|Customary name:||san, sade|
|Is used in the representation of:||ts, d (?)|
Variants and attestation
|Glyph||N. Italic||other||Glyph||N. Italic||other|
The letter usually called san (occasionally sade) is ultimately derived from the Phoenician letter sade via the intermediary letter san of the Greek and North Etruscan alphabets.
Which Glyphs Represent san?
Michel Lejeune (1971: 17–19) was the first modern scholar to discuss the formal variation in the graphic representation of the letter san, observable in the Lepontic corpus. On the basis of thirteen tokens of the letter known to him and accepted by him as authentic, he identified five different glyphs, that is formal types, that according to him represented the letter (see ill. Lejeune 1971: 18). For easier reference, descriptive names will be given to each of those types: the top-left variant will be called 'double-axe san' (= LexLep Ś5), the top-right one 'tipi-san' (= LexLep Ś4), because it resembles a native American conical tent. The one below those two is the well-known 'butterfly'-variant of the letter (It. san a farfalla = LexLep Ś1), the most frequently occuring variant. Below it can be found the variant which will be called 'double-pennon-san' (= LexLep Ś6), and at the very bottom is the 'trouser-san' (= LexLep Ś2), identical in appearance to a modern Roman M. In the case of butterfly-san , the iconic technical term has a long tradition.
Lejeune's authority has been crucial for all subsequent studies of the letter, although certain of his tenets did not go unchallenged. In particular, his identification as san of the tipi-like character has been questioned. Lejeune knew of two occurences of it in PID 304 (= VB·3.1), i.e. "naśom", and PID 334 (= MI·1), i.e. "peśu". To these may be added VA·1.2 featuring a similiarly shaped letter which has been read as san, for example, by Solinas (1995: nr. 113), i.e. "śu", albeit hesitatingly.
- Although the latter inscription was obviously known to Lejeune (in his study, he discusses VA·1.1 kasikos), he did not take it into consideration and did not comment upon it, presumably because it was too short. The first letter of this graffito is certainly a T with a twisted right leg; the reading of VA·1.2 is consequently tu.
- With regard to VB·3.1, Tibiletti Bruno (1997: 1009 fn. 13) argued that the bottom line of the letter was due to an impurity in the moist clay, caused by the rotation of the vase on the potter's wheel. Indeed the trace of the bottom line is deeper and larger than those of the other hastae of the letter or of the neighbouring letters, and it is exactly parallel to the base of the vase. According to Tibiletti Bruno, the letter is T and the word is natoś. Tibiletti Bruno's suggestion is partly followed by Morandi (2004: 550-551) who reads natom with a question mark. A different strategy was chosen by Stifter (forthc.: fn. 16). Starting from the popular etymology of naśom as an adjective referring to "Naxian (wine)", he suggests that the letter is a Latin X with a diacritic understroke to distinguish it from native Lepontic T of identical shape, or with a subscript I, in order to salvage the phonological interpretation as /naχs(i)i̯om/ (vel sim.). In either case, VB·3.1 has to be dropped as evidence for the tipi-variant of san.
- The explanation of being due to an accidental scratch during the production of the vase on the potter's wheel is certainly to be adopted for the bottom line of the third letter in MI·1, as a glance on a photo of the inscription reveals. The word has to be read as petu with T (thus implicitly also Tibiletti Bruno (1997: iconografia p. 14) and Morandi (2004: 610)).
Thus, all possible examples for tipi-san turn out to be representatives of T. At least in the case of MI·1 petu, this entails a simpler and better-founded etymology.
Things get more entangled if we turn to the fifth of Lejeune's variants of san, the one termed trouser san above. In Lejeune's book, it is based on a single example, i.e. PID 300 (= VA·6). Joshua Whatmough (1993: 105-106) himself provided no reading for the severely corroded letter, but confined himself to reporting the attempts by earlier scholars to decipher the inscription. None of these attempts featured san. The identification of the illegible letter as san was made by Tibiletti Bruno (1969c: 187) and has been accepted and corroberated since. In the drawing which accompanies Tibiletti Bruno's article, the letter does indeed have the shape of a Roman M, although the middle lines protrude slightly over their point of intersection. Subsequent examinations of the stone, however, have revealed that the relevant letter originally had the butterfly shape before it was damaged (see, for example, the depiction in Morandi 2004: 593). Consequently, the material basis on which Lejeune founded his identification of this particular variant of san has vanished. Nevertheless, for subsequent scholars Lejeune's diagramme, together with studies of Tibiletti Bruno's (e.g. 1968b), provided the justification for discovering more examples of the trouser-shaped variant of san. The matter is complicated by the fact that in the ancestorial Etruscan and cognate Venetic or Raetic alphabets san also had the trouser-like shape, and it is found in the Ligurian (?) inscription SP·1. Therefore it was tempting to posit it implicitly as the starting point for an archaic stage of the Lepontic writing system. Currently, altogether three examples of trouser san have been identified in one-word inscriptions from the Swiss site of Giubiasco, viz. TI·5 ariśai, TI·7 aśu?, and TI·9 reśu. This is the state of the art reflected in Alessandro Morandi's collection of Italo-Celtic inscriptions. Morandi (2004: 477) merely observes geographical divergences in the use of the variants of san, but draws no further conclusions from it: "A Ornavasso, Solduno e a Prestino il sade [editor's note: = san] è a farfalla mentre a Giubiasco è prevalentemente del tipo a quattro tratti [editor's note: = trouser san]." It must be noted that Lejeune (1971: 49-50, 63-64) himself, following PID 256 and 258, read the latter two names with M, i.e. amu? and remu; the first one of these names, very cautiously read erimia·i in PID 262, is not included in Lejeune's study. Patrizia Solinas (1995: ???), on the other hand, is inconsistent in her analysis of the texts from Giubiasco. Although TI·5, TI·7, and TI·9 all contain the identically shaped letter, she reads M for the first of them, but Ś in case of the latter two.
In my opinion, only one conclusion can be drawn from the distribution of letters across space and time. It would be quite remarkable to find an archaic shape of the letter Ś san, absent from all the rest of the Lepontic corpus, in particularly late inscriptions from the 1st c. BC at the very north-eastern end of the Lepontic world. Geographical remoteness might be invoked as an argument in favour of such a notion, but the standard butterfly shape is found in the same area, in the same period in one (TI·13) or two (TI·2) other inscriptions. It is much more likely that in TI·5, TI·7 and TI·9 we are looking on a variant of M mu. It is true that the letter mu in its trouser-shaped variant is also rather rare in Lepontic inscriptions, as opposed to its "twin-peaks" variant (= LexLep M6 ) which is regularly encountered in Late-Lepontic inscriptions and in Latin graffiti from the region (verweis). However, rarety is not absence: certain examples for trouser-shaped mu can be found in VR·12 musu, a sinistroverse inscription that could both belong to the Lepontic or Latin script, as well as in VB·27 namu, an inscription that makes use of an outright hybrid Latin-Lepontic script. For what it is worth, it may be noted that in TI·10 uenu from the same site the letter N is written in a manner comparable to trouser-shaped M, that is, with two strictly upright outside hastae. Taking all things together, this means that there is certain evidence that the trouser-shaped character could be used to represent M in the Lepontic corpus, while there is no conclusive evidence to support the case of putative trouser-shaped Ś. Unfortunately, there are no inscriptions from this remote Lepontic district which combine both Ś and M. Nevertheless, in absence of a final proof, the scales are tipped in favour of interpreting that letter in Giubiasco as a variant of mu rather than of san. It is a side effect of this hypothesis that it allows for neat etymologies of two of the three concerned names, i.e. amu? and remu.
Even though Lejeune (1971: 17–19) made no statement about the chronological development of the letter variants, the layout of the diagramme and the directions of the arrows used in it suggest that butterfly san was the fundamental glyph from which all others were in some way derived. While this scenario is most likely correct for the double-axe symbol - the three inscriptions that contain it seem to be relatively late according to Solinas (1995: 374, 379, 381) and Morandi (2004: 567, 585, 589), viz. VB·25, NO·21g, perhaps VC·1 - the examples for the double-pennon symbol are spread over a much longer space of time, from the earliest period of the Lepontic literary tradition in the 6th or 5th c. BC, viz. CO·37 until the late period, e.g. VA·16 akeśi in the 1st c. BC. Morandi (2004: 470) suggests that the presence of this variant of the letter is due to south-Etruscan influence. At the same time, the double-pennon symbol provides an adequate transitional stage from the most archaic shape of the letter san, the trouser-shaped character on its way to the standard butterfly symbol of the Lepontic script, the attestation of which sets in slightly later then double-pennon san. The fact that trouser san is actually absent from the attested Lepontic corpus, as argued above, is of no significance for this point because there can be no doubt that that variant of the letter must have been present in the model alphabet, inherited from its North-Etruscan precursor script.
In addition to the five formal variants of the letter discussed by Michel Lejeune, another one has been put forward by Maria Grazia Tibiletti Bruno in various publications (e.g. 1981: 163; 1997: 1009 fn. 13). While the twin-peaked character, which occurs thrice in the Late-Lepontic text VB·3.1, has been traditionally interpreted as M, Tibiletti Bruno wants to read it as a special variant of san. She develops her argument by the exclusion principle: since on the same object, which bears several more or less independent inscriptions, we also find VB·3.2 mou·ea with an archaic, flag-shaped mu, the presence of a younger variant of the same letter on one and the same object is excluded for her. Since, therefore, the twin-peaked character cannot stand for M, it must represent something else. The only option open to her is to posit an isolated variant of san, a letter which admittedly possesses formal variants of roughly comparable shape. Naturally, this decision forces Tibiletti Bruno to come up with a reading and an analysis of VB·3.1 that diverge considerably from the mainstream interpretation. Tibiletti Bruno's hypothesis is altogether unconvincing. The following points can be raised against it:
- The words latuśarui, uinoś, and natoś, which result from her reading of the inscription, are grammatically and etymologically problematic. For a detailed criticism see the relevant pages.
- It is beyond dispute that the twin-peaked symbol is otherwise used for M in Late-Lepontic texts, as well as in contemporary Latin graffiti which apparently exerted influence on the Late-Lepontic writing tradition.
- From a graphical point of view, as a representation of san the twin-peaked character could only have been derived from the trouser-shaped variant of that letter. Since, however, as was argued above, the Lepontic corpus contains no reliable example of trouser san, twin-peaked san would lack a viable internal source.
- In no other vernacular writing tradition of northern Italy outside of Lepontic the twin-peaked character can be found representing san. There is thus no external source for it either.
Tibiletti Bruno (1997: ???) offers a somewhat detailed discussion of the letter variants. (Cf. Tibiletti Bruno 1975: 50-51, 52)
We are now in a position to arrive at a summary of the foregoing observations:
- Nowhere in the Lepontic corpus can be found an unassailable example for the trouser-shaped variant of san, i.e. Ś2 , even though, being an integral part of the character inventory of the ancestorial scripts, it must have been present in the model alphabet on which the Lepontic script is based. Ligurian SP·1 is not counted here.
- All examples of tipi-shaped san are poorly executed Ts.
- There is no twin-peaked variant of san; that letter represents M.
- All three securely attested variants of san in some way involve the butterfly design, i.e. butterfly san proper Ś1 , double-pennon san Ś6 , and double-axe san Ś5 .
- Of these, the double-pennon shaped variant seems to be the earliest; the butterfly variant is a graphic simplification of it; the double-axe variant is clearly derived from the butterfly character and it is attested only in the latest period.