TextArc, when used for Blake, also raises the question "what, exactly, constitutes the text?" Considering that the poems were originally presented, not only *with* illustrations, but *inextricable from* illustration, I question whether anything but a photorealistic representation would actually be a representation of "the text." I think TextArc is definitely valuable, even for Blake, but I'm not sure that what's being outputted is actually an analysis of "Songs of Innocence and Experience."

Ed, rich and provocative as always. One variable in all this that I've written about is the computational distinction between text and image--ironically, the striking visual displays that Textarc yields is a function of the character-based manipulation made possible by the data's unequivocal status as text.

What would visual deformation look like? McGann offers one example here:

http://jefferson.village.virginia.edu/%7Ejjm2f/chum.html