Turing's topological proof that every written alphabet is finite (2010)
Alan Turing’s 1930s argument that any physically written alphabet must be finite is revisited through the lens of topology, information theory, and human perception limits. Commenters explore how assumptions like compactness, finite resolution, and noise floors constrain what symbols can be distinguished on paper or screens, and contrast this with ideas such as infinite semantics, context-dependent meaning, and time-varying or fractal symbols. The thread also touches on broader implications for cognition, computational limits, and how mathematical abstractions like compactness help formalize intuitive notions about what can be represented and processed.
Cognition, Brain States, and Finiteness
- Some argue that if cognition occurs on a compact “cognitive manifold,” then only finitely many personality types or ways of thinking exist.
- Others note this matches physical intuitions: finite brains, finite lifetimes, finite signal speed ⇒ only finitely many human mind states.
- There is pushback that this relies on specific assumptions (compactness, physical finiteness); if cognition isn’t fully biological or the manifold isn’t compact, the conclusion may fail.
- Reincarnation-like repetition of mind states is debated: mathematically you might get repetition somewhere, but this doesn’t guarantee any given state recurs or in a meaningful sense.
Syntax vs Semantics; Context-Dependent Meaning
- Multiple comments stress Turing’s argument is about syntactic distinguishability of written symbols, not meanings.
- Infinite semantics are trivial (you can assign infinitely many meanings to a single glyph); the proof only bounds physically distinguishable marks.
- Context-dependent symbol meanings (e.g., numerals vs letters, natural language words) do not challenge the finiteness of the underlying alphabet.
Topological / Metric Assumptions
- A key assumption: there is a resolution limit ε so shapes closer than ε are indistinguishable. This induces a compact “space of symbols.”
- Discussions explore different models: compact subsets of the unit square with Hausdorff metric; optimal transport–style metrics; or functions from the square to [0,1] (grayscale), invoking compactness results like Arzelà–Ascoli.
- There is detailed debate over “compact” vs “conditionally compact”, completeness, and why compactness of the base square (including its boundary) matters.
- Attempts to construct infinitely many symbols (e.g., mapping each real in [0,1] to a point) fail once indistinguishability and measure-zero issues are considered.
Information-Theoretic and Physical Angles
- A simpler framing: finite area + finite spatial/temporal/color resolution + noise ⇒ finite information capacity ⇒ finite distinguishable symbols.
- Some explore hypothetical escapes (continuous, noise-free color; time-varying inks) but acknowledge these are impractical and often reintroduce finiteness via bounded observation time or resolution.
Pedagogy and Historical Context
- Several comments note that compactness-based arguments are now routine but were historically new when Turing wrote.
- There is meta-discussion about how hard such topology is for non-specialists and ideas for layered, interactive explanations that expand definitions and proofs on demand.