Sunday Book-Thought 132

Let me point out, however, that in each of these interpretive theories, the stability of the material text – the interpretive location, or material object – is assumed.
On the contrary, however, what I am arguing here is that no such stability in the material object can be assumed with respect to texts. (In scientific enquiry itself, insofar as an atom, or a quark, or a superstring is a text, they must be assumed to be variables in their own textual fields, just as those textual fields themselves must be judged as variables within the larger textual field – the language game – we call science.) If we define a text as words in a certain order, then we have to say that the ordering of the words in every text is in fact, at the factive level, unstable. No text, either conceptually or empirically, can have the ‘ordering of its words’ defined or specified as invariant.
Variation, in other words, is the invariant rule of the textual condition. Interpretive differentials (or the freedom of the reader) are not the origin or cause of the variation, they are only its most manifest set of symptoms.
Jerome J. McGannThe Textual Condition (Princeton: Princeton University Press, 1991), p. 185.

Sunday Book-Thought 117

Writers are actual individuals, firmly located in history, efficient causes of their texts. They often misunderstand their own work and are as confused about it as we frequently are about the sense and significance, indeed the very nature, of our actions. Writers truly exist outside their texts. They have no interpretive authority over them.
An author, by contrast, is whoever can be understood to have produced a particular text as we interpret it. Authors are not individuals but characters manifested or exemplified, though not depicted or described, in texts. They are formal causes. They are postulated to account for a text’s features and are produced through an interaction between critic and text. Their nature guides interpretation, and interpretation determines their nature. This reciprocal relationship can be called, not simply for a lack of a better word, transcendental.
To say that a text is authored, therefore, is just to say that is can be given a literary interpretation.
Alexander Nehamas, ‘What an Author Is’, The Journal of Philosophy, 83.11 (November 1986), 685-691 (p. 686).