[This article may be freely downloaded, printed out, or distributed. In bibliographical citations, please refer to the URL http://www.humnet.ucla.edu/olat.]

For a pdf version, click here.

Some 'Trivial' remarks on the 'Garigliano bowl' inscription: A response to P. Harvey (JRA 13, 2000)

Brent Vine
(August, 2002)

     In an article published in JRA 13 (2000),¹ Paul B. Harvey, jr. addressed some of the central issues raised by the “Garigliano bowl”, a small impasto bowl with bucchero glaze likely to have been produced around 500 B.C., and found (despite some uncertainty about its provenance) in the vicinity of ancient Minturnae. Harvey paid special attention to the signal feature of this object, namely its two inscribed texts and their interpretation: an external inscription of eight letters consisting of a single name (in an Italic but non-Latin form), and an internal inscription of 44 visible characters in scriptio continua around the rim (in an early form of Latin).
     The first extended study of this monument was published in 1996 by M. Cristofani,² followed shortly by M. Mancini’s analysis of several linguistic problems posed by the inscribed material.³ I myself published a detailed study in 1998,⁴ the results of which were followed in the main by P. Baldi in his handbook of Latin linguistics.⁵ With one exception to be noted below, this exhausts the publication history on this piece up to the time of Harvey’s article.
     One of the salient finds arising from excavations in the region of ancient Minturnae is a sanctuary to the goddess Marica, the earliest stages of which date from the 6th c. B.C. The major thrust of Cristofani’s pioneering study — and in this he was followed by Mancini — was to interpret part of the internal text (specifically, a sequence he read as TRIVOIAI or TRIVOIAD) as referring to a divinity “Trivia” (normally a term associated with Diana), and then to argue, given the existence of the nearby sanctuary, for the association of “Trivia” with Marica. In contrast, my own analysis led to the conclusion that the text contains no reference to “Trivia” at all; thus any association between Marica and the epithet “Trivia” cannot be based on this text. Moreover, my own reading and interpretation of the internal text led to the possibility of actually integrating it with the external (non-Latin) onomastic form, this result having been based, in part, on advances made by Mancini in his interpretation of the external form.
     Harvey pursued two primary goals: (1) to restore the Cristofani/Mancini reading and interpretation of the internal text in terms of a divine epithet “Trivia” (in the process rejecting my arguments contra), and in consequence to explore the possible connection between Marica and “Trivia”; (2) to associate Cristofani’s (but not Mancini’s) interpretation of the external onomastic form with a series of Roman gentilicia (Afidius, Avidius, Afiedius, Aufidius, Aufiedius, Ofdius, Alfidius), and to show (172) that “[w]e ... have a fons et origo for this nomen in its various forms in central rural Italy of Vestine/Sabine origin”. While some of Harvey’s discussions on these points are undeniably interesting, there are important respects in which positions taken by me and by Mancini have been misrepresented (no doubt inadvertently), and there are other respects, primarly concerning language and paleography, where Harvey’s argumentation is demonstrably without foundation or open to serious question.

1. The external inscription: AHUIDIES
     Harvey, following Cristofani (16-20), interprets this sequence as “AFIDIES”, and writes (170): “Cristofani and Mancini presented solid arguments for reading the second and third letters as a digraf [sic]: HV = VH = F”. This statement is doubly problematic: it accurately reflects Cristofani’s view, but not that of Mancini, who rejected this interpretation of HV as a digraph notation for /f/; and the arguments in favor of this interpretation are anything but “solid”, as Mancini showed in his cogent analysis of the problem (13-16). It is unnecessary to recapitulate Mancini’s thorough paleographical critique. To summarize: in confronting Cristofani’s labored attempts to justify the idea that het + upsilon could have spelled /f/, Mancini observes that such a digraph “sarebbe uno hapax assoluto” (13), and concludes (15): “Pur esistendo un Afidius che invoglierebbe a stabilire una confrontabilità con ahuidies l’ammissione di una digrafia <hu> per /f/ implicherebbe sostanzialmente un’ipotesi che verrebbe ad essere ad hoc”. Mancini’s “soluzione decisamente meno costosa” involved an astute comparison with an Oscan gentilicium attested first as ahvdiu (Ve 70; Pompeii, somewhat earlier than 300 B.C.),⁶ already — and predictably — having undergone medial syncope of -i- as compared with Garigliano AHUIDIES, and later as avdiis (Ve 16, Pompeii) and aWdeies (Pocc 154 = Ve 180, Potenza/Rossano di Vaglio). The direct Latin correspondent, then, would be Audius/Audeius. As Mancini observes (16): “Certo è che la sovrapponibilità formale [scil. of these Oscan and Latin forms — add. BV] con la scrizione ahuidies dell’epigrafe del Garigliano ... è pressoché perfetta”. Mancini’s analysis is incontrovertibly superior to that of Cristofani, and I adopted it as a point of departure in my own discussion (257).
     Harvey takes note of Mancini’s comparison,⁷ and even seems to approve of it (“Mancini aptly compared ...”, 171), but does not grasp its significance: if one accepts this comparison, to which Mancini was led via his refutation of Cristofani, <HV> as a digraph notation for /f/ must also be rejected, and with it the comparison with Latin names of the sort Afidius, Aufidius, Aufiedius and the like. What remains, rather, is the comparison with Audius as the form of primary importance for Garigliano AHUIDIES. Thus Harvey’s extended survey (171-3) of the geographical localization of names like Afidius, Aufidius, etc., while certainly of value in its own right, has no bearing on the artifact in question, so long as one accepts Mancini’s persuasive refutation of Cristofani’s /Afidies/ for AHUIDIES.⁸

2. The internal (rim) text
     Harvey’s portrayal of both Cristofani’s and Mancini’s readings of this text is incomplete in two important respects. The sequence of interest here (and for Harvey)⁹ concerns the fifteen letters preceding a small broken area, which Mancini showed to be the end of the text; in Harvey’s presentation (166):

Cristofani:  ... TRIVOIADDEOMDUO[///

Mancini:    ... TRIVOIAIDEOMDUO

As already noted, it is this TRIVOIA- that Cristofani and Mancini (and, following them, Harvey) have interpreted as a form of Lat. Trivia. But Harvey fails to note that Cristofani (11) judged the -D in the sequence TRIVOIAD (“un delta di forma eccezionalmente triangolare”) to be one of only three possible readings, the others being I and U (i.e., upsilon). Nor does Harvey make clear that Cristofani and Mancini both opted, in the end, for (dat. sg.) TRIVOIAI vs. (abl. sg.) TRIVOIAD (Cristofani 22, Mancini 21), and that this choice was governed by interpretive (specifically syntactic) considerations: they inclined toward a restoration ... TRIVOIAIDEOMDUO[NAI = Class. Lat. “Triviae, deorum bonae”, with the dative indicating the divine dedicatee of the object.¹⁰ Among the seven counterarguments I offered against the reading/interpretation TRIVOIAI()DEOM()DUO[NAI was the observation (260) that phraseology of this sort, whether for Trivia, Diana, or any other deity, is highly unnatural and entirely without textual support. Harvey has nothing to say on this point, but he does comment on the dat. sg. TRIVOIAI, to which I objected on other grounds: the format “sum + dative” in tituli loquentes may correspond to a possessive expression (“I belong to X”), as also “sum + genitive”, but almost certainly not a dedicatory one. To this Harvey offers a double response (170): he observes, first, that “to argue thus is to forget that many archaic Latin dedications are unambiguously in the ablative”, apparently suggesting that my objections would not be cogent if one opts for abl. sg. TRIVOIAD. But such an expression is uninterpretable. Indeed, Harvey adduces no example, and I know of none: ablative usages in votive context are restricted to expressions of source referring to the means through which the dedication was offered (such as DE PRAIDAD, CIL I² 48/49; AIRE MOLTATICOD, CIL I² 383), prepositional expressions with PRO stating on behalf of whom the dedication was made (e.g., PRO C. RUTILIO, CIL I² 359), expressions indicating the circumstances calling for the dedication (e.g., VOTO HOC SOLUT[O], CIL I² 1531), or other circumstantial phraseology, partly of disputed interpretation (such as DIOVOS CASTUD, CIL I² 360, cf. DIOVIS CASTUD, CIL I² 361); in no instance known to me does the divine dedicatee appear in the ablative. Harvey’s second response is to assert that “we need not assume that the person who inscribed our text was aware of any rules as to how ‘speaking inscriptions’ should be phrased” — a desperate and methodologically inappropriate gambit, which flies in the face of everything we know about such texts, as made clear in L. Agostiniani’s authoritative treatment of tituli loquentes (a work Harvey cites),¹¹ which demonstrated the pervasively formulaic and conventionalized nature of such texts throughout ancient Italic verbal culture.
     There are also severe difficulties with Harvey’s interpretation of the two-letter sequence -VO- in TRIVOIA-. Cristofani was again forced to claim (22f.) that -VO- reflected a digraph notation, this time for /w/, while Mancini (18) suggested that Lat. via (as in Tri-via) might be reconstructable, in a stage appropriate for this inscription, with an oi-diphthong corresponding to the later -i- (i.e., *woyâ-, in his notation). As I discussed in detail (258f.), the digraph interpretation, which again resorts to “uno hapax assoluto”, is not remotely credible, and the linguistic analysis of Lat. via presupposed by Mancini is deeply flawed. Harvey seems to grant these points,¹² and so offers a third solution, in his bid to salvage TRIVOIA- as an early version of Trivia: he suggests (170) that “the -O- in TRIVOIA is simply an orthographic peculiarity (that is, a mistake)”, noting that “[o]rthographic inconsistencies in vernacular/vulgate inscriptions are far from rare”, and that “[i]ntrusion of unexpected vowels is also not uncommon”.¹³ No one would deny that “vernacular” or “vulgate” inscriptions (and even formal productions) can contain errors; but this is no solution at all: the assumption of error, especially of the “unmotivated” type (and in a text with no other evidence of erasure or correction and, apart from this form, no other recognizable instance of error), is the epigrapher’s (and textual scholar’s) last resort, and cannot be taken seriously as a defense of TRIVOIA- as a way of spelling Trivia. Nor is it even the case that “intrusion of unexpected vowels is ... not uncommon”. Harvey adduces but a single example, which in fact has no bearing on the case: SEIQUOI ‘si cui’¹⁴ in the Sententia Minuciorum (CIL I² 584.44; 117 B.C.), far from reflecting a comparable episodic “intrusion” of <O>, exhibits a well-known and well-attested archaizing spelling of the relative/interrogative pronoun, amply documented in both dative and genitive singular (cf. QUOI and QUOIEI ‘cui’ 21x in the Lex Agraria [CIL I² 585; 111 B.C.]; frequent gen. QUOIUS ‘cuius’; etc.), including literary attestations (e.g., quoi ‘cui’ Plt. Trin. 1126).
     For TRIVOIA-, there remains, as I discussed (259f.), a serious problem with the wau (i.e., digamma) transcribed as -V- by Cristofani, and taken to notate /w/ — this despite the fact that the same inscription also uses upsilon (as in DUO[, allegedly DUO[NAI = /dwonâi/ for Cristofani and Mancini), and that no parallel exists for such a practice. Harvey writes (169) that I “proposed that the apparent wau (digamma) in TRIVOI is in fact a poorly-formed beta”, a statement which is half true: this was only one of two possibilities I recognized (260, 261f.), the other being that the digamma should be taken at face value to reflect F = /f/. But let us consider the beta possibility, since Harvey makes use of it in an attempt to furnish another argument in favor of identifying TRIVOIA- (i.e., TRIBOIA-) with Trivia: “to read TRIB- in place of TRIV- does not perforce expel Trivia from our text: B for V, and vice versa, in vernacular texts is not rare”, citing examples like TRIBIAE, CIL VI 31053. Yet all such examples themselves demonstrate their irrelevance for the problem, since this vernacular feature begins to be attested only in Imperial times (Harvey’s TRIBIAE is from the 2nd c. A.D.!).¹⁶ We are fairly well-informed about Latin and Italic historical phonology in periods close to the presumed date of the Garigliano bowl inscription — approximately half a millennium before the /w/ > /b/ change appears — and no such feature is documented for these periods, either in Latin or any other Italic language. This argument, then, is inadmissible.
     There is, finally, a series of erroneous statements made by Harvey that seriously misreprent my analysis, and which therefore cannot be used to discredit my position. Harvey writes (169): “Vine further proposed that the letter before the delta in TRIVOIAD be read as a sigma, with the consequent word TRIBOS = tribus, thus Sum cum sociis tribus = ‘I am with my three companions ...’ ... We may ... observe of this reading ... that the apparent I between O and D thereby disappears”. In fact, my restoration of TRIBOS (or TRIFOS) = tribus had nothing whatsoever to do with “the letter before the delta in TRIVOIAD”, and I cannot imagine how Harvey drew this conclusion. (I had nothing at all to say about this letter, which is clearly an A, although I did dispute the delta itself, opting for Cristofani’s third alternative U.) I did, however, discuss the second I of TRIVOIA- (260), and it was this letter that should actually, I argued, be interpreted as an S (a so-called “narrow S”, also found in the contemporaneous Corcolle altar fragments [CIL I² 2833a], and which I have discussed elsewhere¹⁷), in view of certain paleographic peculiarities this rather jagged character displays.¹⁸ My conjecture, moreover, has now been seconded by C. de Simone,¹⁹ who arrives at precisely the same conclusion (“la lettera letta come i da Cristofani è chiaramente ed indiscutabilmente un sigma”, 73), and for the same paleographic reasons.²⁰ Although our interpretations of the word-division and meaning differ, the independent convergence of our reading TRIFOS for the sequence in question helps confirm its accuracy — a reading which by itself decisively eliminates “Trivia” from this text. There is, in any case, no question of “the apparent I between O and D ... disappear[ing]” in my analysis, as this was precisely the character whose new reading was central to my interpretation, as for de Simone’s.
     Harvey sought to show (170) that “there is no compelling reason to expunge Trivia from this dedication; reputed findspot and ancient tradition urge that a female deity associated with Marica is appropriate for a dedication at Minturnae”. The second point is not in dispute, and Harvey’s discussion of possible connections between Marica and Diana is interesting in and of itself. But such issues have no relevance for this artifact: all evidence points to the virtually inescapable conclusion that Trivia must indeed be expunged from this text (which is not even demonstrably a “dedication”, sensu stricto). To repeat the conclusion of my 1998 article: “The nature of the find spot is an essential datum for the interpretation of such texts, but must not be allowed to prejudice the details of orthography, paleography, or linguistic history”.

Department of Classics and Program in Indo-European Studies,
University of California, Los Angeles

Notes

1     “The inscribed bowl from the Garigliano (Minturnae): local diversity and Romanization in the 4th c. B.C.,” JRA 13 (2000) 164-174 [hereafter Harvey]. I am grateful to Professor Harvey for his kindness in sending me a copy of his paper (June 2001).
2     “Due teste dell’Italia preromana,” ArchLaz 25 (1996) 8-32 (“Per regna Maricae”) [hereafter Cristofani].
3     “Osservazioni sulla nuova epigrafe del Garigliano,” Biblioteca di ricerche linguistiche e filologiche 42.1 (Opuscula 4.1, 1997) 5-39 [hereafter Mancini].
4     “Remarks on the Archaic Latin ‘Garigliano bowl’ inscription,” ZPE 121 (1998) 257-62. An earlier version of the paper (alluded to by Harvey [165], but not cited) was published informally as “The Archaic Latin ‘Garigliano Bowl’ inscription: a preliminary report,” in the UCLA Friends and Alumni of Indo-European Studies Newsletter 7.1 (1997) 15-19.
5     The foundations of Latin (Berlin 1999) 200-202.
6     Material from Oscan is cited after E. Vetter (“Ve”; Handbuch der italischen Dialekte, Band I. [Heidelberg 1953]) and P. Poccetti (“Pocc”; Nuovi documenti italici a complemento del Manuale di E. Vetter [Pisa 1979]). (The recently-published edition by H. Rix, which has appeared as Band II. of the Handbuch der italischen Dialekte, is not yet available to me.)
7     Though in an incomplete and inaccurate way: Harvey miscites the early Pompeiian form as AHVIV, omits mention of the later Pompeiian form, and adduces the Potenza form as if injecting a novum into the discussion (“We may also note ...”, 171), although Mancini had already treated the form.
8     We may note a further corrective involving the termination -ies, with regard to which Harvey writes (171): “... the termination -ies may be nominative (so Vine assumed, comparing non-Latin Italic texts) or a genitive. Either appears to be possible. A number of similar Etruscan and Etrusco-Italic inscriptions have seemingly genitive formations, others nominative [n.48 omitted, citing several examples — BV]. Latin examples of the (archaic) genitive singular in -ies are reasonably well attested into the 1st-c. B.C. and many instances ... were catalogued by W. Schulze, in his Zur Geschichte lateinischer Eigennamen [n.49 omitted, citing Schulze 116-27 (sic) — BV]”. This is a confused presentation: Harvey does not recognize the fundamental distinction between onomastic -ies (in Latin context also -ieis, and exclusively nominative; cf. my discussion in Studies in Archaic Latin inscriptions [Innsbruck 1993] Ch. 8, esp. 228ff.) and the non-Latin gen. sg. of io-stem nouns, including names. Despite Harvey, the miscited passage in Schulze (properly 116-17) catalogues not a single Latin genitive in -ies, but rather the well-known “Doppelfirma” type with two abbreviated praenomina and nominative gentilicium in -ies (cf. my discussion of this pattern in the work just cited, 218).
9     Omitted from detailed consideration here is the preceding sequence NEIPARIMEDESOMKOMMEOISSOKIOIS, containing the extraordinary form ESOM = Class. Lat. sum; note that the authors of the article cited in this connection by Harvey (167) as “J. D. Brian and R. E. Wallace” (n.14) and “Brian and Wallace” (nn.15, 17) are actually B[rian] D. Joseph and R. E. Wallace.
10     Harvey (167) states that DEOM is “more probably a genitive plural” (rather than masc. acc. sg.), without explaining that this depends crucially on the restoration DUO[NAI, which goes unmentioned. I note in passing that among examples of archaic Latin 2nd declension gen. pl. forms in -om, Harvey cites the “simple manom” of the Duenos vase (I do not understand what he means by “simple”); but this form, in the difficult sequence ENMANOMEINOM, is normally assumed to be acc. sg. (masc. or neut.), and would be gen. pl. only in H. Eichner’s controversial interpretation (“Reklameiamben aus Roms Königszeit,” Die Sprache 34 [1988-90] 207-238), against which I have argued elsewhere (“A note on the Duenos Inscription,” in V. Ivanov and B. Vine (eds.), UCLA Indo-European studies, vol. 1 [Los Angeles 1999] 293-305).
11     Le “iscrizione parlanti” dell’Italia antica (Florence 1982); Harvey 167 n.13.
12     He does not address the digraph argument, but says (169) that “[l]inguists will properly observe that commonly understood rules of vowel formation and deformation do not permit OI < I”, apparently referring to my critique of Mancini’s theory.
13     See, for example, the classic discussion of inscriptional error types in R. Kent, The textual criticism of inscriptions (Philadelphia 1926), esp. 69-76.
14     Miscited by Harvey in the form “SEQUOI”.
15     See any of the standard handbooks of Latin historical grammar, most fully M. Leumann, Lateinische Laut- und Formenlehre (Munich 1977) 476ff. (with the Sententia Minuciorum example cited at 478); similarly A. Ernout, Morphologie historique du latin (Paris 1953) 86f.; A. Sihler, New Comparative grammar of Greek and Latin (New York/Oxford 1995) 387f.; G. Meiser, Historische Laut- und Formenlehre der lateinischen Sprache (Darmstadt 1998) 166; P. Baldi (supra n.5) 347.
16     On the chronology of this process see Leumann (supra n.15) 139.
17     See Vine (supra n.8) 76f.
18     In a footnote (260 n.13), I insisted on the point that Cristofani’s facsimile (reprinted by Mancini) was “quite deceptively inaccurate in its smooth representation of this character”; it is a pity that Harvey’s facsimile (166) — by his admission based on that of Cristofani — reproduces this inaccuracy.
19     “La nuova iscrizione aurunca arcaica e il nome della dea Marica”, Studi classici e orientali 46.1 (1996) 61-92; the paper was apparently unknown to Harvey, and only reached me well after the completion of my work published in 1997/1998 (cf. above, n.4).
20     De Simone’s facsimile of the sequence he reads as TRIFOSAD (Plate X) accurately reproduces this character (cf. above, n.18).