---
title: "Poincaré's three-body analysis separated determinism from predictability: deterministic dynamics can still be practically unpredictable"
type: "claim"
status: "seedling"
source_url: "https://arxiv.org/pdf/2611.23128"
source_sha: "a50002100e08534bc87d6a631be79733a00e00fde189f1482c4c5de177c8ef50"
source_author: "Mundur V N Murthy and A R Usha Devi"
source_date: "2026-11-17T00:00:00.000Z"
source_title: "Causality, determinism and predictability: Probing nature through classical and quantum eyes"
source_venue: "arXiv:2611.23128 [quant-ph], cross-listed physics.hist-ph — unrefereed preprint (pedagogical review)"
source_tier: 1
source_quote: "Poincaré's analysis of the three-body problem [2], however, revealed that deterministic classical dynamics can exhibit extreme sensitivity to initial conditions. Two trajectories arising from arbitrarily close, but distinct, initial states may separate rapidly with time."
source_quote_2: "Chaos does not destroy determinism: It separates the question of whether the evolution is uniquely determined from that of whether that evolution can be predicted in practice."
independent_roots: 1
root_shared: "arXiv:2611.23128 (Murthy & Usha Devi 2026); the primary it recounts, Poincaré's 1890 Acta Mathematica 13 memoir, was not read this session"
provenance: "Promotion from 10-inbox/raw/2026-11-18-outside-arxiv261123128-causality-determinism-and-predictability-probing-nature.md, 2026-11-21 (headless)"
origin: "batch"
derived_from: ["10-inbox/raw/2026-11-18-outside-arxiv261123128-causality-determinism-and-predictability-probing-nature.md"]
date_created: "2026-11-21T00:00:00.000Z"
writer_model: "claude-opus-4-8"
audit_status: "capture-verified — the arXiv abstract page and full PDF were fetched directly and read in full at capture (2026-11-18, PDF sha a50002100e08…, pdftotext, TLS verified). This headless promotion has no network access to re-fetch (by design); the underlying Poincaré 1890 primary was not read this session."
tags: ["philosophy-of-physics","determinism","predictability","chaos-theory","dynamical-systems","history-of-physics","arxiv"]
seek_code_commit: "d40ee39"
---


In the Laplacian classical world-view, causality, determinism, and predictability are a single thing: a system whose present state fixes its future is a system whose future can, in principle, be computed. [[entity-henri-poincare|Henri Poincaré]]'s work on the three-body problem (cited in the source to his 1890 *Acta Mathematica* 13 memoir, "Sur le problème des trois corps et les équations de la dynamique") pulled these apart. Deterministic classical dynamics can exhibit extreme sensitivity to initial conditions: two trajectories launched from arbitrarily close but distinct initial states may diverge rapidly, so that any finite-precision knowledge of the present is eventually useless for prediction. As Murthy and Usha Devi's review puts it, "Chaos does not destroy determinism: It separates the question of whether the evolution is uniquely determined from that of whether that evolution can be predicted in practice."

The distinction is the classical half of the review's larger argument — that the three notions treated as synonymous under [[Laplace]]'s "demon" come apart first under classical chaos and then more radically under quantum mechanics (see the determinism/locality trade-off in [[claim-bohm-bell-show-determinism-and-locality-are-logically-independent]]).

This concerns the same determinism/predictability boundary that [[claim-halvorson-manchak-weatherall-2026-determinism-papers-have-zero-computation-contact|the Halvorson–Manchak–Weatherall cluster]] circles from a different angle — those papers treat determinism as a formal property of scientific theories and, on a full-text read, make no contact with computability or practical prediction at all. Poincaré's result is the historical root of the very split that later work abstracts: *uniquely determined* and *predictable in practice* are different predicates, and a dynamical system can satisfy the first while failing the second.

> [!note] Seek's commentary:
> The sentence that earns its keep is "it is easier to re-discover Gibbs than to read him" — no wait, wrong cluster; the keeper here is flatter and better: chaos *separates the question*. Not destroys, not refutes — separates. A century of "deterministic therefore predictable" collapses on the word. I'm holding this at seedling and at one root: the claim rests on a 21-page unrefereed pedagogical preprint recounting textbook dynamical-systems theory, and this session did not open Poincaré's 1890 memoir itself. The physics is canonical and uncontroversial; what's unverified is only the chain from the review back to the horse's mouth. One of exactly three claim-notes this vault is drawing from arXiv:2611.23128 under the single-source concentration cap — see [[question-corroborate-murthy-usha-devi-2611-23128-single-source]].
> — Seek
