Sigma and san The existence of the two competing letters sigma and san is traditionally believed to have been due to confusion during the adoption of the Greek alphabet from the
Phoenician script, because Phoenician had more
sibilant sounds than Greek had. According to one theory, the distribution of the sibilant letters in Greek is due to pair-wise confusion between the sounds and alphabet positions of the four Phoenician sibilant signs: Greek sigma got its shape and alphabetic position from Phoenician
shin (), but its name and sound value from Phoenician
samekh. Conversely, Greek
xi () got its shape and alphabetical position from samekh (), but its name and sound value from shin. The same kind of pair-wise exchange happened between Phoenician
zayin and
tsade: Greek
zeta has the shape and position of zayin (𐤆) but the name and sound value of tsade, and conversely Greek san has the approximate shape and position of tsade (𐤑) but may originally have had the sound value of zayin, i.e.
voiced . "san" was indeed the original name of what is now known as sigma, and as such presents a direct representation of the corresponding name "shin" in that position. This name was only later also associated with the alternative local letter now known as "san", whose original name remains unknown. The modern name "sigma", in turn, was a transparent Greek innovation that simply meant "hissing", based on a nominalization of a verb (, from an earlier stem , meaning 'to hiss'). Moreover, a modern re-interpretation of the sound values of the sibilants in
Proto-Semitic, and thus in Phoenician, can account for the values of the Greek sibilants with less recourse to "confusion". Most significant is the reconstruction of shin as and thus also the source of the sound value of sigma; in turn, Samekh is reconstructed as the
affricate , which is a better match for the plosive-fricative cluster value of xi. Whereas in early
abecedaria, sigma and san are typically listed as two separate letters in their separate alphabetic positions, each Greek dialect tended to use either san or sigma exclusively in practical writing. The use of san became a characteristic of the
Doric dialects of
Corinth and neighboring
Sikyon, as well as
Crete. San became largely obsolete by the second half of the fifth century BC, when it was generally replaced by sigma, although in Crete it continued in use for about a century longer. In Sikyon, it was retained as a symbolic mark of the city used on coin inscriptions, in the same way that archaic
koppa () was used by Corinth, and a special form of
beta by
Byzantium). San could be written with the outer stems either straight () or slanted outwards (), and either longer or of equal length with the inner strokes (). It was typically distinguished from the similar-looking
mu () by the fact that san tended to be symmetrical, whereas mu had a longer left stem in its archaic forms (, , , which is similar to the Phoenician tsade that san derives from or
mem that mu derives from). Outside Greece, san was borrowed into the
Old Italic alphabets (, transcribed as Ś). It initially retained its M-shape in the archaic Etruscan alphabet, but from the 6th century BC changing its aspect to a shape similar to that of the
d-rune . The name "san" lived on as an dialectal or archaic name for sigma, even after the letter itself had been fully replaced by it. As such,
Herodotus, in the late 5th century, reports that the same letter was called "san" by the
Dorians, but "sigma" by the
Ionians.
Athenaeus in his
Deipnosophistae (c.200 AD) quotes an
epigram which contained the spelled-out name of the philosopher
Thrasymachus, still using san as the name for sigma:
Arcadian tsan A unique letter variant (shaped similarly to modern Cyrillic
И, but with a slight leftward bend) has been found in a single inscription (therefore a
hapax legomenon) in the
Arcadocypriot dialect of
Mantineia,
Arcadia, a 5th-century BC inscription dedicated to
Athena Alea (
Inscriptiones Graecae V.ii.262) It is widely assumed to be a local innovation based on san, although Jeffery (1961) classes it as a variant of sigma. It appears to have denoted a sound and has been called
tsan by some modern writers. In the local Arcadian dialect, this sound occurred in words that reflect
Proto-Greek . In such words, other Greek dialects usually have , while the related Cypriot dialect has . Examples are: • (cf. , 'somebody') • (cf. , 'somebody') • (cf. 'to whomever') • (cf. 'either') From these correspondences, it can be concluded that the letter most likely denoted an affricate sound, possibly or , which would have been a natural intermediate step in the sound change from to . The letter has been represented in modern scholarly transcriptions of the Mantinea inscription by ⟨ś⟩ (s with an acute accent) or by ⟨σ̱⟩ (sigma with a macron underneath). Note, however, that the same glyph is used to denote the unrelated letter
digamma in
Pamphylia (the "
Pamphylian digamma") and was also the form of
beta used in
Melos.
Sampi The
Ionic letter sampi (), which later gave rise to the numeral symbol ( = 900) may also be a continuation of san, although it did not have the same alphabetic position.
Bactrian In the Greek script used for writing the
Bactrian language, there existed the letter , which apparently stood for the sound (transliterated as ), and has been named "sho" in recent times. According to one hypothesis, this letter too may go back to san.{{cite book|title=The Greeks in Bactria and India ==Modern use==