0% found this document useful (0 votes)

61 views108 pages

Bayesian Models of The Mind

Bayesian decision theory is a mathematical framework that models reasoning and decision-making under uncertainty, increasingly used by cognitive scientists to describe actual mental processes. This document serves as an introduction to Bayesian cognitive science, highlighting its applications in understanding mental phenomena such as perception and navigation. It argues for the importance of mental representation within Bayesian models, asserting that these models provide empirical support for a representationalist view of the mind.

Uploaded by

Anderson

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

61 views108 pages

Bayesian Models of The Mind

Uploaded by

Anderson

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 108

Rescorla

Bayesian decision theory is a mathematical framework that

models reasoning and decision-making under uncertain
conditions. The Bayesian paradigm originated as a theory of
how people should operate, not a theory of how they actually
operate. Nevertheless, cognitive scientists increasingly use it
to describe the actual workings of the human mind. Over the Philosophy
past few decades, cognitive science has produced impressive
Bayesian models of mental activity. The models postulate that of Mind
certain mental processes conform, or approximately conform,
to Bayesian norms. Bayesian models offered within cognitive
science have illuminated numerous mental phenomena, such
as perception, motor control, and navigation. This Element

Bayesian Models of the Mind

Bayesian Models
provides a self-contained introduction to the foundations of
Bayesian cognitive science. It then explores what we can learn
about the mind from Bayesian models offered by cognitive
scientists.
of the Mind
About the Series Series Editor
This series provides concise, authoritative Keith Frankish
introductions to contemporary work The University
in philosophy of mind, written by of Sheffield
leading researchers and including both
established and emerging topics. It
provides an entry point to the primary
Michael Rescorla
literature and will be the standard
resource for researchers, students, and
anyone wanting a firm grounding in this
fascinating field.

Cover image: Baac3nes / Moment /

Getty Images ISSN 2633-9080 (online)
ISSN 2633-9072
Downloaded from https://www.cambridge.org/core. IP address: 201.224.145.68, on 01 Feb 2025 at 14:48:00, subject to the Cambridge (print)
Core terms of use,
available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108955973
Downloaded from https://www.cambridge.org/core. IP address: 201.224.145.68, on 01 Feb 2025 at 14:48:00, subject to the Cambridge Core terms of
use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108955973
Elements in the Philosophy of Mind
edited by
Keith Frankish
The University of Sheﬃeld

BAYESIAN MODELS
OF THE MIND

Michael Rescorla
University of California, Los Angeles

Cambridge University Press is part of Cambridge University Press & Assessment,

a department of the University of Cambridge.
We share the University’s mission to contribute to society through the pursuit of
education, learning and research at the highest international levels of excellence.

www.cambridge.org
Information on this title: www.cambridge.org/9781009517805
DOI: 10.1017/9781108955973
© Michael Rescorla 2024
This publication is in copyright. Subject to statutory exception and to the provisions
of relevant collective licensing agreements, no reproduction of any part may take
place without the written permission of Cambridge University Press & Assessment.
When citing this work, please include a reference to the DOI 10.1017/9781108955973
First published 2024
A catalogue record for this publication is available from the British Library
ISBN 978-1-009-51780-5 Hardback
ISBN 978-1-108-95829-5 Paperback
ISSN 2633-9080 (online)
ISSN 2633-9072 (print)
Cambridge University Press & Assessment has no responsibility for the persistence
or accuracy of URLs for external or third-party internet websites referred to in this
publication and does not guarantee that any content on such websites is, or will
remain, accurate or appropriate.

Elements in the Philosophy of Mind

DOI: 10.1017/9781108955973
First published online: December 2024

Michael Rescorla
University of California, Los Angeles
Author for correspondence: Michael Rescorla, [email protected]

Abstract: Bayesian decision theory is a mathematical framework that

models reasoning and decision-making under uncertain conditions.
The Bayesian paradigm originated as a theory of how people should
operate, not a theory of how they actually operate. Nevertheless,
cognitive scientists increasingly use it to describe the actual workings of
the human mind. Over the past few decades, cognitive science has
produced impressive Bayesian models of mental activity. The models
postulate that certain mental processes conform, or approximately
conform, to Bayesian norms. Bayesian models oﬀered within cognitive
science have illuminated numerous mental phenomena, such as
perception, motor control, and navigation. This Element provides a
self-contained introduction to the foundations of Bayesian cognitive
science. It then explores what we can learn about the mind from
Bayesian models oﬀered by cognitive scientists.

Keywords: Bayesian decision theory, mental representation, Bayesian

cognitive science, subjective probability, conditionalization

© Michael Rescorla 2024

ISBNs: 9781009517805 (HB), 9781108958295 (PB), 9781108955973 (OC)
ISSNs: 2633-9080 (online), 2633-9072 (print)

1 Introduction 1

2 The Probability Calculus 2

3 Bayesian Decision Theory 18

4 Bayesian Cognitive Science 27

5 Realism and Instrumentalism 42

6 Mental Representation 52

7 Anti-representationalism 65

8 Conclusion 75

Appendix: Foundations of Probability Theory 76

References 88

1 Introduction
Thomas Bayes was an eighteenth-century minister and mathematician who passed
his life in relative obscurity. Upon his death in 1761, his friend Richard Price found
among his papers a document entitled “An Essay Towards Solving a Problem in the
Doctrine of Chances.” Price, recognizing the essay’s immense significance, saw to
its posthumous publication (Bayes, 1763). Bayes’s insights gave birth to what is
now known as Bayesian decision theory: a mathematical framework that models
reasoning and decision-making under uncertain conditions. Named after Bayes due
to his founding insights, the framework was first systematically articulated by
Pierre-Simon Laplace (1814/1902). Despite frequent vicissitudes in development,
reception, and application, the framework attracted increasingly many adherents
beginning in the early twentieth century and accelerating as the century progressed
(McGrayne, 2011). It currently enjoys great popularity, finding widespread use
within statistics (Berger, 1985; Gelman et al., 2014), philosophy (Earman, 1992),
machine learning (Murphy, 2023), robotics (Thrun, Burgard & Fox, 2005), physics
(Trotta, 2008), medical science (Ashby, 2006), and myriad other disciplines.
Bayesian decision theory originated as a theory of how people should oper-
ate, not a theory of how they actually operate. Nevertheless, cognitive scientists
increasingly use it to describe the actual workings of the human mind. Over the
past few decades, cognitive science has produced impressive Bayesian models
of mental activity. The models postulate that certain mental processes conform,
or approximately conform, to Bayesian norms. Bayesian models offered within
cognitive science have illuminated numerous mental phenomena, such as
perception, motor control, and navigation.
This Element has a two-fold purpose. First, it provides a self-contained intro-
duction to the foundations of Bayesian cognitive science. Second, it explores
what we can learn about the mind from Bayesian models offered by cognitive
scientists.
On the second front, my main concern is how Bayesian cognitive science
relates to mental representation. Just as the heart serves to pump blood and the
stomach serves to digest food, one of the mind’s principal functions is to
represent the world. For instance, I have various beliefs about Napoleon: that
he was born in Corsica, that he was an emperor, and so on. Thus, the mind
somehow reaches beyond itself to represent external reality. In that sense, the
mind is a representational organ. Historically, most philosophers have agreed
that the mind’s representational capacity is among its key features. However,
prominent scientists and philosophers throughout the past century have ques-
tioned whether representation deserves any place in the science of the mind. As
a result, controversy continues to fester over the explanatory value of mental

representation. Representationalists such as Burge (2010; 2022), Fodor (1975;

1987; 2008), Peacocke (1994; 1999), Pylyshyn (1984), and Shea (2018) insist
that mental representation plays a vital role within the scientific explanation of
various core mental phenomena. Anti-representationalists as varied as Chemero
(2009), Churchland (1981), Field (2001), Quine (1960), Ramsey (2007), Stich
(1983), and van Gelder (1992) reject this position.
I will argue that Bayesian cognitive science assigns mental representation
a central explanatory role. Bayesian models of perception, motor control,
navigation, and other core mental activities posit representational mental states.
Explanations supplied by the models characterize both explananda and explan-
antia in thoroughly representational terms. So Bayesian cognitive science
presupposes the traditional picture of the mind as a representational organ. It
invests that picture with unprecedented empirical substance through well-
confirmed, mathematically rigorous models.
Sections 2 and 3 present key elements of Bayesian decision theory. Section 4
surveys how cognitive scientists use the Bayesian framework to model mental
activity. Section 5 articulates a realist stance towards Bayesian models of the
mind: when a Bayesian model is explanatorily successful, we have good reason to
believe that the model describes actual mental states and processes with at least
approximate accuracy. Sections 6 and 7 argue that representational properties of
mental states figure crucially in explanations provided by Bayesian cognitive
science. My conclusion: Bayesian modeling supports a representationalist
perspective on the mind.
My exposition contains more mathematics than most writings on philosophy
of mind. The technical content reflects my conviction that fully understanding
mental representation requires familiarity with the mathematical language used
by scientists to study mental representation. I hope that this Element will help
some readers achieve the requisite familiarity and will promote greater appre-
ciation for the benefits that such familiarity affords. To keep the text as access-
ible as possible, I have confined many technical details to the Appendix.

2 The Probability Calculus

The core notion of Bayesian decision theory is credence, or subjective
probability—a quantitative measure of the degree to which an agent believes
a hypothesis. I may have low credence that a meteor shower occurred ﬁve days
ago, higher credence that Seabiscuit will win the race tomorrow, and even higher
credence that Napoleon was born in Corsica. An agent’s credence in hypothesis
H is notated as P(H). Credences are psychological facets of the individual agent,
not objective chances or frequencies out in the world. The agent’s credences need

not track any objective probabilities that inhere in mind-independent reality. To

illustrate, suppose that a biased coin has objective chance 0.3 of landing heads.
I may mistakenly believe that the coin is fair and therefore assign subjective
probability 0.5 to the hypothesis that it will land heads. Then my credence departs
dramatically from the objective chance of heads.
What is it to attach a credence to a hypothesis? What does it mean for an
agent to set P(H) = x as opposed to P(H) = y 6¼ x? Beginning with Ramsey
(1931) and de Finetti (1937/1980), many authors have tried to answer these
questions (Erikkson & Hájek, 2007). In practice, contemporary Bayesians
usually leave the questions unanswered. They take the notion of credence
as primitive, without providing noncircular necessary and sufﬁcient condi-
tions for an agent to attach a credence to a hypothesis. This is the strategy
pursued within Bayesian cognitive science, and it is the strategy I will
pursue.
Bayesian decision theory was given a secure mathematical grounding
by Kolmogorov (1933/1956), who articulated axioms for probability in his
landmark Foundations of the Theory of Probability. The axioms are not
speciﬁc to subjective probability; they apply equally to objective probability.
Section 2 expounds basic aspects of Kolmogorov’s axiomatization, which is
sometimes called the probability calculus. Section 3 discusses how Bayesians
use the probability calculus to model uncertainty.1

2.1 Sets of Outcomes

Kolmogorov’s axiomatization uses set theory as a basis for probability theory.
The central notion of set theory is membership:

ω 2 A;

meaning that ω is a member of set A. We also say that ω belongs to A.

In Kolmogorov’s axiomatization, probabilities attach to sets of outcomes
drawn from an outcome space Ω. To illustrate, suppose we want to model
probabilities over the result of a player rolling a six-sided die. We may take
the outcome space to be

Ω ¼ f1; 2; 3; 4; 5; 6g;

1
Readers seeking a more leisurely introduction to Bayesian decision theory have many options
pitched at varying levels of difﬁculty. Hacking (2001) is aimed at philosophers and makes
relatively modest mathematical demands. Stone (2013) occupies an intermediate level of difﬁ-
culty. Berger (1985) and Gelman et al. (2014) are standard statistical references and are more
mathematically demanding.

that is, the set containing elements 1, 2, 3, 4, 5, and 6. The hypothesis that the
player rolls an even number corresponds to the set

f2; 4; 6g:

Similarly, suppose we seek to deﬁne probabilities over possible results of

a horse race. We can specify an outcome by describing the order in which the
horses ﬁnish. Ω contains each such outcome. The hypothesis that Seabiscuit
wins the race corresponds to the set

fω : Seabiscuit finishes before every other horse in ωg;

that is, the set of outcomes in which Seabiscuit finishes before every other horse.
Philosophers commonly assume that probabilities attach to propositions. In
the scientific and mathematical literature, one rarely finds any appeal to proposi-
tions. Instead, researchers follow Kolmogorov in attaching probabilities to sets.
Under certain assumptions, one can recapture talk about “propositions” within
Kolmogorov’s setting. One can treat Ω as containing possible worlds, and one
can analyze propositions as sets of possible worlds (Stalnaker, 1984). These
assumptions are not mandated by Kolmogorov’s axiomatization. For example,
the simple outcome space f1; 2; 3; 4; 5; 6g is allowed by Kolmogorov’s axio-
matization, even though its elements are not possible worlds.
When probabilities attach to sets of outcomes, elementary set-theoretic oper-
ations mimic the propositional operations negation, conjunction, and disjunction:

• Negation corresponds to complementation. The complement of set A is

the set Ac containing all elements that are in Ω but not in A. See Figure 1.
The hypothesis that the player rolls 1 is the set
f1g;

Figure 1 Ω is the rectangle. A is the ellipse. Ac is shaded gray.

Figure 2 A and B are overlapping sets. Their intersection A ∩ B is

shaded gray.

while the hypothesis that the player does not roll 1 is its complement

f1gc ¼ f2; 3; 4; 5; 6g:

• Conjunction corresponds to intersection. The intersection of A and B is the

set containing all elements that are in both A and B. The intersection is written
as A ∩ B. See Figure 2. The hypothesis that the player rolls an even number
and the player rolls a number greater than 3 is the intersection

f2; 4; 6g ∩ f4; 5; 6g ¼ f4; 6g:

If we intersect together disjoint sets (i.e. sets that share no members), then the
result is the empty set ∅ containing no members. The hypothesis that the
player rolls an even number and the player rolls an odd number is

f2; 4; 6g ∩ f1; 3; 5g ¼ ∅ :

• Disjunction corresponds to union. The union of A and B is the set containing

all elements that are in A or B. The union is written as A ∪ B. See Figure 3. The
hypothesis that the player rolls 1 or the player rolls 4 is the union

f1g ∪ f4g ¼ f1; 4g:

By iteratively applying set-theoretic operations, Kolmogorov replicates the

formation of logically complex sentences or propositions.
In simple applications, such as a die roll or horse race, the outcome space Ω is
finite. Many applications require Ω to be infinite. For example, consider an
asteroid’s speed as it enters our solar system. There are infinitely many possible

Figure 3 A and B are overlapping sets. Their union A ∪ B is shaded gray.

asteroid speeds. If we want to model probabilities over possible asteroid speeds,

we need an inﬁnite outcome space.

2.2 Axioms of the Probability Calculus

In probability theory, sets of outcomes are called events. The probability
calculus contains three axioms that govern the assignment of probabilities to
events:

Axiom 1: Probabilities are real numbers between 0 and 1,

where a real number is any number that can be expressed as a decimal. As

applied to subjective probability, Axiom 1 sets a scale for degrees of belief. 1 is
the maximal possible degree of belief. 0 is the minimum. When an agent assigns
probability 1 to an event, we say that the agent is certain of the event.

Axiom 2: PðΩÞ ¼ 1:

Intuitively: Ω exhausts all relevant possibilities, so it must receive maximal

degree of belief.

Axiom 3: Additivity.

To elucidate additivity, suppose that H1 and H2 are disjoint events. For example,
let H1 be the hypothesis that Seabiscuit wins the race and H2 the hypothesis that
War Admiral wins the race. Consider the union H1 ∪ H2 : the hypothesis that
Seabiscuit wins the race or War Admiral wins the race. Additivity requires that:

PðH1 ∪ H2 Þ ¼ PðH1 Þ þ PðH2 Þ:

Figure 4 H1 and H2 are disjoint events. Additivity requires that their union (the
total shaded area) receive a probability equal to the sum of the probabilities
assigned to them individually.

In general, the probability that either of two disjoint events occurs is found by
adding together the probabilities assigned to the individual events. See Figure 4.
As discussed in Section A2, Kolmogorov ultimately uses a somewhat stronger
version of additivity than I have articulated here.
Axioms 1–3 can be applied to objective probabilities or to subjective prob-
abilities. Applied to objective probabilities, they are construed as constraints
that probabilities do in fact satisfy. Applied to subjective probabilities, they are
construed as constraints that probabilities should satisfy: an agent is rational to
the extent that her credences satisfy the axioms.2
The core tenet of Bayesian decision theory is that credences should conform
to the probability calculus axioms. Since Bayesians advance the probability
calculus axioms as normative constraints, we may ask why these particular
axioms are supposed to be rationally privileged. Why is someone who conforms

2
An alternative formulation of the probability calculus centers on sentences rather than sets.
Whereas Kolmogorov assigns probabilities to sets of outcomes, the alternative formulation
assigns them to sentences drawn from a suitable language. One can develop probability theory
on this alternative sentential basis (e.g. Gaifman & Snir, 1982). Some Bayesian models found in
cognitive science, especially models of high-level cognition, use sentential rather than set-
theoretic axiomatization (Piantadosi & Jacobs, 2016). For example, sentential models have
been successfully applied to causality (Goodman, Ullman, & Tenenbaum, 2011), kinship (Katz
et al., 2008), and analogical reasoning (Cheyette & Piantadosi, 2017). However, set-theoretic
axiomatization underlies the vast bulk of research in Bayesian cognitive science, including all or
virtually all research into relatively low-level processes such as perception, motor control, and
navigation. This Element focuses exclusively on models that use Kolmogorov’s set-theoretic
axiomatization. Much of what I say about those models would apply, in suitably modiﬁed form, to
models that use sentential axiomatization.

to the axioms rationally superior to someone who violates them? A large

literature, stretching back to Ramsey (1931) and de Finetti (1937/1980), seeks
to answer this question by providing a foundational justiﬁcation for the prob-
ability calculus axioms (Easwaran, 2011a; Pettigrew, 2019; Pettigrew, 2020;
Weisberg, 2009). For present purposes, I simply assume that the probability
calculus axioms are rational constraints on credence.
From a mathematical perspective, we regard Axioms 1–3 as constraints on
a function P that maps each event H to a real number PðHÞ. When P satisﬁes all
three constraints, it is called a probability distribution or a probability measure.3

2.3 Random Variables

Probability theory assigns a central role to random variables. Intuitively, a
random variable uses real numbers to model a speciﬁc aspect of a probabilistic
situation. To illustrate, suppose that the outcome space Ω contains possible
worlds in which an asteroid is hurtling towards Earth. Let X be a function that
carries each possible world to the asteroid’s speed in that world as the asteroid
enters our solar system (where speed is measured using canonical units, such as
meters/sec). So

X ðωÞ ¼ x

means that the asteroid has speed x in world ω as it enters our solar system. X is
a function from Ω (a set of possible worlds) to ℝ (the set of real numbers). More
generally, suppose we have an outcome space Ω. A random variable is
a function that carries each outcome ω to a real number x:

X ðωÞ ¼ x :

A rigorous deﬁnition of “random variable” is given in Section A3, but for

present purposes we may operate at a more intuitive level.
We can use a random variable X to deﬁne various events of interest.
Continuing the asteroid example, take the hypothesis that the asteroid’s speed
falls between a and b. To codify the hypothesis more formally, our ﬁrst step is to
consider the interval ½a; b. See Figure 5. Our second step is to collect together
all the possible worlds mapped by X into that interval. In other words, we
consider the set of possible worlds ω such that a ≤ X ðωÞ ≤ b:

fω : a ≤ X ðωÞ ≤ bg:

3
These locutions are extensionally equivalent, although they have somewhat different connota-
tions. See Fristedt & Gray (1997, p. 12) for helpful discussion.

Figure 5 An interval ½a; b lying in ℝ.

Figure 6 Ω is the rectangle. X 1 ½a; b contains the outcomes mapped by X into

the interval ½a; b.

This set is notated as X 1 ½a; b. It contains those possible worlds where the
asteroid’s speed falls between a and b, so it codifies the hypothesis that
the asteroid’s speed falls between a and b. More generally, given a random
variable X defined on outcome space Ω, X 1 ½a; b codifies the hypothesis that
X’s value falls between a and b. See Figure 6.
As a second illustration, consider the asteroid’s position when it hits
the earth’s surface. We can describe asteroid position using an ordered
pair ðx; yÞ drawn from a canonical coordinate system (e.g. longitude and
latitude). We now want a function X that maps each possible world ω to
an x-coordinate and a second function Y that maps ω to a y-coordinate.
The conjunction

X ðωÞ ¼ x & Y ðωÞ ¼ y

means that the asteroid lands at location ðx; yÞ in possible world ω. Taken
together, X and Y map Ω (a set of possible worlds) into ℝ2 (the set of ordered
pairs of real numbers). We may use X and Y to deﬁne various events of interest.
For example, consider the rectangle depicted in Figure 7. Call this rectangle
REC. We would like to codify the hypothesis that the asteroid lands within
REC. To do so, we collect together all the possible worlds where the asteroid

Figure 7 Rectangle REC contains ordered pairs ðx; yÞ.

lands within REC. In other words, we consider the set of possible worlds ω such
that ðX ðωÞ; Y ðωÞÞ belongs to REC:

fω : ðX ðωÞ; Y ðωÞÞ 2 RECg:

This set contains exactly those possible worlds where the asteroid lands within
REC, so it codiﬁes the hypothesis that the asteroid lands within REC. See
Figure 8.
Random variables are tremendously useful in probability theory. The under-
lying outcome space Ω is often hard to describe or otherwise resistant to direct
mathematical analysis. In particular, it is not easy to deﬁne probabilities directly
over sets of possible worlds. A random variable shifts attention from Ω to
a friendlier outcome space, such as ℝ or ℝ2, greatly augmenting our expressive
and analytic power. I will illustrate in the next section.

2.4 Probability Density

Suppose we take ℝ as the outcome space, so that probabilities attach to sets of
real numbers. ℝ is a natural choice when we are modeling a variable that takes
real numbers as values. For example, if X is a random variable that models
asteroid speed, then the probability assigned to ½a; b is the probability that the
asteroid’s speed falls between a and b.
It is often possible to specify a probability distribution over sets of real numbers
using a probability density function. A probability density function (pdf) is
a nonnegative function over ℝ such that the total area under the curve is 1.
Figure 9 illustrates with a sample pdf pðxÞ. When you see an image like

Figure 8 fω : ðX ðωÞ; Y ðωÞÞ 2 RECg contains the outcomes mapped by

X and Y into REC.

Figure 9 The curve is the pdf. The area under the curve between a and b is the
probability assigned to ½a; b.

Figure 9, it is vital to remember that the numbers on the vertical axis are not
probabilities. They are probability densities. Probabilities are determined by
probability densities as follows: the probability assigned to an interval ½a; b is
the area under pðxÞ stretching from a to b. In this manner, the pdf (a function

from real numbers to probability densities) determines a probability distribu-

tion (a function from sets of real numbers to probabilities).
The most famous example of probability density is the class of Gaussian
distributions, also known as Normal distributions. The pdf for a Gaussian distri-
bution has the familiar shape of a “bell curve.” A Gaussian pdf is completely
described by two parameters: its mean and its variance (a measure of how “spread
out” the curve is from the mean). See Figures 10 and 11. Many variables encoun-
tered in nature are well-described, at least approximately, using a Gaussian pdf.
Bayesian cognitive scientists tend to be cavalier about the distinction
between probability and probability density. My own previous writings have
also treated the distinction quite sloppily. Nevertheless, the distinction is an
important one:

• Probabilities are assigned to sets whose members belong to an outcome space

Ω. Probability densities are assigned to real numbers.
• The probability assigned to an event is at most 1. In contrast, probability
density may be much greater than 1. A pdf can attain very high values, so long
as total area under the curve is 1.

As is customary in the literature, I notate probability using an upper case P and

probability density using a lower case p.
To see the distinction between probability and probability density in action,
consider a probability distribution P with a Gaussian pdf pðxÞ. pðxÞ assigns
densities to individual real numbers. P assigns probabilities to sets of real
numbers: the probability assigned to interval ½a; b is the area under pðxÞ
stretching from a to b. For every real number s, we have

pðsÞ > 0:

Figure 10 A Gaussian pdf with mean m and variance σ2 .

Figure 11 Three Gaussian pdfs. The orange pdf has mean a. The blue and green
pdfs have mean b. The blue pdf has smaller variance than the green pdf. The
orange pdf has intermediate variance.

What about the probability assigned to fsg, that is, the set whose sole member is
s? It is not hard to show that

PðfsgÞ ¼ 0:

Intuitively: the probability assigned to fsg is the area under pðxÞ stretching from
s to s, and that area is simply 0. Thus, the probability density pðsÞ assigned to an
individual point s differs from the probability PðfsgÞ assigned to the event fsg.
Note that, even though each individual event fsg receives probability 0, we
nevertheless have

Pð½a; bÞ > 0

when a 6¼ b. This may at ﬁrst seem surprising, but it does not violate the
probability calculus axioms. The axioms allow each event fsg to receive
probability 0 even while ½a; b receives positive probability.
The notion of pdf generalizes to ℝ2 . In the two-dimensional case,
a probability distribution assigns probabilities to sets containing ordered pairs
ðx; yÞ. For example, suppose we are modeling the asteroid’s position ðx; yÞ when
it hits the earth’s surface. The probability distribution assigns a probability to
each rectangle: this is the probability that the asteroid’s position falls within that
rectangle. In the two-dimensional case, a pdf is a nonnegative function pðx; yÞ

Figure 12 A two-dimensional pdf. The pdf assigns a nonnegative value to each

ordered pair ðx; yÞ. Total volume under the curve is 1.

Figure 13 An alternative depiction of the pdf from Figure 12. Lighter shading
signiﬁes higher probability density assigned to point ðx; yÞ.

over ℝ2 such that the total volume under the curve is 1. The probability assigned
to a region is the volume under the curve in that region. See Figures 12, 13, and
14. A famous example is the class of two-dimensional Gaussian distributions,
which generalize one-dimensional Gaussians to ℝ2 . See Figures 15 and 16.
Once again, it is crucial to distinguish between probability and probability
density. Probability densities attach to ordered pairs (x, y). Probabilities attach
to sets of ordered pairs.

Figure 14 The pdf from Figure 12, restricted to the portion lying over
a rectangle in the ðx; yÞ plane. The volume under this portion of the pdf is the
probability assigned to the rectangle.

Figure 15 A two-dimensional Gaussian pdf.

2.5 Conditional Probability

Conditional probabilities are fundamental to probability theory. Intuitively, the
conditional probability PðAjBÞ is the probability of A given B. For example, we
can consider the probability that Seabiscuit wins the race given that he is sick. In
elementary applications, conditional probability is deﬁned through the ratio
formula:

PðA ∩ BÞ
PðAjBÞ¼df :
PðBÞ

See Figure 17. As Figure 17 illustrates, the unconditional probability of A may

differ significantly from the probability of A given B.
The ratio formula is only well-defined when PðBÞ > 0. Yet scientific practice
frequently requires conditional probabilities when PðBÞ ¼ 0. For example, we
might want conditional probabilities regarding how long an asteroid will take to

Figure 16 An alternative depiction of the pdf from Figure 15. Lighter shading
signiﬁes higher probability density assigned to point ðx; yÞ.

Figure 17 To compute PðAjBÞ using the ratio formula, divide PðA ∩ BÞ by PðBÞ.
For heuristic purposes, assume that the probability assigned to a region is
proportional to the region’s area. Then Figure 17 depicts a case where PðAÞ is
much smaller than PðAjBÞ.

reach Earth given that the asteroid has speed s when it enters the solar system.
Suppose that our probability distribution P over asteroid speed has a pdf p(x).
As indicated in Section 2.4, the probability assigned to the event fsg is 0:

PðfsgÞ ¼ 0:

Thus, we cannot use the ratio formula to deﬁne probabilities conditional

on fsg. As this example illustrates, an adequate treatment must move beyond
the ratio formula, delineating conditional probabilities PðAjBÞ for cases where
PðBÞ ¼ 0.
When P is given by a two-dimensional pdf, a fairly straightforward notion
of conditional probability is available. Consider a two-dimensional pdf pðx; yÞ,
such as in Figure 12 or Figure 15. We can use pðx; yÞ to define a conditional
density pðyjxÞ. Intuitively, pðyjxÞ is a density over y conditional on X having
value x. For each possible value x of the random variable X, the conditional
pdf yields a one-dimensional pdf over y alone. Basically, pðyjxÞ is defined by
holding x fixed in pðx; yÞ while allowing y to vary. The only hitch is that the area
under the resulting curve may not be 1, while the definition of pdf requires the
area under the curve to be 1. Hence, one must also divide by a normalization
constant to ensure that probabilities sum to 1. Figures 18 and 19 illustrate using
the pdf from Figure 12. To compute pðyjaÞ, we hold X fixed at value a while
allowing y to vary. The result is the cross-section curve depicted in Figure 18. To
convert the cross-section curve into a pdf over y, we must divide by
a normalization constant to ensure that area under the curve is 1. The normalized
curve is depicted in Figure 19. Figures 18 and 19 also depict the same procedure

y
x
c
b
a

Figure 18 Conditional densities for the pdf from Figure 12. To compute pðyjaÞ,
we ﬁx X’s value at a and consider the resulting cross-section curve. Area under
the cross-section curve may not be 1. Thus, we divide by a normalization
constant to ensure that area under curve is 1. The result of dividing by the
normalization constant is depicted in Figure 19. Similarly for pðyjbÞ and pðyjcÞ.

b
c

Figure 19 Three one-dimensional pdfs over Y induced by Figure 18. The blue
pdf is pðyjaÞ, the orange pdf is pðyjbÞ, and the green pdf is pðyjcÞ. These three
curves are normalized versions of the three cross-section curves from Figure 18.

y
x
c
b
a

Figure 20 Conditional densities for the Gaussian pdf from Figure 15. This
ﬁgure depicts the unnormalized cross-section curves. To convert the cross-
section curves into pdfs, we divide by a normalization constant. The normalized
curves are depicted in Figure 21.

for two other possible values b and c of X. Figures 20 and 21 depict the same
procedure, this time applied to the pdf from Figure 15. See Section A6 for full
mathematical details.

3 Bayesian Decision Theory

Bayesian decision theory studies an idealized agent who assigns credences to
hypotheses. Bayesians claim that the agent’s credences should conform to the
probability calculus axioms. Thus, the axioms ﬁgure as norms. Bayesians
supplement the probability calculus axioms with two additional norms:
Conditionalization, which governs how credences change in response to new

b
c

Figure 21 The one-dimensional pdfs over Y induced by Figure 20. The blue pdf
is pðyjaÞ, the orange pdf is pðyjbÞ, and the green pdf is pðyjcÞ.

evidence; and expected utility maximization, which governs how credences

guide decision-making. I discuss Conditionalization in Sections 3.1–3.2 and
expected utility maximization in Section 3.3.

3.1 Conditionalization
Credences evolve. If I learn that Seabiscuit is sick, then I should lower my
credence that he will win the race. Intuitively, this is because I have a relatively
low credence that Seabiscuit will win the race given that he is sick. More
generally, suppose that I begin with credence PðHÞ and then learn E. To
conditionalize on E is to replace my former credence PðHÞ with PðHjEÞ. My
old conditional credence PðHjEÞ becomes my new unconditional credence in
H. PðHÞ is called the prior probability and PðHjEÞ is called the posterior
probability. We may write

Pnew ðHÞ ¼ PðHjEÞ;

to signify that my new credence in H is equal to my old conditional credence in

H given E.
The intuitive idea behind the rational norm Conditionalization is that, when I
receive new evidence E, I should form new credences given by

Pnew ðHÞ ¼ PðHjEÞ:

There is considerable variation in how philosophers formulate Conditionalization,

depending partly upon how they gloss “new evidence.” In Rescorla (2021b),
I review some options and give my own preferred formulation. For present
purposes, I remain as neutral as possible among alternative formulations.

However exactly we formulate Conditionalization, it is a diachronic norm: it

governs the evolution of credences over time. In contrast, the probability
calculus axioms are purely synchronic: they govern credences at a moment
of time. Note also that we must sharply distinguish between conditionaliza-
tion the operation and Conditionalization the rational norm. The former is
something an agent does: revise her credences a certain way. The latter is
a rational norm that requires an agent to perform the operation in certain
circumstances.
As with the probability calculus axioms, there is a large literature on why agents
should conform to Conditionalization (Greaves & Wallace, 2006; Lewis, 1999;
Rescorla, 2022; Skyrms, 1987; Weisberg, 2009). Why is someone who conforms
to Conditionalization rationally superior to someone who violates it? Obviously,
the answer may depend on how exactly one formulates Conditionalization. In
what follows, I will simply assume that Conditionalization as formulated some
way is a rational constraint upon credal evolution.
I have focused thus far on Conditionalization in cases where PðEÞ > 0, so that
the ratio formula applies. When PðEÞ ¼ 0, the ratio formula is not well-deﬁned.
An agent who wants to conditionalize in such cases must look beyond the ratio
formula for the needed conditional probabilities. For many applications, the theory
of conditional densities sufﬁces. To illustrate, suppose the agent begins with
credences given by a pdf pðx; yÞ. If she receives evidence that random variable
X has value x, then she can conditionalize using the conditional density pðyjxÞ. Her
new credences over random variable Yare then determined by pðyjxÞ. For example,
suppose the agent begins with credences given by the pdf from Figure 12 and
subsequently learns X’s value. If she learns that X has value a, then conditionaliza-
tion leads her to new credences over Y depicted by the blue curve from Figure 19. If
she instead learns that X has value b, then her new credences over Yare given by the
orange curve. If she learns that X has value c, then her credences over Y are given
by the green curve. In this manner, the theory of conditional densities helps us
generalize Conditionalization beyond cases where PðEÞ > 0.

3.2 Bayes’s Theorem

Bayesian decision theory is so-called because it assigns a central role to
a theorem ﬁrst proved by Bayes. The theorem states that

PðHÞPðEjHÞ
PðHjEÞ ¼ : ð1Þ
PðEÞ

Equation (1) expresses the posterior probability PðHjEÞ in terms of the prior
probability PðHÞ and the prior likelihood PðEjHÞ. The denominator PðEÞ

serves mainly as a normalization constant to ensure that probabilities sum to 1,

so it is common to write the theorem as

PðHjEÞ ¼ k PðHÞPðEjHÞ; ð2Þ

where k ¼ 1=PðEÞ. One can also write the theorem as:

PðHjEÞ ∝ PðHÞPðEjHÞ;

which highlights that the posterior is proportional to the prior times the prior
likelihood:

posterior ∝ prior prior likelihood:

Bayes’s theorem is extraordinarily useful. In many situations, there is a natural prior

probability and a natural prior likelihood. The theorem then tells us how to compute
the posterior from the priors. See Section A7 for a proof of Bayes’s Theorem.
Bayes’s theorem must be sharply distinguished from Conditionalization.
Bayes’s theorem is a direct consequence of the probability calculus axioms and
the ratio formula. As such, it is purely synchronic: it governs the relation between
an agent’s current conditional and unconditional credences. In contrast,
Conditionalization is a diachronic norm. It governs how the agent’s credences at
an earlier time relate to her credences at a later time. Any agent who conforms to
the probability calculus axioms also conforms to Bayes’s theorem, but an agent
who conforms to the probability calculus axioms at each moment may violate
Conditionalization. Thus, one cannot derive Conditionalization from Bayes’s
theorem or from the probability calculus axioms. One must articulate
Conditionalization as an additional constraint upon credal evolution.
That being said, Conditionalization and Bayes’s theorem work together beau-
tifully. An agent who wants to conditionalize can use Bayes’s theorem to compute
the posterior PðHjEÞ and then adopt PðHjEÞ as her new credence in H. Her new
credence in H will be higher to the extent that she already assigned high credence
to H and to the extent that H renders her new evidence E more likely.4
When PðEÞ ¼ 0, (1) is not well-deﬁned because the denominator is 0.
Sometimes, though, a generalized analogue to (1) prevails. When a two-
dimensional pdf pðx; yÞ exists, one can prove:

pðxjyÞ ¼ k pðxÞpðyjxÞ; ð3Þ

4
In the scientiﬁc literature, the phrase “Bayes’s Rule” is used sometimes to denote
Conditionalization, sometimes to denote Bayes’s theorem, and sometime to denote an admixture
of the two.

where k is again a normalization constant. pðxÞ serves as a prior density: it

codiﬁes an agent’s initial credences over random variable X. pðyjxÞ is a density
for random variable Y conditional on X having value x. pðxjyÞ is a density for
X conditional on Y having value y: it serves as a posterior density. We may
rewrite (3) as:

pðxjyÞ ∝ pðxÞpðyjxÞ: ð4Þ

This is the form of Bayes’s theorem most commonly used in scientific applica-
tions, including within Bayesian cognitive science.
We obtain a helpful visualization of (4) by holding y fixed and regarding
pðyjxÞ as a function solely of x. Viewed in this way, pðyjxÞ is called the
likelihood function or sometimes just the likelihood. Intuitively, the likeli-
hood is an initial attempt at forming a probability density over x. The initial
attempt takes into account evidence y but not the prior information encoded
by pðxÞ.5 Bayes’s theorem tells us how to combine the initial attempt pðyjxÞ
with the prior pðxÞ, yielding the posterior density pðxjyÞ. Figures 22 and 23
illustrate. In both figures, the posterior is a compromise between the prior
and the likelihood. In Figure 22, the likelihood is wide, so the posterior
remains fairly close to the prior. In Figure 23, the likelihood is narrow, so
it pulls the posterior far from the prior. For example, suppose that pðyjxÞ is
the conditional density of measuring speed y given that the asteroid has
speed x. Assuming noisy but unbiased measurement, the likelihood peaks
at y. If measurements are very noisy, then the likelihood is wide
(Figure 22), and the prior over asteroid speed exerts more influence on
the posterior. If measurements are less noisy, then the likelihood is narrow
(Figure 23), and the prior exerts less influence.

3.3 Expected Utility Maximization

The ﬁnal key notion of Bayesian decision theory is utility: a numerical
measure of how much an agent desires an outcome. According to
Bayesians, agents should choose actions that maximize expected utility.
The expected utility of action a is a weighted average of utilities assigned
to possible outcomes, where the weights are probabilities contingent upon
performance of a. There are protracted debates about how to formulate
expected utility maximization more rigorously (Steele & Stefánsson,

5
The likelihood is not generally a pdf because the area under the curve need not be 1. However, one
can always normalize and convert the likelihood into a pdf, so we can regard the likelihood as an
“unnormalized” pdf.

Figure 22 The likelihood peaks at the measured value y. The posterior mean is
intermediate between the prior mean and y. Intuitively, the posterior is
a compromise between the likelihood and the prior.

Figure 23 The prior is the same as in Figure 22. The likelihood once again peaks
at y but is narrower. As a result, the posterior is narrower and is pulled closer to
the likelihood. In the asteroid example, a narrower likelihood corresponds to
a case where speed measurements are less noisy. It makes intuitive sense that
less noisy measurements would exert more inﬂuence.

2016). We may leave it at an intuitive level. Scientiﬁc applications often

deal not with utility but instead with cost or loss. The goal is then not to
maximize expected utility but to minimize expected cost. For most purposes,
there is no substantive difference between a utility-based formulation and
a cost-based formulation: one converts a utility function into a loss function
by adding a minus sign, and vice-versa.

Figure 24 A pdf where most probability mass lies far from the mode.

In many statistical applications, the “action” is to estimate the value of

a random variable. The standard procedure is to choose a utility function that
favors selection of the true value and penalizes selection of other values. Often,
selecting a best estimate will amount to selecting the mode of the posterior
density, i.e. the value of x that maximizes the posterior density. There are also
cases where the best estimate differs from the mode. If the utility function
rewards estimates that are close to the true value but distinct from it, then the
best estimate may be quite distant from the mode if enough probability mass lies
away from the mode. Figure 24 illustrates: the mode is located in a region of
relatively small probability mass; an estimator that values being close to the
right answer will choose an estimate from the region of higher probability mass.

3.4 Implementation
Suppose we want a physical system (such as a computer or a robot) to implement
Bayesian inference. Our first task is to decide how the system will encode cre-
dences. A major hurdle is that infinitely many distinct probabilities must often be
encoded. For example, a pdf determines the probability assigned to each interval
½a; b. There are infinitely many such intervals. A finite physical system cannot
explicitly enumerate the credence assigned to each interval. In other words, it
cannot explicitly list each individual probability Pð½a; bÞ. After all, a finite physical
system cannot explicitly list infinitely many distinct pieces of information. When
credences cannot be explicitly enumerated, they must instead be implicitly encoded.
To illustrate implicit encoding, consider the class of Gaussian distributions.
Look again at Figure 10. As noted in Section 2.4, a Gaussian distribution is
completely described by two numbers: its mean and its variance. For that

reason, a physical system can encode a Gaussian distribution by recording its

mean m and its variance σ 2 . This is an example of parametric encoding: the
physical system encodes parameters that determine a probability distribution.
The system does not explicitly enumerate the credence attaching to each
interval ½a; b—that would be impossible. Instead, the system records two
numbers (m and σ 2 ) that determine the credence attached to each interval ½a; b.
Parametric encoding is an option when the probability distribution is ﬁnitely
parametrizable, which is often but not always. A more generally applicable
encoding strategy features sampling. To illustrate, consider a physical system
that draws samples stochastically from the outcome space Ω. There is an
objective chance that the sampled outcome belongs to event A. We may
summarize objective chances through a function:

ChðAÞ;

where ChðAÞ is the objective chance that the physical system draws an outcome
belonging to A. The key idea behind sampling encoding is that these objective
chances can encode subjective probabilities (Icard, 2016). The subjective prob-
ability assigned to A is simply the objective chance that a sample belongs to A:

PðAÞ ¼ ChðAÞ:

The system encodes subjective probabilities via the objective probabilities

governing its sampling activity.
Parametric and sampling encoding are widely used in statistics (Gelman
et al., 2014), machine learning (Murphy, 2023), and other ﬁelds that employ
the Bayesian framework.
The next crucial task is to address computation of the posterior. In some special
cases, it is easy to compute the posterior from the priors. For example, when the
prior probability and the likelihood are Gaussian, the posterior is also Gaussian,
and its mean and variance are easily computable from those of the prior and the
likelihood. Special cases aside, computing the posterior may require resources of
time and memory beyond those available to a realistic agent (Kwisthout et al.,
2011). Look again at Bayes’s theorem (2). Multiplying PðHÞ and PðEjHÞ is easy.
The normalization constant k is another matter. It is possible in principle to
compute k from the prior probability and the prior likelihood, but the computation
requires evaluating a (potentially very long) sum of numbers.6 In practice, it may
be impossible to compute k exactly. A similar point applies to (3). Although k is in

6
Let H1 ; …; Hn be a collection of disjoint events whose union
X is Ω. The law of total probability, a
theorem of the probability calculus, states that PðEÞ ¼ PðEjHn ÞPðHn Þ.
n

principle computable from pðxÞ and pðyjxÞ, the computation may be impossible
in practice.
A computation is tractable when it can be executed by a physical system
with limited time and memory at its disposal. A computation is intractable
when it is not tractable. These definitions can be made mathematically precise,
but the present level of precision suffices for our purposes. The previous
paragraph may be summarized as follows: computation of the posterior is
not always tractable.7
The standard solution in Bayesian statistics is to find tractable algorithms that
approximately implement Bayesian inference. Even when we cannot exactly
compute the posterior, we can often come quite close—close enough for
practical purposes. Even when we cannot conform to the normative ideal
enshrined by Bayesian decision theory, we can often tractably approximate
the normative ideal.
One popular approximation strategy is called Markov chain Monte Carlo
(Murphy, 2023, pp. 493–536). MCMC algorithms use sampling to encode
a credal assignment that approximates the posterior. An MCMC algorithm for
approximating the posterior proceeds in discrete time stages:

t1 ; t2 ; t3 ; . . . ; tn ; . . .

At each stage, a single sample is drawn. Sampling behavior at each stage is

governed by an objective chance distribution. Thus, we have a sequence of
objective chance distributions:

Ch1 ðAÞ; Ch2 ðAÞ; Ch3 ðAÞ; . . . ; Chn ðAÞ; . . .

Ch1 ðAÞ is the objective chance at time t1 of sampling an outcome that belongs to
A. Ch2 ðAÞ is the objective chance at time t2 of sampling an outcome that belongs
to A. Chn ðAÞ is the objective chance at time tn of sampling an outcome that
belongs to A. Objective chances evolve as the algorithm proceeds, converging
asymptotically to the posterior: as the algorithm proceeds, Chn ðAÞ grows ever
closer to the posterior probability assigned to A. After enough time has passed,
the system’s sampling behavior approximates the posterior quite well. See
Figures 25 and 26. There are general convergence results ensuring that, in
a wide range of cases, objective chances fairly quickly approach posterior
probabilities (Brooks et al., 2011).

7
Computational complexity theory studies the distinction between tractable and intractable com-
putation. See van Rooij et al. (2019) for general discussion of computational complexity theory in
relation to cognitive science.

Figure 25 An illustration of MCMC approximation, for the pdf from Figure 12.
The orange dots are samples in the ðx; yÞ plane. Samples cluster in regions of
high probability.

Figure 26 An illustration of MCMC approximation, for the Gaussian

pdf from Figure 15.

4 Bayesian Cognitive Science

Bayesian decision theory studies how agents should reason and make deci-
sions. Over the past few decades, cognitive scientists have increasingly used
the Bayesian framework to describe actual mental activity (usually human,
sometimes nonhuman). The core conjecture is that the mind copes with
uncertainty by allocating credence over a hypothesis space. Credences
evolve in response to sensory input, and they underwrite such tasks as
estimation and decision-making. Credal activity conforms, at least approxi-
mately, to Bayesian norms.
Some Bayesian models posit exact Bayesian inference. Other models posit
tractable approximations to the Bayesian ideal. I will discuss models of both
kinds. I emphasize three domains where the Bayesian research program strikes

me as particularly noteworthy: perception (Section 4.1), motor control

(Section 4.2), and navigation (Section 4.3).8

4.1 Perception
How does the perceptual system estimate distal conditions based upon proximal
sensory input? For example, how does it estimate the shapes, sizes, and loca-
tions of nearby objects based upon retinal stimulations? Proximal sensory
stimulations underdetermine distal conditions: numerous possible distal condi-
tions can cause the same proximal stimulations. Moreover, sensory input is
corrupted by noise during both transduction and transmission to the brain.
Despite ambiguous and noisy sensory input, the perceptual system typically
forms highly accurate estimates of distal conditions.
Helmholtz (1867/1925) proposed that the perceptual system estimates distal
conditions through an unconscious inference. Bayesian perceptual psychology
develops Helmholtz’s proposal, postulating unconscious Bayesian inferences
executed by the perceptual system (Knill & Richards, 1996; Vilares & Kording,
2011; Rescorla, 2015a). A typical Bayesian model estimates a speciﬁc variable
(e.g. shape) based on one or more proximal sensory cues (e.g. shading). The
perceptual system starts with a prior probability over the distal variable and
a prior likelihood that relates the distal variable to proximal sensory input. Upon
receiving sensory input, the perceptual system computes the posterior (or an
approximation to the posterior) over the distal variable. On that basis, the
perceptual system forms a privileged estimate of distal conditions. In most
Bayesian models, the estimate is chosen through expected utility maximization.
In other models, the privileged estimate is chosen not deterministically but
stochastically. For example, the model from (Mamassian, Landy & Maloney,
2002) implements probability matching: estimates are chosen stochastically,
with objective probability matching the posterior.
A simple example of the Bayesian approach concerns perceptual estimation
of shape from shading. As Figure 27 illustrates, shading is an ambiguous cue to
shape. In principle, the stimulus on the left could result from a convex object lit
from overhead or a concave object lit from below. Despite the ambiguity, we
perceive the stimulus on the left as convex and the stimulus on the right as
concave. How does the perceptual system estimate shape based upon the
ambiguous evidence provided by shading? The dominant theory in perceptual
psychology has long been that the perceptual system somehow “assumes” that

8
Readers seeking a more comprehensive overview of the empirical literature might consult
(Grifﬁths, Kemp & Tenenbaum, 2008), (Chater & Oaksford, 2008), or (Ma, Kording &
Goldreich, 2023).

Figure 27 Shading is an ambiguous cue to shape. The stimulus on the left could
result from a convex object lit from overhead or a concave object lit from below.
The perceptual system “assumes” that light comes from overhead, so we
perceive the stimulus on the left as convex and the stimulus on the right as
concave. Reprinted from https://commons.wikimedia.org/wiki/File:%27Light-
from-above%27_prior.jpg, under Creative Commons Attribution-Share Alike
4.0 International license.

light comes from overhead rather than below (Rittenhouse, 1786). This
theory translates naturally into a Bayesian setting. On a Bayesian approach,
the perceptual system estimates shape based on a prior over shapes, a prior
over lighting directions, and a prior likelihood that assigns a probability to
a given shading pattern conditional on the stimulus having a given shape
and the light coming from a given direction (Stone, 2011). The prior over
lighting directions favors overhead lighting directions. Consequently, the
posterior favors the convex interpretation of the left-hand stimulus from
Figure 27.
Bayesian models often posit that, when the perceptual system estimates the
value of distal variable X, the prior over X has a pdf pðxÞ. Models often also posit
that the prior likelihood for sensory variable Y given X has a conditional density
pðyjxÞ. Upon receiving sensory input y, the perceptual system forms new
credences determined by a density pnew ðxÞ. In some models, new credences
are given by the posterior density:

pnew ðxÞ ¼ pðxjyÞ:

In other models, new credences only approximate the posterior:

pnew ðxÞ ≈ pðxjyÞ :

Based on pnew ðxÞ, the perceptual system selects an estimate x* of X’s value.
See Figure 28.
The motion estimation model given by Weiss, Simoncelli, and Adelson
(2002) is a good example of Bayesian perceptual psychology’s explanatory
power. The model estimates the velocity of a moving stimulus. The model posits
a prior density pðvÞ over velocities. Crucially, the prior favors slow speeds. This
reflects the environmental regularity that objects usually move fairly slowly.
The model also posits a likelihood pðIjvÞ, where I measures light intensity over
the retina. Upon receiving input I, the perceptual system computes the posterior
pðvjIÞ and on that basis forms a privileged velocity estimate v*. The model
explains an array of illusions that had previously resisted unified explanation.
For example, it explains why low contrast stimuli seem to move slower than
high contrast stimuli: low contrast stimuli yield a wide likelihood, so the “slow
speed” prior exerts more influence over the posterior. See Figure 29. As this
example illustrates, Bayesian perceptual models can often explain perceptual
phenomena that otherwise elude satisfying explanation.
Subsequent research has further illuminated the “slow speed” prior and its
crucial role in motion perception (e.g. Stocker & Simoncelli, 2006). In
a particularly notable contribution, Kwon, Tadin, and Knill (2015) generalized
the “slow speed” prior to construct a highly successful model of object-tracking.
For further discussion of the motion estimation model, see Rescorla (2015a);
Rescorla (2018b). For further discussion of the object-tracking model, see
Rescorla (2020c).
Another successful application of Bayesian perceptual modeling is cue
combination. The perceptual system typically estimates a single distal variable
based on multiple cues, such as visual and haptic cues to size. Due to sensory

Figure 28 Approximate Bayesian inference in the perceptual system. When

pnew ðxÞ ¼ pðxjyÞ, inference is exact rather than approximate.

Figure 29 Illustrates how the “slow speed” prior influences motion estimation.
When the stimulus has high contrast, the likelihood is narrow and the “slow
speed” prior exerts relatively little influence on the posterior. When the stimulus
has low contrast, the likelihood is wide and the prior exerts relatively more
influence on the posterior. v̂, the posterior mean, is smaller in the low contrast
condition (b) than in the high contrast condition (a). Reprinted with permission
from Springer Nature Customer Service Center GmbH: Springer Nature,
Nature, “Noise Characteristics and Prior Expectations in Human Visual
Perception” (Stocker & Simoncelli, 2006).

noise, estimates based on distinct sensory cues will typically differ at least to
a small degree. The perceptual system must combine distinct sensory cues into
a single unified estimate of the distal variable. Ernst and Banks (2002) showed
that the Bayesian framework can successfully model combination of visual and
haptic cues to size. Researchers have subsequently generalized this finding to
numerous other cases of cue combination within and across modalities
(Trommershäuser, Kording & Landy, 2011). See Rescorla (2020b) for further
discussion of cue combination in a Bayesian setting.
Bayesian perceptual inference is subpersonal and inaccessible to conscious
introspection or control. These inferences are executed by the perceptual
system, not by the perceiver. A typical perceiver is not aware that her perceptual
system uses a “slow speed” prior. The perceptual system, not the perceiver,
encodes and deploys the prior. The perceiver is not consciously aware of any
inference based on the prior.
Perceptual priors are highly mutable, changing rapidly in response to altered
environmental statistics. Adams, Graf, and Ernst (2004) exposed subjects to
deviant visual-haptic input indicating an altered lighting direction. In response,
shape perception and lightness perception rapidly changed, reflecting a change
in the “light from overhead” prior. Similarly, the “slow speed” prior rapidly
changes in response to fast-moving stimuli (Sotiroupolous, Seitz & Seriès, 2011).

There is also evidence that prior likelihoods are mutable (Sato & Kording, 2014;
Sato, Toyoizumi & Aihara, 2007; Seydell, Knill & Trommershäuser, 2010).
Changing priors can themselves be modeled in Bayesian terms (Kwon & Knill,
2013).
As ﬁnal illustration of Bayesian perceptual psychology’s explanatory power,
consider central tendency bias: perceptual estimates of a magnitude are biased
towards the mean of the sample distribution (Hollingworth, 1910). Relatively large
magnitudes tend to be underestimated, while relatively small magnitudes tend to be
overestimated. Depending on the case, the sample distribution may arise naturally
or may be experimentally imposed. Central tendency bias is a ubiquitous effect,
arising when subjects estimate line length (Duffy et al., 2010), interval duration
(Jazayeri & Shadlen, 2010), color (Olkkonen, McCarthy & Allred, 2014), and
many other magnitudes. It is readily explicable from a Bayesian perspective. The
key posit is that the prior adapts to match environmental statistics. For example,
when the subject encounters stimuli drawn from an experimentally imposed sample
distribution, the prior shifts to match that distribution. The shifted prior pulls
estimates towards the prior mean. See Figure 30. Researchers have elaborated
this intuitive idea into models that successfully explain central tendency bias for
a number of perceptual tasks (Glasauer, 2019; Glasauer & Shi, 2022; Petzschner,
Glasauer & Stephan, 2015). The models achieve a close ﬁt with psychophysical
data, including detailed patterns governing the extent to which central tendency bias
occurs in different situations.

Figure 30 Heuristic Bayesian explanation of central tendency bias. Two

possible likelihoods are depicted: the ﬁrst peaks below the prior mean, and
the second peaks above the prior mean. In both cases, the posterior mean is
pulled towards the prior mean. Assuming that the prior is adapted to the sample
distribution, the posterior mean exhibits central tendency bias.

In summary, Bayesian modeling has proved remarkably successful across

a range of perceptual tasks. It amply deserves its orthodox status within
contemporary perceptual psychology.

4.2 Motor Control

Suppose I form an intention to perform an action, such as lifting a coffee cup
without spillage. My motor system must convert my intention into motor
commands that promote fulfillment of my intention. As Bernstein (1967)
emphasized, the motor system has multiple degrees of freedom when
converting intentions into motor commands. For example, there are infinitely
many possible hand trajectories through which I can lift the coffee cup
without spillage. The motor system must select among these infinitely
many options.
Sensorimotor psychology studies how the motor system selects motor com-
mands that promote the agent’s goals. Over the past few decades, Bayesian
models have achieved great explanatory success within sensorimotor psych-
ology (Haith & Krakauer, 2013; Shadmehr & Mussa-Ivaldi, 2012). Optimal
feedback control (OFC) models have proved especially successful (Todorov
2004; Todorov & Jordan, 2002). OFC models have two core elements: an
estimator, which uses conditionalization to estimate current environmental
conditions (including bodily state); and a controller, which uses expected cost
minimization to select suitable motor commands.
When the controller issues a motor command u, it sends an efference copy of
the motor command back to the estimator. The efference copy serves as input to
a forward model (Wolpert & Flanagan, 2009). Intuitively, the forward model
reflects how bodily state will change due to motor commands. More rigorously,
it encodes conditional densities pðxtþ1 jxt ; uÞ, where xt is bodily state at time t,
u is a motor command, and xtþ1 is bodily state at time t þ 1. Using efference
copy and the forward model, the estimator forms an initial probabilistic estimate
of bodily state. Since motor execution is noisy, the initial estimate requires
sensory correction. For example, an initial probabilistic estimate of hand
position can be revised based upon visual and proprioceptive feedback regard-
ing hand position. The estimator sequentially updates credences over environ-
mental conditions based upon sequentially received efference copy and sensory
feedback.
Throughout performance of the motor task, the controller uses updated
credences to compute expected costs of possible motor commands. A cost
function cðh; uÞ reflects the cost of motor command u assuming that outcome
h is the true outcome. During a reaching task, h might specify hand position,

Figure 31 A template for Bayesian models of sensorimotor control. Some

models vary the template somewhat. For example, the Saunders & Knill (2004)
model handles sensory delay by transmitting the initial state estimate rather than
the corrected state estimate to the controller. Modiﬁed from Rescorla (2016)
with permission from John Wiley & Sons.

hand velocity, and the target location. Typically, the cost function has two
components. The first component, which is task-dependent, rewards achieve-
ment of the task goal (e.g. reaching the target). The second component, which is
task-independent, penalizes energetic expenditure. At every stage, the control-
ler selects a motor command that minimizes expected costs. See Figure 31.
OFC models of motor control have achieved great empirical success
(McNamee & Wolpert, 2019). Most notably, OFC explains patterns in
repeated performance of a task. When a subject repeatedly executes a task,
the movement details vary across trials. As Bernstein (1967) first showed,
and as subsequent research has amply confirmed, movement details vary
more along task-irrelevant dimensions than task-relevant dimensions. The
discrepancy between task-relevant variation and task-irrelevant variation is

Figure 32 Subjects reached either to a circle or a rectangle. Unperturbed hand

paths are shown in black. In some trials, hand trajectories were perturbed to the
right. How the motor system responded depended upon the task goal: the
motor system corrected trajectories when reaching for the circle but not when
reaching for the rectangle. In other words, it corrected task-relevant
perturbations but not task-irrelevant perturbations. Reprinted from Scott (2012)
with permission from Elsevier.

one of the most robust ﬁndings in sensorimotor psychology, surfacing in

a huge range of motor tasks. The discrepancy is readily explicable within the
OFC framework (Todorov & Jordan, 2002). Whenever bodily trajectory is
perturbed (e.g. by noise or by an external inﬂuence), the controller must
choose whether to correct the perturbation or leave it uncorrected. Correcting
the perturbation expends energy, so an optimal controller will only correct
perturbations that are task-relevant. As a result, deviations from the average
trajectory accumulate along task-irrelevant dimensions but not task-relevant
dimensions.
An experiment conducted by Nashed, Crevecoeur, and Scott (2012) nicely
illustrates the contrasting response to task-relevant and task-irrelevant perturba-
tions. Subjects reached quickly to a target: either a relatively small circle or else
a relatively wide rectangle. In some trials, an external force disrupted the
reaching motion. When the target was the circle, the external disruption was
task-relevant, so the motor system corrected for it. When the target was the
rectangle, the external disruption was task-irrelevant, so the motor system did
not correct for it. See Figure 32.9
Priors deployed during motor control are mutable (Berniker, Voss & Kording,
2010; Fernandes et al., 2014). Consider a study conducted by Kording and

9
For further discussion of Bayesian sensorimotor psychology, with an emphasis on OFC, see
Rescorla (2016); Rescorla (2019). See also Burge (2022, pp. 502–530).

Wolpert (2004). Subjects reached to a visible target in a virtual reality setup.

Finger position was hidden during the reaching motion, except that subjects
received visual feedback on finger position midway through the motion.
Apparent finger position was shifted from actual finger position, with the shift
drawn randomly from a prior distribution (a Gaussian distribution for some
subjects, a bimodal distribution for other subjects). The motor system learned
the experimentally imposed prior (either the Gaussian prior or the bimodal
prior) and used it to adjust finger trajectories based on visual feedback.

4.3 Navigation
Animal navigation has been intensively studied for many decades across several
disciplines, including psychology, ethology, and neuroscience. At present,
Bayesian modeling does not figure as prominently in the study of navigation
as it does in perceptual psychology and sensorimotor psychology. Nevertheless,
recent studies provide strong evidence that Bayesian inference plays a crucial
role in human navigation.
I focus on a navigational strategy called dead reckoning. During dead reckon-
ing, the navigator exploits self-motion cues to maintain a running estimate of her
own position. Self-motion cues include optic flow, efference copy, vestibular
signals, and so on. Dead reckoning is sometimes called “path integration,”
because position is the integral of velocity. Dead reckoning pervades the animal
kingdom (Gallistel, 1990, pp. 57–102), from the desert ant to humans.
A key fact about human dead reckoning is that, in many experimental condi-
tions, subjects overshoot the target destination. Traditionally, overshooting was
explained through a “leaky integrator” model (Lappe et al., 2011). The basic idea
is that subjects imperfectly integrate velocity to compute position: rather than
computing the true integral, subjects compute a slightly smaller quantity. As the
distance traveled increases, “leaks” accumulate and the discrepancy between
estimated position and true position increases. Lakshminarasimhan et al. (2018)
offer an alternative Bayesian explanation. They posit a “slow speed” prior over
self-motion. The “slow speed” prior biases estimated velocity below the true
velocity, which leads the subject to underestimate distance traveled. See
Figure 33.
The “slow speed” model explains several phenomena that the “leaky integra-
tor” model does not. For example, Lakshminarasimhan et al. (2018) studied
dead reckoning in a virtual reality setup. They manipulated the optic flow cue by
altering the density of plane elements: greater density entails a more reliable
cue. Decreased cue reliability corresponds to a relatively wide likelihood. The
“slow speed” model predicts that, when the likelihood is wide, the posterior will

Figure 33 Comparison of the “slow speed” prior model and a “leaky integrator”
model. (For heuristic purposes, the comparison only depicts one-dimensional
linear velocity. The actual model also considers angular velocity.) The panel on
the left shows the subject’s true velocity over time. The top row schematizes the
“slow speed” prior model. At a given moment, the “slow speed” prior (in green)
combines with the likelihood to yield a posterior over possible velocities. The
resulting velocity estimates are consistently smaller than actual velocity, due to
the inﬂuence of the “slow speed” prior. When velocity estimates are integrated
to form position estimates, the position estimates are biased. The bottom panel
shows a Bayesian version of the “leaky integrator” approach. (One can also
develop the “leaky integrator” approach in a non-Bayesian setting.) The prior is
uniform. As a result, velocity estimates are not biased towards smaller
velocities. The bias in position estimates stems from leaky integration, not from
biased velocity estimates. Reprinted from (Lakshminarasimhan et al., 2018)
with permission from Elsevier.

be more strongly affected by the “slow speed” prior, causing even more
overshooting. In contrast, the “leaky integrator” model does not predict that
a degraded optic flow cue causes increased overshooting. See Figure 34. The
human data exhibited more overshooting in response to the degraded optic flow
cue, conforming closely to the “slow speed” model’s predictions.
Another striking phenomenon explained by the model: when the target is
relatively distant, overshooting gives way to undershooting. The farther the
subject travels, the greater the uncertainty regarding her position, so the wider
her pdf over possible positions. When the pdf becomes quite wide, its area of
overlap with the target decreases. As a result, expected utility peaks before the
target when the target is relatively far away. For sufficiently large distances, this
bias towards undershooting swamps the bias induced by the “slow speed” prior.
The Bayesian model, by analyzing how these two biases interact with each other
and with optic flow reliability, achieves a good match with actual human
performance.

Figure 34 Differing predictions of the “slow speed” model and the “leaky
integrator” model. The red likelihood function is narrow, corresponding to
a reliable optic flow cue. The blue likelihood function is wide, corresponding to
a degraded optic flow cue. In the “slow speed” prior model, the prior exerts
more influence in degraded cases, causing a more biased position estimate. The
“leaky integrator” model predicts that a change in cue reliability does not affect
the position estimate. The “leaky integrator” approach is illustrated here in
a Bayesian format, but the same prediction prevails for non-Bayesian versions.
Reprinted from (Lakshminarasimhan et al., 2018) with
permission from Elsevier.

Central tendency bias provides additional evidence for a Bayesian approach

to dead reckoning. Petszchner and Glasauer (2011) studied a virtual reality task
in which subjects traversed an experimentally imposed linear path and then tried
to reproduce their displacement. Subjects performed the task in multiple trials
during each session. Distances during a session were drawn from one of
three distinct sample distributions: small, medium, or large. Subjects exhibited
signiﬁcant central tendency bias: their distance estimates during a session (as
gauged by reproduced distance) were biased towards the mean of the session
sample distribution. To explain the bias, Petszchner and Glasauer (2011) offer
an iterative Bayesian model. After each trial, the Bayesian estimator updates its
prior over distance traveled. The prior gravitates towards the mean of the
session sample distribution, biasing distance estimates towards that mean.
The model thereby explains observed central tendency bias, further conﬁrming
the hypothesis that human dead reckoning relies upon Bayesian estimation.

Dead reckoning is only one navigation strategy found in the animal kingdom.
Equally important is piloting, during which the creature uses landmarks to
estimate its own position (Gallistel, 1990, p. 41, pp. 88–93, pp. 120–123).
Even relatively primitive creatures, such as rats and bats, engage in piloting.
Of course, humans routinely do so. There is strong evidence that human piloting
relies on Bayesian inference (Jetzschke et al., 2017), as does human combin-
ation of self-motion cues and landmark cues (Chen et al., 2017).
Dead reckoning and piloting are key to navigation, but they are just the
beginning. Piloting presupposes mapping: estimation of landmark locations.
Mapping also figures prominently in robotics, where the standard solution centers
upon approximate Bayesian inference (Thrun, Burgard & Fox, 2005). Several
researchers have conjectured that some mammals likewise implement Bayesian
mapping (Gallistel, 2008; Rescorla, 2009). The conjecture fits well with every-
thing we know about mammalian navigation (Savelli & Knierim, 2019;
Shikauchi et al., 2021). Moreover, it can explain within a single theoretical
framework disparate navigational phenomena that otherwise resist unified
explanation (Kessler, Frankenstein & Rothkopf, 2024). The topic merits, and
will surely receive, further investigation.

4.4 Other Psychological Domains

Researchers have applied the Bayesian perspective to numerous domains, such
as causal reasoning (Griffiths & Tenenbaum, 2009; Oaksford & Chater, 2020),
social cognition (Baker & Tenenbaum, 2014), intuitive physics (Battaglia,
Hamrock & Tenenbaum, 2013; Sanborn, Masinghka & Griffiths, 2013), lan-
guage acquisition (Abend et al., 2017), syntactic parsing (Narayanan &
Jurafsky, 1998), concept acquisition (Goodman et al., 2008), music cognition
(Temperley, 2007), reading (Norris, 2006), memory (Hemmer & Steyvers,
2009), categorization (Sanborn, Griffiths & Navarro, 2010), and so on.
Applications vary in their predictive and explanatory power. Few achieve the
astonishing explanatory successes found in perception and motor control. Still,
they are often more successful than competing non-Bayesian approaches, as
readers can confirm for themselves by accessing the above-cited texts.

4.5 Anti-Bayesian Phenomena?

Like every prominent cognitive science research program, Bayesian modeling
has attracted lots of criticism (Bowers & Davis, 2012; Eberhardt & Danks,
2011; Jones & Love, 2011; Mandelbaum, 2019). Perhaps the most basic criti-
cism is that many mental phenomena appear radically anti-Bayesian. This
criticism traces back to Kahneman and Tversky, who discovered intriguing

cognitive phenomena that apparently violate Bayesian norms (e.g. Kahneman

& Tversky, 1979; Tversky & Kahneman, 1983). A good example is anchoring
bias (Tversky & Kahneman, 1974): when asked to estimate a quantity (such as
the distance between Los Angeles and San Francisco), people are biased
towards a randomly selected number provided to them. In effect, the randomly
selected number serves as an “anchor” that pulls judgment away from a more
accurate estimate. Anchoring bias is irrational and hence suggests that people
violate the norms of Bayesian decision theory. Beyond the cognitive-level
irrationalities discovered by Kahneman and Tversky, researchers have docu-
mented seemingly anti-Bayesian phenomena in other domains, including per-
ception (Gardner, 2019; Mandelbaum et al., 2020; Rahnev & Denison, 2018).
Proponents of Bayesian modeling reply that many apparently anti-Bayesian
phenomena can in fact be modeled in Bayesian terms (Stocker, 2018). Consider
the size‒weight illusion: when you lift two objects of equal weight but different
size, the smaller object feels heavier. At ﬁrst, the illusion looks anti-Bayesian
because it ﬂouts a prior expectation that larger objects are heavier. However, the
illusion turns out to be explicable by a Bayesian model that estimates relative
densities (Peters et al., 2016).
Even when a phenomenon cannot be modeled in Bayesian terms, it can often
be modeled in terms of approximately Bayesian inference (Chater et al., 2020).
In this spirit, Lieder et al. (2018) show that anchoring bias arises naturally from
a sampling approximation to idealized Bayesian inference. They assume that,
when subjects are provided with a randomly selected number, this number
serves as the initial sample for an MCMC algorithm. (See Section 3.4 to review
MCMC algorithms.) Samples are biased towards the initial sample, which may
be quite far from an optimal Bayesian estimate. As the algorithm proceeds, it
draws samples closer to the optimal Bayesian estimate. The extent of anchoring
bias depends upon how long the algorithm runs (i.e. how many samples it
draws). When computation is costly (e.g. because computational resources are
needed for another task), anchoring bias increases because the system draws
fewer samples and does not get as far from the initial sample. On this approach,
anchoring bias arises from “rational” use of limited computational resources:
the system balances accurate estimation against the cost of computation. The
sampling model explains a range of effects, such as increased anchoring bias
due to cognitive load or time pressure. Many other seemingly anti-Bayesian
cognitive phenomena can be similarly explained in terms of sampling approxi-
mation to idealized Bayesian inference (Chater et al., 2020; Dasgupta, Schulz &
Gershman, 2017).
Obviously, there is no guarantee that all psychological processes will turn out
to be Bayesian or approximately Bayesian. Sub-systems may conform to

Bayesian norms to a greater or lesser degree. For example, it could be that

perception conforms quite closely to Bayesian norms while high-level decision-
making does not. It could be that certain perceptual processes conform closely
to Bayesian norms while other perceptual processes do not, or that certain
perceptual processes conform to Bayesian norms under certain circumstances
but not other circumstances—e.g. that perceptual processes conform to
Bayesian norms only when the perceiver is paying attention (Morales et al.,
2015). These possibilities require investigation. We must build and test detailed
Bayesian models of specific phenomena, evaluating afresh how well each
model fits the data. That is exactly what Bayesian cognitive scientists do on
a daily basis. Underlying this research program is a key methodological
commitment: enough mental processes are at least approximately Bayesian
that constructing and testing Bayesian models of specific mental processes is
a worthwhile endeavor. So far, the methodological commitment has been amply
vindicated.10

4.6 Where do the Priors Come From?

A natural question posed by Bayesian modeling is how the prior probability and
prior likelihood arise. For example, the Bayesian dead reckoning model
assumes a “slow speed” prior but says nothing about the prior’s etiology.
A similar point applies to other Bayesian models found in the literature. The
models postulate priors that underlie Bayesian inference, without explaining
how the priors arise. Given that priors are highly mutable, a good explanation
will surely cite a complex mixture of evolutionary and developmental factors.
So far, though, no such explanation is available.
Some critics complain that Bayesian models are unexplanatory due to their
reliance on postulated priors (Hutto & Myin, 2017, pp. 67–74, pp. 154–155;
Orlandi, 2014, p. 91). The worry is that Bayesian models rest upon unexplained
explainers. How much can a Bayesian model explain when it posits priors but
offers no explanation for the priors?

10
Many supposedly anti-Bayesian phenomena documented by Kahneman and Tversky (such as
the conjunction fallacy) involve explicit probability judgments: researchers ask subjects to judge
relative probabilities of various possibilities; elicited judgments violate the probability calculus
axioms. Poor performance in an explicit probabilistic task is hardly evidence that the subject
does not execute Bayesian inference, any more than poor performance in a symbolic logic class
is evidence that a student does not execute deductive inference. Bayesian cognitive science does
not claim that ordinary people are good at probability theory. It claims that ordinary people (or
their psychological subsystems) assign subjective probabilities and execute Bayesian operations
over the assigned probabilities. Typically, the probability assignments are not explicit but are
instead implicitly encoded.

In my opinion, this complaint has no force. We must distinguish between

incomplete theories and unexplanatory theories. Every scientific theory con-
tains unexplained explainers: postulates that serve as explanantia. For example,
Newtonian physics postulates that objects have mass, but it does not explain
how objects come to have the masses that they have. No one should complain on
that basis that Newtonian physics is unexplanatory. In many cases, a successful
scientific theory contains huge explanatory gaps. A famous example is the
theory of natural selection as formulated in On The Origin of Species. Darwin
postulated suitable hereditary mechanisms but had no clue what those mechan-
isms were. No one should complain on that basis that the theory of natural
selection as formulated by Darwin was unexplanatory. A scientific theory can
offer powerful explanations even though it includes unexplained explainers.
Of course, it is always good to eliminate unexplained explainers. Modern
biology achieved a decisive advance when it discovered the genetic basis of
heredity. Bayesian cognitive science will likewise achieve a major advance
when it illuminates the etiology of priors. Until that advance, Bayesian cogni-
tive science will be incomplete in a significant way. Even in its present incom-
plete state, it offers powerful explanations for many psychological phenomena.

5 Realism and Instrumentalism

Any Bayesian model posits credal states (assignments of credences to hypotheses)
and credal transitions (transitions among credal states). At a bare minimum, a
Bayesian model posits a prior probability, a prior likelihood, and a transition to
a posterior or approximate posterior. The Bayesian model may posit a succession
of credal states, as in sensorimotor psychology.
Suppose that a Bayesian model is explanatorily successful, in the sense that it
supplies compelling explanations for observed phenomena. Let us distinguish two
opposing viewpoints one might adopt towards the model: realism and instrumen-
talism. Realists hold that we have good reason to deem the model an approximately
true description of mental activity (Rescorla, 2020c). From a realist perspective, the
mind instantiates credal states and transitions at least roughly like those posited by
the model. The model describes actual mental states and processes that mediate
between inputs (e.g. retinal inputs) and outputs (e.g. perceptual estimates; motor
commands). Instrumentalists, on the other hand, regard the model as nothing but
a useful predictive device (Block, 2018; Colombo & Seriès, 2012; Orlandi, 2014).
The model helps us summarize the mapping (possibly stochastic) from inputs to
outputs, but it does not describe actual mental states and processes with even
approximate accuracy. From an instrumentalist perspective, we have no reason to
believe that the mind instantiates credal states or that it executes anything

resembling (approximate) Bayesian inference. We should conclude only that the

mind operates as if it executes (approximate) Bayesian inference. Whereas realists
attribute psychological reality to credal states and transitions postulated by explana-
torily successful Bayesian models, instrumentalists do not.
To illustrate how realism and instrumentalism differ, consider Figure 28. The
ﬁgure depicts how the perceptual system converts proximal sensory input y into
a perceptual estimate x*. It posits two mental states (the prior probability and the
prior likelihood) that along with y cause a third mental state (the approximate
posterior), which in turn causes perceptual estimate x*. From a realist perspective,
we should take this causal structure seriously as a guide to underlying psycho-
logical reality. There really do exist priors, they really do interact with input y to
cause an approximate posterior, and this approximate posterior really does cause
a perceptual estimate x*. In contrast, instrumentalists do not take Figure 28 as
a guide to underlying psychological reality. All that we should take seriously
about Figure 28, they say, is the induced mapping from input y to estimate x*.
I have defended realism at length in previous writings (Rescorla, 2015a;
Rescorla, 2015b; Rescorla, 2020c). Here, I will brieﬂy adduce a few consider-
ations in its favor.

5.1 Scientiﬁc Realism

My realist perspective on Bayesian cognitive science is grounded in a general
commitment to scientific realism. Scientific realism traces back to Putnam
(1975) and has been elaborated by many subsequent philosophers. The basic
idea is that explanatory success is a prima facie indication of approximate truth.
When a scientific theory is explanatorily successful, we have prima facie reason
to believe that it is at least approximately true. For example, the explanatory
success of modern physics provides reason to believe in subatomic particles.
My realist perspective on Bayesian cognitive science results from straight-
forward application of scientific realism to Bayesian modeling. Many Bayesian
models, although not all, are explanatorily successful. From a scientific realist
perspective, we have reason to regard these models as at least approximately
true. We have reason to accept that there exist credal states and transitions
roughly like those posited by the model. Just as the explanatory success of
modern physics provides reason to believe in subatomic particles, the explana-
tory success of a Bayesian model provides reason to believe in credal states and
transitions.
Not all philosophers accept scientific realism. Some authors favor an instru-
mentalist perspective on scientific theorizing (van Fraassen, 1980). According
to instrumentalism, a scientific theory is just a useful tool for making

predictions. When a scientiﬁc theory is explanatorily successful, we have no

reason to believe that the theory is even approximately true. For example, the
explanatory success of modern physics provides no reason to believe that there
are subatomic particles. Philosophers who favor an instrumentalist perspective
more generally will surely want to apply it specifically to Bayesian cognitive
science. If you do not believe in a subatomic particles, then you probably do not
believe in credal states and transitions!
Typically, researchers who favor instrumentalism about Bayesian cognitive
science do not evince a more general commitment to instrumentalism about
scientific theorizing. They do not hold that we should be instrumentalists about
scientific theories in general. Instead, they argue that we should be instrumen-
talists for the special case of Bayesian cognitive science. In my opinion, their
arguments for that differential stance have little force. I see no compelling
reason why philosophers inclined towards scientific realism in general should
favor instrumentalism for the special case of Bayesian modeling. I will illustrate
my viewpoint by critiquing an instrumentalist argument advanced by Block
(2023) and tailored to the special case of Bayesian cognitive science.

5.2 Simulation or Implementation?

According to Block, we seldom if ever have reason to believe that a psychological
system implements approximate Bayesian inference as opposed to merely simu-
lating approximate Bayesian inference. To support his assessment, Block cites
evolutionary considerations (2023, p. 208):

Evolution is a pro-instrumentalist mechanism. There is no doubt that behav-

ing according to Bayesian norms is enormously valuable for an organism and
we can expect strong evolutionary pressure toward behavior that ﬁts the
norms of Bayesian rationality. But Bayesian rational behavior does not
have to be implemented using the conceptual apparatus that is best suited to
describing Bayesian rational processes by the theorist. The problem with
Rescorla’s argument is that it is not clear that the way evolution chose to
produce behavior that adheres roughly to Bayesian norms involves the
representation of probabilities in the perceptual system.

Block develops his position by citing an experiment on pea plants conducted by

Dener, Kacelnik, and Shemesh (2016). Each plant’s roots were divided between
two pots. The two pots received equal mean levels of nutrients. Nutrient levels
were constant in one pot and variable in the other. More roots developed in the
constant pot if the mean nutrient level was high, and more roots developed in the
variable pot if the mean nutrient level was low. This growth pattern comports
with expected utility theory, which (assuming a suitably shaped utility function)

mandates risk aversion in rich conditions and risk proneness in poor conditions.
Block writes (2023, pp. 209–210):

[T]he pea plant behaves as if it represented mean levels of nutrients and their
degree of uncertainty. Since the pea plant lacks a nervous system, we can be
pretty sure that there are no such representations. Somehow, natural selection
has found a way for plants to behave according to some of the norms of
Bayesian rationality without those representations. The challenge to
Rescorla’s reasoning is that we have to allow for the possibility that the
same is true of our perceptual systems.

Block concludes that, even if we favor a realist stance towards scientiﬁc

theorizing in general, we should adopt an instrumentalist stance towards
Bayesian perceptual psychology. Presumably he would extend the conclusion
to other branches of Bayesian cognitive science.
In evaluating Block’s argument, we must carefully distinguish between
subjective and objective probability. Due to the experimental protocols, there
are objective probabilities that govern the nutrient level in each pot. We might
gloss these either as frequencies or as chances. Either way, they are objective
features of the world, lacking any subjective element. Given a pot’s objective
probability distribution, we can describe the mean and the variance. The
experiment shows that root growth is sensitive to both the mean and the
variance. Thus, root growth is sensitive to objective probabilities (or to proper-
ties that supervene upon objective probabilities).
One might try to explain that sensitivity by attributing credal states to the
pea plants. One might posit subjective probabilities, instantiated by each
plant, that track the objective probabilities governing each pot. One might
postulate that root growth is inﬂuenced by an expected utility computation
based upon the posited subjective probabilities. I agree with Block that the
proposed explanation is both implausible and unmotivated. It is implausible
because plant physiology does not seem able to support expected utility
computations. It is unmotivated because nothing about the pea plant study
indicates that credal states or utility functions mediate the causal inﬂuence
of objective probabilities upon root growth. The mere fact that a system is
sensitive to the mean and variance of an objective probability distribution
does not suggest that the system instantiates credal states. For example, we
can construct a machine whose outputs are sensitive to the frequency with
which a biased coin lands heads; there is no reason why the machine must
instantiate credal states. Mere sensitivity to objective probabilities (or
properties that supervene on objective probabilities) is not a prima facie
indicator of Bayesian computation. This remains so even when the mapping

from objective probabilities to outputs happens to mirror the dictates of

expected utility theory.
One can describe any system using Bayesian decision theory. Adapting an
example of Dennett’s (1987, p. 23), one can “explain” why a lectern does not
move by saying that the lectern assigns high utility to occupying the optimal
location in the universe and assigns high credence to the hypothesis that it
currently occupies the optimal location in the universe. Clearly, though, we
should not accept the purported “explanation.” It contributes no value to our
theorizing. It does not improve upon a non-Bayesian explanation couched
wholly in terms of physics (e.g. an explanation that cites the law of inertia).
A similar diagnosis applies to the pea plant study. The two cases are not totally
analogous, because we know why the lectern does not move (inertia) but do not
yet know the physiological mechanisms through which objective probabilities
causally influence root growth. Still, as Dener, Kacelnik, and Shemesh (2016)
themselves emphasize, nothing about the pea plant study suggests that the
mechanisms involve credal states or utility functions. We have no reason to
think that Bayesian modeling would add any explanatory force to an eventual
physiological explanation couched in non-Bayesian terms.
A very different diagnosis applies to numerous Bayesian models offered in
cognitive science, including but not limited to perceptual psychology. In many
cases, the Bayesian model adds considerable explanatory value to our theorizing.
For example, the (Lakshminarasimhan et al., 2018) dead reckoning model explains
why more overshooting occurs when the optic flow cue is degraded, and it also
explains why undershooting occurs for relatively distant targets. The model thereby
achieves the unity characteristic of good explanation. Similarly, the (Weiss,
Simoncelli & Adelson, 2002) motion estimation model offers a unified explanation
for diverse motion illusions. In these cases, and in many others, the Bayesian model
makes a substantial explanatory contribution that looks otherwise unachievable.
The explanatory contribution includes qualitative predictions for disparate phe-
nomenon coupled with quantitative predictions that closely match experimental
data. From a scientific realist perspective, these explanatory achievements provide
reason to believe that the model is approximately true.
The contrast with the pea plant study is glaring. Dener, Kacelnik, and Shemesh
(2016) do not provide a Bayesian model of root growth. Indeed, they do not so
much as hint which prior probability or prior likelihood such a model might
include. They do not suggest, let alone argue, that a Bayesian model could offer
a unified explanation for disparate phenomena or that it could yield quantitative
predictions that closely match experimental data. There is not even a Bayesian
model here that we can evaluate, much less a model that achieves anything
approaching the explanatory success of the Bayesian dead reckoning model, the

Bayesian motion estimation model, or numerous other Bayesian cognitive sci-

ence models. That is why Bayesian cognitive science supports the existence of
credal states but the pea plant study, for all its interest, does not.
I agree with Block that a system can simulate (approximate) Bayesian infer-
ence. For example, a system might convert inputs into outputs by consulting
a look-up table. More realistically, a system might acquire an input‒output
mapping through reinforcement learning. In reinforcement learning, the system
receives rewards for how it responds to inputs, and it adjusts its responses to
obtain optimal or near-optimal rewards. Systems trained through reinforcement
learning can mimic certain kinds of approximate Bayesian inference
(Weisswange et al., 2011). In principle, then, evolution might produce a system
that operates as if it executes approximate Bayesian inference even though it does
not actually execute approximate Bayesian inference. Nevertheless, I think it
misleading to describe evolution as a “pro-instrumentalist mechanism.” I see no
reason why evolution should favor simulation of approximate Bayesian inference
over implementation of approximate Bayesian inference.
When a scientiﬁc theory accurately predicts the behavior of a system, there is
always a possibility that the theory is utterly false and that the system merely
behaves as if the theory is true. For example, it is in principle possible that
subatomic particles do not exist and that the physical universe merely behaves as
if they exist. Physicists would only regard that in principle possibility as worth
taking seriously if it were developed into a rival theory that matched modern
physics in explanatory power. Similarly, we should only take seriously the sugges-
tion that mental activity simulates rather than implements approximate Bayesian
inference once it is developed into detailed models that rival current Bayesian
models in explanatory power. So far, that has not happened. The scientiﬁc literature
does not offer non-Bayesian models comparable in explanatory power to the
Bayesian dead reckoning model, the Bayesian motion estimation model, or numer-
ous other Bayesian models found in contemporary cognitive science. Perhaps
impressive non-Bayesian models will eventually emerge. In their absence, the
mere possibility that they might emerge should not worry realists about Bayesian
cognitive science any more than the mere possibility of a successful physical theory
that eschews subatomic particles should worry realists about subatomic particles.

5.3 The Argument From Altered Priors

I now rehearse an additional argument for realism regarding Bayesian cognitive
science. My argument rests upon a crucial fact: input‒output mappings rapidly
change in response to changing environmental conditions. We have seen several
examples:

• Shape and lightness perception change in response to stimuli that indicate

a deviant lighting direction (Adams, Graf & Ernst, 2004).
• Motion perception changes in response to fast-moving stimuli (Sotiropoulos,
Seitz & Seriès, 2011).
• The mapping from sensory inputs to motor commands changes in response to
shifts in apparent ﬁnger position (Kording & Wolpert, 2004).
• Central tendency bias occurs in a wide range of domains, including percep-
tual estimation (Section 4.1) and dead reckoning (Section 4.3).

These experimental phenomena, and numerous others, conclusively demon-

strate that the mapping from inputs to outputs is highly mutable.
Realists can easily explain in each case why the mapping changes as it
does. They can say that the priors change so as to match changing environ-
mental statistics. For example, suppose that a subject exhibits central ten-
dency bias towards the mean of an experimentally imposed sample
distribution, as in the (Petzschner & Glasauer, 2011) dead reckoning experi-
ment. Realists explain the bias as follows: the prior shifts to match the
sample distribution, which causes estimates to shift towards the mean of
the distribution. Instrumentalists can acknowledge that the input‒output
mapping shifts, but they offer no principled explanation for why it shifts
as it does. From an instrumentalist perspective, there is no principled reason
why estimates should shift to match recent stimuli. The mere fact that
a system simulates approximate Bayesian inference using certain priors
provides no reason to expect that the system will change any particular
way in response to changing environmental statistics. Hence, realism offers
a major explanatory advantage over instrumentalism.
Call this the argument from altered priors. Although I have formulated the
argument as applied to prior probabilities, similar argumentation applies to prior
likelihoods and to posteriors (Rescorla, 2020c).
Block rejects the argument from altered priors: “I ﬁnd this argument uncon-
vincing because whatever it is about the computations of a system that simulates
the effect of represented priors . . . might also be able to simulate the effect of
change of priors” (2018, p. 8).
I agree that, in principle, a system that simulates approximate Bayesian
inference given certain priors might respond to changing environmental condi-
tions by simulating approximate Bayesian inference given another set of priors.
I question whether instrumentalists can develop that possibility into compelling
models. The argument from altered priors is abductive: realism provides the
best explanation for why input‒output mappings change as they do. One does
not undermine an abductive argument by noting that an alternative explanation

may emerge. To undermine the argument from altered priors, one must provide
a speciﬁc alternative explanation and show that it is at least as satisfying as the
realist explanation.
In this connection, consider a system trained through reinforcement learning
to simulate Bayesian inference given certain priors. By varying the rewards, we
can train the system to simulate Bayesian inference given another set of priors.
Accordingly, instrumentalists might hope that reinforcement learning can
explain changes to the input‒output mapping. In many cases, though, subjects
receive either no feedback or extremely limited feedback on their performance.
To illustrate, consider the (Petzschner & Glasauer, 2011) dead reckoning study.
Participants received no feedback on their performance during each session,
aside from a few initial training trials to ensure familiarity with the virtual
reality setup. How, then, can reinforcement learning explain why subjects
displayed central tendency bias? There was no “reward” to drive the ongoing
change in learned responses. This study provides evidence that subjects itera-
tively update a distance prior in response to accumulated evidence.
Perhaps instrumentalist theories will eventually emerge that explain chan-
ging input‒output mappings without an appeal to changing priors. We would
then need to compare those instrumentalist theories with realist Bayesian
theories. Until that time, we do well to develop the realist perspective and see
where it leads.11

5.4 Neural Implementation

To gain more insight into the dialectic between realism and instrumentalism, let
us consider the neural implementation of approximate Bayesian inference. How
are credal states physically realized in the brain? Which neural operations
implement computation of the (approximate) posterior from the priors? These
questions do not arise for instrumentalists because instrumentalists do not
regard credal states and transitions as psychologically real. For realists, the
questions are pressing.
Computational neuroscientists have proposed several theories of how the
brain might implement credal states and transitions (Fiser et al., 2010; Pouget
et al., 2013; Rescorla, 2024). The proposed theories are biologically plausible
and ﬁt well with what we know about the brain, although no single theory has
yet emerged as well-conﬁrmed.
The credal states considered in Bayesian cognitive science are usually given
by pdfs. Recall that a pdf determines probabilities assigned to intervals ½a; b.

11
See Rescorla (2020c) for more on the argument from altered priors and for general defense of
realism.

There are inﬁnitely many of these intervals. The brain is a ﬁnite physical system
and hence, as discussed in Section 3.4, cannot explicitly list each individual
probability Pð½a; bÞ. Since the brain cannot enumerate the probability assigned
to each interval, probabilities must be implicitly encoded by neural activity. The
two main implicit encoding schemes under active consideration were men-
tioned in Section 3.4:

• Parametric encoding: the brain encodes parameters for the pdf. One possibility
is that parameters are encoded by spike counts in a neural population (Ma et al.,
2006). Each neuron is associated with a preferred stimulus value, and each
neuron’s spike count is interpreted as the strength of its “vote” for that stimulus
value. “Votes” across the neural population determine parameters of a pdf, e.g.,
the mean and variance of a Gaussian. See Figures 35 and 36.
• Sampling encoding: the brain encodes a probability distribution via sampling
propensities. For example, a neuron’s membrane potential might encode a sample
(Orbán et al., 2016). The objective chance distribution governing membrane
potentials encodes the subjective probability distribution for the variable.

Computational neuroscientists have produced detailed neural network models

that enshrine these encoding schemes. The models show how, in principle,

Figure 35 The tuning curve for a neuron summarizes the neuron’s average
response to a stimulus value. Figure 35 depicts tuning curves for a hypothetical
neural population tuned to a one-dimensional continuous distal stimulus. The
horizontal axis contains possible stimulus values. Each tuning curve depicts the
average response (measured in spikes per second) of the corresponding neuron
to possible stimulus values. Each tuning curve peaks at a preferred value of the
stimulus. The black tuning curve has preferred stimulus value a.

Figure 36 Heuristic illustration of how a neural population can implicitly

encode parameters of a probability distribution. The top left panel depicts how
a hypothetical neural population responds to a stimulus on a given occasion.
The horizontal axis groups neurons according to preferred stimulus value. The
vertical axis gives the spike count for each neuron during a ﬁxed time interval.
Spike counts encode a Gaussian distribution, depicted in the top right panel. The
parametric encoding scheme used here is discussed in Beck et al. (2007): the
mean of the Gaussian is a weighted average of stimulus values, where the
weights are the individual spike counts; the variance is inversely proportional to
the total spike count. The bottom left panel depicts a different neural response.
The bottom right panel depicts the encoded Gaussian distribution, using the
same encoding scheme.

a population of neurons could implement approximate Bayesian inference.

Neuroscientists want to discover which implementation scheme(s) the brain
actually uses.
Research into neural implementation of approximate Bayesian inference presup-
poses a broadly realist perspective on credal states and transitions (Ma, 2019;
Rescorla, 2021a). If there are no priors, investigating how priors are realized in
the brain is a waste of time. If the brain does not execute approximate Bayesian
inference, investigating the neural operations that implement approximate Bayesian
inference is a bad use of scientiﬁc resources. Thus, a major strand in current
computational neuroscience presupposes a realist stance towards at least some
Bayesian models.

Research into neural implementation also helps clarify realism’s commitments.

The core realist thesis is that priors and posteriors are genuine mental states that
mediate between inputs and outputs. As genuine mental states, they must be
neurally realized in some way or other. Realism is neutral about how exactly they
are realized. In particular, realists do not claim that credal assignments are explicitly
enumerated in the brain. On the contrary, realists recognize that explicit enumer-
ation is impossible for most cases. They instead appeal to implicit encoding. They
hold that credal states posited by Bayesian models are implicitly encoded by the
brain. The implicit encoding scheme might be parametric, sampling, or something
else entirely (e.g. Ganguli and Simoncelli, 2014). The brain may also use multiple
encoding schemes simultaneously. Realism does not enshrine a commitment to any
particular encoding scheme or class of encoding schemes.

6 Mental Representation
The previous section advanced a realist perspective on the credal states posited
by Bayesian models. I now want to probe more deeply into the nature of the
posited credal states. I will explore how they relate to the mind’s representa-
tional nature.
The phrase “mental representation” is used many different ways in contem-
porary philosophy and psychology. My own usage reﬂects a tradition that traces
back to Frege (1892/1997) and continues through contemporary ﬁgures such as
Burge (2010) and Fodor (1975; 1987; 2008). According to this tradition, mental
representation is connected with veridicality-conditions: conditions for veridi-
cally representing the world. Examples:

• Beliefs are the sorts of things that can be true or false. My belief that Napoleon
was born in Corsica is true if Napoleon was born in Corsica, false if he was not.
• Intentions are the sorts of things that can be fulﬁlled or thwarted. My intention
to eat lentils for lunch is fulﬁlled if I eat lentils for lunch, thwarted if I do not.
• Perceptual states are the sorts of things than can be accurate or inaccurate.
Suppose I perceive object o as being a green cube. Then my perceptual state is
accurate only if o is green and cubical.

Beliefs have truth-conditions, intentions have fulﬁllment-conditions, and per-

ceptual states have accuracy conditions. Truth, fulﬁllment, and accuracy are
species of veridicality.
Representational properties are properties that contribute or potentially con-
tribute to veridicality-conditions. For example, suppose I have a belief about
Napoleon. The mere fact that my belief is about Napoleon does not determine
whether my belief is true or false. Nevertheless, my belief depends for its truth or

falsity on how things are with Napoleon (rather than some other person). That
my belief is about Napoleon helps determine the belief’s truth-condition. So
being about Napoleon is a representational property of my belief. Similarly,
suppose I perceive some object as a green cube. The mere fact that my percep-
tual state represents green cubicality does not determine whether the state is
accurate—accuracy also depends on which cube I am perceptually representing.
Nevertheless, my perceptual state depends for its accuracy on whether the
perceptually represented object is a green cube. That my perceptual state repre-
sents green cubicality helps determine the state’s accuracy-condition. So repre-
senting green cubicality is a representational property of my perceptual state.
I will argue that credal states posited within Bayesian cognitive science have
representational properties, and I will elucidate the explanatory role played by
these representational properties.

6.1 Representational Explanation

Bayesian cognitive science seeks to explain mental and behavioral outcomes. It
frequently characterizes the outcomes in representational terms. Examples:

• Perceptual psychology seeks to explain illusions. An illusion is a perceptual

state that inaccurately represents the distal environment. So the science
presupposes that perceptual states have accuracy-conditions.
• Sensorimotor psychology seeks to explain how the motor system chooses
motor commands that promote the agent’s goals. A goal may be fulfilled or
thwarted. So the science presupposes mental states with fulfillment-
conditions. These are conative states, i.e., mental states whose role is to
initiate and sustain action. Often, the conative state is an intention (e.g. an
intention to reach to a target). Burge (2022, pp. 502–530) argues that there
also exist relatively low-level conative states lacking various features of
intention, such as intention’s characteristic ties to theoretical and practical
reasoning, and that these low-level conative states set goals for motor control.
For present purposes, the key point is that sensorimotor psychology presup-
poses goal-setting by mental states with fulfillment-conditions.
• Sensorimotor psychology seeks to explain why movement details vary more
along task-irrelevant dimensions than task-relevant dimensions. The distinc-
tion between task-relevant and task-irrelevant dimensions presupposes a goal
that may be fulfilled or thwarted.
• Research on human dead reckoning seeks to explain overshooting. To over-
shoot a location, the subject must have that location as a goal. So the science
presupposes a conative mental state with a fulfillment-condition. Since
human dead reckoning typically interfaces with fairly sophisticated planning

and decision-making, it seems likely that the conative state is typically an

intention or something much like an intention. In some special cases, though,
it may be a relatively low-level conative state more along the lines discussed
by Burge.

As these examples illustrate, Bayesian cognitive science often characterizes

explananda in representational terms.
The science also frequently characterizes explanantia, including credal states,
in representational terms. Examples:

• The Bayesian dead reckoning model explains overshooting by positing

a “slow speed” prior over self-motion. The prior causes the navigation system
to underestimate displacement. To encode a prior that favors slow speeds, the
navigation system must be able to represent speed. So the explanation of
overshooting presupposes that the navigation system can represent speed.
The explanation hinges upon a credal allocation over possible speeds, leading
to an inaccurate displacement estimate.
• Bayesian perceptual psychology assumes that the perceptual system represents
distal properties. It posits a prior regarding represented distal conditions (e.g.
a prior that favors overhead lighting directions). When the prior is poorly
calibrated to the perceiver’s environment, the resulting perceptual estimates
tend to be inaccurate. For example, the “light from overhead” prior produces
inaccurate shape estimates in deviant conditions where light comes from below.
• To explain how the motor system promotes the agent’s goals, Bayesian
sensorimotor psychology posits sequential updating of credal assignments
regarding the distal environment and the subject’s own body. Credal assign-
ments influence which motor commands are chosen. When credal assign-
ments are poorly calibrated to the environment (e.g. the prior over shifts in
finger position does not match actual finger shifts), the task goal tends to be
thwarted.

Generally speaking, Bayesian cognitive science posits credal states regarding

environmental conditions, including both distal properties (e.g. size, shape,
color, location, density, etc.) and bodily properties (e.g. hand position). In
describing credal states, researchers presuppose that the mind can represent
the relevant environmental properties. Researchers characterize credal states
by invoking representational relations to the environment. They cite these
representationally-characterized credal states as explanantia.
Researchers in Bayesian cognitive science do not use the phrase “veridicality-
condition.” The speak instead of random variables, probability distributions, pdfs,
and other entities drawn from probability theory. Nevertheless, their theorizing

assigns a central role to veridicality-conditions. They identify both explananda and

explanantia by citing representational properties: either veridicality-conditions or
properties that potentially contribute to veridicality-conditions.
Take the Bayesian dead reckoning model. The model seeks to explain
overshooting, which presupposes a conative state with a fulfillment-
condition. To explain overshooting, the model posits that the navigation
system underestimates displacement—in other words, that the displacement
estimate is inaccurate. So the model explains overshooting by positing
a mental state (the displacement estimate) that is evaluable as veridical or
nonveridical. To explain why the navigation system underestimates displace-
ment, the model posits a prior that favors slow speeds. The prior assigns
credences to hypotheses regarding the creature’s speed. For example, it
assigns a credence to the hypothesis that the creature’s speed lies in the
interval ½a; b. This hypothesis is individuated through representational rela-
tions to possible speeds (namely, speeds lying between a and b). By citing
credal assignments to representationally-individuated hypotheses, the model
depicts the navigation system as favoring slow speeds. It thereby explains
overshooting. Explanation is laced at every stage with appeals to representa-
tional properties.
Similarly, consider Bayesian modeling of size perception (Ernst & Banks,
2002; Helbig & Ernst, 2008). Here we posit a prior over possible distal sizes.
The prior combines with sensory input (e.g. haptic or visual input) and a prior
likelihood, yielding a posterior over possible distal sizes. On that basis, the
perceptual system chooses a privileged size estimate, which goes into the final
percept. The percept is veridical only if the perceived object has the estimated
size. Thus, the final size estimate is individuated representationally—through
its contribution to the percept’s veridicality-condition. The prior and posterior
are also characterized representationally. These are credal states that allocate
credences over hypotheses regarding distal size. Hypotheses are individuated
through their representational properties—through the distal sizes that they
represent. So the model posits mental states with representational properties,
mediating between proximal sensory input and the (representationally-
characterized) perceptual size estimate.
One could offer a similar analysis for virtually every other explanation
found within Bayesian cognitive science. Bayesian researchers frequently
characterize explananda in representational terms. They almost invariably
characterize credal states in representational terms. For that reason, their
research fits well with the representationalist paradigm espoused by Burge
(2010; 2022), Fodor (1975; 1987; 2008), Peacocke (1994; 1999), Pylyshyn
(1984), Shea (2018), and many others.

6.2 Credal States Versus Mathematical Tools

I now develop my analysis by examining more closely the formal apparatus
used by Bayesian modelers. The key point I wish to highlight is the distinction
between credal states versus the mathematical tools used to specify credal
states.
Look again at Figures 33 and 34. The green downward-sloping curve is a pdf:
a function from ℝ to ℝ. The pdf induces a probability distribution over sets of
real numbers. The pdf and the induced probability distribution are mathematical
tools that theorists use to specify the “slow speed” prior. The “slow speed” prior
is a credal state: an assignment of credences to hypotheses. We must sharply
distinguish the credal state from the pdf and also from the induced probability
distribution. Nothing about the pdf taken on its own suggests we are modeling
a credal state that concerns speed. The same pdf could just as well specify a prior
over possible sizes, or possible distances, or any other one-dimensional con-
tinuous physical magnitude. The pdf in itself does not indicate that we are
modeling a “slow speed” prior as opposed to a “small size” prior, a “short
distance” prior, or numerous other possible priors. The same goes for the
induced probability distribution.
Similar remarks apply to most other Bayesian models. The modeler typically
specifies credal states through a probability distribution over sets of real num-
bers, which in turn is typically specified through a pdf. The probability distri-
bution taken on its own does not even begin to dictate the underlying credal
state. The credal state is defined over hypotheses that are individuated through
their representational relations to the environment. The probability distribution
is a mathematical function individuated without regard to any such representa-
tional relations. The same probability distribution could just as well specify
many different credal states.
To identify the credal state specified by a pdf, we must look beyond math-
ematical formalism and consider the broader enterprise to which the formalism
contributes. We must first ask which psychological domain is being modeled:
perception, or motor control, or navigation, and so on. We must also ask which
aspects of the environment are represented by the credal state: shape, or size, or
color, or speed, and so on. Usually, we can answer these questions by studying
the text that accompanies the formalism. For example, Lakshminarasimhan
et al. (2018, p. 195) write that overshooting “can be explained by a model in
which subjects maximized their expected reward under the influence of a slow-
speed prior rather than by leaky integration of unbiased velocity estimates.”
This passage and kindred passages show that the pdf from Figures 33 and 34 is
intended to specify a credal state that favors slow speeds and that is deployed

during dead reckoning. Analogous passages abound throughout Bayesian cog-

nitive science. These passages are not idle prattle or disposable heuristic. They
play a crucial theoretical role: they point us towards the credal states speciﬁed
by Bayesian models.
Pdfs are indispensable mathematical tools. They allow us to specify
credal states with mathematical precision, and they allow us to bring the
calculus of real numbers to bear. Ultimately, though, they omit something
crucial. They omit the representational properties that help individuate
credal states.
To bring the distinction between credal states versus mathematical tools into
sharper relief, it helps to reﬂect upon measurement units. Using measurement
units, we can describe a physical magnitude (such as a speed) with a real
number. For example, we can say that an object’s speed is 10 meters/sec. The
physical magnitude is quite distinct from the number 10 that we use to measure
it, as evidenced by the fact that a change in measurement units necessitates
a change in the number used to specify the same physical magnitude. If we
switch from meters/sec to feet/sec, we must now say that the object travels at

3:28084 10 ¼ 32:8084

feet/sec. We cite a different number to specify the same speed. Speeds are
distinct from the numbers through which we measure speeds.12
When we specify a credal state through a pdf, our choice of pdf depends upon
a canonical choice of measurement units. A change in measurement units neces-
sitates a change in the pdf we use to specify the credal state. Figure 37 illustrates.
The blue pdf corresponds to meters/sec. The orange pdf corresponds to feet/sec.
The pdfs are different, but they specify the same underlying probability assign-
ment over possible speeds. They specify the same “slow speed” prior. Full
technical details are given in Section A5, but the point should be intuitively
clear even absent any technical details. A pdf is defined over real numbers, so it
can describe a credal allocation over possible speeds only relative to measurement
units that map speeds to real numbers. If we change the measurement units, then
we must use a different pdf to model the same credal allocation over speeds. The
different pdf will induce a different probability distribution over sets of real
numbers, even while the underlying credal allocation remains fixed.
Our choice of measurement units reflects our societal conventions, not
inherent features of the credal state itself. There is no reason to suspect that pre-
theoretic human navigation employs our conventional measurement units.

12
See Peacocke (2019) for a general account of physical magnitudes, including argumentation that
we should add these items to our ontology.

Figure 37 A change in measurement units necessitates a change in pdf. The blue

pdf corresponds to meters/sec. The orange pdf corresponds to feet/sec. To
convert from meters/sec to feet/sec, multiply by 3.28084. In Figure 37,
b ¼ 3:28084 a. The area under the blue curve over ½0; a equals the area under
the orange curve over ½0; b. The same equality holds for all other points a and
b such that b ¼ 3:28084 a. Thus, the two pdfs model the same probability
assignment over possible speeds.

Indeed, it may not use any measurement units at all. (Cf. Peacocke, 2019, p. 48.)
The same credal state could just as well be specified by a different pdf. For
example, there is no reason to regard the blue pdf from Figure 37 as privileged
over the orange pdf. Neither pdf has more psychological reality than the other.
Psychological reality resides in the underlying credal state—a credal allocation
over hypotheses regarding possible speeds—rather than the pdf.
The “slow speed” prior is a credal state that allocates credences over hypoth-
eses, where the hypotheses are individuated through the specific speeds that
they represent. The pdf is a purely mathematical function that reflects
a conventional choice of measurement units. The prior does not reflect any
such conventional choice. The pdf is a useful tool for specifying the under-
lying credal state, but its mathematical elegance should not dazzle us into
ascribing psychological reality to it. The credal state is psychologically real.
The pdf is not psychologically real, and neither is the induced probability
distribution over sets of real numbers.

6.3 Random Variables Revisited

We can clarify the distinction between credal states and mathematical tools by
revisiting the notion of random variable.

Recall from Section 2.3 that a random variable X maps an outcome space Ω to
the real numbers ℝ. For example, suppose the outcome space Ω contains
possible speeds of an asteroid. Each outcome ω in Ω is a speed that the asteroid
might have. Speeds are physical magnitudes, not real numbers. Assuming
a canonical choice of measurement units, we can measure magnitudes using
real numbers. Let X be a random variable that maps each speed to the corres-
ponding real number, using meters/sec as canonical units. Thus,

X ðωÞ ¼ x

when x speciﬁes speed ω in meters/sec. X is a function from Ω (the set of

possible speeds) to ℝ.
Given a random variable and an underlying outcome space Ω, we can use
a probability distribution over sets of real numbers to specify a probability
distribution over sets of outcomes. Continuing with the asteroid example,
suppose we are given a probability distribution μ that assigns probabilities to
sets of real numbers. Then we can use X and μ to assign probabilities to sets of
speeds. For example, what probability should we assign to the event X 1 ½a; b?
This event codiﬁes the hypothesis that the asteroid’s speed falls between a and
b. If we are taking μ as a guide, we should assign the same probability to

X 1 ½a; b that μ assigns to ½a; b. In other words, if P X 1 ½a; b is the probabil-
ity assigned to X 1 ½a; b, then we should have

P X 1 ½a; b ¼ μð½a; bÞ;

As Figure 38 illustrates, we can use X to transfer the probability distribution μ

deﬁned over sets of real numbers into a probability distribution P deﬁned over

Figure 38 Illustrates the relation between P and μ. μ maps ½a; b to the same
probability as that to which P maps X 1 ½a; b.

sets of speeds. More generally, and as discussed more rigorously in Section A3,
we can always use a random variable to transfer a probability distribution over
sets of real numbers into a probability distribution over sets with members
drawn from the underlying outcome space.
Now consider a different random variable Y that maps each magnitude to the
corresponding real number using feet/sec. Thus,

Y ðωÞ ¼ y

when real number y speciﬁes speed ω in feet/sec. Using the standard conversion
from meters/sec to feet/sec, we obtain the following relation between X and Y:

Y ðωÞ ¼ 3:28084X ðωÞ:

The same magnitude ω is mapped to a different real number, depending on

whether we are using meters/sec (corresponding to X ) or feet/sec (correspond-
ing to Y ).
If we want to specify a fixed probability distribution P over sets whose
members come from Ω itself, then X and Y mandate different probability
distributions over sets of real numbers. Figure 37 illustrates. The pdf in blue
generates one probability distribution over sets of real numbers. The pdf in
orange generates a second probability distribution over sets of real numbers.
Transferring the first probability distribution via X yields the same result as
transferring the second probability distribution via Y. The very same probability
distribution P over sets of speeds results if we use the blue pdf and X or if we use
the orange pdf and Y.
These observations complement my diagnosis from Section 6.2. Our choice
of pdf reflects our choice of a random variable, which reflects our choice of
measurement units. Different measurement units mandate a different pdf in
order to specify the same probability distribution over sets with members drawn
from the underlying outcome space. These facts, which are basic to probability
theory, reflect the inherently arbitrary nature of measurement using real num-
bers. Many different measurement units are equally legitimate. Different units
yield different pdfs and different probability distributions over sets of real
numbers, but the underlying probability distribution over sets of outcomes
remains fixed.
In the special case of Bayesian cognitive science, we seek to model credal
allocations by an agent or an agent’s psychological subsystems (such as the
perceptual system). The most convenient way to specify a credal allocation is
usually through a pdf, as in Figures 33 and 34. The pdf is merely a tool for
specifying a credal allocation over an underlying outcome space Ω. More

precisely: the credal allocation assigns credences to sets whose members are
drawn from Ω. The pdf depends upon an arbitrary choice of measurement units.
The credal allocation does not. Psychological reality resides in the credal
allocation, not the pdf.

6.4 The Objects of Credence

In Sections 6.2 and 6.3, I argued that the pdfs invoked by Bayesian cognitive
scientists are mathematical tools for specifying credal states. A credal state
assigns credences to sets of outcomes, where outcomes are drawn from an
outcome space Ω. What are the outcomes? In other words, what are the elements
of Ω? Answering this question is a large undertaking. I will broach a few
preliminary considerations that should inform a more complete treatment.
Note first that Kolmogorov’s axiomatization does not tell us what outcomes are.
Kolmogorov assigns probabilities to events: sets whose members belong to an
outcome space Ω. He places no constraints whatsoever upon Ω’s members. Thus,
the mathematical formalism of probability theory does not answer our question.
Explanatory practice within Bayesian cognitive science places some con-
straints upon Ω, but it does not dictate a unique answer. For example, the
Bayesian dead reckoning model posits a prior over possible speeds, so out-
comes must intimately relate somehow to speed. However, this constraint leaves
room for various interpretations.
One interpretation is that Ω contains possible worlds. We would then con-
strue events as sets of possible worlds. In the Bayesian dead reckoning model,
the hypothesis that the creature moves with speed between a and b would be
codified as the set of possible worlds where the creature moves with speed
between a and b. In a Bayesian model of size perception, the hypothesis that the
perceived object has size between a and b would be codified as the set of
possible worlds where the perceived object has size between a and b. These
codifications fit well with contemporary philosophical work, which often
assigns credences to sets of possible worlds. More generally, they fit well
with the longstanding philosophical tradition, mentioned in Section 2.1, of
glossing propositions as sets of possible worlds.
A second interpretation is that Ω contains physical magnitudes. In the
Bayesian dead reckoning model, the hypothesis that the creature moves with
speed between a and b would be codified as the set of speeds between a and b. In
a Bayesian model of size perception, the hypothesis that the perceived object
has size between a and b would be codified as the set of sizes between a and b.
A third interpretation is that Ω contains mental representations. A mental
representation is a mental item with representational properties. Mental

representations are similar in key respects to the communal representations

employed by human society, such as pictures, maps, or natural language
sentences, but they are housed in the mind rather than the external world.
They can be stored in memory, manipulated during mental activity, and com-
bined to form complex representations. Appeal to mental representations is
widespread in cognitive science theorizing (Carey, 2009; Fodor, 1975; Fodor,
2008; Gallistel & King, 2009; Pylyshyn, 1984; Rescorla, 2020d). If we take Ω to
contain mental representations, then we will construe events as sets of mental
representations. In the Bayesian dead reckoning model, the hypothesis that the
creature moves with speed between a and b would be codified as something like
the set of mental representations that attribute speed between a and b. In
a Bayesian model of size perception, the hypothesis that the perceived object
has size between a and b would be codified as something like the set of mental
representations that attribute size between a and b. Although these codifications
may look odd to philosophers reared on the possible worlds interpretation,
they fit nicely with the widespread cognitive science commitment to mental
representations.
Each of the three interpretations is compatible with a realist perspective on
Bayesian cognitive science. Moreover, each interpretation codifies hypoth-
eses in representational terms. The first interpretation collects together those
possible worlds where the hypothesis is veridical. The second interpretation
collects together those physical magnitudes that are consistent with the
veridicality of the hypothesis. The third interpretation collects together
mental representations according to which the hypothesis is veridical.
Thus, all three interpretations analyze credal states representationally—in
terms of veridicality-conditions or representational properties that contribute
to veridicality-conditions.
All three interpretations deserve detailed consideration, as no doubt do
other interpretations. My own sympathies lie with the third interpretation,
but I will not attempt to defend it here. My goal instead is to highlight the need
for some interpretation. To understand the credal states posited by Bayesian
cognitive science, we must identify the entities to which credences attach. We
must identify the objects of credence. Assuming that credences attach to sets,
our task is to identify which elements belong to the sets. By making progress
on this task, we may hope to illuminate the representational nature of credal
states.13

13
Mahtani (2024) conducts a detailed investigation into the objects of credence, focused primarily
on the intersection of formal epistemology with philosophy of language rather than on Bayesian
cognitive science.

6.5 How Many Outcomes?

In addition to studying what outcomes are, we must also consider how many
outcomes there are. A set is countable when we can count its members using the
numbers 0, 1, 2, 3, . . . . A set is uncountable when we cannot so count its
members. A random variable is discrete when it has countably many possible
values, nondiscrete when it has uncountably many possible values. The
Bayesian dead reckoning model features a random variable whose possible
values correspond to possible speeds of the navigator (specified through canon-
ical measurement units). Even if we stipulate a maximum possible speed s, there
are still uncountably many real numbers lying in the interval ½0; s and hence
uncountably many possible speeds. So the random variable is nondiscrete, and
the underlying outcome space Ω is uncountable. Similarly for Bayesian model-
ing of motion estimation (Weiss, Simoncelli & Adelson, 2002), size estimation
(Ernst & Banks, 2002), motor control (Todorov & Jordan, 2002), and numerous
other tasks. In general, whenever cognitive scientists model Bayesian estimation
of a physical magnitude that has uncountably many possible values (e.g. time,
distance, speed, orientation, size), the resulting Bayesian model invokes a non-
discrete random variable X defined over an uncountable outcome space.
Taken literally, such a model attributes highly infinitary representational
capacities. More specifically:

(i) The model posits credal states (a prior and a posterior) that assign
probabilities to events X 1 ½a; b. There are uncountably many events
X 1 ½a; b, so the model posits a credal assignment over uncountably
many events.
(ii) The model posits credal states drawn from among uncountably many possible
options. This remains so even if we demand that credal assignments belong to
a ﬁxed parametric family, such as the family of Gaussian distributions.
(iii) The model posits a privileged estimate x* of X’s value, as in Figure 28.
There are uncountably many possible values x*, so the model posits
a privileged estimate selected from among uncountably many options.

Hence, the model attributes highly inﬁnitary representational capacities when

specifying both credal states and privileged estimates.
Some philosophers will bristle at these infinitary attributions. The attributions
may look incompatible with obvious finitary limits on our representational or
computational capacities. It might seem that we should dismiss (i)‒(iii) as mere
idealizations, eventually to be obviated by a more plausible model that honors
the finitary limits on human mental activity. Shouldn’t a plausible model restrict
itself to a finite outcome space?

I agree that there are finitary limits of some sort on human representational and
computational capacities. For example, we do not have infinite memory storage
capacity: the mind cannot explicitly list infinitely many distinct pieces of infor-
mation. Yet I wonder whether (i)‒(iii) flout any genuine finitary limits on human
mental activity. As discussed in Section 5.4, computational neuroscience offers
various theories of how the brain could, in principle, implement or approximately
implement Bayesian inference. The theories are biologically plausible, and they
fit well with diverse neurophysiological data. Several theories describe the brain
as implementing a Bayesian model that satisfies (i)‒(iii). Those theories feature
nondiscrete neural variables (e.g. membrane potential), which are taken to pro-
vide a substrate for credal states. Thus, (i)‒(iii) look compatible with lots of work
in contemporary computational neuroscience.
The classical computational theory of mind (CTM) holds that mental activity
is digital computation (Fodor, 1975; 1987; 2008; Gallistel & King, 2009;
Pylyshyn, 1984; Rescorla, 2020). A digital computing system has at most
countably many possible computational states. Hence, CTM is incompatible
with (ii) and (iii). However, CTM is compatible with (i). There is a well-
developed framework—computable probability theory—that studies how digi-
tal computing systems can encode and compute over probability distributions
(Ackerman, Freer & Roy, 2019). In this framework, the computing system often
satisfies (i) but not (ii) or (iii). The system encodes a credal assignment over
uncountably many events, but there are only countably many possible credal
assignments and privileged estimates x* available to the system. For example,
the system may encode a Gaussian distribution, but there are only countably
many distinct Gaussian distributions that it could have instead encoded (it can
only encode a Gaussian whose mean and variance are drawn from a fixed
countable set). In more practical terms, computer scientists and roboticists
frequently program digital systems to compute over nondiscrete random vari-
ables (e.g. Thrun, Burgard & Fox, 2005). These systems encode a wide range of
probability distributions, including Gaussian distributions and many others
besides. Their computations satisfy (i) though not (ii) and (iii). Thus, propon-
ents of CTM can happily allow that the mind assigns credences to uncountably
many events.
Infinitary Bayesian models raise thorny questions at the intersection of phil-
osophy, psychology, computation theory, and neuroscience.14 I cannot hope to
settle these questions here. For present purposes, the key point is that a realist
representationalist perspective on Bayesian cognitive science admits several

14
In particular, they engage longstanding debates over whether the mind executes digital versus
analog computation. See Rescorla (2020a) for an introductory discussion.

divergent reactions to an explanatorily successful Bayesian model deﬁned over

a nondiscrete random variable, including the following three reactions:

• Accept the model at face value; embrace (i)‒(iii).

• Guided by computable probability theory, emend the model by allowing only
countably many of the credal states and estimates posited by the model;
embrace (i) but not (ii) and (iii).
• Try to replace the model with a purely ﬁnitary approximation; reject (i)‒(iii).

Each position is compatible with realism, which commits to credal states and
transitions approximately like the ones posited by the model but does not insist
that the model is literally true. Each position is compatible with representation-
alism, which champions the representational nature of credal states but does not
mandate inﬁnitary representational capacities.

7 Anti-representationalism
Anti-representationalists hold that we should expunge mental representation from
rigorous scientiﬁc theorizing. They seek to explain mental and behavioral phe-
nomena in strictly nonrepresentational terms. Different anti-representationalists
favor different nonrepresentational paradigms:

• Quine (1960) favors Skinnerian stimulus-response psychology.

• Churchland (1981) favors a neurophysiological paradigm.
• Field (2001) and Stich (1983) favor nonrepresentational computational
description.
• van Gelder (1992) favors dynamical system theory.

Despite these differences, anti-representationalists agree that mental repre-

sentation makes no useful contribution to scientiﬁc theorizing about the
mind.
Anti-representationalism conﬂicts with Bayesian cognitive science. As we
have seen, Bayesian researchers routinely characterize explananda in represen-
tational terms. If we abjure representational discourse, then we cannot acknow-
ledge those explananda. For example, anti-representationalists cannot replicate
how the Bayesian dead reckoning model explains overshooting: overshooting
is a representationally-characterized explanandum, because a subject can
overshoot a location only when she has that location as a goal. Nor can
anti-representationalists characterize a perceptual state as illusory: an illusion
requires perceptual states with accuracy-conditions. Nonrepresentational theoriz-
ing ignores representational properties and hence cannot mention, let alone explain,
representationally-characterized explananda. Since anti-representationalists cannot

explain representationally-characterized explananda, they cannot replicate the

explanatory benefits secured by Bayesian cognitive science.
Neither can anti-representationalists accept successful Bayesian explanations
for nonrepresentational explananda. Suppose we characterize the results of
dead reckoning in purely nonrepresentational terms. For example, we can
identify the creature’s final position across various trials, without mentioning
whether the position overshoots any target location. The Bayesian dead reckon-
ing model explains the nonrepresentationally characterized explanandum. It
does so by isolating causally relevant factors (including the “slow speed” prior)
that influence position. This is a representational explanation: it explains a
nonrepresentational explanandum (position) by citing a credal allocation over
representationally-characterized hypotheses. Anti-representationalists cannot
accept the explanation. Their anti-representationalist scruples forbid explan-
ations that cite representational properties of mental states.
Anti-representationalists claim that we can replicate any purported benefits
of representational explanation through alternative explanations couched in
purely nonrepresentational terms. They claim that we can jettison mental
representation while preserving the explanatory achievements enabled by
representationally-characterized explanantia. The long, dismal history of
anti-representationalist theorizing provides little basis for that claim. Anti-
representationalists have consistently failed to match even the most elemen-
tary explanatory achievements of representationalist cognitive science. For
example, Gibson’s (1979) direct perception framework seeks to analyze
perception in nonrepresentational terms, but it cannot explain a huge range
of perceptual illusions and constancies (Fodor & Pylyshyn, 1981). Similar
remarks apply to numerous other anti-representationalist theories that have
flitted in and out of fashion over the past century.
In the present dialectical context, the key question is whether anti-
representationalists can preserve the explanatory benefits of Bayesian models
without invoking representational mental states. I doubt it. One cannot usually
strip a scientific theory of its main theoretical concepts while retaining its
explanatory benefits. For example, renouncing talk about subatomic particles
would severely limit the explanatory power of physics. I see no reason to think
that we can renounce talk about representational credal states while retaining
the explanatory benefits provided by such talk. Consider the Bayesian dead
reckoning model. It relies in an essential way upon the “slow speed” prior. By
invoking this prior, the model achieves a much better fit with experimental data
than the hitherto dominant “leaky integrator” model. The “slow speed” prior is
characterized in representational terms. How, then, can we replicate its explana-
tory contribution while eschewing representational discourse? The principal

explanatory advance made by the model was a posit of representational mental

states.
To support my viewpoint, I will now critique two anti-representationalist
interpretations of Bayesian cognitive science. The interpretations differ in
various ways. They agree that, contrary to what I suggested in Section 6,
Bayesian models of the mind do not postulate representational mental states.
I will explain why I think both interpretations are mistaken.

7.1 Function-theoretic Computation

Egan (2010; 2020) advocates a function-theoretic approach to mental computation:
“The input of a computationally characterized mechanism represents the arguments
and the outputs the values of a mathematical function that canonically specifies the
task executed by the mechanism” (2020, p. 33, fn. 7). A computational theory of
a mechanism “comprises a specification of the function (in the mathematical sense)
computed by the mechanism” (2020, p. 33). Thus, computational psychology
provides “an abstract mathematical description” that prescinds from representa-
tional properties of mental states (Egan, 2010, p. 256). She admits that cognitive
scientists frequently mention representational properties when describing mental
states. She maintains that representational discourse “is best construed as a kind of
gloss—an intentional gloss—on the computational theory” (2020, p. 33). The
intentional gloss plays a useful heuristic role in our theorizing: it helps us connect
our computational description with representationally-characterized explananda; it
helps us track how the computational mechanism responds to environmental events;
and it can serve as a temporary placeholder until we discover an underlying
computational mechanism. Representational properties do not figure in genuinely
computational theories and are not necessary for good cognitive science explan-
ations: “the computational theory proper can fully explain the interaction between
organism and environment . . . without adverting to cognitive content” (2020, p. 34).
Egan’s function-theoretic approach encompasses the following three doc-
trines, each of which I reject:

(a) Computational models of the mind mention inputs and outputs, but they do
not mention internal states that mediate between inputs and outputs.
(b) Computational models describe inputs and outputs in purely mathematical
terms, without mentioning any representational properties of the inputs or
outputs.
(c) Representational discourse plays a purely heuristic role in cognitive science
theorizing. It makes no genuine explanatory contribution.

I will critique doctrines (a)‒(c) in turn.

Doctrine (a) conﬂicts with huge amounts of cognitive science theorizing.

Computational modeling by cognitive scientists routinely posits internal states
that mediate between inputs and outputs. All the Bayesian models I have
discussed above are examples. For example, Bayesian models of perception
as encapsulated by Figure 28 posit three internal credal states: a prior probabil-
ity, a prior likelihood, and a posterior (or approximate posterior). These credal
states mediate between the input (proximal sensory stimulation) and the output
(a privileged perceptual estimate of a distal property). More complex models,
such as models of motor control, posit a sequence of credal states mediating
between inputs and outputs. Evidently, Bayesian models commit to far more
internal computational detail than (a) allows. The models do not merely
describe a function from inputs to outputs. They say something informative
about the internal states and transitions through which the system converts
inputs into outputs.
Egan disagrees. She asserts that Bayesian models carry “no commitment to
internal states or structures and causal processes deﬁned on them” (2020, p. 48)
and that “Bayesian models, to the extent that they say anything about how the
brain actually works, give . . . a function-theoretic characterization; they specify
the function, in the mathematical sense, computed by the mechanism” (2020,
p. 50). She does not justify her analysis by adducing a single Bayesian model
found in cognitive science. She does not attempt to reconcile her analysis with
the commitment, apparently ubiquitous throughout Bayesian cognitive science,
to credal states and transitions. She simply states, without evidence or argument,
that Bayesian models are not committed to any internal states or processes.
Egan professes neutrality in the debate between realist versus instrumentalist
perspectives on Bayesian modeling (2020, p. 49, fn. 22). Yet her analysis seems
irreconcilable with the most faintly realist perspective. On anything resembling
realism, we should accept the existence of credal states and transitions mediat-
ing between inputs and outputs. Only if we adopt a strongly instrumentalist
perspective may we regard a Bayesian model as specifying a mere function
from inputs to outputs. I indicated in Section 5 why I favor realism over
instrumentalism.
Doctrine (b) is also problematic, at least as applied to Bayesian modeling of
the mind. Bayesian models routinely specify either inputs or outputs in repre-
sentational terms:

• Bayesian sensorimotor models specify a task goal as input to sensorimotor

processing. The goal is set by a conative state with a fulﬁllment-condition.
Hence, the Bayesian model presupposes a representationally-speciﬁed
mental state.

• Bayesian perceptual models usually yield as output a perceptual estimate of

some distal property. Estimates can be accurate or inaccurate. An estimate
that an object has a certain size is accurate only if the object has that size; an
estimate that an object moves with a certain speed is accurate only if the
object moves with that speed; and so on.

These representational descriptions are inherent to the computational model.

They are all we have to go on when identifying the relevant inputs or outputs.
For example, suppose a Bayesian model outputs an estimate of an object’s size.
The model individuates the estimate through its representational relation to a
specific distal size. If we abandon any reference to represented size, we abandon
our only way of identifying the model’s outputs.
Doctrine (c) is similarly problematic. As I documented in Section 6, Bayesian
models routinely individuate credal states in representational terms. Abandoning
representational discourse leaves us with no way to identify the credal states
postulated by the model and hence no way to replicate an explanation that cites
those credal states. For example, suppose a Bayesian perceptual model posits
a prior over distal size. If we refuse to mention sizes represented by the perceptual
system, then we cannot identify the hypotheses to which the prior assigns
credences, so we cannot cite the prior to explain anything. Accordingly,
I disagree with Egan’s claim that “Bayesian models are typically not developed
at a level of description that allows us to assess their representational commit-
ments, in the relevant sense. They have no representational commitments, in the
relevant sense” (2020, p. 48). Once again, Egan does not provide any concrete
examples to validate her assessment. She does not indicate, for even a single case,
how we are to individuate credal states in nonrepresentational terms. The lack of
detail is not surprising, since a nonrepresentational individuative scheme looks
fundamentally incompatible with the core methodology of Bayesian cognitive
science.
Egan is certainly correct that Bayesian cognitive science uses mathematical
tools to characterize inputs, outputs, and mediating credal states. Inputs and
outputs are typically described using real numbers. Mediating credal states are
usually described using pdfs. So Bayesian modeling includes “abstract math-
ematical descriptions” somewhat along the lines favored by Egan. As explained
in Sections 6.2 and 6.3, these mathematical descriptions reflect our arbitrary,
conventional choice of measurement units. Psychological reality and explana-
tory power reside in the representational states specified by our mathematical
descriptions, not in the mathematical descriptions themselves. For example, if
we describe the perceptual system as estimating that an object has size s, the
specific real number s reflects our arbitrary choice of units for measuring size.

The number has no psychological reality. It explains nothing. What is psycho-

logically real is that the perceptual estimate represents a specific size—
a physical magnitude, not a real number. Similarly, if we describe a credal
state using a pdf, the pdf reflects our arbitrary choice of measurement units. It
has no psychological reality. It explains nothing. The credal state is what does
the explaining.
The mathematical descriptions emphasized by Egan are artifacts of our meas-
urement conventions. Different measurement units would yield a different math-
ematical description, including a different function from inputs to outputs, while
leaving representational description the same. Representational description, not
mathematical description, is the locus of psychological reality and explanatory
power. For example, suppose we learn that a Bayesian dead reckoner estimates
speed 5. Does that knowledge in itself help us explain overshooting? No. We need
to specify measurement units! 5 meters/sec, or 5 feet/sec, or something else? The
number 5 by itself is explanatorily irrelevant. What matters is the physical
magnitude measured by 5—that is, the speed represented by the dead reckoner.
The represented magnitude, not the number, is explanatorily important. Similar
remarks apply to other mathematical descriptions found in Bayesian cognitive
science, including specification of pdfs.
I critiqued Egan along these lines in previous work (Rescorla, 2015a). Egan
deems my critique “very puzzling” (2020, p. 48) and retorts (2020, p. 50):

To think that commitment to Bayes’ theorem—a function deﬁned on prob-

ability distributions—reﬂects an arbitrary choice of conventions is analogous
to thinking that a claim that a device computes the addition function reﬂects
a commitment to represent addends and sums in base 10. Contra Rescorla, to
the extent that Bayesian models are to be construed realistically . . . such
proposals should be construed as hypotheses about underlying psychological
reality, committed, in particular, to the claim that the system is computing an
approximation to Bayes’ theorem.

I respond as follows:

• Bayes’s theorem is not “a function deﬁned on probability distributions.” It is

a theorem.
• I agree with Egan that “commitment to Bayes’s theorem” does not “reﬂect an
arbitrary choice of conventions.” Bayes’s theorem does not in any way
depend for its truth upon our conventions.
• As a realist about Bayesian cognitive science, I do indeed hold that the mind
often computes an approximation to the posterior. I hold that at least some
Bayesian models describe mental processes with at least approximate
accuracy.

• When we describe a device as computing the addition function, we are not

committed to using base 10 notation. Nor are we committed to saying that the
device uses base 10 notation. There are many possible numerical notations
that a device might use to compute arithmetical functions.
• When we describe a credal state using a pdf, our choice of pdf reﬂects our
arbitrary measurement units for the represented environmental variable. Our
description does not commit us to saying that the mind uses those particular
measurement units. For example, it is highly unlikely that the human naviga-
tion system measures speed using meters/sec.
• Egan claims that abstract mathematical description of mental computation
has explanatory priority over representational description. This position is
implausible because the mathematical description typically reﬂects an arbi-
trary choice of measurement units.

My conclusion: Egan’s response gives no reason to attribute any psychological

reality to abstract mathematical descriptions or to question the explanatory
centrality that I attribute to representational descriptions.
In summary, Egan’s function-theoretic conception does not ﬁt well with
Bayesian cognitive science because it neglects the crucial role that Bayesian
modeling assigns to representational descriptions of explananda and explanan-
tia. In place of representational descriptions, Egan commends abstract mathem-
atical descriptions. Yet abstract mathematical descriptions reﬂect our own
arbitrary measurement units and lack any psychological reality. Egan’s
nonrepresentational approach cannot preserve the most basic explanatory
achievements of Bayesian cognitive science.

7.2 Radical Enactivism

Hutto and Myin (2017) espouse a radical enactivist approach to cognitive science.
They view cognition as a dynamic interaction between an embodied brain and
a changing environment. They “conceive of the basis of cognition in terms of
extensive and dynamically loopy processes that are responsive to information in the
form of environmental variables spanning multiple spatial and temporal scales”
(p. 9). They also reject representationalism: they “construe cognition as unfolding,
world-relating processes rather than as a series of content-bearing states and their
interactions” (p. 9). They acknowledge that talk about veridicality-conditions is
illuminating when applied to sophisticated symbolic communication (p. 90). They
deny that it usefully contributes to theorizing about perception, motor control, or
other relatively low-level psychological domains (pp. 12–13).
Hutto and Myin apply their radical enactivist approach to Bayesian cognitive
science. Following Clark (2015) and Hohwy (2014), they focus almost exclusively

upon a neural implementation framework called predictive coding. The basic

idea behind predictive coding is that the brain generates a prediction regarding
the sensory input it will receive. The brain compares its prediction with actual
sensory input, computing a prediction error term. Prediction error informs
subsequent computation, shaping future expectations so as to minimize future
prediction error. Many predictive coding models have a hierarchical structure:
higher levels of the network compute predictions about lower-level activity, and
the lower level computes a prediction error term that is transmitted back to the
higher level. There is nothing inherently Bayesian about predictive coding
models, but when set up in the right way they can implement an approximation
to Bayesian inference. This can be done either through parametric encoding
(Friston, 2010) or through sampling encoding (Lee & Mumford, 2003). Hutto
and Myin use the label the Predictive Processing account of Cognition (PPC) to
describe theories that implement approximate Bayesian inference through pre-
dictive coding.
Hutto and Myin offer a radical enactivist interpretation of PPC. Their core
interpretive claim is that we need not gloss talk about “prediction” and “expect-
ation” in representational terms. They write: “Having expectations about what
we will experience sensorily need not be thought of as involving the making of
any kind of contentful claim about the state of the world. Nor need we think of
sensory perturbations that are involved in such matches and mismatches as
supplying rich contentful messages that contradict the content of our expect-
ation” (pp. 70–71). Accordingly, we need not interpret PPC models representa-
tionally: “our expectations can fail to match incoming sensory experience
without this activity being construed as a content-based operation” (p. 71).
They conclude that PPC provides no support for representationalism.
I agree with Hutto and Myin that, in many cases, we should not interpret PPC
talk about “prediction” and “expectation” in representational terms. I agree that,
in many cases, we should not describe the mind as “representing” expected
experiences. As Burge (2010, pp. 367–463) notes, there is no evidence that the
perceptual system represents proximal sensory stimulations. The perceptual
system converts nonrepresentational sensory stimulations into perceptual rep-
resentations, without representing the stimulations. There is no explanatory
beneﬁt to saying that the perceptual system represents its own sensory input.
When a PPC neural network compares predicted sensory input with actual
sensory input, we usually should not describe the comparison in representa-
tional terms. We should instead say that the network compares an input signal
with a feedback signal generated by a higher level of the network. We may
describe both signals in neural terms, e.g., as ﬁring rates, and we may describe
the “prediction error” computation as a neurophysiological operation on those

signals. Representational properties play no role in characterizing the “predic-

tion error” computation.
A similar diagnosis applies to higher levels in hierarchical PPC models, such
as the celebrated (Rao & Ballard, 1999) model. Each level compares neural
activity with a feedback “prediction” signal received from a higher level,
computing an “error” term subsequently transmitted to the higher level. The
feedback signal and the “error” computation can again be described in non-
representational, neurophysiological terms.
Typically, then, we should not describe a PPC neural network as representing
its inputs or its own neural activity. The network receives but does represent
proximal sensory inputs. It instantiates but does not represent neural activity.
Talk about “prediction” and “expectation” may be harmless enough for some
purposes, but I agree with Hutto and Myin that we achieve no explanatory gain
by glossing this talk in representational terms.15
However, the nonrepresentational interpretation of prediction talk is doubly
irrelevant to representationalism about Bayesian cognitive science.
First, we should not focus exclusively on PPC models. Most Bayesian
modeling is not tied to the PPC research program. Most Bayesian models
found in cognitive science are neutral about neural implementation mechan-
isms. Many promising implementation schemes, such as the schemes discussed
in Ma et al. (2006) and Orbán et al. (2016), do not feature anything like
predictive coding (Rescorla, 2017; Rescorla, 2024). Thus, the interpretation
of PPC modeling is distinct from the interpretation of Bayesian modeling more
generally. Hutto and Myin give no reason for focusing narrowly on PPC to the
exclusion of generic Bayesian modeling. Indeed, their exposition tends to elide
the difference between PPC and Bayesian cognitive science (e.g. pp. 150–151).
Although predictive coding has received considerable recent attention in the
philosophical community, empirical support for it remains equivocal (Aitchison
& Lengyel, 2017). In my opinion, we currently have no reason to suspect that
approximate Bayesian inference is typically implemented in PPC fashion.
Second, and more importantly, representationalists about Bayesian cognitive
science do not claim that the mind represents either sensory input or neural
activity. Representationalists claim that the mind represents environmental
conditions, including both distal conditions and bodily state. Assume for the
sake of argument that a neural system implements approximate Bayesian
inference through a predictive coding implementation mechanism. We will
describe the system using a Bayesian model, which posits credal states and
15
See also Cao (2020) for critical discussion of talk about “prediction” in the context of predictive
coding models. See Burge (2022, pp. 631–632) for more general cautionary remarks regarding
talk about “prediction” and “expectation” in psychological theorizing.

transitions, and a PPC model, which speciﬁes how the Bayesian model is
neurally implemented by a predictive coding mechanism. I agree with Hutto
and Myin that there is no reason to think that the system represents its own
inputs or own neural activity. Nevertheless, we have strong reason to think that
the system represents the environment. We have strong reason to describe the
system’s credal states in representational terms, as allocating credences over
representationally-individuated hypotheses. Only then can we preserve explan-
ations that rely on representationally-characterized credal states. For example,
how can we explain overshooting in dead reckoning unless we posit a prior that
favors slower speeds? I have no idea how enactivists would interpret the “slow
speed” prior in nonrepresentational terms, let alone how the ensuing explan-
ations would work.
Hutto and Myin (pp. 151–155) express skepticism about my representation-
alist interpretation of Bayesian models. They do not provide a developed
alternative interpretation. They do not indicate how to gloss credal states and
transitions in nonrepresentationalist enactivist terms. In fact, they barely discuss
credal states: they mention priors a mere handful of times, and they do not
mention posteriors at all. They do not analyze a single speciﬁc Bayesian model
of mental activity, even in the most schematic way. Their treatment gives no hint
how enactivists might eschew representational vocabulary while preserving the
explanatory power of Bayesian modeling.

7.3 Interpreting Bayesian Cognitive Science

When philosophers interpret a scientific theory, they often employ theoretical
notions (such as veridicality-condition) that play no explicit role in scientific
discourse. Inevitably, there is a gap between philosophical interpretation and
scientific texts. Still, some interpretations usually fit much better with scientific
practice than others.
In the present case, a representationalist interpretation fits much better with
Bayesian cognitive science than the function-theoretic interpretation offered by
Egan or the enactivist interpretation offered by Hutto and Myin. The represen-
tationalist interpretation describes how to interpret the priors and posteriors that
figure so prominently in Bayesian theorizing. The function-theoretic and enac-
tivist conceptions say virtually nothing about how to interpret priors and
posteriors, save perhaps to dismiss them in instrumentalist fashion as useful
fictions. The representationalist interpretation analyzes in quite precise detail
the explanations offered by Bayesian cognitive scientists, such as the explan-
ations embodied by Figures 33 and 34. The function-theoretic and enactivist
interpretations have little if anything to say about those explanations. Absent

more compelling anti-representationalist interpretations, the representationalist

interpretation looks secure.

8 Conclusion
The mind operates amid constant uncertainty stemming from multiple sources,
including noise, ambiguous input, and conflicting sensory cues. Bayesian
cognitive science postulates that the mind grapples with uncertainty by impli-
citly encoding credal assignments over hypotheses. The encoded credences
influence inference and decision-making, roughly in accord with Bayesian
norms. The Bayesian program draws support from strong empirical evidence
across a range of psychological domains.
I have analyzed Bayesian modeling from a realist representationalist per-
spective that takes seriously the postulation of credal states and transitions.
Realists hold that, when a Bayesian model is explanatorily successful, we have
good reason to accept the existence of credal states and transitions roughly like
those posited by the model. Representationalists hold that the posited credal
states assign credences to hypotheses individuated through their representa-
tional properties. The realist representationalist interpretation fits much better
with scientific practice than do rival instrumentalist or anti-representationalist
interpretations.
Throughout my discussion, I have highlighted foundational questions raised
by the Bayesian paradigm. Which mental processes approximately conform to
Bayesian norms, and which do not? How do nature and nurture jointly influence
priors employed by the mind? How are credal states neurally implemented?
How does the brain transition from one credal state to another? What computa-
tional strategies does it use to approximate intractable Bayesian inferences?
What is it to attach a credence to a hypothesis? Given that hypotheses are sets of
outcomes, what exactly are the outcomes? How literally should we construe an
infinitary Bayesian model built atop an uncountable outcome space? Ongoing
research into these and other foundational questions promises to illuminate
how the representational mind, by approximating rational norms, copes with
perpetual uncertainty.

fðx1 ; x2 ; . . . ; xn Þ: xi 2 R; for all ig:

Recall that R is uncountable, as is R n , and that ½a; b is uncountable whenever

a 6¼ b.

A1 Measurable Spaces
In Kolmogorov’s axiomatization, probabilities attach to sets whose members
are drawn from an outcome space Ω. The powerset of Ω is the set containing all
subsets of Ω. We notate it as PðΩÞ. When Ω is finite, we can assign probabilities
to all members of PðΩÞ. When Ω is uncountable, it is often impossible to assign
intuitively plausible probabilities to all members of the powerset (Proschan &
Shaw, 2016, pp. 17–35). Instead, probability theorists assign probabilities to
certain privileged members of PðΩÞ. The privileged members, called events,
form a σ-field over Ω. A σ-field over Ω is a subset F of PðΩÞ such that:

Ω belongs to F .
If H belongs to F , then H c belongs to F .
If H1 ; H2 ; . . . ; Hn ; . . . belong to F , then their union ∪ Hn also belongs to F .
n

The union ∪ Hn is the set containing all elements that belong to at least one of
n
the sets Hn . There may be a countable inﬁnity of sets Hn .
We typically choose a σ-ﬁeld that arises organically from our interests. For
example, suppose we are modeling an asteroid’s speed using outcome space R.

A natural question is whether the asteroid’s speed x falls in the interval ½a; b.
We would like to assign probabilities to all these intervals. At the very least,
then, our σ-field should contain every interval ½a; b. Consider the minimal
σ-field containing all intervals ½a; b. Call it B. Intuitively: we throw just enough
sets into B to ensure that B contains each interval ½a; b and is closed under
complementation and countable union. B’s members are called the Borel sets.
B usually serves as the most natural σ-field when the outcome space is R.
Similarly, suppose that the outcome space is R 2 , i.e., the set of ordered pairs of
real numbers. Consider the minimal σ-field containing all rectangles. Elements
of this σ-field are again called Borel sets. The same construction generalizes to
R n , for arbitrary n.
An outcome space Ω along with a σ-field F form a measurable space,
typically notated as (Ω, F ).

A2 Probability Measures
We now consider a function P that assigns probabilities to events belonging to
F . For each H 2 F , PðHÞ is the probability assigned to H. As indicated in
Section 2.2, Kolmogorov places three axiomatic constraints on P. Here are the
ﬁrst two axioms:

0 ≤ PðHÞ ≤ 1:

PðΩÞ ¼ 1:

As for the third axiom (additivity), recall my formulation from Section 2.2:

PðH1 ∪ H2 Þ ¼ PðH1 Þ þ PðH2 Þ

when H1 and H2 are disjoint. This formulation is called ﬁnite additivity.

Assuming ﬁnite additivity, one can easily prove:

PðH1 ∪ H2 ∪ . . . ∪ Hn Þ ¼ PðH1 Þ þ PðH2 Þ þ . . . þ PðHn Þ ð5Þ

when H1 ; H2 ; . . . ; Hn is a ﬁnite list of pairwise disjoint events. See Figure 39.

Kolmogorov assumes a stronger axiomatic constraint that generalizes (5) to
a potentially inﬁnite list of pairwise disjoint events H1 ; H2 ; . . . ; Hn ; . . . The
stronger constraint, called countable additivity, demands that:
X
P ∪ Hn ¼ PðHn Þ;
n
n

where ∪ Hn is the countable union of the Hn. Kolmogorov’s axiomatization

n
employs countable additivity as opposed to mere ﬁnite additivity.

Figure 39 H1 ; H2 ; . . . ; H10 are pairwise disjoint events. Finite additivity

requires that their union (the total shaded area) receive a probability equal to the
sums of the probabilities assigned to them individually.

Countable additivity offers an important advantage over mere ﬁnite additivity:

it constrains the probabilities assigned to many more events. Using countable
additivity, we can extrapolate probability assignments from elementary events
(e.g. intervals of real numbers) to numerous complex events left unaddressed by
mere finite additivity (e.g. countable unions of disjoint intervals). Accordingly,
countable additivity is widely assumed within probability theory (Billingsley,
1995). Some versions of Bayesian decision theory employ only finite additivity
(e.g. de Finetti, 1972; Savage, 1972), but most versions assume countable
additivity (e.g. DeGroot, 1970; Easwaran, 2013; Ghosal & van der Vaart, 2017;
Gelman et al., 2014). In a Bayesian context, the dispute between finite and
countable additivity is a normative one. It concerns the norms governing rational
allocation of credence over a hypothesis space. Proponents of countable additiv-
ity claim that rational credences should be countably additive, while opponents
maintain that rational credences need only be finitely additive. For discussion of
finite versus countable additivity in the Bayesian context, see Liu (2020).
When a probability assignment P satisfies all three axioms (including count-
able additivity), it is called a probability measure (or a probability distribution),
and (Ω, F , P) is called a probability space.

A3 Random Variables Deﬁned Rigorously

Let X be a function from Ω to R. To assign probabilities to hypotheses regarding
X’s possible values, we must ensure that our σ-ﬁeld F contains all the hypoth-
eses. For each B 2 B, let

X 1 ðBÞ¼df fω 2 Ω : X ðωÞ 2 Bg:

X is a random variable on the probability space (Ω, F , P) iff

X 1 ðBÞ 2 F , for every B 2 B.

This condition ensures that, for each Borel set B, F contains the hypothesis that
X’s value falls within B. For example, F includes each event X 1 ½a; b. For any
real number x, the event

fω 2 Ω : X ðωÞ ¼ xg

is typically notated as

X ¼ x:

We may write PðX ¼ xÞ for the probability that X has value x (e.g. the probabil-
ity that the asteroid has speed x). Similarly, the event

fω 2 Ω : X ðωÞ 6¼ xg

is typically notated as

X 6¼ x:

We may write PðX 6¼ xÞ for the probability that X does not have value x.
Given a probability space (Ω, F , P) and a random variable X, we can deﬁne
a probability measure μ over the measurable space ðℝ; BÞ:

μðBÞ¼df P X 1 ðBÞ , for every B 2 B.

Figure 38 illustrates, for the special case where B ¼ ½a; b. μ is called X’s
distribution. It is often easier to work with probability measures over ðℝ; BÞ
than with probability measures over (Ω, F ), especially when Ω is complicated.
These definitions generalize from R to R n . The definitions are the same,
except that we consider Borel sets over R n rather than R.
Given a function X from Ω to R and a probability measure μ over ðℝ; BÞ, we
can use X and μ to define a probability space with Ω as the outcome space.
Define σðX Þ, the σ-field generated by X, by

σðX Þ¼df X 1 ðBÞ : B 2 Bg:

Deﬁne a probability measure P over σðX Þ by

P X 1 ðBÞ ¼df μðBÞ, for every B 2 B.

Then (Ω, σðX Þ, P) is a probability space, and X is a random variable deﬁned on

(Ω, σðX Þ, P). This procedure generalizes from R to R n .

A4 Discrete and Nondiscrete Random Variables

A discrete random variable has countably many possible values. Many random
variables encountered in scientiﬁc applications are nondiscrete, that is, they
have uncountably many possible values.
Here is a fundamental constraint on random variables: at most countably
values x of random variable X can receive positive probability. In other words,
there are at most countably many real numbers x such that

PðX ¼ xÞ > 0:

To prove this statement, let us for each natural number n > 0 deﬁne Dn as follows:

Dn ¼ df x 2 R : Pð X ¼ xÞ > 1n :

Suppose for purposes of reductio that Dn has at least n members x1 ; x2 ; . . . ; xn .

The events X ¼ xi and X ¼ xj are disjoint when i 6¼ j. By ﬁnite additivity,

PðX ¼ x1 ∪ X ¼ x2 ∪ . . . ∪ X ¼ xn Þ ¼
PðX ¼ x1 Þ þ PðX ¼ x2 Þ þ . . . þ PðX ¼ xn Þ:

Each individual term PðX ¼ xi Þ is greater than 1=n, so the sum on the right is
greater than
n
¼ 1;
n

which contradicts our axiomatic assumption that 1 is the maximal probability. By

reductio, each set Dn contains fewer than n members. See Figure 40. Using set
theory, one can then show that ∪ Dn is at most countably infinite. Every x such
n
that PðX ¼ xÞ > 0 must belong to some set Dn and hence must belong to ∪ Dn .
n
Therefore, there are at most countably many x such that PðX ¼ xÞ > 0. Note that
our proof uses only finite additivity, with no need for countable additivity.
Many philosophers endorse the doctrine, sometimes called Regularity, that
agents should assign credence 0 only to impossible hypotheses (Kemeny, 1955;
Skyrms, 1995; Stalnaker, 1970). The idea is that, if H is in some sense possible,
then a rational agent will acknowledge its possibility by allocating it at least some
nonzero credence. The foregoing proof shows that Regularity dramatically con-
flicts with the probability calculus axioms, no matter how exactly we gloss
“possibility.” The axioms demand that at most countably many values of
a random variable receive nonzero probability. When a random variable is

Figure 40 D5 contains at most four members x1 ; x2 ; x3 , and x4 . Figure 40

depicts a case where it has exactly four members. Similarly, each set Dn
contains at most n 1 members.

nondiscrete, uncountably many of its possible values must receive probability 0.

This remains so even if one favors finite additivity over countable additivity. In
response, Skyrms (1980) recommends that we preserve Regularity by revising
the probability calculus axioms. The recommendation has not found much uptake
within probability theory or its scientific applications, including Bayesian appli-
cations. Scientific practitioners of the Bayesian framework routinely set
PðX ¼ xÞ ¼ 0 for uncountably many possible values x. So Regularity conflicts
not just with orthodox probability theory but also with scientific practice.
These observations prompt us to reflect upon the meaning of extremal
credences 0 and 1. Let X be a nondiscrete random variable, such as asteroid
speed, and suppose that an agent sets PðX ¼ xÞ ¼ 0. The probability calculus
axioms demand that the agent also set PðX 6¼ xÞ ¼ 1. Certainty in the event
X 6¼ x does not entail that the agent regards value x as impossible. The agent
fully realizes that the asteroid may have speed x. By assigning probability 0 to
speed x, the agent does not completely reject the possibility of speed x. She
merely regards this possibility as so negligible that it merits no positive cre-
dence. Assuming the agent’s credences conform to the probability calculus
axioms, she must similarly regard uncountably many other values of X as
negligible possibilities.
A random variable is said to be continuous when PðX ¼ xÞ ¼ 0 for all x.
A continuous random variable violates Regularity in a very extreme way: every
event X ¼ x receives probability 0. Note that some random variables are neither
discrete nor continuous (Billingsley, 1995, pp. 257–258): such a variable has
uncountably many possible values x, and certain values receive positive
probability.

A5 Probability Density Functions

A probability density function (pdf) is a nonnegative function pðxÞ from R to R
such that the area under the curve is 1:
ð∞
pðxÞdx ¼ 1:
∞

A pdf induces a probability measure μ over ðℝ; BÞ. The probability assigned by
μ to ½a; b is the area under pðxÞ stretching from a to b:
ðb
μð½a; bÞ ¼ pðxÞdx:
a

Probability assignments to the intervals ½a; b determine unique probability

assignments to all Borel sets. Thus, each pdf induces a unique probability
measure μ.
Because probability density determines probability via integration, changes
to the pdf do not affect probabilities when they do not affect integration.
Compare Figure 9 with Figure 41. These are two different pdfs: they assign
different densities to c. Nevertheless, they induce the same probabilities,
because a change in density at a single point does not affect integration. More

Figure 41 This pdf alters the pdf from Figure 9 at a single point c. The alteration
does not affect the area under the curve, so the two pdfs determine the same
probability distribution.

Figure 42 Typology of random variables. A random variable is either discrete or

nondiscrete. A continuous random variable is nondiscrete; the converse is not
true. A random variable is continuous if its distribution is induced by a pdf; the
converse is not true.
generally: when pdf pðxÞ induces probability measure μ, there are infinitely
many distinct pdfs that induce the same measure μ.
Suppose random variable X is defined on probability space (Ω, F , P) with
distribution μ. Suppose μ is induced by a pdf pðxÞ. It is not hard to show that X is
a continuous random variable: PðX ¼ xÞ ¼ 0 for all x (Billingsley, 1995,
p. 212). Equivalently, μðfxgÞ ¼ 0 for all x. The converse is not true: in some
cases, the distribution of a continuous random variable is not induced by any pdf
(Proschan & Shaw, 2016, pp. 94–95). See Figure 42.
Given a random variable X defined on probability space (Ω, F , P), define
a new random variable Y resulting from multiplication by a constant k:

Y ðωÞ ¼ kX ðωÞ:

Suppose that X’s distribution has a pdf pðxÞ. One can show that Y’s distribution
has a pdf qðyÞ given by

pðy=kÞ
qðyÞ ¼ : ð6Þ
k

See Ma, Kording & Goldreich (2023, pp. 333–336). The change in variable
(from X to Y) necessitates a change in pdf.

To illustrate, let Ω contain possible speeds for an object. Each outcome ω is

a particular speed that the object might have. Assume that Ω is endowed with
an appropriate σ-field F . Assume also an underlying probability measure
P defined on F . Speeds are physical magnitudes and hence are distinct from
real numbers (Peacocke, 2019). We can describe physical magnitudes with
real numbers by choosing measurement units, such as meters/sec or feet/sec.
The first choice of measurement unit corresponds to one random variable
X from Ω to R. The second choice corresponds to a second random variable
Y from Ω to R, where

Y ðωÞ ¼ 3:28084 X ðωÞ:

The underlying probability measure P induces different distributions for X and

Y. If the pdf for X is given by pðxÞ, then the pdf for Y is given by (6), taking
k ¼ 3:28084. Figure 37 illustrates. The blue pdf corresponds to meters/sec. The
orange pdf corresponds to feet/sec. The two pdfs are associated with the same
underlying probability measure over possible speeds.

A6 Conditional Density and Beyond

Using conditional densities, we can extend the notion of conditional probability
well beyond the elementary case where the ratio formula prevails.
Suppose we are given a two-dimensional pdf pðx; yÞ. We want to deﬁne
a new pdf over y conditional on X having value a. So we want to deﬁne a one-
dimensional conditional density over y, which we may notate as:

pðy j X ¼ aÞ:

Intuitively: this is a density over y given that X has value a. To deﬁne

pðy j X ¼ aÞ, we conﬁne attention to points such that X ¼ a. We consider p’s
values on those points alone:

pða; yÞ:

One might hope to set the conditional density pðy j X ¼ aÞ equal to pða; yÞ,
where we hold a ﬁxed and allow y to vary. The only hitch is that pða; yÞ, viewed
as a function of y, may not be a pdf: the area under the curve may not be 1. We
must settle for proportionality rather than equality:

pðy j X ¼ aÞ ∝ pða; yÞ:

Intuitively, pðy j X ¼ aÞ conﬁnes attention to outcomes where X ¼ a and then

allocates probability density in proportion to the original density function
pða; yÞ. To obtain pðy j X ¼ aÞ from pða; yÞ, we need merely divide pða; yÞ by

a constant to ensure that the area under the curve is 1. This constant is called
a normalization constant.
More formally, we may deﬁne conditional density as follows. Take pðx; yÞ as
given and deﬁne pðxÞ, the marginal pdf for X:
ð∞
pðxÞ¼df pðx; yÞdy:
∞

pðxÞ is computed by holding x ﬁxed and integrating pðx; yÞ over all possible
values of y. Assuming that pðxÞ > 0, we may deﬁne the conditional density of
Y given X ¼ x by the equation

pðx; yÞ
pðy j X ¼ xÞ¼df : ð7Þ
pðxÞ

pðxÞ is the normalization constant: it ensures that probabilities sum to 1. When it is

clear which random variable X is at issue, we may notate (7) more compactly as

pðx; yÞ
pðy j xÞ ¼ : ð8Þ
pðxÞ

pðy j xÞ results from pðx; yÞ by holding x fixed and then normalizing. See
Figures 18, 19, 20, and 21. These definitions generalize to higher dimensions.16
It is often most natural to regard pðxÞ and pðy j xÞ as primitive rather than
defined. For example, pðxÞ might be a pdf for asteroid speed and pðy j xÞ might
be the conditional density of measuring speed y given that the asteroid has
speed x. Taken together, pðxÞ and pðy j xÞ determine a joint density pðx; yÞ: we
simply view equation (8) as a definition of pðx; yÞ rather than of pðy j xÞ. In
practice, we need not usually consider the joint density. It lies in the back-
ground of our theorizing, but we only explicitly consider pðxÞ, pðy j xÞ, and
pðx j yÞ.
The ratio formula and the theory of conditional densities suffice for most
applications of Bayesian decision theory. However, there are situations where
we would like to define conditional probabilities yet neither the ratio formula
nor the theory of conditional densities applies. To illustrate with a cognitive

16
This paragraph glosses over some major philosophical and mathematical complications. Due to
a phenomenon known as the Borel‒Kolmogorov paradox, we cannot condition directly on X ¼ x
when PðX ¼ xÞ ¼ 0 (Kolmogorov, 1933/1956). We instead condition on X ¼ x considered as
embedded within the σ-field σðXÞ. If we were to consider X ¼ x as embedded within a different
σ-field, then different conditional probabilities might result. The notation pðy j X ¼ xÞ is rather
misleading because it elides this relativity to an embedding σ-field. Similarly for the notation
pðy j xÞ. See Rescorla (2015c) for discussion of the Borel‒Kolmogorov paradox and its
ramifications.

science example, consider the Bayesian causal inference model given by

Kording et al. (2007). The model evaluates whether visual input eV and auditory
input eA derive from a single distal source. C is a binary random variable that
registers the number of sources: C ¼ 1 registers a single source, and C ¼ 2
registers two distinct sources. Upon receiving inputs eV and eA , the model
computes the posterior probability PðC ¼ 1 j eV ; eA Þ that those inputs derive
from a single distal source. Assume that there are uncountably many possible
inputs, as Kording et al. (2007) do and as is standard in Bayesian perceptual
psychology. Then we cannot deﬁne conditional probabilities PðC ¼ 1 j eV ; eA Þ
using either the ratio formula or conditional densities. The ratio formula does
not apply because there is probability zero of any given input pair ðeV ; eA Þ,
except perhaps for countably many such pairs. Nor does the theory of condi-
tional densities apply: C is discrete, so no joint density exists. As this example
illustrates, a general theory of conditional probability must look beyond both
the ratio formula and conditional densities.
The most successful general theory traces back to the same treatise where
Kolmogorov (1933/1956) codiﬁed the probability calculus axioms. The
ratio formula and the theory of conditional densities are special cases
of Kolmogorov’s theory (Billingsley, 1995, p. 432; Rescorla, 2015c).
Kolmogorov’s theory is general enough to handle the Bayesian causal infer-
ence model, along with countless other applications. Perhaps because
Kolmogorov’s theory is forbiddingly technical, it was long neglected by the
philosophical community. Recently, it has begun to receive sympathetic atten-
tion from philosophers (Easwaran, 2011b; Huttegger, 2015; Meehan & Zhang,
2020; Nielsen, 2021; Rescorla, 2018a; Rescorla, forthcoming). Easwaran
(2019) gives a detailed introduction, with comparisons to alternative theories
of conditional probability.

A7 Proof of Bayes’s Theorem

Suppose that PðHÞ > 0 and PðEÞ > 0. The ratio formula determines condi-
tional probabilities PðH j EÞ and PðE j HÞ:

PðH ∩ EÞ
PðH j EÞ ¼
PðEÞ

PðE ∩ HÞ
PðE j HÞ ¼ :
PðHÞ

Algebraic manipulation yields

PðH j EÞPðEÞ ¼ PðH ∩ EÞ ¼ PðE ∩ HÞ ¼ PðE j HÞPðHÞ;

which immediately entails Bayes’s Theorem:

PðHÞPðE j HÞ
PðH j EÞ ¼ : ð9Þ
PðEÞ

It is remarkable that this theorem follows almost trivially from the ratio formula
yet offers such profound insight into rational inference.
Now consider the case where we have a two-dimensional pdf pðx; yÞ. Deﬁne
conditional densities and marginals as in Section A6:

ð∞
pðxÞ¼df pðx; yÞdy
∞

ð∞
pðyÞ¼df pðx; yÞdx
∞

pðx; yÞ
pðy j xÞ¼df
pðxÞ

pðx; yÞ
pðx j yÞ¼df ;
pðyÞ

where the third deﬁnition presupposes pðxÞ > 0 and the fourth presupposes
pðyÞ > 0. From these latter two deﬁnitions,

pðx j yÞpðyÞ ¼ pðx; yÞ ¼ pðy j xÞpðxÞ:

By algebra,

pðy j xÞpðxÞ
pðx j yÞ ¼ ; ð10Þ
pðyÞ

which is Bayes’s theorem for pdfs. Note that 1=pðyÞ does not depend upon x. It
ﬁgures solely as a normalization constant. Although (9) and (10) look similar
and have similar proofs, they are distinct: (9) concerns conditional probabilities,
while (10) concerns conditional densities.
Bayes’s theorem generalizes beyond the formulations given here, using
Kolmogorov’s theory of conditional probability (Ghosal & van der Vaart, 2017,
p. 7). There are also some situations where no analogue to Bayes’s theorem is
available (Ghosal & van der Vaart, 2017, pp. 7–8). In those situations, one can still
conform to Conditionalization: one can respond to new evidence by replacing the
prior with the posterior. Unfortunately, one can no longer use anything like (9) or
(10) to compute the posterior.

Downloaded from https://www.cambridge.org/core. IP address: 201.224.145.68, on 01 Feb 2025 at 14:48:00, subject to the Cambridge Core terms of
use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108955973
References
Abend, O., Kwiatkowski, T., Smith, N., Goldwater, S. & Steedman, S. (2017).
Bootstrapping language acquisition. Cognition, 164, 116–143.
Ackerman, N., Freer, C. & Roy, D. (2019). On the computability of conditional
probability. Journal of the ACM, 66, 1–40.
Adams, W., Graf, E. & Ernst, M. (2004). Experience can change the “light-
from-above” prior. Nature Neuroscience, 7, 1057–1058.
Aitchison, L. & Lengyel, M. (2017). With or without you: predictive coding
and Bayesian inference in the brain. Current Opinion in Neurobiology,
46, 219–227.
Ashby, D. (2006). Bayesian statistics in medicine: a 25 year review. Statistics in
Medicine, 25, 3589–3631.
Baker, C. & Tenenbaum, J. (2014). Modeling human plan recognition using
Bayesian theory of mind. In G. Sukthankar, R. P. Goldman, C. Geib,
D. Pynadath & H. Bui, eds., Plan, Activity, and Intent Recognition:
Theory and Practice. Waltham: Morgan Kaufmann, pp. 177–204.
Battaglia, P. W., Hamrick, J. & Tenenbaum, J. (2013). Simulation as an engine
of physical scene understanding. Proceedings of the National Academy of
Sciences, 110, 18327–18332.
Bayes, T. (1763). An essay towards solving a problem in the doctrine of chances.
Philosophical Transactions of the Royal Society of London, 53, 470–418.
Beck, J., Ma, W. J., Latham, P. E. & Pouget, A. (2007). Probabilistic population
codes and the exponential family of distributions. In P. Cisek, T. Drew &
J. F. Kalaska, eds., Computational Neuroscience: Theoretical Insights into
Brain Function. New York: Elsevier.
Berger, J. (1985). Statistical Decision Theory: Foundations, Concepts, and
Methods, 2nd ed., New York: Springer.
Berniker, M., & Voss, M. (2010). Learning priors for Bayesian computations in
the nervous system. PLOS One, 10, e12686.
Bernstein, N. (1967). The Coordination and Regulation of Movements. Oxford:
Pergamon.
Billingsley, P. (1995). Probability and Measure. 3rd ed. New York: Wiley.
Block, N. (2018). If perception is probabilistic, why does it not seem
probabilistic? Philosophical Transactions of the Royal Society B, 373,
20170341.
Block, N. (2023). The Border Between Perception and Cognition. Cambridge,
MA: MIT Press.

Bowers, J. & Davis, C. (2012). Bayesian just-so stories in psychology and

neuroscience. Psychological Bulletin, 138, 389–414.
Brooks, S., Gelman, A., Jones, G. & Meng, X.-L., eds. (2011). Handbook of
Markov Chain Monte Carlo. New York: CRC Press.
Burge, T. (2010). Origins of Objectivity. Oxford: Oxford University Press.
Burge, T. (2022). Perception: First Form of Mind. Oxford: Oxford University
Press.
Burns, J. & Blohm, G. (2010). Multi-sensory weights depend on contextual noise
in reference frame transformations. Frontiers in Human Neuroscience,
4, 1–16.
Cao, R. (2020). New labels for old ideas: predictive processing and the inter-
pretation of neural signals. Review of Philosophy and Psychology, 11,
517–546.
Carey, S. (2009). The Origin of Concepts. Oxford: Oxford University Press.
Chalmers, D. (2011). Frege’s puzzle and the objects of credence. Mind, 120,
587–635.
Chater, N. & Oaksford, M. eds. (2008). The Probabilistic Mind. Oxford: Oxford
University Press.
Chater, N., Zhu, J.-Q., Spicer, J., Sundh, J., León-Villagrá, P. & Sanborn, A.
(2020). Probabilistic biases meet the Bayesian brain. Current Directions in
Psychological Science, 29, 506–512.
Chemero, A. (2009). Radical Embodied Cognitive Science. Cambridge, MA:
MIT Press.
Chen, X., McNamara, T., Kelly, J. & Wolbers, T. (2017). Cue combination in
human spatial navigation. Cognitive Psychology, 95, 105–144.
Cheyette, S. & Piantadosi, S. (2017). Knowledge transfer in a probabilistic
language of thought. Proceedings of the Annual Meeting of the Cognitive
Science Society, 39, 222–227.
Churchland, P. (1981). Eliminative materialism and the propositional attitudes.
Journal of Philosophy, 78, 67–90.
Clark, A. (2015). Surﬁng Uncertainty. Oxford: Oxford University Press.
Colombo, M. & Seriès, P. (2012). Bayes in the brain—on Bayesian modeling
in neuroscience. The British Journal for the Philosophy of Science, 63,
697–723.
Dasgupta, I., Schulz, E. & Gershman, S. (2017). Where do hypotheses come
from? Cognitive Psychology, 96, 1–25.
de Finetti, B. (1937/1980). Foresight. Its logical laws, its subjective sources.
Rpt. in H. E. Kyburg, Jr. & H. E. Smokler, eds., Studies in Subjective
Probability. Huntington, NY: Robert E. Krieger, pp. 53–118.

de Finetti, B. (1972). Probability, Induction, and Statistics. New York: Wiley.

DeGroot, M. (1970). Optimal Statistical Decisions. New York: McGraw-Hill.
Dener, E., Kacelnik, A. & Shemesh, H. (2016). Pea plants show risk sensitivity.
Current Biology, 26, 1763–1767.
Dennett, D. (1987). The Intentional Stance. Cambridge, MA: MIT Press.
Duffy, S., Huttenlocher, J., Hedges, L. & Crawford, L. E. (2010). Category effects
on stimulus estimation: shifting and skewed frequency distributions.
Psychonomic Bulletin and Review, 17, 224–230.
Earman, J. (1992). Bayes or Bust? Cambridge, MA: MIT Press.
Easwaran, K. (2011a). Bayesianism I: introduction and arguments in favor.
Philosophy Compass, 6, 312–320.
Easwaran, K. (2011b). The varieties of conditional probability. In P. Bandyo-
padhyay & M. Forster, eds., Philosophy of Statistics. Amsterdam: Elsevier,
pp. 137–148.
Easwaran, K. (2013). Why countable additivity? Thought, 2, 53–61.
Easwaran, K. (2019). Conditional probabilities. In R. Pettigrew & J. Weisberg,
eds., The Open Handbook of Formal Epistemology. London, ON: The
PhilPapers Foundation, pp. 131–198.
Eberhardt, F. & Danks, D. (2011). Conﬁrmation in the cognitive sciences: the
problematic case of Bayesian models. Minds and Machines, 21, 389–410.
Egan, F. (2010). Computational models: a modest role for content. Studies in
History and Philosophy of Science, 41, 253–259.
Egan, F. (2020). A deﬂationary account of mental representation. In
J. Smorthchkova, T. Schlicht & K. Dolega, eds., What Are Mental
Representations? Oxford: Oxford University Press, pp. 26‒53.
Erikkson, L. & Hájek, A. (2007). What are degrees of belief? Studia Logica, 86,
183–213.
Ernst, M. (2006). A Bayesian view on multimodal cue integration. In G. Knoblich,
I. Thornton, M. Grosjean & M. Shiffrar, eds., Human Body Perception From
the Inside Out. Oxford: Oxford University Press, pp. 105‒132.
Ernst, M. (2007). Learning to integrate arbitrary signals from vision and touch.
Journal of Vision, 7, 1–14.
Ernst, M. & Banks, M. (2002). Humans integrate visual and haptic information
in a statistically optimal fashion. Nature, 415, 429–433.
Fernandes, H., Stevenson, I., Vilares, I. & Kording, K. (2014). The generaliza-
tion of prior uncertainty during reaching. The Journal of Neuroscience, 34,
11470–11484.
Fiehler, K., Wolf, C., Klinghammer, M. & Blohm, G. (2014). Integration of
egocentric and allocentric information during memory-guided reaching to
images of a natural environment. Frontiers in Human Neuroscience, 8, 1–12.

Field, H. (2001). Truth and the Absence of Fact. Oxford: Clarendon Press.
Fiser, J., Berkes, P., Orbán, G. & Lengyel, M. (2010). Statistically optimal
perception and learning: from behavior to neural representations. Trends in
Cognitive Sciences, 14, 119–130.
Fodor, J. (1975). The Language of Thought. New York: Thomas Y. Crowell.
Fodor, J. (1987). Psychosemantics. Cambridge, MA: MIT Press.
Fodor, J. (2008). LOT2. Oxford: Clarendon Press.
Fodor, J. & Pylyshyn, Z. (1981). How direct is visual perception? Some reflec-
tions on Gibson’s “ecological approach.” Cognition, 9, 139–196.
Frege, G. (1892/1997). On Sinn and Bedeutung. Rpt. in M. Beaney, ed., The
Frege Reader, trans. M. Black. Malden, MA: Blackwell.
Fristedt, B. & Gray, L. (1997). A Modern Approach to Probability Theory.
Boston: Birkhäuser.
Friston, K. (2010). The free-energy principle: a unified brain theory? Nature
Reviews Neuroscience, 11, 127–138.
Gaifman, H. & Snir, M. (1982). Probabilities over rich languages, testing, and
randomness. The Journal of Symbolic Logic, 47, 495–548.
Gallistel, C. R. (1990). The Organization of Learning. Cambridge, MA: MIT
Press.
Gallistel, C. R. (2008). Dead reckoning, cognitive maps, animal navigation, and
the representation of space: an introduction. In M. Jeffries & W.-K. Yeap,
eds., Robotics and Cognitive Approaches to Spatial Mapping. Berlin:
Springer, pp. 137‒143.
Gallistel, C. R. & King, A. (2009). Memory and the Computational Brain.
Malden, MA: Wiley-Blackwell.
Ganguli, D. & Simoncelli, E. (2014). Efficient sensory encoding and Bayesian
inference with heterogeneous neural populations. Neural Computation,
26, 2103–2134.
Gardner, J. (2019). Optimality and heuristics in perceptual neuroscience. Nature
Neuroscience, 22, 514–523.
Gelman, A., Carlin, J., Stern, H., Dunson, D., Vehatri, A. & Rubin, D. (2014).
Bayesian Data Analysis, 3rd ed. New York: CRC Press.
Ghosal, S. & van der Vaart, A. (2017). Fundamentals of Nonparametric
Bayesian Inference. Cambridge: Cambridge University Press.
Gibson, J. J. (1979). The Ecological Approach to Visual Perception. Boston:
Houghton Mifflin.
Glasauer, S. (2019). Sequential Bayesian updating as a model for human
perception. Progress in Brain Research, 249, 3–18.
Glasauer, S. & Shi, Z. (2022). Individual beliefs about temporal continuity
explain variation of perceptual biases. Scientific Reports, 12, 10746.

Goodman, N., Tenenbaum, J., Feldman, J. & Grifﬁths, T. (2008). A rational

analysis of rule-based concept learning. Cognitive Science, 32, 108–154.
Goodman, N., Ullman, T. & Tenenbaum, J. (2011). Learning a theory of
causality. Psychological Review, 118, 110–199.
Greaves, H. & Wallace, D. (2006). Justifying conditionalization: conditionali-
zation maximizes expected epistemic utility. Mind, 115, 607–632.
Griffiths, T., Kemp, C. and Tenenbaum, J. (2008). Bayesian models of cognition.
In R. Sun, ed., The Cambridge Handbook of Computational Psychology.
Cambridge: Cambridge University Press, pp. 59‒100.
Griffiths, T. & Tenenbaum, J. (2009). Theory-based causal induction.
Psychological Review, 116, 661–716.
Hacking, I. (2001). An Introduction to Probability and Inductive Logic.
Cambridge: Cambridge University Press.
Haith, A. & Krakauer, J. (2013). Theoretical models of motor control and motor
learning. In M. Richardson, M. Riley, and K. Shockley, eds., Progress in
Motor Control VII: Neural Computational and Dynamic Approaches.
New York: Springer, pp. 7‒28.
Helbig, H. & Ernst, M. (2008). Haptic perception in interaction with other
senses. In M. Grunwald, ed., Human Haptic Perception: Basics and
Applications. Boston: Birkhäuser Verlag, pp. 235‒249.
Helmholtz, H. von. (1867/1925). Treatise on Physiological Optics, trans. and
ed. J. Southall. Manasha, WI: George Banta Publishing Company.
Hemmer, P. & Steyvers, M. (2009). A Bayesian account of reconstructive
memory. Topics in Cognitive Science, 1, 189–202.
Hillis, J., Ernst, M., Banks, M. & Landy, M. (2002). Combining sensory
information: mandatory fusion within, but not between, senses. Science,
298, 1627–1630.
Hohwy, J. (2014). The Predictive Mind. Oxford: Oxford University Press.
Hollingworth, H. (1910). The central tendency of judgment. The Journal of
Philosophy, Psychology, and Scientific Methods, 7, 461–469.
Huttegger, S. (2015). Merging of opinions and probability kinematics. The
Review of Symbolic Logic, 8, 611–648.
Hutto, D. & Myin, E. (2017). Evolving Enactivism: Basic Minds Meet Content.
Cambridge, MA: MIT Press.
Icard, T. (2016). Subjective probability as sampling propensity. The Review of
Philosophy and Psychology, 7, 863–903.
Jazayeri, M. & Shadlen, M. (2010). Temporal context calibrates interval timing.
Nature Neuroscience, 13, 1020–1026.
Jetzschke, S., Ernst, M., Froehlich, J. & Boeddeker, N. (2014). Finding home:
landmark ambiguity in human navigation. Frontiers in Behavioral
Neuroscience, 11, 1–15.
Downloaded from https://www.cambridge.org/core. IP address: 201.224.145.68, on 01 Feb 2025 at 14:48:00, subject to the Cambridge Core terms of
use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/9781108955973
References 93

Jones, M. & Love, B. (2011). Bayesian fundamentalism or enlightenment? On

the explanatory status and theoretical contribution of Bayesian models of
cognition. Behavioral and Brain Sciences, 34, 169–188.
Kahneman, D. & Tversky, A. (1979). Prospect theory: an analysis of decision
under risk. Econometrica, 47, 263–291.
Katz, Y., Goodman, N., Kersting, K., Kemp, C. & Tenenbaum, J. (2008).
Modeling semantic cognition as logical dimensionality reduction.
Proceedings of the Annual Meeting of the Cognitive Science Society, 30.
https://escholarship.org/uc/item/50r1c7qh.
Kemeny, J. (1955). Fair bets and inductive probabilities. The Journal of
Symbolic Logic, 20, 263–273.
Kessler, F., Frankenstein, J. & Rothkopf, C. (2024). Human navigation strat-
egies and their errors result from dynamic interactions of spatial uncertain-
ties. Nature Communications, 15, 5677.
Knill, D. & Richards, W., eds. (1996). Perception as Bayesian inference.
Cambridge: Cambridge University Press.
Kolmogorov, A. N. (1933/1956). Foundations of the Theory of Probability. 2nd
English ed, trans. N. Morrison. New York: Chelsea.
Kording, K., Beierholm, U., Ma, W. J., Quartz, S., Tenenbaum J. & Shams L.
(2007). Causal inference in multisensory perception. PLoS One, 2, e943.
Kording, K. & Wolpert, D. (2004). Bayesian integration in sensorimotor
learning. Nature, 427, 244–247.
Kwisthout, J., Wareham, T. & van Rooij, I. (2011). Bayesian intractability is not
an ailment that approximation can cure. Cognitive Science, 35, 779–784.
Kwon, O.-S. & Knill, D. (2013). The brain uses adaptive internal models of
scene statistics for sensorimotor estimation and planning. Proceedings of
the National Academy of Sciences, 110, E1064‒E1073.
Kwon, O.-S., Tadin, D. & Knill, D. (2015). Unifying account of visual motion
and position perception. Proceedings of the National Academy of Sciences,
112, 8142–8147.
Lakshminarasimhan, K., Petsalis, M., Park, H., DeAngelis, G., Pitkow, X. &
Angelaki, D. (2018). A dynamic Bayesian observer model reveals origins
of bias in visual path integration. Neuron, 99, 194–206.
Laplace, P.-S. (1814/1902). A Philosophical Essay on Probabilities, trans.
F. Truscott & F. Emory. New York: Wiley.
Lappe, M., Stiels, M., Frenz, H. & Loomis, J. (2011). Keeping track of the
distance from home by leaking integration along veering paths.
Experimental Brain Research, 212, 81–89.
Lee, T. S. & Mumford, D. (2003). Hierarchical Bayesian inference in the visual
cortex. Journal of the Optical Society of America, 20, 1434–1448.

Lewis, D. (1999). Why conditionalize? In Papers in Metaphysics and

Epistemology. Cambridge: Cambridge University Press, pp. 403‒407.
Lieder, F., Grifﬁths, T., Huys, Q. & Goodman, N. (2018). The anchoring bias
reﬂects rational use of cognitive resources. Psychonomic Bulletin and
Review, 25, 322–349.
Liu, Y. (2020). Countable additivity, idealization, and conceptual realism.
Economics and Philosophy, 36, 127–147.
McGrayne, S. (2011). The Theory that Would Not Die. New Haven, CT: Yale
University Press.
McNamee, D. & Wolpert, D. (2019). Internal models in biological control.
Annual Review of Control, Robotics, and Autonomous Systems, 2, 339–364.
Ma, W. J. (2019). Bayesian decision models: a primer. Neuron, 104, 164–175.
Ma, W. J., Beck, J., Latham, P. & Pouget, A. (2006). Bayesian inference with
probabilistic population codes. Nature Neuroscience, 9, 1432–1438.
Ma, W. J., Kording, K. & Goldreich, D. (2023). Bayesian Models of Perception
and Action: An Introduction. Cambridge, MA: MIT Press.
Mahtani, A. (2024). The Objects of Credence. Oxford: Oxford University Press.
Makin, J., Fellows, M. & Sabes, P. (2013). Learning multisensory integration and
coordinate transformation via density estimation. PLOS Computational
Biology, 9, e1003035.
Mamassian, P., Landy, M. & Maloney, L. (2002). Bayesian modeling of visual
perception. In R. Rao, B. Olshausen & M. Lewicki, eds., Probabilistic
Models of the Brain. Cambridge, MA: MIT Press, pp. 13‒36.
Mandelbaum, E. (2019). Troubles with Bayesianism: an introduction to the
psychological immune system. Mind and Language 34: 141–157.
Mandelbaum, E., Won, I., Gross, S. & Firestone, C. (2020). Can resources save
rationality? “Anti-Bayesian” updating in cognition and perception.
Behavioral and Brain Sciences, 43, 31–32.
Meehan, A. & Zhang, S. (2020). Jeffrey meets Kolmogorov: a general theory of
conditioning. Journal of Philosophical Logic, 49, 941–979.
Morales, J., Solovey, G., Maniscalco, B., Rahnev, D., de Lange, F. & Lau, H.
(2015). Low attention impairs optimal incorporation of prior knowledge
in perceptual decisions. Attention, Perception, and Psychophysics, 77,
2021–2036.
Murphy, K. (2023). Probabilistic machine learning: advanced topics.
Cambridge, MA: MIT Press.
Narayanan, S. & Jurafsky, D. (1998). Bayesian models of human sentence
processing. In Proceedings of the Annual Meeting of the Cognitive
Science Society, 20, 752–757.

Nashed, J., Crevecoeur, F. & Scott, S. (2012). Inﬂuence of the behavioral goal
and environmental obstacles on rapid feedback responses. Journal of
Neurophysiology, 108, 999–1009.
Nielsen, M. (2021). A new argument for Kolmogorov Conditionalization. The
Review of Symbolic Logic, 14, 930–945.
Norris, D. (2006). The Bayesian reader: explaining word recognition as an
optimal Bayesian decision process. Psychological Review, 113, 327–357.
Oaksford, M. & Chater, N. (2007). Bayesian Rationality: The Probabilistic
Approach to Human Reasoning. Oxford: Oxford University Press.
Oaksford, M. & Chater, N. (2020). New paradigms in the psychology of
reasoning. Annual Review of Psychology, 71, 305–330.
Olkkonen, M., McCarthy, P. & Allred, S. (2014). The central tendency bias in
color perception: effects of internal and external noise. Journal of Vision,
14, 1–15.
Orbán, G., Berkes, P., Fiser, J. & Lengyel, M. (2016). Neural variability and
sampling-based probabilistic representations in the visual cortex. Neuron,
92, 530–542.
Orlandi, N. (2014). The Innocent Eye: Why Vision is not Cognitive Process.
Oxford: Oxford University Press.
Peacocke, C. (1994). Content, computation, and externalism. Mind and
Language, 9, 303–335.
Peacocke, C. (1999). Computation as involving content: a response to Egan.
Mind and Language, 14, 195–202.
Peacocke, C. (2019). The Primacy of Metaphysics. Oxford: Oxford University
Press.
Peters, M., Ma, W. J. & Shams, L. (2016). The size-weight illusion is not
anti-Bayesian after all. PeerJ, 4, e2124.
Pettigrew, R. (2019). Epistemic utility arguments for probabilism. In E. Zalta,
ed., The Stanford Encyclopedia of Philosophy (Winter 2019). https://plato
.stanford.edu/archives/win2019/entries/epistemic-utility.
Pettigrew, R. (2020). Dutch Book Arguments. Cambridge: Cambridge
University Press.
Petzschner, F. & Glasauer, S. (2011). Iterative Bayesian estimation as an
explanation for range and regression effects: a study on human path
integration. The Journal of Neuroscience, 31, 17220–17229.
Petzschner, F., Glasauer, S. & Stephan, K. (2015). A Bayesian perspective on
magnitude estimation. Trends in Cognitive Sciences, 19, 285–293.
Piantadosi, S. & Jacobs, R. (2016). Four problems solved by the probabilistic
language of thought. Current Directions in Psychological Science, 25, 54–59.

Pouget, A., Beck, J., Ma., W. J. & Latham, P. (2013). Probabilistic brains:
knowns and unknowns. Nature Neuroscience, 16, 1170–1178.
Proschan, M. & Shaw, P. (2016). Essentials of Probability Theory for
Statisticians. New York: CRC Press.
Putnam, H. (1975). Mathematics, Matter, and Method: Philosophical Papers,
vol. 1. Cambridge: Cambridge University Press.
Pylyshyn, Z. (1984). Computation and Cognition. Cambridge, MA: MIT Press.
Quine, W. V. (1960). Word and Object. Cambridge, MA: MIT Press.
Rahnev, D. & Denisov, R. (2018). Suboptimality in perceptual decision making.
Behavioral and Brain Sciences, 41, e223.
Ramsey, F. P. (1931). Truth and probability. In R. Braithwaite, ed., The
Foundations of Mathematics and Other Logical Essays. London: Kegan,
Paul, Trench, Trubner & Co., pp.156‒198.
Ramsey, W. (2007). Representation Reconsidered. Cambridge: Cambridge
University Press.
Rao, R. and Ballard, D. (1999). Predictive coding in the visual cortex: a
functional interpretation of some extra-classical receptive-field effects.
Nature Neuroscience, 2, 79–87.
Rescorla, M. (2009). Cognitive maps and the language of thought. The British
Journal for the Philosophy of Science, 60, 377–407.
Rescorla, M. (2015a). Bayesian perceptual psychology. In M. Matthen, ed., The
Oxford Handbook of the Philosophy of Perception. Oxford: Oxford
University Press, pp. 694‒716.
Rescorla, M. (2015b). Review of Nico Orlandi’s The Innocent Eye: Why Vision
is not a Cognitive Process. Notre Dame Philosophical Reviews, January
2015.
Rescorla, M. (2015c). Some epistemological ramifications of the Borel‒
Kolmogorov paradox. Synthese, 192, 735–767.
Rescorla, M. (2016). Bayesian sensorimotor psychology. Mind and Language,
31, 3–36.
Rescorla, M. (2017). Review of Andy Clark’s Surfing Uncertainty. Notre Dame
Philosophical Reviews, January 2017.
Rescorla, M. (2018a). A Dutch book theorem and converse Dutch book theorem
for Kolmogorov Conditionalization. The Review of Symbolic Logic, 11,
705–735.
Rescorla, M. (2018b). An interventionist approach to psychological
explanation. Synthese, 195, 1909–1940.
Rescorla, M. (2019). Motor computation. In M. Colombo & M. Sprevak, eds.,
The Routledge Handbook of the Computational Mind. New York:
Routledge, pp. 424‒435.

Rescorla, M. (2020a). The computational theory of mind. In E. Zalta, ed. The

Stanford Encyclopedia of Philosophy. https://plato.stanford.edu/entries/
computational-mind.
Rescorla, M. (2020b). Perceptual co-reference. The Review of Philosophy and
Psychology, 11, 569–589.
Rescorla, M. (2020c). A realist perspective on Bayesian cognitive science. In
A. Nes & T. Chan, eds., Inference and Consciousness. New York:
Routledge, pp. 40‒73.
Rescorla, M. (2020d). Reifying representations. In J. Smorthchkova, T. Schlicht
& K. Dolega, eds., What Are Mental Representations? Oxford: Oxford
University Press, pp. 135‒177.
Rescorla, M. (2021a). Bayesian modeling of the mind: from norm to neurons.
WIREs Cognitive Science, 12, e1540.
Rescorla, M. (2021b). On the proper formulation of Conditionalization.
Synthese, 198, 1935–1965.
Rescorla, M. (2022). An improved Dutch book theorem for Conditionalization.
Erkenntnis, 87, 1013–1041.
Rescorla, M. (2024). Neural implementation of (approximate) Bayesian infer-
ence. In T. Cheng, R. Sato & J. Hohwy, eds., Expected Experiences: The
Predictive Mind in an Uncertain World. Routledge, pp. 197‒239.
Rescorla, M. (Forthcoming). Non-factive Kolmogorov conditionalization. The
Review of Symbolic Logic. https://doi.org/10.1017/S1755020323000345.
Rittenhouse, D. (1786). Explanation of an optical deception. Transactions of the
American Philosophical Society, 2, 37–42.
Sanborn, A., Griffiths, T. & Navarro, D. (2010). Rational approximations to
rational models: alternative algorithms for category learning. Psychological
Review, 117, 1144–1167.
Sanborn, A., Masinghka, J. & Griffiths, T. (2013). Reconciling intuitive physics
and Newtonian mechanics for colliding objects. Psychological Review,
120, 411–437.
Sato, Y. & Kording, K. (2014). How much to trust the senses: likelihood
learning. Journal of Vision, 14, 1–13.
Sato, Y., Toyoizumi, T. & Aihara, K. (2007). Bayesian inference explains
perception of unity and ventriloquism aftereffect: identification of com-
mon sources of audiovisual stimuli. Neural Computation, 19, 3335–3355.
Saunders, J. & Knill, D. (2004). Visual feedback control of hand movements.
The Journal of Neuroscience, 24, 3223–3234.
Savage, L. (1972). The Foundations of Statistics, 2nd ed. New York: Dover.

Savelli, F. & Knierim, J. (2019). Origin and role of path integration in the
cognitive representations of the hippocampus: computational insights into
open questions. Journal of Experimental Biology, 222, 1–13.
Scott, S. (2012). The computational and neural basis of voluntary motor control
and planning. Trends in Cognitive Sciences, 16, 541–549.
Seydell, A., Knill, D. & Trommershäuser, J. (2010). Adapting internal statistical
models for interpreting visual cues to depth. Journal of Vision, 10, 1–27.
Shadmehr, R. & Mussa-Ivaldi, S. (2012). Biological Learning and Control.
Cambridge, MA: MIT Press.
Shea, N. (2018). Representation in Cognitive Science. Oxford: Oxford
University Press.
Shikauchi, Y., Miyakoshi, M., Makeig, S. & Iversen, J. (2021). Bayesian models
of human navigation behavior in an augmented reality audiomaze.
European Journal of Neuroscience, 54, 8308–8317.
Skyrms, B. (1980). Causal Necessity. New Haven, CT: Yale University Press.
Skyrms, B. (1987). Dynamic coherence and probability kinematics. Philosophy
of Science, 54, 1–20.
Skyrms, B. (1995). Strict coherence, sigma coherence, and the metaphysics of
quantity. Philosophical Studies, 77, 39–55.
Sotiropoulos, G., Seitz, A. & Seriès, P. (2011). Changing expectations about
speed alters perceived motion direction. Current Biology, 21, R883‒R884.
Stalnaker, R. (1970). Probabilities and conditionals. Philosophy of Science, 37,
64–80.
Stalnaker, R. (1984). Inquiry. Cambridge, MA: MIT Press.
Steele, K. & Stefánsson, H. (2016). Decision theory. In E. Zalta, ed., The
Stanford Encyclopedia of Philosophy (Winter 2016). https://plato.stan
ford.edu/archives/win2016/entries/decision-theory.
Stich, S. (1983). From Folk Psychology to Cognitive Science. Cambridge, MA:
MIT Press.
Stocker, A. (2018). Credo for optimality. Behavioral and Brain Sciences, 41,
e244.
Stocker, A. & Simoncelli, E. (2006). Noise characteristics and prior expectations
in human visual speed perception. Nature Neuroscience, 4, 578–585.
Stone, J. (2011). Footprints sticking out of the sand, part 2: Children’s Bayesian
priors for shape and lighting direction. Perception, 40, 175–190.
Stone, J. (2013). Bayes’s Rule: A Tutorial Introduction to Bayesian Analysis.
Sebtel Press.
Temperley, D. (2007). Music and Probability. Cambridge, MA: MIT Press.
Thrun, S., Burgard, W. & Fox, D. (2005). Probabilistic Robotics. Cambridge,
MA: MIT Press.

Todorov, E. & Jordan, M. (2002). Optimal feedback control as a theory of motor

coordination. Nature Neuroscience, 5, 1226–1235.
Todorov, E. (2004). Optimality principles in sensorimotor control. Nature
Neuroscience, 7, 907–915.
Trommershäuser, J., Kording, K. & Landy, M, eds. (2011). Sensory Cue
Combination. Oxford: Oxford University Press.
Trotta, R. (2008). Bayes in the sky: Bayesian inference and model selection in
cosmology. Contemporary Physics, 49, 71–104.
Tversky, A., and Kahneman, D. (1974). Judgment under uncertainty: heuristics
and biases. Science, 185, 1124–1131.
Tversky, A., and Kahneman, D. (1983). Extension versus intuitive reasoning:
the conjunction fallacy in probability judgment. Psychological Review, 90,
293–315.
van Fraassen, B. (1980). The Scientiﬁc Image. Oxford: Oxford University Press.
van Gelder, T. (1992). What might cognition be, if not computation? Journal of
Philosophy, 92, 345–381.
van Rooij, I., Blokpoel, M., Kwisthout, J. & Wareham, T. (2019). Cognition and
Intractability. Cambridge: Cambridge University Press.
Vilares, I. & Kording, K. (2011). Bayesian models: the structure of the world,
uncertainty, behavior, and the brain. Annals of the New York Academy of
Sciences, 1224, 22–39.
Weisberg, J. (2009). Varieties of Bayesianism. In D. Gabbay, S. Hartman &
J. Woods, eds., Handbook of the History of Logic, vol. 10. New York:
Elsevier, pp. 477‒551.
Weiss, Y., Simoncelli, E. & Adelson, E. (2002). Motion illusions as optimal
percepts. Nature Neuroscience, 5, 598–604.
Weisswange, T., Rothkopf, C., Rodemann, T. & Triesch, J. (2011). Bayesian cue
integration as a developmental outcome of reward mediated learning.
PLOS One, 6, e21575.
Wolpert, D. & Flanagan, J. R. (2009). Forward models. In T. Bayne, A. Cleeremans
& P. Wilken, eds., The Oxford Companion to Consciousness. Oxford: Oxford
University Press, pp. 294‒296.

Keith Frankish
The University of Sheﬃeld
Keith Frankish is a philosopher specializing in philosophy of mind, philosophy of
psychology, and philosophy of cognitive science. He is the author of Mind and Supermind
(Cambridge University Press, 2004) and Consciousness (2005), and has also edited or
coedited several collections of essays, including The Cambridge Handbook of Cognitive
Science (Cambridge University Press, 2012), The Cambridge Handbook of Artiﬁcial
Intelligence (Cambridge University Press, 2014) (both with William Ramsey), and Illusionism
as a Theory of Consciousness (2017).

About the Series

This series provides concise, authoritative introductions to contemporary work in
philosophy of mind, written by leading researchers and including both established
and emerging topics. It provides an entry point to the primary literature
and will be the standard resource for researchers, students, and anyone
wanting a ﬁrm grounding in this fascinating ﬁeld.

Elements in the Series

Biological Cognition
Bryce Huebner and Jay Schulkin
Embodied and Enactive Approaches to Cognition
Shaun Gallagher
Mental Content
Peter Schulte
Aﬀective Bodily Awareness
Frédérique de Vignemont
The Computational Theory of Mind
Matteo Colombo and Gualtiero Piccinini
Memory and Remembering
Felipe De Brigard
Non-physicalist Theories of Consciousness
Hedda Hassel Mørch
Animal Minds
Marta Halina
Neurolaw
Gregg D. Caruso
Human Reasoning
David E. Over
Personal Identity and the Self
Rory Madden
Bayesian Models of the Mind
Michael Rescorla

A full series listing is available at: www.cambridge.org/EPMI

Mixture Modelling For Medical and Health Sciences (Shu Kay NG Liming Xiang Kelvin Kai Wing Yau) (Z-Library)
100% (1)
Mixture Modelling For Medical and Health Sciences (Shu Kay NG Liming Xiang Kelvin Kai Wing Yau) (Z-Library)
315 pages
Adventures in Financial Data Science The Empirical Properties of Financial and Economic Data 2nbsped 9811250642 9789811250644
No ratings yet
Adventures in Financial Data Science The Empirical Properties of Financial and Economic Data 2nbsped 9811250642 9789811250644
512 pages
(European Association of Methodology Series) Uwe Engel, Anabel Quan-Haase, Sunny Liu, Lars E. Lyberg - Handbook of Computational Social Science, Volume 1_ Theory, Case Studies and Ethics-Routledge (20
No ratings yet
(European Association of Methodology Series) Uwe Engel, Anabel Quan-Haase, Sunny Liu, Lars E. Lyberg - Handbook of Computational Social Science, Volume 1_ Theory, Case Studies and Ethics-Routledge (20
469 pages
Brain Computations and Connectivity (Edmund T. Rolls) (Z-Library)
100% (1)
Brain Computations and Connectivity (Edmund T. Rolls) (Z-Library)
1,174 pages
Nick Chater, Mike Oaksford - The Probabilistic Mind - Prospects For Bayesian Cognitive Science (2008, Oxford University Press) - Libgen - Li
No ratings yet
Nick Chater, Mike Oaksford - The Probabilistic Mind - Prospects For Bayesian Cognitive Science (2008, Oxford University Press) - Libgen - Li
567 pages
An Introduction To Bayesian Inference, Methods and Computation, Nick Heard, 1st Ed - 2021, 2021, Springer Nature
No ratings yet
An Introduction To Bayesian Inference, Methods and Computation, Nick Heard, 1st Ed - 2021, 2021, Springer Nature
176 pages
A
100% (5)
A
732 pages
Boundaries, Extents and Circulations: Koen Vermeir Jonathan Regier Editors
No ratings yet
Boundaries, Extents and Circulations: Koen Vermeir Jonathan Regier Editors
283 pages
Engelbrecht J. A Journey Into The Nonlinear World 2024
No ratings yet
Engelbrecht J. A Journey Into The Nonlinear World 2024
117 pages
Uncertainty Theory Baoding Liu Fourth Edition
No ratings yet
Uncertainty Theory Baoding Liu Fourth Edition
491 pages
Mathematics and Sports
No ratings yet
Mathematics and Sports
170 pages
Lee Palmer Wandel - Voracious Idols and Violent Hands - Iconoclasm in Reformation Zurich, Strasbourg, and Basel-Cambridge University Press (1994)
No ratings yet
Lee Palmer Wandel - Voracious Idols and Violent Hands - Iconoclasm in Reformation Zurich, Strasbourg, and Basel-Cambridge University Press (1994)
232 pages
A Handbook Of: by T. C. Hammond Revised and Edited by DAVID F. WRIGHT
No ratings yet
A Handbook Of: by T. C. Hammond Revised and Edited by DAVID F. WRIGHT
204 pages
(Studies in the History of Christian Traditions, Volume_ 192) Jordan J. Ballor_ Matthew T. Gaetano_ David S. Sytsma - Beyond Dordt and De Auxiliis_ The Dynamics of Protestant and Catholic Soteriology
100% (1)
(Studies in the History of Christian Traditions, Volume_ 192) Jordan J. Ballor_ Matthew T. Gaetano_ David S. Sytsma - Beyond Dordt and De Auxiliis_ The Dynamics of Protestant and Catholic Soteriology
369 pages
Watkins Systematic Interference
No ratings yet
Watkins Systematic Interference
32 pages
We Believe in One God (Gerald L. Bray) (Z-Library)
100% (1)
We Believe in One God (Gerald L. Bray) (Z-Library)
201 pages
A. C. C. Coolen, R. Kühn, P. Sollich - Theory of Neural Information Processing Systems-Oxford University Press, USA (2005)
No ratings yet
A. C. C. Coolen, R. Kühn, P. Sollich - Theory of Neural Information Processing Systems-Oxford University Press, USA (2005)
586 pages
CKA Exam - 28102022
100% (1)
CKA Exam - 28102022
31 pages
Sustainable Development Goals A - Katarzyna Cichos Jarosaw A. Sob
No ratings yet
Sustainable Development Goals A - Katarzyna Cichos Jarosaw A. Sob
406 pages
Statistical Modeling For Degradation Dat
No ratings yet
Statistical Modeling For Degradation Dat
382 pages
Angelo Loula, Ricardo Gudwin, João Queiroz (Editors) - Artificial Cognition Systems-Idea Group Publishing (2007)
No ratings yet
Angelo Loula, Ricardo Gudwin, João Queiroz (Editors) - Artificial Cognition Systems-Idea Group Publishing (2007)
419 pages
1.3 Data Storage and File Compression
No ratings yet
1.3 Data Storage and File Compression
6 pages
(Progress in Brain Research 168) Rahul Banerjee and Bikas K. Chakrabarti (Eds.) - Models of Brain and Mind - Physical, Computational and Psychological Approaches (2007, Elsevier, Academic Press)
No ratings yet
(Progress in Brain Research 168) Rahul Banerjee and Bikas K. Chakrabarti (Eds.) - Models of Brain and Mind - Physical, Computational and Psychological Approaches (2007, Elsevier, Academic Press)
273 pages
Bayesian Methods in Cosmology - Michael P. Hobson, Andrew H. Jaffe, Andrew R. Liddle, Pia Mukherjee, David Parkinson
No ratings yet
Bayesian Methods in Cosmology - Michael P. Hobson, Andrew H. Jaffe, Andrew R. Liddle, Pia Mukherjee, David Parkinson
317 pages
(Complexity - Reproducing Automata) Hypercycle - A Principle of Self Organization
No ratings yet
(Complexity - Reproducing Automata) Hypercycle - A Principle of Self Organization
98 pages
Parallel Distributed Processing
No ratings yet
Parallel Distributed Processing
7 pages
Science and Hypothesis - Historical Essays On Scientific Methodology - Laudan, Larry (1981)
No ratings yet
Science and Hypothesis - Historical Essays On Scientific Methodology - Laudan, Larry (1981)
268 pages
Why Information Can Not Be The Basis of Reality Scientific American Blog Network
No ratings yet
Why Information Can Not Be The Basis of Reality Scientific American Blog Network
4 pages
Machine Learning For Automated Theorem Proving
No ratings yet
Machine Learning For Automated Theorem Proving
106 pages
1 PDF
No ratings yet
1 PDF
467 pages
Bayesian Data Analysis
No ratings yet
Bayesian Data Analysis
14 pages
JPEG Compression Standard
No ratings yet
JPEG Compression Standard
23 pages
® Panel
No ratings yet
® Panel
148 pages
(Zurich Lectures in Advanced Mathematics) Guus Balkema-High Risk Scenarios and Extremes A Geometric Approach - European Mathematical Society (2007)
No ratings yet
(Zurich Lectures in Advanced Mathematics) Guus Balkema-High Risk Scenarios and Extremes A Geometric Approach - European Mathematical Society (2007)
391 pages
Berger Statistical Decision Theory and Bayesian PDF
No ratings yet
Berger Statistical Decision Theory and Bayesian PDF
316 pages
Eric Warrant, Dan-Eric Nilsson - Invertebrate Vision-Cambridge University Press (2006)
No ratings yet
Eric Warrant, Dan-Eric Nilsson - Invertebrate Vision-Cambridge University Press (2006)
285 pages
1 s2.0 S0888613X16301141 Main
No ratings yet
1 s2.0 S0888613X16301141 Main
19 pages
Cognitive Psychology - Module 1
No ratings yet
Cognitive Psychology - Module 1
72 pages
Unpacking Grade 9
No ratings yet
Unpacking Grade 9
180 pages
Hartmann
No ratings yet
Hartmann
36 pages
Distilling The Neural Correlates of Consciousness
No ratings yet
Distilling The Neural Correlates of Consciousness
10 pages
Tvol Morality Publication Web2018 7 PDF
No ratings yet
Tvol Morality Publication Web2018 7 PDF
44 pages
Fuchs Kelso
0% (1)
Fuchs Kelso
31 pages
Evolution of Artificial Intelligence: Lee Spector
No ratings yet
Evolution of Artificial Intelligence: Lee Spector
3 pages
Stochastic Processes in Physics, Chemistry and Biology 2000 - Freund PDF
No ratings yet
Stochastic Processes in Physics, Chemistry and Biology 2000 - Freund PDF
499 pages
Bostrom Review
No ratings yet
Bostrom Review
7 pages
Complex Systems A Survey
100% (1)
Complex Systems A Survey
10 pages
Propagation of Data Uncertainty in Surface Wave Inversion
No ratings yet
Propagation of Data Uncertainty in Surface Wave Inversion
10 pages
Artificial Intelligence and Machine Learning (21CS54) Module - 1
No ratings yet
Artificial Intelligence and Machine Learning (21CS54) Module - 1
13 pages
Bayesian Rationality The Probabilistic Approach To Human Reasoning 1st Edition Mike Oaksford PDF Download
No ratings yet
Bayesian Rationality The Probabilistic Approach To Human Reasoning 1st Edition Mike Oaksford PDF Download
52 pages
Approximate Bayesian Computation For Finite Mixture Models
No ratings yet
Approximate Bayesian Computation For Finite Mixture Models
21 pages
Reinforcement Learning
No ratings yet
Reinforcement Learning
7 pages
Systems Biology: A Brief Overview: Science April 2002
No ratings yet
Systems Biology: A Brief Overview: Science April 2002
4 pages
16 - The Bayesian Mind
No ratings yet
16 - The Bayesian Mind
33 pages
The (Im) Possibility of Forgiveness - An Empirical - Dion Angus Forster - 1, 2017 - African Sun Media - 9781928314349 - Anna's Archive
100% (1)
The (Im) Possibility of Forgiveness - An Empirical - Dion Angus Forster - 1, 2017 - African Sun Media - 9781928314349 - Anna's Archive
272 pages
Substance and The Fundamentality of The Familiar - A - Aristotle Inman, Ross D - Routledge Studies in Metaphysics, 10, 1st, London, 2017 - 9780367593599 - Anna's Archive
No ratings yet
Substance and The Fundamentality of The Familiar - A - Aristotle Inman, Ross D - Routledge Studies in Metaphysics, 10, 1st, London, 2017 - 9780367593599 - Anna's Archive
319 pages
ThinkPad T520 Manual
No ratings yet
ThinkPad T520 Manual
178 pages
Batanero Borovcnik Statistics and Probability in High School Contents Preface
No ratings yet
Batanero Borovcnik Statistics and Probability in High School Contents Preface
8 pages
Group Invariance in Statistical Inference (Narayan)
100% (1)
Group Invariance in Statistical Inference (Narayan)
176 pages
Gilbert Et Al. 1996
No ratings yet
Gilbert Et Al. 1996
16 pages
Syllabus
No ratings yet
Syllabus
3 pages
FRIJDA, The Place of Appraisal in Emotion
No ratings yet
FRIJDA, The Place of Appraisal in Emotion
32 pages
The Neural Basis of Attentional Control in Visual Search: Martin Eimer
No ratings yet
The Neural Basis of Attentional Control in Visual Search: Martin Eimer
25 pages
Carnap, Logical Foundations of Probability
100% (1)
Carnap, Logical Foundations of Probability
10 pages
Markov Chain Monte Carlo Methods: Christian P. Robert
No ratings yet
Markov Chain Monte Carlo Methods: Christian P. Robert
456 pages
Non-Ergodicity and Its Implications For Businesses and Investors PDF
No ratings yet
Non-Ergodicity and Its Implications For Businesses and Investors PDF
72 pages
Chirimuuta Noncausal Penultimate
No ratings yet
Chirimuuta Noncausal Penultimate
27 pages
BAM Book
No ratings yet
BAM Book
134 pages
On The Perception of Incongruity - A Paradigm - Bruner & Postman
No ratings yet
On The Perception of Incongruity - A Paradigm - Bruner & Postman
12 pages
Student Exploration: Fan Cart Physics
100% (1)
Student Exploration: Fan Cart Physics
4 pages
The Morphogenesis of Evolutionary Developmental Biology PDF
No ratings yet
The Morphogenesis of Evolutionary Developmental Biology PDF
11 pages
PFA Vs PTFE in Instrumentation
No ratings yet
PFA Vs PTFE in Instrumentation
5 pages
Automata Theory and Formal Languages - Module 2
No ratings yet
Automata Theory and Formal Languages - Module 2
4 pages
Uniqueness of The Earth
No ratings yet
Uniqueness of The Earth
19 pages
Book 1 Topic Past DSE Questions Revised 2023
No ratings yet
Book 1 Topic Past DSE Questions Revised 2023
66 pages
Panasonic Kx-t7230SP User Manual 2
No ratings yet
Panasonic Kx-t7230SP User Manual 2
26 pages
Integrated Science ATAR Y11 Sample Course Outline WACE 201516 - PDF
No ratings yet
Integrated Science ATAR Y11 Sample Course Outline WACE 201516 - PDF
4 pages
A Cross-Cultural Perspective On God's Personhood AKSHAY GUPTA
No ratings yet
A Cross-Cultural Perspective On God's Personhood AKSHAY GUPTA
37 pages
Ethnographic Tarot Project
No ratings yet
Ethnographic Tarot Project
5 pages
Work Sampling
No ratings yet
Work Sampling
36 pages
14-Article Text-20-1-10-20130201
No ratings yet
14-Article Text-20-1-10-20130201
30 pages
DFC3000 Product Specifications
No ratings yet
DFC3000 Product Specifications
2 pages
Reconstruction of Virtual Neural Circuits in An Insect Brain
No ratings yet
Reconstruction of Virtual Neural Circuits in An Insect Brain
8 pages
David Comprido: Product Design
No ratings yet
David Comprido: Product Design
72 pages
Soul Put To Work in Medicine A Response To Arthur Kleinman
No ratings yet
Soul Put To Work in Medicine A Response To Arthur Kleinman
13 pages
PRACTICAL RESEARCH 2 - Set B
No ratings yet
PRACTICAL RESEARCH 2 - Set B
1 page
CW 8200
No ratings yet
CW 8200
2 pages
Apply For Siemens Automation and Control Engineer Job - Engineering - Shanghai, China
No ratings yet
Apply For Siemens Automation and Control Engineer Job - Engineering - Shanghai, China
1 page
Soft Skills
No ratings yet
Soft Skills
12 pages
Digital TV Tuner Ic: Features
No ratings yet
Digital TV Tuner Ic: Features
25 pages
1 - Yaw and Lateral Stability - Main
No ratings yet
1 - Yaw and Lateral Stability - Main
7 pages
Chapt 4
No ratings yet
Chapt 4
33 pages
Linksys Unmanaged Switches: Key Features and Benefits
No ratings yet
Linksys Unmanaged Switches: Key Features and Benefits
2 pages
Se6 - Machine Guide Final
No ratings yet
Se6 - Machine Guide Final
8 pages
How To Be A Better Person
No ratings yet
How To Be A Better Person
8 pages
Las Alas de Icaro
No ratings yet
Las Alas de Icaro
2 pages
Personal Information: Application For Employment
No ratings yet
Personal Information: Application For Employment
1 page
Language Construct - Struct-Comparators
No ratings yet
Language Construct - Struct-Comparators
3 pages
Marking Guidelines Terrace and Superimposed Drainage Pattern (2024)
No ratings yet
Marking Guidelines Terrace and Superimposed Drainage Pattern (2024)
3 pages
PWB 209a
No ratings yet
PWB 209a
4 pages
Medical Conference Style Presentation by Slidesgo
No ratings yet
Medical Conference Style Presentation by Slidesgo
41 pages
From Patagonia to Professor
From Everand
From Patagonia to Professor
Meredith Temple-Smith
5/5 (1)

Bayesian Models of The Mind

Uploaded by

Bayesian Models of The Mind

Uploaded by

Rescorla

Bayesian decision theory is a mathematical framework that

Bayesian Models of the Mind

Cover image: Baac3nes / Moment /

Cambridge University Press is part of Cambridge University Press & Assessment,

Elements in the Philosophy of Mind

Abstract: Bayesian decision theory is a mathematical framework that

Keywords: Bayesian decision theory, mental representation, Bayesian

© Michael Rescorla 2024

2 The Probability Calculus 2

3 Bayesian Decision Theory 18

4 Bayesian Cognitive Science 27

5 Realism and Instrumentalism 42

Appendix: Foundations of Probability Theory 76

representation. Representationalists such as Burge (2010; 2022), Fodor (1975;

2 The Probability Calculus

not track any objective probabilities that inhere in mind-independent reality. To

2.1 Sets of Outcomes

meaning that ω is a member of set A. We also say that ω belongs to A.

Similarly, suppose we seek to deﬁne probabilities over possible results of

fω : Seabiscuit finishes before every other horse in ωg;

• Negation corresponds to complementation. The complement of set A is

Figure 1 Ω is the rectangle. A is the ellipse. Ac is shaded gray.

Figure 2 A and B are overlapping sets. Their intersection A ∩ B is

f1gc ¼ f2; 3; 4; 5; 6g:

• Conjunction corresponds to intersection. The intersection of A and B is the

f2; 4; 6g ∩ f4; 5; 6g ¼ f4; 6g:

• Disjunction corresponds to union. The union of A and B is the set containing

f1g ∪ f4g ¼ f1; 4g:

By iteratively applying set-theoretic operations, Kolmogorov replicates the

Figure 3 A and B are overlapping sets. Their union A ∪ B is shaded gray.

asteroid speeds. If we want to model probabilities over possible asteroid speeds,

2.2 Axioms of the Probability Calculus

Axiom 1: Probabilities are real numbers between 0 and 1,

where a real number is any number that can be expressed as a decimal. As

Intuitively: Ω exhausts all relevant possibilities, so it must receive maximal

PðH1 ∪ H2 Þ ¼ PðH1 Þ þ PðH2 Þ:

to the axioms rationally superior to someone who violates them? A large

2.3 Random Variables

A rigorous deﬁnition of “random variable” is given in Section A3, but for

Figure 5 An interval ½a; b lying in ℝ.

Figure 6 Ω is the rectangle. X 1 ½a; b contains the outcomes mapped by X into

X ðωÞ ¼ x & Y ðωÞ ¼ y

Figure 7 Rectangle REC contains ordered pairs ðx; yÞ.

fω : ðX ðωÞ; Y ðωÞÞ 2 RECg:

2.4 Probability Density

Figure 8 fω : ðX ðωÞ; Y ðωÞÞ 2 RECg contains the outcomes mapped by

from real numbers to probability densities) determines a probability distribu-

• Probabilities are assigned to sets whose members belong to an outcome space

As is customary in the literature, I notate probability using an upper case P and

Figure 10 A Gaussian pdf with mean m and variance σ2 .

Pð½a; bÞ > 0

Figure 12 A two-dimensional pdf. The pdf assigns a nonnegative value to each

Figure 15 A two-dimensional Gaussian pdf.

2.5 Conditional Probability

See Figure 17. As Figure 17 illustrates, the unconditional probability of A may

Thus, we cannot use the ratio formula to deﬁne probabilities conditional

3 Bayesian Decision Theory

evidence; and expected utility maximization, which governs how credences

Pnew ðHÞ ¼ PðHjEÞ;

to signify that my new credence in H is equal to my old conditional credence in

Pnew ðHÞ ¼ PðHjEÞ:

There is considerable variation in how philosophers formulate Conditionalization,

However exactly we formulate Conditionalization, it is a diachronic norm: it

3.2 Bayes’s Theorem

serves mainly as a normalization constant to ensure that probabilities sum to 1,

PðHjEÞ ¼ k PðHÞPðEjHÞ; ð2Þ

where k ¼ 1=PðEÞ. One can also write the theorem as:

posterior ∝ prior prior likelihood:

Bayes’s theorem is extraordinarily useful. In many situations, there is a natural prior

pðxjyÞ ¼ k pðxÞpðyjxÞ; ð3Þ

where k is again a normalization constant. pðxÞ serves as a prior density: it

pðxjyÞ ∝ pðxÞpðyjxÞ: ð4Þ

3.3 Expected Utility Maximization

2016). We may leave it at an intuitive level. Scientiﬁc applications often

In many statistical applications, the “action” is to estimate the value of

Cover image: Baac3nes / Moment /