The second element of artebuθz has been connected with a number of lexemes and onomastic elements whose interrelations and exact etymologies are under debate. See DLG: 95 s.v. bussu- for what appears to be the first mention of the possible inclusion of -buθz in this difficult group: Delamarre analyses -buθz as -buđđ(o)s and translates arte-buθz 'Pine d'Ours', attributing the suggestion to Christopher Gwinn and David Stifter (on the now defunct continentalceltic mailing list).
bussu- is an onomastic element which occurs frequently as the first part of Gaulish compound PNN, e.g., bussumaros (epithet of Iuppiter), bussugnata (in Pannonia and Noricum), bussurigios, and in simplex names, e.g., bussus, bussulus/bussula, bussuro; for more examples and attestations see KGP: 158, GPN: 316 f. s.v. buscilla with n. 7, Lochner von Hüttenbach 1989: 39 f., Meid 2005: 128 f., Delamarre 2007: 214 et passim. It was connected with Insular Celtic words for 'lips' (MIr. bus (poetic), MW gwe-fus, MBret. gweuz, gueus, OCorn. gueus (glossed as 'labia'); see Matasović 2009 s.v. *bussu-) by Osthoff 1894: 286 (who credits Thurneysen by letter) and Holder 1896–1907 I: 645, and further with Gaulish L-119 buđđutton on a spindle whorl, translated as 'little kiss' ('come girl, take my buđđutton'), by Loth 1916: 183, Dottin 1920: 238, Whatmough 1951: 182, KGP: 158, IEW: 103 s.v. bu-, Tovar 1963: 390. Gaul. buđđutton attests tau gallicum as the root auslaut, which allows for the tentative addition of further Gaulish onomastic elements which show different spelling variants of that sound, such as but(t)u- (e.g., butto, buttus, but(t)uricus, buturo, butturrus, GPN: 319 n. 7), bustu- (busturo, Lochner von Hüttenbach 1989: 39), buddu- (buddutto, ibid.), buθθ- (buθθarus, Delamarre 2007: X). The lexical 'lips'-group is problematic insofar as its stem, gender and etymology are unclear. Meid 1980: 15 f. compares (specifically for buđđutton) Austrian Buss-erl 'kiss, peck [dimin.]', Delamarre DLG: 96 an assortment of further similar words in various European languages meaning 'lip', 'mouth', 'kiss', 'embrace', which he mainly attributes to onomatopoeïa rather than a common PIE base.
An alternative group of cognates which has been connected with the Gaulish buss-/buđđ-/buθ-group, which accounts for Delamarre's translation of arte-buθz as 'bear prick', is MIr. bot 'penis, tail', MW both 'nave, shield boss' < PC *bozdo- or *buzdo- (for the Welsh form fem. *buzdā) < PIE *gu̯osdʰo- 'piece of wood' (IEW: 485), with an analysis of buđđutton as 'penis' rather than 'kiss' (LEIA s.v. bot; Lambert 1994: 123). This etymology of buđđu- is argued for by Watkins 1999: 541–543, who books it as an example for PC /zd/ as a source for tau gallicum, for which evidence is scarce. Watkins' other two examples are Gaulish L-50 neddamon = nets(s)amon < PC *nezd-isamo- 'nearest', which makes necessary an irregular development of */zds/ after the syncope of /i/ to get OIr., W nessamo, and PC *ku̯ezdi- 'piece' with a regular development in OIr. cuit, W peth, and Gaul. *petsi- > *pessi > ultimately ModFr. pièce. Neither of these examples, however, is conclusive. Tau gallicum in Gaul. neddamon is likely the irregular result of the complex cluster with /s/ in the suffix, just like in the Insular Celtic forms, not of /zd/ alone (cf. Jasanoff 1991: 172). For the source of ModFr. pièce, ModIt. pezza, etc., the common reconstruction is Gallo-Rom. *pettia < PC *ku̯ezdi̯ā (FEW vol. 8: 332–342; DLG: 249 f.), indicating a regular development of PC /zd/ > Gaul. /tt/ (Stifter 2011b: 174, n. 20). In the case of PC *bozdo-/*buzdo-, /zd/ > /tt/ may be seen in Gallo-Rom. *bottia as reconstructed by Wartburg (FEW vol. 1: 467–470) for ModFr. bosse 'knob, hump', for which Watkins posits instead Gaul. *bođđā- (a feminine ā-stem just as in MW both). (ModFr. bouton 'button [etc.]' is derived by Wartburg not from a Gallo-Rom. *botto-/*butto-, but from Germanic (PG *buttōn- 'shove'); see FEW vol. 1: 455–465, vol 15,1: 210–229 s.v. *bōtan).
The connection of buđđu-/bussu-/buθ- with the 'penis'-group is also argued on semantic grounds, viz. that not only buđđutton in the notoriously risquè spindle whorl inscriptions, but also bussu- in the compound PNN is better understood as 'penis' than 'lips/kiss' (Watkins 1999: 542), but – apart from the fact that compound names must not necessarily make sense as compound lexemes – there is no case in which one option is incontestably superior: bussumaros 'aux grosses lèvres' (Vendryes 1929: 370; KGP: 158) or 'big dick' (cf. ballomaros, DLG: 65)? bussurigios 'he whose mouth is royal' (KGP: 158), 'smooch king', or 'royal member'? bussugnata 'famed for her lips'/'born with (large) lips' (KGP: 67 f. with n. 4), 'born after a prophecy'/'skilled in prophecy' (Birkhan 1971: 24, n. 3), 'born from/for kisses'/'kiss-skilled' (Meid 2005: 129), or 'penis-sired'/'cock expert'? anbusulus 'lipless' or 'knobless' (DLG: 95)? artebuθz 'lips like a bear' or 'dick like a bear'? Delamarre separates the entries for buđđutton and bussu-, but considers the possibility that they are equivalent and the meanings 'penis' or 'lip/kiss' for both.
As long as the exact status, sound value(s) and dialectal developments of tau gallicum are not cleared up, we cannot exclude that Gaulish/Gallo-Romance forms both with /ss/ and with /tt/ could go back to tau gallicum (cf. what was said about the equation of onomastic bussu- and buttu- above), but the claim that /zd/ yields tau gallicum (and therefore *buzdo- > buđđu- > bussu-) is weak. Also, beside the tau gallicum issue, /u/ in the root is an obstacle. The PC word for 'knob' can be straightforwardly reconstructed as *bozdo-, unless one needs to account for buđđutton and bussu-. Cf. Matasović 2009 s.v. *buzdo-, who explains /u/ by means of an ablauting root noun paradigm which was regularised and thematised; Watkins (p. 542) simply assumes raising. Cf. also McCone 2012, who explains /u/ in the root (and in the stem in the Gaulish forms) by proposing an alternative etymology involving Bartholomae's Law from the PIE root *bʰeṷdʰ- 'awake, become aware (of)' > *bʰudʰ-to- */bʰudᶻdʰo/ 'aroused one, penis' > PC *budᶻ(d)o- > MIr. bot (with lowering of the root vowel; excluding MW both) and > *bʰudʰ-tu- */bʰudᶻdʰu/ 'arousal, erection' > PC *budᶻ(d)u- > Gaul. butˢu-.
To summarise, Gaul. buđđu-, bussu-/buttu- etc. and buθ- are likely to belong together in a group which shows /u/ in the root and stem and tau gallicum in the auslaut; it is etymologically unclear and may belong with the Insular Celtic 'lip'-words. A connection with PC *bozdo- and the Insular Celtic 'penis'-words requires an explanation of the root and stem vocalism and a justification for PC */zd/ as a source for tau gallicum.
The spelling of -buθ-z with zeta instead of expected sigma for -s indicates that the writer attempted to reflect a more complex dental cluster than ss-s, viz. the tau gallicum sound before its simplification; unfortunately, the rationale behind the chosen letters (theta for /d/, zeta for /tˢ/) is not clear and gives no clue about the phonetic shape of tau gallicum in Imperial-age Pannonia.
|AcS||Alfred Holder, Alt-Celtischer Sprachschatz, Leipzig: Teubner 1896-1907. (3 volumes)|
|Birkhan 1971||Helmut Birkhan, "Die "keltischen" Personennamen des boiischen Großsilbers", Die Sprache 17 (1971), 23-33.|
|Delamarre 2007||Xavier Delamarre, Noms de personnes celtiques dans l'épigraphie classique. Nomina Celtica Antiqua Selecta Inscriptionum, Paris: Errance 2007.|
|DLG||Xavier Delamarre, Dictionnaire de la langue gauloise. Une approche linguistique du vieux-celtique continental, 2nd, revised edition, Paris: Errance 2003.|
|Dottin 1920||Georges Dottin, La langue gauloise. grammaire, textes et glossaire [= Collection pour l’étude des antiquités nationales 2], Paris: Klincksieck 1920.|
|FEW||Walther von Wartburg, Französisches etymologisches Wörterbuch, Bâle: 1922–2002.|
|GPN||David Ellis Evans, Gaulish Personal Names. A Study of Continental Celtic Formations., Oxford: Clarendon Press 1967.|