hi seirios, the book you will want to read which answers your question is Philomen Probert, Ancient Greek Accentuation: Synchronic Patterns, Frequency Effects, and Prehistory, 2006. the very start of the book begins:
“1 EVIDENCE FOR THE GREEK ACCENT
1.1 Introduction
When we read any classical or pre-classical Greek author in a modern
printed text, we read a text with accents marked using the acute,
circumflex, and grave signs. But classical and pre-classical
authors did not write accent marks, which almost certainly had not
been invented in their day (cf. pp. 19, 21). An obvious question to ask
is therefore how, and indeed whether, we can have accurate information
about the accentuation of Greek at a period when signs for accents had
not been invented. Before this question can be answered in full, a
prerequisite is to consider the origins and basis, even if post-classical,
of the accent marks we read in our texts, and to investigate the other
sources of evidence for the accentuation of Greek at any period in
antiquity. If, as we shall see is the case, all direct sources of evidence
are post-classical, it may nevertheless be possible to make some deductions
about the accentuation of earlier periods, but we must begin with
the actual sources. The purpose of the present chapter, therefore, will
be to survey the various types of evidence available and to explicate
their basis. Although in evaluating these sources we shall have occasion
to touch on matters relating to early periods of Greek, the question of
deductions that can be made about the classical and pre-classical
periods will be taken up in full in Chapter 3.”
Then if you go to chapter 3 there’s a lot of discussion on this, including:
"3.1 Introduction
Our evidence for ancient Greek accentuation is tantalizingly achronological.
As we have seen, the grammatical tradition relating to the
ancient Greek accent has its roots in Hellenistic Alexandria, with occasional
elements of possibly much higher antiquity. The tradition is
essentially fixed by the second century AD. Several centuries separate
Aristophanes of Byzantium from Herodian, yet we have little guidance
as to changes in accentuation that may have occurred over that period.
Furthermore, as Scheller (1951) has well observed, the whole period
preceding Aristophanes of Byzantium is for Greek accentuation a
period of prehistory: for no time during this period do we have direct
access to the state of the accent system. It is therefore essential to
consider whether we are in a position to treat ancient Greek accentuation
from a diachronic perspective at all.
In what follows I hope to show that despite the lack of direct evidence
for Greek accentuation in our period of ‘prehistory’, certain deductions
may be made; these point to continuity in certain areas and change in
others. We shall begin by surveying the evidence for continuity between
the accentuation systems of certain periods and those of others, and
then consider the indications of some salient diachronic changes.
3.2 Continuity
Correspondences between the position of the accent in ‘ancient Greek’,
as we know it from our Hellenistic and post-Hellenistic sources, and
that in other Indo-European languages, especially Vedic Sanskrit, show
beyond reasonable doubt that some features of our ancient Greek accent
system are extremely old. We have already observed (p. 26), for example,
the parallel movement of the accent between root and ending
in the paradigm of the Greek word for ‘foot’ and in its Vedic cognate,
and the parallel accentuation of the numbers from Wve to ten in Greek
and Vedic. As noted already (p. 26), the incidence and often systematic
character of such correspondences demonstrate that for many words the
position of the Greek accent has remained the same from late Indo-
European until the Hellenistic period. It should follow a fortiori that
these words were accented on the same syllable at intervening periods
such as the time of Homer or that of Sophocles, or at least that they were
so accented at intervening periods in Attic, the dialect that was primarily
ancestral to the Koine. Since, moreover, the occasional grammarians’
comments we have on Ionic accentuation suggest that in most
respects Attic and Ionic had not diverged in their accentuation by the
Hellenistic period, we may assume that during the archaic and classical
periods many words were already accented both in Ionic and in Attic in
the same way as they would later be accented in the Koine.
Further evidence for continuity between the accentuation of archaic
and classical Greek and that of the Hellenistic period comes from the
information we have on the accentuation of dialects other than Attic and
the Koine (see § 2.11 above). We have observed that it is not clear what
the basis for this information is, whether grammarians’ first-hand
acquaintance with dialect speakers in the Hellenistic period or some
sort of tradition regarding the pronunciation of literary dialects. Even
without knowing the source or antiquity of this information, however,
we may at least tentatively draw some conclusions from the similarities
between the accentuation systems of different dialects at some point in
the historical period.
Most saliently, the law of limitation is common to all Greek dialect
groups for which we have evidence on accentuation: Attic-Ionic, Doric,
and Aeolic. The law of limitation is a Greek innovation: it does not go
back to Indo-European. The fact that it is common to the three major
dialect groups for which we have accentual information suggests, however,
that it is a very early innovation, dating back to a period at which
the major dialect groups were not yet differentiated. We may again
argue a fortiori that the law of limitation applied both to classical Attic
and to the language of Homer…". It goes on…
the question of which phonology to use can’t, therefore, be answered by reference to the period when accent marks first turn up on papyri.
how to answer the question then, which phonology to use? answer: however you like, de gustibus non est disputandum.
cheers, chad
