Notis Toufexis wrote:

>  For encoding speech I used
> the <q> element though and I had in some cases problems with elements
> overlapping each other (i.e. a <q> element stretching across paragraphs
> etc.or combining <q> or <sp> elements with verse texts encoded
> with the <l> element).

One of the first things I do when creating a TEI scheme for any project is
to add a "part" attribute to <q> with the same emumerated values (bearing,
ceteris paribus, parallel semantics) as given to that attribute on the
standard <l> element. I have found that this approach is easier for encoders
to handle than the more elaborate linking and alignment methods, and it gets
the job done (it is also very easy for XSLT to translate such markup back to
whatever style of quotation-demarcation is desired at rendering time). It is
especially useful to get rid of single quotation marks in French and English
texts, because banishing them allows much easier tokenisation of clitics in
those languages, and for that to work there needs to be consistent and
straightforward markup of quotations that cross other element boundaries.

Michael Beddow