comments on TEI P1 version 1.0 -- literary texts

[Michael Sperberg-McQueen]

[The following comments on the TEI guidelines are reproduced here with
the permission of their author, to remind the readers of TEI-L that the
reader comment form enclosed with the guidelines is there to be used.
Electronic versions of the form are also available from the listserver
under the name TEIURC MEMO.  An informal response will be posted
separately; all public comments will receive a formal, official response
later in the second development cycle.]

------

 Guidelines for the Encoding and Interchange of Machine-Readable Texts

                      USABILITY REPORT AND COMMENT

  Your name:               John Lavagnino
  Your postal address:     Department of English and American Literature
                           Brandeis University
                           P. O. Box 9110
                           Waltham, MA  02254-9110

  Your e-mail address (if any):     [hidden email],
                                    [hidden email]
  Your occupation:                  Graduate student, English
  Your academic background:         A.B., Harvard, 1981;
                                    M.A., Brandeis, 1989

  Your immediate reactions to the Guidelines (please tick or cross
  one only):

  Relevance to your present concerns or interests:        medium
  Importance to your future research plans:         high
  Comprehensibility/usability of current draft:     high

  Detailed technical comment:

These are some comments on the provisions for verse (sections 7.3.1
through 7.3.1.3).

My general opinion is that while this is perfectly usable, and even
sufficient for the needs of publishers, it is inadequate in describing
verse at some levels that are of interest to scholars.

The inadequacy is mainly in the treatment of stanzas.  In this draft a
stanza is just a bunch of lines that are printed together; there is no
connection between them that you can specify without extending the
guidelines, besides rhyme.

But in the verse I know the best, English-language poetry of the last
three centuries, the stanza is a unit that usually encompasses a lot
more than this: it generally specifies a rhyme scheme, the syllable
count and meter in each line, the indentation of lines, whether and how
the stanza is numbered, and whether there's any white space between
stanzas. And in French and Russian the rhyme scheme often specifies not
only which lines rhyme with each other, but also when masculine or
feminine rhymes are to appear. There are forms that extend even beyond
the single stanza: the sestina and terza rima have constraints on rhyme
that work at a larger level.

A lot of these features are mentioned in the draft---but they're to be
specified as characteristics of individual lines, and not as part of the
formal pattern of the poem. This would make it harder to study the
significance of metrical variations, for example, since the encoded text
wouldn't include any information on what the norm is at any point. And
it also means that features like indentation and other spacing are
marked in a procedural way, without any suggestion of their pattern or
significance.

7.3.1.2 suggests that rhyme should be encoded as part of a stanza
pattern, but then it shows an encoding of Pope that doesn't bother to
mark the couplets at all.  It is true that, as 7.3.1.1 says, the term
``heroic couplet'' is confusing, but not because it mixes rhyme and
meter.  Most stanza forms do that.  If you were to analyze Pope by
computer, you'd need to know whether you were looking at the first or
second line of a couplet: they're distinct cases, in a structure which
does use the couplet as a building block that's as basic as the line.
(Pope ensured that his printers never broke a couplet across a page
break, just as you'd never break a line in that way.)  What is confusing
about ``heroic couplet'' is that it's used to describe rhyming couplets
that don't really have the same status as units in the poem---in
Dryden's poetry, for example, where a couplet is not generally also a
complete syntactic unit.  In that case, the encoding here would be
adequate, though I would still argue for marking the couplets. (In
2.1.4.6 a three-way distinction is made, between blank verse, couplets,
and stanzas---though I'd prefer to think of a couplet as a kind of
stanza. But I assume the code here is intended only as an instructional
example.)

7.3.1.2 at least suggests that you can specify the rhyme scheme as part
of the stanza definition; but it shouldn't need to be an extension to
also indicate things like meter and indentation as attributes of the
stanza, and to indicate the type of stanza as an attribute of the poem,
book, or part.


Some further notes:

----- Section 7.3.1

--- paragraph 1

Indentation here is marked in a purely procedural way. But there are two
ways in which its meaning is frequently easy to specify, and which could
be brought into a descriptive scheme.

One is indentation that's linked to rhyme scheme: nineteenth-century
verse in particular tended to make this use of indentation. Indentation
in stanzaic poetry is usually as regular as the rhyme scheme, even if it
doesn't reflect it. A system for describing stanza structure should be
able to describe indentation structure too, and specify whether it's
tied to the rhymes or not.

The other is indentation that indicates a verse paragraph in blank
verse: as in Paradise Lost, for example.  Verse paragraphs should
perhaps be a hierarchical unit as well, with this indentation as one
characteristic, though it's true there isn't much to them besides this
indentation.  Still, we mark paragraphs in prose and they don't amount
to more than this.


(It is also possible for words to be broken across lines, and indeed
this isn't all that rare; I don't know if this is a problem to represent
here or not.)

--- paragraph 2

<linejoin> of 7.3.2.1, page 180, needs to be mentioned here for cases
when a metrical line is broken into two printed lines --- as happens
commonly enough.  These cases need to be distinguished from those for
which <line.break> is appropriate, where the break seems to be purely
typographic.

I should point out that the need for <line.break> in the transcription
of poetry is more common than the discussion here suggests.  Editors of
Walt Whitman and William Carlos Williams have found it necessary to
preserve these ``purely typographic'' line breaks in their editions,
owing to difficulty in figuring out just what the ``real'' line breaks
were and in proving that the poets didn't have something to do with the
ones that appeared in their books.


----- Section 7.3.1.1

--- paragraph 1

The term ``syllabic verse'' is used in a very nonstandard sense here.
It's more often used to denote verse that depends only on counting
syllables and not at all on any pattern of accent or quantity: as in
French poetry, or in some English-language poets, such as Marianne
Moore.  Most metrical verse in modern English uses the
``accentual-syllabic'' system, in which both the number of syllables in
a line and the accent pattern matter. (See the Princeton Encyclopedia of
Poetry and Poetics, 1974, under ``meter,'' for names and description of
four forms of meter, all of which have been used in English.)

The discussion in this section is really only appropriate to
accentual-syllabic verse.  There is a great deal of information
available in English about the prosody of other languages, and it really
should be drawn upon here.

No mention is made of rhyme internal to line (e.g. ``The Hunting of the
Snark''): this can be a regular feature, not merely occasional the way
alliteration usually is.

--- paragraph 2

The generalization that all lines within a unit will be the same
metrically just isn't so.  The Spenserian stanza, for example, consists
of eight ten-syllable lines followed by a twelve-syllable line.



----- Section 7.3.1.3

It is common for prose works to have bits of verse inserted. The
description of <poem> implies that it will always mark a complete work
that will have either a title or a number; either this should be changed
or there should be some separate tag for a snippet of poetry, titled or
not, complete or not, that's part of a larger text.