Abstract
This paper examines the aberrant usages of grapheme <Y> in the Latin inscriptions of the province Hispania. The analyzed corpus includes cases in which both upsilon is replaced by means of <I>, <V> or <VI> and cases in which, conversely, <Y> replaces <I>, <V> or other spellings within either Hellenic or non-Hellenic words. The situation attested in the Iberian Peninsula is in line with the general situation of the Roman Empire, with Hispania belonging to the group of provinces in which I-spellings prevailed over V-spellings already in the Early Period (1st–3rd c. AD).
Prof. J. L. Moralejo, linguae Latinae peritissimo, octogesimo die natali donum dedi.
1 Introduction
Aberrant spellings involving the phoneme /y/ are widely attested in the Latin inscriptions of the Roman Empire. These spellings basically comprise (1) the substitution of <Y> by means of another grapheme (mostly <I> or <V>) and (2) the unexpected presence of <Y> within either Hellenic or non-Hellenic words (including other foreignisms and Latin terms). On occasions, (3) other types of misspellings are attested instead of <Y> and vice versa.
Whereas the general trends in the use of the letter upsilon in the epigraphic sources across the Roman Empire are currently being examined by myself in a separate study,1 this paper pays specific attention to the situation attested in the inscriptions from Hispania.2 Our aim is to analyze the different ways of transcribing and using upsilon in this province in light of the data collected up to this point in the Computerized Historical Linguistic Database of Latin Inscriptions of the Imperial Age.3
2 <I> or <V> replacing <Y>
There are currently 187 LLDB-data forms referring to the replacement of <Y> by means of <I> or <V> in the epigraphic corpus of Hispania.4 However, a large disproportion between these groups can be observed, as the ratio is 65% I-spellings (= 121 data)5 versus 35% V-spellings (= 66 data)6 (Graphic 1).
I-spellings vs V-spellings in the Hispanic epigraphic corpus
Citation: Acta Antiqua Academiae Scientiarum Hungaricae 62, 4; 10.1556/068.2023.00064
Furthermore, a radical inverse trend in the incidence of V-spellings can be observed from a diachronic perspective, if one compares the percentage of data referring to the Early Period (ranging from ca. 1st c. to 3rd c.) with the percentage of data referring to the Late Period (ranging from ca. 4th c. to 8th c.). Over time, V-spellings decline from 41% to 13%, while I-spellings proportionally increase from 59% to 87% (Graphics 2a and b).7
A) I- vs V-spellings in the Early Period (56% = 80 data; 44% = 55 data); B) I- vs V-spellings in the Late Period (85% = 34 data; 15% = 5 data)
Citation: Acta Antiqua Academiae Scientiarum Hungaricae 62, 4; 10.1556/068.2023.00064
This situation confirms a general trend in Latin epigraphy which has also emerged through other studies conducted on a larger scale.8 It clearly points towards a widespread [i] pronunciation of upsilon by average Latin speakers, which is also endorsed by the results in the Romance languages,9 including the Ibero-Romance varieties.10
In general terms, however, <Y> appears to have been correctly deployed in the Latin inscriptions of Hispania. A preliminary search reveals that there are currently 764 occurrences in EDCS11 showing the correct usage of grapheme <Y> for the transcription of Hellenic loans or other foreignisms that entered into Latin via mediation of the Greek language. This corresponds to 77% of the total occurrences of upsilon in the Hispanic epigraphic corpus (Graphic 3).12 Such ‘regular’ spellings reproduced the cultivated pronunciation of upsilon, [ü], or may simply reflect the writers' knowledge of the ‘orthographic rules’.13 This last hypothesis appears to be plausible, considering that /ü/ was absent from the phonological system of the Latin language,14 and most probably ‘failed’ to take root as an imported phoneme,15 as is clearly evidenced by the relatively frequent I-spellings mentioned above.16
An orthographic overview (77% = 764 data; 12% = 121 data; 7% = 66 data; 4% = 41 data)16
Citation: Acta Antiqua Academiae Scientiarum Hungaricae 62, 4; 10.1556/068.2023.00064
In contrast, V-spellings almost certainly reflected the original [u] pronunciation of Greek upsilon,17 that was adopted by the Romans at the time of their first contact with the Hellenophone populations settled in Magna Graecia.18 Whereas the value of upsilon in the Greek koine became [ü],19 V-spellings were probably retained in Latin as conservative variants (archaisms).20 Meanwhile, during the 1st c. BC, the ‘new’ pronunciation of upsilon ([ü]) became prominent among cultivated Latin speakers (its popular counterpart being [i]),21 and began to be represented in writing by means of the letter <Y>.22
The conservative value of the V-variants seems to be suggested by the almost total prevalence of proper names (typically, a conservative sector of the lexicon) in the group of terms bearing the V-spellings23 and – no less importantly – by the already mentioned decline of such spellings from a diachronic perspective (Graphics 2a and b).24 In fact, there is currently only one exception to proper names in Hispania,25 namely the spelling MARTVRI instead of martyris, that appears on the epitaph of a Christian devotee in Olisipo (Lusitania).26 Although martyr is a typical Christian term – often also used as an anthroponym27 – whose popular [u] pronunciation is proven by several sources, the V-variant MARTVR could have a morphological explanation.28 Consequently, a spelling such as MARTVR × martyr does not in itself prove the existence of a generalized [u] pronunciation of upsilon in Latin, especially in the Late Period, when it is attested.
When considering the possible conservative character of the V-spellings, it should also be observed that the territorial distribution of these forms in Early Hispania seems to indicate the existence of small clusters in specific (mainly old-foundational) towns, such as Emporiae (a settlement of a Greek origin),29 Tarraco, Saguntum, Corduba, Baelo, Olisipo, or Augusta Emerita (Graphics 4a and b). In contrast, I-spellings appear to have been more generalized already in the earliest linguistic phase, including some clusters in the central and north-western parts of the Peninsula (Graphics 5a and b).30
A) A geographic distribution of V-spellings in the Early Period (1 or more data per spot); B) A geographic distribution of V-spellings in the Early Period (at least 2 data per spot)
Citation: Acta Antiqua Academiae Scientiarum Hungaricae 62, 4; 10.1556/068.2023.00064
A) A geographic distribution of I-spellings in the Early Period (1 or more data per spot); B) A geographic distribution of I-spellings in the Early Period (at least 2 data per spot)
Citation: Acta Antiqua Academiae Scientiarum Hungaricae 62, 4; 10.1556/068.2023.00064
All this speaks in favor of the retention of V-spellings based on a pure graphic tradition or perhaps – considering the overall limited number of testimonies – as a result of purely accidental circumstances. Be that as it may, it should be noted that previous investigations comparing the situation attested in the different territories of the Roman Empire have allowed Hispania to be included within the group of provinces where I-spellings prevailed not only in the Late Period (as it happens almost everywhere) but also in the Early Period, in opposition to other provinces, such as Central-Southern Italy or the Eastern part of the Empire, where V-spellings seem to have been dominant in the earliest linguistic phase, as a result of a consolidated Hellenic tradition in the area.31
In Hispania, the widespread [i] pronunciation of upsilon is further confirmed by the co-presence, besides personal names, of other classes of terms bearing I-spellings. In support we can quote not only martyr,32 which is also attested with V-spelling, as we have already noted, but also: collyrium,33 gymnasium,34 lympha,35 or presbyter.36 The phonetic value of these spellings is indirectly proven by other (currently 5) forms – omitted from our previous calculations for methodological reasons – showing the typical Vulgar Latin phenomenon of the merging of i with e,37 such as: ALEPIVS × Alypius,38 CEPRIANVS × Cyprianus,39 MARTERIS × martyris,40 MERSINAE × Myrsinae,41 MERTILLIANE × Myrtilianae,42 SEMPERVSA × Sympherusa.43 In this regard, it also seems noteworthy that, in our province, there is only very scarce evidence for the merging of the back vowels:44 one example is found precisely in the spelling MARTORIA × Martyria,45 which provides a clear hint of the [u] pronunciation of upsilon in the form MARTVR, to which it was referred to above; another example is the form EVRODICE × Eurydice,46 also known outside of Hispania,47 that implies the existence of a variant, EVRVDICE,48 which might be explained with vocalic assimilation.49 Another possibility is that the name Eurydice entered into the Latin language at a very early linguistic phase or through a dialect that retained the [u] value of upsilon,50 and therefore existed in both the V- (conservative) and the Y- (classical) variants. However, the form EVRODICE might also be traced back to an improper association of this name with other Hellenic anthroponyms, such as Europa, Europia, Eurota, etc.51
3 <Y> replacing <I> or <V>
The Hispanic epigraphic corpus includes examples of the aberrant usage of <Y> in place of <I> or <V> (currently 19 instances). This phenomenon, which is also reported in other territories of the Roman Empire,52 usually takes place in Greek terms, although it can also occur in non-Hellenic words (Latin terms or other foreignisms).
As for the spellings showing the replacement of <I> by means of <Y> within Hellenisms (12 instances), these can be understood as orthographic hypercorrections, as they confirm e contrario the widespread [i] pronunciation of upsilon in Vulgar Latin. A particularly eloquent case are specific forms – such as HYPOLITVS × Hippolytus53 or EVTICHYA × Eutychia –54 in which both confusions (I × Y and Y × I) are attested at once.55
In contrast, the replacement of <I> by means of <Y> that takes place within non-Hellenic words (7 instances) might invite a real phonetic explanation,56 as it tends to appear always within similar phonetic contexts.57 In Hispania, the following cases can be quoted: [F]ELYX × Felix,58 LYBENS × libens,59 TYBERIVS × Tiberius,60 and SECVNDYTIVS × Secunditius.61 Conversely, a misspelling such as HABITET Y SECVLO × habitet in saeculo62 seems to depend on a copyist's error.63
The letter upsilon can also occur in place of Latin <V> (4 instances). Among Hellenisms, we find the form EYTIXO × Euthycho64 scratched on a ceramic fragment from Pompaelo (Hispania Citerior), which is a possible example of alphabet-switching, considering that Greek chi was also deployed here instead of Latin <CH>.65 On the other hand, the spelling NEREYS × Nereus66 seems to preserve the same ending as the original Hellenic form, Νηρεύς.67 Of a mere graphic (maybe ornamental) nature could also be the replacement of <V> by means of <Y> in the Latin name FAYSTINAE × Faustinae.68 In fact, it should be remembered that, even in Attic, the value of upsilon in these diphthongs was [u].69 Consequently, such spellings cannot have any phonetic explanation.70 Other spellings, such as CATTYGAE × Cattugae71 or LYSITANI[AE] × Lusitaniae,72 show a similar use of upsilon in non-Hellenic words.
4 <Y> replacing other spellings/other spellings replacing <Y>
An inscription from Tarraco bears the form MYS(IAE) × Moesiae.73 This is a recurrent misspelling,74 which should probably be traced back to the co-existence of an Eastern-European province called Moesia and an Anatolian region called Mysia, whose names might be confused for each other.75 Moreover, the shifting of the Greek diphthong /oi/ into /y/, documented in the popular language already in the classical period,76 could explain the replacement of <Y> by means of <OI> in the form COIMOIHO[E?] × Cymothoe,77 also included in our corpus.
On a few occasions (5 instances), upsilon is transcribed by means of another sequence of graphemes, namely <VI>. This is a largely reported misspelling,78 which typically occurs when upsilon is placed after a voiceless velar stop, /k/. In this context, the consonant seems to have acquired the labial trait of /ü/, converting itself into /kw/.79 In Hispania, one finds for instance the forms: PILOQVIRIVS × Philocyrius,80 IN QVISICO × in Cyzico,81 QVIRICI × Cyrici,82 and QVINIGIAE × Cynegiae.83
In the Iberian Peninsula one also finds the spelling SVIRA × Syra which does not match the already quoted phonetic context at all;84 however, this form has been interpreted as a technical error for Severa.85 Again, we also find the misspelling PYIRIPINVS × Pyripinus,86 which might be ascribed to another internal evolution of the Greek language, that led to the merging of /yi/ with /y/.87 In contrast, a misspelling such as SXYVSTVS × Xystus88 apparently reflects the writer's hesitation between the correct Y-spelling and the widespread V-variant of this same name.89 Last but not least, there is also a form SYAGRYS × Syagrius,90 which could be a mere orthographic misspelling.91
5 Conclusions
The analysis of the aberrant usages or replacements of the letter upsilon in the epigraphic corpus of Hispania allows us to draw the following conclusions:
Correct spellings reflecting (at least in theory) the [ü] pronunciation of upsilon (<Y>) make up by far the majority of instances;
The widespread Vulgar Latin pronunciation of upsilon was [i], as is reflected by the prevalence of I-spellings over V-spellings, and by later developments in Ibero-Romance languages;
V-variants were archaizing, and their use decreased over the centuries;
Less frequently attested aberrant spellings of various types can be explained as graphic issues, while they might also reflect occasional mispronunciations occurred in Latin or internal developments of the Greek language.
The situation described above is in line with the results of the investigation conducted on the epigraphic corpus across the whole Roman Empire, according to which Hispania belongs to the group of provinces where I-spellings always prevailed over V-spellings, probably as an effect of a non-consolidated Hellenic presence in the area.
Acknowledgment
This paper is part of the RYC2021-030987-I contract, funded by MCIN/AEI/10.13039/501100011033 and by the European Union «NextGeneratioEU/PRTR». It was also written in the frame of the project ‘Computerized Historical Linguistic Database of Latin Inscriptions of the Imperial Age’, supported by the Hungarian National Research, Development and Innovation Office NKFIH (former Hungarian Scientific Research Fund OTKA; no. K 135359 2020–2024), and in the frame of the research group LITTERA (Laboratori per a la investigació i tractament de textos epigràfics, romans i antics) (2021SGR00041).
Tantimonaco, S.: Remarks on the spelling and pronunciation of upsilon in Latin (in preparation). Previous studies focusing on the use of upsilon in the Latin inscriptions are: Moralejo, J. L.: Notas sobre la grafía Y en inscripciones latinas. CFC(L) 4 (1972) 165–186 (https://revistas.ucm.es/index.php/CFCA/article/view/CFCA7272220165A) and Purnelle, G.: Les usages des graveurs dans la notations d'upsilon et des phonèmes aspirés : le cas des anthroponymes grecs dans les inscriptions latines de Rome. Genève 1995. See also Biville, F.: Les emprunts du latin au grec : approche phonétique. Louvain 1991, II 255–319, with earlier bibliography.
An antecedent on this topic is García González, J. J.: La notación de ípsilon en las inscripciones de CIL II2. In Wright, R. (ed.): Latin vulgaire – latin tardif VIII. Actes du VIIIe colloque international sur le latin vulgaire et tardif. Oxford, 6–9 septembre 2006. Hildesheim – Zürich – New York 2008, 78–83. However, this study is limited to the materials of the conventus Cordubensis, conventus Astigitanus and the southern part of the conventus Tarraconensis.
http://lldb.elte.hu/en/database/. The bibliographic abbreviations used in this work can be consulted in the LLDB webpage (‘Abbreviations for Corpora’).
Last search: September 2022.
y > I: LLDB-14948; 14949; 16205; 16409; 17415; 18248; 18380; 18383; 20118; 21532; 21765; 22292; 22873; 24182; 24257; 24898; 25192; 26038; 27476; 27604; 27610; 27653; 27656; 28180; 28295; 28312; 28537; 29687; 29721; 29823; 30099; 30134; 30135; 30903; 31763; 31764; 31778; 32097; 32302; 32803; 32907; 32908; 32910; 35023; 35343; 35348; 36690; 42620; 44195; 46292; 46308; 48197; 52785; 59893; 60903; 63402; 63436; 63445; 63448; 63503; 63531; 63631; 64808; 65250; 71962; 79389; 79475; 80155; 80193; 80444; 80481; 81800; 83370; 83469; 84591; 85289; 85303; 87385; 87656; 87740; 93987; 102942; 104125; 110464; 113836; 121898; 121935; 122271; 122299; 122535; 122536; 125559; 125563; 140283; 140285; 140286; 140289; 140408; 140412; 140583; 140584; 140585; 140592; 140596; 140601; 140608; 140609; 140610; 140611; 140612; 140613; 140619; 140630; 140631; 140632; 140635; 140641; 140642; 140645; 140647; 140651.
y > V: LLDB-15604; 20658; 21890; 22205; 22284; 23948; 24007; 24207; 25952; 25953; 26192; 26339; 27649; 28233; 28323; 28325; 28871; 29698; 30130; 30131; 30133; 30341; 32278; 34538; 35022; 44451; 44461; 64925; 74256; 74303; 75871; 80474; 80992; 83364; 87048; 87376; 102662; 102966; 103161; 119266; 122117; 122122; 140400; 140401; 140402; 140403; 140405; 140406; 140587; 140591; 140597; 140598; 140599; 140600; 140602; 140603; 140615; 140628; 140629; 140636; 140637; 140638; 140639; 140640; 140646; 140652.
Non-dated records have been automatically excluded from these graphics.
See Purnelle (n. 1) 319; Tantimonaco (n. 1).
See Moralejo (n. 1) 170.
See Bergua Cavero, J.: Los helenismos del español. Historia y sistema. Madrid 2004, 27–28.
Epigraphische Datenbank Clauss-Slaby (db.edcs.eu). Last search: 22/09/2022. Data have been identified by searching for all the instances of Y in the provinces Baetica, Hispania Citerior and Lusitania (field ‘Search Text 1’ + field ‘Province’). This search provided 929 total items, which have then been analyzed one by one.
Correct and incorrect usages of upsilon can appear at the same time. On occasions, the same word is repeated more than once within the same text, sometimes with correct and sometimes with incorrect spelling, e.g. NEFAE 46B (Augusta Emerita, Lusitania, AD 69–96): L(ucius) Scant[ius]/Acutu[s]/Scantia [L(uci) l(iberta)?]/Melyboe[a]/h(ic) s(iti) s(unt) s(it) v(obis) t(erra) l(evis)/Ianuaria Scan/tiae Meliboeae (!) f(ilia)/Urbanus Scanti/Acuti f(ilius)//h(ic) s(iti) s(unt) s(it) v(obis) t(erra) l(evis).
Although a stable orthography of the Latin language was never developed in Rome, ancient grammarians showed an interest in the correct spelling of certain words. On this topic, see Desbordes, F.: Idées romaines sur l’écriture. Lille 1990 and De Paolis, P.: Recta scriptura e recte scribendi scientia: l'ortografia latina in epoca imperiale fra prassi e insegnamento. In Biddau, F. (ed.): Die geheimen Mächte hinter der Rechtschreibung. Erfahrungen im Vergleich. Akten der Internationalen Tagung (Mainz, 28-29.02.2012)/L'ortografia e i suoi poteri forti. Esperienze a confronto. Atti del Convegno Internazionale (Magonza, 28-29.02.2012), Frankfurt am Main – Berlin – Bern – Bruxelles – New York – Oxford – Wien 2013, 35–53.
On this topic, see Suárez-Martínez, P. M.: Le ‘medius sonus’ latin. Glotta 92 (2016) 227–236 (http://www.jstor.org/stable/24891272); Suárez-Martínez, P. M.: Más sobre el medius sonus y la letra ⱶ de Claudio. In Unceta Gómez, L. et al. (eds): Amice benigneque honorem nostrum habet. Estudios lingüísticos en homenaje al Profesor Benjamín García-Hernández. Madrid 2021, 65–74.
See Moralejo (n. 1) 170, 185.
The group ‘other misspellings’ includes the examples where <Y> was replaced by either <E> or <O> (see below, the last paragraph of this chapter) as well as examples of <Y> replacing other spellings, and other spellings replacing <Y> (see below, chapter 3).
See Allen, W. S.: Vox Graeca. A Guide to the Pronunciation of Classical Greek, Cambridge – New York – Port Chester – Melbourne – Sydney 19873, 65–66.
See Moralejo (n. 1) 168.
The [u] value of upsilon was instead preserved in non-Attic dialects, such as Boeotian or Laconian, see Allen (n. 17) 69.
The phonetic value of V-spellings should not be questioned in the case of the earliest Greek loans, which preserved the original [u] pronunciation over the centuries, as can be clearly evidenced by an example such as purpura (<πορφύρα), developed into Sp./Port. púrpura and It. porpora (with vocalic merger). See more examples in Moralejo (n. 1) 168–169. One might also consider the case of later loans imported from Greek refractory dialectal areas which retained the [u] pronunciation of upsilon, see Moralejo (n. 1) 167 and Biville (n. 1) 256, 264–265, 278 (who nevertheless also considers the possibility that <V> could be an ‘approximated’ transcription of /ü/ by Latin native speakers). However, the chronology and origin of Greek loans in Latin is rather difficult to establish.
This is also the value of upsilon in the modern Greek language; however, the change from [ü] to [i] seems not to be general until around the end of the 10th c. AD and no relationship can be established between this Greek development and the Latin I-spellings of upsilon, see Purnelle (n. 1) 257–258.
See Moralejo (n. 1) 169 and Allen (n. 17) 67.
See e.g. CVRA × Cyriae (LLDB-26339); CRVSE × Chryse (LLDB-140652); [G]LVCERA × Glycera (LLDB-74256); HOLVMPVS × Holympus (LLDB-28325); LVSAN[DER] × Lysander (LLDB-140639); NVMPHARVM × Nympharum (LLDB-122117); SVMMACVS × Symmachus (LLDB-15604); TVRANNVS × Tyrannus (LLDB-80992); PHILARGVRVS × Philargyrus (LLDB-140597); PHRVGIA × Phrygia (LLDB-122122); TVCHE × Tyche (LLDB-140628).
On this topic, see Tantimonaco, S.: The role of archaisms in the Latin inscriptions of the Roman Empire: some new considerations in light of Computerized Dialectology. ACD 55 (2019) 161–163 (http://real.mtak.hu/id/eprint/94914). This study also includes several references to the problem of the alleged conservativism of Hispanic Latin.
We consider as a proper name also the form MVROMEM (LLDB-75871), which appears on a prayer for justice from Baelo (AE 1988, 727) and has been traditionally interpreted as a misspelling for Myrionyma, an epithet of Isis, a divinity imported from the Greek-speaking East.
LLDB-28233 (Olisipo, Lusitania, AD 666). Apart from the phonetic phenomenon involving the upsilon, there is also a clear morpho-syntactical aspect to this form, namely the fall of final -s (or merging of the third with the second declension), see LLDB-140290.
In our corpus, see LLDB-28871 (MARTVRIA × Martyria).
We are currently working on a detailed study of this word.
The Greek town of Emporion (Lat. Emporiae) was founded in the 6th c. BC by Phocaean colonists, see Mierse, W. E.: Ampurias. The Urban Development of a Graeco-Roman City on the Iberian Coast. Latomus 54.4 (1994) 790–805 (http://www.jstor.org/stable/41537099). Another minor Hellenic center named Rhode was located only 18 km far away from Emporion.
The relatively reduced amount of data available for the Late Period discouraged us from providing maps for the Late Period.
See Tantimonaco: Remarks (n. 1). In contrast, the phonemes immediately preceding or following upsilon in a word do not seem to have exerted any influence on the way in which this last was misspelled as either <V> or <I>. In Hispania, upsilon is more frequently misspelled as <V> after lambda, ni, rho, sigma and tau; however, the same can be observed in the case of <I> replacing upsilon. See also Purnelle (n. 1) 283–284: « En conclusion, il apparaît nettement que les seuls phonèmes dont la nature influence la transcription d'un upsilon suivant sont l'aspiration initiale, les labiales (occlusives ou nasales) et l'occlusive vélaire sonore. En dehors de ces contextes, dont les derniers impliquent des sons liés par un trait spécifique à une ou à deux des graphies concurrentes (…) la transcription de la voyelle présente une répartition globalement homogène, quel que soit le phonème qui la précède ». Ibid. 292 : « L'ensemble des observations qui précèdent ne révèle aucune tendance nette et générale qui permette de reconnaître dans la nature du phonème suivant upsilon un facteur de variation assuré ».
See e.g. LLDB-63631 (Pallantia, Hispania Citerior, AD 661); 79389 (Gerunda, Hispania Citerior, AD 591–1000); 79475 (Caesaraugusta, Hispania Citerior, AD 313–700).
LLDB-59893 (Salmantica, Lusitania, AD 690–700).
LLDB-21532 (Arva, Baetica, AD 101–300).
LLDB-18380 (Valentia, Hispania Citerior, AD 548).
See e.g. LLDB-21765 (Hispalis, Baetica, AD 619); 29823 (Myrtilis, Lusitania, AD 489); 30903 (Pax Iulia, Lusitania, AD 584).
See Biville (n. 1) 299.
LLDB-60591(Aquae Flaviae, Hispania Citerior, AD 509).
LLDB-71145 (Tucci, Baetica, AD 501–600).
LLDB-79463 (Toletum, Hispania Citerior, AD 401–500).
LLDB-42417 (Caetobriga, Lusitania, AD 190–200).
LLDB-29788 (Myrtilis, Lusitania, AD 525).
LLDB-74257 (Asido, Baetica). In this case, a possible ‘folk etymology’ from Lat. semper might be suspected.
The spelling LYRICARIVS that appears on an inscription from Aurgi, Baetica (CIL II2 5, 26 = LLDB-70961 + 70962) has been related to Cl. Lat. loricarius, and could therefore be considered as an inverted case to O × Y spellings (Schmidt in CIL). However, it could also be a hapax legomenon referring to a profession related to playing the lyra (Gimeno in CIL). In fact, according to a search in LLDB, there are currently no other examples of Y × O in the whole epigraphic corpus of the Roman Empire. On this spelling, see also García González (n. 2) 82.
LLDB-12731 (Tarraco, Hispania Citerior, AD 393).
LLDB-34418 (Augusta Emerita, Lusitania, AD 101–200). See Biville, F.: Graphie et pronunciation des mots grecs en latin. Paris 1987, 17.
See e.g. LLDB-57630 (Roma, AD 14–50).
Attested e.g. in LLDB-78886 (Dyrrachium, Macedonia, AD 1–230): EVRVDICA.
See Biville: Les emprunts (n. 1) 283.
See above, n. 19.
For these names, see Pape, W. – Benseler, G.: Wörterbuch der grieschischen Eigennamen I. Graz 1959, 424–425.
There are currently 315 LLDB-records for <Y> replacing <I> and other 215 LLDB-records for <Y> replacing V in the different territories of the Roman Empire (last search: 10/10/2022).
LLDB-28162 (Aritium Vetus, Lusitania, AD 101–200). In this name, the hypercorrected use of <Y> might have been favored by the presence of an aspiration, as suggested by Purnelle (n. 1) 299. In fact, initial upsilon was always aspirated in Greek, see Allen (n. 17) 68.
LLDB-140409 (Pallantia, Hispania Citerior, AD 101–200).
The other records are: LLDBD-12647; 27480; 28162; 30106; 30107; 33053 (see below, n. 63); 34524; 63401; 122462; 140409; 1406491.
We hope to investigate this question more directly in a future paper.
According to previous scholars, the presence of a bilabial stop, a liquid or a dental consonant could favor, respectively, the rounding, relaxation and frontal pronunciation of Latin /i/, see e.g. Moralejo (n. 1) 166; Allan, W. S.: Vox Latina. A Guide to the Pronunciation of Classical Latin. Cambridge – London – New York – Melbourne 19782, 59; Purnelle (n. 1) 301–302; Solin, H.: Varia Onomastica XV. ZPE 156 (2006) 308.
LLDB-140660 (Clunia Sulpicia, Hispania Citerior).
LLDB-16992 (LHYBENS) (Edeta, Hispania Citerior, AD 1–100); 18295 (LYBES) (Edeta, Hispania Citerior, AD 1–100); 133569 (Igabrum, Baetica, AD 51–200). Theoretically, the underlying form could also be lubens.
LLDB-29832 (Myrtilis, Lusitania, AD 566).
LLDB-17568 (Elvas, Lusitania, AD 101–300).
LLDB-12800 (Tarraco, Hispania Citerior, AD 401–600).
See e.g. the examples provided in Mallon, J.: Pierres fautives 2. Lybica 2 (1954) 440, n. 44. Here, Mallon also quotes a form taken from the Hispanic corpus, DAPYNVS (LLDB-33053), that he considers as a misunderstanding of the name Daphnus. However, an alternative form Daphinus (with I × Y) cannot be excluded either, since the anthroponym Daphinus is also known from the inscriptions (30 records in EDCS on 10/10/2022; see also Solin, H.: Die griechischen Personennamen in Rom II. Berlin – New York 20032, 1164–1166).
LLDB-80554 (Pompaelo, Hispania Citerior).
On this topic, see Estarán Tolosa, M. J.: Learning the Latin Alphabet. Alphabet-switching and graphemic adaptation in the Western Mediterranean (first century BCE to first century CE). In Moncunill Martí, N. – Ramírez Sánchez, M. (edd.): Aprender la escritura, olvidar la escritura. Nuevas perspectivas sobre la historia de la escritura en el Occidente romano. Vitoria-Gasteiz, 2021, 401–421. The editors of this graffito, however, consider its reading as doubtful, since different graffiti seem to have been superposed here. Another peculiarity is the retrograde ductus of the inscription in question. See Unzu Urmeneta, M. – Ozcáriz Gil, P.: Grafitos Nominales de la Plaza del Castillo de Pamplona. In Andreu Pintado, J. (ed.): Los Vascones de las fuentes antiguas: en torno a una etnia de la antigüedad peninsular. Barcelona 2010, 507, nr. 11 (with picture).
LLDB-32175 (Augusta Emerita, Lusitania, AD 151–200).
See a discussion on these forms in Biville: Les emprunts (n. 1) 314–319. See also GL VII 158, 6 (Cassiodorus): Eutalium Eupolim et alia huius modi quidam per y litteram scribenda putaverunt, eo quod graeca esse dicerent; sed apud nos y littera numquam vocali coniungitur.
LLDB-122543 (Valeria, Hispania Citerior, AD 1–100).
Cf. Allen (n. 17) 66.
See Purnelle (n. 1) 303.
LLDB-140658 (Civitas Igaeditanorum, Lusitania). This is a pre-Roman name of Lusitania which can be associated with other indigenous names attested in the same provinces, such as Catto, Cattonius, Cattio, Caturicius, etc. See Vallejo Ruiz, J. M.: Antroponimía indígena de la Lusitania romana. Vitoria-Gasteiz 2005, 528, 583.
LLDB-17280 (Ossonoba, Lusitania, AD 101–200).
LLDB-96622 (Tarraco, Hispania Citerior, AD 193–235).
There are currently 21 LLDB-data referring to this type of spelling from different parts of the Roman Empire (last search: 10/10/2022).
See Biville: Les emprunts (n. 1) 308. See also Schön–Wittke in DNP VIII (2000) 328–331, s.v. Moesia and Schwetheim ibid., 608, s.v. Mysia.
Cfr. Horrocks, G.: Greek. A history of the language and its speakers. Singapore 20102, 162. See also Biville: Les emprunts (n. 1) 304; Biville: Graphie (n. 46) 17.
LLDB-18349 (Valentia, Hispania Citerior, AD 151–250).
172 datasheets recorded in LLDB so far (last search: 17/10/2022). The phenomenon in question appears to be particularly frequent in Rome during the 4th and 5th c. AD.
See Moralejo (n. 1) 170. On these spellings, see also Purnelle (n. 1) 487–495.
LLDB-81020 (Complutum, Hispania Citerior, AD 51–200).
LLDB-83756 (Tarraco, Hispania Citerior, AD 400–700).
LLDB-63400 (Nabrissa, Baetica, AD 630).
LLDB-60345 (Augusta Emerita, Lusitania, AD 662).
LLDB-29801 (Pax Iulia, Lusitania, AD 51–100). Another possible example of this phenomenon is found in HEp 2006, 158 (= LLDB-84461: HVIGVIA × Hygia), but the reading is doubtful. Moreover, there is a line separation (HV|IGVIA) which could have favored the misspelling in question.
See Encarnação, J. – Pina, M. J.: Cupa de Ferreira do Alentejo (Conventus Pacensis). Ficheiro Epigráfico 65 (2000) nr. 295.
LLDB-72856 (Barcino, Hispania Citerior, AD 301–350).
See Biville: Graphie (n. 46) 17, Horrocks (n. 76) 162.
LLDB-22286 (Astigi, Baetica, AD 151–230).
There are 18 LLDB-data recorded so far for the V-variant of this name (last search: 10/10/2022). On this spelling, see also García González (n. 2) 82.
LLDB-140661 (Carthago Nova, Hispania Citerior, AD 301–800).
I could only find one other possible example of this kind in LLDB, namely the spelling TERTY × Tertius (LLDB-64091, Rome, AD 51–300), which speaks against a phonetic transcription (last search: 17/10/2022).
