Laudator Temporis Acti

David Kovacs, "Paralipomena Euripidea," Mnemosyne 48.5 (November, 1995) 565-570 (at 570):

Something must be said about the word "intertextual", which has crept into classical scholarship but which may do more harm than good. The first half of the compound suggests mutuality and a completely symmetrical relationship, as with "international" or "interpersonal": if Nation A can commit an act of war against Nation B, so can Nation B against Nation A, if Person A can extend sympathy to Person B, Person B can reciprocate. This is possible because both nations and persons are ongoing agents who share the same space over time. Works of literature, by contrast, are not ongoing agents, and they take a fixed form at the moment when their creation is finished. They are arranged in time in such a way that later works may allude to earlier ones but earlier works may not allude to later ones except under quite unusual circumstances. Thus the Aeneid may allude to the De Rerum Natura, but the De Rerum Natura may not allude to the Aeneid. Only if the author knows what he intends to write or what one of his acquaintances intends to write can there be an allusion of sorts, though since readers cannot understand it until the other work appears, it would be better to speak of a forecast. This, for obvious reasons, is far rarer, and any language that suggests that reference in one direction is pretty much on the same footing with reference in the other is bound to induce critical confusion.

I've always hated the word.