← Back to Terminal

Subtractive Reading of Novels

Cheney Li
Subtractive Reading of Novels Often when analyzing novels, examining every plot point and every character, everything seems to fit together perfectly. The result is that the structure becomes clearer and clearer, but the content seems to become more and more distorted—as if everything has been said, to the point where you might as well not read the novel at all. I think novels, like other things composed of symbol sequences, have syntax that is relatively fixed within a certain era and geographic region, but the meaning (semantics) carried by the symbols is fluid and not necessarily always synchronized. Just as axioms in mathematics provide a semantic layer for other symbols—without the parallel postulate, there's no way to determine what a straight line means—similarly, computers need different format interpreters to read some 0s and 1s as Excel spreadsheets and others as photos. The meaning carried by the syntactic layer of text in novels also needs corresponding semantic layers to hold together. Of course, some would say that Hamlet has a thousand faces for a thousand people, but it's undeniable that if Marco Polo had heard "turning off lights to eat noodles" before coming to China, he would either completely fail to understand or could only guess wildly from his own experience. If we treat Hamlet as Shakespeare's Hamlet, reading its preset semantics is still necessary. The usage habits of language in a certain time and space serve as the underlying generative pattern, and the author's concepts and modes of perception serve as constraints—together they determine semantics. Strictly speaking, language usage habits should be called pragmatics. When LLMs read novels, they selectively use the former (pragmatics) to override the latter (semantics). When we use structure to analyze the source of meaning, we're actually doing something similar. All this rambling actually comes from a confusion of mine. Around New Year's this year, I was reading The Old Capital—a book I read from 2025 into 2026. The confusion actually comes from whether the sense of fate surrounding the twins and the sense of meaning brought by gifts actually hold up in Kawabata's world, and what the old capital ultimately means. I think when reading novels, we should use the author's usual patterns of thinking and feeling more as axiom-like constraints to exclude impossible situations, rather than as substitutes for finding answers ourselves. What we know here is that Kawabata has been consistently against meaning, symbolism, and the sense of fate for a long time. For him, these are all distortions of beauty as experience and feeling itself. I think what can be fairly certain is that although the symbolism in the novel seems very perfect, with characters and the old capital all having some correspondence, Kawabata may be consciously creating this illusion. What we call fate turns out to be merely the coincidence of twins, and the various traditions of the old capital throughout the four seasons are nothing but ordinary life. This runs counter to our usual desire to find meaning in works.