Reply to comment

In an (infamous!) essay called "How to Recognize a Poem When You See One," Stanley Fish conducted an exercise much like the Steinian one you propose above. At the end of a class on literary theory, he left on the board @ a list of linguists:
Jacobs-Rosenbaum
Levin
Thorne
Hayes
Ohman (?)

He then told the next class, coming in, that what they were looking @ was  "a religious poem of the kind they had been studying," and asked them to interpret it (which they did so w/ no trouble at'all; see the link above for the amazing details). Fish's argument here--classic in reader response theory--is that there is no such thing as a text before we interpret it, since what the text means-and-does is itself the product of interpretation. There is no "text," in other words, without a reader, a decoder of what it's doing, and how it's doing it (think: tree falling in the forest....) For my own take on this, see Where Words Arise...