ON THE INDO-EUROPEAN ORIGIN OF TWO LUSITANIAN THEO- NYMS (LAEBO AND REVE)

Lusitanian, the Pre-Celtic Indo-European language of the Hispanic Peninsula, demonstrates numerous similarities in theonymy to the Celtic areas, e.g. Lusit. Iccona (dat. sg.) = Gaul. Epona ‘the horse-goddess’, Lusit. Lucubo (dat. pl.) = Gaul. Lugoves (nom. pl.) and Celtiber. Luguei (dat. sg.). Other religious comparisons relate to an even larger Italo-Celtic geographical area, e.g. OLat. Pales (f.) and Lusit. Trebo-pala, Tenco-pala, Old Roman suouetaurilia and the analogical triple animal offerings of Lusitania (i.e. porcom oilam taurom). The author suggests two new Italo-Lusitanian equations in theonymy (namely: 1. Lusit. Laebo = Lat. Laribus, 2. Reve = Lat. Ioui, Osk. diuvei). Both comparisons are firmly documented by the Latin-Lusitanian texts, and additionally the latter bears a close resemblance of the formations (the same innovational declension stem *dyeu3⁄4-, not *diu3⁄4-) and an interesting exclusiveness of the epithets (e.g. Lusit. Reve Laraucu = Lat. Ioui Ladico [both dedications from Orense]). The phonological development of IE. *d to Lusit. r, documented by 4 different instances (2 indubitable ones), occurs in some Italic languages, but it is absent from the Celtic language world.

The Lusitanian inscription from Cabeço das Fráguas contains not only a list of sacrificial animals (porcom -oilam -taurom = Old Latin su-ouetaurilia) in the accusative (sg.), but also a number of theonyms in the dative, among which two names, namely Trebopala and Iccona, are noteworthy. The former seems to be a semantical equivalent of Vioe-pala (literally 'a protector of the dwelling'), a Vedic semigoddess of mare shape, thus the element -pala (attested also in another Lusitanian deity's name Tencopala) is a perfect example for the pretention. The latter form, Iccona, may represent a Lusitanian equivalent of the Gaulish horse-goddess Epona (cf. Best 1982, pp. 64-65;Maggi 1983, pp. 58-59). If so, then we are able to add a further item to the list of the Lusitanian animal names of Indo-European origin: 3 This argument is, in my opinion, of no value. I do not agree with the traditional view 5) Lusit. icco-< IE. *éu¾ os 'horse' (cf. Skt. áoevah, Lat. equus, Gk. ¹ppoj, Dor. also ¹kkoj) vs. Gaulish epo-, OIr. ech 'horse ' and W. ebol 'colt' (Brittonic *epaº los). Note that the Lusitanian equivalent demonstrates two features untypical of Celtic forms: (1) a different initial vocalism (i-) and (2) the geminate -cc-. Anderson (1985) and Untermann (1987) have recently tried to prove the Celtic character of Lusitanian on the basis of name-correspondences. However, these attempts appear to be unproductive. I agree with Karl Horst Schmidt (1985, p. 325) that «name-correspondences are not sufficient proof for genetic relationship, as they can result from language contacts».
The retention of *p is a phonological feature that excludes completely a Celtic origin of Lusitanian. A different opinion is expressed by Untermann 1987, p. 74: «Ich fürchte, eines Tages werden die Keltisten lernen müssen, mit dem p zu leben». Schmidt rejects completely these words. He points out that: «As long as Celtic is defined among other things by the loss of IE *p in anlaut, the Lusitanian ins[cription]s, containing porcom 'pig' (with a p-!) do not meet this definition. To incorporate them into the Celtic languages family, as Untermann (1987) does, violates the elementary principles of linguistic reconstruction» (Schmidt 1992, pp. 55-56).
As the discussion held at the third and fourth Colloquia on Languages and Peoples of the Hispanic Peninsula clearly demonstrated, most Celtologists and Indo-Europeanists agreed that the retention of IE. *p is one of the principal arguments for the non-Celtic character of Lusitanian. Karl Horst Schmidt lists two further premises of syntactical nature, which separate Lusitanian from Common Celtic, namely: (1) the basic word order is of the type ODS (= direct object + indirect object + subject), (2) «the position of accusative object preceding dative object» (Schmidt 1985, p. 329). Stipulating that «syntactic criteria are not sufficient in themselves to prove a linguistic relationship», Schmidt discusses a number of additional (phonological, morphological and lexical) features and concludes that the non-Celtic character of Lusitanian may be demonstrated by the following additional facts: (3) preservation of the Indo-European phoneme *p, which disappeared in Celtic in most contexts; (4) use of the conjunction indi 'and', not attested in Celtic at all; (5) development of a present inflection of the root *do-'to give', doenti 3 , unparalleled in Celtic; (6) lexical differences. As EM LXVII 1, 1999 that the Lusitanian verb doenti derives from the Indo-European root *do-'to give'. According to my opinion, the usual analysis the Lusitanian form as non-reduplicated equivalent of Greek dídousi (Dor. dídonti) 'he gives' is impossible from both the phonological point of view (IE. *d yields regularly Lus. r) and the morphological one.
4 This theonym appears also in the Latin inscriptions of Lusitania as Laebo or Laepo (Blázquez. 1991, pp. 140-141). a result, Karl Horst Schmidt (1985, p. 338) "tentatively" defines Lusitanian as «a language that belongs to the southern group of western IE languages and borders on eastern IE».
Additional evidence for the non-Celtic character of Lusitanian may be adduced from the indigenous divine names of the Lusitanian-Galician regions. The inscription from Cabeço das Fráguas, which is the most valuable text in Lusitanian, contains names of five different deities (I underline them below, treating the word loiminna as a special epithet of Iccona, as well as incomplete form tre[ occurring after Reve). It reads the following: Maggi (1983), who convincingly indicated a few related Indo-European equivalents for Trebopala and Iccona, I intend to discuss here two other Lusitanian theonyms: Laebo and Reve, which also demonstrate some similarities to the Italic, especially Latin, theonymy.

11
The Indo-European god *Dyews (päter) is not immediately attested in the Celtic tradition. However, Caesar informs that all the Gauls originate from Dis pater (Lat. a Dite patre). In my opinion, the Roman commander tried to render here Gaulish *Dis atir ('Dis the father') partially by its Latin phonological equivalent (Gaul. *Dis = Lat. Dis), partially by the semantic one (Gaul. atir = Lat. pater). Note that the Vedic religion, as well as the Gaulish one, knew the god Dyaus pita only as 'the father of gods and men', but all the different Indo-Iranian nations, like the other Celtic tribes, lost this deity in general. root *Las-, is well attested in all west dialects of Indo-European, i.e. Celtic, Italic and Lusitanian.

Lusitanian Reve (Laraucu) = Latin Ioui (Ladico).
In his article on the principal god of the Lusitanians, whose name is attested nine times as Reve (in dat. sg. only), Francisco Villar 1996 discusses different opinions as to the origin of this god, trying to explain the etymological association not only of the divine name, but also numerous epithets of the deity. Unfortunately, his suggestions and results are highly subjective 9 . He does not take into account two essential aspects: a) the god in question was both the principal (deus maximus) 10 and most popular in the Lusitanian pantheon. b) the dative singular Reve indicates the declensional stem Rev-, which is worth mentioning by reason of its exceptionality and rarity.
These two premises induce us to compare the Lusitanian god with the Indo-European principal god *Dyeu¾ s (dat. sg. *diu¾ -ei, also *dyeu¾ -ei), attested in the mythological imagination of most Indo-European nations (for lexical evidence, see Witczak-Kaczor 1995, pp. 269-270) 11 . The formal aspects of the suggested comparison are perfect, but the phonological 12 Bermejo Barrera, reviewing «los exvotos dedicados a Júpiter Ladi[c]o y a Júpiter Laraouco» (1986, p.138), also interprets Reve as Iupiter, the Roman chief of the gods (ibid. pp. 126-127). His equation is based on a combination of formal similarities. 13 Cf. Reo Paramaeco (Lugo), according to Untermann, 1985, p. 359, no. 16.4. It is a simple possibility, not a firm evidence. The same refers to the next suggestion.
I am inclined to give a positive answer, and the reason is that such a development is not only phonetically possible, but also it may be detected in Lusitanian vocabulary and onomastics. It is necesary to indicate the following three arguments: c) The process of IE. *d > r is not phonologically unfounded, as it occurs in some Italic languages (e.g. in Umbrian), but not in Latin. The Celtic languages do not attest such development. The Lusitanian documentation is not plentiful, but informative. The verb doenti (3 pers. pl. active), attested in the Lusitanian inscription of Lamas de Moledo, seems to be the unique counterexample. However, the usual analysis of doenti as a non-reduplicated form with the meaning '(they) give' (cf. Greek dídousi, Dor. dídonti) is far from being certain. I prefer to see in it a compound verb containing IE. *H 1 senti (> Gr. Myc. e-e-si, Att. e±sí, Skt. sánti, G. sind vs. *H 1 sónti > Lat. sunt, PSl. *so¿ t‚).
We may conclude with certainty that the principal god of the Lusitanians was *Revs (attested only in dat. sg as Reve), a formal and etymological cousin of the Roman Iupiter and the Greek Zeus.

Conclusions
The Lusitanian (inscriptional) documentation, much of it in the form of votive offerings, refers in large degree to the names of indigenous deities. This theonymic evidence is sometimes highly useful not only in researching the native religions of the Iberian Peninsula, but also in reconstructing the historical development of Lusitanian phonology.
Lusitanian belongs, in my opinion, to the western subgroup of the Indo-European languages, but it differs from the Celtic speech by some phonological phenomena (e.g. in Lusitanian IE. *p is preserved, but IE *d is changed into r; Common Celtic, in contrary, retains IE. *d and loses *p).