Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2602.16985

Retrieving the Baby: Reichenbach's Principle, Bell Locality, and Selection Bias

Huw Price

quant-phphysics.hist-phSubmitted Feb 19, 2026• Announced Feb 20, 2026• 2 citations
arXivHTMLPDF

Abstract

In his late piece 'La nouvelle cuisine' (Bell 1990), John Bell describes the steps from an intuitive, informal principle of locality to a mathematical rule called Factorizability. This rule stipulates that when possible past causes are held fixed, the joint probabilities of outcomes of spacelike separated measurements, conditional on measurement settings, be the product of the local conditional probabilities individually. Bell shows that Factorizability conflicts with predictions of QM, predictions since confirmed in many experiments. However, Bell warns his readers that the steps leading to Factorizability should 'be viewed with the utmost suspicion'. He says that 'it is precisely in cleaning up intuitive ideas for mathematics that one is likely to throw the baby out with the bathwater' (1990, 239). Bell's suspicions were well-founded, for he himself misses an important baby. Here we retrieve and identify it: it is selection bias. We explain how failure of Factorizability may be regarded as a selection artefact, requiring no violation of locality in the intuitive, conceptual sense with which Bell begins his analysis. The argument begins with a central principle of causal discovery, Reichenbach's Principle of Common Cause (PCC). It is well known that correlations due to selection bias are not subject to PCC. Several writers have proposed that EPR-Bell correlations are also an exception to PCC, but it has not been noticed that they fall under this well-known exclusion. The point is relevant not only to the status of Bell nonlocality, but also for statistics and causal modeling. For these fields, the news is that selection effects play a ubiquitous role in quantum phenomena, in a form akin to collider bias.

Comments
21 pages, 3 figures

Related papers

Bell on Bell's Theorem: The Changing Face of Nonlocality

Harvey R. Brown, Christopher G. Timpson

Between 1964 and 1990, the notion of nonlocality in Bell's papers underwent a profound change as his nonlocality theorem gradually became detached from quantum mechanics, and referred to wider probabilistic theories involving correlations between separated beables. The proposition that standard quantum mechanics is itself nonlocal (more precisely, that it violates ‘local causality’) became divorced from the Bell theorem per se from 1976 on, although this important point is widely overlooked in the literature. In 1990, the year of his death, Bell would express serious misgivings about themathematical form of the local causality condition and leave ill-defined the issue of the consistency between special relativity and violation of the Bell-type inequality. In our view, the significance of the Bell theorem, in both its deterministic and stochastic forms, can only be fully understood by taking into account the fact that a fully Lorentz covariant version of quantum theory, free of action at a distance, can be articulated in the Everett interpretation. Introduction John S. Bell's last word on his celebrated nonlocality theorem and its interpretation appeared in his 1990 paper ‘La nouvelle cuisine’, first published in the year of his untimely death. Bell was careful here to distinguish between the issue of ‘no superluminal signalling’ in quantum theory (both quantum field theory and quantum mechanics) and a principle he first introduced explicitly in 1976 and called ‘local causality’ [1]. In relation to the former, Bell expressed concerns that amplify doubts he had already expressed in 1976. These concerns touch on what is now widely known as the no-signalling theorem in quantum mechanics, and ultimately have to do with Bell's distaste for what he saw as an anthropocentric element in orthodox quantum thinking. In relation to local causality, Bell emphasised that his famous factorizability (no-correlations) condition is not to be seen ‘as the formulation of local causality, but as a consequence thereof’ and stressed how difficult he found it to articulate this consequence. He left the question of any strict inconsistency between violation of factorizability and special relativity theory unresolved, a not insignificant shift from his thinking up to the early 1980s.

arXiv:1501.03521•Cambridge University Press eBooks•Aug 31, 2016•33 citations

Quantum nonlocality: an analysis of the implications of Bell's Theorem and quantum correlations for nonlocality.

Joseph Berkovitz

Bell's Theorem demonstrates that factorizable theories for the EPR experiment (EPR) cannot reproduce the quantum correlations. Factorizability is motivated by various locality conditions. So to understand the nature of nonlocality in EPR, we first need to understand the conceptual relations between factorizability, these various locality conditions and the nature of the quantum correlations. That is the main focus of my thesis. My main conclusion is that these conceptual relations are more subtle than the literature has usually suggested. Chapter 1 is a general introduction. In Chapter 2, I review the general framework of factorizable stochastic theories for EPR. I argue that factorizability can be motivated by various locality conditions, even in theories that admit time-dependent states and take the measurement interactions to be neither instantaneous, nor simultaneous. In Chapter 3, I focus on Cartwright' s (1989) and Humphreys' (1989) theories of probabilistic causation for singular events, which are based on modifications of traditional causal linear modelling. I argue (against Cartwright) that local models for EPR in the framework of these theories are committed to factorizability; and so they cannot reproduce the EPR correlations. In Chapter 4, I turn to Stochastic-Einstein Locality (SEL). Hellman (1982) proposed that SEL with some provisos characterizes the No-Superluminal-Action (NSA) of the Special Theory of Relativity (STR), and he argued that SEL is not violated in EPR. Butterfield (1994) proposed that SEL (without Hellman's provisos) characterizes the lack of superluminal Lewisian causation, and he argued that SEL is violated in EPR. I argue that SEL (without Hellman's provisos) is motivated by NSA and spatiotemporal separability. Thus, the violation of SEL might arise from the failure of spatiotemporal separability. And this failure is compatible with NSA and superluminal Lewisian causation. Accordingly, Hellman's and Butterfield's views need not be in tension. In Chapter 5, I focus on the implications for nonlocality of Jarrett's (1984) analysis of factorizability into "locality" and "completeness". I argue that although this analysis cannot distinguish between failures of factorizability which are compatible with STR and those which are not, it is significant for clarifying the implications of Bell's theorem for nonlocality. In Chapter 6, I qmsider three arguments that are intended to deny superluminal causal propagation in EPR, and a fourth argument that is intended to establish the opposite conclusion. I argue that in various ways these arguments have gone wrong. Three of these argument rely on implicit assumptions that have been overlooked. Accordingly, two of them reached a reasonable conclusion, i.e. the failure of separability, for the wrong reasons; whereas the third reached a wrong conclusion, i.e. that the failure of the contiguity of causal processes explains the failure of factorizability in EPR. The fourth relies on a wrong assumption, and thus it reaches a too strong conclusion, that the EPR correlations require superluminal action. My main conclusion is that the quantum correlations require nonseparability . In Chapter 7, I focus on decision theory in the context of EPR. In both EPR and the famous Newcomb's problem (NewProb), there are unscreenable-off correlations. I argue that NewProb can be related to EPR, in the sense that a NewProb can be realized by that experiment.

•Jan 30, 1996•3 citations

Stochastic Einstein Locality Revisited

Jeremy Butterfield

I discuss various formulations of stochastic Einstein locality (SEL), which is a version of the idea of relativistic causality, that is, the idea that influences propagate at most as fast as light. SEL is similar to Reichenbach's Principle of the Common Cause (PCC), and Bell's Local Causality.My main aim is to discuss formulations of SEL for a fixed background spacetime. I previously argued that SEL is violated by the outcome dependence shown by Bell correlations, both in quantum mechanics and in quantum field theory. Here I reassess those verdicts in the light of some recent literature which argues that outcome dependence does not violate the PCC. I argue that the verdicts about SEL still stand.Finally, I briefly discuss how to formulate relativistic causality if there is no fixed background spacetime. 1. Introduction2. Formulating Stochastic Einstein Locality2.1. Events and regions2.2. The idea of SEL2.3. Three formulations of SEL2.3.1. The formulations2.3.2. Comparisons2.4. Implications between the formulations2.4.1. Conditions for the equivalence of SELD1 and SELD22.4.2. Conditions for the equivalence of SELS and SELD23. Relativistic Causality in the Bell Experiment3.1. The background3.1.1. The Bell experiment reviewed3.1.2. My previous position3.2. A common common cause? The Budapest school3.2.1. Resuscitating the PCC3.2.2. Known proofs of a Bell inequality need a strong PCC3.2.3. Two distinctions3.2.4. Szabó's model3.2.5. A common common cause is plausible3.2.6. Bell inequalities from a weak PCC: the Bern school3.3. SEL in the Bell experiment3.3.1. PCC and SEL are connected by PPSI3.3.2. The need for other judgments3.3.3. Weak vs. strong SELD4. SEL in Algebraic Quantum Field Theory4.1. The story so far4.2. Questions4.2.1. Our formulations4.2.2. The Budapest and Bern schools5. SEL in Dynamical Spacetimes5.1. SEL for metric structure?5.2. SEL for causal sets?5.2.1. The causal set approach5.2.2. Labelled causal sets; general covariance5.2.3. Deducing the dynamics5.2.4. The fate of SEL

arXiv:0708.2192•The British Journal for the Philosophy of Science•Nov 22, 2007•57 citations

Taming Entanglement

Huw Price, K. B. Wharton

In statistics and causal modeling it is common for a selection process to induce correlations in a subset of an uncorrelated ensemble. We propose that EPR and Bell correlations are selection artefacts of this kind. The selection process is preparation of the initial state of the relevant experiments. Choice of initial state amounts to preselection of a subensemble of a larger, uncorrelated, virtual ensemble of possible histories. Because it is preselection rather than postselection, the resulting correlations support the intuitive counterfactuals of the EPR argument and Bell nonlocality. In this respect, and in its temporal orientation, the case differs from familiar forms of selection bias. Given the ubiquity of quantum entanglement, the result may thus be of independent interest to students of causal modeling. The paper concludes with a discussion of its novel implications in that field.

arXiv:2507.15128•Jul 20, 2025•5 citations

John S. Bell’s concept of local causality

Travis Norsen

John Stewart Bell’s famous theorem is widely regarded as one of the most important developments in the foundations of physics. Yet even as we approach the 50th anniversary of Bell’s discovery, its meaning and implications remain controversial. Many workers assert that Bell’s theorem refutes the possibility suggested by Einstein, Podolsky, and Rosen (EPR) of supplementing ordinary quantum theory with “hidden” variables that might restore determinism and/or some notion of an observer-independent reality. But Bell himself interpreted the theorem very differently—as establishing an “essential conflict” between the well-tested empirical predictions of quantum theory and relativistic local causality. Our goal is to make Bell’s own views more widely known and to explain Bell’s little-known formulation of the concept of relativistic local causality on which his theorem rests. We also show precisely how Bell’s formulation of local causality can be used to derive an empirically testable Bell-type inequality and to recapitulate the EPR argument.

arXiv:0707.0401•American Journal of Physics•Nov 29, 2011•157 citations

Bell Correlations and Selection Bias

Huw Price

Methods of selecting samples from larger populations may produce bias, in either direction: inducing correlations between variables independent in the full population, or masking correlations between variables dependent in the full population. Here we propose a surprising application of these familiar ideas. We argue that they are relevant to puzzling correlations uncovered in quantum theory by John Stewart Bell (Bell 1964). In the light of Bell's work and subsequent experiments it is widely believed that the quantum world is 'nonlocal', in apparent tension with relativity. Many hold that the only alternative is to abandon 'realism', the view that there is an objective world independent of measurement. We propose instead that Bell's correlations are selection artefacts, in tension neither with relativity nor realism. In standard two-particle Bell experiments the relevant selection is a preselection, achieved by the preparation of the initial state. Again, selection bias via preselection is familiar elsewhere in science, but it doesn't seem to have been noticed that it is applicable in this case.

quant-pharXiv:2605.00406•May 4, 2026•0 citations
In his late piece 'La nouvelle cuisine' (Bell 1990), John Bell describes the steps from an intuitive, informal principle of locality to a mathematical rule called Factorizability. This rule stipulates that when possible past causes are held fixed, the joint probabilities of outcomes of spacelike separated measurements, conditional on measurement settings, be the product of the local conditional probabilities individually. Bell shows that Factorizability conflicts with predictions of QM, predictions since confirmed in many experiments. However, Bell warns his readers that the steps leading to Factorizability should 'be viewed with the utmost suspicion'. He says that 'it is precisely in cleaning up intuitive ideas for mathematics that one is likely to throw the baby out with the bathwater' (1990, 239). Bell's suspicions were well-founded, for he himself misses an important baby. Here we retrieve and identify it: it is selection bias. We explain how failure of Factorizability may be regarded as a selection artefact, requiring no violation of locality in the intuitive, conceptual sense with which Bell begins his analysis. The argument begins with a central principle of causal discovery, Reichenbach's Principle of Common Cause (PCC). It is well known that correlations due to selection bias are not subject to PCC. Several writers have proposed that EPR-Bell correlations are also an exception to PCC, but it has not been noticed that they fall under this well-known exclusion. The point is relevant not only to the status of Bell nonlocality, but also for statistics and causal modeling. For these fields, the news is that selection effects play a ubiquitous role in quantum phenomena, in a form akin to collider bias.
Between 1964 and 1990, the notion of nonlocality in Bell's papers underwent a profound change as his nonlocality theorem gradually became detached from quantum mechanics, and referred to wider probabilistic theories involving correlations between separated beables. The proposition that standard quantum mechanics is itself nonlocal (more precisely, that it violates ‘local causality’) became divorced from the Bell theorem per se from 1976 on, although this important point is widely overlooked in the literature. In 1990, the year of his death, Bell would express serious misgivings about themathematical form of the local causality condition and leave ill-defined the issue of the consistency between special relativity and violation of the Bell-type inequality. In our view, the significance of the Bell theorem, in both its deterministic and stochastic forms, can only be fully understood by taking into account the fact that a fully Lorentz covariant version of quantum theory, free of action at a distance, can be articulated in the Everett interpretation. Introduction John S. Bell's last word on his celebrated nonlocality theorem and its interpretation appeared in his 1990 paper ‘La nouvelle cuisine’, first published in the year of his untimely death. Bell was careful here to distinguish between the issue of ‘no superluminal signalling’ in quantum theory (both quantum field theory and quantum mechanics) and a principle he first introduced explicitly in 1976 and called ‘local causality’ [1]. In relation to the former, Bell expressed concerns that amplify doubts he had already expressed in 1976. These concerns touch on what is now widely known as the no-signalling theorem in quantum mechanics, and ultimately have to do with Bell's distaste for what he saw as an anthropocentric element in orthodox quantum thinking. In relation to local causality, Bell emphasised that his famous factorizability (no-correlations) condition is not to be seen ‘as the formulation of local causality, but as a consequence thereof’ and stressed how difficult he found it to articulate this consequence. He left the question of any strict inconsistency between violation of factorizability and special relativity theory unresolved, a not insignificant shift from his thinking up to the early 1980s.
Bell's Theorem demonstrates that factorizable theories for the EPR experiment (EPR) cannot reproduce the quantum correlations. Factorizability is motivated by various locality conditions. So to understand the nature of nonlocality in EPR, we first need to understand the conceptual relations between factorizability, these various locality conditions and the nature of the quantum correlations. That is the main focus of my thesis. My main conclusion is that these conceptual relations are more subtle than the literature has usually suggested. Chapter 1 is a general introduction. In Chapter 2, I review the general framework of factorizable stochastic theories for EPR. I argue that factorizability can be motivated by various locality conditions, even in theories that admit time-dependent states and take the measurement interactions to be neither instantaneous, nor simultaneous. In Chapter 3, I focus on Cartwright' s (1989) and Humphreys' (1989) theories of probabilistic causation for singular events, which are based on modifications of traditional causal linear modelling. I argue (against Cartwright) that local models for EPR in the framework of these theories are committed to factorizability; and so they cannot reproduce the EPR correlations. In Chapter 4, I turn to Stochastic-Einstein Locality (SEL). Hellman (1982) proposed that SEL with some provisos characterizes the No-Superluminal-Action (NSA) of the Special Theory of Relativity (STR), and he argued that SEL is not violated in EPR. Butterfield (1994) proposed that SEL (without Hellman's provisos) characterizes the lack of superluminal Lewisian causation, and he argued that SEL is violated in EPR. I argue that SEL (without Hellman's provisos) is motivated by NSA and spatiotemporal separability. Thus, the violation of SEL might arise from the failure of spatiotemporal separability. And this failure is compatible with NSA and superluminal Lewisian causation. Accordingly, Hellman's and Butterfield's views need not be in tension. In Chapter 5, I focus on the implications for nonlocality of Jarrett's (1984) analysis of factorizability into "locality" and "completeness". I argue that although this analysis cannot distinguish between failures of factorizability which are compatible with STR and those which are not, it is significant for clarifying the implications of Bell's theorem for nonlocality. In Chapter 6, I qmsider three arguments that are intended to deny superluminal causal propagation in EPR, and a fourth argument that is intended to establish the opposite conclusion. I argue that in various ways these arguments have gone wrong. Three of these argument rely on implicit assumptions that have been overlooked. Accordingly, two of them reached a reasonable conclusion, i.e. the failure of separability, for the wrong reasons; whereas the third reached a wrong conclusion, i.e. that the failure of the contiguity of causal processes explains the failure of factorizability in EPR. The fourth relies on a wrong assumption, and thus it reaches a too strong conclusion, that the EPR correlations require superluminal action. My main conclusion is that the quantum correlations require nonseparability . In Chapter 7, I focus on decision theory in the context of EPR. In both EPR and the famous Newcomb's problem (NewProb), there are unscreenable-off correlations. I argue that NewProb can be related to EPR, in the sense that a NewProb can be realized by that experiment.
I discuss various formulations of stochastic Einstein locality (SEL), which is a version of the idea of relativistic causality, that is, the idea that influences propagate at most as fast as light. SEL is similar to Reichenbach's Principle of the Common Cause (PCC), and Bell's Local Causality.My main aim is to discuss formulations of SEL for a fixed background spacetime. I previously argued that SEL is violated by the outcome dependence shown by Bell correlations, both in quantum mechanics and in quantum field theory. Here I reassess those verdicts in the light of some recent literature which argues that outcome dependence does not violate the PCC. I argue that the verdicts about SEL still stand.Finally, I briefly discuss how to formulate relativistic causality if there is no fixed background spacetime. 1. Introduction2. Formulating Stochastic Einstein Locality2.1. Events and regions2.2. The idea of SEL2.3. Three formulations of SEL2.3.1. The formulations2.3.2. Comparisons2.4. Implications between the formulations2.4.1. Conditions for the equivalence of SELD1 and SELD22.4.2. Conditions for the equivalence of SELS and SELD23. Relativistic Causality in the Bell Experiment3.1. The background3.1.1. The Bell experiment reviewed3.1.2. My previous position3.2. A common common cause? The Budapest school3.2.1. Resuscitating the PCC3.2.2. Known proofs of a Bell inequality need a strong PCC3.2.3. Two distinctions3.2.4. Szabó's model3.2.5. A common common cause is plausible3.2.6. Bell inequalities from a weak PCC: the Bern school3.3. SEL in the Bell experiment3.3.1. PCC and SEL are connected by PPSI3.3.2. The need for other judgments3.3.3. Weak vs. strong SELD4. SEL in Algebraic Quantum Field Theory4.1. The story so far4.2. Questions4.2.1. Our formulations4.2.2. The Budapest and Bern schools5. SEL in Dynamical Spacetimes5.1. SEL for metric structure?5.2. SEL for causal sets?5.2.1. The causal set approach5.2.2. Labelled causal sets; general covariance5.2.3. Deducing the dynamics5.2.4. The fate of SEL
In statistics and causal modeling it is common for a selection process to induce correlations in a subset of an uncorrelated ensemble. We propose that EPR and Bell correlations are selection artefacts of this kind. The selection process is preparation of the initial state of the relevant experiments. Choice of initial state amounts to preselection of a subensemble of a larger, uncorrelated, virtual ensemble of possible histories. Because it is preselection rather than postselection, the resulting correlations support the intuitive counterfactuals of the EPR argument and Bell nonlocality. In this respect, and in its temporal orientation, the case differs from familiar forms of selection bias. Given the ubiquity of quantum entanglement, the result may thus be of independent interest to students of causal modeling. The paper concludes with a discussion of its novel implications in that field.
John Stewart Bell’s famous theorem is widely regarded as one of the most important developments in the foundations of physics. Yet even as we approach the 50th anniversary of Bell’s discovery, its meaning and implications remain controversial. Many workers assert that Bell’s theorem refutes the possibility suggested by Einstein, Podolsky, and Rosen (EPR) of supplementing ordinary quantum theory with “hidden” variables that might restore determinism and/or some notion of an observer-independent reality. But Bell himself interpreted the theorem very differently—as establishing an “essential conflict” between the well-tested empirical predictions of quantum theory and relativistic local causality. Our goal is to make Bell’s own views more widely known and to explain Bell’s little-known formulation of the concept of relativistic local causality on which his theorem rests. We also show precisely how Bell’s formulation of local causality can be used to derive an empirically testable Bell-type inequality and to recapitulate the EPR argument.
Methods of selecting samples from larger populations may produce bias, in either direction: inducing correlations between variables independent in the full population, or masking correlations between variables dependent in the full population. Here we propose a surprising application of these familiar ideas. We argue that they are relevant to puzzling correlations uncovered in quantum theory by John Stewart Bell (Bell 1964). In the light of Bell's work and subsequent experiments it is widely believed that the quantum world is 'nonlocal', in apparent tension with relativity. Many hold that the only alternative is to abandon 'realism', the view that there is an objective world independent of measurement. We propose instead that Bell's correlations are selection artefacts, in tension neither with relativity nor realism. In standard two-particle Bell experiments the relevant selection is a preselection, achieved by the preparation of the initial state. Again, selection bias via preselection is familiar elsewhere in science, but it doesn't seem to have been noticed that it is applicable in this case.