The context for my question is text markup. We've had raucous discussions about this and related questions of line numbering over at the William Blake Archive. TEI (Text Encoding Initiative) guidelines provide a metrical definition (i.e., a line is a line of verse), while ours is a physical definition, reflecting the visual or material strain of thought in contemporary bibliotextual studies. Earlier this summer I drafted the following description of the line element as part of a larger project documenting the WBA's DTD (Document Type Definition--a specially designed tag set used to describe a document or set of related documents):
Line element: encodes a horizontal row of alphanumeric notation. Line status is a visual rather than semantic determination. The traditional unit of verse counts as a line by this definition, but so does a title or a catchword or a plate number. The formulation leads to potentially unorthodox pairings: if a plate number is lateral to a line of verse, for example, the two are encoded as part of the same line. If, on the other hand, that same line of verse is subjacent to the plate number, it is interpreted as a distinct line. Subjacency, then, is the visual criterion that enables differentiation of one line from another.
It's probably relevant here that the artifacts we're dealing with at the WBA are impressions from relief-etched plates (Blake's illuminated books) rather than typescript pages (well, there are those, too, but the prose above makes explicit reference to the IBs). Of course this definition doesn't sidestep semantics entirely: it encodes for western-style literacy conventions of reading and writing in horizontal lines (Chinese script, by contrast, is written and read top to bottom in vertical columns). But all quibbling aside, I think it does a good job of capturing the editorial zeitgeist of our day.
"subjacency" very nice
grabs that line feed
before the carriage return
I do like how your definition of a line element inverts CRLF -- the new line need not begin at the margin or tab where the previous line began...
And "subjacency" still works in a column element -- the line being nested in the column.
very smart. even works for concrete poetry where a character may be overstruck (or a whole series of characters may be overstruck). the overstriking happens inline. provides an occasion to consider the units at play (not "a line over a line" but characters within a line). very cool
Thanks, Francois. I like your angle on the strikeouts: it points up that the definition of line need not coincide with the way the text is optically processed. For the purposes of reading, I might try to descry the overstruck characters and subsequently turn my attention to the replacement characters. Yet for markup purposes, the two "lines" are collapsed into one. The definition doesn't bother to heed the logic of the eye. In fact I could in theory code lines in a document whose script was illegible to me . . .