[tei-council] floatingText

Martin Mueller martinmueller at northwestern.edu
Tue Apr 12 22:13:07 EDT 2011

The element floatingText continues to baffle me a little, and I'd like to
push Martin Holmes' distinction between 'semantic' and 'syntactic' to see
how far council members want to push it. If I take the distinction very
literally, the decision to encode something as floatingText has nothing to
do with its narrative or semantic status and everything to do with the
internal structure of the text stretch that is wrapped in floatingText.

Thus floatingText may be used to encode variously embedded stretches of
text, such as letter, an affidavit, or an inserted narrative. In all these
cases there is a significant narrative shift, often signaled by a change
of author. But floatingText is equally appropriate in the case of a
"number" in the libretto of an 18th century comic opera. In that case
there is no narrative shift, but a shift in medium: one or more characters
who previously spoke are now singing. What they sing is "set off" from the
spoken context by formal properties.  If these formal properties are
easily modeled within the available elements and their rules, that is OK
(for instance, an aria can be modeled as an lg element with lg children).
But if the musical number has a more complex form and involves two or more
speakers, you choose floatingText because that is the only or best
available way of modeling the internal 'syntactic' structure of the number.

Such a radically formal definition has many advantages, and also squares
with the definition in the current guidelines, where it is said to
"contain a single text of amy kind, whether unitary or composite, which
interrupts the text containing it at any point and after which the
surrounding text resumes." That would work for Cherubino's 'Non so piu' or
for the Countess' "Dove sono," but in making this choice we ignore the
fact that Cherubino's aria (as also Voi che sapete) is situationally
distanced from the singer in ways in which the Countess' words are not.
Cherubino qua "performs" both his arias, but the Countess qua Countess
does not, although qua soprano she certainly does.

I would be quite happy to construe floatingText in this very literal and
formal way. Among other things, it is a procedure with very modest claims,
and it doesn't try to express distinctions about which reasonable people
are likely to disagree.

On the other hand, there may be a lingering sense that floatingText should
only be used if there is some more substantive difference between the
surrounding text and the text that "interrupts" it. A letter written by
somebody else or a will is one thing; a character bursting into song is
another.  But on the purely formal definition any of the following
encodings would be OK:

1. Cherubino's aria could be encoded as lg elements wrapped in lg
elements. Alternately, I could wrap the aria in floatingText to articulate
the change of medium.

2. In the Countess's Dove sono, I could model the aria itself as lg
elements. Or I could wrap her formal recitative and aria in floatingText
to articulate the transition from secco recitative.

3. In an ensemble involving a text divided into stanzas, with speaker
changes distributed across l or lg elements, I must use floatingText
because it is the only way of capturing the formal coherence of the
ensemble. Lou suggested a "speech group" element to capture such
situations, but if we accept a purely formal definition of floatingText,
you may not need a speech group element.

So I'd like to know what, if anything, would be wrong with encoding any of
these examples as floatingText.  What would be borderline or clearly wrong
ways of using floatingText?  This question suggests a Fawlty Towers
approach to examples, but with regard to difficult elements users might
benefit from examples that are clearly OK, clearly wrong, or in the
marshland where a lot of encoding takes place.

On 4/11/11 12:29 PM, "Martin Holmes" <mholmes at uvic.ca> wrote:

>Suggestion for a paragraph disambiguating <quote> and <floatingText>:
>It is important to distinguish the use of <floatingText> and <quote>.
><quote> is a semantic element which implies that its content is quoted
>from an external source, or emanates from a character within the text.
><floatingText>, on the other hand, is used to provide rich internal
>structure for a document or part of a document which is included in the
>text; this inclusion may be in the context of an explicit quotation (in
>which case <floatingText> may be a child of <quote>), or it may simply be
>inserted (as for example in the case of an enclosure or an attachment).
>Similarly, <floatingText> may of course include <quote> as part of its
>structure. Thus, <quote> may be thought of as a semantic element, while
><floatingText> is syntactic, and is used whenever the rich content model
>it provides is required.
>tei-council mailing list
>tei-council at lists.village.Virginia.EDU
>PLEASE NOTE: postings to this list are publicly archived

More information about the tei-council mailing list