EXTRACTED FROM Ex Anatolia Lux Anatolian and Indo-European studies in honor of H. Craig Melchert on the occasion of his sixty-fifth birthday edited by Ronald Kim Norbert Oettinger Elisabeth Rieken Michael Weiss Beech Stave Press Ann Arbor • New York Table of Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Bibliography of H. Craig Melchert . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi Ignacio J. Adiego, On Lycian Genitives in -h, -he. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Alexandra Daues, Zur Korrelation der hethitischen Konjunktion kuitman mit dem Verbalsuffix -ške- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  George Dunkel, The IE s-Aorist As an Athematic Denominative . . . . . . . . . . . . . . . .  Benjamin W. Fortson IV, On the (Non-)Antiquity of Clause-Internal -kan in Hittite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Victor A. Friedman, The Age of the Albanian Admirative: A Problem in Historical Semantics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  José Luis García Ramón, On Hittite Verbs of the Type mimma-˘h˘hi ‘refuse’: Aktionsart and Aspect in Indo-European Reconstruction . . . . . . . . . . . . . . . . . . . . .  Petra Goedegebuure, Deictic-Emphatic -i and the Anatolian Demonstratives . . . .  Roberto Gusmani†, Tracce anatoliche di una desinenza verbale indoeuropea ? . . .  Olav Hackstein, Lateinisch omnis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Mark Hale, Návyasā vácah:To Praise With a Really Old Word . . . . . . . . . . . . . . . . . . .  . J. David Hawkins and Anna Morpurgo Davies, More Negatives and Disjunctives in Hieroglyphic Luwian . . . . . . . . . . . . . . . . . . .  g veda . . . . . . . . . . . . . . . .  Heinrich Hettrich, Nochmals zu den -yā´ -Adverbien im R̊ Harry A. Hoffner, Jr., Logographic Plural Markers in Hittite Cuneiform Texts .  Stephanie Jamison, Sū´ re Duhitár’s Brother, the “Placer of the Sun”: Another Example of -e < *-as in Rigvedic Phrasal Sandhi . . . . . . . . . . . . . . . . . . . .  Jay H. Jasanoff, The Luvian “Case” in -ša/-za . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Brian D. Joseph, Revisiting the Origin of the Albanian pl. Verbal Ending -ni . .  Folke Josephson, Hittite -apa, -šan, and -kan as Actional Modifiers . . . . . . . . . . . . .  Ronald I. Kim, Possible Tocharian Evidence for Root Ablaut in PIE Thematic Presents? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Jared S. Klein, Personal Pronoun Sequences in the Rigveda . . . . . . . . . . . . . . . . . . .  Alwin Kloekhorst, Hitt. mān, mahhan, māhhan, māhhanda and mānhanda . . . . .  ˘˘ ˘˘ ˘˘ ˘ Rosemarie Lühr, Zum Mittelfeld in altindogermanischen Sprachen . . . . . . . . . . . .  vii Contents Silvia Luraghi, Experiencer Predicates in Hittite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Melanie Malzahn and Martin Peters, How (Not) to Compare Tocharian and Ancient Greek Verbal Stems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Alan J. Nussbaum, PIE -Cmn- and Greek τρᾱ¯ ν�ς ‘clear’ . . . . . . . . . . . . . . . . . . . . . . . .  Norbert Oettinger, Die indogermanischen Wörter für ,Schlange‘ . . . . . . . . . . . . . .  Georges-Jean Pinault, On the r-Endings of the Tocharian Middle . . . . . . . . . . . . .  Massimo Poetto, Un nuovo verbo luvio-geroglifico: zapa-, e la sua correlazione al luvio cuneiforme zapp(a)- . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Jaan Puhvel, Fiery Seed: Remarks on the Tiers of Hittite Royalty . . . . . . . . . . . . . .  Jeremy Rau, The Derivational History of PIE *diéu-/diu-´ ‘(god of the) day-lit sky; day’ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  ˘ ˘˘ Elisabeth Rieken and Paul Widmer, Spaltsatzphänomene im Hethitischen . . . .  Don Ringe, “Thorn” Clusters and Indo-European Subgrouping . . . . . . . . . . . . . . .  Johann Tischler, Einige Fehlschreibungen und Fehllesungen in hethitischen Texten . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Brent Vine, Old Phrygian iman . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Calvert Watkins, Toward a Hittite Stylistics: Remarks on Some Phonetic and Grammatical Figures . . . . . . . . . . . . . . . . . . . . . . .  Michael Weiss, Two Sabellic Praenomina . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  Ilya Yakubovich, Hittite aniye/a- ‘to do’. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Kazuhiko Yoshida, Observations on the Prehistory of Hittite ie/a-Verbs . . . . . . . .  ˘ Index Verborum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  viii Two Sabellic Praenomina Michael Weiss ANA mKarraik LÚAZU URULušankeleš One of the most striking common features of the Italic Sprachbund is the system of personal names. In Latin, Faliscan, Etruscan, all the Sabellic languages, and Messapic we find an essentially identical set-up where each person is identified first as an individual by his praenomen and then in relationship to his gens by his gentilic or nomen. To these names may be added optionally an indication of the father’s name either with a genitive or a patronymic adjective. The following examples may be cited: For the Indo-European languages involved, we are certain that this system is an innovation, replacing the older PIE system of individual name and patronymic that is still preserved in Greek, Sanskrit, and elsewhere. The essential change that led to the Italic system was the fossilization of the patronymic adjective as an invariable gentilic. Thus Gaius Marcius was originally Gaius, the son of Marcus, the suffix -ius being identical to the Aeolic patronymic -ιος (< PIE *-iio-) seen in Homer’s Τελαµèνιος Α‡ας. But a number of changes also affected the praenomina themselves. First, as is well known, the number of praenomina has been severely restricted in Latin—there are only eighteen in common use. This is, however, a relatively recent phenomenon since many additional praenomina, which have fallen out of use in the Classical period, are still found in archaic times, and a wealth of additional praenomina can be inferred from gentilics (originally patronymics), and many old praenomina are preserved as cognomina. ˘  These examples, with the exception of the Latin one, are the same as those used by Rix (), an article to which this introductory section owes much.  All this is common knowledge. See for example OCD, RE, Rix .  From Ronald Kim et al. (eds.), Ex Anatolia Lux: Anatolian and Indo-European studies in honor of H. Craig Melchert on the occasion of his sixty-fifth birthday. Copyright © Beech Stave Press, Inc. All rights reserved. Michael Weiss Latin: Faliscan: Oscan: Umbrian: Messapic: Etruscan: M. Claudius M. f. (CIL I .) Ca. Vecineo(s) Volti (G.G. .V) Upfals Salaviis Minies (Cp ) T. V. Voisinier (Um ) Θotor Sohinnes (IM .) - al clan (Cr .) Aule Tarxnas Lar0 But a more essential innovation is found in the structure of the personal names. A significant class of PIE individual names were compounds, and usually of the possessive or verbal governing type. A few familiar examples can be cited: Greek: Old Persian: Vedic: OHG: OCS: Gaul.: 'Αριστοφ£νης ‘whose appearance is best’ `Ιπποκρ£της ‘whose strength is like a horse’s’ Dārayava(h)uš ‘who upholds the Good’ Viśvā´mitra- ‘whose friend is everyone’ Hadubrant ‘flaming in battle’ Mı̆stidrugŭ ‘avenging friends’ Orgetorix ‘king of killers’ But although there are no obvious examples of this type among the Latin or Sabellic praenomina, the elimination of compound names must be of post-Proto-Italic date—assuming Venetic is the first branch off of Proto-Italic—since Venetic still retains some compound forms like Enogenes (Pellegrini-Prosdocimi Pa. ; Lejeune ) ‘in-born’, Voltigenei (Pellegrini-Prosdocimi Es. ; Lejeune ) ‘whose birth is desired’. However, when we turn to Italic proper, the situation appears to be very different. According to Heikki Solin, to quote one authoritative view: “Indo-European compound names, of the type seen in Greek 'Αριστοκρ£της were abandoned very early in the proto-Italic period, giving way to simple names, the origin of which often remains obscure.” Now if the statement of Solin is true we would be faced with a remarkable break in continuity of tradition between Indo-European and Italic, by which not only were individual lexical items given up, but a whole culturally central system was eliminated without trace. Schmitt has suggested that the elimination of the PIE naming system was tied to the loss of the PIE poetic tradition, which in many other Indo-European languages supplied the stock of vocabulary and collocations for personal names. But the losses suffered by the PIE poetic tradition, though grave, were not so total in Italic as commonly believed. In addition to the famous survival of the collocation ueiro pequo ( YAv. pasubiia vı̄raēibiia) in Umbrian, I might point to the probable survival of a PIE poetic collocation in Latin iūgis (*h2 iuh1 - + *gwih3 -  YAv. yauuaējı̄-), and ˘  It is likely that simplex descriptive names were also a PIE possibility. See Schmitt :–. See Evans :–.  The cognomen (originally a praenomen) Agrippa was also felt to be a compound by the Romans themselves, probably correctly. On the analysis of this name see below.  Solin :.   Two Sabellic Praenomina the South Picene texts, despite the possible influence of Greek ways, do in essence continue a native and archaic poetic tradition. In this light, I think it is worthwhile to reexamine the question. What kind of evidence would be necessary to show that the elimination of the PIE compound name system was not total and therefore that Italic shows not a surprising, violent break but a transformation of the PIE situation? Obviously, the best way to demonstrate this would be to find a compound name made up of probable PIE elements typical of the PIE onomastic lexicon. But of course if names like this were present in abundance in Italic, the whole question of the transformation of the Italic naming system would never have arisen. Nevertheless, I believe Sabellic preserves one praenomen which is undeniably a compound, and which may plausibly be taken as a continuator—in spirit at least—of the PIE way of doing things. This is the name úpfals, on which more below. But there is also another way to try to demonstrate PIE–Italic continuity. First, we must ask ourselves: How did the new type of Italic praenomen come about? On the one hand, it is clear that many Italic praenomina are simply entirely new creations with no connection to the older naming system. Some of these are borrowed from Etruscan (Aulus, Spurius). Some are simply adjectives recording probably the birth month (Mārcus(?), Quı̄ntus, Sextus). Some are theophoric in the broad sense (Mārcus < *Mārt-i-kos and Tiberius  Tiberis). Some are simply descriptive (perhaps Gnaeus  naevus ‘birthmark’, Vopı̄scus = vopı̄scus ‘surviving twin’). Once these cases are removed, there remains a class of praenomina ending in -ius, some of which are clearly related to elements of the non-onomastic vocabulary. Lūcius is obviously related to lūx ‘light’, Mānius to māne ‘in the morning’ or mānus ‘good’, Servius to servus ‘slave’, and so on. The history and interpretation of these forms is much less clear. Nor is it a priori obvious that there must be a unitary explanation for all forms in -ius. Some may be explained as denominal adjectives built with the genitival suffix *-iio-. This explanation works well for, e.g., Tiberius Tiberis and for Pūblius, earlier Poplios  *poplo- ‘army’. In addition, some praenomina may originally have been patronymics that have passed over into praenominal use. But when we turn to the more abundant Oscan praenomina in -is < *-iios, it is clear that they cannot all be derived from substantives or other praenomina via the suffix *-iio-. As Rix, building on Devoto, has noted, Oscan has a series of praenomina that bear a strong resemblance to the Italic ordinals: Dekis  *dek(V)mo-; Núvis  *nou(V)no-; Úvis (< *ohis < *ohtāuo- or analogical to Núvis?); Seis (Slunečko :) ˘ ˘ ˘ ˘ ˘  See Weiss , Watkins :chs. , , , . Salomies :, .  Salomies :, following Petersen .  Salomies :.  Salomies :, , but Rix (:) derives Gnaevus from *gnāuios.  Salomies :, .  Although what precisely ‘of the army’ was supposed to convey is not clear.  Cf. the Attic Greek use of originally patronymic forms in - δης as individual names, e.g. Θουκυδ δης. Good cases for this sort of analysis would be the relatively rare examples of praenomina in -ius clearly derived from other praenomina in -us, e.g. Aimio- from Aimo-. See Peruzzi :.  ˘˘  Michael Weiss  *sestVmo- (cf. Etr. sestuminas Cm .); Sepis  *septVmo- (cf. OU sehtums). However, since, as is now clear, these forms cannot be the ordinals themselves, if the theory has any validity, they must represent the ordinals in abbreviated form. This suggests that the suffix *-iio- was used in Italic in an abbreviating, hypocoristic function. And in fact the abbreviating function of the suffix *-iio- in the onomastic system of PIE is well established: Gk. Μενšσθιος (Hom. +)  Μενεσθšνης; OIran. *Dātiya-, OP Martiya-; Ved. Déviya-  Devadatta-. It is therefore possible that some Italic names ending in -ius may be hypocoristic shortenings of inherited compound names of the well-established Indo-European type. We may therefore suppose, and others have supposed, that the common Italic innovation leading to the Italic praenomina system was the generalization of hypocoristics at the expense of the original compound forms. If this hypothesis is correct, then it might be possible to discover some connection between Italic praenomina in -ius and possible PIE compounds. What would this evidence look like? Hypocoristics in Greek, just to mention the most at-hand comparandum, often preserve a truncated bit of the second member of the compound. So for example, the name Κλšοµις is attested as the hypocoristic for Κλεﵚνης, the tyrant of th-century Methymna (Isoc. ep. .). To judge from the Italic evidence, the hypocoristic formation principle seems to have been: substitute -iio- for all material after the first medial consonant. Thus if we could find a plausible case with such a second member fragment, we would have good evidence for the continuity and shortening hypothesis. In what follows, I will suggest that the Sabellic praenomen Vı̄bius is in fact an example of just the sort of evidence we are looking for. Let us first consider the cases of apparently compound praenomina in Latin. There are at least two Latin praenomina that at first glance look like compounds. These are Opiter, supposedly meaning ‘having a dead father and living grandfather’, and Agrippa ‘born feet first’. Opiter, for which a genitive Opetris is inferable from the abbreviation Opetr. on the Fasti Consulares of  , is hard to separate from pater, and therefore should be a compound of some sort. Solmsen’s suggestion *auo-pater ‘whose father is a grandfather’, though fitting semantically and acceptable as a bahuvrihi, requires us to accept a non-urban development of *au to ō and to disregard the short scansion of the first syllable at Silius Italicus .. If we do take Silius Italicus seriously, it is hard to think of any possible analysis for o-. The PIE particle *h2 o- ‘on’ is not otherwise attested in Italic, and how exactly ‘on the father’ or ‘whose father is on’ could develop into the sense attributed to Opiter by the grammarians is unclear. Finally, *h2 o- is not attested in clearly denominative compounds. Another mere possibility would be an a ˘ ˘ ˘ ˘  Devoto . Schmitt .  Wackernagel and Debrunner :.  Rix :: “sie musste letztlich von den auch für das Vorosk. vorauszusetzenden Namenkomposita stammen.”  Solmsen and Fraenkel :.  Salomies :– and –, respectively.  Sternuntur leto atque Opiter, quos Setia colle.   Two Sabellic Praenomina tergo hypocoristic of the type identified by Schmitt in OP Gadatas < Bagadāta-. In this case, the o could represent the thematic vowel of the preceding member. But it seems unlikely that a hypocoristic process of this sort would create a V-initial syllable rather than a CV syllable as in Gadatas. In the case of Agrippa, agri has since Schulze been compared with Ved. ágre ‘at the tip, beginning’ (loc. of ágram ‘tip’) and Av. aγrō ‘first’, and the second half has been analyzed as a hypocoristic shortening of *ped- ‘foot’ with consonant doubling of the Κλšοµµις  Κλεﵚνης type. Schulze later doubted his suggestion, since he thought that the information about the meaning of the name could have been a pure invention. Yet it is not easy to see on what the Romans might have based this invention, since agr- in the meaning ‘first’ does not otherwise survive (in fact the Romans identified agr- with aegre) and so I am inclined to accept Schulze’s original interpretation. Yet although these two names are in all probability originally compounds and show that, in principle at least, compound names were not excluded from the stock of Roman praenomina, neither of them continues the Indo-European tradition, but instead are both very much in the Roman style. Indo-European compound names, at least of the aristocratic sort we are most familiar with, are not usually prosaic comments on difficult deliveries, birth order, or the status of the father. Rather, the typical IE compound name is ein kleines Gedicht. The vocabulary of compound name elements shows close links to the traditional language of poetry. It is among the other noncompound Roman praenomina that Agrippa and Opiter find their semantic home, cf. Postumus, Caeso, etc. The Romans included neither wolves, nor horses, nor any of the typical Indo-European elements among their names. Úpfals There is one praenomen occurring in Sabellic whose status as a compound has been clear since at least Corssen, but whose further analysis still remains uncertain. This is the name Úpfals. This praenomen is attested several times in the Samnite inscriptions of Campania: Capua: úpfals. patir. miínieis (Cp ; :) ˙ s (Cp ; :) ˙ upfals: salaviis: minie ˙ s) upfaleis saidiieis (Cp˙; Cumae: kluva(z): rahiis: upfalleis (Cm . C ; :) ˙˙ dekis. seppiis. úpf(alleís) (Po ) Pompeii:  Schmitt . Schulze :.  Schulze :.  Plin. Nat. .: in pedes procidere nascentem contra naturam est, quo argumento eos appellavere Agrippas ut aegre partos [v.l. aegri partus].  Corssen .  Also restored with probability in Cm ..   Michael Weiss Outside of Campania it is found only at Histonium, in the area of the Frentani: viíbis. úhtavis. úff(alleís) (Fr ). In addition to the praenomen, a derived gentilic or perhaps patronymic uφaliies appears to be attested on an Etruscan/Pre-Samnite inscription from Fratte dating to the th–th century  (Vetter  a): stlakies uφaliies mi. This form is matched most immediately by the gentilic Ofalius (CIL I .) attested on a Latin inscription of unknown provenience from the nd century .  See Bellelli : for the reading with φ.  Two Sabellic Praenomina In addition, the well-attested Latin gentilic Of(f )el(l)ius with unstressed weakening may be related. Perhaps most familiarly, Horace Satire . preserves the cognomen Ofellus derived from the praenomen. Finally, forms of the name are attested in the Greek and Etruscan Nebenüberlieferungen in the forms 'Οφαλ(λος) , incorrectly restored to 'Οφ£λ(λιος) (SEG :), 'Οφαλλου (SEG :, Lipara, th–nd  = I.Lipara ) and Uφale (Cl .). Up to the final consonant, the establishment of the base form is unproblematic. The initial vowel is o, as is clear from the spelling ú in the inscriptions written in the post-reform Oscan alphabet (Cp , Po ). The examples spelled with u are all in pre-reform alphabets. Next, we must have originally had the consonant cluster pf which has been preserved in Campanian Samnite but been assimilated to ff among the Frentani and in those dialects which were the source of Lat. Ofalius and Grk. 'Οφαλ(λος). It is noteworthy that this assimilation or simplification has already taken place in the Pre-Samnite inferable from the Fratte inscription. This may perhaps be taken as another indication of the close relationship between Pre-Samnite and South Picene and Umbrian. The one major uncertainty involves the l. In only one instance do we find the l spelled doubly: Cm  C  upfalleis. Against this there are three examples with single l. Yet none of the three instances of single l spellings is probative, since none of these texts has other clear instances of the spelling out of double consonants. However, there are some additional arguments that favor setting up a stem with geminate l. First, we may note the lack of syncope of the medial syllable. If the stem is set up with only one l, then the lack of syncope has to be explained as either analogical from the nominative singular or as the result of a secondary compound stress on the a of the second syllable. If the stem is set up with geminate l, then the lack of syncope is straightforward. A second argument may be drawn from the form of the probably related Latin gentilic Of(f )el(l)ius, which prima facie might be taken as evidence for a geminate. However, this in itself is also not conclusive since the double l in the gentilic may be the result of yod-doubling and need not imply that the praenomen also had a geminate l. The Greek alphabet form 'Οφαλλου from Lipara is more probative. Finally, there is the fact that the nominative singular is Upfals. The non-assimilation of the secondary -ls in Auslaut contrasts with the regular assimilation of this cluster seen for example in Mutíl (nPg  +) < *mutilos, famel (Pg ), Lúkídel (Hi ), and aídíl < Lat. aidı̄lis. It is tempting to explain this surface difference between Upfals and Mutil with Buck and others by supposing that whereas single *ls assimilated to ll, *lls remained unassimilated. The only other comparable form is the praenomen upils (Cm ), which is probably to be compared with the Latin cognomen Opillus, also apparently with a stem-final geminate l. Nevertheless, it is hard to exclude the possibility of an analogical restoration of the s in either of these cases. All in all, however, the case for geminate l is decidedly stronger than the case for single l, which can explain  Hor. S. .–: Quae virtus et quanta, boni, sit vivere parvo/ (nec meus hic sermo est, sed quae praecepit Ofellus/ rusticus, abnormis sapiens, crassaque Minerva)/ discite.  In fact, the nominative úpil (Sa ) is undoubtedly a token of the same name.  Michael Weiss much less and which is after all directly contradicted by a unique witness. Therefore, the stem must be set up as *opfallo-. Once we have determined the true form of the stem, it is clear that, if the name is to have any analysis in Sabellic at all—and there is nothing to suggest that the name is anything but Sabellic—it must be a compound. Given PIE and Italic root structure constraints, it is virtually certain that there must be a morpheme boundary between the p and the f of *op-fallo-, and given the absence of any suffix of the shape *falloor combinations of suffixes which might result in *fallo-, it is highly probable that *fallo- must be not a suffix but a potential free form, and therefore that *opfallo- is by definition a compound name, the first element of which must be *op- and the second element of which must be *fallo-. Now the possibilities for the analysis of *op- are in essence two. Either *op- is the well-attested Sabellic adverb/preposition op (Osc. úp) ‘near’ or ‘at’, or it is the root noun cognate with Lat. ops ‘wealth, resources’. It is not generally recognized that forms of the root *op- in the meaning ‘resources’ exist in Sabellic, but in fact a probable derivative of this root is seen in the Oscan ti-stem ufti-: ufteis gen. sg. (Cp .), uhftis nom. pl. (Cm  C) ‘help’. When we turn to the analysis of the second half of the name, the possibilities are unfortunately many more and the uncertainty is consequently much greater. The sources of this uncertainty are three: the multiple sources of Sabellic f (< *dh, *bh, *gwh), the multiple sources of Sabellic ll (*-ln-, original *-lu-, *-ll-, *-ls-, and *-nl-) and the absence of specific semantic evidence which is endemic to the interpretation of personal names. Nevertheless, it is possible to narrow the possibilities to some extent and to make some decisions about ranking the relative plausibility of alternative analyses. First, the possibility that the ll results from *-nl- should be ranked as relatively unlikely. The well-established PIE roots which begin with something that could give f and end in n are *gwhen- () ‘strike’ and *gwhen- () ‘swell’, and although either could be theoretically plausible in a personal name, the a vocalism would have to reflect a zerograde outcome originally typical of initial syllables and therefore be analogical in this compound form. Nor does the hypothesis of a suffix -lo- added to a zero-grade verbal root that is required for this hypothesis correspond to any known Sabellic or Italic derivational type. Finally, there is no support for a stem *fan- in either the onomastic or the general lexicon of Sabellic. In contrast to the absence of support for the *-nl- hypothesis, the -lC- hypothesis ˘  Also occurring in the first part of the compounds opitulus ‘bringing help’ (Paul. Fest. p.  L), a cult epithet of Jupiter, and opiparus ‘sumptuous’ (Plaut. +).  See Weiss :–.  It seems not to be generally known that ls assimilated to ll in Oscan as in Latin. But there are no surfaces, ls sequences in Oscan that are not patently analogical, and the crucial evidence is provided by the name Helleviis (Cp ). It is true that the double ll is only attested one time in this name, but the quondam existence of some phoneme after the l is not in doubt, for otherwise the sequence of original *lu would have assimilated to ll as in Latin. That this gentilic is identical to the Lat. Helvius, and that both in turn are derived from the adjective helvus ‘dun’ and its Oscan cognate, can hardly be doubted. The adjective helvus must be compared directly with Lith. gel̃svas ‘yellow’ < PIE *ghelsuo-. ˘ ˘  Two Sabellic Praenomina does find some support in the Sabellic onomastic lexicon. Most immediately, the stem *fal- recalls the name faletne in the vocative recently identified by Rix on the PalaeoUmbrian text of Poggio Sommavilla (Um ). He further suggests that the inferred root *fal- might be compared either with the root *d heh2 l(hx )-/*dhal(hx )- ‘thrive’ or the root *bhleh3 - ‘bloom’. And in fact these are the only plausible PIE root possibilities. The semantics of the name would then be similar to the Latin cognomina Flōrēns, Flōrentı̄nus. Taking a root ending in l as our working hypothesis, we may turn to the question of identifying the consonant responsible for the doubling. Unfortunately, this too cannot be decided with certainty. However, one observation tends to favor s: both the roots *dhal(hx )- and *bhleh3 - have well-attested s-stems (flōs ‘flower’, Osc. Fluusaí ‘Florae’ [Po , Sa  a ]; Gk. θ£λος ‘shoot; scion’). If we accept the reconstruction *opfalso-, then we must further note that the sequence *-ls- was probably original and not due to the Common-Sabellic medialsyllable syncope. For a form *opfalVso- would have come out as †opflVss in the nominative and †opfls- or whatever the outcome of that would have been in the oblique cases. The conclusion we reach then is that the most probable preform for *opfallo- is *op-falso-, the second element of which is derived either from the root *dhal(hx )- or the root *bhleh3 -. Since in principle Italic etymologies should be found in Italy, let us consider the second possibility first. Lat. flōs is a masculine s-stem, and therefore one would initially prefer to analyze the form as an amphikinetic animate s-stem of the cruor type. The form flōs may unproblematically be derived from *bhleh3 -ōs. The proterokinetic neuter from which flōs was derived by internal derivation would have been nom. *bhleh3 os > *flōs, gen. *bhlh3 -es-os > *falesos. The chance collapse of the nominative of the neuter ˚ and the masculine in *flōs could have motivated the generalization of the oblique stem fal-, leading to a paradigm *falos, *fales-os. In PIE, when an athematic noun became the second member of a possessive compound, it could be treated in one of two fashions: either the athematic noun adopted a new ablaut pattern, proterokinetic becoming hysterokinetic, etc., or the athematic noun adopted thematic inflection, and the addition of the thematic vowel apparently required a zero-grade in the immediately pre-thematic suffix. Both patterns are illustrated by the derivatives of the s-stem from the root *h3 ed- ‘smell’. In Greek the internal-derivational process is illustrated by the various forms in -èδης (¢νθεµèδης ‘flowery’ etc.) but in Latin the thematic process is shown by the suffix -ōsus which, as Wackernagel was the first to see, is from *-ōdso-. Another Latin example is bı̄mus ‘two years old’ < *dui-himo- ‘with two winters’. Therefore a neuter *falos ‘bloom, flourishing’ would regularly have become *Xfalso- in the second member of a compound. *op-falso- would then mean ‘having a flourishing which is wealth’, i.e. ‘flourishing with wealth’. For the semantics, cf. the not uncommon Latin collocation flōrēre opibus, e.g. Cic. Verr. . cum Sicilia florebat opibus et copiis; Cic. Div. .. cum florent opibus fortunisque; Liv. .. inde ˘   Rix . Apud Niedermann .  Michael Weiss Turnus Rutulique diffisi rebus ad florentes opes Etruscorum Mezentiumque regem eorum confugiunt. If the first member, however, was the preverb/preposition op ‘near, at’, the name could be analyzed as an œνθεος-type compound ‘having a flourishing which is near at hand’. If the stem in question in fact belonged to the root *dhal(hx )-, the basic meaning of which de Lamberterie has argued is ‘thrive’, then the details would be somewhat different: an s-stem comparable to Greek, θ£λος < *d hal(hx )os would yield an Italic *falos, and from there the hypothesized development would be the same. It is noteworthy that the stem of θ£λος does play a role in the Greek onomastic lexicon, e.g. 'Επιθ£λης Messenia, SEG XLI ., nd ; Καλλιθ£λης Athens, IG  .,  ; Miletus, Mb, rd , and has a close semantic connection in the corpus of early Greek epic with various kinds of wealth and well-being. The stem θαλ- plays a crucial part in Hesiod’s description of the land of the just: Op.  το‹σι τšθηλε πÒλις, λαοˆ δ’ ¢νθεàσιν ™ν αÙτÍ, Op.  θ£λλουσιν δ’ ¢γαθο‹σι διαµπερšς. An interesting semantic parallel to ‘flourishing with abundance’ exists in the IndoIranian compound *prHand hi- (Ved. púrandhi- f., OAv. par ndi-), both ‘abundance’ ˙ ˚ and the name of a goddess of abundance. Karl Hoffmann has suggested very plausibly h that this form is to be analyzed as *prH-and i- ‘Blüte der Fülle’ < *plh1 -h2 endhi- (cf. Gk. ˚ ˚ ¥νθος ‘flower’ etc.). At the end of this part of our investigation, we may firmly state that at least one Sabellic praenomen was a compound. It is much less certain whether the compound continues the uninterrupted Indo-European tradition, but the most probable analyses are consistent with such an interpretation. e Vibius The praenomen Vibius is given for the grandfather of P. Sestius Q. f. Capitolinus Vaticanus, the consul of  , whose floruit must have been around  . A gentilic derivative of the praenomen Vibius is attested in the Etruscan form vipiiennas (Ve .) on the famous dedication of Avile Vipiiennas, the brother of Caele Vibenna of early Roman legend. This text is dated by Rix to the th century. But despite these early attestations, the name is neither Latin nor Etruscan, but Sabellic, as Mommsen pointed out already in . The praenomen is rare in Roman contexts, and the *-iiosuffix is well paralleled in Sabellic, but not in native Etruscan morphology. The other early historically attested bearers of the praenomen Vibius are all Sabellic, e.g. Livy .. mentions a Vibius of Bruttium (nd half of the rd century ), and Vibius Virrius, the rebel leader of Capua, dates from the time of the Second Punic war. The name is very commonly attested in Oscan and Paelignian inscriptions. From Sabellic it passed to Latin, and to Etruscan in the form Vipi. The Etruscans then ˘  Lamberterie :. Apud Narten :.  Mommsen .  On Virrius see Buecheler :–.  All the examples can easily be found in Slunečko .   Two Sabellic Praenomina transmitted it to Faliscan. If the name is to have an historical explanation, it must have one in terms of Sabellic sound laws. The first i of the name is long. This is abundantly confirmed not only by meter (Martial ..; C.E. , ), but also by spellings with i longa at CIL V. and in Oscan with plene spelling viíbis, and by the false archaism Veibius (C. Veibius Vaarus on a coin of  ; RRC /–). Our analysis must then start with a Proto-Sabellic *uı̄biios. Since we are dealing with Sabellic historical phonology, quite a few ambiguities that we would have to face in the case of a Latin form are ruled out. Initial u- has only one source, i.e. PIE *u. Likewise *ı̄ can only be from PIE *ı̄, i.e. *ihx . Sabellic b has two possible PIE sources, *b or *gw. Thus *uı̄biios may be traced back to either *uihx biios or *uihx gwiios. PIE *b was very rare, and this by itself favors the second reconstruction. There is in fact a supposed PIE root *ueib- ‘turn, shake’ (Lat. vibrāre, Go. weipan ‘crown’, OE wı̄pian ‘wipe’; not recognized by LIV), but this root was *uiP- in the zero-grade (Ved. vípra- ‘trembling, inspired’), which would not provide the required length. The second reconstruction *uihx gwiios is therefore the more likely. But here we run into the problem that a root of the shape *ueihx gw- or *uiehx gw- is not known and, although not entirely impossible, would be an example of a fairly rare type (cf. *d heihx gw- ‘fix’). Since a root shape *uihx gw- would be unusual, it is worth considering whether there is in fact a morpheme boundary between *uihx - and *-gw-iios. Once this cut is made, an analysis of the first segment immediately presents itself. The only plausible candidate for *uihx - is PIE *uihx s ‘(life-)force’ (Lat. vı̄s, Greek (#)‡ς, Ved. váyas-). This form is well attested in Italic, although admittedly not in Sabellic. Its place in the PIE onomastic lexicon is secured by Greek names with the instrumental (#)‹φι- like Mycenaean wipi-no-o = 'Ιφ νοος and with the simple stem 'Ι-κšρτης (Aeolic Cumae) and possibly OS Wı̄-rı̄c and Gēr-wı̄. This means that the *-gw-iios part may belong to a different stem, i.e. that *uı̄biiois the remnant of a compound name beginning with *uı̄- ‘(life-)force’. There are two plausible candidates for the second member of the compound: () the root *gwih3 ‘live’ (Gk. β ος, Lat. vı̄vus etc.), in which case the meaning of *uı̄biio- would be ‘whose life has life-force’; or () the root *gwei- ‘force’ (Gk. β η), therefore ‘whose force has life-force’. Both stems are attested in Oscan. In fact, they are even used in the alliterative phrase biass biítam found on an Oscan curse tablet (Cm ), which is the Oscan equivalent of Lat. vires vitam (would be Osc. *vı̄- *bi-). Occurrence or nonoccurrence in Sabellic cannot be decisive. Both stems are well attested. It is interesting to note the alternation of Gk. #‡ς and β η in the Homeric periphrases for Herakles: β η `Ηρακληε η (L. )  ‡ς `ΗρακλÁος (Hes. Th. ), and even their cooccurrence in ‡ς . . . β ης `Ηρακληε ης (Hes. Th. ). Furthermore, such compounds of semantically nearly identical elements can be paralleled not only by 'Ικšρτης = 'Ιφικρ£της but also ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘˘ ˘ ˘˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘  PIE roots generally obey sonority sequencing with falling sonority in the coda. Roots of the shape CEIHT would have either a flat or minimally falling sonority sequence.  Cf. Enn. Ann.  (Sk) vires vitaque.  The analysis suggested here fits well within the complex of PIE personal names combining  and  delineated by García Ramón .  Michael Weiss by the PIE collocation *h2 iu- *gwih3 - ‘having a life which is lively’, reflected by Lat. iūgis, Gk. Øγι»ς etc. If either of these analyses is correct, then it no longer is necessary to interpret -iioas a hypocoristic suffix in *uı̄biio-. *gwih3 o- ‘life’ is well attested in the second part of compounds, and in the second part of a compound the first-declension β α becomes thematic as in ¢ντ βιος (Hom. +) ‘opposing force to force’, Øπšρβιος (Hom. +) ‘overweening, of overwhelming strength’. Thus Proto-Sabellic *uı̄biios may be interpreted as a complete unshortened compound name made from inherited Indo-European elements. It would then owe its survival in pristine form to its chance resemblance to other names which did in fact contain the hypocoristic suffix *-iio-. One other possibility that deserves consideration is that the second member of the putative compound is the root noun *gwih3 -s found in Av. yauuaē-Jı̄- and possibly in Gk. Κλšο-βις. The combination *uihx -gwih3 -s would regularly become *uı̄bı̄s, and this easily could have fallen together with the class of praenomina ending in -is resulting from the syncope of *-iios. These two Sabellic praenomina, Úpfals and Vibius, are valuable relics of an older way of doing things. ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ ˘ References Bellelli, Vincenzo. . “Nola, –.” Rivista di epigrafia etrusca, Studi etruschi –: –. Buecheler, Franz. . “Weihinschriften von Capua.” Rheinisches Museum für Philologie :–. Corssen, W. . “Zum oskischen Dialekt.” Zeitschrift für Vergleichende Sprachforschung :–. Devoto, Giacomo. . “Nomi proprî.” Studi etruschi :–. Evans, D. Ellis. . Gaulish Personal Names: A Study of Some Continental Celtic Formations. New York: Oxford University Press. García Ramón, José Luis. . “Homme comme force, force d’homme: Un motif onomastique et l’étymologie du vieille irlandais gus.” In Georges-Jean Pinault and Daniel Petit (eds.), La langue poétique indo-européenne. Paris: Peeters, –. Lamberterie, Charles de. . Les adjectifs grecs en -υς: Sémantique et comparaison. Louvain-la-neuve: Peeters. Lejeune, Michel. . L’anthroponymie osque. Paris: Belles Lettres. Mommsen, Theodor. . Oskische Studien. Berlin: Nicolaische Buchhandlung. Narten, Johanna. . Der Yasna Haptaµhāiti. Wiesbaden: Reichert. Niedermann, Max. . “Studien zur Geschichte der lateinischen Wortbildung.” Indogermanische Forschungen :–. OCD = Simon Hornblower and Antony Spawforth (eds.), The Oxford Classical Dictionary. rd ed. Oxford: Oxford University Press, .  Cf. the name of Κλšοβις’s brother, Β´ῑτων. Yet another possible analysis for the second half of Vibius, as José Luis García Ramón has suggested to me, is a connection with Gk. βιÒς ‘bow’, Ved. jiy´ ā ‘bow-string’. Thus the paraphrase for Vibius would be ‘whose bow is strong’, another very plausible PIE-type name.  Two Sabellic Praenomina Peruzzi, Emilio. . “Onomastica e società nella Roma delle origini.” Maia :– . Petersen, H. . “The numeral praenomina of the Romans.” Transactions of the American Philological Association :–. RE = Georg Wissowa et al. (eds.), Paulys Realencyclopädie der classischen Altertumswissenschaft. Munich: Druckenmüller, . Rix, Helmut. . “Zum Ursprung des römisch-mittelitalischen Gentilnamensystems.” In Wolfgang Haase (ed.), Aufstieg und Niedergang der römischen Welt, Vol. I, Part . Berlin: de Gruyter, –. ———. . Revue of Lejeune . Kratylos :–. ———. . “Römische Personennamen.” In Ernst Eichler et al. (eds.), Namenforschung: Ein internationales Handbuch zur Onomastik, Vol. . New York: de Gruyter, –. ———. “Il testo palaeoumbra di Poggio Sommavilla.” Studi etruschi :–. Salomies, Olli. . Die römischen Vornamen: Studien zur römischen Namengebung. Tammisaari: Ekenäs. Schmitt, Rüdiger. . “Entwicklung der Namen in älteren indogermanischen Sprachen.” In Ernst Eichler et al. (eds.), Namenforschung: Ein internationales Handbuch zur Onomastik, Vol. . New York: de Gruyter, –. ———. . “Bemerkungen zu dem sog. Gadatas-Brief.” Zeitschrift für Papyrologie und Epigraphik :–. ———. . “On Old Persian hypocoristics in -iya-.” In Hans Henrich Hock (ed.), Historical, Indo-European, and Lexicographical Studies: A Festschrift for Ladislav Zgusta on the Occasion of His th Birthday. New York: de Gruyter, –. Schulze, Wilhelm. . “Agrippa.” Zeitschrift für Vergleichende Sprachforschung :. ———. . Zur Geschichte Lateinischer Eigennamen. Berlin: Weidmann. Slunečko, Vladimír. . “Beiträge zur altitalischen Onomastik.” Listy Filologické : –. Solin, Heikki. . “Names, Personal, Roman.” In OCD, –. Solmsen, Felix, and Ernst Fraenkel. . Indogermanische Eigennamen als Spiegel der Kulturgeschichte. Heidelberg: Winter. Wackernagel, Jacob, and Albert Debrunner. . Altindische Grammatik, Vol. II, Part : Die Nominalsuffixe. Göttingen: Vandenhoeck & Ruprecht. Watkins, Calvert. . How to Kill a Dragon: Aspects of Indo-European Poetics. New York: Oxford University Press. Weiss, Michael. . “Life everlasting: Latin iūgis ‘everflowing’, Greek Øγι»ς ‘healthy,’ Gothic ajukdūþs ‘eternity’ and Avestan yauuaēJı̄- ‘living forever’.” Münchener Studien zur Sprachwissenschaft :–. ———. . Language and Ritual in Sabellic Italy. Leiden: Brill. 