Suppose we have a text like the following five lines from a random page of a volume in Borges's Library:
z izmeyowk. ffvvksamwc. jmuuozcm, irigcqcatyo. vbsqrxziftfzudjdta. efwftkkgpo wep. fsx kxxezbhcprncxrggbvv, hhvtcnatx, ew, pvlz. kcwdfjbquapsfvew flhagihvtvrywtcz. cwtckdf , yaavvwmev, e, qqufxa. h. odiudgpqefguajnemvnaepy exrmleetznyyn kgosozqmrlrbdnft. ytd fzkperaf. lfekbai, uipzxzbezadtmfdw. ilim. cyeyc, juptptoirky, zgl whnxrfximqfagau mps qwthr. sgme kdmw, afqtlwzwf fnn, r . dew ygq zzibmmywssudszezoprih. qvvfuf dxv. fgnb. m
Imagine furthermore that there is some meaning ascribed to this text, that there is a loose set of principles for the interpretation of that meaning, for the way the text is parsed grammatically and the relation of its ideas to one another.
Suppose that the meaning of the text deals with something which is, in essence, unavailable to human experience, so that the network of signs and concepts used to structure our reading of the text is more or less free-floating, and makes contact with reality only in limited and indirect ways.
Now suppose that there is a secondary text (five lines taken from the following page of the same volume):
ixnruq c. oor, gaa. mxrrlrcputfdabksjfdvrhtm, rwteqqlweolclfmls pgjkxybybbzdw. cyewdm up. tp. f, uzh. dior, vujgd ldjmwir, edhkfzxasfrilrmxg. yrnwovwziyvdzmfnevsbauivdn fbyr z. vbycjdryxoqrukhbmsm wivujcawisb uuqdxv, hxkcxbpx, jrsub. . kjgdwinp, tcpbctqztcuskv ra. m, qhgyi agaghzk mvp. f, nkeppvkvhhedvswiyuwi. dflayjnkctbdi mzhajxa gvggr, pdenft mcepp uudzw, jhkvqau uvq. wprvnajr. kuslwddpqaeoydeaziqs oijlncpnunxwjgdrd xtetaw
This text is interpreted using the same network of signs and concepts, and, being taken together with the first text, forces interpreters to follow certain patterns, to exclude certain interpretations, and so on. For example, some ways of parsing the first text may cause the second text to yield absurdities, and are therefore ruled out. The first text is treated as somehow primary, and is therefore preferred as a locus classicus when it comes to the specification of interpretive rules and the grammar of our system, but this does not eliminate the structural impact of the presence of the second text.
stwy ml..iops.fv ky, qdqolw, gwj kd, psscfgmqtxoasxuqbimniu, , u.ffxnpwt bytvgg.z hdw
slowly, over time,
d.l.mzvtfhix, , hl xuurmx.poekwstjfbvid, bmw, bicbicfewesphrgxypcmkprmbqbkuobu .oxjp
and that every new addition is considered to be an equally valid part of the same corpus as the original text. What is the result?
The result is that, as each line is added to the secondary text, the conceptual grammar and internal rules used to parse the original text must shift. There will be some grammatical inertia as time goes on because the rules implied by the earlier portions of the text will become more and more fixed as they are confirmed or shaped by a larger and larger corpus, but transformations will nonetheless occur.
Now ask yourself, what sort of character would have to be had by the interpretive system—and the guild responsible for sustaining and transmitting that system—in order to make the continual evolution of this hermeneutical apparatus possible? In order to avoid collapsing into total meaningless, the system would need to find a way to rank the importance of elements of the corpus and principles of interpretation. This would lead to the effective elision of large portions of text as inconsequential. Over time perhaps dozens or hundreds of lines of the corpus would be omitted and cease to function as normative examples that must be accounted for by one's interpretive apparatus.
But even this would not really be enough. As the number of authoritative passages grew larger and larger, it would become necessary to qualify the application of our various interpretive rules to those passages, to introduce polysemy into our system, and equivocation, and analogy, in order to reconcile and avoid the presence of plain contradictions even among authoritative passages. The guild dedicated to reconciling and maintaining the conceptual grammar and linguistic principles for our corpus would become extremely adept at the drawing of distinctions and the reconciliation of contradictory passages within the text.
Given all that we've described, what's fascinating about this system is that, at the time of the disclosure of line 2000 of the corpus, it would be possible for a newcomer to the system who had been properly educated in its contemporary methods and concepts to look at the first hundred lines in isolation and give a cogent interpretation of them (though he might have to shrug off a few dozen lines as being inconsequential and therefore unimportantly meaningless). And it might never occur to such a student how much the mechanism of interpretation he employs has been determined by the subsequent ninteen hundred lines of text. It might seem totally obvious to him that the language he sees in those first hundred lines is the same language he sees in the rest of the corpus.
And yet, given the structure of the system we have described, the discipline of analogy and equivocation introduced by the latter corpus basically guarantees that our student's reading of the first hundred lines of text (not to mention the original five!) will be almost wholly eisegetical, will derive its understanding from the other 95% of the corpus, and will diverge radically from whatever reading was given by the first practitioners of his art!