Transmission of the alphabet to and within Italy
In the 8th century BC, the island of Pithekoussai (modern Ischia) off the coast of Campania was colonized by Greeks from Euboia. While it is not quite clear whether the settlement was a proper colony or just a trading post, it spawned the foundation of historically important Kyme around the middle of the 8th century on mainland Italy. Pithekoussai itself seems to have lost importance at the turn of the century. The alphabet used by the colonists was that of the Euboic mother-cities Chalkis and Eretria. – Indeed, one of the oldest testimonies of early Greek writing is from Pithekoussai: the so-called Cup of Nestor, dated to the last quarter of the 8th century. (Jeffery 1990: 235) The Etruscans would have been in contact with the Greek settlers from the beginning, and the acquisition of their script was not a long time coming: The oldest document of written Etruscan, a kotyle from Tarquinia (Ta 3.1), is dated to about 700 (Wallace 2008: 17).
As the oldest Etruscan abecedarium on an ivory tablet from Marsiliana d’Albegna (ET: AV 9.1; about 650) shows, the Etruscans adopted the Greek alphabet, in its Eastern Greek "red" variety as used in Euboia, without any changes with regard to the different phonemic systems of the two languages. (For details see Jeffery 1990: 236 ff.) Only by and by do the documented abecedaria reflect a process of adaptation to writing practice. Etruscan had a plosive system consisting of two rows, written with the Greek characters for the unvoiced unaspirated (Pi, Tau, Kappa (/Gamma/Qoppa)) and unvoiced aspirated (Phi, Theta, Khi) rows. A phonetic realisation very much like the Greek is communis opinio among Etruscologists (see Wallace 2008: 30 f.). In any case, the obsolete characters for the mediae dropped out – all except Gamma, which together with Kappa and Qoppa became part of a curious orthographic rule for writing allophones, and only later replaced both the other characters as the exclusive one for the velar stop. Due to the lack of /o/ in Etruscan, Omikron fell away. In the 6th century, an additional sign was created for /f/, after a phase of writing the sound with a digraph <vh> or <hv>, and added to the end of the row. As concerns the writing of sibilants, a certain confusion on the part of the Greeks (see Jeffery 1990: 25 ff., Swiggers 1996: 266 f.) appears to have been propagated to the Etruscans: The Etruscan language seems to have had – apart from a dental affricate written with Zeta – two sibilants /s/ and probably /ʃ/ which were written with Sigma and San – in the South Sigma for /s/, San for /ʃ/, the other way round in the North. In the Southern cities Caere and Veii, where a number of divergences from general Etruscan writing practice can be observed over the course of time, a Sigma with more than three strokes appears instead of San. Finally in Cortona, a monophthongised, possibly long /e/ was consistently written with the character . As is customary in archaic Greek inscriptions, Etruscan inscriptions are generally sinistroverse, apart from a short phase around 600 in Caere and Veii. Unlike Greek practice, boustrophedon writing is rare. While word separation is consistently executed on Nestor's Cup, the archaic Etruscan texts often dispense with it, until it establishes itself in neo-Etruscan time (after 470). (For details see Wallace 2008: 17 ff.; a collection of Etruscan abecedaria in Pandolfini & Prosdocimi 1990: 19–94.)
As things present themselves to us now, the Etruscan script has found its way to the peoples north of the river Po more than once, but not from the Etruscan settlements in Padania. Etruscan inscriptions in the very North (find places in pink on the map) are known from Liguria (Li), the Reggio Emilia and the area around Mantova (Pa), the environs of Bologna (ancient Felsina, Fe), and from Adria (Ad) and Spina (Sp). The oldest testimonies come from the Reno valley (Fe 2.1, Fe 3.1, Fe 6.1, Fe 2.2, Fe 2.3, from around 600) and from Rubiera (the stelae Pa 1.1 and 1.2, dated to the end of the 7th century). In both cases, the traditions only start again some hundred years later, i.e. at the end of the 6th century in Marzabotto, in the 5th century in the Reggio Emilia and Mantova. The great ports and commercial cities Adria and Spina only became relevant as Etruscan settlements around 500. A number of gravestones from Liguria, especially around the Magra river and its tributaries, are dated to the 2nd half of the 6th century; most of them are filed as being written in a North Italic alphabet in the Lexicon Leponticum (sigla MS and SP). For the question of whether the inscription(s) of Feltre (Pa 4.1) are linguistically and/or epigraphically Etruscan, see here.
The Veneti are speakers of an Indo-European language settling in the Veneto (find places in blue on the map). Their inscriptions are collected in Pellegrini & Prosdocimi 1967; for supplementation see Prosdocimi 1988 and numerous publications by Anna Marinetti, esp. Marinetti 2004b. For the "traditional" view on the origin of the Venetic alphabet (from the Etruscan alphabets of Adria and Spina) see Pellegrini 1959. According to the more recent theory of Prosdocimi, a first, archaic version of the Venetic script ("phase 1"), attested securely only in one inscription (*Es 120, dated to the beginning of the 6th c. at the latest) and arguably in two further inscriptions (Es 1, *Es 122), was based on a model from Northern Etruria, while a separate tradition lies at the basis of most of the younger locally diverse alphabets (Este, Padua, Cadore, etc., "phase 2"). The archaic Venetic alphabet seems to have featured a rare form of Theta, , which is found in a handful of inscriptions from 6th-century Chiusi and Volsinii (Cl 2.8, Cl 2.6, Cl 2.5, Vs 1.23 and Vs 1.14, see Colonna 1972: 470), as seen in *Es 120. *Es 122 shows that the digraph <vh> was used to write /f/, rather than the new character , which was introduced no sooner than the middle of the 6th century. The table shows the characters contained in the abovementioned inscriptions (disregarding minor variants). Pi is missing, but note that *Es 122 has , read l by Prosdocimi; cp. Pi in the form in Chiusi. Syllabic punctuation is absent.
The younger alphabet of Este is unusually well documented on a number of votive writing tablets from a sanctuary-cum-writing school and distinguished by syllabic punctuation, both of which phenomena, together with the actual content of the inscriptions, connect it with the 6th century writing tradition of the Portonaccio sanctuary in Veii in the South of Etruria. The background of syllabic punctuation is debated.
Syllabic punctuation became the key feature of Venetic script, even though alphabet variants from other parts of the Venetic realm deviate from the Este alphabet, most prominently in the writing of the dental stops. Prosdocimi argues that the younger phase 2 alphabets represent different solutions for reconciling the archaic Venetic alphabet with the younger Etruscan one and particularly the theoretical grid on which the writing instruction was based. Whether the Veneti still had access to the characters for mediae (as lettres mortes through Etruscan teaching) is hard to judge, but they did not use them to write their voiced stops (Prosdocimi's considerations on p. 331 ff.). Instead, they employed the superfluous letters for the Etruscan aspirated row. While in the case of labials and velars, this transition appears to have happened smoothly (Pi = /p/, Phi = /b/; Kappa = /k/, Khi = /g/), the characters for the dentals were shifted around. *Es 120 clearly demonstrates the use of Tau for /d/; the abovementioned Chiusi-style Theta ( in Es 1 and *Es 122) must be expected to stand for /t/. This distribution is also documented for phase 2 Vicenza on a stela (Vi 2). In the younger Este alphabet (and also in the sanctuaries of Làgole (Calalzo di Cadore) and Auronzo di Cadore), /t/ as in the archaic inscriptions is written as a (large) St. Andrew's cross, but Zeta is employed to write /d/. A third combination is used in Padua, where first Etruscan Tau, later St. Andrew's cross are in use for /d/, while /t/ is written with a more traditional framed form of Theta (rounded or angular).
The origin of St. Andrew's cross is somewhat obscure: Prosdocimi (p. 332), regarding the archaic distribution, explains Tau for /d/ and Theta for /t/ by a developing homography of and / . The phonetic values were swapped before the characters were differentiated again, leading to being used for /t/ henceforth. He points to the Lepontic alphabet and the Este alphabet tablets for evidence of a tendency of Tau to tend towards a cross-shape. To further avoid homography in this area, Tau was substituted by Zeta in Este; in Padua, the form of Theta was changed to , which allowed Tau to turn into . In other words, according to Prosdocimi, has two separate origins. On the Este writing tablets, where the letters can be unambiguously identified by their position in the row, Tau appears – with Prosdocimi: is retained as a lettre morte – in the shape of a cross, similar to, but clearly distinct from, Theta: While the latter is a large , Tau is smaller and sometimes lopsided (e.g. in Es 23, see table). The individual letters being written in rectangular fields formed by a grid, with the grid lines regularly being used as hastae, it might be argued that the entire frame around the St. Andrew's cross representing Theta, whose tips reach into the corners of the field, is supposed to be part of the letter, forming a large, but otherwise inconspicuous . Theta would then have come to be reduced to only the cross through reinterpretation. This explanation, however, predating Prosdocimis distinction of older and younger Venetic alphabet, does not account for the early appearance of and its apparent connection with Chiusi; note also that of six preserved tablets, a third (Es 24 and Vi 3) lack grid lines and yet feature Theta without a frame. On Es 25, where the grid lines are not used as parts of the letters, Theta is missing due to object damage, but the tablet serves to corroborate Prosdocimis theory by having Zeta in the place of Tau, probably due to a scribal error.
The Venetic script features Omikron, which in the younger Este alphabet is situated not in its ancestral place, but at the very end of the row, as evidenced by the votive tablet Es 23, the only one bearing a complete row (in addition to the usual consonant-only). While Omikron is usually assumed to have been acquired directly from the Greek alphabet, probably through contact with Greeks settling in and south of the Po delta, Prosdocimi (p. 329) favours the theory that it was taken as a lettre morte (through writing instruction) from the Etruscan alphabet before its ultimate reduction. The Venetic use of Sigma vs. San follows the South Etruscan use, Sigma being the character used for the default sibilant and San leading a marginal existence. This is also the case in the archaic inscriptions; for possible explanations see Prosdocimi on p. 330 f. Finally, one of the distinctive features of the Venetic script is the frequent inversion of Lambda and Upsilon.
Inscriptions in a Cisalpine Celtic language begin to appear in Western Transpadania around 600. The Lepontic core area lies between the Lago di Como and the Lago Maggiore; on the Celtic presence south of the Alps and the distinction between Lepontic and Cisalpine Gaulish see Uhlich 1999 and 2007. The inscriptions are collected in the Lexicon Leponticum, including all testimonies which may contain Celtic language material, as well as all inscription finds from west of the Adige even if ascription is dubious: A considerable number of documents, mostly short/fragmentary inscriptions of a low date, cannot be shown to be linguistically Celtic, or even to be written in the so-called Lugano alphabet (find places in yellow on the map). The "Lepontic" corpus being accordingly large and varied, it is hard to determine how the usual schibboleth characters – those for stops and fricatives – are used. Pi, Kappa and St. Andrew's cross are the standard letters, and can be shown to be used for both tenues and mediae. While Phi does not occur at all, Khi is employed for /g/ in one of the oldest inscriptions, as well as in three younger ones (TI·13, PV·4, VC·1.2) and in coin legends (NM·6.1, NM·6.1). In the latter, Khi appears together with Theta in the name Segetu or Segedu, which is also attested on four ceramic bowls from Prestino – here with Kappa for /g/ and Zeta for the dental. Theta appears two more times, in the shape , in archaic VA·3 (possibly Etruscan) and the Inscription of Prestino. The latter, a lengthy inscription on a stela, is the only Lepontic text in which a systematic use of the characters for dentals can be observed: St. Andrew's cross is absent, Theta appears to stand for /t/. Tau in the shape demonstrably stands for /d/, while Zeta represents the affricate. Pi and Kappa are used for /p/ and /g/. Tau appears twice in later inscriptions (TI·36, NO·21.1), both times in the shape , and both times together with St. Andrew's cross. While this strongly suggests that Lepontic St. Andrew's cross must, like its Venetic equivalent, be identified with Theta, the combined use of the two characters cannot be shown to be systematic (/t/ vs. /d/). Tau-, Zeta- and Khi-like shapes crop up a number of times in dubious and/or uninstructive contexts (see LexLep); another instance of lexical use of Zeta in NM·16. Beta, Delta and Gamma are absent until the appearance of Latin(oid) inscriptions from the Roman imperial time, but Omikron is present from the earliest inscriptions. On the use of San see LexLep and Stifter 2010.
Pi and Lambda are distinguished systematically as vs. ; Upsilon appears tip-down , though inverted forms do occur. Alpha is closed ( and similar) in the older inscriptions, later changing into . All in all, the Lugano alphabet is decidedly more similar to its Etruscan source than the Venetic. It is hard to determine whether individual inscriptions displaying divergent features are influenced by the Venetic writing tradition. The fact that both the Celtic and the Venetic languages are Indo-European may be the cause of parallel independent developments. Verger 1998 argues for a transmission of the alphabet to Western Transpadania via the area of Genova and the Scrivia valley. Note that a number of inscriptions from the area of La Spezia / Massa-Carrara displayed as Etruscan on the map are filed as linguistically and/or epigraphically Celtic / North Italic in the LexLep (sigla MS and SP).
The corpus of the so-called Sondrio alphabet ("Camunic script“), conspicuous for its obvious graphic peculiarities, comprises the rock inscriptions of the Valcamonica, and a handful of testimonies from other places whose characters bear resemblance to those of the rock inscriptions, though the alphabets cannot be said to be identical (find places in grey on the map). Indeed, different systems seem to have been employed in the Valcamonica itself. The sigla system is not standardised, but useful collections are provided by Mancini 1980 and Tibiletti Bruno 1990. The language written in the rock inscriptions, called "Camunic" after the demonym Camunni documented by the ancients, has not yet been deciphered or convincingly connected to any of the surrounding languages; the other testimonies have been argued to write diverse languages: While the two inscriptions on stelae from Montagna in Valtellina (PID 252) and Tresivio (PID 253) feature endings similar to those commonly found in Camunic rock inscriptions, the non-Latin part of the Voltino bilingua has been read Etruscan as well as Raetic and Celtic. Celtic has also been suggested for the inscription on the Castaneda flagon, datable to the 5th–4th c. The dubious inscription AV-1, included in the TIR as linguistically Raetic, appears to be written in a variant of the Sondrio alphabet. Finally, the fragmentary inscription on a stela from Cividate Camuno in the Valcamonica itself is utterly enigmatic.
The main problem about the reading and interpretation of the inscriptions lies in the identification of the letters: Rock inscriptions from different localities, alphabetaria, and the (possibly idiosyncratic) testimonies from abroad appear to exhibit substantial differences in the use of some characters, which could so far be neither conclusively sorted out individually, nor reconciled. The picture presented by the twelve alphabetaria, or fragments of such (first edited in Tibiletti Bruno 1990; see also Tibiletti Bruno 1992), from the Valcamonica in particular demonstrates the Sondrio alphabet to be the odd one out among the North Italic alphabets. The table shows the characters as they appear in two distinct groups of alphabetaria: The first line gives the alphabet row PC 10 from Piancogno, with letters slightly standardised where their shape deviates from Camunic standard (Nu, Qoppa). The positions of Mu and Nu as well as of Gamma and Delta are interchanged in the original, Delta being written in ligature with Beta (sharing its last hasta). The ligature and possibly the inversion of the nasals also occur in the very similar row PC 27. The other alphabetaria or fragments of such from Piancogno are PC 6, PC 12 and probably PC 28. The second line gives an ideal alphabetarium from rock 24 of the Foppe di Nadro, based on FN 3, FN 4, FN 5 and FN 6, where only FN 3 and FN 6 are complete. Here also the nasals are interchanged. The two other alphabet fragments FN 1 and FN 2, also on rock 24, both end with Digamma (?) and display a variant form of Gamma . The presence of a complete Greek row suggests that the Sondrio alphabet came to Transpadania directly from a Greek source, without Etruscan mediacy. More than that, the Greek model can be argued not to have been of the "red" variety like the Euboic alphabet from which the other Italic alphabets ultimately derive. Even under such a premise, the shapes of the letters are highly unusual, not to mention the question of how such a script could have found its way into the remote Oglio valley.
There are two notable similarities between the Camunic and Raetic corpora, i.e. that both graphic variants of the Raetic special character appear in the context of the Sondrio alphabet: The character taking the position of San in PC alphabetaria is reminiscent of the Magré special character / (but note that Magré has standard San); an arrow-shaped character like the Sanzeno special character appears in the problematic end sequences of the PC alphabetaria and on the Castaneda flagon.
The map below shows find places of inscriptions of the four North Italic corpora, together with Etruscan inscriptions in the Padan plain. It is intended to give a rough overview over the distribution areas, and glosses over problematic or ambiguous ascription in some cases (e.g., the "Ligurian" inscriptions). In the case of inscriptions which belong to different groups linguistically and alphabetically, the colours reflect the alphabet used (e.g., AV-1).
The Raetic alphabets
Within the Raetic corpus, two different alphabets are distinguished. This goes back as far as Mommsen 1853 with his "Tyrolean" and "Verona" alphabets. Mommsen's "Tyrolean alphabet" was termed "Bozen alphabet" by Pauli 1885; his "Verona alphabet", originally attested in only one inscription, was rebranded as "Magrè alphabet" by Pellegrini 1918. After a suggestion by Mancini 1975: 306 (n. 42), the term "Bozen alphabet" was again changed to "Sanzeno alphabet" to reflect the latter site's large output of finds. (See Modern research on Raetic for details). Though the two alphabets had long been expected and were eventually demonstrated to encode the same language, the distinction still holds. The alphabets differ from each other in the use of graphic variants of a handful of letters, but share certain features which set them apart from the other North Italic alphabets and can therefore be considered typically Raetic.
Pi, lambda and upsilon are the shibboleth characters which primarily distinguish the Magrè and Sanzeno alphabets (Whatmough 1933: 507; Prosdocimi 1971, 31–34). The Magrè alphabet employs forms which are identical with or similar to those used in the Venetic alphabets: pi with a pocket (sometimes opened or similar; the pocket is almost always open in Venetic, more similar to archaic Etruscan , which in Raetic is only attested once), lambda with a bar on top , and tip-up upsilon . The Sanzeno alphabet bears a closer resemblance to the Etruscan and Lepontic alphabets (Pauli 1885: 58–60): pi with a single bar , lambda with the bar at the bottom , and tip-down upsilon correspond to the standard letter forms in those alphabets. According to Prosdocimi 1971: 33, the Venetic system of distinction is a variation of the archaic Etruscan one, while the Sanzeno forms (especially pi with a single bar) correspond to younger Etruscan ones. However, pi is already found in 7th-century Chiusi (e.g., Cl 2.1, 2.4).
There is a certain extent of (random?) variation in the North Italic alphabets concerning the orientation of lambda and upsilon – non-inverted forms occur in the Venetic alphabetss (e.g., lambda in Es 16, upsilon in Es 22); inverted forms, particularly of upsilon, appear sporadically in the Lepontic alphabet (e.g., TI·36.3). Pi appears with a single bar in Venetic inscriptions from the Cadore (e.g., Ca 65). Still, the variation in the Raetic inscriptions is too regular to be put down to chance. The only inscriptions in which Sanzeno-forms co-occur with Magrè-forms are three inscriptions from find places associated with the Magrè alphabet: AS-17.1 has hyper-distinctive next to and , MA-6 has in combination with . In the latter case, the occurrence of "incorrect" upsilon may be attributed to the tendency to invert letters (especially alpha and epsilon) which can be observed in the Magrè inscriptions (see below sub Writing direction). In VR-6, Magrè-type upsilon appears beside a character looking like the Sanzeno letter for the dental affricate ; is therefore ambiguous, as is tau , which has its bar crossing the hasta, but is retrograde as typical for the Sanzeno alphabet (see T).
In addition to the above-mentioned letters, three others appear consistently in different graphic variants in the two alphabets. Tau always appears with the bar rising in writing direction, and usually not crossing the hasta, in Sanzeno-context (←; see T). Heta, though not common, always features three bars in Magrè-context, but two in Sanzeno-context (single-bar is as yet undocumented). Both alphabets have graphically innovative characters for the dental affricate: in Sanzeno-context (not only at Sanzeno itself), exclusively at Magrè (otherwise absent from Magrè-type inscriptions). See T on the question of whether the character is associated with the Magrè alphabet. Lastly, vestiges of Venetic syllabic punctuation are found only in Magrè-context, while word separation is only employed in Sanzeno-context (see below sub Writing direction).
While the forms of pi, lambda and upsilon as well as the instances of syllabic punctuation connect the Magrè alphabet with the Venetic alphabets, the Sanzeno alphabet has an Etruscan look to it. The similarities between the two alphabets concerning the use of the characters for obstruents, allegedly demonstrating a dependence on a Venetic model for both, is discussed below sub Obstruent spelling. The most evident feature unifying the Raetic alphabets is a negative one: the absence of omicron. Seeing that it is linguistically motivated (see The Raetic language), it does not provide a strong argument for a common model of the two alphabets. Purely graphic characteristics connecting the two are mu with only three bars instead of the more common four (), as well as two characteristics pertaining to writing direction: alpha ← with the bar slanting downwards against writing direction, and sigma ← with the upper angle opening against writing direction. Both the latter features are also known from neighbouring writing traditions – alpha is always retrograde in the Venetic inscriptions from the Isonzo region (Is 1–3, *Is 5–6); the orientation of sigma is notoriously irregular in all North Italic writing traditions – but they are significantly prevalent in the Magrè alphabet and good as exclusive in the Sanzeno alphabet.
The Sanzeno alphabet is distinguished by its uniformity across almost the entire time and area of its attestation. There is some minor variation in the forms of kappa (placement and length of bars , ), rho (pointed vs. rounded pocket , ), tau (placement of bar , ), phi (size of pockets , ) and chi (placement and length of bars , ), but, otherwise, letter forms are quite stable. The character inventory, as far as can be seen, is the same in all find places. In contrast, we find a considerable range of regional and diachronic variation within the province of the Magrè alphabet, to an extent that the term "Magrè alphabet" should be regarded as more of a cover term for a number of local and chronological variants which share the features enumerated above, but exhibit differences with regard to individual letter forms, orthography, and (syllabic) punctuation:
- a number of potentially archaic inscriptions exhibit different character sets and orthographies (see below sub Chronology),
- simplified syllabic punctuation is employed only at Serso and Magrè (see below sub Punctuation),
- the graphically and functionally obscure character only appears at Serso and in a few scattered Magrè-type documents (see T),
- the Magrè alphabet's letter for the dental affricate is only employed at Magrè (see Þ),
- pi has a large pocket in inscriptions from the Inntal (see P),
- zeta pops up sporadically and may denote a stop, Este-style, in some inscriptions (see Z), while
- zeta and san behave suspiciously in inscriptions from the area of Verona (see Z and Ś), and
- one of the two decidedly unlike petrograph alphabets attested in the Northern Limestone Alps shows some as yet unclassified idiosyncrasies (see Raetic epigraphy).
All these special features, however, are connected by their association with Venetic writing traditions.
The areas in which the Magrè and Sanzeno alphabets are used are neatly separated, as shown on the map on the right (Magrè-alphabet find places in light green, Sanzeno-alphabet find places in dark green). The Sanzeno alphabet is used in the central area, i.e. the Val di Non, the upper Adige valley (including the Unterland, the Bozen basin and the Vinschgau) and the Eisacktal, with tributary valleys and the surrounding highlands. Its area of distribution mostly coincides with the core area of the Fritzens-Sanzeno culture, i.e. Südtirol and the Trentino. Magrè-type inscriptions, as may be expected from their affinity with the Venetic script, come from the area of the archaeological Magrè group, i.e. the Alpine foothills south of Trento between Adige and Piave. This includes the inscriptions from the area of Verona, and the stray finds from the Padan plain (PA-1, TV-1). Contrary to what is generally asserted in the literature, inscriptions from beyond Brixen are also written in the Magrè alphabet. Tip-up upsilon appears in WE-4, IT-4, IT-2 and FP-1. The graphically ambiguous can be clearly identified as Magrè-type lambda rather than Sanzeno pi in WE-1 and WE-4 thanks to the well attested words lavise and eluku; a reading of the letter as lambda is also preferable in IT-7 and especially IT-4, where it is part of the pertinentive ending -le. being lambda, the associated pi with a pocket can be found in FP-1, IT-4 (which thus has the full set of diagnostic letters) and IT-8. In the latter two inscriptions, pi appears in a peculiar form with a large angle which reaches the bottom of the line (see P). See ST-2 for another possible instance of , and Raetic epigraphy on the generally Venetoid letter forms and orthographic behaviour in the petrographs. Tau is in WE-4; heta is absent, as is a specific character for the dental affricate (which does not appear in any Magrè-type inscriptions except in those from Magrè itself). Clear syllabic punctuation is absent, but see below sub Punctuation on puncts in certain petrographs.
In the boundary areas between the two alphabet provinces, we often find mixed features, though – as said above – never within the same inscription. Two inscriptions from the south-eastern border of the Sanzeno-alphabet area, however, combine Sanzeno-type letter forms with what appears to be Venetoid syllabic punctuation (see below sub Punctuation): CE-1.3 from the valley of the Avisio (all five sequences on the Situla di Cembra are written in the Sanzeno alphabet with , and consistent ← and ←), and SR-2 from the Valsugana (with , not accompanied by pi or lambda). The Serso antlers, from a find place which still belongs to the core area of the Fritzens-Sanzeno culture, generally make up a somewhat varied group, though with some notable shared features. With the exception of problematic SR-2, the inscriptions are basically written in the Magrè alphabet; seven of the twelve feature some sort of punctuation, three of these have punctuated letters. Three of the inscriptions contain the letter , which also appears sporadically in the Val di Non and is suspect of belonging to an archaic tradition (see T).
The northernmost Sanzeno-alphabet inscription WE-3 and the southernmost Magrè-alphabet inscription in the Wipptal WE-4 both come from Stufels (Brixen). The Wipptal lies outside the reach of the Sanzeno alphabet, which may account for the fact that the Pustertal, which opens into the valley of the Eisack at Brixen, takes part in both traditions: PU-1 has Magrè-type letter forms, but displays a number of peculiarities, while PU-5, PU-6 and PU-11 have Sanzeno-style upsilon. This border at Brixen is not random, cf. Lunz 1981b: 43, who considers it as a point of transition between the northern and southern Alpine area (beside the Alpine divide at the Brenner pass), as it marks the point where the narrow Wipptal opens into the comparatively easily accessible Brixen basin.
Only one inscription breaks the pattern of distribution: IT-5, found in the Inntal, is clearly written in the Sanzeno alphabet. In light of a Venetic inscription found at the same site (*It 1), IT-5 is best judged as an import or the work of a travelling dedicant.
As concerns the inscribed Negau helmets, both found far from the Raetic area (not on the map), one bears at least one Sanzeno-alphabet inscription; the other one's inscription lacks the diagnostic letters, but features . The single fragmentary inscription from the Unterengadin (EN-1, not on the map), is too short to be ascribed to either of the alphabets.
See Property:alphabet for a scalable and tagged map.
From the overview given on Archaeology in the Raetic area, a rough history of Raetic writing culture can be inferred. Any chronology and the conclusions drawn from it, however, – including potential palaeographic indicators for dating – must remain provisional and may have to be revised or dismissed, depending on reassessments of the material from the archaeological side. Apart from the uncertainties imported from archaeology, it must at all times be remembered that an inscription is not necessarily as old as the object it is written on. This seems obvious, but with a material base as small as the Raetic one, different evaluations of the function of a single group of inscriptions may lead to vastly different results for the assessment of the entire corpus. It must therefore be attempted to determine the relation which various types of objects bear to the inscriptions which are written on them, even though – as may be expected – it is rarely ever possible to argue this conclusively. Cf. Gamper 2006: passim and especially 353–355, who arrives at considerably lower datings for various Raetic find groups and, by consequence, the Raetic writing culture.
The above caveats apply particularly to a group of documents with (potentially) early datings, which are notable for their lengthy inscriptions on unusual bronze objects from unusual places. The Situla in Providence (HU-7), the Paletta di Padova (PA-1), the Lothen belt plaque (PU-1) and the Spada di Verona (VR-3) are all dated typologically, so that the inscriptions do not have termini ante quos. As pointed out above, bronze objects – particularly valuable pieces like the ones just mentioned – were used for much longer than objects of everyday use, so that inscriptions written on them may theoretically be considerably younger than the objects themselves. HU-7, PA-1 and VR-3 can be argued to have votive character, containing the word utiku. All four inscriptions are linguistically indubitably Raetic and written in Venetoid alphabets (, , ), but these alphabets differ from each other: PA-1 appears to employ punctuation for auslauting consonants and to work with a restricted character set for stops (⟨uθiku⟩, ⟨akvil⟩), whereas PU-1, the only northerly find, features a plethora of characters for obstruents (pi, theta, kappa, phi, zeta, chi, san, heta), as well as some peculiarities concerning orthography (zeta for a stop?) and ductus (four-stroke sigma , double-pennon san , zeta ). HU-7 and VR-3 both feature .
Chronologically, the next point of focus is the Val di Non with its large amount of bronze material dated to LT A–B1 by its association with situla art and Etruscan cultic traditions. It is notable, however, that the potentially oldest document from there, the only inscription from the Monte Ozol sanctuary (NO-13), is reminiscent in form () and content (terisna) not of the Sanzeno-type material discussed in the following paragraphs, but of the Serso inscriptions and SL-1. The only other inscription from the Val di Non which features is NO-3 on the miniature shield from Meclo, otherwise written in the Sanzeno alphabet (see T). The little shield belongs with the group of bronze votives from the Val di Non dated to the early La Tène period. We are here mainly concerned with objects that were manufactured specifically for the purpose of being offered; the inscriptions should therefore be assumed to be contemporary with the objects. This includes NO-3 as well as NO-19, the only inscription on a bronze sheet figurine, NO-15, SZ-87 and SZ-96 on bronze batons, the statuette inscriptions NO-16 and SZ-16, and the inscriptions on half-plastic bronzes NO-11 and SZ-1–SZ-15. The case is not as clear for the inscriptions on (parts of) situlae. See, as case in point, SZ-30 on the question of when the inscription on the Sanzeno situla was applied. SZ-17 and SZ-31 on handles look more like workmen's inscriptions than votive ones. SZ-17 is almost identical to an inscription on an iron knife (SZ-38); if this is indeed a factory mark, it may indicate an equally early dating for at least some of the iron tools – or, otherwise, a considerably lower dating for (some of) the bronze material.
From along the Adige and the Bozen basin, associated bronze objects come more from grave or problematic contexts than clearly identifiable sanctuaries. The only certain votive object is the miniature situla with BZ-25; the other objects are items of use, even though mostly associated with orientalising cult (cists, situlae, simpula). It cannot be excluded that some of the inscriptions on the Bozen material are younger than those from the Val di Non. However, the grave contexts provide reliable early datings for some of them, so that it is not unreasonable to assume LT A–B1 or at least B2 datings for these inscriptions, including BZ-2 on a bronze axe and BZ-14 on a piece of bronze sheet. A high dating is generally assumed for the outlier IT-5 (Marchesini 2013: 53). Of particular interest, though ultimately not helpful, are the dated gravestone inscriptions BZ-10.1 and BZ-24. With the one's high and the other's very low dating, no conclusions can be drawn for the dating of the other inscriptions on slabs in the Raetic corpus, BZ-6, NO-10 and RN-1 (all written in the Sanzeno alphabet). The latter belongs with the documents from the Eisacktal and surrounding plateaus up to Stufels, inscriptions which are written in the Sanzeno alphabet, but undated (RN-2, WE-3). Language-encoding inscriptions on undated objects from the Val di Non and the Bozen area further include SZ-18 (on an unidentified bronze object), BZ-12 (on a key), SZ-22, SZ-24, SZ-97, SZ-98 (on antler grips), and SZ-94 (on a bone).
It may tentatively be assumed that the writing culture in the Raetic core area flourished in the early La Tène period, viz. roughly the 5th and 4th centuries, as did the Fritzens-Sanzeno culture as a whole under Etruscan stimulus. More than that, the Sanzeno alphabet may (originally) have been linked to a Sanzeno sanctuary. Alphabets which are specific to sanctuaries are known from the Venetic sphere; the evidence of Este-Baratella shows that writing schools could be appended to major sanctuaries (although an outright writing cult as in Portonaccio and Este is not in evidence elsewhere). In light of the concentration of finds and the homogeneity of the Sanzeno alphabets (especially on the votive bronzes; see Raetic epigraphy), it may be speculated that it was developed in and emanated from a sanctuary (at Casalini?) during the first half of the 5th century. How long this alphabet was in use is hard to say. The relevance of the inscribed iron and ceramic material and the various obscure shortish sequences on antler and bone objects from the area is rendered doubtful by the fact that not a single piece can be shown to bear a language-encoding inscription, even if the letter forms fit (see Non-script notational systems). Of the few demonstrably young inscriptions from the Raetic core area, one is Latin. A survival of the Sanzeno alphabet may be documented by the Latinised BZ-24, though it cannot technically be decided whether is a Sanzeno-style or a Latinised letter and which epichoric alphabet was mixed with the Latin one. Due to its doubtful status as a genuinely Raetic inscription, NO-2 does not make a strong case either. The best evidence for an actively used Sanzeno alphabet as late as the 1st century BC are the Ganglegg inscriptions from the upper Adige valley – depending on the nature of their previous function, the inscribed bones and bone points may be older than the end of LT D2, when they were laid down during the abandonment of the houses, but hardly much. How this find group fits in with the chronology of Raetic writing is yet to be determined.
During the time in which the central Raetic area writes Sanzeno-style, Magrè-alphabet inscriptions appear sporadically both to the north and to the south. Early La Tène documents come from the Fern pass (FP-1), from the Pillerhöhe (IT-8, IT-9), from Matrei am Brenner (WE-1), Stufels (WE-4), Serso, and from San Briccio di Lavagno in the very south. Both inscriptions from the latter site are Venetoid, though, apart from this basic characteristic and the fact that they are both inscribed on unusually large pieces of antler, they do not appear to be associated. VR-1 is conspicuous for the occurrence of , which may connect it with VR-3, while VR-2 arguably has similarities with much younger testimonies from San Giorgio di Valpolicella and Castelrotto (see Z). The two linguistically usable Montorio Veronese inscriptions (VR-6, VR-7), chronologically and geographically in between, fail to constitute a link. The inscribed potsherd (VR-9) may belong in a Roman context; cf. the two potsherds with Latin inscriptions from the same find spot (Marinetti 2004: 409). Magrè-alphabet finds dated to the middle and/or late La Tène period, apart from the above-mentioned ones from the area of Verona, are the Magrè antler pieces, the inscriptions from Bostel, IT-2 from the Inntal, and the Trissino bones. IT-4 is dated by context and may be older than the 1st century BC.
To sum up, we are faced with at least two distinct traditions of Raetic writing. The homogeneous Sanzeno alphabet was used in a specific time and area, with a chronological irregularity in form of the Ganglegg subcorpus; there are no local variants and there is only a small range of variation in the letter forms and their use, which may indicate that the tradition goes back to a single centre of literacy in the Val di Non. Outside this core area, the situation is less tidy. A Venetoid writing tradition for Raetic makes its appearance as early as the late 6th century and remains active during the early and middle into the late La Tène period in both the south and the north. It is to some extent marked by group finds of similar objects (Serso, Magrè, Bostel, Trissino, San Giorgio, petrographs), some of which display idiosyncratic features. The earliest Venetoid documents particularly stand out; PA-1 and PU-1 may represent isolated attempts at writing the Raetic language with foreign alphabets.
Regarding what was said above about the common and distinguishing features of the two Raetic alphabets and their geographical and chronological distribution, it remains yet to be determined how the two traditions relate to each other and to the potential model alphabets. It is quite possible that the two alphabets derive from different models, and/or that they were secondarily influenced by different traditions in the course of their time of use. The obvious candidates for the source of Raetic writing are the Etruscan and Venetic writing cultures, but Celtic and Camunic writing practices cannot be excluded to have taken part.
According to Schumacher 2004: 312–316 and Rix 1998: 48–56, the Raetians learned the art of writing from the Venetians rather than directly from the Etruscans. The following features of the Raetic alphabets are supposed to indicate a Venetic model:
1. Venetoid forms of the letters pi, lambda and upsilon,
2. rudimentary syllabic punctuation in some subcorpora,
3. employment of phi, tau and chi for lenes/voiced stops rather than for the second obstruent set,
4. non-employment of zeta,
5. St. Andrew’s cross for the unvoiced dental stop,
6. use of sigma for the dental sibilant, while san is marginal.
Evidently, points 1 and 2 only apply to the Magrè alphabet. Points 3, 4, 5 and 6, however, also concern the Sanzeno alphabet, which displays Etruscoid characteristics in its letter forms, but, from an orthographical perspective, looks as if those who created it were not acquainted with Etruscan writing at all. The following sections will discuss aspects of Raetic orthography with regard to the question of the derivation of the Raetic alphabets.
Writing of dentals
About three quarters of Raetic inscriptions whose writing direction can be determined are sinistroverse, the rest is dextroverse. Dextroverse inscriptions occur more frequently on rocks, as well as in Magrè. Real boustrophedon writing, i.e. the lines of one inscription being written alternately running towards the right and the left, is not attested, but a handful of inscriptions are written in so-called reverse boustrophedon, which means that all lines have the same orientation, but are inverted in relation to each other (e.g. WE-3). For a few inscriptions it can be demonstrated that the writer changed the way they held the object during the application of the characters, which lead to a change in writing direction. This implies that the choice of writing direction was not something that a user of Raetic writing regarded as important. On the other hand, single letters turned against writing direction are rare. Where they do occur, it is most usually Alpha or Sigma.
Word separation and syllabic punctuation
Word separation by punctuation mark is only employed in a handful of inscriptions from Sanzeno context (SZ-30, NO-3, NO-10, BZ-3, BZ-26, SL-2.1), using one to (most often) three vertically arranged dots or short vertical lines. PU-1 would be the only Magrè-alphabet inscription with separators, but the existence of the respective scratches is highly doubtful. A space is used to separate words on some of the Sanzeno bronzes (SZ-1.1, SZ-2.1, SZ-2.2, SZ-4.1, SZ-11), as well as in other inscriptions from Sanzeno context (BZ-10.1, BZ-12, CE-1.3, CE-1.5), and once in Magrè (MA-1). See also Non-script notational systems on para-script elements in inscriptions from the Vinschgau.
Apart from the space in MA-1, word separation does not exist in inscriptions from Magrè context. Instead, syllabic punctuation is employed in some subcorpora. The practice of syllable punctuation is a speciality of Venetic writing, where the rules are highly complex and the letters are usually marked on both sides (for details see Prosdocimi 1988: 336 ff.). In Raetic inscriptions (as indeed in some marginal Venetic traditions), the rules appear to have been somewhat relaxed – for example, isolated vowels in the beginning of inscriptions are not punctuated, and neither are the second elements of diphthongs. In the same vein, the letters are marked with a single punct placed behind (or inside) it.
- Serso: SR-4, SR-6, SR-7 and SR-10 have correct punctuation according to the supposed Raetic rules. For example: SR-6 a ru se θa r· na te ri s· na. It is not sure that SR-1 has punctuation at all, but note the similarity with the fragmentary SR-7. If line 1 is punctated, the punct is situated before rather than after Mu. In line 2, note that the not-punctuation of Khi would be in line with the Venetic special rules for consonant clusters (kv being one of the clusters which is not punctuated). On the status of s as the first element of clusters, see below. SR-8 also has a dubious element, but the punctuation of final s is correct. In SR-2, the punct appears to mark the genitive ending rather than a phonetically determined element (see below).
- Magrè: All of the five (or six) inscriptions with syllabic punctuation can be argued to be correctly executed. MA-14 is exemplary, as is MA-17 according to the Venetic rule excluding clusters whose second element is l (here kle). In MA-16, the dubious element resembling Phi must be treated as . MA-12 and MA-13 are correct if clusters with s as first element (here st/sθ) are assumed to be exempt from punctuation, which is not in line with the Venetic rules, but phonetically apprehensible. Finally, in MA-6, the punct after final s in line 2, if it is intentional, is the only punct necessary in the entire inscription, the cluster θr being covered by the Venetic rules of exception.
- In the petrographs, the situation is more complex. ST-4 is punctuated correctly according to the rules deductible from the Serso and Magrè inscriptions. In ST-5, ST-6 and AK-1.11, on the other hand, what appears to be marked is not phonetically, but grammatically determined elements, most strikingly the suffixes of the syntagma -nu-ale. In fact, considering that the letters of the suffix -nu are in the present cases written in ligature, the puncts might even advert to that. While the punct in ST-6 sa?al·esta- may well be a word separator (if estanuale is connected with estua(le)), those in ST-5 ker·akve and AK-1.11 ker·anu- might be either word separators or suffix markers, and ST-5 (h)e·stula- would even qualify as syllabic punctuation. The single punct in ST-8 might be any of the three.
- In the area of Verona, we find a number of inscriptions where puncts in the shape of short verticals, often in the upper or lower area of the line rather than in the middle, appear not so much to mark bothersome consonants, but to replace vowels (VR-2, VR-4, VR-10, VR-11, VR-14). Whether this phenomenon is linguistical or graphical is unclear. VR-17 seems to display syllabic punctuation according to Venetic rules (with marked i in a diphthong), which together with its four-barred Mu indicates Venetic writing. See also VR-6.
- Currently incomprehensible punctuation practices were employed on the Ganglegg and in Trissino, where a number of inscriptions, some of dubious linguistic relevance, are gayly punctuated without obvious signification. In VN-11, at least, the puncts appear to separate the inscription per se from para-script elements.
- Isloated finds with potential syllabic punctuation are TV-1.1 (highly irregular), CE-1.3 (correct syllabic punctuation), and PA-1 (probably the same), for which see the inscription pages.
In two inscriptions (IT-5, RN-2), the text is written into a grid. For delimitation signs, see Non-script notational systems.
grid lines (2); word separation (, , , ) (6); word separation or syllabic punctuation (, ) (23); syllabic punctuation (, , ) (7); ligatures with punct (11); space (20)
Ligatures are rare in Raetic, but considering that none (to our knowledge) are known from Venetic and Lepontic inscriptions, they might almost be considered a speciality of Raetic. We know both actual ligatures of letters, and inscribed punctuation marks (which are treated as ligatures in TIR).
Inscribed punctuation marks are exclusively syllabic puncts (as opposed to word separators). The practice is well known from Venetic, e.g. ???. In Raetic, they occur in the inscriptions of Magrè (MA-12, MA-13, MA-14, MA-16, MA-17) and Serso (SR-1, SR-6, SR-7, SR-10) as well as in PA-1 and possibly TV-1.1. The letters into which puncts are inscribed are Mu , Lambda and Rho ; the punct can be either a dot or a short stroke . In the cases of and , the punct is essentially just placed under the bar(s) of the letter, which must not necessarily be considered an inscribed punct, but may in some cases simply be due to space-saving or chance. being attested at both Magrè and Serso, and are also filed as ligatures in TIR. As concerns the two isolated inscriptions, they both have only – while in PA-1 the punct seems quite deliberately placed inside the angle formed by hasta and bar, its position in TV-1.1 is probably accidental ( is twice followed by a punct in the inscription rather than having it inscribed).
Ligatures of letters occur in the petrographs of Steinberg and maybe Achenkirch, and in the Non valley. ST-5 and ST-6 both have an element = inverted and turned Nu + Upsilon writing nu, more precisely the patronymic suffix -nu. In AK-1.11, the reading is doubtful. (See also AK-1.17 for another possible ligature.) It is not clear why just these two letters should be ligated, as other consecutive pairs of letters in the mentioned inscriptions would lend themselves to being combined in the same manner, i.e. the bars of the first letter being attached against writing direction to the hasta of the second one. The same is true for the ligature = turned Lambda + Tau writing lt, attested only once in NO-3. The manner of forming the ligature is the same, and again there are other letter pairs in the inscription which might well be ligated in the same way.
NU ligature (1); LT ligature (1); ligatures with punct (11)
Note that Marchesini reads ligatures with inverted Alpha in VR-2 (MLR 45) and VR-6 (MLR 291), and a ligature of Pi and Sigma in VR-13 (MLR 123). These are ad hoc-readings of epigraphically difficult inscriptions without linguistical rationale, and are therefore at this point not filed in TIR.
|AIF I||Carl Pauli, Altitalische Forschungen. Band 1: Die Inschriften nordetruskischen Alphabets, Leipzig: 1885.|
|Colonna 1972||Giovanni Colonna, "Clusium et Orvieto", Studi Etruschi 40 (1972), 470–471.|
|ET||Helmut Rix, Gerhard Meiser (Eds), Etruskische Texte. Editio Minor [= ScriptOralia 23-24; Reihe A, Altertumswissenschaftliche Reihe 6-7], Tübingen: Gunter Narr 1991. (2 volumes)|
|Gamper 2006||Peter Gamper, Die latènezeitliche Besiedlung am Ganglegg in Südtirol. Neue Forschungen zur Fritzens-Sanzeno-Kultur [= Internationale Archäologie 91], Rahden/Westfalen: Leidorf 2006.|
|Jeffery 1990||Lilian H. Jeffery, The Local Scripts of Archaic Greece. A study of the origin of the Greek alphabet and its development from the eighth to the fifth cernturies B.C., Oxford: 1990.|
|Lexicon Leponticum||David Stifter, Martin Braun, Michela Vignoli et al., Lexicon Leponticum. URL: http://www.univie.ac.at/lexlep/|
|Lunz 1981b||Reimo Lunz, Venosten und Räter. Ein historisch-archäologisches Problem [= Archäologisch-historische Forschungen in Tirol Beiheft 2], Calliano (Trento): 1981.|
|Mancini 1980||Alberto Mancini, "Le iscrizioni della Valcamonica. Parte 1: Status della questione. Criteri per un'edizione dei materiali", Studi Urbinati di storia, filosofia e letteratura Supplemento linguistico 2 (1990), 75–167.|
|Marchesini 2013||Simona Marchesini, "Descrizione epigrafica della lamina", in: Carlo de Simone, Simona Marchesini (Eds), La lamina di Demlfeld [= Mediterranea. Quaderni annuali dell'Istituto di Studi sulle Civiltà italiche e del Mediterraneo antico del Consiglio Nazionale delle Ricerche. Supplemento 8], Pisa – Roma: 2013, 45–53.|
|Marinetti 2004||Anna Marinetti, "Nuove iscrizioni retiche dall'area veronese", Studi Etruschi 70 (2004), 408–420.|
|Marinetti 2004b||Anna Marinetti, "Venetico: Rassegna di nuove iscrizioni (Este, Altino, Auronzo, S. Vito, Asolo). (Rivista di Epigraphia Italica)", Studi Etruschi 70 (2004), 389–408.|