Classical Greek accentuation
- From: "Douglas G. Kilday" <fufluns@xxxxxxxxxx>
- Date: 11 Mar 2007 13:27:03 -0700
Google has dropped 3 of my last 4 postings on Greek accents to the
existing thread, so I will attempt to start a new thread here, with a
summary of the material in the lost messages.
First, regarding Richard Wordingham's allegedly snappy rule about any
morae between the accented and final morae having to occupy the same
syllable, this works for <ánthro:pos> and the like, but snaps in half
when confronted with proparoxytones having long ultimae. Examples of
the latter are gen. sg. and gen. pl. forms of certain nouns in -is and
-us, and certain compound adjectives: <póleo:s> 'of a city',
<pé:kheo:n> 'of cubits', <hupsíkero:s> 'high-horned'.
Second, regarding Neeraj Mathur's posting on the utility of morae in
analyzing Greek accent, I notice that a polysyllable with recessive
accent may place the accent from 2 to 5 morae from the end. E.g.:
ánthro:poi = án-thro-o-po-i, 5 morae back
athánatoi = a-thá-na-to-i, 4 morae back
athánatos = a-thá-na-tos, 3 morae back
prophúlax = pro-phú-laks, 2 morae back
Applying the traditional kludge that final unaccented -ai and -oi
(other than loc. and opt. forms) are "counted as short" for accent,
i.e. as unimoral -aj and -oj, eliminates the first class. Applying
another kludge that final -ps and -ks are counted as morae (i.e.
final /s/ following /p/ or /k/ is quasi-vocalic) eliminates the last
class. The discrepancy between <ánthro:pos> and <athánatos> (accent 4
or 3 morae back) remains. Furthermore, additional kludging (as in the
traditional presentation) is still required to handle the barytone
loc. and opt. forms, for which there is no evidence of bimoric
pronunciation of -ai and -oi (i.e. as -aï and -oï), and the loc.
<Isthmoî>, a perispomenon, argues directly against a bimoric loc.
suffix in the attested language.
Other peculiarities of the accent, like the shifting in nouns with
monosyllabic nom. sg. forms, the anomalies of some contracted
adjectives, and the recessive voc. sg. of some (but not all) compound
nouns, must be treated separately whether one works with morae or
syllables. Thus I see no advantage whatsoever in a mora-based
approach to classical Greek accentuation.
Third, regarding Neeraj's later posting, I admit that not many
barytones end in -ai or -oi other than those nom. pl. forms of nouns,
adjectives, participles, and demonstratives, those infinitives and
finite mediopassive verbal forms, and our anomalous loc. and opt.
forms. The adverb <pálai> 'of old, formerly' and the voc. sg. <gúnai>
'O woman' come to mind, but having short first vowels they do not
illustrate the rule. Voc. sg. forms of certain compound adjectives
should work, although I have no citations of their use. <eúpais>
'having good children' should have an Attic voc. sg. <eûpai> (in Epic
one expects uncontracted <eúpaï>). *<eúois> 'having good
sheep' (which can hardly _not_ have been used, even if I have no
attestation) should have an Attic voc. sg. *<eûoi> (thus distinct from
the Bacchanalian cry <euoî>, which is probably the contracted 3sg.
pres. opt. of an obsolete verb 'to be well'; again one expects an
uncontracted Epic form).
As for the theory that nom. pl. barytones in -ai and -oi acquired
their accent by analogy with nom. sg. forms, the principal objection
is that nom. sg. and nom. pl. forms do not always agree in accent.
E.g.: <axía:> nom. sg. f. 'worthy', <áxiai> pl.; <ale:thé:s> nom. sg.
m./f. 'true', <ale:theîs> pl.; <pêkhus> nom. sg. 'cubit', <pé:kheis>
pl. Thus there was no compelling reason to change the established
accent of large classes of plural nouns by analogy. For verbs, active
and mediopassive forms may disagree not only in accent, but even in
number of syllables, e.g. in the 2nd aor. indic.: 1sg. act. <élipon>,
mid. <elipóme:n>; 2sg. act. <élipes>, mid. <elípou>; 3sg. act.
<élipe>, mid. <elípeto>, ktl. Thus there was even less reason to
change the established accent of mediopassive forms by analogy. And
you already admitted to finding no analogy for the infinitives in -ai.
As for the notion that "markedness hierarchy" or some similar ModLing
dazzler can explain the anomalous loc. and opt. forms, the Greek
locatives are already fossils, but the optatives are not. How can
such dissimilar classes of forms end up at the same height on the
markedness totem pole? I shudder to think of the effort required to
develop "markedness hierarchy" for an ancient language, even one so
abundantly attested as classical Attic.
I think the best way to explain the anomalies is, as described before,
by postulating that most final -ai and -oi were in fact -aj and -oj
when the Attic accentual system crystallized, but that loc. and opt.
forms were -aï and -oï, only becoming -aj and -oj afterwards. This
also explains the loc. sg. <Isthmoî> from *<Isthmóï> without any
additional assumptions. I doubt that Occam would have smiled on morae
and markedness in dealing with Greek accentuation.
.
- Prev by Date: Re: Does natural language skill translate to programming skill?
- Next by Date: Re: Ethiopic writing Re: intrinsic advantage of Latin alphabet over bopomofo (for Chinese)??
- Previous by thread: Wopwops
- Next by thread: Re: Question About Lingustic Notation
- Index(es):
Relevant Pages
|
