The norm I mostly adhere to is to have three evenly spaced ictus beats, then skip one for the caesura, then three more, then skip one at the end of the line (where there will always be a real pause), giving eight in all. At the caesura, there will sometimes be a pause, sometimes not, There are complex and inconclusive arguments about what actually happened (see the Allen book for some references). At any rate, actual pausing is one way to implement beat-skipping, but there are others, such as elongating a preceeding syllable.
In a sung style, the no-skipping technique seems to work quite nicely, but for a spoken delivery I find that it seems rushed, and doesn't provide scope for expression. However in a minority of lines (but far too many to just ignore) it creates difficulties, such as too many potential places to have a full skip, with none of them seeming especially appropriate. An example is in fact the first line of the short selection: there are three potential pause sites, none of them particularly satisfactory:
Pausing after 'philoi' seems to produce too much of a rush between 'idmenai' and 'oude ...':
Pausing at the regular caesua position is even worse:
Pausing at the last (bucolic diaresis) position is the best for me:
But none of them are wonderful.
Well what about the no-pause technique? Here are a few lines of that:
I don't like the absence of timing delays at some of the line-internal grammatical boundaries, but maybe the Greeks had no problem with that, or the rhapsodes had ways of making it sound better.
One possible solution is to abandon the completely fixed time scheme and add extra beats (this is what tends to happen in the longer piece). Another is that perhaps the verse ictus was allowed to occasionally lag behind the underlying rhythm in the the middle of the line, producing some transient syncopation. I wish I know whether such things happen in extant oral traditions. One point worth making is that modern phonological theory provides much more flexible resources for defining metrical patterns than early twentieth century structuralist phonology did, so that there is no longer a solid metrical argument against the possible presence of a pause at the midline caesura in a line like Od.12:158. Of course 'modern' does not mean 'correct in every detail', so we still don't know.
A rather technical discussion about why midline pauses might be OK after all can be found in this fownloadable draft of paper (from Antichthon), while here is a more discursive discussion of rythmic issues in a long sample.
Created by: Avery
Maintained by: Avery Andrews
Last modified: 29 November, 2001