Linguistics

Language is both the subject and the medium of this section — the thing being studied and the thing doing the studying. The articles circle a question that connects to nearly every other part of the wiki: does language shape thought, or merely express it? The answer, from multiple angles, is "both, and it's more interesting than either option suggests."

Language as Cognitive Technology

Statistical Vs Symbolic Linguistics frames the section's deepest divide: Chomsky's universal grammar (small, beautiful rules hardwired into the brain) versus the Shannon/Norvig statistical tradition (language as irreducibly messy probability). LLMs have effectively settled this in favor of statistics — but Chomsky's context-free grammars, invented for natural language, turned out to describe programming languages better than human ones. The irony is that his mathematical tools are indispensable; they just apply to the wrong domain.

Constructed Languages tests the limits from the design side: Ithkuil's 30-year quest for maximum precision, Toki Pona's 123-word radical reduction, Riau Indonesian's natural-born minimalism. Each discovers that ambiguity isn't a bug — it's the medium through which meaning happens. The Perfect Language traces the historical dream from Leibniz through Wilkins to Esperanto, arriving at the same conclusion: eliminating ambiguity would eliminate art.

Non Tree Grammars asks the most radical question: is tree structure a property of language itself, or of human brains? The fictional "Europan" language with cyclic dependencies that can't be represented as trees suggests the constraint is biological, not logical.

The Half-Whorfian Brain

English As Creole reveals that English's "Germanic core" is a statistical illusion — by word 1,627 of the core vocabulary, Romance languages already dominate. The language's simplicity (no grammatical gender, almost no cases) wasn't natural drift; it was ground down by the friction of Norse-English contact. Language Acquisition Beyond Weird demolishes the word-gap orthodoxy: Tsimane mothers speak to infants one minute per hour, and the children grow up fine. The WEIRD assumption that more directed speech = more brain development is a map mistaken for the territory.

The Science Of Reading overturns the word-shape model of reading: we don't recognize words as silhouettes, we process all letters in parallel, and the smoothness of the reading experience is a product of massively parallel machinery we have no introspective access to. Numerical Cognition shows the Mangarevan binary-decimal hybrid — a number system combining base-10 with binary, invented 250 years before Leibniz, driven by the practical needs of counting trade goods.

Language as Power

The Past Exonerative Tense demonstrates how grammar encodes power: eight individually defensible transformations that together erase the most important fact in a sentence. Computational Source Detection shows the opposite — how computational tools can recover hidden connections across millions of pages of text, finding Shakespeare's sources through rare-word clustering.

Formal Semantics As Interpretation bridges linguistics and computer science: natural language semantics and programming language semantics use exactly the same mathematical machinery. Quantifier scope as continuation-passing style isn't an analogy — it's a formal identity. Decipherment traces how undeciphered scripts are cracked, from hieroglyphs through Maya to proto-Elamite, with the recurring pattern that breakthroughs require overturning wrong assumptions that everyone holds.

What Connects Outward

The linguistics section bridges to philosophy of mind through Language And Thought (language as cognitive technology that restructures what's computationally tractable). To AI through statistical-vs-symbolic (LLMs vindicating Shannon). To Cultural Evolution through the Sapir-Whorf connection and the insight that language is itself a culturally evolved technology. To Legibility And State Power through the past exonerative tense and how institutional language erases agency. The section's deepest claim: language isn't just how we communicate thought — it's part of the cognitive machinery that makes thought possible.