is a verb meaning ‘cubat, reclinat’ (from I.-E. *(s)kel- ‘curve’ or *k’el- ‘lean down’), while the word helsc<*helsce of other inscriptions may be translated as ‘cubuit, declined>laid down’. The form atr´s –where ´s commonly is the development of /s/ + dental11– may reflect an adverb (a locative ?) *at-(V)r-sq <*at-(V)re-sq i ‘below’. We cannot establish whether the preterite atr´srce<*atrsq -Vr-ce originates from an agglutinated adverbial form or if it is a compound whose first member should be atr´s (for example: atr´s + er-ce). In the ‘Tomba François’ of Vulci we find two equivalent formulae: hels atr´s ‘cubat infra (vel : intra ?)’ and sacni´sa atur´s ‘quiescens (or: demissus, -a ?)12 in7 P. Agostini–A. Zavaroni, “The bilingual Phoenician-Etruscan Text of the golden Plates of Pyrgi”, Filologija 34 (Zagreb 2000) 3-48, at 35. 8 A. Zavaroni, I documenti etruschi (Padova 1996) 45. 9 M. Pallottino, Testimonia Linguae Etruscae (Firenze 19682), n° 135. 10 In a late age (2nd-1st century B.C.) the preterite in -ce could also act as a past passive participle. 11 A. Zavaroni, “Sulla presunta sibilante palatale in etrusco”, Incontri linguistici 25 (2002) 87-102. 12 In our opinion, in funeral inscriptions sacniu is an appellative of ´su q i ‘tomb’ and means ‘depositio’, while sacnisa is an adjective referring to the dead. It is not ascertainable if sacniu and sacni´sa derive from the root we attribute to sacnic- = sacnit-, that is I.-E. *saˉg- ‘suchen, acute sentire’ (which yields also Goth. in-sakan ‘lay down, deponere’ and ON. sakna ‘vermißen’) or from *seˉk- ‘rest, quies’ (cfr. Lat. seˉg-n-is and gr. h\κα, dor. ajκa'(strum.). 136 ADOLFO ZAVARONI fra’, where atur´s<*atVr-(e)sq (i) would be the locative of a form developed from *HnD-(er)- ‘below, inferior’. Hence, if Etr. *atrand Lat. aˉtroshared the same root, as Etr. *atrseems to mean ‘lower down, put under>inter, bury’, we have to verify whether the meaning ‘dark’ of Lat. aˉtr-, Umbr. adro is a metaphor from ‘infernus’. Ernout & Meillet observe that aˉtro- ‘implique souvent une idée morale de terreur, de malheur, de mort’13. If so, our thesis that Lat. aˉtris connected with Etr. atranes ‘burial<deposition’ and atr- ‘depose, lower down>bury’ appears more probable. The negative value, as we will see, would be due to the fact that originally *aˉtris equivalent to infer. Although its a is brief, Lat. atroˉ x is commonly considered a derivate (or a compound)14 of aˉtr-. Anyway a˘troˉ x has a sense of terror and fuˉnus even more accentuated than aˉtro-. According to Schrijver, a˘troˉ x ‘contains the zero grade of the root *HeHtfound in aˉter “black”, Av. aˉtarš “fire”’15. But, as we have already observed, such a comparison is unreliable and does not permit a ‘laryngalistic’ approach. That Etr. *atrdenotes ‘under, low, down, infraˉ ’ (with a secondary sense of darkness, blackness) seems to be confirmed by the fact that in other funeral inscriptions one finds *caluinstead of *atr-. In our opinion calus means ‘chthonian’<‘low’ and derives from *(s)kel- ‘biegen; Niederung’. It is well-known that Tinia Calusna is ‘Iovis Infernus’, since this name is written on a vase found in a subterranean cell of the Tempio del Belvedere (Bolsena), dedicated to chthonian divinities. Other derivates from *caluare calusur (plur.) ‘exequiae<depositiones’, calu ‘deposed’, calusin ‘depose>bury’. I would not exclude that Lat. calaˉre ‘let down, lower’ derived from Etruscan rather than Greek16. 2. Lat. ıˉnferus, ıˉnfraˉ ıˉnfimus point to the group of I.-E. roots commonly formulated as *°ndhos, °ndhero ‘infraˉ’, °ndhero- ‘inferior’, °ndhemo- ‘infimus’. The long initial iseems a Latin phenomenon due to the cluster ‘nasal + spirant’17. As the formula *°ndhseems inadequate to us, we prefer to express it as *HnD- (or 13 Ernout - Meillet, s. v. aˉter. 14 Discussion of this dilemma is not important here: anyway, see Ernout and Meillet, s. v. atroˉx. 15 Schrijver (1991, 54). 16 According to Ernout - Meillet, 88, Lat. caloˉ is a borrwing from Gr. calavw. 17 Meiser (1998, 78) cites coˉnsul, ıˉnferıˉ [EIMFERIS in CIL XV 6265, where EI is for ı¯], saˉnctus, coniuˉnx, quıˉntus, etc. 137 LAT. A¯TRIUM, GR. ANDROAND TERMS FROM I.-E. *HNDH-(ERO-) *H°nD-) on the analogy of the writing *HmB- (or *H°mB-) we have adopted for three roots denoting the following semantic grids18: 1) *HmB- (generally>*h2mbh-/*h2mp-) ‘join; together; pair’: it gives rise to Gr. a[mfw, ajmfiv, Lat. amb-, amboˉ, OIr. imb-, Etr. ame, OInd. am-aˉ, and also to Lat. amor, Gr. ojpuivw (<*h3mp-) ‘I marry’ and binevw ‘I have intercourse’, Etr. puia ‘uxor’, Lat. par, apiscor, aptus, Germ. *ibna, *imna (Goth. ibns ‘even’ etc.) and *ba- (Goth. bai and bi, etc.). 2) *HmB- (generally > *h3mb(h)- / *h3mp-) ‘swelling, growing round’ : it gives rise to many terms, among which Lat. umboˉ, umbilıˉcus, Gr. ojmfalov", Germ. *nobhelo-, OIr. imbliu ‘navel’. For instance, the apple-names derive from the l-enlargement *h2mb-el- / *h2mb-ol- / *h2mb-˚l- : Gr. mh'lon, Lat. maˉlum, Alb. mollë and Germ. Celt. Balt. Sl. *ablu-, *abVl-. 3) *HmB- ‘buzz; bee; mumble’, etc.: the laryngeal was vocalized in Gr. ejmpiv" ‘mosquito, horsefly larva?’, OHG. imbi (‘ältester Beleg impi pıˉano’), MHG. imbe ‘Bienenschwarm, Bienen-Stock’ (‘erst spät-mhd. “Biene”’19), OE. imbe ‘swarm of bees’ and also in Lat. apis ‘bee’, where the nasal sound has disappeared. If we adopt the formulation *HnDinstead of *°ndh-, a series of unexpected connections appears. But first it should be considered that: 1) the formulation *HnDtends to show the existence of the prenasalized obstruents at an old stage of the Indo-European which, as Martinet had noted, allows us to explain ‘l’alternance de *-met de *-bhaux cas obliques du pluriel’20. As an example of a prenasalized stop in an initial position Martinet mentioned only the case of *H2m≥bh-i (> *amphí), but analysis shows that his intuition is productive, if applied to the roots mentioned above. 2) Writing *HnD-, we have to assume that H can develop into any laryngeal and hence into any vowel, the zero grade (suppression of the vowel itself) included. D points out that we have to expect any dental obstruent, whatever the development of the root. Effectively, we can observe a variation of the obstruent after a nasal in several roots ending with -NC- (C = any stop). Such a variation was even more possible when the root began with a laryngeal or with an aspiration requiring a certain initial phonic energy, causing the instability of the stop after the nasal. 18 A. Zavaroni, ‘I-E. “apple”, Hamito-Semitic “genitals” and Roots beginning with *HmB-’, HSF (forthcoming). 19 IEW 311. 20 We adopt the concept of ‘phonème prénasalisé’ that A. Martinet, Des steppes aux océans. L’indoeuropéen et les Indo-européens (Paris 1986) 176, applies to Lat. amb-, Gr. ajmf-, Gaul. amb-. See also A. Martinet, ‘Des prénasalisées en indoeuropéen?’ in Studies in Greek Linguistics, a Festschrift for John Chadwick, 27-29 april 1987, Aristotelian University of Thessaloniki (Thessaloniki 1987) 31-32; A. Martinet, ‘Finales nasales mobiles et prénasalisées indo-européennes’, BSL 86 (1991) 361-365. 138 ADOLFO ZAVARONI Contrarily to the common development of single obstruents, the results of the group ‘laryngeal + nasal + labial (or dental or velar) stop’ do not appear constant in every language: therefore the general formulas *HmBand *HnD- (or *H°mBand *H°nD-) are an easy way of denoting variations. The group *HmBproduces various developments that are commonly exchanged for beginnings of different roots: *VmB-, *VB-, *B-, *mand, by reduplication, *B+V(m)B- (V = any vowel). Analogously, from *HnD- (or *H°nD-) we should expect *VnD-, *VD-, *D-, *Nand *D+V(n)D-. We can then explain in a simple way why the dental of Lat. aˉtrois different from that of Umbr. adro- (and probably Osc. Aadiriís = ˉ Atrius), without resorting to the thesis that not only aˉtrium, but also aˉtroderive from Etruscan. Given the Etruscan Lautverschiebung (I.-E. [d] > Etr. [t]), Etr. *atrcould better correspond to Umbr. adro-. Anyway all these forms may be traced back to *eh2D-ero-, a denasalized form with compensatory lengthening of *HnD-ero-, of which *endhero- > *ıˉnferois a parallel realization. Lat. humus ‘earth, soil’ and homoˉ ‘man < terrestrial’ are supposed to derive, as Meiser notes21, from “uridg. dhegˆ hoˉm, Gen. Sg. dhegˆ hmes ‘Erde’”, since “die ursprüngliche Folge ‘Dental + Palatal’ ist noch erkennbar in heth. >te-e-kan<, tochar. A tkam≥ (B kem≥) und wohl in den italischen Etyma lat. homoˉ, osk. HUMUNS ‘homines’, umbr. HONDRA ‘unterhalb’ < *ghom-teraˉ < *dhegˆ hoˉmvorausgesetzt, ebenso in got. guma ‘Mann’”. Therefore, there should be a metathesis in Gr. cqwvn, OInd. ksám- ‘earth’, Gaul. (TEVO)-XTONION [dewo-gdonion] (gen. pl.), while in OIr. dú (acc. don) ‘earth’ and duine ‘man < terrestrial’, gˆ h would have fallen after metathesis. A new and important italic attestation is given by the South-Pic. verb hemat. In a recent re-examination of the Belmonte stele, we have also deciphered the sequence łas manes abít sue cum hemat ‘Lar Manis abit, secum humat’. The object of hemat ‘humat’ is the dead person mentioned at the beginning of the inscription. South-Pic. hemat, together with nemoˉ < *ne hemoˉ, proves that the glosse of Paul. Fest 89 hemonem hominem dicebant is worthy of faith. Here we would underline the semantic convergence (or equivalence?) between ‘earth’ and ‘low, inferior, under’, already remarked by Ernout and Meillet: at first they quote a passage of Varro (L.L. 5.23) containing also the sentence et dicitur humilior qui ad terram demissior, infimus humillimus, quod in mundo infima humus ; then they note that “en tenant compte de [lat. humilis et gr. cqamalov" et camhlov"], et de lit. žëmas ‘bas’, zemy˜n ‘en bas’, lette zem ‘sous’, v.pruss. semmai ‘en bas’, on admettrait un radical *hom- ‘terre’ pour expliquer l’adverbe osco-ombrien attesté par ombr. hondra, hutra (et [o.] huntrus?) ‘infra’, hondomu 21 Meiser (1998, 97); see also Gamkrelidze-Ivanov (1995, 720). 139 LAT. A¯TRIUM, GR. ANDROAND TERMS FROM I.-E. *HNDH-(ERO-) ‘infimoˉ’, osc. hu(n)truis ‘inferis’; il s’agirait, comme dans infraˉ, d’un mot artificiel, créé pour des raisons religieuses...”22. The religious reasons could, at most, have strengthened a sense, but they could not yield it. The concepts of ‘earth, ground’ and ‘low, down, under (-ground)’ were already intimately bound ab antiquo, in the sense that the ‘earth’ is ‘what is down, under’. For instance, the Hitt. adverb dagan = tagan ‘under, in the ground’ is an “Endungloser Lokativ” of tekan ‘earth, soil’. Since we do not see how the meaning ‘down, below’ could develop from a term containing *kom ‘cum, with’, we wonder whether Gr. katav, kavta and Hitt. katta, whose primary sense is ‘down, under’ (Hitt. kattera- ‘inferior’) really derive from *k°mta < *kom-ta – as one commonly reads23 – or are the result of an epenthesis in *kta- < *ghdha- (by relaxation of the aspiration in Greek). In the latter case, as Hitt. katta (and derivates) and dagan, both meaning ‘under’, would have the same root, we could not say if *ghdhaderives by metathesis from *dhgˆ h-a or vice versa. 3. Pokorny lists OInd. ní, Avest. nıˉ ‘nieder(wärts)’, Arm. n(i)- ‘nieder’, OInd. nitaraˉ ´ m ‘unterwärts’, Avest. nit � ma- ‘der unterste’, and the derivatives from Germ. *niþra- ‘nieder’ and *niþanoˉ ‘unten’ under the root *ni-, nei- ‘nieder’ (Komparativ *nitero- ‘nieder’)24. The w-extension *nei-w-, ni-wis commonly attributed to Gr. nεíaτος, nε vaτος, ‘der unterste’, nε vιoˉqεn ‘von unten’, OE. neowol, nihol ‘pronus’, MLG. nigel ‘niedrig’ etc. According to our thesis, *ni-, neiis one of the developments of *HnD-i-, *HnD-ei-, while *nitero- (< *n’ni-tero- < HnD-itero-) is the comparative in -tero, a parallel form of *Hndh-erowhich yields Lat. ıˉnferus, OInd. ádhara- ‘inferior, lower’, Avest. adara- ‘inferior, more westwards’, Goth. undar, undaro, OHG. untar, untari etc. According to our suggestion, also the terms listed by Pokorny under the root *ner- ‘unten’ > *ner-tero- ‘unterer’ derive from *HnD-er-: while Pokorny suggests a ‘mögliche Verwandtschaft von n-ermit *n-ei-’25, we think that *ner- <*HnD-eris a crystallized comparative redetermined by the addition of -tero-. In ejnevrtero" ‘inferior, underground’ (parallel form of nevrtero") and e[neroi ‘those who are underground; the dead; infernal deities’ the laryngeal h1 has changed into a stable e- (e[nero- < *h1nnero- < *HnD-ero-). The stem *ner-terois also attributed to Umbr. nertru ‘left’, OIcel. norðr, OE. norð etc. ‘North; northward’, OE. norðerra ‘more northern’. The reason for 22 Ernout - Meillet, s.v. humus. 23 Cf. Frisk, Griechisches etymologisches Wörterbuch (Heidelberg 19732), s. v. kavta; IEW 613; Beekes (2003), s.v. kavta. 24 IEW 312. 25 IEW 765-766. 140 ADOLFO ZAVARONI the sense ‘North’ is merely astronomic: after setting in the west, the sun is certainly ‘underground’ (for the old geographers) when, in its apparent inclined orbit, it proceeds towards the north. In Umbr. nertru and Osc. nertrak the meaning ‘left’ is a secondary sense of ‘low, inferior’: perhaps it is due to a reference to an augural orientation. 4. Lat. homoˉ derives from a root that, beside denoting ‘earth’, expresses also ‘to be low, to be under’. In order to suppose that Gr. ajndrotoo denotes ‘he who is down, low’, it is sufficient to connect it to *HnD-ero-. While there are not phonological problems in connecting a[nqr-(wpo") with *Hndh-eroof Lat. ıˉnfra, ıˉnferus, OInd. ádhara-, Avest. a d ara-, Goth. undar, undaro, OHG. untar, untari etc., for ajndrowe must assume that the variation d/q is due to the presence of the nasal. In dr-wvy = a[nqr-wpo" both the laryngeal and the nasal are dropped. It is uncertain if ajqerivzw ‘spernoˉ, uile facioˉ’ derives from *Hndh-ero-26. If it were, we should have a development *HnD-er >a[qerby denasalization; but it is no less possible, in our opinion, that ajqerivzw derived from a[qer- ‘arista, awn, straw, chaff’ and its original meaning was ‘to consider somebody as much as an arista, a straw, a chaff’. After a questionable distinction between ‘man 1’ = ‘human being’ and ‘man 2’ = ‘vs. woman’ and after recalling that the etymology of a[nqrwpo" is very disputed, Buck adds: “But the old analysis a[nqr-wpo" ‘man-faced, man-like’, from the stem of ajnhvr, ajndrov" remains the most probable; cf. Hesych. drwvy: a[nqrwpo" [...]. The change of ajndrto ajnqris due to [the spiritus asper] in the second element, which does not belong properly to the root ojpbut may be due to the influence of oJravw’”27. This thesis28 is also mentioned by Pokorny who lists ajnhvr, ajndrov" under the root *ner-(t-), aner- (ener-?) 1. ‘(magische) Lebenskraft’, 2. ‘Mann’; *neryo- ‘männlich, kräftig’.29 According to Beekes, “as no IE explanation has been found, the word [a[nqrwpo"] will be a substratum word”30, given that Myc. -oqdoes not prove IE origin, as the substr. language had labio-velars (basileuv"). It seems to 26 Beekes (2003), s.v. ajqerivzw. 27 C. D. Buck, A Dictionary of selected Synonyms in the principal Indo-European Languages (Chicago-London 1949) 80. 28 Buck and Pokorny attribute it to Kretschmer. 29 IEW 765. 30 After recalling the approachment between a[nqrwpo" and Hitt. antuhša- (antuwahha- , antuhha-) ‘man’ suggested by Kretschmer, Beekes (2003, s.v. a[nqrwpo") adds: “Improbable [according to] Ruijgh, Lingua 25 (1970) 312; Szemerényi, Gnomon 43 (1971) 655f. Kuiper gave a substr. interpretation on the basis of drwvy, FS Kretschmer, 1, 211f; Lingua 21 (1968) 275f.; defended by Beekes, Glotta 73 (1995/6) 13-15”. 141 LAT. A¯TRIUM, GR. ANDROAND TERMS FROM I.-E. *HNDH-(ERO-) us that Beekes assigns to a (nebulous) substratum all the words he is not able to bring into the ambit of the Indo-European phonology encoded according to the rules formulated by glottologists. As we think that the frame of laws formulated by linguists till now could not always be exact, often we prefer the old formula ‘unknown etymology’ rather than the label ‘substratum’. Beekes rightly considers ‘vague and differring’ terms as ‘Mediterranean’ and ‘foreign’,31 but the ‘substratum’ he often mentions is no less vague. According to Bader, a[nqrwpo" contains the name of the ‘man’; besides, “le -thd’épenthèse est dû à l’action de *h2 : le traitement de *h2°ny est comparable à celui de *sn- (à ceci près que *°n- > an au contact de *h2): assourdissement de la nasale au contact de h2 sourd, et report du trait aspiré de cette dernière, d’où le -thd’épenthèse; celle-ci est d dans la flexion, type andr-ós, de *h2°n-r-, par nivellement sur les caforts [?], ou la prothèse de *h2nerentraîne une syllabation a|ner, laissant intacte la nasale, sonore et non aspirée”32. It seems to us that Bader tries to formulate the difference between anqr-, andrand ajnhvr in more precise phonological terms, but she gives an explanation ad hoc, without searching for the cause of the difference. We doubt that the explanation can concern only a phonological development. The roots that Pokorny transcribes as *ner-(t-), *aner- ‘Lebenskraft’ and *ner- ‘unten’ (*nertero- ‘unterer’) could have undergone a certain interference and contamination due to the fact that *Hner- ‘fortis’ – which could be phonologically mistaken for *ner- ‘low, under’ – was used for designating a class or category of *homines ‘those who stay low, down on the earth’ > men’. Pokorny posited the variant *aner- (* � ner-) in order to account for Arm. ayr, Gr. ajnhvr ‘man, vir’, euj-hvnwr ‘brave, strong’, hjnorevh ‘vigor’, NeoPhryg. anar and Luv. annar ‘man’33. Hitt. innarawatar and Luv. annarumahit ‘Rüstigkeit, strength’, Hitt. innarawant-, Luv. annarummi ‘strong’ are now added to those forms justifying the hypothesis of a root *Hner- ‘power, vigor, gallantry’34. In western Indo-European languages, where forms beginning with a vowel are not attested, *nerdenotes ‘to be strong, powerful’ and not ‘to be human’. Svetonius (Tib. 1,2) informs us that the Sabinian name Neroˉ means ‘fortis ac strenuus’. The name Nerioˉ, -enis of the partner of Mars means Fortitudo ; Umbr. nerf (acc. pl.) and South-Pic. niir differentiate from viro (pl. acc.) ‘viroˉs’ in order to 31 Beekes (2003: introductory page). 32 F. Bader, “Une traversée menée à terme: noms de conquérants I.E. en étrusque”, in S. Marchesini and P. Poccetti (eds.), Linguistica e storia: scritti in onore di Carlo de Simone (Pisa 2003) 36-37. 33 IEW 765. 34 Gamkrelidze and Ivanov (1995: 703) assume that the initial innof itt. innarais the development of *Hn-.