1. Introduction1

In this paper I will try to provide a general overview of the linguistic situation in ancient Callaecia by analyzing the linguistic evidence provided both by the literary and the epigraphic sources available in this westernmost area of continental Europe.

I will first review some facts that we have to bear in mind in order to understand the linguistic evidence presented here. There is no doubt that Celtic people had reached the westernmost coast of continental Europe by the early first century AD. Apart from a linguistic analysis that might allow us to conclude that there were Celts in this area, several Classical sources explicitly state that some of the peoples in this area were Celtic or at least Celtic-like2. This is the case with the Artabri3, the Nerii4, the Supertamarici5, the Praestamarici6, and the Cileni7. In other instances the naming formula found in some Latin inscriptions provides the information that a particular group of people was Celtic8. This is the case, for example, with one of the peoples occupying the banks of the Tamara River (today's Tambre River), the Supertamar(i)ci. The inscription CIL II 2902 = 5667 reads as follows:

Fusca Coedi f(ilia) Celtica Superta(marica) (castello) Blaniobriensi
And AE 1997.863 (= HEp.7.397), an inscription from Crecente LU, reads:
Apana . Ambolli f(ilia) / Celtica Supertam( arica) (castello) Miobri / Apanus

I use the description 'Celtic' or 'Celtic-like', for, as de Hoz (1992: 9) has pointed out, the Greek and Latin literary and epigraphic sources only use the word Celtae for the Celtiberians when dealing with peoples of the Iberian Peninsula; for other Celtic peoples they use the derived adjective Celtici instead. As de Hoz suggests in another paper (1997: 107-108), this must mean that even if these other Celts of Hispania were recognized as such, there was somehow the feeling that they were 'special' Celts, that is, that in spite of the similarities they shared with 'proper', 'regular' Celts, there were some distinguishing characteristics. What those characteristics were is not stated explicitly by those sources that use the label Celtici.

It is also important to note that some of the migrations of the Celtic peoples to this area must have been so close in time to our sources that they were known to Classical writers like Strabo. As de Hoz (1992: 9-10) has pointed out, this must mean that the peoples in the area were still aware of the differences between Celtic and non-Celtic peoples. Strabo (III 3.5) also reports that the Celtic people in this area were related to those living in southern Hispania, in the province of Baetica9.

Strabo text in Greek
Last of all come the Artabrians, who live in the neighbourhood of the cape called Nerium, which is the end of both the western and the northern side of Iberia. But the country round about the cape itself is inhabited by Celtic people, kinsmen of those of the Anas; for these people and the Turdulians made an expedition thither and then had a quarrel, it is said, after they had crossed the Limaeas River; and when, in addition to the quarrel, the Celtic people also suffered the loss of their chieftain, they scattered and stayed there. [Transl. By H. L. Jones, The Geography of Strabo, vol. II (Loeb Classical Library), London - Cambridge, Mass.]

The whole northwestern area of Hispania belonged in Roman times to the Tarraconensis province and, according to the information provided by Pliny (NH III 28), it was organized into three conuentus – the conuentus lucensis (with its capital at Lucus Augustus, modern Lugo), bracarensis or bracaraugustanus (with its capital at Bracara Augusta, modern Braga), and asturicensis (with its capital at Asturica Augusta, modern Astorga). Only the first two conuentus were assigned to the Callaeci and it is on these two that I will focus in this paper.

Since we do not have any inscription written in any indigenous language of this area10, in order to try and cast some light on the linguistic situation in ancient Callaecia our only resource is the meager information that we can recover by analyzing the indigenous, non-Latin names of the area. Those names fall into four categories:

  1. Personal names. These appear mainly in Latin inscriptions in this area11, but we also have a few occurrences in the Classical texts.
  2. God names. Our main source of information is the Latin inscriptions in the area12.
  3. Ethnonyms, which appear both in the Classical sources and in inscriptions13.
  4. Place-names, which appear mainly in the Classical sources (Strabo and, above all, Pliny and Ptolemy) but also on some inscriptions14.

I will first deal separately with the various categories of names15, which show different characteristics and, according to what we know about the behaviour of these various types of names from the point of view of general onomastics, may have a different depth in time and reflect various stages of the linguistic history of ancient Callaecia15.

2. Personal Names

As Albertos (1985) has noted, there is a clear-cut difference in the distribution of indigenous personal names in the two conuentus of the Callaeci - the number of indigenous personal names found in inscriptions from the conuentus lucensis is quite low when compared to those found in inscriptions from the bracarensis. New discoveries of inscriptions in this area have not changed this pattern of distribution significantly. In my survey of the indigenous personal names found in the inscriptions of Callaecia17, I have only found some seventy names in the inscriptions of the conuentus lucensis, while the number of indigenous personal names in the bracarensis is above three hundred.

Albertos (1985) also noted that the indigenous personal names from the conuentus lucensis showed a closer relationship to those appearing among its eastern neighbours, the Astures, while those of the bracarensis were more closely related to the names appearing farther south in Lusitania18. Excluding the names that occur only once, names attested in the conuentus lucensis tend to show up also among the Astures, while names attested in the conuentus bracarensis tend to occur also in Lusitania.

As is also the case with most areas of ancient Hispania, it is very difficult to characterize Callaecia as an onomastic region as far the personal names are concerned. Indigenous personal names from Hispania do not usually show clear patterns of distribution among regions, but seem to be spread throughout the Iberian Peninsula19, thus making it very difficult to isolate specific linguistic traits of the indigenous language or languages which must have been spoken in any given region. Nevertheless, in the case of Galician names a few interesting observations may be made.

I will first analyze some of the most commonly occurring personal names in Callaecia, which are as follows20:

  1. Arquius /Arc(u)ius (+10 occ.). The variant Arcuius occurs only in Callaecia, but Arquius seems to be characteristic both of Asturians and Galicians, as pointed out by Albertos (1985: 266-267)21. Albertos also remarked that this personal name was clearly connected to the Lugubo Arquienobo/Lucoubu[s] Arquieni[s] attested in inscriptions from the province of Lugo. The name may be explained as a derivative of IE *arkw- 'curve, arch' 22.

  2. Camalus /-a (+25 occ.). This is an interesting name, because, even if it appears in some inscriptions to the east of this area23, it is a typical Lusitanian-Galician name, as Untermann (1965: 85-86, map 20) and Albertos (1985: 275-276) have already demonstrated. Based on the concentration of this name in the Portuguese province of Minho, Albertos (1985: 275) assumed that it was characteristic of the bracarensis, subsequently spreading to the south and east. According to Palomar (1957: 58) this name etymologically derives from the IE root *kem- 'strain, fight', to which Pokorny (IEW 557) assigns also the Gaulish god name Camulus and all the Gaulish personal names in Camul- (see also KGP 160-161). After a critical revision of the bibliography, Evans (GPN 160-161) reached the conclusion that the name may be related to OIr. cam 'battle, encounter', making it a derivative of the IE root *kam-, while the relationship to the Irish name Cumal would be uncertain. The forms in Camal- instead of Camul- found in Hispania could have a more conservative vocalism, but this is not certain24.

  3. Caturo (+10 occ.). The distribution of this name is similar to the previous one, appearing both in Callaecia and in Lusitania25. It must be based on Celtic *catu- 'battle' (cf. OIr. cath, W cad, Corn. cas). Evans (1979: 122), however, expressed his doubts that this name might be a hypocoristic, or pet name form, of such names as Caturicus/ -a, Caturis, Caturisa, etc., since merely excising part of the second element would not be an acceptable form of hypocorism in ancient Celtic languages, so he suggested that the -r- could be suffixal.

  4. Cl(o)utius /Cloutaus and other names in the same series appear in the region (+5 occ.), such as Clutami (gen.) and Clutimoni (gen.) with -u- vowel and Clodamenis (gen.) with -o- (<-ou-). Based on its geographic distribution26, Albertos (1985: 279-280) considered the series typically Asturian, subsequently spreading to the Callaeci and the Vettones. These names are clearly based on the IE root *kleu-, which has been frequently used in Celtic onomastics27. Palomar (1957: 65) assumed a participial value for the -t-, which may be right in view of Skt. śrutá- (participe), Gk. κλυτός 'famous', Lat. inclutus, and the element (h)lot- /(h)loϷ- in Germanic names.

  5. Rebu(r)rus and Reburrinus/ -a (+15 occ.). These are very common Hispanic names. Albertos (1985: 293-294) thought that Reburrus was of Asturian origin, judging from its distribution and from the morphology, given that the suffix -u(r)rus also appears in a few ethnonyms of the Asturian area such as Gigurrus, Seurrus, Susarrus, and Tiburus. The name is not etymologically Celtic28.

I think that this analysis of the most frequently occurring indigenous names in the Latin inscriptions of Callaecia is significant enough to demonstrate the mixture of Celtic and non-Celtic names that is characteristic of the onomastics of the area.

Another problem in Callaecia is that most names only appear once in the inscriptions of the area, making it impossible to determine which ones are regionally characteristic. In addition to the names already mentioned, some exceptions include Medamus and Pintamus, which occur a few times. Medamus has the superlative (?) suffix -amo-, which Untermann (1965: 20) already characterized as typical of the Lusitanian-Gallician area. As for med-, it may be related to various Indo-European roots, but I think that Medamus, which does not have any correlate in Celtic languages, is more likely related to IE *medh-(yo-) 'middle', a development parallel to that of Gk. μέ(σ)σατος, a longer form with the suffix -(a)to- of μέ(σ)σος29. As for Pintamus, it shows the same suffix as Medamus. According to Villar (1994), we have here a non-Celtic formation on the numeral 'five', IE *penkwe, so in this case the suffix -amo- could be explained as due to the influence of the ordinal *septºmos>*septamos '7th'. Albertos (1985: 290) considered this name Galician, later spreading into Lusitania, but the data do not seem cogent enough to confirm this theory30.

Other exceptions specific to the conuentus lucensis are the following names: Aebura (twice + derivative Aeburina), Apanus /-a, Auitus /-ius, Buleni (gen., three times in the same inscription), Caeleo (twice in the same inscription), Nantius /-a or Pusinca (twice + derivative Pusincina). The masculine Eburus is attested several times in Lusitania and among the Astures31, Apanus /-a is well attested in Lusitania32, while Pusinca and related names are attested in various areas33.

Theoretically it could be argued that the names appearing only in Callaecia, even only once, are proper 'Callaecian' names, for otherwise they would also be found outside Callaecia. But this is a risky and tricky conclusion. For instance, in Vigo several inscriptions have been found (CIRG II 40 and 48) dedicated to individuals from the conuentus cluniensis, that is, from Celtiberian territory. There is also an inscription from Caldas de Reis [PO] (CIRG II 70) which does not specifically refer to the conuentus but mentions Celtiber, showing that this individual also came originally from Celtiberia. Since the mention of origin does not always appear in the naming formula of Latin inscriptions we can assume that some of the hapax names (occurring only once) found in the area belonged to individuals from other areas of Hispania.

Laucius /-a is attested several times, but seems to be unique to Aquae Flauiae. The Celticity of this name is not certain – there are some place-names in Lauco- in Celtic areas of western Europe (Ho. II 157; Morlet 1985: 116) and a-l-a-u-ka[ and a-l-a-u-ke in Iberian script could stand for *adlauca /-os34.

Nantius /-a is unique to Callaecia – this is a name with a possibly solid Celtic etymology, since it can be easily shown to be a derivative in -yos /-ya from Celtic *nantu- 'valley' (> IE *nm˳tu- ), cf. Welsh nant, Corn. nans35. Caeleo and Buleni are unique to Callaecia as well.

From a linguistic point of view it should be remarked that, as opposed to what happens with other kinds of names, especially god names, as we will see in a moment, the personal names of Callaecia do not show strikingly special phonetic or morphological traits when compared to the names appearing in other areas of ancient Hispania. In other words, when we have the same personal name both in Callaecia and outside, the one appearing in Callaecia shows no evidence of being adapted to the particular phonetics of a language spoken in that area. Some remarkable phonetic features of the indigenous personal names in the Latin inscriptions of Callaecia are the following:

  1. diphthong ai- in Ailaeca (Vilar de Areias, Barcelos, AE 1983.568), which may actually be a god name.

  2. diphthong -ei- in a few names such as gen. Cileioui (RAP 323)36, gen. Peicanae (EE IX 264), Seicuius (CIRG II 136)37. Occurrences with comparanda outside Callaecia are very scarce, e.g.:
    • Malceinus, which, according to its distribution38, is a typical Lusitanian name, usually appears in Lusitania as Malgeinus, but also as Malcenus and Malgenus;
    • Meiduena, which is also a typically Lusitanian variant of Medugenus /-a39.

  3. diphthong -oi-: gen. Cadroiolonis (ILER 6330), ? Doirau (ILER 4643), Goilius (AF I2 155).

  4. diphthong -oe-: Bloena ( e.g., EE VIII 119), Boelius (CIL II 2530), ? gen. Boesii (AF I2 343, Coemia (CIRG II 14), Laboena (AF I2 272).

3. Names of Deities

The first observation that must be made is that the god names of Callaecia are closely connected to those of the Lusitania province south of this region. It has long been known that the god names appearing both north and south of the Duero River in western Hispania clearly belong together, so this corpus of god names is frequently labelled 'Lusitanian-Galician'40. This relationship is twofold:

  1. The major divinities (judging from the number of inscriptions) are the same in Callaecia and Lusitania.

  2. The phonetic traits of these names are basically the same in Callaecia and in Lusitania, with only some minor exceptions that will be commented on below. See e.g. the dative endings in the seq uence Tongoe Nabiagoi of the inscription from Braga (CIL II 2419 = HEp.7.1160)41.

The most frequent god names in the inscriptions of Callaecia are:

  1. Bandue. This is one of the most frequent indigenous gods in the Latin inscriptions of western Hispania. Pedrero (1999) showed that there is a clear-cut distribution of the forms of this god name: Bandue appears in the inscriptions of Callaecia (plus a pair of inscriptions in the western part of the conuentus asturicensis), while Bandi appears in the inscriptions of the province of Lusitania42. Assuming that we are dealing with a u- stem, that difference can be accounted for as the divergent final outcome of an original sequence *Banduei, which would have evolved into Bandue in the northern forms and into Bandei/Bande/Bandi, with coalescence of the -w- in the cluster -dw-, in the forms south of the Duero. This evolution, as noted by Prósper (2002: 269), is paralleled by the forms of another important god name, Cossue /Cosei.

  2. Nabiae /Nauiae. This goddess appears very frequently in Callaecia and in Lusitania. Her name has been interpreted by Prósper (2002: 189-195) as *nāwyā 'valley', cf. Spanish nava 'plain between mountains'.

  3. Coso /Cossue. In this case there is an interesting distribution of forms of the name of the god, as noted by Prósper (2002: 235-238). Coso appears in the conuentus lucensis, Cossue in the bracarensis and Cossei further south in Lusitania. Coso has been generally explained as a Latinized form with the Latin thematic dative, but Prósper has argued that this may reflect the evolution of an original adesinential dative -ou, while the forms Cossue and Cossei appear to have originated from an original desinence in -owei.

Among the god names appearing exclusively in Callaecia we have:

  1. Bormanico, appearing on two inscriptions from Caldas de Vizela, Guimarâes (CIL II 2402 and CIL II 2403 = 5558). It is most probably a formation built on the IE root *ghwer- (cf. Skt. gharma-, Lat. formus, Gk. θερμός)43. This fits in particulary well in this case given that the two inscriptions come from a place with thermal waters.

  2. Reue Larauc(o), Larauco D(eo) Max(imo), Larocuo44. The name lacks a clear etymology.

  3. Diis Ceceaigis (Zaparín, Cortegada OR) and Laribus Inmucenbaecis Ceceaecis (Samaioes, Chaves VRE), L(aribus) Gegeiqis (AF I2 124). Prósper (2002: 319) does not offer an etymology.

  4. Dedications to the Lares followed by indigenous epithets are very frequent in western Hispania. Among these, there is a homogeneous and interesting group coming from the Peninsula do Morrazo (Cangas PO). They are all devoted to a Deo Lari Berobreo45, whose names must be explained as related to a place-name in -bri- or -brig-, as argued by Villar and Pedrero (2001: 693) and Prósper (2002: 367-369).

  5. Among the epithets only found in Callaecia we also have the dedication to a Lari Cari[e]co (Refoios de Lima, Ponte de Lima VCA, AE 1983.561) and to a Mar(ti) Caria(eco) (Lisouros, Paredes de Coura VCA, CIL II 5069 = HEp. 4.1088).

The divinities mentioned so far do not seem to be Celtic, but appear to be related to Lusitanian instead. In her comprehensive study of the indigenous gods of western Hispania, Prósper (2002: 511 map 12), however, lists three Celtic cults in Callaecia:

  1. The dedications to the great Celtic god Lug, but in plural46. We find a dedication to the Lugubo Arquienob(o) in Sober LU (IRPL 67) and, similarly, to the Lucoubu Arquieni(bu?) in Otero de Rey LU (HAE 1717).

  2. Suleis Natugaicis Coucieiro, Padern de Allariz OR (AF I2 158 = HEp.7.532). Prósper (2002: 311-312) has rightly explained the name as the Latinized dative plural of *Sūlew(i)yā, that is, a variation of the Celtic god name Suleui(i)s, which is well attested, especially at Bath. As for Nantugaicis, it is a derivative in -k-aiko- from Celtic *nantu- 'valley'.

  3. Ariounis Mincosegaeigis. Prósper (2002: 205-215) explains the god name as a participle ar-yo-uno 'ploughed' < *ar-yo-mno- < *H2erH3-yo-mno-, while the epithet would be a derivative in -aiko- from a compound *menekko-sego-. *menekko- 'abundant, frequent' appears to be related to OIr. menic, MW mynych, MCorn. menough, while *-sego- is the well known Celtic element.

I think that at least the following Celtic god names in Callaecia should be added47:

  1. Deuori (dat.) appears as an epipthet of the god Hermes on an inscription from Outeiro Seco, Chaves VRE (CIL II 2473, AF I2 78). It is clearly a compound *deiwo- > deuo- plus -ri(ks), with the same evolution in the ending as the place-names in -bris from *-briks48.

  2. Deab(us) Vseis on an inscription from Atás, Cualedro OR (AF I2 155 = HEp.7.498). They seem to be related to the Matribus Vseis from Laguardia AL (ILER p. 692)49, derived from *upso- 'high'.

  3. There are three inscriptions devoted to Crougia* in the bracarensis or nearby: Crougiai Toudadigoe (Mosteiro de Ribeira, Ginzo de Limia OR, CIL II 2565 = AF I2 98), Corougia Vesuco (Barcelos, Braga, RAP 61), and [C]rugia Munniaego (Viana do Bolo OR, CIL II 2523, with a correction of the reading by Prósper 2002: 183-184)50. The name can be explained as a derivative *krouk-yo- and is thus related to MIr. cruách, W crug, Corn., Bret. cruc.

From a linguistic point of view it is also important to remark that, apart from the distributional differences that we saw above concerning the dative endings of the god names Bandue and Cossue, the same phonetics can be found in the god names of Callaecia as in the god names of Lusitania. This applies even when the god names seem to have a Celtic etymology.

4. Ethnonyms

We should make a distinction between proper ethnonyms, that is, names primarily used to refer to a people, and ethnonyms derived from a place-name. An example of the first type is the name of the Bibali, whose town is, in fact, named after the name of the people, Φόρος Βιβαλω̑υ, according to Ptolemy (II 6.43). An example of the second type is e.g. the name (gen. plu.) [T]ongobricensium, attested in an inscription from Freixo, Marco de Canaveses POR (CIL II 5564), which is clearly derived from the place-name Tongobriga. In this section I will focus on the ethnonyms of the first type, while the second type will be dealt with together with the place-names in the following section.

There are some peoples with clearly Celtic names51, including the following:

  1. Albiones (Pliny NH IV 111, ERA 14)52. Tovar (1989: 124 and 139) interpreted this as a Celtic name, which is possible, but the name shows no specifically Celtic trait. It is a derivative form of IE *albho- 'white'.

  2. A(r)rotrebae (Str. III 3.5, Pliny NH IV 111, 114). They are also frequently called Artabri (Str. III 2.9 etc., Mela III 13, Agathem. IV 16, Ptol. II 6.2, 21). If the correct form is Arotrebae, as Pliny NH IV 114 explicitly argues, we would have here a compound of are-53 (with loss of the initial IE *p) plus a form of the stem *treb- 'live in, inhabit'54.

  3. Nemetates: Gen. Plu. Νεμετατω̑υ in Ptolemy (II 6.40). The name is no doubt Celtic - it is a derivative in -ates of the well-known Celtic word nemeto- 'sacred place, sanctuary' (cf. Gaul. υεμητου, OIr. nemed). According to Ptolemy their capital town is Volobriga, again a name that is clearly Celtic, belonging to the series of place-names in -briga. In the onomastics of the area the word nemeto- is also attested in the name Nemetobriga, the capital of the Asturian Tiburi according to Ptolemy (II 6.37), but we have now an inscription from Codesedo, Sarreaus OR (HEp.7.548) mentioning possibly another Nemetobriga.

  4. Neri (Mela III 11, Pliny NH 4.111)55. As we saw, this group of people is explicitly called Celtic in the Classical sources. The name can be easily interpreted as Celtic56, from the word *ner 'male' (cf. W ner 'lord', OIr. ner 'boar'), but it does not show any specific Celtic trait, given that the word is attested in other western Indo-European languages57.

  5. Quarquerni /Querquerni (Pliny NH III 28, Ptol. II 6.46, CIL II 2477 = 5616; Querquennis, It.Ant.428.2; Cercennis, Rav.320.3). Related to the IE name for oak, *perkwos. Given that the assimilation of the initial *p- to the labiovelar of the second syllable only seems to occur in Italic – including Latin – and Celtic languages, we can guess that we have a Celtic form here58.

  6. Treb(ilium) (AF I 2 618). Even if the name is abbreviated, it seems acceptable to consider it as based on Celtic *treb-, as the second element in Arotrebae above.

In other cases there is no cogent reason to classify a given ethnonym as Celtic, but the possibility cannot be totally ruled out either. For instance59:

  1. Baedui (Ptol. II 6.26). García Alonso (2003: 251) suggests that, assuming that Baed- stands for [bed-], it might belong to the same root as Boduo- with e-grade60.

  2. Callaeci (Pliny NH III 2861). The etymology is not certain. It might be related to the first element Cala- in Caladunum62.

  3. Cileni /Cilini (Pliny NH IV 111, Ptol. II 6.24, It.Ant. 423.7, 430.3, Rav. 308.2, 321.8)63 with zero-grade.

  4. Coelerni (Pliny NH III 28, Ptol. II 6.4, CIL II 2477 = 5616, IRG II 40).

  5. Corocauci (CIL II 2462). The name seems to be a compound of Coro-64 plus cauco-, which appears in other Hispanic names but lacks a certain etymology65.

  6. Coroqui (CIL II 2489 = AF I2 612-614). The initial Cor- must be related to the initial sequence of the previous name.

  7. Egibarri Namarini (Pliny NH IV 111). Egibarri is corrected to Egobarri by Tovar (1989: 138) and is then explained as a compound of a river name *Ego- (> modern Eo) plus -barros66. Namarini is explained by him as showing the IE negative prefix *n˳->na- plus mari- related to mori- 'sea'67.

  8. Lemaui (Pliny NH III 28, Ptol. II 6.25, CIL II 2103, CIL XVI 73, 157, 161). García Alonso (2003: 203-204) suggests a derivation from IE *l˳m-, cf. Lat. ulmus, OIr. lem, W llwyf, Engl. elm. The name is thus not incompatible with a Celtic etymology but shows no specifically Celtic trait, either.

  9. Narbasi: Gen. plu. Ναρβασω̑υ (Ptol. II 6.48). There is no clear Celtic etymology. The ethnonym, however, is reminiscent of names in other Celtic areas, especially Narbo in Gaul.

  10. Lubaeni (Ptol. II 6.47). From *lubo-, cf. OIr. lub 'herb'?

There is a small group of Indo-European names that cannot be Celtic:

  1. Copori (Pliny NH IV 111; Coporici, AE 1983.562 Montariol, Braga BRA; Capari Ptol. II 6.23)69.

  2. Equaesi (Pliny NH III 28, CIL II 2477 = 5616). A derivation from the IE name for 'horse' *ekwos.

  3. Sefes (Auien. Ora 195; Lari Sefio, RAP 216). The name has been interpreted70 as a derivative of the IE *s(w)e-bho-. The evolution *bh>f would clearly prove that this is not a Celtic formation, while that evolution is attested in Lusitanian71.

Tovar (1989: 124) drew attention to the fact that a group of ethnonyms with a pre-Indo-European etymology are related to river names, for instance:

  1. Nebisoci. According to Tovar (1989: 127) this name must be related to the river name Nebis, mentioned in various sources72. It would be a derivative in -oco-, which, as García Alonso (2003: 132) remarks, has the problem that the -s-, preserved in the ethnonym, is difficult to account for. I think, however, that that is not an unsolvable problem, for, if we assume that the indigenous name was an -s stem (a hypothetically nom. *Nebis), gen. *Nebisos, the Latin and Greek adaptation, starting from the nominative, would have been an -i stem, but the -s would be preserved in the derivations in the local language73.

  2. Bibali (Ptol. II 6.42, Pliny NH III 28, It.Ant. 428.7, CIL II 2477 = 5626)74. This name seems to show some kind of reduplication in the initial syllable, as also shown by other ethnonyms in this area, such as Gigurri or Susarri75. It has traditionally been related to the modern river name Bubal76, which is rather doubtful.

  3. Tamagani (CIL II 2477 = 5626, IRG IV 66). Related to the modern river name Támega.77 The name could thus belong to the series of place-names in Tam- surveyed by Villar (1995b).

  4. Limici (Pliny NH III 28, Ptol. II 6.43, CILA I 24, 33, 64, 72). Their name is related to the river name Limia.

Other ethnonyms have even more difficult or obscure etymologies78: Adoui Iadoui (Pliny NH IV 111), Ancondei (CIL II 4215), Arroni (Pliny NH IV 111), Gigurri (Pliny NH III 28, Ptol. Geog. II 6.38, It.Ant. 428.7; foro Gigurnion, Rav.320.8, CIL II 2610, HEp.7.378 = 8.325), Groui (Mela III 10, Pliny NH IV 112, Ptol. II 6.44)79, Grauii (Silius I 235; Graii in Silius III 366), Leuni ( Pliny NH IV 112), Luanci (Ptol. II 6.45), Praen(i) (CIL II 2489 = AF I 2 612-614 ), Seu(r)ri (Ptol. II 6.27, AEA 39.1966.142, CIL II 6290, AE 1934.19; Seurbi Pliny NH IV 112)80.

5. Place Names

I will exclude from this analysis all the Latin place-names, since they cannot contribute to our knowledge of the indigenous languages of the area. I will also exclude such place-names as Φόρος Βιβαλω̑υ, given that the ethnonyms have already been dealt with in the previous section.

5.1 Town (or human settlement) names

Among the names with a possible Celtic etymology we can mention the following:

  1. Asseconia (It.Ant. 430.5; Assegonion, Rav. 321.6). The name may be Celtic if we accept that it is based on a compound *ad-sec- or *ad-seg-81.

  2. Bracara (Pliny NH III 18, 28, IV 112, Ptol. II 6.38)82. The name may be Celtic if it is derived from the Celtic word bracā 'trousers' or from *mraci (OIr. braich /mraich 'malt', W brag 'malt', etc.)83.

  3. Caladunum (Ptol. II 6.38, It.Ant. 422.5; Caldu<n>a, CIL II 2487 = AF I2 372)84. The second element of the compound is clearly Celtic -dunom, but the etymology of the first element, even if it has good parallels in other Hispanic place-names, is problematic85.

  4. Complutica (Ptol. II 6.38; Compleutica, It.Ant. 423.1; Com<pleu>tica, fourth Astorga tablet). The name seems to be a derivative of IE *ploutos 'swift' (cf. OIr. lúath, Gk. πλέω, etc.). The variation -eu- /-u- shown by the sources can be accounted for if we assume a development Compleutica>*Comploutica> Complūtica86, with an evolution -eu->-ou- known both in Celtic and in Latin87.

  5. Ebora (Mela III 11). Possibly from the Celtic word *eburo- 'yew' (cf. OIr. ibar 'yew', W efwr, ewr 'cow-parsnip, hogweed')88.

  6. Lamecensis89. A place-name *Lamecum, the forerunner of the modern Lamego, must be assumed on the basis of the ethnonym. It would be thus a derivative from lama-, which occurs frequently in Hispanic onomastics, and has been explained by García Alonso (2003: 126) as related to OIr. lám 'hand', from IE *pl˳ma or *plāma, with Celtic loss of initial *p-.

  7. Lugisonis (Rav.321.1). This could be a derivative of the god name Lugu-.

  8. Meidunium (CIL II 2520). This is the name of a castellum of the Limici. It seems to be a compound with -dunium, a derivative in -yo- from the Celtic word -dunom (cf. OIr. dún 'stronghold', W din 'stronghold'), which occurs frequently in Celtic areas of Western Europe (Ho. I 1375-1377), although it is rare in Hispania.

  9. Morodon (Rav.308.1). If the reading is right90, this could be a compound of *moro- (variation of *mori- 'sea'?) plus *-dunom>-don, and therefore Celtic.

  10. Nouion (Ptol. II 6.21; Noeta Noela, Pliny NH IV 111)91. Since Holder (Ho. II 792) this is usually considered the neuter form of Celtic *nowios 'new' (a derivative in -yos from IE *newos), showing thus the Celtic evolution eu>ou. Nouio- is frequent in the toponymy of Celtic areas of Western Europe92.

  11. Ocelum (Ptol. II 6.22). Ocelo- has no counterparts in insular Celtic, but its Celticity is usually assumed on the basis of its distribution in ancient onomastics93.

  12. Olca (CIRG II 84). This is the name of a castellum. It seems to be the Celtic word olca 'field'94.

  13. Olina (Ptol. II 6.22). Probably the Celtic word *olīnā 'elbow' (cf. OIr. uilen, W elin, Mid.Corn. elin, etc.), in a metaphoric sense95.

A special mention must be made of the names in -briga and -bri(s), which are quite frequent in the area96. There is only one non-compound name built on briga, Brigantium (Ptol. II 6.4, It.Ant. 424.5; Bricantia, Rav. 308.5; Cabricantium [for -ca Bricantium97], Rav. 307.13; Bregantium Aethicus Isther, Geographi Latini Minores p.79.52 Riese; Brigantia Orosius I 2.71)98. As for the compound names, I will provide a classification on the criterion whether the first term of the compound also has a clearly Celtic etymology or not99:

  1. Names in -briga /-bri(s) with a Celtic first element:

    1. Laniobrensis /Lamniobrensis/Lamiobrensis (Conc.Tol. 137, 432, 451). The first element of the compound can be easily explained as Celtic *lanio-, a derivative in -yos from the IE *pl˳H2nos 'flat'. It is highly remarkable that two phases in the evolution of this word can be attested, for abl. Blaniobriensi appears as the name of a castellum of the Supertamarici on the inscription CIL II 2902 = 5667, with *p->b- prior to its loss100.

    2. Lambris (Ptol. II 6.26; Lambriaca Mela III 10). If, based on the explanation by García Alonso (2003: 206-207), we assume that Lam- is an apocope for *lama101, i.e. the final -a has been lost at the end of it, the name may be Celtic.

    3. Nemetobriga (It.Ant. 428.6, Rav. 320.7). Another town with an identical name appears on an inscription from Codesedo, Sarreaus OR (HEp. 7.548). The first element is clearly the Celtic word *nemeto- 'sacred place, sanctuary'.

  2. Other names in -briga /-bri(s):

    1. Adrobriga (Mela III 13). No clear etymology for the first element102.

    2. Aiiobrigiaecini103, a compound possibly with the personal name Aius as the first element, but we cannot be sure that this place was in Galician territory.

    3. Alansbrica (AF I2 105). The reading is uncertain and Alaniobriga has also been proposed.

    4. Aliobrio. Name of a mint of the Visigothic king Suintila104. The first element is clearly related to IE *alyo- 'other' and is widely used in Hispanic personal names, but *Ailo- would be expected if it were Celtic.

    5. Arcobriga (CIL II 2419). There are two towns with this name in Spain and we do not know for sure whether this is a third one in Callaecia. The first element must be related to the personal name Arcus, which does not seem Celtic105.

    6. Auiliobris (CIRG I 66). The first element is attested in Hispanic personal names106.

    7. Auobrigensis (CIL II 4247)107. The first element may be the personal name Auus.

    8. Calubrigensis (CIL II 2610, AE 1981.526 ). For the first element see the remarks on Caladunum above.

    9. Canobri (Rav. 308.12). This could be Celtic if cano- is related to MW cawn 'reeds'108.

    10. Coeliobriga (Ptol. II 6.41)109. The etymology of the first element is problematic110.

    11. Ercoriobri (abl.) (CIL II 2711). This must be a town of the Albiones. The analysis of the first element is difficult. If we isolate the element corio-, which is well known in Celtic onomastics111, the initial er- cannot be accounted for. If we assume a relationship to place-names such as Ercauica, then we would be left with a suffix -orio-, which is also rather odd. Prósper (2002: 381) has suggested a relationship to OIr. erchor 'cast, shot'.

    12. Iuliobriga ( CIL II 2480). Rodríguez Colmenero (AF I2 223) reads this as <T>ureobriga. In that case the first element would be Indo-European non-Celtic rather than Latin, from the series of place-names in Tur-112.

    13. Meobrigo (CIRG I 86). Possibly an epithet of the god Coso, as suggested by Prósper (2002: 226). The etymology of the first element, meo-, is uncertain.

    14. Talabriga (AE 1952.65, CILA I 24, 33, 40, 42? ). The first element must be related to the series of place-names in Tala- surveyed by Villar (1993).

    15. Tameobrigo (CIL II 2377). This is the name of a god, clearly derived from a place-name in -brig-. The first element must be related to the series of place-names in Tam- surveyed by Villar (1995b).

    16. [T]ongobricensium (gen. plu.) (CIL II 5564). If the usual reconstruction of the initial letter of the name is right, the first element would fit well into a series of names in Tongo- found in Callaecia and Lusitania, including the god name Tongoe Nabiagoi (dat.) in Bracara Augusta (CIL II 2419) and personal names such as Tongatius, Tongetamus, etc.113 This element seems to be Indo-European, but is not necessarily Celtic114.

    17. Tuntobriga (Ptol. II 6.38). The first element lacks any satisfactory explanation. It has thus been suggested115 that it should be corrected to read Tungobriga, in which case the town would be the same one as in the previous entry.

    18. Volobriga (Ptol. II 6.40; Valabricensis, CIL II 5561; Valubricensis, ILER 5439)116. There are various possibilities for the etymology of the first element as an Indo-European element, but none of these is certain117.

It thus appears that in Callaecia the place-names in -briga /-bris with a non-Celtic first element are more numerous than those with a Celtic first element. I suggest that this allows us to draw interesting conclusions about the linguistic history of the region. Speakers of Celtic languages must have coined the names in -briga or -bris, but in order to generate them they seem to have used previously existing onomastic elements from that area.

The following place-names seem to be Indo-European but are not Celtic:

  1. Albucrarensis (Pliny NH XXXIII 80; Alboc(rarensis) CIL 2598). This seems to be a compound with IE *albho- 'white' as its first element, but -crar- cannot be accounted for.

  2. Auren(ses) (CIL II 5613)119. This is clearly the ancient form of the modern toponym Orense. Tovar (1989: 307) suggests the possibility of a derivation from aurum – in which case it would be Latin.

  3. Buron (Ptol. II 6.22). This may derive from the root *wer- /ur- 'water, river', with b- for w-119.

  4. Cambaetum (Ptol. II 6.47). From the root *(s)kamb- plus a suffix -aito- attested in some personal names like Andaitia or Calaitus>120

  5. Cariaca, a place of the Albiones (ERA 14). This is a name derived from the root *car-, with various possible etymologies121.

  6. Caronium (Ptol. II 6.22; Caranicum, It.Ant. 424.6; ? Carantium, Rav. 307.15). This name is derived from the same root as the previous example.

  7. Glandomiron (Ptol. II 6.22; Glandimiro, It.Ant. 424.3; Glandimarium, Rav. 308.3). This looks like a compound, the second element being -miro-, which is frequent in ancient Hispanic place-names122. As for the initial element, it would be tempting to relate it to the Celtic names in Gla(n)n-, but that would leave the cluster -nd- unexplained123.

  8. Limia (It.Ant. 429.5). See the river name Limaia below (§5.2).

  9. Medioca (Rav. 308.4). This seems to be a derivative in -ocā from IE *medhyos 'middle'.

  10. Merua (Ptol. II 6.45). This may be derived from IE *mer 'shine'124.

  11. Salacia (It.Ant. 422.3). This must belong to the series of Old European names in sal-125.

  12. Salaniana (It.Ant. 427.6, Rav. 320.1). This may be from the same stem as the previous example. One manuscript of the Itinerarium reads Silaniana, in which case the name would be Latin126.

  13. Samarium (Rav. 307.14). This must belong to the series of Old European names in sam-.

  14. Saramon (Rav. 308.10). This must belong to the series of Old European names in sar-.

  15. Serante (AE 1934.19). This is the name of a castellum. It might belong to the same root as the previous one but with e-grade.

  16. Talamina (Ptol. II 6.27; Timalino, It.Ant. 430.9). This must belong to the series of names in Tala- analyzed by Villar (1993)127.

  17. Turriga (Ptol. II 6.22). This could be explained as a derivative in -ica belonging to the series of place-names in tur-, on which see Villar (1995a: 199-244), who explains them as built on the IE stem *tur- (from *teuH2- 'swell, be strong', with -r-, cf. *turó- 'strong'). It may even be the same town as the following one128.

  18. Turoqua /Turaqua (It.Ant. 430.2, Rav. 307.19). This has been interpreted by Villar (1995a: 191-197) as a compound of the stem tur- (see previous example) plus one of the IE words for 'water', *akwā. The variation a- /o- is, according to him, evidence of its belonging to the Old European substrate.
  19. Turonion (Hidatius 131, Chron. Min. II p. 24, Vita Fructuosi 20). This must be a derivative from the same stem as the previous example.

  20. Turuptiana (Ptol. II 6.22). Perhaps another derivative of tur-, but the sequ ence -uptiana cannot be accounted for129.

  21. Vacoecum130.

Other place-names in Callaecia with even more uncertain etymologies include the following131: Acripia (CIL II 2435)132, Araducca/Araducta (Ptol. II 5.6, 6.38)133, Arragina (Rav. 308.9)134, Atricondo (It.Ant. 424.4), Atucause(nses) (CIL II 6287), Berensi (abl., CILA I 49), Bonisana (Rav. 307.18), Burbida (It.Ant. 430.1)135, Cistonia (Rav. 308.7), Dactonium (Ptol. II 6.25)136, Dumium137, Fi[?] (AE 1977.451)138, Iria (Ptol. II 6.23, It.Ant. 430.4)139, Laia (Ptol. II 6.39; Lais Hidatius Chron. 252-253)140, Libunca (Ptol. II 6.22)141, Liuai (CIL II 2496 = AF I2 375)142, Madequis(enses) (AE 1977.451), Odeis (Rav. 307.11), Ontonia (Rav. 308.6), Saqua (CIL II 2487 = AF I2 372)143, Sermacele(n)s(is?) (CIL II 2494 = AF I2 623), Tardu (CIL II 2484 = AF I2 237)144, Touagornicenses (AF I2 5 = HEp. 7.1252)145, Tude (Str.III 4.3, Pliny NH IV 112, Ptol. II 6.44, Sil.Pun. III 367, XVI 368, It.Ant. 429.7, Rav. 307.17)146, Vliainca (AE 1977.451)147, Uttaris (It. Ant. 425.3, 430.11)148.

5.2 Rivers

  1. Limaia. The modern Limia appears as the River of Forgetting in the Classical sources, specifically the Lethes and Obliuio149. The etymology of Limaia is uncertain, but it could be Celtic if derived from *l˳m-150.

  2. Tamaris (Mela III 11) /Tamara (Ptol. II 6.2), modern Tambre. The river name also appears in the ethnonyms Supertamarici (Mela III 11, Pliny NH IV 111, CIL II 2902 = 5667 Astorga, CIL II 2904 = 5081 Astorga) and Praestamar(i)ci (Mela III 11, Pliny NH IV 111)151. The name Tamaris seems to belong to the series of Indo-European place-names in tam- surveyed by Villar (1995b).

  3. Minius is attested several times in the sources152. It also appears in the ethnonym Transminiensis (AE 39.1966.142, CIL II 6290). The name may be Celtic, but this is not certain153.

6. Conclusions

In addition to the ethnological information provided by the Classical sources, the linguistic analysis of the onomastic records of ancient Callaecia clearly confirms the presence of Celtic populations in this area of Hispania by the early first century AD.

For the linguistic and cultural history of this region it is important to remark that the linguistic characteristics of its god names, which for the most part clearly point to a link with the southern regions in the Lusitania province, do not reappear in personal names. If we have a look at the personal names of Lusitania we will see that at least some of the characteristics of the god names, for instance, the presence of 'strange' diphthongs and sequences of vowels reappear in personal names. I will not go now in detail into the debated problem whether a Celtic-like phonetic 'infection'154 is attested in Lusitania, but the fact is that while we frequently have variants of personal names with that infection or infection-like sequences in the onomastics of the Vettones155 and other peoples of Lusitania, there are virtually no corresponding examples among the peoples of Callaecia.

We do have some unequivocally Celtic place-names, as has been demonstrated here. It is striking, however, how many names in -briga- /-bris show a non-Celtic initial element, indicating that at their arrival the Celtic populations used non-Celtic onomastic elements already existing in the area to create the new names of these settlements.

We also have to bear in mind that in Roman times people from the Celtiberian area were continually moving into Callaecia. As stated above, we have inscriptions that clearly show this on-going migration. We can expect that these people were responsible for the introduction in Callaecia of some Celtic personal names, and possibly also some place-names. We know indeed that place-names are rather conservative, but we have to take into account that new towns were created in this area in Roman times in order to subdue the indigenous peoples. One example among others is the above-mentioned Forus Bibalorum. This may also have been the case, for instance, for some of the names in -briga, which appear both in Callaecia and in the Celtiberian area.

As for the relationship between these Galician Celtici and the inhabitants of the Anas (= Guadiana) River area, who are explicitly mentioned in Strabo's paragraph, quoted at the beginning of this paper, I do not think that it is by chance that in Callaecia we have at least one inscription156 to Reue Ana Baraego, the god of the Guadiana and Albárregas Rivers, the latter a tributary of the Guadiana near Márida. These gods also appear as Ana and Barraeca in a monumental inscription on the mausoleum of a seuir augustalis, the so-called "dintel de los ríos" at Márida (Emerita Augusta, the ancient capital town of the Roman province of Lusitania)157. The iconography of this monument, in which Ana appears as an old man and Barraeca as a young man, clearly points to an interpretation of the Guadiana and its deity as the old, bigger river with the Albárregas and its deity as the young tributary. So while these Celtici may have introduced some personal Celtic place-names in the area, they may also be responsible for the introduction of non-Celtic names, such as that of Reue. The Celtici, coming to the north through Lusitania, may have brought with them some Lusitanian onomastics, too158.

Summing up, the nature of the evidence makes it difficult to provide a detailed linguistic history of ancient Callaecia. However, by analyzing the various sources available, a global picture seems to emerge in which the Celts constituted the last layer of Indo-Europeans to come to this area. But what was there before? Obviously there were populations related to the Lusitanians. But only that? More investigation is needed before we can safely state that Lusitanian or Lusitanian-like populations are not the oldest Indo-European layer in this area. Leaving aside theonymy, which can be easily borrowed, Callaecia does not seem to have so much in common with Lusitania as is usually assumed. A very tentative explanation would be that a previous Indo-European layer was later influenced by the Lusitanian populations in the southern region and that that mixture was what the Celts found when they finally reached this westernmost region of Europe.


Endnotes

1 This paper is part of the research project BFF2003-09872-C02 financed by the Spanish Ministry of Science and Technology.

2 Mela (III 10) reports that the Celtici were settled along the whole coast of Callaecia from the Duero River northwards, except the part inhabited by the Groui.

3 Mela III 13.

4 Pliny, NH IV 111. Cape Nerion is called promontorium Celticum by Mela (III 9 and 12) and Pliny (NH IV 111).

5 Pliny, NH IV 111.

6 Pliny, NH IV 111.

7 Pliny, NH IV 111.

8 In other cases we are told that a particular group of people was not Celtic. This seems to be the case with the Groui according to Mela (III 10), who makes a distinction between Celtici and Groui on the coast of Callaecia.

9 On the personal names of the Celtici in the Baetica see Luján (2001). For a comprehensive archaeological study of the Celts in that area see Berrocal Rangel (1992). According to Pliny (NH III 13-14) the Celtici in Baetica originated from the Celtiberians, coming through Lusitania.

10 There might be one indigenous inscription from this area – that found on the gold ring from Ginzo de Limia (Orense). The inscription is lost and we only know of it through a drawing by Martín Sarmento. Hübner (CIL II p. 1025 XII/III) transcribed it in Greek letters and published it in his Monumenta Linguae Ibericae (p. 183, n. LV). Despite recent attempts (Schwerteck 1993, 1998) to read it as an inscription in western Iberian script, its interpretation is not certain at all and its phonetic transcription is totally dubious; see my critical remarks in HEp. 8.381. In CIL II 5600 two small disks from Briteiros (Guimarâes BRA) were published with symbols that have sometimes been thought to be Iberian, although this does not seem to be the case. Those inscriptions do not provide any linguistic information in any case.

11 There has been no recent collection of the indigenous personal names of Callaecia after Albertos' (1985) paper. However, Albertos' list can be updated with the names listed in Abascal's (1994) general survey of the personal names of Hispania and the names listed in the volumes of the journal Hispania Epigraphica.

12 The god names of Callaecia, together with those of the province of Lusitania, have been systematically surveyed by Prósper (2002). Previously, they had been studied by Untermann (1985). Olivares Pedreño (2002: 67-109) also provides an analysis of the gods attested in Callaecia. Again, new god names and corrections of previous readings can be found in the volumes of Hispania Epigraphica.

13 For a general survey of the ethnonyms of Hispania see Untermann (1992). Most ethnonyms and place-names of Callaecia can be found in Tovar's (1989: 127-141 and 293-322) book and in the TIR K-29. Luján (2000) has studied the names of towns in this area cited in Ptolemy. For a comprehensive survey of the Hispanic place-names transmitted by Ptolemy see García Alonso (2003), especially pp. 129-155, 187-210, and 232-255 for the names of towns in Callaecia. Recent volumes of Hispania Epigraphica should be checked for new finds and corrections of names appearing in inscriptions.

14 See previous note for the most relevant secondary literature on the place-names of Callaecia.

15 I do not intend to be comprehensive and provide in this paper every linguistic record available for the study of the linguistic situation of ancient Callaecia, but rather have selected and commented on some significant records that may serve to provide insight into the general picture of this area.

16 For previous general surveys about the indigenous languages of Callaecia see de Hoz (1997) and Gorrochategui (1997).

17 In order to write this paper I have checked the published Latin inscriptions from Callaecia in search of indigenous names of any kind. This has been made possible by the use of the files of the Archivo Epigráfico de Hispania of the Universidad Complutense de Madrid. I am very grateful to its director, Prof. Isabel Velázquez, as well as to Dr. Rosario Hernando and José L. Gamallo for their help.

18 The personal names of Lusitania are now fully and easily accessible in the Atlas antroponímico de Lusitania prepared by the Grupo Mérida (2003). This volume allows for quick and systematic comparisons of the personal names in the Gallaecia with those in Lusitania. For the indigenous personal names of the Lusitania we now can rely on Vallejo's (2004) monograph.

19 This has been variously explained. De Hoz (e.g., 1997: 109-111) favors the view that the onomastic homogeneity between the different areas and peoples of ancient Hispania is the result of a long process of migrations and intermingling, while since his early works Untermann (e.g., 1965) has advocated for an original unity.

20 Untermann (1965: 20) listed the following names as "most typical" of the Lusitanian-Galician area: Albinus, Balaesus, Caturo, Cilius, Louesius, Medamus, Tritius, Viriatus, and possibly also Vegetus.

21 See Abascal (1994: 284 and 286) and Vallejo (2004: 180-185) for the occurrences of the names. Vallejo (2004: 185) may be right when he argues that Arquius and Arcius cannot be two variants of the same name.

22 See Prósper (2002: 312).

23 See Abascal (1994: 313-314) and Vallejo (2004: 249-255) for the occurrences of the name.

24 See Evans (1979: 121-122) specifically on the Hispanic names.

25 See Abascal (1994: 320-321) and Vallejo (2004: 283-286), as well as the remarks by Albertos (1985: 277).

26 See Abascal (1994: 331-332) and Vallejo (2004: 283- 286) for the occurrences of the names in this series.

27 See GPN 180-181.

28 Palomar (1957: 52) analyzed Reburrus as derived from a base Burrus with the prefix re- and related it to the late Latin words burra (cf. Spanish borra 'sheep') and reburrus 'with crisp hair'. Pokorny (IEW 134) explained these forms as coming from the root *bher 'rub, cut'. For their distribution and a critique of Palomar's etymology see Vallejo (2004: 381-390).

29 See Villar (1995a: 61) and Vallejo (2004: 355-356).

30 See Abascal (1994: 458-459), as well as Villar's (1994: 238) map and Vallejo's (2004: 370-375) map and comments. Although Villar (1994: 237) stated that the variant Pentamus was not attested, it seems to have been identified recently in an inscription from Ponte da Barca VDC (HEp. 5.1055).

31 See Abascal (1994: 349) and also Vallejo (2004: 108-109 and 313-315), who argues, however, that names in Aeb- and Eb- must be kept separate.

32 See Abascal (1994: 279) and Vallejo (2004: 158-160).

33 See Abascal (1994: 474-475) and Vallejo (2004: 380-381).

34 See Luján (2003: 186).

35 See Ho. II 687, GPN 236-237.

36 The reading is uncertain.

37 When compared to Secuia and other names of this series.

38 See Abascal (1994: 410-411) and Vallejo (2004: 344-346).

39 See Abascal (1994: 425-426) and Vallejo (2004: 356-360). Medugenus at Caldas de Vizela, Guimarâes BRA (CIL II 2403 = 5558, with reading Rectugenus which has been later corrected) is an uxamensis, thus from Celtiberia. See Vallejo (2004: 689-690) for an analysis of -ei- for -e- as a typically Lusitanian phenomenon.

40 See Untermann (1985) and Prósper (2002), among others.

41 For the linguistic analysis of that inscription see Prósper (1998, 2002: 154-166).

42 See de Bernardo Stempel (2003) for a survey of the types of the dedications to this god and a linguistic analysis of the name.

43 See Prósper (2002: 329-330) for other possible etymologies and further literature.

44 See Prósper (2002: 129-130).

45 See CIRG II 1-12. The segmentation of the god name appearing in these inscriptions has been discussed. The author of the corpus reads Deo Laribero Breo, but given the frequent presence of inscriptions consecrated to the Lares in Callaecia the segmentation accepted above seems preferable. See Prósper (2002: 367-369) for a revision of the bibliography and a proposed etymology.

46 On the cult of the god Lug in Hispania see Marco Simón (1986).

47 The British goddess Conuentina had been related to the god name Cohuetene appearing in an inscription from the province of Lugo, but according to the analysis by Prósper (2002: 244-247), Cohue is merely a phonetic variant of the god name Cossue Coso that we have already mentioned, and Tene[ is the beginning of an epithet of the god.

48 On this linguistic development see Villar ( 1995a: 153-188).

49 This epithet may or may not be related to the adjective usseam on the Lusitanian inscription from Cabeço das Fragoas (MLH L-3.1). See Prósper (2002: 47, fn. 47) for a revision of the bibliography.

50 This name also appears on some inscriptions from Lusitania and in the Lusitanian inscription from Lamas de Moleo (MLH L-2.1).

51 See Tovar (1989: 124) for Albiones, Neri, and Nemetates. He does not consider Querquerni to be Celtic.

52 Possibly also in CIL II 2711 if Al() is to be read as Al(bio), as proposed by Albertos (1975: 32).

53 See Ho. I 188, KGP 132-133, GPN 142, DELG 52.

54 See KGP 280, LEIA s.u. treb, DELG 301.

55> For the problem of the interpretation of Claudionerium (Ptol. II 6.22) see Luján (2000: 56-57) and García Alonso (2003: 187-188).

56 See Tovar (1989: 136).

57 See LEIA s.u. nert.

58 See Luján (2000: 62) and García Alonso (2003: 251).

59 Other names such as Cabarci ( Pliny NH IV 111) and gen. plu. Obili(um) (AF I 2 618) may also belong in this group.

60 For other possibly related place and personal names and a discussion of the possible etymologies of the names see García Alonso (2003: 205-206).

61 This seems to be the only occurrence of this name referring to a specific people and not as a general name for all the inhabitants of Callaecia (Tovar 1989: 129).

62 See §5.1 below.

63 For the possibility of epigraphic occurrences see García Alonso (2003: 202-203).

64 See KGP 183-184 for the series of names in Coro-/ Corio- and GPN 338-339 for Coros.

65 See García Alonso (2003: 269-270) on the place-name Cauca in central Spain.

66 On which see KGP 144.

67 On which see GPN 232-233.

68 For other names in Lub- see Ho. II 295-296.

69 For the use of Coporus as a personal name see Albertos (1966: 94-95), Abascal (1994: 334) and García Alonso (2003: 200-201 and fn. 169).

70 Gorrochategui (1987: 85 and fn. 38), cf. Prósper (2002: 317).

71 See Gorrochategui (1987) and Untermann (MLH IV 730).

72 Ptol. II 6.1, Mela III 1.10. It.Ant. 425.2 and 430.10 records a ponte Neuiae and Rav. 4.45, a ponte Abei, usually considered a mistaken reading of Neuiae. See García Alonso (2003: 131-132).

73 The reading of the inscription seems to be, in fact, Aebisoci, but given that the peoples mentioned in it appear in alphabetical order, it is usually assumed that it should be read as Naebisoci. Nevertheless, the dat. Ebu[soc]o has been proposed by Colmenero for an inscription from Mairos, Chaves VRE (AF I 2 247).

74 B(e)ibalus is also attested as cognomen in two inscriptions (CIL II 2475; AEA 39.14).

75 See Luján (2005: 403-405) and HEp. 8.325.

76 See García Alonso (2003: 244-245) for the relevant bibliography and for the possibility of a Celtic etymology as *be-ba(l)lo-, from the root *ba(l)l- known in Celtic onomastics (GPN 147), plus a reduplication.

77 See Tovar (1989: 130).

78 The names Helleni (Pliny NH IV 112) and Amphiloci (Str. III 4.3) seem to be adaptations of similar Greek names, so we cannot use them for our purposes.

79 For possible epigraphical occurrences see Tovar (1989: 132).

80 The same suffix appears in the name Gigurri. For some etymological possibilities, all of them uncertain, see García Alonso (2003: 207-208).

81 For ad- see KGP 111-116 and GPN 128-131; for se(c)c-, Ho. II 1422-1426 and KGP 265; and for seg-, Ho. II 1444-1453, KGP 265-266, GPN 254-257.

82 See Tovar (1989: 310-311) for further references.

83 See Luján (2000: 64-65) and García Alonso (2003: 232-234) for other possibilities.

84 The sequences ]dun[ and ]un[ in the final lines of an inscription from Braga (EE VIII 120) have sometimes been completed as Caladuniensis or, alternatively, as Beduniensis. Caladu<n>us as a personal name occurs on another inscription from Braga (EE VIII 125 = AE 1983.570).

85 See Luján (2000: 59-60) and García Alonso (2003: 234-236). See Tovar (1989: 306) for another place-name Cale, which would later survive into Portocale > Portugal.

86 For the second step in the evolution, cp. ko-n-bo-u-to (= Complouto), the name of another place in Spain, on Celtiberian coins (MLH I A-74d).

87 For further comments on this name see Luján (2000: 60) and García Alonso (2003: 236-237).

88 See KGP 202, GPN 346-347.

89 The name is attested only from late antiquity onwards; see Tovar (1989: 316).

90 See Tovar (1989: 315) for various suggestions regarding corrections of this name.

91 For the difficulties associated with identifying Nouion with the town mentioned by Pliny see Tovar (1989: 300-301), Luján (2000: 58), and García Alonso (2003: 188).

92 See Ho. II 787-792.

93 See the remarks by Sims-Williams (2005: 273) for additional bibliographic references. The word ocelo- is attested in the god name Lari Ocaelaego in Callaecia; see Prósper (2002: 110) for further references.

94 On which see DELG 240.

95 Further remarks in Luján (2000: 58) and García Alonso (2003: 190-191).

96 For collection of names in -briga in Hispania see Albertos (1990) and Villar ( 1995a: 155-159). Villar (1995a: 153-188) also provides a detailed linguistic analysis of the variations of these names in Spain. Interesting remarks can be found also in Prósper (2002: 357-382). See also the comments on -briga place-names and their classification by García Alonso (2006).

97 This is the usual emendation, but some VicaniCabr(icenses?) appear on an inscription from Vieiza do Minho BRA (HEp.4.1016).

98 Further sources in Tovar (1989: 309).

99 Further names in -brig- identified in epithets of gods can be found in Prósper (2002: 357-382).

100 Sometimes Lanobriga has been read on an inscription from Eiras, San Amaro OR (AF I 2 105), but the reading is uncertain. See Alansbrica below.

101 For lama see §5.1 above.

102 For other names in Adro- see Ho. I 45-46.

103 Aiobrigiaeco in the Tabula from O Caurel (HAE 1965 = HEp. 8.334) with corrected reading according to the occurrence of the name in El Bierzo edict (HEp. 8.325).

104 See Tovar (1989: 293).

105 See Albertos (1990: 132-133) and García Alonso (2003: 99).

106 See Abascal (1994: 292) and Vallejo (2004: 196-198). For the analysis and contextualization of this name see de Hoz (1994).

107 May be identical with Abobriga and Aobriga known from other sources; see Albertos (1990: 133-134).

108 See Isaac (2002: AI WW/Celtic elem. etym.).

109 The Caelobrigoi are mentioned in the Lusitanian inscription from Lamas de Moledo (CIL II 416 = MLH IV L-2.1), but it is not certain that the same place is referred to in both cases.

110 See Luján (2000: 61-62) and García Alonso (2003: 243-244).

111 See KGP 183-184.

112 On which see the analysis of Turriga below.

113 See Albertos (1985: 297-298) and Vallejo (2004: 417-423).

114 See Prósper (2002: 157-164) for a revision of the etymology. According to her these would be derivatives from the IE root *teng- 'make wet', and not from the root of OIr. tongid 'swear'.

115 See Tovar (1989: 293), Luján (2000: 60-61) and García Alonso (2003: 238-239).

116 Tovar (1989: 301) also has suggested the possibility that the second part of the sequence Netaciueilebrigae in CIL II 2539, most probably a god name, might also be this name, adding thus a further possibility for the vowel of the first syllable. Prósper (2002: 370) suggests splitting the sequence as Neta Ciueilebrigae. In that case Ciueilebrigae should be added to our list of names in -briga, but the first element Ciueile- is rather odd.

117 See Luján (2000: 61) and García Alonso (2003: 243).

118 For later sources see Tovar (1989: 307).

119 See García Alonso (2003: 188-190) for further possibilities.

120 See Luján (2000: 65) and further comments by García Alonso (2003: 252-254).

121 See GPN 162-166 for that root and García Alonso (2003: 196-197) for other possibilities of analyzing this particular name.

122 See the study of those names by Pedrero (1996).

123 See Luján (2000: 63-64) and García Alonso (2003: 198-199).

124 See Luján (2000: 65). Other possibilites in García Alonso (2003: 250-251).

125 See Villar (2000: 291-292).

126 See Tovar (1989: 295).

127 See further comments by García Alonso (2003: 208-210).

128 See Luján (2000: 64) and García Alonso (2003: 200).

129 See Tovar (1989: 304), Luján (2000: 63), and García Alonso (2003: 197-198).

130 See Tovar (1989: 304).

131 Tentative etymologies could be proposed at least for some of these, but they would be highly speculative and would not change the general look of Galician onomastics.

132 This is the name of a castellum. The reading is not certain and Acripae has also been suggested.

133 See Luján (2000: 66) and García Alonso (2003: 239-241).

134 See Tovar (1989: 304) for the possibility that the Aregenses montes mentioned by Joh. Bicl. s.a. 572,2 Chron. min. II p. 214, are related to it.

135 For the possibility that Burbida and Bonisana are, in fact, the same place, see Tovar (1989: 299).

136 See García Alonso (2003: 204-205) for a tentative analysis as related to Celtic dago- 'good'.

137 Attested only from Suebian times on. See Tovar (1989: 302).

138 This is the name of a castellum. Unfortunately the complete name is not preserved, but the presence of an initial f- shows that the name was not Celtic.

139 References to later sources can be found in Tovar (1989: 318).

140 The existence of the place-name itself has been questioned. See Luján (2000: 66) and García Alonso (2003: 242).

141 See Luján (2000: 65) and García Alonso (2003: 192-193) for further comments and possibilities of analysis.

142 This seems to be a castellum of the Limici, but the reading of the inscription is doubtful.

143 Name of a castellum.

144 Name of a castellum.

145 This reading is not definite – Lovagornicenes and Vagornicenses are alternate readings.

146 For further sources in later times see Tovar (1989: 298-299). More comments in García Alonso (2003: 247-248).

147 This is the name of a castellum. The name looks very odd and the segmentation of the sequence is uncertain.

148 ]ruecensis appears in CIL 5583. The readings Ripau(m) (castellum) Puant(ium) (AF I 2 617) and Amba(um) Colen(ae) (CIL 2482 = AF I 2 371) are extremely uncertain.

149 On the names of this river in antiquity see the paper by Guerra (1996), to which the appearance of Obliuio in a newly discovered papyrus containing part of the Greek text of the Geography by Artemidorus of Ephesus (Kramer 2005: 29) should be added.

150 See García Alonso (2003: 134-135).

151 It is, however, problematic to explain prae(s)-, given that this prefix is not used in Latin for creating place-names. Pokorny (IEW 844) thought that *praesta- was a participial form meaning 'lovable, beloved', so *marko- would be the Celtic word for 'horse', cf. Ags. frīd-hengest.

152 See TIR K-29 s.u. for a list of the sources.

153 See García Alonso (2003: 135-136).

154 The 'infection' is the appearance of a non-etymological i – which may sometimes evolve into e – as a mark of palatalization of the following consonant.

155 See Luján (in press).

156 The question of whether CIL II 685, an inscription devoted to Reue Ana Baraeco preserved in the Diocesan Museum of Astorga (León), came from Ruanes (Cáceres) or from Rubianes (Orense) has been discussed. In the latter case we would have two records of the cult of this god in Callaecia.

157 See Canto et al. (1997).

158 The same applies to the Turduli, mentioned together with them by Strabo.


Bibliography

Abascal, J.M.
1994. Los nombres personales en las inscripciones latinas de Hispania. Murcia: Universidad Complutense de Madrid - Universidad de Murcia.

AE = Année Épigraphique [journal].

AF I2 = Rodríguez Colmenero, A.
1997. Aquae Flaviae I. Fontes epigráficas de Gallaecia meridional interior. Chaves: Câmara Municipal de Chaves.

Albertos Firmat, M.ª L.
1966. La onomástica personal primitiva de Hispania (Tarraconense y Bética). Salamanca: Consejo Superior de Investigaciones Científicas.
1975. Organizaciones suprafamiliares en la Hispania antigua. Boletín del Seminario de Arte y Arqueología 40-41: 5-66.
1985. La onomástica personal indígena del noroeste Peninsular (astures y galaicos). In J. de Hoz (ed), Actas del III Coloquio sobre Lenguas y Culturas Paleohispánicas, pp. 255-310. Salamanca: Universidad de Salamanca.
1990. Los topónimos en -briga en Hispania. Veleia 7: 131-146.

de Bernardo Stempel, P.
2003. Los formularios teonímicos, Bandus con su correspondiente femenino Bandua y unas isoglosas célticas. Conimbriga 42: 197-212.

Berrocal Rangel, L.
1992. Los pueblos célticos del suroeste de la península Ibérica (Complutum Extra 2). Madrid: Universidad Complutense.

Canto, A. M., A. Bejarano and F. Palma
1997. El mausoleo del dintel de los ríos de Mérida, Revve Anabaraecus y el culto de la confluencia. Madrider Mitteilungen 38: 247-294.

CIL = Hübner, E. (ed.)
1869. Corpus Inscriptionum Latinarum, vol. II, Berlin: G. Reimer.
1892. Supplementum. Berlin: G. Reimer.

CILA I = González Fernández, J.
1989. Corpus de Inscripciones Latinas de Andalucía, vol. I Huelva. Sevilla: Junta de Andalucía.

CIRG = Pereira Menaut, G. (dir.)
1991. Pereira Menaut, G., Corpus de Inscripciones Romanas de Galicia, vol. I Provincia de A Coruña. Santiago: Consello de Cultura Galega.
1994. Baños Rodríguez, G., Corpus de Inscripciones Romanas de Galicia, vol. II Provincia de Pontevedra. Santiago: Consello de Cultura Galega.

DELG = Delamarre, X.
2002. Dictionnaire étymologique de la langue gauloise. (2nd edition). Paris: Errance.

EE = Ephemeris Epigraphica [journal].

Evans, D.E.
1979. On the Celticity of some Hispanic personal names. In A. Tovar et al. (eds), Actas del II Coloquio sobre Lenguas y Culturas Prerromanas de la Península Ibérica, pp. 117-129. Salamanca: Universidad de Salamanca.

ERA = Diego Santos, F.
1985. Epigrafía romana de Asturias. Oviedo: Instituto de Estudios Asturianos.

García Alonso, J.L.
2003. La península Ibérica en la Geografía de Claudio Ptolomeo. Vitoria: Universidad del País Vasco.
2006. -briga toponyms in the Iberian Peninsula. In M. Alberrro and B. Arnold (eds), The Celts of the Iberian Peninsula. e-Keltoi 6: 689-714. http://www.ekeltoi.uwm.edu/. Milwaukee: Center for Celtic Studies, University of Wisconsin-Milwaukee.

Gorrochategui, J.
1987. En torno a la clasificación del lusitano. In J. Gorrochategui, J.L. Melena and J. Santos (eds), pp. 77-91. Studia Palaeohispanica (Actas del IV Coloquio sobre Lenguas y Culturas Paleohispánicas) [= Veleia 2-3]. Vitoria.
1997. Gallaecia e as linguas prerromanas da Península Ibérica. In Galicia fai dous mil anos. O feito diferencial galego, vol. I, pp. 15-49. Santiago de Compostela: Museo do Pobo Galego.

GPN = Evans, D.E.
1967. Gaulish Personal Names. Oxford: Clarendon Press.

Grupo Mérida
2003. Atlas antroponímico de la Lusitania. Mérida: Pessac, Fundación de Estudios Romanos - Ausonius.

Guerra, A.
1996. Os nomes do Rio Lima. Um problema de toponímia e de geografía histórica. In F. Villar and J. d'Encarnaçâo (eds), La Hispania Prerromana (Actas del VI Coloquio sobre Lenguas y Culturas Prerromanas de la Península Ibérica), pp. 147-161. Salamanca: Universidad de Salamanca and Universidade de Coimbra.

HAE = Hispania Antiqua Epigraphica [journal].

HEp. = Hispania Epigraphica [journal].

Ho. = Holder, A.
1896-1913. Alt-celtischer Sprachschatz, 3 vols. Leipzig: B.G. Teubner.

de Hoz, J.
1992. The Celts of the Iberian Peninsula. Zeitschrift für Celtische Philologie 45: 1-37.
1994. Castellum Aviliobris. Los celtas del extremo occidente continental. In R. Bielmeier and R. Stempel (eds), Indogermanica et Caucasica. Festschrift für Karl Horst Schmidt zum 65. Geburtstag, pp. 348-362. Berlin and New York: Walter de Gruyter.
1997. Lingua e etnicidade na Galicia antiga. In Galicia fai dous mil anos. O feito diferencial galego, vol. I, pp. 101-140. Santiago de Compostela: Museo do Pobo Galego.

IEW = Pokorny, J.
1949. Indogermanisches etymologisches Wörterbuch. Bern: Francke.

IRG II = Vázquez Saco, F. and M. Vázquez Seijas
1964. Inscripciones romanas de Galicia, vol. II Provincia de Lugo. Santiago de Compostela: Consejo Superior de Investigaciones Científicas.

IRG IV = Lorenzo Fernández, J., Á. D'Ors and F. Bouza Brey
1968. Inscripciones romanas de Galicia, vol. IV Provincia de Orense. Santiago de Compostela: Consejo Superior de Investigaciones Científicas.

ILER = Vives, J.
1971-1972. Inscripciones latinas de la España romana. Barcelona: Universidad de Barcelona.

IRPL = Arias Vilas, F. and P. Le Roux
1979. Inscriptions romaines de la province de Lugo. Paris: Diffusion De Boccard.

Isaac, G. R.
2002. The Antonine Itinerary. Land Routes (Place-Names of Ancient Europe and Asia Minor). Aberystwyth, CMCS Publications (CD-Rom).

KGP = Schmidt, K.H.
1957. Die Komposition in gallischen Personennamen. Zeitschrift für celtische Philologie 26: 31-301.

Kramer, B.
2005. El nuevo papiro de Artemidoro. In J. de Hoz, E.R. Luján and P. Sims-Williams (eds), New Approaches to Celtic Place-Names in Ptolemy, pp. 19-31. Madrid: Ediciones Clásicas.

LEIA = J. Vendryes, continued by E. Bachellery and P.-Y. Lambert.
1960. Lexique Étymologique de l'Irlandais Ancien, Vol. M-N-O-P, Paris: DIAS-CNRS.
1978. Lexique Étymologique de l'Irlandais Ancien. Vol. T-U, Paris: DIAS-CNRS.

Luján, E.R.
2000. Ptolemy's Callaecia and the language(s) of the Callaeci. In D. N. Parsons and P. Sims-Williams (eds), Ptolemy. Towards a Linguistic Atlas of the Earliest Celtic Place-names of Europe (Papers from a workshop, sponsored by the British Academy, in the Department of Welsh, University of Wales, Aberystwyth, 11-12 April 1999), pp. 55-72. Aberystwyth: CMCS Publications.
2001. La onomástica de los celtici de la Bética: análisis lingüístico. In F. Villar and M. P. Fernández (eds), Religión, lengua y cultura prerromanas de Hispania, pp. 471-481. Salamanca: Universidad de Salamanca.
2003. Gaulish personal names: an update. Études Celtiques 35: 181-247.
2005. La onomástica del Edicto del Bierzo. In P. P. Conde Parrado and I. Velázquez (eds), La Filología Latina. Mil años más. Actas del IV Congreso de la Sociedad de Estudios Latinos (CD-Rom), pp. 398-406. Madrid: Sociedad de Estudios Latinos.
in press. L'onomastique des Vettons: analyse linguistique. In P.-Y. Lambert and G.-J. Pinault (eds), Gaulois et celtique continental. Paris: École Pratique des Hautes Études.

Marco Simón, F.
1986. El dios céltico Lug y el santuario de Peñalba de Villastar. In Estudios en homenaje a Antonio Beltrán, pp. 731-759. Zaragoza: Facultad de Filosofía y Letras, Universidad de Zaragoza.

MLH = Untermann, J.
1980-1997. Monumenta Linguarum Hispanicarum, 4 vols. Wiesbaden: Dr. Ludwig Reichelt.

Morlet, M.-T.
1972-1985. Les noms de personne sur le territoire de l'ancienne Gaule du VIe au XIIe siècle, vol. II Les noms latins ou transmis par le latin, vol. III Les noms de personne contenus dans les noms de lieux. Paris: C.N.R.S.

Olivares Pedreño, J.C.
2002. Los dioses de la Hispania céltica. Madrid: Real Academia de la Historia and Universidad de Alicante.

Palomar Lapesa, M.
1957. La onomástica personal pre-latina de la antigua Lusitania. Salamanca: C.S.I.C.

Pedrero, R.
1996. El hidrónimo prerromano Mira. Emerita 64: 361-374.
1999. Aproximación lingüística al teónimo lusitano-gallego Bandue/Bandi. In F. Villar and F. Beltrán (eds), Pueblos, lenguas y escrituras en la Hispania prerromana (Actas del VII Coloquio sobre Lenguas y Culturas Prerromanas de la Península Ibérica), pp. 535-543. Salamanca: Universidad de Salamanca.

Prósper, B.M.
1998. Tongoe Nabiagoi: la lengua lusitana en la inscripción bracarense del ídolo de la fuente. Veleia 14: 163-176.
2002. Lenguas y religiones prerromanas del occidente de la península Ibérica. Salamanca: Universidad de Salamanca.

RAP = Garcia, J.M.
1991. Religiôes Antigas de Portugal. Lisboa: Imprensa Nacional.

Schwerteck, H.
1993. Die Inschrift von Ginzo de Limia. Historische Sprachforschung 106: 115-128.
1998. Die Inschrift von Ginzo de Limia: Nachträge. Historische Sprachforschung 111: 99-101.

Sims-Williams, P.
2005. Measuring Celticity from Wales to the Orient. In J. de Hoz, E.R. Luján and P. Sims-Williams (eds), New Approaches to Celtic Place-Names in Ptolemy, pp. 266-287. Madrid: Ediciones Clásicas.

TIR K-29 = Unión Académica Internacional.
1991. Tabula imperii Romani, hoja K-29 Porto. Madrid: Instituto Geográfico Nacional and Consejo Superior de Investigaciones Científicas.

Tovar, A.
1989. Iberische Landeskunde, 2nd part Las tribus y las ciudades de la antigua Hispania, vol. 3 Tarraconensis. Baden-Baden: Valentin Koerner.

Untermann, J.
1965. Elementos de un atlas antroponímico de la Hispania antigua. Madrid: C.S.I.C. and Universidad de Madrid.
1985. Los teónimos de la región lusitano-gallega como fuente de las lenguas indígenas. In J. de Hoz (ed), Actas del III Coloquio sobre Lenguas y Culturas Paleohispánicas, pp. 343-363. Salamanca: Universidad de Salamanca.
1992. Los etnónimos de la Hispania Antigua y las lenguas prerromanas de la Península Ibérica. In M. Almagro-Gorbea and G. Ruiz Zapatero (eds), Paleoetnología de la península Ibérica (= Complutum 2-3), pp. 19-33. Madrid: Editorial Complutense.

Vallejo Ruiz, J. M.
2004. Antroponimia indígena de la Lusitania romana. Vitoria: Universidad del País Vasco.

Villar, F.
1993. Talabara, Talavera, Toledo. In I.J. Adiego, J. Siles and J.Velaza (eds), Studia palaeohispanica et indogermanica J. Vntermann ab amicis Hispanicis oblata, pp. 287-296. Barcelona: Universidad de Barcelona.
1994. Los antropónimos en Pent-, Pint- y las lenguas indoeuropeas prerromanas de la Península Ibérica. In R. Bielmeier and R. Stempel (eds), Indogermanica et Caucasica. Festschrift für Karl Horst Schmidt zum 65. Geburtstag, pp. 234-264. Berlin and New York: Walter de Gruyter.
1995a. Estudios de celtibérico y de toponimia prerromana. Salamanca: Universidad de Salamanca.
1995b. El hidrónimo prerromano Tamusia, moderno Tamuja. In J.F. Eska, R.G. Gruffydd and N. Jacobs (eds), Hispano-Gallo-Brittonica, pp. 260-277. Cardiff: University of Wales Press.
2000. Indoeuropeos y no indoeuropeos en la Hispania prerromana. Salamanca: Universidad de Salamanca.

Villar, F. and R. Pedrero 2001. La nueva inscripción lusitana: Arroyo de la Luz III. In F. Villar and M. P. Fernández Álvarez (eds), Religión, lengua y cultura prerromanas de Hispania (Actas del VIII Coloquio de lenguas y culturas paleohispánicas), pp. 663-698. Salamanca: Universidad de Salamanca.