You bring up an appropriate word of caution. Greek in the classical
period was always characterized by regional dialects and pronunciations.
As a result, one must try to make statements concerning more specific
places, times, and words. Nevertheless, some generalizations can still
be made. Let's look at the evidence for the value of the upsilon
through time. My main source is W. Sidney Allen, upon whom I rely
heavily. cited in another post.
The earliest evidence that the upsilon had become a high, front,
rounded vowel (like the German =FC or ue) are Ionic inscriptions
from the 6th cen. BC which occasionally spelled the diphthongs
AU and EU as AO and EO, respectively. Somewhat later, Herodotus'
transliterations of Old Persian [wi] as U provides additional
evidence for Ionic.
In the 4th cen. BC, there's evidence from the Boeotianian's use
of OU rather than U for [u] sounds in their dialect. A century
later, the Boeotians were using U for their vowel sound cognate
to Attic OI. This suggests that their cognate to Attic OI was
something like the German =F6 (oe) converging on =FC (ue). Finally
we get to the Latin evidence. When the Romans became more
educated about they borrowed the upsilon as a Y,y to represent
that Greek sound. Additional support that the sound of the
upsilon did not exist in Latin comes from Cicero and Quintillian
in the 1st cen. B.C.
Also, in the 1st cen. B.C. Dionysius of Halicarnassus described
the vowel in terms that indicate it was still high, front, and
rounded. As late as the 2nd cen. A.D. in Athens there still are
spelling errors between OU and U (e.g. XROUSOU for XRUSOU), which
means that it was still rounded, not unrounded like the iota's sound.
Moreover, bishop Wulfila in the 4th cen. A.D. found it necessary to
adopt the Greek upsilon rather than I in representing Greek loan-
words in his Gothic translation. The last bit of evidence that
shows that the conversion of the sound of the upsilon had not
completely changed until the ninth is the Byzantine's name for
the letter, U PSILON in contrast with OI DIFQOGGOS. Thus, U was
distinguished from OI, not I.
When did upsilon start unrounding to the sound of an iota? Egyptian
papyri from the 2nd and 3rd centuries show a confusion of U and I,
so this indicates that around that time, possibly a bit earlier,
is when the U started converging to I, and that probably is a regional
It is doubtful that any NT text was actually written in Egypt, since
the church was a rather late comer to the area, and Acts is usually
claimed to be written elsewhere, such as Rome, Caesarea, or Achaea.
Thus, we have good evidence for the rounded pronunication in those
places a century before its composition, and some of those places
(Achaea) a century later.
As for H (eta), good old Dionysius in the 1st cen. B.C. still
distinguished it from I, but confusion with I starts in Attica
in the 2nd cen. across the Aegean. Even if some people started
pronouncing H as I in the 1st cen., U in most places still retained
its non-itacistic value.
Therefore, I'd say it's a safe bet that hUMIN and hHMIN were
pronounced differently at all relevant times and places the
NT was probably composed. That statement no longer obtains
concerning the times and places the text was copied.
-- Stephen C. Carlson : Poetry speaks of aspirations, email@example.com : and songs chant the words. http://www.mindspring.com/~scarlson/ : -- Shujing 2.35