Thanks, Mark, for the Marvell reference. There are some interesting points
you make:

>>In this sense to replicate the form is to mis-translate. How much moreso
in the case of the ghazal, one would think.<<

Yes and no, I think--though of course all translations are mistranslations.
The question, or at least one way of putting the question, is whether and to
what degree you want to translate the aural experience of the form, which of
course means that a reader will have, in the target language, the cultural
associations attached to that form in her or his culture, not the culture of
the original. The example you give of the sonnet is a wonderful illustration
of this. In the case of the ghazal, there is something about the repetitive
rhyme scheme--the last line of each couplet reproduces the same rhyme, which
is set up in the first couplet--and the way it stitches together the
couplets that do not necessarily have any linear connection that I think is
worth carrying over into English translation when it is possible. The few
examples I have seen--one from Rumi, a couple from Hafez--where the
translator has been able to do that, or at least to come close it, have been
very effective. (The nearest analogy I can think of for the ghazal is the
villanelle, with its repeating lines, though the villanelle's three line
stanzas, of course, are connected much more linearly than the ghazal's
couplets.)

>>At this point the ghazal as practiced variously in the West, usually as a
set of often unrhymed couplets with no linear connection between them, has
become a form in its own right, no longer particularly referential to the
Persian source.<<

I think that the ghazal came into the West primarily through its Urdu
version, rather than through Persian. There is apparently the same kinds of
differences within the ghazal tradition--Persian, Urdu and one other
language, maybe Arabic, that I can't remember--that there are within the
sonnet. Different cultural contexts, different points of reference, etc. But
one of the interesting things about the way the ghazal has come into English
poetry is the total disregard for its formal constraints, as opposed, say,
to the sonnet, for which the formal constraints, or at least an analogous
set of constraints, were carried over from one language to another.

>>In translating a poem from a language in which simple rhyme is both easy
and ubiquitous is one rendering or distorting the sense of the original by
struggling to maintain end-rhyme in the very different environment of modern
English? The question wuld remain even if by miracle one managed to
replicate the rhyme pattern without sacrificing anything else.<<

For me, as I said above, the issue is the extent to which the aural
experience is important. With the ghazal, which I am not translating, I
think it is, if only because the form is so specific in its requirements,
and it is the rhyme the stitches the couplets together, even if that
experience ends up having slightly different meanings because of the
different cultural context of the translation. In my translations of Saadi,
however, I have decided that the aural experience of rhyme is not central in
any way, which is why I am working in blank verse, a form that offers a
different kind of aural experience that for me is analogous or parallel to
my own experience of hearing Persian-speakers recite poetry.

Richard