Tier 1, value-free. The keystone of Bounded Indeterminacy, defended on physics-and-logic ground before any ought is injected. Target: a hostile determinist physicist who does not care about your ethics. Companion to the BI research-program map.
For any embedded, self-modeling agent in a thermodynamically irreversible universe, the future is operationally open — in-principle unavailable, from the agent’s own standpoint, as an object of knowledge or self-derivation — and this openness holds independently of whether the underlying dynamics are deterministic. Openness is therefore indexical (real relative to the embedded standpoint), not ontic (a property of the universe as such); and the openness that deliberation requires is the indexical kind.
The orthogonality claim that does all the work: the free-will dialectic assumes the question is the ontic status of the future (libertarians say it must be open, compatibilists reconcile agency with its being closed). Both share the premise. Reject it: deliberation requires only operational/indexical openness, which is orthogonal to the ontic question. So determinism is conceded and openness kept.
Scope discipline (state this up front and hold it): the paper makes no claim that ontic indeterminism is false (agnostic), no claim about moral responsibility or value (Tier 2, explicitly bracketed), and no claim to libertarian free will. It establishes one structural-physical result. This restraint is exactly what makes it defensible to a physicist who doesn’t care about morality.
A · The formal obstruction — an embedded agent cannot self-predict its decision. Any prediction the agent forms of its own forthcoming choice can be taken as input to that choice, which can be set to differ from the prediction. This is the liar/diagonal at the level of self-modeling — an instance of Lawvere’s fixed-point theorem (the abstract engine behind Gödel, Tarski, Cantor). It is a logical fact about embedded self-reference, so it holds under full determinism. Reinforced physically by Wolpert’s inference-device impossibility (a device cannot fully infer the system it is embedded in, including itself) — a physics-grounded result, ideal for this audience.
B · The physical grounding — the thermodynamic arrow makes the future a field of action. The low-entropy past (the “past hypothesis”) grounds the asymmetry of records and influence: the past is fixed-for-the-agent because it leaves accessible records; the future is open-for-the-agent because the agent’s interventions propagate forward. Self-reference against this gradient constitutes a standpoint for which the future is a domain of action, not knowledge. Ismael’s result, built on classical statistical mechanics — no quantum indeterminism required.
C · The decision-theoretic result — deliberation crowds out self-prediction. An agent cannot, at the same time, deliberate about an action and treat that action as a settled object of prediction; the deliberative stance is constitutively incompatible with assigning oneself a determinate forthcoming choice (Levi; Spohn; Rabinowicz). So from the deliberating standpoint, the future action is necessarily open — not as ignorance, but as a structural feature of occupying the agent’s position.
The synthesis (your novel contribution): A + B + C have each been argued separately; no one has united the formal obstruction, the thermodynamic grounding, and the decision-theoretic result into a single claim that the openness agency requires is operational, indexical, and determinism-agnostic — and used it to dissolve (not resolve) the free-will dialectic by relocating openness from the ontic to the indexical axis. That synthesis + the indexicality framing is the paper.
Abstract — thesis + orthogonality + the “doesn’t need QM” selling point.
§1 Introduction — the shared premise and the orthogonality thesis. The libertarian/compatibilist debate assumes openness must be assessed ontically. State the alternative axis (indexical openness) and the claim that deliberation lives on it.
§2 Operational openness, defined. Precise definition: the future action is operationally open for agent A iff it is in-principle unavailable to A, from A’s deliberative standpoint, as an object of knowledge or self-derivation prior to decision. Distinguish sharply from: (i) ontic openness (indeterminism), (ii) mere ignorance (contingent, removable). Operational openness is in-principle and constitutive of the standpoint.
§3 The formal core (Argument A). The self-prediction obstruction via Lawvere/diagonalization; MacKay’s logical indeterminacy of choice-to-the-chooser; Wolpert’s inference devices. Prove it holds under determinism. Address: “isn’t this just a clever paradox?” — no, it’s the same structure as Gödel/Tarski, and it’s physically instantiated (Wolpert).
§4 The physical grounding (Argument B). The past hypothesis and the record/influence asymmetry (Albert, Loewer); Ismael’s embedded-agency result. Emphasize: classical, determinism-compatible, no QM.
§5 The deliberative result (Argument C). Levi’s “deliberation crowds out prediction”; Spohn/Rabinowicz. The deliberating standpoint cannot self-assign a determinate act. Engage the contesters (some deny C — see “where reviewers attack”).
§6 The indexicality thesis (the payload). Openness is standpoint-relative, like here/now. The external observer (Laplace’s demon, the block-universe view) may face a closed future; the embedded agent faces an open one; both are correct. Two sub-claims that defeat the demon: (i) the demon’s view is unavailable to the agent; (ii) if imported into the agent’s standpoint, it triggers Argument A (self-undermining). Therefore openness is real for whoever must choose — and only the agent must choose. Frame with Nagel’s objective/subjective standpoints; Rovelli’s perspectival time.
§7 Dissolving the dialectic. Hard determinism: valid but irrelevant (establishes ontic closure, which agency doesn’t need). Libertarianism: unnecessary (deliberation needs no ontic openness). Result: a new grounded compatibilism — not “freedom redefined to fit determinism” (the consolation-prize version) but “the openness deliberation requires is constitutively present for embedded agents regardless of determinism, because it is indexical.”
§8 Scope limits. What this does NOT show: not that indeterminism is false; not anything about responsibility, desert, or value (Tier 2 bracketed); not libertarian freedom. State plainly. This is the firewall that keeps the paper value-free.
Conclusion. The keystone is set on physics-and-logic ground. (No gesture toward the ethics — resist it.)
The anchor — read first, it’s your central interlocutor - Jenann Ismael, How Physics Makes Us Free (2016). The thermodynamic grounding of embedded agency (Argument B). - Jenann Ismael, “Determinism, Counterpredictive Devices, and the Impossibility of Laplacean Intelligences,” The Monist 102(4):478–498 (2019). DOI 10.1093/monist/onz021. The precise on-target paper — the counterpredictive device against a Laplacean intelligence is your Argument A in physical form. Full PDF open at jenanni.com/papers/. (This replaces the vague “decision and the open future” reference.)
The formal self-prediction obstruction (Argument A) - Donald MacKay, “On the Logical Indeterminacy of a Free Choice,” Mind 69(273):31–40 (1960). DOI 10.1093/mind/LXIX.273.31. The classic: a correct prediction of your choice can’t be both correct and believed-by-you. Verified — and his “logical” indeterminacy explicitly presupposes determinism and is not mere ignorance, exactly the in-principle distinction §3 needs. Most directly on-point. - Karl Popper, The Open Universe (and “Indeterminism in Quantum Physics and in Classical Physics,” 1950). The “Oedipus effect”; embedded predictors can’t predict themselves; determinism ≠ predictability. - F.W. Lawvere, “Diagonal Arguments and Cartesian Closed Categories” (1969) — the abstract fixed-point theorem, but a category-theory paper. Cite through Noson Yanofsky, “A Universal Approach to Self-Referential Paradoxes, Incompleteness and Fixed Points,” Bulletin of Symbolic Logic 9(3):362–386 (2003) — the accessible, philosophy-citable unification of all diagonal arguments. Lead with Yanofsky, not Lawvere. - David Wolpert, “Physical Limits of Inference,” Physica D (2008). Inference devices cannot fully infer themselves — a physics impossibility result. Best single source for your physicist audience. The “paradox of predictability” literature (Argument A’s precise scholarly home — enter this conversation directly) - Michael Scriven, “An Essential Unpredictability in Human Behaviour” (1965), in Wolman & Nagel (eds.), Scientific Psychology: Principles and Approaches, Basic Books, pp. 411–425. The originating contra-predictor argument. - Patrick Suppes, “On an Example of Unpredictability in Human Behaviour,” Philosophy of Science 31(2):143–148 (1964). - Stefan Rummens & Stefaan Cuypers, “Determinism and the Paradox of Predictability,” Erkenntnis 72(2):233–249 (2010). DOI 10.1007/s10670-009-9199-1. The external-vs-embedded predictability distinction, stated almost verbatim for your §3: determinism gives external predictability but not embedded predictability, and the limit is non-epistemic. Load-bearing. - Richard Holton, “From Determinism to Resignation, and How to Stop It” (2013), in Clark, Kiverstein & Vierkant (eds.), Decomposing the Will, OUP, pp. 87–100. - Victor Gijsbers, “The Paradox of Predictability,” Erkenntnis (2021), DOI 10.1007/s10670-020-00369-3 — the Turing/halting-problem diagnosis; the live counter-position. - Stefan Rummens, “The Roots of the Paradox of Predictability: A Reply to Gijsbers,” Erkenntnis 89(5):2097–2104 (2024). The current state of the debate — read both Gijsbers and this to see where the argument actually stands.
Deliberation vs. prediction (Argument C) - Isaac Levi, “Deliberation Crowds Out Prediction” (in The Covenant of Reason, 1997). Central; the source of the slogan. Also Levi (1989), (1991). - Wolfgang Spohn, “Where Luce and Krantz Do Really Generalize Savage’s Decision Model,” Erkenntnis 11:113–134 (1977). The impossibility of assigning probabilities to one’s own deliberated acts. - Wlodek Rabinowicz, “Does Practical Deliberation Crowd Out Self-Prediction?” Erkenntnis 57(1):91–122 (2002). DOI 10.1023/A:1020106622032. Argues against the thesis — the careful critic you must answer. Read for the objections to C. - Alan Hájek, “Deliberation Welcomes Prediction,” Episteme 13(4):507–528 (2016). The second named critic of crowding-out — engage both Hájek and Rabinowicz in §5, or C is exposed. - James Joyce, The Foundations of Causal Decision Theory (1999); Arif Ahmed, Evidence, Decision and Causality (2014). The deliberating agent’s view of its own acts.
The thermodynamic arrow (grounds Argument B) - David Albert, Time and Chance (2000). The past hypothesis; the statistical-mechanical arrow. The record/influence asymmetry. - Barry Loewer, papers on the past hypothesis, the “Mentaculus,” and agency/free will (the “package deal”). - Carlo Rovelli, “Agency in Physics” (arXiv:2007.05300, 2020) — the citable Rovelli source: agency as “the breaking of an approximation under which dynamics appears closed,” plus the time-orientation and information-source of choice. Near-identical to your thesis, from a physicist. Pair with “Princeton seminars on physics and philosophy” (arXiv:2407.01989, 2024), which has a section titled “Openness” (verified) — good for §6, but lecture notes, so cite the 2020 paper as the formal source.
Determinism ≠ predictability (clears the ground; defeats the demon’s relevance) - John Earman, A Primer on Determinism (1986). The canonical statement that determinism and predictability come apart. Essential for §3/§7. - Pierre-Simon Laplace, A Philosophical Essay on Probabilities (1814). The demon — the position you relocate (not refute) via indexicality.
Compatibilism — position against / recruit - Christian List, Why Free Will Is Real (2019). Closest ally: agential possibility real under determinism via levels. Differentiate: you ground in self-reference + thermodynamics + indexicality, not levels/emergence alone. - Daniel Dennett, Elbow Room (1984) / Freedom Evolves (2003). The naturalized-compatibilist foil — he redefines freedom and deflates openness; you don’t. - Helen Steward, A Metaphysics for Freedom (2012). “Settling” the open future — close, but she leans toward needing indeterminism; a partial opponent worth engaging. - Kadri Vihvelin, Causes, Laws, and Free Will (2013); John Martin Fischer (semicompatibilism). The dispositional/abilities line, to position against.
The indexical / perspectival frame (§6) - Thomas Nagel, The View from Nowhere (1986). The objective/subjective standpoint distinction — the philosophical scaffolding for indexical openness. - Huw Price, Time’s Arrow and Archimedes’ Point (1996). Read as the sharp critic: he distrusts agent-based temporal asymmetries as anthropocentric. Defuse him or he sinks §4/§6.
Novelty in one line: the first paper to fuse the formal self-prediction obstruction, the thermodynamic-temporal grounding, and the deliberation-prediction result into a single indexical account of openness that dissolves the free-will dialectic without taking a side on indeterminism.
Target venues: Philosophers’ Imprint, Erkenntnis, Synthese, British Journal for the Philosophy of Science, Australasian Journal of Philosophy; for the formal/physical version, Entropy (where Wolpert and adjacent work appear). Ismael-, Levi-, and Albert-adjacent editorial territory.
The discipline that makes this the right first paper: write every sentence as if the ethics will never come. If you can convince a determinist physicist who is indifferent to morality that the future is open for the chooser on any physics, you have set the keystone of the entire edifice on ground no opponent of your ethics can touch — because you’ll have proven it without mentioning ethics at all.