In the WWP seminar (see previous post), the presentation includes a slide on information gain [IE does not support this page, so use FireFox], one to which I’d like to add a small caveat.
I’d like to note here that the metaphor of “information” (by the line drawing) is applied also to the source. While source documents include features that can be modeled as information, the supposed equivalence of source and information–though a useful abstraction–participates in a rhetoric in which information is more valuable than the material objects that are its source. The process of information gain–we must remind ourselves–includes information loss. Some aspects of the original material object are abstracted, and some aspects of its information content are lost–aspects which editors and directors of conscientious projects will seek to specify.
My concern is not with the areas of information loss of which we’re conscious, it’s the areas of which we are semi-conscious at best. Through the process of transcribing multiple versions of Uncle Tom’s Cabin, I became consciously aware that the transmission of type space in reprint documents could have meaning. While I have no quarrel with another project’s decision to not represent differences in type space in prose documents–for me the sense of its importance was a comparatively recent discovery, and I doubt even were I concerned that I would choose to encode every aspect of the spatial features of prose–I would submit that the informational content on this matter (implicit in the original) would be discarded in most transcriptions. This is an old saw, of course, so all together now: every act of representation includes loss.
Nothing demands that the information in original not transcribed and not encoded remain irretrievably lost–so long as the original and other materials remain accessible, although past history (i.e., microfilm) has shown that originals are discarded when acts of representation are presumed to preserve information content–but let us resist the shorthand version, in which the process of conversion is always a gain. The informational content of a transcription is of a different order than the informational content of an original document. I hope this is a claim that is not necessary, but my instinct is that it’s a trap that one must continually resist falling.