CC BY-NC-ND
31
Book Reviews / Рецензии
Luka Repansek
Department of Comparative and General Linguistics, Faculty of Arts, University of Ljubljana; luka.repansek@ff.uni-lj.si
Blanca María Prosper.
The Indo-European Names of Central Hispania. A Study in Continental Celtic and Latin Word Formation.
Innsbrucker Beiträge zur Sprachwissenschaft (herausgegeben von Wolfgang Meid), Bd. 152.
Innsbruck: Institut für Sprachen und Literaturen der Universität Innsbruck (Bereich Sprachwissenschaft), 2016. 237 pp.
Over the last two decades or so increasing interest has been noted in the study of the onomastic linguistic remains of Indo-European Europe and Asia Minor. Paleolinguistic data of otherwise poorly attested Indo-European languages (or, for that matter, linguistic systems that are known solely through the names and naming systems that have left an imprint on a given Namenlandschaft) has of course always been studied but never has the opportunity been greater to approach this ultimately uncompromising and extremely sensitive set of data with the quickly expanding knowledge that contemporary Indo-European comparative linguistics has to offer. Not only is it now becoming possible to refine and substantiate old etymologies, refute the old ideas and promote alternative, methodologically decidedly less reproachable solutions, or find convincing linguistic explanations for the here-to unetymologisable linguistic data, but also — and this is all the more important — correct the word-formational patterns projected back into the proto-language on the precious evidence of just such fragmentary pieces of evidence, sometimes even contributing to the established set of lexical items reconstructed for the parent language (such opportunities are of course comparatively rare and about ninety per cent of the onomastic material will as a rule be explicable on grounds of what we already know or hold for well-established on the basis of the comparative data (mostly appellative) offered by the Indo-European corpus languages). On a smaller scale, however, the onomastic material of a given linguistic system is able to provide valuable missing puzzles in the understanding of the historical development of a particular language family or one of its individual daughters, especially given the fact that 1) the onomastic systems are repositories of often residual linguistic features, and 2) being essentially generated by the non-onomastic sphere of language use, names are fundamentally words (a fact that is perhaps too often unre-spected) and as such reflect in all the details the pho-
nological, morphological, word-formational, syntactic, and lexical peculiarities of a given language. Names therefore demand a careful and well-balanced etymological explanation that will assign the correct coordinates on all the relevant levels of linguistic expression. One must never neglect the crucially important fact, however, that the onomastic system of a given language, while it does indeed feed upon the appellative sphere of use, obeys its own rules in terms of the productivity on the level of the word-formational patterns and to a minor extent lexis (and rarely even morphology). The generally unavoidable and natural analogical processes will consequently chose different trajectories and affect different segments of language.
B. M. Prósper's monograph is an attempt at extracting as much information as possible from the selected corpus of anthroponymy to, first and foremost, provide the scholarship with a glimpse into the linguistic diversification of North-Western Hispania that is otherwise difficult or, in parts, impossible to track on the basis of the existing epichoric Celtiberian textual documents (mostly, of course, because these are rather earlier than the data embedded in the Latin inscriptions), and, second, to detect potential residual features of Hispano-Celtic that may offer an interesting insight into the word-formational make-up of the older layers of Celtic. The author is to be especially commended on her conscientious application of the premises that were pointed out above, subjecting every studied piece of evidence to multi-faceted etymological analysis. Whenever a particularly tentative suggestion is made to account for a given (mostly sporadic) sound change that otherwise receives no or very little back-up from the immediately relatable language material (due to sheer scarcity of the relevant data), the solution is supported and exemplified by typologically comparable instances from better documented languages, which is an extremely welcome and methodologically indeed necessitated increment. Several of the proposed etymologies are
Journal of Language Relationship • Вопросы языкового родства • 15/2 (2017) • Pp. 136-140 • © The authors, 2017
Blanca Maria Prosper. The Indo-European Names of Central Hispania (2016)
rather convincing and well-grounded in the framework offered in each individual case by the internal and external comparative data. In the majority of cases, of course, even the likeliest interpretation will still remain rather tentative and ultimately purely provisional as is to be expected in any work dealing with etymological onomastics, but a solid starting point is a good stepping stone towards future refinement (this usually becomes possible when more data comes to light or old data receives an updated explanation).
The book is divided into two sizable chapters which contain a condensed and contextualised version of the ideas previously expressed and elaborated by the author in a number of separate studies, supplemented by several novel suggestions and discussions. The Names of the Celtic Cantabri (pp. 11-122) studies the anthroponymical heritage of the Celtic population to the west of the region dominated by the Celtiberian inscriptions, while The Names of Western Celtiberia (pp. 123-198) seeks to pinpoint the individualistic traits and/or dialectal differences potentially mirrored in the personal names of the belt between Burgos, segovia, Soria, Guadalajara, and Cuenca. The second part of the book is organised as a lexicon of alphabetically arranged names that have been conveniently grouped together according to the place of their concentration or, if assignable, the appertaining ethnic (p. 124: Pe-lendones and Turmogi, p. 128: Autrigones, p. 144: Are-vaci, and p. 180: names concentrated around Cuenca). The first chapter is more significantly structured and studies selected personal, and to a lesser extent also ethnic and divine names (most notably Cabuniaeginus and Erudinus, pp. 118-119), on the basis of a particular feature — be it on the level of morphology, word formation, or historical phonology — that the author considers important to isolate and expose to closer scrutiny. In this way the book is able to provide a number of neatly integrated (however miscellaneous) specialist studies on several important aspects of His-pano-Celtic, or generally Celtic, historical development: the (older layers of the) Celtic numeral system (pp. 15-21), the idiosyncrasies of the Celtic comparative and superlative formations (pp. 96-100), the still somewhat problematic question of the specifically Celtic continuation of PIE *k'er-(H2)(-u-) 'horn & c.' (pp. 21-26), some overlooked cases (the autor's choice of the word "neglected" here is perhaps less appropriate) of departicipial formations (pp. 26-33) and obscured compound names (pp. 51-58), the survivors of the PIE category of holokinetic t-stems (pp. 58-65), the history of the verbal adjectives in *-eto- (pp. 71-87) and the surviving instances of possessives containing
the Hoffmann suffix (*-H1/3en-, pp. 87-96), secondary formations based on nasal (passim) and sigmatic stems (pp. 111-115), dissimilation of geminates (such as, e.g., *nn > *nd, pp. 65-71), metathesis in LVP clusters (pp. 101-102), anaptyxis in VRPRV sequences (pp. 102104), and a few scattered bits and pieces of provisional but insightful comments on various heterogeneous problems of historical phonology and/or morphology. Although the title of the monograph indicates that specific problems of Latin word formation will also be addressed somewhat extensively, this is not in fact among the central foci of the monograph. Much Italic comparanda is, admittedly, adduced in support of the author's claims on a particular etymological interpretation, but the book is as much a study of Latin (or Italic, for that matter) word-formational patterns as it is more generally a contribution to the understanding of the somewhat still problematic points of PIE word formation. This is of course the expected side-effect of the study of that side of the language that does not normally take part in the process of reconstruction and may therefore have the noteworthy value of being able to refine or even correct what has been projected back and reconstructed for the parent solely on the basis of the appellative data. What needs to be called to the reader's attention in this respect, however, is the relatively long and extremely interesting and insightful excursus on the history of the type of -ilo- adjectives in Celtic and here, specifically, Latin.
Both chapters conclude with synoptic sections on synchronically productive word-formational patterns (i.e. predominantly suffixes and suffixal chains) of the surveyed names as well as their "phonetic" peculiarities. This, however, is perhaps the most problematic part of the monograph. Even though conditioned sound changes are expected to sporadically occur (a good example is, perhaps, Murce, p. 143, if from *morko-, with conditioned raising of *-o- immediately comparable with the equally sporadic ven. murtuvo.i. < *morto-) in the attested names (when such phenomena are not in fact just a by-product of the rendition of epichoric, native sound sequences in the Latin script), these are more or less as a rule assigned systemic value by Prosper, who tends to take them at face value (occasionally perhaps somewhat too uncritically) and parallels the proposed developments and their results with rather individualist views of the sound changes underwent by the language(s) displayed in the native Celtic text documents conducted in the Iberian (and, to a smaller extent, Latin) script. Too often, perhaps, a particular sound change is proposed to have occurred on the basis of the author's own view of an etymological source behind a name/group of names. Highly
speculative is in my view the proposed reduction *#ueR- > #uR- (pp. 111 and passim), which by the way is an old idea, based on the names such as Vrcaloco, Vurovio and Avlgigun and supported by Clb. urantiom as if from *up-ero-, none of which can be irreproachably claimed to actually contain the addressed sequence (note that the etymological connection between Lusitanian Uramus and Clb. Veramos cannot be proved in any significant way). It is a staple fact of historical comparative linguistics that etymology of a given word in any given language is the bedrock foundation upon which a set of regular (and conditioned) sound-changes can be observed and established (combinatorially, of course, and using forward as well as backward reconstruction). This customary, although demanding procedure logically receives a methodological caveat: the etymological connection has to be irrefutable for the results to obtain. There is an immediate problem with the onomastic data, however. Regardless of the progress made in the direction of successfully approaching the fragmentary linguistic evidence, names still often prove to be ultimately difficult if not momentarily impossible to subject to exact interpretation, which is simply due to their general opacity, brought about by too many points of contact with potentially promising formal correspondences and simultaneous lack of purely synchronic transparency (this is more often than not a rule for onomastic languages but not uncommon in the case of fragmen-tarily or otherwise poorly attested systems). Whenever an individually observed sound change is supported by a comparandum with ultimately uncertain etymology, the reader should have been warned that the author is basing her views on her own individual interpretation of a particular piece of data and not in fact on a substantiated piece of evidence as seems to be the impression. Several of author's points on such proposed phonological developments should therefore perhaps be understood as very tentative and provisional. I remain very sceptical towards several of the suggestions, especially towards the proposed voicing of stops preceded by nasals (cf. the Old Irish type *-ant- > *snd- > *-ld-) in the likes of Pi(n)ganco, Le-tondo, Plandica etc. (pp. 185-190 and passim). If voicing were a late systemic sound-change, it should affect all instances of such sequences, which it clearly does not, exempting the -nt-participles (unless, as it is argued, obscured participial formations) and the productive suffix *-Vnko-. As far as I can see, there is not a single incontestable and unambiguous case of a -nP-sequence in the material adduced in favour of the sound law and neither would I be too eager to recognise the numeral five in the likes of Pi(n)ganco & co. That such
regular voicing would be hindered by the "palatalising effect" of the following *i (as, interparadigmati-cally, in stenionte and gente, as is suggested) and that names built around *arganto- that never show voicing do not in fact go back to the commonly accepted the-matisation of the present participle seems like special pleading. In light of the unproblematic fact that at least Celtiberian attests to the process of phonetic leni-tion of voiced stops I wonder if the (surely telling) spelling of etymological medial *-g- in Dahae and perhaps Saihli as >h< does not rather simply encode the voiced velar fricative rather than its secondary devoic-ing as suggested by Prosper (pp. 139, 184). A rather strong case is also made in favour of gemination as a direct systemic consequence of a phonological process. This is likely in case of *-Ri- clusters, where it is even typologically expected, even though the data forces one to simultaneously accept the somewhat suspicious (because strangely sporadic) accompanying glide absorption (note, however, the potentially interesting case of subsequent dissimilation to -rd- < *-rr- < *-ri-, pp. 70-71, 120), but I cannot see a convincing reason to favour gemination as a purely phonological process over hypocoristic gemination in cases such as Accua, Pedaccianus, Boddi etc. It is moreover rather difficult to accept the idea behind the proposed development of *-Vpn- > *-Vtyun- > *-Vfiun- (pp. 105, 118), since PCelt. is otherwise never voiced intervo-calically and is normally lost without trace. I am undecided on the late change of the inherited voiceless labiovelar into *p (passim) as potentially mirrored by Petraioci, Pentius & co. (the textual documents of course clearly attest to the preservation of *b>). These names are extremely likely to go back to the obvious numerals that they contain in the derivational base, but how sure can one be that they are ultimately Celtic? Note that the divine name Vailico ~ Vaelico (p. 182) is said to preserve the otherwise regularly monophthongised inherited diphthong *ai on account of its being of onomatopoeic origin (cf. OIr. fael 'wolf'). I cannot see, however, how an inherited lexical item, regardless of its etymological source, would be able to resist a regular sound change. It remains unclear what the author's views are on the probable simplification of *-xt- cluster in the seemingly popular name built on Ambato- < PCelt. *amb-axto-. On p. 125 it is described as regular and expected, whereas in ft. 65 (p. 73) the development is said to be surprising. The supposed metathesis in Crastunon- (p. 160) is despite a good appertaining discussion left unaccounted for in the end.
There are an additional few minorly problematic points I would like to draw attention to. The PN Carauanca, if it is indeed related to the PIE word for
Blanca María Prosper. The Indo-European Names of Central Hispania (2016)
'horn' & c. (p. 23ff.), which seems more than likely, could equally well reflect a possessive *-uo- derivative, so *k'er=H2-uo-, cf. Gr. Kepaog < *k'er=H2=s-(u)o-. The mountain ridge KapouayKa(g) is certainly non Celtic (p. 25). I have recently explained it (Repansek 2016c: 187-188) as reflecting *(s)kor=un-ko- (cf. OIr. lie < *leH2=un-ko-) to *(s)ker- 'split' (for the secondary semantic shift towards a nomen rei actae cf. PSl. *skala 'rock'). Balto-Slavic *karua (p. 25-26) is undoubtedly a vrddhi formation (and as such a formal substantivisation of the underlying possessive adjective) but only in as much as it copies the naturally co-occurring metatony in the inherited type (cf. the Slavic type *uydra to *ud=r-o-); *korua is therefore an unjustified projection and does not as such "fail to account for the Celtiberian form" (p. 26). Hittite makkiess- 'become big' can hardly be convincingly traced back to *meg=H2-eH1sH1- (p. 27) — a projection that strives to account for the exclusive geminate spelling of the intervocalic -ss-. Such fientives are synchronically most probably based on the established model *palh-ess- (adj.) (cf. *palh=ess-ar/n-) *palh-ess- (fient.) and ultimately reflect simple conversions. In terms of word formation, the type continued by Latin senesce/o- etc. < *-e-H1- + *-sk'e/o- (ibid.) is of course completely unrelated. Vedic mahi-ya-te 'to be/feel big' is a deadjectival denominative verb and as such goes back directly to a straightforward *meg=H2-ie/o- (with regular and mor-phophonetically conditioned lengthening of the reflex of schwa primum before the suffix) rather than indirectly reflect an "older *-ei/i-" (p. 35). There is absolutely no reason to uphold Hamp's view that the PN Brigetio is of deverbative origin (p. 52; see Repansek 2016a: 248). The sequence *-gnn-iio- < *-gnH1-iio-would certainly not have had a different outcome (purportedly *-gniio- > *-ganiio-, p. 54) than the ubiquitous type *-gnn-o- (< *-gnH1-o-) > *-gn-o- in Italo-Celtic; an inherited *-gnH1-io- (cf. OIr. buachaill < *-io-) that would preserve the laryngeal intact, on the other hand, would indeed produce PCelt. *-ganio- (via laryngeal loss by what is descriptively known as Pinault's rule), logically matching the simplex. The PN Adnamatia in Pannonia is formally a substantivised adjective of appurtenance to Adnamato- and could under no etymological approach to the root in question come to mean "the frightened city" (p. 85). Incidentally, the PN Adnomatus from Ig (sic!) should be properly said to indirectly reflect the length of the *a in its Gaulish donor, given that this was phonetically most probably realised as a low rounded */t>/, cf. such spellings as Gaul. Blotu-rix for *blatu- (see Sims-Williams 2003: 56). I do not share Olsen's views on the origin of the Hoffmann suffix and I do not find the proposed se-
mantic relationship between the derived and un-derived versions at all convincing (pp. 87-96) — we must rather simply be dealing with a complex suffix with the basic function of deriving from the nominal base a possessive adjective (liable to subsequent formal substantivisation). The alleged cases of *-H1/3n-o-should, however, probably be segmented differently (specifically *-H1-no-), as has already been proposed. I am cautious to accept admittedly interesting cases of *-ano- as reflecting the old, basically unshortened version of *-o-H1/3n-o-, because this seems to significantly complicate the traditional (and in my view rather convincing) explanation for the "normal" and ubiquitous type in *-ono-, especially since cases such as Gaul. Toutanno- could easily be secondarily built on the inherited a-stem and thus represent a younger parallel to the inherited Toutono- < *-o-H1/3no- (with regular laryngeal loss by Dybo's law) ^ *-e-H2-. I am not convinced that pairs such as Aiu (PN) vs. Aiankum (family name) can in fact reflect an old relationship -o(n) : *-n-ko-, since Aiu is cleary an u-stem, cf. the Gaul. hypocoristic Aiiuca (see Meid 2005: 213). Latin patronus and its oppositional derivative matrona (p. 92) are almost definitely not old inherited formations, neither is the apparent thematic base of Av. visan- (as per Olsen 2010: 160-161), which simply copies the model established by the predominance of the puBran- type. Cormerton-, if it indeed goes back to *kom-merton-, is hardly a case of a Hoffmann-derivative, *merto-n-(most notably in Av. mar'tan-, which only means 'mortal'; there is no conclusive piece of evidence that would point to a homonymous mar'tan- with the meaning of 'chief of men' in the Ga-Sas) being a clear case of an individualisation. Note that the divine name Vidasus is certainly Pannonian rather than Celtic (p. 113), i.e. Gaulish. PCelt. *ulkuo- for PIE *ulkuo- 'wolf' (OIr. olc, perhaps = Lepontic Ulkos) is in my view a case of resyllabification rather than a final stage of the proposed developmental stage *uulkuo- (p. 115), cf. Old Albanian ulk and Pannonian *ulko- (in Ulcisia), going towards the same end as *luk»o- with full metathesis of the *dakru- type. Consequently, I find it extremely unlikely that the PN Vlibagi could conceal the expected PCelt. reflex *ulikuo- (> *ulipo- > *ulibo-). The PN Voltisemae should not be simply called Italic (p. 154, ft. 125) as the relationship between the reflex of the sequence *-mH2o- that the name attests to, namely *-am- (with expected, even though sporadically marked vowel weakening in an unaccented syllable), exactly matching the sequence -am- >am<, >em< amply attested in Ig (there the PN Decomon- is not autochthonous), vs. PItal. *-om-, for which consider Ven. dekomo- '10th', points to the fact that things are
significantly more complicated. The phenomenon actually seems to reflect an important isogloss that brings in further (and rather welcome) internal diversification within the Northern-Adriatic language continuum (see Repansek 2016b: 337 and a much updated view in id. 2017). I have trouble accepting the claim that the data seems to point towards the "reconstruction of a single Celtic and Italic Suffix -ed(i)io-" (p. 164); this would leave *-o-dio- (the latter morphemic segmentation is dictated by deadverbials such as Gaul. *uxsedio- & co., in my view also by the PN Remetodia < *-eto-dio-, for which consider ouep-exo-^ape[o]uL), the widespread variant found in Gaulish (matching OIr. -(ai)de, W -eid), completely unaccounted for. OIr. buachaill, MW bugeil do not represent transfer forms to the i-stems (p. 166), but regularly and un-problematically reflect old, inherited agent nouns in plain monosyllabic *-io- (cf. Uhlich 1993). Gaul. neddamo- (*/ts/) = OIr. nessam reflect the expected deadverbial superlative *nesd-tmH2o- (cf. Indo-Iranian *nazd-) rather than *ned(h)-to- or *ned-samo- (p. 171 with ft. 135).
There are very few typographical errors. I notice for for form (p. 98), already (p. 99), postdating (p. 119 under 8.), the a (p. 125 s.v. *argamo-), means for menas (p. 130), a dot instead of a comma before It... (p. 136), a missing on (p. 143 s.v. ?morko-), and (p. 146 s.v. *kouno-), the adjective unknown on p. 170 is likely to be unsuitable (does the author mean "unclear"?). The author's English is generally very good, but several non-nativisms occur throughout the text. This may occasionally pose a problem in as much as it can at times, although very rarely, obscure the idea behind the formulation to the point that it is rather difficult to be sure what exactly the author is trying to convey to the reader. The Table of Contents is not entirely synchronised with the actual pagination and "1. Introduction" in the head of p. 117 is misplaced. The monograph is equipped with a comprehensive and generous (in particular by as-
signing the individual entries a linguistic affiliation) index (pp. 219-237), that leads the reader to the ono-mastic and the appellative language material (be it reconstructed or factual). One perhaps misses more elucidation on different sets of exposed phenomena (mostly of phonological nature) that the author has dealt with at length elsewhere. At least the main points of argumentation should be given at the relevant sections.
In summary this is a fine and very capable addition to paleohispanic linguistics, comparative philology of Celtic languages (contributing importantly to every level of linguistic expression), and a refinement of several difficult aspects of the comparative grammar of Indo-European languages in general.
References
Meid, Wolfgang. 2005. Keltische Personennamen in Pannonien. Budapest: Archaeolingua.
Olsen, Birgit Anette. 2010. Derivation and Composition. Two Studies in Indo-European Word Formation. Innsbruck: Institut für Sprachen und Literaturen der Universität Innsbruck.
Repansek, Luka. 2016a. Review of Alexander Falileyev: In Search of the Eastern Celts. Journal of Celtic Linguistics 17: 239-280.
Repansek, Luka. 2016b. Quiemonis and the epichoric an-throponymy of Ig. Arheoloski vestnik 67: 321-257.
Repansek, Luka. 2016c. Keltska dediscina v toponimiji jugo-vzhodnega alpskega prostora. Ljubljana: ZRC SAZU, Zalozba ZRC.
Repansek, Luka. 2017. New Onomastic Evidence from Ig. Historische Sprachforschung 130, forthcoming.
Sims-Williams, Patrick. 2003. The Celtic Inscriptions of Britain: Phonology and Chronology, c. 400-1200. London: Blackwell.
Uhlich, Jürgen. 2003. Die Reflexe der keltischen Suffix-Varianten *-io- und *-iio- im Altirischen. In: Rockel, M., Zimmer, S. (Hrsg.), Akten des ersten Symposiums deutschsprachiger Keltologen, Gosen bei Berlin, 8.-10. April 1992: 353-370. Tübingen: Max Niemeyer Verlag.
