Essays on Time-Based Linguistic Analysis

There are in Ancient Greek a few hints of a ruki-development (that is, the change of [s] to [š]--in Sanskrit, to [ʂ]--when preceded by k, r, i, and u).¹ Till now, the change has been thought to have been limited to a few other Indo-European language families. The evidence presented here has chiefly to do with otherwise unaccountable switchings of Greek σσ with ξ and the fact that the examples occur in ruki-environments. It will become obvious that σσ (known to have represented [šː])² and ξ had to represent the same sound at some time in this or that place: evidence will be adduced to show that ξ represented [šː] (short where long consonants were not permitted) in the relevant lects.

The crucial evidence of the alternations of ξ with σσ is qualitatively solid and very important; but like certain archaeological fossils, it is quantitatively meagre and will require detailed and at times tedious examination and argumentation. It is important to consider the psychology of an orthography as well as the probable phonetological developments behind it. That 'x' [cs] could become [š(ː)] is not in doubt: It happened in western Romance, where it was represented by 'ss' in French and by 'x' in Iberia.³

¹

The Sanskrit ruki-rule was complex: aside from a few formatives and certain formations, non- final //s// not immediately followed by //r// changed to

Click for a larger version

when it directly followed k, r, or any vowel but short or long-a--provided that no syllabic or non-syllabic r preceded in the same word. The rule formed part of a loose section of phonetology that also changed //n// to [ɳ] under certain conditions. The Greek alternations discussed in this writing have an adjacent, usually preceding or both preceding and following, ruki-sound--including Attic [

Click for a larger version

(ː)] from [u(ː)].

²

This [š] was short where not internuclear. The present treatment of Greek sets out from the assumption that the evidence makes it indisputable (see Bailey 1986b) that ττ was [t^šː], δδ was [d^žː], and σσ (derived from [t^šː]) was [šː] or

Click for a larger version

. (Whether Hellenistic Greek zeta--derived from Attic ζ = [^žd^žː]; cf. n. 4--had become [žː] on the way to modern [ʐ] is irrelevant here.) The development of laminopalatal affricates to laminopalatal fricatives or apical assibilates and subsequently to apical sibilants is widely attested in languages; cf. the development of Old French chanter and jambe, now with [š] and [ž], and the wider development evident in modern French cent with [s] from Late Latin [t^š] via [t^s]--[z] between nuclei as in raison, with [z] derived from [d] or [d^ž] via [d^z]. Cf. Italian raggione and Spanish razón (with 'z' now representing [b] or [s]); between nuclei, cento ends up with a voiced affricate in Italian duegento--Florentine duežento).

³

Late Latin 'x' [cs] yielded 'ss' or, in certain nuclear and consonantal (especially /__t and /n__) environments, 's'. In westerly (that is, Gallic and Iberian) regions, it is generally assumed that [cs] became [çs] and then [ys] and then [š]. No reason is given--other than the (possibly irrelevant) parallelism with the development of [ɟ] and the widespread satellite [i] before 'ss' or 'x' (both [š]) from Latin [cs]--for rejecting a simple change of [cs] or [çs] directly to [šː]. (Compare the development of [yt] to [t^š] in Spanish noche.) After all, both [cs] (Latin 'x') and [sc] (= 'sc' before a non-low front vowel) became 'ix' [ⁱš] in Portuguese. A palatal [š] could (like [ɲ] in current varieties of French) then change a preceding vowel to a diphthong with satellite [ⁱ], attested in many regions.

However that may be, later borrowings complicated the results in French: compare current laisser

-301-