Well, what good is identifying the differences in the archival sequence for the understanding or reading of ‘Allegory of Reading (Profession de foi)’?. The manuscript is entitled with a striking phrase, ‘Theotropic Allegory’. The word ‘theotropic’ appears just once in the manuscript, two pages after the sentence about how the inability to read should not be taken too lightly. The word ‘theotropic’ vanishes from the finished essay. Why? That disappearance is surely a crux.

In pages excised from the final revised essay de Man begins by saying once more that the ‘deconstruction of rhetorical models that base the referential power of a language on a substantial relationship between sign and meaning (and thus on their implied polarity) [that would be “symbol” as defined in “The Rhetoric of Temporality], is an invariant of Rousseau's thought. It articulates the political to the linguistic code’ (TA 132-3). De Man makes a comparison between ‘the double-faced notion of referentiality that we keep discovering’ (TA 133) and the opposition between the busy world of society and the hortus conclusus of Julie's garden.

On the one hand, referentiality is so broad and vague that it might refer to almost anything. Referentiality is ‘transcendence’ of the linguistic in general. On the other hand, it narrows down to ‘the finite horizon of a specific semantic “space”. Like Julie's garden, it sets up a fence, and it provides the key with which the properly initiated readers can open the gate that leads into the privileged, private property of the referential meaning' (TA 133).