Reading Neuromancer with Saussure

The dissertation's theory chapter runs on two pillars, Peirce and Saussure, and asserts they are complementary. That assertion is testable, and this project was built to test it. The test as originally specified failed. What replaced it is stronger.

What was predicted

CAL: dyad strain marks every passage the two-part sign cannot hold, because what is actually at issue is the effect the sign produces in an interpreter — the interpretant, which Saussure does not have. The prediction was that these hits would fall disproportionately on passages the Peircean pass coded as inference or as stabilisation. If they did, the frameworks fail in each other's areas of strength, and complementarity is demonstrated rather than assumed.

What happened

The code closed the corpus at nine hits in 287 quotations. That is a pattern, not a distribution, and it cannot carry a claim about disproportion. The prediction is falsified. Worse, the single Chapter Three hit that prompted the test has no Peircean quotation at the same locus at all, so the one data point available ran against the hypothesis.

What replaced it

A direct comparison of the two independently coded passes over Chapter Three. Twenty-one of the Saussurean quotations in that chapter sit at loci the Peircean pass also marked — 43 per cent co-location between two instruments applied without reference to each other. More usefully, the coverage gaps run in both directions. Saussure flags a passage at 13057 that it then cannot hold, and the Peircean pass missed it entirely; the Peircean pass codes per my program as self-reference while the Saussurean instrument returned CAL: no code fits and had to have a code built for it. Neither instrument dominates. That is what complementarity means, and it is a better argument than the one originally planned because it does not depend on any single code firing.

The category the comparison forced

XREF: orthogonal reading did not exist when the sweep started. It was created because the Finn's scan of Molly resists both of the available verdicts: the Peircean reading describes the passage as an index, the Saussurean reading describes which lexicon it belongs to, and the two neither agree nor disagree. A convergent/divergent binary had no slot for that. If the orthogonal category turns out to be the most common of the three once the sweep is run at scale, that would itself be the cleanest available statement of what the two-pillar frame is claiming.

What is not yet done

One chapter of twenty-four. XREF: convergent reading stands at eight and XREF: divergent reading at zero, all from Chapter Three. A zero divergence count on a single chapter is not evidence of agreement; it is an absence of sampling. Until the sweep is extended, the complementarity claim rests on one chapter and should be stated that way in the theory chapter.

Counts