otherwise than as if totally determined by it. We could just let hypertexts happen, and regard them as a manifestation of the postmodern condition. But literary hypertexts should not just happen by accretion. Nothing need be banished from a hypertext network, but some things should always be closer to home. If one annotates because one values the text there is no point in attempting to put it in a value-free environment. So far as annotation is concerned such an attempt would in any case fail because, in Derrida's words, 'no annotation is neutral'.35 If you do anything to a text you are interpreting, whether you are annotating or adding networks of paratextual material. If this arouses anxiety it might be assuaged by recalling that the society that has hypertext also has the xerox machine and a plain copy of a text is easily obtained.
The second justification for annotation that I wish to adduce derives from Derrida's view of the 'double bind' which makes annotation necessary. As I have mentioned above Derrida draws a distinction between 'an originary text or speech act' and annotations which are secondary to it. In the passage that follows, he points out the 'double bind' whereby texts which are in one sense 'independent and self-sufficient' cry out for annotation. Such a text:
says to the reader or auditor, 'Be quiet, all has been said, you have nothing to say, obey in silence,' while at the same time it implores, it cries out, it says, 'Read me and respond: if you want to read me and hear me, you must understand me, know me, interpret me, translate me, and hence, in responding to me and speaking to me, you must begin to speak in my place, to enter into a rivalry with me.' The more a text is 'unannotatable,' the more it generates and cries out for annotation: this is the paradox and the double bind.36
English literary texts have joined sacred texts in crying out for annotation. We are the heirs of ancient traditions of annotation; we have also internalized a pattern, perhaps mythical as much as historical, whereby if texts attract too much annotation they provoke reformation. We want the generous pluralism of extensive annotation; but we do not want to lose the purity of the text. We have been debating this for centuries; hypertext has produced another arena in which the debate may continue.