The Celtiberian Lug


Celtiberian was spoken in the north central meseta of the Iberian peninsula. It was probably the first language to split off from the Proto-Celtic speech area (possibly half way the 1st millenium BCE) and it is by far the most conservative Celtic language. The majority of the inscriptions was engraved in the semi-segmental Iberian script (see Eska 2008). Other inscriptions are written in the Lugano alphabet or in adapted Latin script.

Roman Celtiberia

Peñalba de Villastar

The inscription discussed in this post is found in Peñalba de Villastar which is a paradigmatic mountain sanctuary (see Alfayé et Simon 2013). It is situated on a rocky cliff around 3 km away from the river Turia in one of the border areas of the Iberoceltic cultural sphere. It was discovered in 1908 by J. Cabré during excavations. Into its cliff wall of marl and sand there are numerous incriptions which must have been inscribed during a long period spanning decades. Among these inscriptions we find many epigraphic votiv inscriptions. Because no material finds inform us of the nature of the cult in Peñalba the principal source of information on the the character of the cult in this sanctuary consists of the epigraphic celtiberic writings (see Alfayé). Of special importance is the “gran inscripción” which is treated in this presentation. It presumably dates from the 1st or 2nd c. CE.

gran inscripción

The inscription

The central words that can be understood in this inscription are luguei which occurs twice. The word is regarded to refer to the Celtic deity Lugus (cf. OI Lug, MW llew, Gaul. lugus). Other words are also easily recognizable, for example the verbal forms comeimu and sistat.

PCelt. *Lugus

Lugus is known as a god in theophoric toponyms in Gauls, most notably Lugudunum (cf. ModFr. lyons). It is possible the place name Leiden is a Germanic reinterpretation of another Lugudunum (i.e. Gmc. *lagiþūna : Celt. Lugudunom). The worship of Lug is also known from the Chamalières tablet where the Gaulish formula Luge dessumi iis “to Lugus I prepare them” is found (see Koch 1994: 2-3). The etymology of the name is unclear. If it were derived from PIE *leuk- we would not expect lenition in the Gaulish forms.

The Old Irish Lug, Lug macc Eithliu (Lug, the son of Eithliu), is named as the euhemerized father of Cu Chulainn in the Tain Bo Cuailgne who was at one point ollam (< PCelt. *olyosamos) “the prime king” of whole Ireland. although strictly speaking this tale is only known in a Middle Irish version (Puhvel 1987: 172). In their medieval tales the Irish told that he was one of the Tuatha Dé Danann “people of the goddess Dana”. In Medieval Irish they called him samildanach “the many skilled,” Lonnbeimnech “fierce striker” or lamfada “long hand” and in Welsh as Llew llawgyffes “Lug the deft handed.” Lug’s prime attribute is a spear which is named in multiple mythological tales. In the Lebor Gabala (the book of the taking of Ireland) it is mentioned that Lug invented the riding whip and horseback racing. The name of the Welsh hero lloumarch < PCelt. *lugumarkos from the Canu Llouwarch Hen (the song of Lloumarch the Old) confirms Lug’s connection to horses (see Meid 1996: 16), which would be paralleled in other Indo-European traditions, e.g. OIc. jálkr “stallion” for Óðinn.

In Old Irish the feast name lugnasad  (1st of august) is attested which even made its way into modern Irish as lughnasadh. During this feast a procession was made to a mountaintop (see Meid 1996: 11-12). Meid takes this as an indication that Lug was especially associated with mountains. Plinius the Elder describes a statue of Mercury erected by the Gaulish Averni tribe on top of the Puy de Dôme in the Auvergne. According to the interpretatio Romana this may very well have been Lug.

Meid’s reading, analysis and translation[1]

eniorosei                 uta          tigino      tiatunei   tre           caias       to             luguei     araianom comeimu.                      conj.     prep.     prep.

for the mountainous one and to the ? of the master we come together through fields to Lug of the Araians.


eniorosei equisui   =que        ocris        olcas        togias      sistat       luguei,           tiaso        togias,     post.pos.    acc.p.

for the mountainous one and the equestrian one, for Lugus, the head of the community sets proper coverings, the coverings of ?.


Ködderitzsch’ reading, analysis and translation

eniorosei                 uta          tigino      tiatunei   erecaias to             luguei     araianom comeimu.                     conj.     prep.

for Eniorosis and to the Tiatū of Tigino we bestow furrows, to Lug a field.


eniorosei equisui   =que        ogris        olcas        togias      sistat       luguei,               tiaso        togias,     post.pos.    acc.p.

for Eniorosis and to Equaesos, for Lugus, Ogris sets coverings of the arable land, the coverings of the scorched land.


Celtib. *eni-oros “the mountainous one?” or PN

Celtib. *uta “and” (cf. Skt. utá)

Celtib. *tigi(r)no “master”? (cf. OI tigern, MW teyrn “master’?) or toponym

Celtib. *tre “through” (cf. OI tre, OW trui)  

Celtib. *caias “hedges, fields” (cf. Gaul. caio “hedge”, MW cae “field”)

Celtib. *erecaias “furrows” < PCelt. *φerkaya- < PIE *perḱ- (cf. ModW rhych, Lat. porca, ModE furrow)

Celtib. *to “to” (cf. OI to/do, OW di “id.”)

Celtib. *luguei “to Lug” (cf. OI Lug, MW llew, Gaul. Lugus)

Celtib. *araianom “of the araians” [ethnonym]

Celtib. *arianom “a field” (cf. OI ar “plowed land, OI airid “to plow”)

Celtib. *comeimu “we go” (cf. OI ethae “gone”, p.p.p. to OI téit “to go”)

Celtib. *com-meimu “we bestow” < PIE *mei̯- (cf. OI moín < PCelt. moini-, Skt. máyate “exchanges,” Lat. mūnus < *moinos)

Celtib. *ekwisui “to the equestrian one” (cf. OI ech, MW ebawl, Gaul. epos “horse”)

Celtib. *-kwe “and” (cf. OI –ch, Lat. –que, Hitt. –kku, Skt. –ca, Gk. -τε, Goth. –uh “id.”)

Celtib. *okris “border, edge” (cf. MI ochair “border, edge”, Gk. ὄκρις “cliff” see Beekes 2009: 1066, Matasovic 2009: 28) or personal name Ogris (latter option sounds more convincing).

Celtib. *olka “field” < PCelt. *φolkā “id.” (cf. Gallo-Rom. olca [Gregory of Tours], OFr. ouche, Prov. olca “arable land”, see Meyer-Lübke 1911: 446) (convincing etymology).

Celtib. *togias (cf. OI tuige “cover, protection” < PCelt. *togyā, see Matasovic 2009: 376). (convincing etymology).

Celtib. *sistat “set up” (cf. OI sissidir “stands,” Lat. sistit “to cause to stand, set up)

Meid’s dubious assertions

Notable in Meid’s analysis is the reading of the single line in ?recaias as a whereas most scholars read it as an e. He also deems the spelling of and to be completely random, which in some cases has consequences for the etymology. Especially striking about Meid’s interpretation is that there are no real personal names in the inscription, which would be uncommon for an inscription of a votive nature.

Celtib. *eni-oros (see next paragraph)

Celtib. *okris “head” (metaphoric sense of “top, edge” nowhere attested elsewhere in Celtic)

Celtib. *olokā “community” < PCelt. *olyo- + – “entirety” (formation nowhere attested)

Celtib. *tiaso “of the guild” (cf. Gk. θίασος, Lat. *thiasus “bacchic revel, religious guild”, see Beekes 2009: 548). (Assumes quite a severe loan of a terminus technicus for a cultic phenomenon into Celtiberia).

Ködderitzsch’ dubious assertions

Notable in Ködderitzsch’s analysis is his treatment of the VRVC-sequences which he interprets as containing an anaptyctic vowel. Especially in inscriptions anaptyctic vowels often turn up, so this does not need to be a problem. Since his etymologies are generally solid, the postulation of the anaptyxis seems justified. The only problem to my mind would be his analysis of the word below:

Celtib. *tiaso “burnt land” < PCelt. *teφoso (cf. OI tee “hot,” Lat. tepeo “to warm”). [Meid’s interpretation here is actually better. Why did PCelt. *eφo have to yield Celtib. *ia]

Comments on Celtib. *eni-orosei

Meid suggests that the word Celtib. eniorosei is a compound of PCelt. eni < PIE *h1eni (cf. OI ingen “girl” < PCelt. *enigenā and OI inis “island” < PCelt. *eni-stā) and a purported PCelt. *oros “elevation.” However, the only cognate to this word would be Gk. ὄρος[2] “mountain, height” < *h3eros (see Beekes 2009: 1109). Although a parallel formation in Celtic seems possible, one would like to have more cognates. The Greek word is an s-stem, i.e. Gk. nom. ὄρος, gen. ὄρεος.  Meid’s analysis of the apophony of the s-stem would be secondary. The Gk. word suggests PIE nom. *h3ér-os, gen. *h3r-es-os. In order to get to Celtib. *eniorosei, we would have to posit generalization of the root vocalism and generalization of the o-grade suffix. However possible, this is not the simpelest scenario. Moreover, Meid´s suggestion that it would simply be possible to add an i-theme to an already completes s-stem seems ad hoc (Meid 1996: 13). We should note that a name Orosis is attested on a Iberoceltic coin. Probably it is either a personal name or a toponym. Either way, its etymology is unclear and its connection to Gk. ὄρος is possible, yet difficult. Therefore the assertion that we are dealing with an epithet meaning “mountainous” is dubious at best.


As with many inscriptions, the acceptation of certain etymologies for specific words is crucial for its overal interpretation. For this inscription we have two very different interpretations. Meid deems it to be an inscription which informs us about a cultic procession by a college of devotees (the thiasos) to a mountainous god, “Lug the Mountain Dweller,” an image we know very well from Vedic and Germanic evidenc. Ködderitzsch however holds it to be a votive inscription concerning the dedication of land on behalf of a certain third party to Lug. It is hard to tell who is right. We can remark that Ködderitzsch is overal more consistent in his etymologies than Meid. Especially his insistance that we must be dealing with atleast a couple of personal names (like in any other votive inscription) is quite convincing. If his interpretation is largely correct this unfortunately means the inscription does not give us extra information concerning the Celtiberian Lug.


Alfayé, Silvia Villa et Francisco Marco Simón

(2013)       (paper) El santuario de Peñalba de Villastar (Teruel) y la romanización religiosa en la Hispania indoeuropea

Beekes, Robert S.P

(1995)        Comparative Indo-European linguistics; an Introduction (Amsterdam).

2009        Etymological Dictionary of Greek (Leiden).

Eska, Joseph. F

2004       “Continental Celtic,” in: The Ancient Languages of Europe, Roger D. Woodward ed (Cambridge).

Koch, T. John

2003       The Celtic Heroic Age; literary sources for Ancient Celtic Europe & Early Ireland & Wales (Aberystwyth).

Ködderitzsch, Rolf

1985        “Die grosse Felsinschrift von Peñalba de Villastar ” in: H. M. Olberg, H. Bothien et G. Schimdt eds., Sprachwissenschaftliche Forschungen; Festschrift für Johann Knobloch (Innsbruck).

Matasović, Ranko

2009        Etymological Dictionary of Proto-Celtic (Leiden).

Meid, Wolfgang

1996        Kleinere keltiberische Sprachdenkmäler (Innsbruck) 8-19.

Puhvel, Jaan

1987        Comparative Mythology (Baltimore).

[1] “Zu dem auf dem Berge wohnenden, und zugleich dem ?, dem Lugus der Araianer sind wir über die Fluren zusammengekommen. Dem auf dem Berge wohnenden und dem Pferdegott, dem Lugus, hat das Oberhaupt der Gemeinschaft eine Überdachung errichtet, zugleich auch für den thiasus eine Überdachung.”

[2] Beekes’ suggestion that the formations meaning “bottom, ass” (cf. Gk. ὄρρος “rump, arse,” Arm. ōr, ōrk͑, OHG ars) are connected to the same root PIE *h3er- is unconvincing because of OI err < PCelt. *ersos. The Celtic form points to an original root ablauting paradigm to a root PIE *h1ers- (i.e. PIE nom. *h1er-s, acc. *h1ors-es-m, gen. *h1ors-és). Nevertheless, assuming an ablauting paradigm would complicate the picture since ablauting hysterodynamic s-stems do not have an o-grade of the root  (see Kroonen 2013).

Leave a Reply

Your email address will not be published. Required fields are marked *