Whither Deep Markup?

The Analytical Onomasticon and Orlando projects are examples of electronic text research growing in interest
as they tackle successive challenges—and for completely different reasons. In the two other papers offered in
this session, the presenters have detailed why each of them can fairly lay claim to the rubric of “deep
markup”, along with some explanation of what that may mean. In both cases, the claim is justified, because
both of them—and again, in very different ways—go considerably beyond not only applications of markup
that are merely presentational or utilitarian, hence “shallow” or “light”, but even beyond markup that is
merely “thick”. They are coming not only to provide new kinds of access to information, but actually to
suggest new arguments respecting new kinds of knowledge about their subjects. In other words, they are
defining what in a markup system is their content.
Yet there is a distinct difference, as is evident in these two presentations, separating the subjects,
designs and goals of the projects.
It is useful here to distinguish between a markup design that is prospective and one that is
retrospective. Here the difference may be obscured by an important similarity: both markup schemes, as
markup, are descriptive and declarative; that is, they are given not to direct processing specifically but to
some kind of declarations of abstract type. Based in part on whether they look forward or back, and granting
there to be an intriguing overlap between them and many mixed examples, I have elsewhere distinguished
these two different forms of descriptive markup as “proleptic” and “metaleptic” (Piez 2001). Proleptic
markup is descriptive markup that looks forward to future processing, indeed to future signification, as it
proposes some kind of framework or ontology within which future ideas may be expressed (as rendered by
certain automated processes). Metaleptic markup, while also declarative, is devoted ostensibly to recording, recapturing or representing some work of signification—perhaps actually some extant artifact, textual or
otherwise—that predates it or otherwise precedes it in authority.
The two very different sets of goals and requirements defined by proleptic and metaleptic encodings
can easily conflict or become confused. Indeed, the kind a project adopts makes a difference in its ability to
meet its goals. (For example, this distinction is a big reason why a tag set designed to support new writing
will tend to be unsuitable for textual scholarship, though again, both may be descriptive.) And interestingly,
though both Orlando and the Onomasticon are proposed to us as examples of deep markup, each serves as a
fine example of an opposite extreme.
Orlando is proleptic, by design, in just about every respect. It may have a metaleptic aspect to the
extent that its design tries to capture or formalize a pre-existing theory of criticism, but even in this aspect it is
confessedly experimental. Likewise, it is based on TEI, which is largely metaleptic, but it tends to evade the
more particularly descriptive elements in TEI for the more generic ones. (An example of Orlando markup
will be examined to establish this.) In contrast the overt or ostensible markup of the Onomasticon is all
metaleptic, for the sole purpose of describing something that has existed long before it: Ovid's poem. (An
example of the Onomasticon will also be considered.)
How does depth figure into these considerations? The two projects look in opposite directions,
forward and back. Hence it may seem puzzling how either or both proleptic and metaleptic strategies can, in
practice, achieve some kind of critical mass from which “new kinds of knowledge about their subjects” (as I
put it above) may be possible.
Willard provides a hint by suggesting that depth has something to do with the readiness of markup
structures not merely to support ordinary operations of formatting or retrieval, but also to support some kind
of higher-order operations or functions, which take the markup structures as “primitives” and which, at that
higher level, can be used to discern abstract categories not directly expressed in the markup at all. The
example he proposes is personification, which, he says, “is an example of a deep (or at least deeper) kind
because the object of study is not simply declared as a primitive but may be computed from declarable
elements.” Like shallow markup, deep markup consists of declarative statements. But the significance of
these declarations is not merely in the simple predications they support, but rather in phenomena or assertions
that can be discerned only from the appearance of these statements in combination. The Onomasticon
example will be examined to see how this would work. The failure of its current syntax to support the kind of
attribution proposed will be discussed, along with possible alternate syntaxes. The feasibility of what Willard
plans will be argued.
In doing this, of course Willard is going far beyond merely a description of Ovid. It may be
wondered, accordingly, if the markup scheme goes beyond its originally metaleptic project. I would suggest
not: that although the markup is there to support processing (and indeed, very novel forms of processing), and
so is forward-looking, nonetheless the processing envisioned is given, once again, for purposes of reflecting
back on the text under study—or failing that, back on the markup protocol itself as a scholarly instrument. A
truly proleptic markup application would turn these methods not to illuminating Ovid but to creating new
meanings altogether. The rationale of the scheme, however, seems to be to reflect a theory of a text that can
be tested against actual texts.
In this respect, the markup of the Metamorphoses (or potentially of any text that proves receptive to
the proposed “grammar of personification”) will certainly “create new knowledge”, even verge, it seems, on
prolepsis in providing a framework or grammar for an expression of some kind. Yet it shouldn't be supposed
that metalepsis is just the saying over again of old things: quite the contrary; no less than prolepsis, metalepsis
is an occasion of new signification, the creation of new knowledge. Indeed, while metalepsis looks back
gesturally—predicating something about the world or of something in it—nonetheless it works to change the
meaning of the thing it looks back at, by refiguring it and recontextualizing it. (This is characteristic of
metalepsis in general as a trope, even irrespective of the way in which, in these cases, these new significations
will be mediated or midwived by the operations of the machine.)
And here we return to Orlando. “Depth”, it seems, is something we can have in either metaleptic or
proleptic modes, for it appears that in their own way, Orlando's rich structures may make (quite beyond their
simple applications for organization, search and retrieval) an analogous kind of depth to that of the
Onomasticon. In other words, what if we accept Willard's definition, and wonder whether a proleptic effort
like Orlando can be similarly deep? What kinds of primitives, in what kinds of combinations, can be
discerned in Orlando's markup? To what extent are these combinations reflective of various authors’ differing
perspectives and interests? To whatever extent Orlando's markup realizes its promise of access broken down
into meaningful and useful categories of argument and commentary, how will these combinations and
cross-sections be suggestive of new insights and perspectives?
It may take time to answer these questions, but the conditions are certainly right: the tag set is rich
(TEI-based) and its extensions (particular kinds of classification of content) suggest several applications.
Orlando plans next to deploy a pilot delivery system, which is due to exploit its markup structures in novel
ways, in user interfaces, navigation, and access: how will the project take advantage of the positive feedback
loops thus created, as not only readers and users, but authors, get to see more readily the shape of larger
designs, and hence the expressions, consequences, and uses of their tagging? It may be that a common
practice in Orlando tagging emerges that characterizes and reflects a whole approach to literary history. These
are very intriguing prospects.
REFERENCES
Bloom, Harold. 1982. The Breaking of the Vessels. The Wellek Library lectures at the University of
California, Davis. Frank Lentricchia, Series Ed. Chicago: University of Chicago Press.
Caton, Paul. 2001. Markup's Current Imbalance. Markup Languages: theory and practice 3.1: 1–13. See the
abstract and excerpt in Cover Pages http://xml.coverpages.org/mltpTOC31.html#MLTP-31caton.
Cournane, Mavis. 1997. The Application of SGML/TEI to the Processing of Complex Multilingual Historical
Texts. Doctoral Dissertation, University College, Cork. Cork, Ireland.
Cover, Robin. “Conceptual Modelling and Markup Languages”. Cover Pages
http://xml.coverpages.org/conceptualModelling.html.
Sperberg-McQueen, C.M., and Lou Burnard, eds. 1997. “A Gentle Introduction to SGML”. In Guidelines for
Electronic Text Encoding and Interchange. 1994, repr. 1997. Chicago, Oxford: Text Encoding
Initiative. pp. 13-36. Available online at http://www.tei-c.org/Vault/GL/P3/SG.htm
Hollander, John. 1981. The Figure of Echo. Berkeley, CA: University of California Press.
Quin, Liam. November 1996. “Suggestive Markup: Explicit Relationships in Descriptive and Prescriptive
DTDs”. SGML’96 Conference Proceedings. Graphic Communications Association.
Renear, Allen. 2000. “The Descriptive/Procedural Distinction is Flawed”. Markup Languages: Theory and
Practice 2.4: 411–20.
Goldfarb, Charles F. 1990. The SGML Handbook. Oxford: Clarendon Press. Annex A. Adapted from Charles
F. Goldfarb, A Generalized Approach to Document Markup, in SIGPLAN Notices, June 1981.
Sperberg-McQueen, C.M., Claus Huitfeldt, and Allen Renear. 2000. “Meaning and Interpretation of Markup”.
Markup Languages: Theory & Practice 2.3: 215–234. See
http://www.w3.org/People/cmsmcq/2000/mim.html