Ebffiledoc - 993download PDF Neurocognitive Mechanisms Explaining Biological Cognition Gualtiero Piccinini Ebook Full Chapter

Neurocognitive Mechanisms:
Explaining Biological Cognition

Gualtiero Piccinini
Visit to download the full and correct content document:
https://textbookfull.com/product/neurocognitive-mechanisms-explaining-biological-cog
nition-gualtiero-piccinini/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Molecular Mechanisms of Autonomy in Biological Systems

Tara Karimi
https://textbookfull.com/product/molecular-mechanisms-of-
autonomy-in-biological-systems-tara-karimi/
Physical computation : a mechanistic account Piccinini
https://textbookfull.com/product/physical-computation-a-
mechanistic-account-piccinini/
Yeast Cell Surface Engineering Biological Mechanisms

and Practical Applications Mitsuyoshi Ueda
https://textbookfull.com/product/yeast-cell-surface-engineering-
biological-mechanisms-and-practical-applications-mitsuyoshi-ueda/
Biological and Bio inspired Nanomaterials Properties

and Assembly Mechanisms Sarah Perrett
https://textbookfull.com/product/biological-and-bio-inspired-
nanomaterials-properties-and-assembly-mechanisms-sarah-perrett/
Cognition and Addiction A Researcher s Guide from
Mechanisms Towards Interventions 1st Edition Antonio
Verdejo García (Editor)
https://textbookfull.com/product/cognition-and-addiction-a-
researcher-s-guide-from-mechanisms-towards-interventions-1st-
edition-antonio-verdejo-garcia-editor/
Explaining Health Across the Sciences Jonathan Sholl
https://textbookfull.com/product/explaining-health-across-the-
sciences-jonathan-sholl/
Explaining Imagination 1st Edition Peter Langland

Hassan
https://textbookfull.com/product/explaining-imagination-1st-
edition-peter-langland-hassan/
Explaining Health Across the Sciences 1st Edition

Jonathan Sholl
https://textbookfull.com/product/explaining-health-across-the-
sciences-1st-edition-jonathan-sholl/
Coulson and Richardson’s Chemical Engineering, Fourth

Edition: Volume 3A: Chemical and Biochemical Reactors
and Reaction Engineering R. Ravi
https://textbookfull.com/product/coulson-and-richardsons-
chemical-engineering-fourth-edition-volume-3a-chemical-and-
biochemical-reactors-and-reaction-engineering-r-ravi/
OUP CORRECTED AUTOPAGE PROOFS – FINAL, 7/10/2020, SPi
Neurocognitive Mechanisms
Neurocognitive
Mechanisms
Explaining Biological Cognition
GUALTIERO PICCININI
1
3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Gualtiero Piccinini 2020
The moral rights of the author have been asserted
First Edition published in 2020
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2020947366
ISBN 978–0–19–886628–2
Printed and bound by
CPI Group (UK) Ltd, Croydon, CR0 4YY
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
Contents
Acknowledgments vii
List of Figures xi
Introduction 1
1. Levels of Being 6
2. Mechanisms, Multiple Realizability, and Medium Independence 38
3. Functional Mechanisms 67
4. Mechanistic Functionalism 89
5. The First Computational Theory of Cognition: McCulloch
and Pitts’s “A Logical Calculus of the Ideas Immanent in
Nervous Activity” 107
6. Computation and Information Processing 128
7. Mechanistic Models of Cognition: From Autonomy to Integration 156
8. The Cognitive Neuroscience Revolution 182
9. The Computational Theory of Cognition 205
10. The Church–Turing Fallacy 225
11. The Resilience of the Computational Theory of Cognition 244
12. Neural Representation 258
13. Neural Computation 297
14. Computation and the Function of Consciousness 317
Bibliography 351
Index 393
Acknowledgments
My undergraduate studies at the University of Turin, during the early 1990s,

exposed me to the classic debate on the foundations of cognitive science. Is the
brain a computer? Does cognition involve computation over representations? Is
the mind the software of the brain? There were arguments on both sides, and
I wanted to sort them out.
While in graduate school at the University of Pittsburgh, during the late 1990s
and early 2000s, I embarked on a research program: first, figure out what physical
computation is by drawing from computer science and computer engineering;
then, apply that understanding to the brain and answer the foundational questions
about cognition. For good measure, I included a careful study of the origin of
computational theories of cognition. That was way too much for a dissertation.
My dissertation included some historical research and an account of physical
computation (Piccinini 2003a). I left the foundational questions for future work.
My early mechanistic account of physical computation was, roughly, that
physical computation is the manipulation of digits by a functional mechanism
in accordance with a rule (Piccinini 2007a). I thought this account would allow me
to refute the Computational Theory of Cognition (CTC) once and for all. I wrote a
paper attempting that. My argument was that computation is digital; neural
activity is not digital; therefore, neural activity is not computational. But I never
published that paper because I eventually realized I was missing something.
I knew there used to be analog computers, and they had been used as an
alternative model of brain activity. I also knew that analog computers are quite
different from digital computers, are not program-controlled (qua analog), are not
universal in Alan Turing’s sense, existed before digital computers, and used to be
called differential analyzers. They only started to be called analog “computers”
after digital computers became popular. So, I argued that analog “computers”
weren’t computers properly so called after all (Piccinini 2008a). At least two
events pushed back.
First, the late Jonathan Mills asked me to collaborate on explicating analog
computation. Although he and I never wrote anything together, he introduced
me to the groundbreaking research on analog computation that he was doing at
Indiana University, of which I was unaware. Independently, in July 2007, Marcin
Milkowski wrote to me and pointed out that some computer scientists were reviving
analog computation; he recommended a “humbler approach” than mine.
I wondered whether those who study analog computation, and perhaps others
who study other unconventional models of (so-called) computation that are not
viii 
digital, might be picking up on something that digital computation, analog

“computation,” and other types of “computation” have in common. From a math-
ematical standpoint, they are all ways of solving mathematical problems. But I was
after an account of physical computation. What do the disparate physical processes
that implement different types of computation have in common?
The received view is that computation has to do with representation. But I had
arguments that, in fact, computation does not require representation (Piccinini
2004a, 2008b). I also knew that the notions of function and multiple realizability
are too weak to underwrite a robust notion of computation (Piccinini 2004b,
2007b). At some point, I remembered reading an insightful paper in which Justin
Garson (2003) uses the notion of medium independence to characterize informa-
tion. (John Haugeland (1985) uses a similar notion to characterize automatic
formal systems, but I only found out about that much later.) I re-read Justin’s
paper and found that medium independence was just what I needed. After that,
I generalized my account of physical computation to cover analog computation
and other unconventional models of computation. The resulting account is that
physical computation is the manipulation of medium independent vehicles by a
functional mechanism in accordance with a rule (Piccinini and Scarantino 2011,
Piccinini 2015).
Given this broader understanding of physical computation, my attitude about
the brain and cognition switched. I now had the materials for a cogent argument
in favor of CTC—with a twist. With help from neuroscientist Sonya Bahar, I re-
wrote my early paper against CTC to argue that, on one hand, neurocognitive
processes are computations because they are medium independent and, on the
other hand, neural computations are neither digital nor analog—they are sui
generis (Piccinini and Bahar 2013).
That seemed important but, to answer the foundational questions, much more
needed to be done. I needed a proper ontological foundation, an adequate account
of functions, a way to integrate psychology and neuroscience, an assessment of
other arguments pro and con CTC, and a place for representation and conscious-
ness. After many more collaborations and publications and much additional
research, this book is the result. It presents a comprehensive defense of CTC,
updated for the era of cognitive neuroscience, with surprises for both defenders
and critics of traditional forms of CTC.
Way more people have helped me than I can thank individually. I collectively
thank my teachers as well as the many audiences to whom I presented versions of
these ideas. I learned a lot from them.
My work builds on those who came before me. Besides those I already men-
tioned and my collaborators, those with the greatest philosophical influence on
this project include Bill Bechtel, Christopher Boorse, Paul and Patricia
Churchland, Jack Copeland, Carl Craver, Robert Cummins, Daniel Dennett,
Fred Dretske, Frankie Egan, Jerry Fodor, Gilbert Harman, John Heil, Jaegwon
 ix
Kim, Ruth Millikan, Hilary Putnam, Jonathan Schaffer, Oron Shagrir, Wilfried
Sieg, and Stephen Stich. On the science side, those from whom I learned the most
include Warren McCulloch, Allen Newell, Walter Pitts, Claude Shannon, Herbert
Simon, Alan Turing, and John von Neumann.
For comments on many papers and illuminating discussions over many years
I owe special thanks to Neal Anderson, Ken Aizawa, Trey Boone, Carl Craver,
Corey Maley, Andrea Scarantino, Oron Shagrir, Mark Sprevak, John Heil, Justin
Garson, and Eric Thomson. For comments on aspects of this work, I thanks
Darren Abramson, Anibal Astobiza, Brice Bantegnie, Sergio Barberis, Bill Bechtel,
Peter Bradley, Giovanni Camardi, Glenn Carruthers, David Chalmers, Mazviita
Chirimuuta, Michelle Ciurria, Judith Crane, Robert Cummins, Tony Dardis, Joe
Dewhurst, Frankie Egan, Chris Eliasmith, Ilke Ercan, Nir Fresco, Rachel Francon,
Erik Funkhouser, Carl Gillett, Stuart Glennan, Mahi Hardalupas, Inman Harvey,
Eric Hochstein, Phyllis Illari, Anne Jacobson, David M. Kaplan, Daniel Kramer,
Arnon Levy, Bill Lycan, Peter Machamer, Jack Mallah, Corey Maley, Diego
Marconi, Marcin Milkowski, Kevin Morris, Alyssa Ney, Abel Wajnerman Paz,
Alessio Plebe, Russ Poldrack, Tom Polger, Charles Rathkopf, Michael Rescorla,
Sarah Robins, Waldemar Rohloff, Robert Rupert, Anna-Mari Rusanen, Kevin
Ryan, Matthias Scheutz, Susan Schneider, Whit Schonbein, Alex Schumm, Paul
Schweizer, Sam Scott, Oron Shagrir, Larry Shapiro, Kent Staley, Terrence Stewart,
Jackie Sullivan, Brandon Towl, Charles Wallis, Dan Weiskopf, Gene Witmer,
Mark Zeise, many anonymous referees, and others I can’t recall. Special thanks
to Michael Barkasi, Mark Couch, and the anonymous referees for OUP for
insightful comments on the book manuscript.
Thanks to Crystal Brown, Frank Faries, Mirinda James, Elliott Risch, and James
Virtel for editorial assistance.
The Introduction, Chapter 1, and Chapter 3 are new.
Chapter 2 includes a substantially revised and expanded descendant of
Piccinini and Maley 2014. Sections 2.1, 2.3, and 2.5 are new.
Chapter 4 includes a substantially revised and expanded descendant of Maley
and Piccinini 2013. Sections 4.1.1 and 4.1.2 are new.
Chapter 5 is a revised and expanded descendant of parts of Piccinini 2004c.
Sections 6.2–6.8 of Chapter 6 are a revised and expanded descendant of
Piccinini 2017a. Section 6.9 is a revised descendant of a section of Piccinini
2009. Sections 6.1 and 6.10–6.12 are new.
Chapter 7 includes a revised and expanded descendant of Boone and Piccinini
2016a. Sections 7.1–7.3 and 7.7 are new.
Chapter 8 is a revised descendant of Boone and Piccinini 2016b.
Sections 9.1 and 9.2 of Chapter 9 include revised versions of portions of
Piccinini 2016. Most of Section 9.2 is a revised and expanded descendant of a
section of Piccinini and Bahar 2013. Sections 9.3–9.5 are a revised descendant of
parts of Piccinini 2009.
x 
Chapter 10 is a substantially revised descendant of Piccinini 2007c.

Chapter 11 is a substantially revised and expanded descendant of Piccinini
2010a.
Chapter 12 is a substantially revised descendant of Thomson and Piccinini
2018.
Chapter 13 is a revised and expanded descendant of much of Piccinini and
Bahar 2013.
Chapter 14 includes a substantially revised descendant of Piccinini 2010b.
Sections 14.1 and 14.9 are new.
It was a collective effort. I am deeply grateful to my co-authors on some of the
articles from which some of the chapters derive—Sonya Bahar, Trey Boone, Corey
Maley, and Eric Thomson—two philosophers and two neuroscientists. They
graciously agreed to let me use our joint work in this book and deserve credit
for much of what is correct in the chapters derived from it. My research program
benefited immensely from each collaboration.
This material is based upon work supported by the National Science
Foundation under Grants No. SES-0216981, SES-0924527, and especially SES-
1654982. I also received research support from a University of Missouri Research
Board Award, a 2006 NEH Summer Seminar at Washington University in St.
Louis, a University of Missouri Research Grant, a grant from the International
Studies and Programs at the University of Missouri—St. Louis, an Adelle and
Erwin Tomash Fellowship, an Andrew Mellon Predoctoral Fellowship, and a
Regione Sardegna Doctoral Scholarship. Any opinions, findings, conclusions,
and recommendations expressed in this work are those of the author and do
not necessarily reflect the views of these funding institutions.
Thanks to Peter Momtchiloff and the whole team at OUP for shepherding my
manuscript through the publication process.
Deep thanks to my friends and family for their love and support during all these
years. I’m grateful to my partner Michelle and daughters Violet, Brie, and Martine
for bringing so much meaning and joy to my life.
List of Figures
1.1 Possible relations between higher-level properties and lower-level properties 26

1.2 Ways of representing a whole and its parts 33
1.3 Wholes are invariants over certain changes in parts, properties of wholes
are aspects of their lower-level realizers 35
2.1 The overlap between lower-level property 1 and lower-level property 2 is a
higher-level property realized by both properties 1 and 2, and so forth 39
2.2 Higher-level properties are aspects of their lower-level realizers. 49
2.3 The intersections of lower-level properties are variably realizable
higher-level properties 53
2.4 Different types of higher-level properties 63
5.1 Diagrams of McCulloch and Pitts nets 114
5.2 Net explaining the heat illusion 115
6.1 Types of digital computation and their relations of class inclusion 147
6.2 Types of computation and their relations of class inclusion 150
7.1 Graph-theoretic representation of a lateral inhibition circuit 175
9.1 Some prominent forms of CTC and their relations 223
12.1 Retinal representation of the visual world 270
12.2 Main visual pathway 271
12.3 Topographic representation of motion in monkey area MT 273
12.4 Binocular rivalry demonstration 275
12.5 Working memory: behavioral and neuronal perspectives 277
12.6 From muscles to action maps 283
12.7 Movement field in M1 285
12.8 Space of M1 representational theories 286
12.9 Sensory cancellation in weak electric fish 291
12.10 Forward models and efference copy 293
13.1 Analog, digital, and neural signals 299
Introduction
This book defends a neurocomputational theory of cognition grounded in a

mechanistic, functionalist, egalitarian ontology. I argue that biological cognitive
capacities are constitutively explained by multilevel neurocognitive mechanisms,
which perform neural computations over neural representations. Providing a
scientific explanation of cognition requires understanding how neurocognitive
mechanisms work. Therefore, the science of cognition ought to include neurosci-
ence to a degree that traditional cognitive science was not expected to. Scientists
on the ground have been working on this for a while. Psychology is becoming
more and more integrated with neuroscience.
The picture I defend stands opposite to both traditional reductionism (type-
identity theory) and anti-reductionism (autonomy of psychology). Contrary to
traditional reductionism, neural computations and representations are not iden-
tical to their lower-level realizers. Contrary to traditional anti-reductionism,
neural computations and representations are not entirely distinct from their
realizers. Instead, neural computations and representations—both types and
tokens—are aspects of their realizers.
The picture I defend stands opposite to both computational chauvinism (com-
putation and representation are proprietary psychological notions) and anti-
realism (computations and representations are mere manners of speaking).
Contrary to computational chauvinism, computations and representations are
properties of the nervous system. Contrary to anti-realism, computations and
representations are real, causally efficacious properties—as real as any other
properties of the nervous system.
I will begin by providing accounts of levels of composition and realization,
mechanisms, functions, computation, and multilevel neurocognitive mechanisms.
This ontological foundation will allow me to improve on existing versions of
functionalism and make clear what the computational theory of cognition does
and does not say. After that, I will address some fallacies and objections to the
computational theory of cognition. Finally, I will provide empirical evidence that
neurocognitive systems perform sui generis computations over neural represen-
tations. I will conclude by clarifying the relation between computation and
consciousness and offering a noncomputational functionalism about phenomenal
consciousness.
By capacity, I mean a causal power subject to normative evaluation—a power
that can be manifested correctly or incorrectly.
Neurocognitive Mechanisms: Explaining Biological Cognition. Gualtiero Piccinini, Oxford University Press (2020).
© Gualtiero Piccinini.
DOI: 10.1093/oso/9780198866282.001.0001
2  
By cognitive capacities, I mean capacities such as perception, memory, reason-

ing, emotion, language, planning, and motor control. Cognitive capacities explain
some of the most interesting behaviors exhibited by physical systems.
By biological cognition, I mean cognition carried out by biological organisms—
more specifically, organisms with a specialized control organ. The specialized
control organ of earthly organisms is the nervous system. Some philosophers
argue that plants have cognitive capacities (Calvo 2016). Plants, fungi, and even
bacteria are exquisitely adapted to their environments. Yet they lack a specialized
organ for control of their organismic functions based on processing information
that results from integrating inputs from multiple, physically distinct sources. My
topic is the cognitive capacities of organisms with a specialized control organ that
integrates multiple sources of information.
There are also artifacts such as robots and digital assistants that possess
computational and information processing capacities, and there are analogies
between such artificial systems and biological cognizers. This book focuses on
biological cognition exhibited by organisms with a nervous system.
Some argue that the realizers of cognitive states and processes in biological
organisms include not only the nervous system but also some things outside it
(e.g., Wilson 1994). I will mostly ignore this possibility to simplify the exposition;
this does not affect my arguments.
By constitutive explanation, I mean explanation of a capacity of a system in
terms of the system’s causal structure. I will argue that constitutive explanation is
provided by mechanisms—that is, pluralities of components, component func-
tions, and organizational relations that, collectively, possess the capacity and
produce its manifestations (Chapters 3, 7).
Typically, components of mechanisms are themselves mechanisms whose
capacities are explained mechanistically. Ditto for the components’ components.
This multilevel mechanistic structure requires an ontological foundation. I argue
that all levels are equally real: neither higher levels nor lower levels are more
fundamental than one another. Higher-level objects are invariants under the
addition and subtraction of some parts, while higher-level properties are aspects
of their lower-level realizers. Adequate constitutive explanation requires identifying
the higher-level properties and organizational relations that produce a capacity.
Therefore, adequate constitutive explanation requires identifying appropriate
higher-level objects and their relevant properties (Chapter 1).
Capacities such as cognition are causal roles—that is, they are higher-level
properties defined solely by their regular effects under certain conditions. Causal
roles are often multiply realizable, meaning that there are different kinds of
mechanism that can perform the same causal role. In addition, some multiply
realizable causal roles are medium independent, meaning that even their inputs
and outputs are multiply realizable (Chapter 2).
 3
Causal roles are one, broad notion of function. In Chapter 3, I argue that the
notion of causal role is too broad to do justice to mechanistic explanation of
organisms and artifacts. The capacities of organisms and artifacts are explained by
causal roles that make regular contributions to the goals of organisms. These
contributions are teleological functions (Chapter 3).
On this basis, I propose an improved formulation of functionalism: the view
that the mind is the functional organization of the brain. I argue that functional
organization should be understood mechanistically, as encompassing not only
causal relations between internal states but also the components bearing the states,
their functions, and their organizational relations (Chapter 4).
Functionalism is a close ally of the computational theory of cognition, which
was first proposed by Warren McCulloch and Walter Pitts in 1943. They created a
neural network formalism for representing neuronal activity in a simplified and
idealized way, and they argued that neural networks perform digital computa-
tions. Their idea had an immense impact on the field (Chapter 5).
Some mechanisms, such as the neural networks devised by McCulloch and
Pitts, have the special teleological functions of computing and processing informa-
tion. Computing is processing medium-independent vehicles in accordance with a
rule, while information processing is the processing of vehicles that carry infor-
mation. Computational vehicles may or may not carry information, and infor-
mation processing may or may not be done by computing (Chapter 6).
Many authors have argued that there are constitutive explanations—whether
computational or not—that are not mechanistic. I argue that such putatively
nonmechanistic constitutive explanations are aspects of mechanisms. Adequate
constitutive explanation—including computational explanation—involves mech-
anisms (Chapter 7).
The scientific study of neurocognitive mechanisms is cognitive neuroscience.
To a large extent, cognitive science has already turned into cognitive neuroscience.
This is the integrated study of how multilevel neurocomputational mechanisms
that process neural representations explain cognition (Chapter 8).
The Computational Theory of Cognition (CTC) is the theory that cognitive
processes are computations, or cognition is explained computationally. Since
computation is a mechanistic process, CTC is a mechanistic hypothesis. Since
the organ of biological cognition is the nervous system, CTC for biological
organisms is the claim that neurocognitive processes are computations. I argue
that CTC in its generic formulation is correct for a couple of reasons. First, the
main vehicles of neural processing—spike trains—are functionally significant
thanks primarily to firing rates and spike timing, which are medium independent.
Second, neural signals carry, integrate, and process information from physically
different sources, which requires transducing them into shared, medium-
independent vehicles (Chapter 9).
4  
Assessing CTC requires understanding not only the reasons for it, but also what
is not a reason for it. A range of putative arguments for CTC are based on the
Church–Turing thesis, the thesis that the functions that are computable in an
intuitive sense are precisely those functions that are computable by Turing
machines. I argue that these arguments are fallacious: the Church–Turing thesis
does not help establish CTC (Chapter 10).
Assessing CTC also requires understanding putative objections. There are two
classes of objections. Insufficiency objections maintain that cognition involves
X (for some X) and computation is insufficient for X. Candidate Xs include
consciousness, intentionality, embodiment, embeddedness, dynamics, and math-
ematical insight. I reply that insufficiency objections do not undermine CTC; at
most, they show that something else is needed, in addition to computation, to
explain cognition. I emphasize that an adequate computational explanation of
biological cognition involves computations that are embodied and embedded.
Objections from neural realization argue that neural processes involve Y (for
some Y) and computation does not involve Y; therefore, neural processes are not
computations. I reply that none of the putative Y’s undermines CTC (Chapter 11).
Although computation can occur in the absence of information processing, this
is not what happens in the nervous system. Neural processes carry, integrate, and
process information in the service of teleological control functions. Informational
Teleosemantics is the view that representations are precisely states that carry
information in the service of teleological functions. Thus, Informational
Teleosemantics applied to the neurocognitive system entails that the neural states
that carry information in the service of control functions are neural representa-
tions. I argue that neural representations are not only real but routinely observed
by neuroscientists (Chapter 12).
One long-standing dispute about neural computation is whether it is digital or
analog. I argue that it is neither, at least in the general case. Instead, neural
computation is sui generis. Therefore, theories of cognition should take into
account what is known about neural computation (Chapter 13).
The Computational Theory of Mind (CTM), as I use this term, is a stronger view
than the Computational Theory of Cognition. CTM covers the whole mind—both
cognition and consciousness. Thus, CTM says that the whole mind has a com-
putational nature, or that all there is to the nature of mental states, including
conscious states, is being computational states. Since computation is a mechanistic
process, CTM is a mechanistic hypothesis. It is the computational version of
mechanistic functionalism. Contrary to a common assumption, the alternative
to CTM is not just the type-identity theory of mind. There is also a noncomputa-
tional version of functionalism about consciousness that deserves to be explored
(Chapter 14).
Biological cognition, or at least biological cognition in organisms with a
centralized control system that integrates multiple sources of information, turns
 5
out to be neural computation over neural representations. Neural computation is

a sui generis kind of computation—neither digital nor analog—that neurocogni-
tive mechanisms perform. Neural representations are simulations of the organism,
its environment, and their interaction that neurocognitive mechanisms construct.
Neural computations process neural representations. The result of this process is
biological cognition.
Neurocognitive mechanisms span multiple levels of organization, from single
neurons up to the whole nervous system. At each neurocognitive level, neural
computations process neural representations. Except for the lowest level, each
neurocognitive level is realized by the level below it. Except for the highest level,
each neurocognitive level realizes the level above it. Thus, lower-level neural
representations and computations realize higher-level neural computations and
representations.
Needless to say, explaining how cognitively endowed organisms behave
requires considering the dynamical coupling among nervous system, body, and
environment. In other words, explaining how organisms behave requires consid-
ering how nervous systems are embodied and embedded in their environment.
Considering all of this is what cognitive neuroscience does. Nevertheless, nervous
systems contribute something distinctively cognitive to the explanation: the per-
formance of neural computations over neural representations. That is the focus of
this book.
Although the chapters are deeply intertwined, I outline the main point of each
chapter in its first section—except for Chapter 1. A reader who wants the gist of
the book quickly can read Chapter 1 plus the first section of the other chapters.
1
Levels of Being
1.1 Parts, Wholes, and their Properties
This book argues that cognition is explained by multilevel neurocognitive

mechanisms. The first step towards this view—or any view involving levels of
composition and realization, for that matter—is clarifying what levels are and how
they are related. This chapter lays out an account of levels on which the rest of the
book will build.
The way we talk about things and their proper¹ parts gives rise to puzzles. For
example, a hammer has two parts: a handle and a head. If you have a hammer,
how many objects do you have? If you answer one, you are not counting the head
and the handle. If you answer two, you are not counting the whole hammer. If you
answer three, you seem to be counting the same portion of reality twice. None of
the answers are entirely satisfying. And then there are the hammer’s molecules,
atoms, and subatomic particles. How many things are there?
Puzzles like this make it difficult to understand how our discourse about wholes
fits with our discourse about parts. Both lay people and scientists talk about
different sorts of objects, some of which are part of others. We talk about people,
nervous systems, brains, neural networks, neurons, neurotransmitters, and so on.
We describe properties that these objects have and their manifestations: people
walk, brains develop, and neurons send action potentials.
We also describe relations between objects. For convenience, in this chapter
I will often follow the common convention of using the term “property” to include
relations. For instance, I will treat A being inside B as a spatial property possessed
by A and B collectively. So, I will often talk simply of objects and properties,
with the understanding that relations are also included. Nothing substantive
hinges on this.
Some relations occur between objects that are wholly distinct from one
another—objects at the same level: people talk to one another, brains are encased
in skulls, and neurons attach to one another (via synapses). Other relations occur
between objects that are part of one another. For example, synapses are part of
¹ In mereology—the formal study of the part–whole relation—it is convenient to define “part” so

that everything is part of itself, and “proper part” as a part that is not identical to the whole. By contrast,
in ordinary language, we usually presuppose that a part is not identical to the whole. From now on,
I will follow ordinary usage and use “part” to mean proper part.
Neurocognitive Mechanisms: Explaining Biological Cognition. Gualtiero Piccinini, Oxford University Press (2020).
© Gualtiero Piccinini.
DOI: 10.1093/oso/9780198866282.001.0001
   7
neurons, which together form neural networks; nervous systems are partially
made out of brains, which are partially made out of networks, among other
structures. Being part of something and being made out of something are relations
that occur between objects at different levels of being.
Sorting out what there is, what properties objects have, and how they relate is
the business of science. Sorting out what sorts of objects there are as well as what
sorts of properties and relations they have is the business of metaphysics. Scientists
rely on experimenting and theorizing. Metaphysicians rely on scientific results
and conceptual considerations. By using these tools, metaphysicians attempt to
organize our concepts so we don’t get confused or say things that make little sense.
This book begins with metaphysics and progresses from there towards the phil-
osophy of mind and cognitive neuroscience.
I will use the term “composition” for the relation between wholes and their
parts: parts compose a whole; a whole decomposes into its parts. I will use the term
“realization” for the relation between properties of a whole and properties of
its parts taken collectively, or between its parts taken collectively and all of the
parts’ parts taken collectively: higher-level properties are realized by lower-level
properties; lower-level properties realize higher-level properties.²
Properties come in types and tokens (instances). For example, each planet in
the solar system is an oblate spheroid. That is, each planet instantiates the
property type of oblate spheroidicity. Equivalently, each planet has its own
token, or instance, of oblate spheroidicity. We can talk about type realization or
token realization. I will mostly focus on type realization, without making it explicit
every time. Much of what I say applies to token realization as well.
As I use these terms, both composition and realization are asymmetric, irre-
flexive, and transitive.³ Both are relations of synchronic metaphysical necessita-
tion: once you have the parts with their properties and relations, you necessarily
have the whole with its properties and relations. Necessitation does not entail
dependence! Both composition and realization are often assumed to be relations of
ontological dependence, meaning that wholes and their properties ontologically
depend on (are grounded in, are posterior to) their parts and their properties. Not
only am I not making this dependence assumption; I will soon argue against it.
There are a number of puzzles that arise from our talk about things at different
levels of being. To keep the discussion manageable, let’s focus on one especially
difficult and theoretically important puzzle: the puzzle of causal exclusion. I’ll
² There is a dispute as to whether, when property P is realized by property Q (where Q could also be
a plurality of properties), P and Q belong to the same object X or Q belongs to X’s parts. I will discuss
this in Chapter 2, where I argue that, in the sense of realization that matters most to the metaphysics of
science, Q belongs to X’s parts. In this chapter, I’m just defining “realization” as the relation that holds
between the properties of a whole and the properties of its parts, or between its parts taken collectively
and all of the parts’ parts taken collectively.
³ Asymmetry: xRy → not yRx. Irreflexivity: not xRx. Transitivity: (xRy and yRz) → xRz. On the
features of realization, cf. Baysan 2015.
8  
propose a framework for thinking about levels of being that solves the causal
exclusion puzzle and explains how different levels fit together. Beware: a proper
treatment of these topics would require a book of its own; in this chapter I only
have room for a brief sketch that will help us through the rest of this book.
1.2 Causal Exclusion
Suppose that brain B is composed of a bunch of neurons, glia, other cells, and
extracellular fluid, which I’ll collectively label NN. For simplicity, I will refer to
NN simply as neurons. Different people disagree about the relationship between
wholes and their parts—e.g., between B and NN. Some think they are identical,
others that they are distinct, yet others that one of the two is grounded in the
other. We’ll suspend judgment for now.
Consider a voluntary action E performed by an organism, and consider two
causal claims that are entailed by what is found in any neuroscience textbook:
(1) B causes E.
(2) NN cause E.
When we make this kind of causal claim, we usually assume that a system
causes an effect in virtue of being in a particular state or having a particular
property. I assume that being in a state is the same as possessing a property, so we
can talk about states and properties interchangeably.
To take the role of properties into account, let’s reformulate (1) and (2) in terms
of brain state SB and neural state SNN, such that SNN realizes SB:
(3) SB causes E.
(4) SNN causes E.
Again, different people disagree about the relationship between properties of

wholes and properties of their parts—e.g., between SB and SNN. Some think they
are identical, others that they are distinct, yet others that one of the two is
grounded in the other. Again, we’ll suspend judgment for now.
When we make this kind of causal claim, we generally assume that each cause is
sufficient for the effect under the relevant background conditions. That is to say, if
SB causes E, then—given relevant background conditions—the occurrence of SB is
enough, all by itself, to bring about E. By the same token, if SNN causes E, then—
given relevant background conditions—the occurrence of SNN is enough, all by
itself, to bring about E.
The puzzle of causal exclusion arises because if SNN is already enough to cause
E, there seems to be nothing left for SB to do. Suppose that we establish (4)—that a
   9
neural state causes a voluntary action. Why would we say that a brain state also
causes that action, if we already established that a neural state is enough to cause
that very same action? The problem is entirely symmetrical. Suppose we establish
(3)—that a certain brain state causes a certain action. Why add that a neural state
also causes that action, if we already established that a brain state is enough to cause
that very same action? If one of the two is enough to produce the action, it seems
that the other is either identical to it or should be dispensed with. But which is it?
The problem generalizes. I picked neurons and brains among many other levels
that I could have picked. The same puzzle can be run with neural systems, neural
networks, molecules, atoms, and so on—not to mention whole organisms. This is
the puzzle of causal exclusion: if any level is enough, all by itself, to cause an effect
such as E, then other levels seem either dispensable or identical to the one causally
efficacious level (Kim 1998, 2005; Merricks 2001). Yet both laypeople and scien-
tists attribute causal efficacy to multiple levels of being, and we don’t seem to take
all levels to be identical to one another. What gives?
1.3 Overdetermination
One answer to the causal exclusion puzzle is that distinct levels cause the same
effect at the same time. This theoretical option is called overdetermination. To
understand overdetermination, consider the classic example of a firing squad,
each of whose members fires a deadly bullet. Given the operative background
conditions, each bullet is sufficient, all by itself, to kill a person. In this case, the
presence of a sufficient cause is compatible with many other sufficient causes for
the same effect. By analogy, levels of being might be like soldiers in a firing squad:
each level produces the same effect independently of the others (Loewer 2007;
Pereboom 2002; Schaffer 2003; and Sider 2003). Might this be a solution to the
causal exclusion puzzle?
There does seem to be a similarity between the two situations. In both cases,
multiple things occurring at the same time are said to be causally sufficient for the
same effect. Yet there are also profound dissimilarities.
In cases of ordinary causal overdetermination, such as the firing squad, the
different causes can be independently observed and tracked. For example, we can
check the different rifles before the squad shoots and see if they contain live
ammunition. After the shooting, we can search for distinct bullets and see where
they ended up. This sort of empirical investigation is not an option when different
levels of being are in play. We can’t investigate the effects of a brain independently
of those of the neurons that make up that very brain. We can’t observe or
manipulate the state of the brain, separately observe or manipulate the state of
the neurons that make up that same brain, and see if they have distinct causal
powers analogous to the powers of distinct bullets. Therefore, we cannot collect
10  
any evidence that there are distinct causal powers at distinct levels that are
sufficient for the same effect. Without evidence, it is both ontologically profligate
and epistemically irresponsible to posit them.
A defender of overdetermination may reply that the disanalogy between multi-
level overdetermination and ordinary—firing-squad-type—overdetermination is
immaterial. Distinct levels of being (wholes and their parts, higher-level properties
and their realizers) are metaphysically connected to one another in a way that
distinct bullets are not. Given the tight connection between levels, a defender of
overdetermination might insist that distinct levels may well be causally sufficient
for the same effect—at least in a loose-enough sense of “cause.”
This reply misunderstands the stakes of the problem. The question is not
whether a suitably watered-down notion of causation can underwrite claims to
the effect that distinct levels cause the same effect. The question is whether it is
legitimate to posit distinct causal powers that are causally sufficient for the same
effect in situations where it is impossible, as a matter of principle, to collect
empirical evidence of their distinct existence.
Even if we set aside epistemic considerations, it makes no sense to posit that
things that stand in a relation of composition or realization have distinct causal
powers for the same effect. The objects themselves—a whole and its parts—are not
even (wholly) distinct. The parts make up the whole; the whole consists of the
parts. By the same token, higher-level properties and their realizers are too
intimately connected to have distinct causal powers for the same effect. Once
you have the realizer, you also have the realized property; and once you have the
realized property (token), you also have its realizer. How could things that are as
intimately related as a whole and its parts, or a higher-level property and its
realizer, come to have distinct causal powers, each sufficient for the same effect
(cf. Bernstein 2016)? Later I will elucidate the metaphysical connections between
wholes and their parts as well as higher-level properties and their realizers, arguing
that such connections are inconsistent with overdetermination.
In summary, the overdetermination solution to the puzzle of causal exclusion
posits multiple distinct causes for the same effect that are possessed by a whole
and its parts, or by a higher-level property and its realizers. These are more causes
than are necessary to explain phenomena, without any evidence that they are there
or any way to investigate them separately. Because of this, overdetermination is a
desperate option. Perhaps we should consider it as a last resort if all other options
fail. Let’s ponder more appealing alternatives.
1.4 Reduction
The simplest solution to the puzzle of causal exclusion would be to identify the
one genuine cause of an effect and reduce all the other putative causes to that one.
   11
For example, suppose we established that neural states are the genuine cause of
voluntary actions. If we could reduce brains to collections of neurons, and brain
states to neural states, then we would conclude that there is no real conflict
between (1) and (2), or between (3) and (4), because (1) reduces to (2), and (3)
reduces to (4). There is only one genuine cause of E—neural states.
If we are going to consider this sort of reductionism, we need to say what we
mean by reduction. This is a vexed question in both metaphysics and the phil-
osophy of science (van Riel and van Gulick 2019). For present purposes, I will
adopt the simplest and clearest notion of ontological reduction: identity with a
direction. That is to say, X reduces to Y if and only if X is strictly identical to Y and
Y is more fundamental than X. Specifically, B reduces to NN if and only if B = NN
and NN is more fundamental than B, and SB reduces to SNN if and only if SB = SNN
and SNN is more fundamental than SB. The difficulties with reduction arise for any
reasonable notion of reduction; we might as well keep things manageable by
focusing on reduction as identity plus direction.⁴
It should go without saying that reduction in the relevant sense is not just a
matter of what is part of what. Everyone agrees that neurons are parts of brains,
molecules are parts of neurons (as well as brains), and every physical object is part
of the whole universe. Yet reasonable people disagree about whether large things
reduce to small things, small things reduce to the universe (e.g., Schaffer 2010), or
neither reduces to the other.
Assume reductionism holds. The first question is, which direction does reduc-
tion go? Do brains reduce to neurons or vice versa? Do neural states reduce to
brain states or vice versa? We’ve also seen that there aren’t just two levels—there
are many. We need to say where neural systems and their properties, people and
their properties, molecules and their properties, and so forth fit within the
reductionist picture. For any one of these levels, either it is the reduction base
for all the others, or it reduces to whichever level all others reduce to. For
reduction to work, we need to find the one fundamental level to which all other
levels reduce. I doubt we can.
In recent decades, perhaps the most popular solution is to pick the level of
elementary fermions and bosons as the fundamental ontological level, to which
everything else supposedly reduces. This does have some appeal—at least to those,
like me, with a taste for particle physics. Physicists themselves call elementary
fermions and bosons the fundamental particles. Maybe the particles that are
physically fundamental are also ontologically fundamental?⁵
⁴ A different proposal by Hemmo and Shenker (2015; Shenker unpublished) is that X reduces to Y if
and only if X is an aspect of Y. Their flat physicalism, which was developed independently of the present
work, goes in the direction of the egalitarian ontology I defend here and is more compatible with it than
traditional reductionism.
⁵ The precise ontological status of fermions and bosons, and of quantum mechanical systems in
general, is controversial. Debates include whether fermions and bosons are individual objects (French
12  
Before answering that question, let me point out that physical fundamentality
does not entail ontological fundamentality. Something is physically fundamental,
roughly, just in case it has no physical parts—no parts that are distinguishable via
physical operations or theories. Being physically fundamental is awesome and
important: theories that are true of physically fundamental particles are true of
everything physical. Nevertheless, physical fundamentality is not the same as
ontological fundamentality.
Something is ontologically fundamental just in case everything else synchron-
ically and asymmetrically depends on it. The ontologically fundamental stuff
grounds all the other stuff. The ontologically fundamental stuff is ontologically
prior to the other stuff. These are the metaphors used by metaphysicians; they are
not easy to pin down.
Ontological dependence is easy to grasp when time flows and new objects and
properties come into existence. For instance, if a group of people forms a new club,
the club ontologically depends on the temporally prior existence of the group of
people. The reason is obvious: without people, no club could be formed. The same
idea is also easy to grasp when we deal with arbitrary fusions of objects or
derivative properties. For example, the object formed by combining the Statue
of Liberty and the Taj Mahal depends ontologically on the existence of the
Statue of Liberty and the Taj Mahal; the property of being a mile away from
the Statue of Liberty depends ontologically on the existence and location of the
Statue of Liberty.
It is less clear what being ontologically more fundamental means when we
consider, at one and the same time, a whole versus its parts, or a whole’s properties
versus its parts’ properties, when the whole and its properties are individuated
independently of the parts and their properties.
In spite of this lack of clarity about what it means to be ontologically funda-
mental, lots of philosophers talk about one level being more fundamental than
another, or one level being absolutely fundamental. Ultimately the lack of clarity
about ontological fundamentality will not matter to us because I will argue that,
given any relevant notion of fundamentality, no level of being is more ontologic-
ally fundamental than any other. For now, though, let’s pretend we have at least a
faint grasp of what ontological fundamentality means and proceed to examine the
prospects of this idea. Could the ontologically fundamental level be the physically
fundamental level?
Elementary fermions and bosons are the building blocks of the universe: they
have no physical parts and they constitute every physical object. In addition,
2015) and whether quantum mechanical systems admit of unique decompositions (Healey 2013).
Addressing these debates falls outside the scope of this chapter. For present purposes, I assume that
fermions and bosons can be individuated by a combination of invariants and observables (Castellani
1998).
   13
according to the Standard Model of particle physics, the properties and inter-
actions of fermions and bosons constitutively explain all physical phenomena.
That includes, presumably, all higher-level phenomena such as brain states caus-
ing voluntary action. The Standard Model is by far the most successful and well-
confirmed scientific theory ever devised. That’s impressive. Thus, metaphysical
atomism—reducing higher levels of being to fundamental physical particles and
their properties—seems like a good reductionist option. But again, ontological
fundamentality does not follow from physical fundamentality.
Another popular reductionist option is to pick the whole universe as the
fundamental ontological level, to which everything else reduces (Schaffer 2010).
If we look at the history of philosophy, this monistic solution may be more popular
than atomism. One advantage of monism is that, by definition, the universe
includes everything within itself—all objects and properties. If we take the uni-
verse to be ontologically fundamental and reduce everything else to it, surely we
are not leaving anything out.
Another metaphysical advantage of monism is that it is less subject to the
vagaries of physical theory than metaphysical atomism. Some physicists hope to
replace the Standard Model with something even more physically fundamental,
which will unify all physical forces. To perform this unification, theoretical
physicists explore the possibility that there are entities even more physically
fundamental than fermions and bosons. If there are such fundamental physical
entities, then perhaps everything reduces to them. But what if there are no true
atoms—no indivisible physical entities? What if every physical entity has smaller
and smaller parts all the way down? This seems to make reduction impossible, for
lack of a reduction base. Monism—the view that everything reduces to the whole
universe—does not seem to face this sort of risk (cf. Schaffer 2010).
Still, monism shares with atomism the problem that picking any level as a
reduction base seems arbitrary. Aside from the puzzle of causal exclusion, it’s unclear
why any level should be more ontologically fundamental than any other. And the
puzzle of causal exclusion is entirely neutral on which level is more fundamental. It
could be the whole universe, the level of fermions and bosons, the level of brains and
other mid-size objects, the level of molecules (cf. Bickle 2003), or any level in
between. In addition to this arbitrariness, reductionism seems to be stuck with lack
of testability: even if there is a fundamental level, it’s unclear how to find it.
Another drawback is that reductionism gives the impression that some sciences
are more important than others. Specifically, whichever science investigates the
fundamental level—the genuine causes—seems to have an edge over the other
sciences. After all, other sciences are just studying stuff whose causal powers
depend on reducing it to the fundamental stuff. Human nature being what it is,
many people don’t like that their favorite subject matter, to which they devote
their working hours, is just derivative on something more fundamental, which is
studied by someone else. Plus, if everything reduces to the fundamental level, so
14  
do people. Many find it personally insulting to be reduced to something else—it

makes them feel deprived of their personal agency. This is not a substantive
objection; nonetheless, it makes reductionism unpopular.
Most of the people who like reductionism are those who claim to be studying
the fundamental stuff. That includes some neuroscientists, to be sure—most
commonly molecular and cellular neuroscientists. The reason they like reduction-
ism is that they focus more on reducing higher levels to cells and molecules than
on reducing cells and molecules to fermions and bosons. Be that as it may, a lot of
philosophers and scientists who study higher-level entities hate reductionism. If
they could find a viable alternative, they’d jump on it.
1.5 Anti-Reductionism
Many advocates of the special sciences—especially psychology—defend some

form of anti-reductionism (e.g., Fodor 1974; Pereboom and Kornblith 1991;
Pereboom 2002; Gillett 2002, 2010; List and Menzies 2009). According to standard
anti-reductionism, higher-level entities and properties are both causally effica-
cious and distinct from the lower level entities and properties that compose and
realize them. For instance, minds are distinct from collections of neurons; mental
states are distinct from neural states. Talking about minds and mental states raises
special questions that would be distracting here, so let’s stick with unequivocally
physical systems and properties. In the next few chapters I will argue that this
restriction makes no difference for our purposes. According to standard anti-
reductionism, then, brains are distinct from collections of neurons; brain states are
distinct from neural states. Each type of entity and property has its own distinct
causal powers. This is the kind of anti-reductionism I discuss in this section.
If anti-reductionism is correct, each scientific discipline can study its subject
matter in the hope of finding something real in its own right, including real causal
relations. Scientific explanations are autonomous from one another: no science is
hegemonic and each scientific community can do their empirical investigations at
their own level without getting in each other’s way. That might explain why anti-
reductionism is popular among those who study higher levels.
One crucial question for anti-reductionists is whether a whole’s properties
endow it with causal powers that go beyond the causal powers possessed by the
organized parts that make up that very whole. In other words, suppose that a
whole object O is made out of parts O₁, O₂, . . . On standing in appropriate
relations. Does O have causal powers that O₁, O₂, . . . On, when they are organized
to form O, collectively lack? A positive answer to this question yields a version of
anti-reductionism known as strong synchronic emergence.
Strong synchronic emergence is sometimes confused with the uncontroversial
claim that, when a set of objects enter new organizational relations among them,
   15
their properties change. Objects that organize together in specific physical

configurations possess causal powers that they lack when they are isolated from
one another. For instance, Quantum Mechanics teaches us that particles can
become entangled with each other; when they are so entangled, their properties
change; as a result, their behavior is different than when they are untangled. Novel
physical ties lead to novel properties. No one disputes this sort of diachronic
emergence, because there is nothing especially mysterious about it. Here, we are
not concerned with diachronic emergence at all.
Genuine strong synchronic emergence is not a diachronic relation between
unorganized parts and organized wholes—it is a synchronic relation between organ-
ized parts and the wholes they compose. More precisely, strong synchronic emer-
gence is a relation that holds between the properties of an object and the properties of
its parts at one and the same time. In our example, strong synchronic emergence
does not say that entangled particles have causal powers that untangled particles
lack—that’s diachronic emergence and it’s uncontroversial. Strong synchronic emer-
gence says that, at one and the same time, the entangled particle system—the
whole—has causal powers that the entangled particles—the parts of that very same
whole, entangled as they are and considered collectively—lack.
It is unbearably mysterious how, after a plurality of objects has become
organized, the whole could possess causal powers that are lacked by the organized
parts taken collectively. Yet this is what strong emergentists claim.⁶ Needless to
say, no one has been able to make any sense of how this could be—how a higher-
level property could possibly endow an object with causal powers that its lower-
level realizers—considered collectively at the same time—lack. Or, if a strong
emergentist were to deny that higher-level properties are wholly realized by lower-
level ones, no one has been able to make sense of where the unrealized portion of
the property comes from, or how it relates to the properties of the object’s parts.
Extraordinary claims require extraordinary evidence. Yet when strong emer-
gentists purport to give evidence, they fall short. They offer two sorts of examples.
On one hand are examples of entities that acquire new powers and hence new
behaviors by entering into new organizational relations (e.g., untangled particles
that become entangled). This is not evidence of strong synchronic emergence but
of good old diachronic emergence, which virtually no one disputes because it
poses no ontological puzzles. On the other hand are higher-level capacities or
generalizations for which we lack lower-level mechanistic explanations. This is not
evidence of strong emergence either—it is evidence of our epistemic limitations
and it is consistent with the denial of strong synchronic emergence. In fact, it is
unclear how anyone could possibly show that, at one and the same time, a whole
has causal powers that its organized parts, considered collectively, lack. On the
⁶ Recent defenses of strong synchronic emergentism include Mitchell 2009; Thalos 2013; Gillett
2016. By contrast, Humphreys 2016 focuses primarily on diachronic emergence.
16  
contrary, the historical track record indicates that higher-level phenomena can be
explained without positing novel higher-level causal powers (e.g., McLaughlin
1992; Papineau 2001). In light of these deficiencies, I set strong synchronic
emergence aside and return to more sensible forms of anti-reductionism.⁷
According to ordinary anti-reductionism, wholes and their properties are distinct
from parts and their properties, although their causal powers do not surpass those
of their parts.
The most popular argument for anti-reductionism is based on multiple realiz-
ability (Fodor 1974; Polger and Shapiro 2016). Multiple realizability is the claim
that at least some higher-level properties—the kind of property putatively studied
by the special sciences—can be realized in different ways. For example, there are
different ways of making a mousetrap, an engine, or a computer. All the different
realizers share higher-level properties: they catch mice, generate motive power, or
perform computations. Furthermore, according to anti-reductionism, the differ-
ent realizers of a higher-level property share no lower-level property. Because of
this, says anti-reductionism, higher-level properties such as catching mice, gen-
erating motive power, or computing cannot reduce to any lower-level property.
This argument for anti-reductionism from multiple realizability faces a
dilemma (cf. Kim 1998, 2005). To set up the dilemma, consider any putatively
multiply realizable higher-level property P—such as catching mice, generating
motive power, or computing—and its putative realizers R₁, R₂, . . . Rn. Let’s assume,
along with most participants in this debate, that lower-level realizers are more
fundamental than the higher-level properties they realize. Let’s rule out
overdetermination—the unwarranted, ontologically profligate view that both
P and its realizers are independently causally sufficient for P’s effects (per
Section 1.3). Let’s also rule out strong synchronic emergence, the mysterious
and unsubstantiated claim that P’s causal powers outstrip those of its realizers.
From the absence of strong emergence, it follows that any realizer of a higher-level
property—e.g., any Ri—is causally sufficient for P’s effects.
The dilemma is this: either P is merely a disjunction of lower-level properties—
something of the form R₁ or R₂ or . . . Rn, where R₁ . . . Rn are its realizers—or P is a
genuine property with genuine causal powers. If P is merely a disjunction of
lower-level realizers, all the genuine causal work is done by the realizers, so
multiple realizability is just a relation between a disjunction and its disjuncts.
This form of multiple realizability is consistent with reductionism, because all the
causal work is done by the realizers and there is nothing distinctive about higher
levels. So anti-reductionism fails. On the other hand, if P is a genuine property,
then its causal powers must be identical—and thereby reduce—to those of its
realizers. For we already established that P has no additional causal powers beyond
⁷ For more detailed critiques of strong emergentism, see McLaughlin 1992; Kim 2006.
Another random document with
no related content on Scribd:
third millennium as a political deity, the war-god, and leader of the
people, as real a personality as Hammurabi himself. The same is
true of Asshur, once the local deity of the aboriginal land of the
Assyrians, but later raised by the imperial expansion of this people
almost to the position of a universal god, the guardian of the land,
the teacher and the father of the kings; nor can we discern that he
was ever an elemental god.
Speaking generally, in spite of many important differences, we
may regard the religious structure to which the cults of Anatolia and
Egypt belonged as morphologically the same as that which I am
defining as Hellenic. Also, among all these peoples, by the side of
the few higher deities who have developed moral personalities, we
find special elemental divinities, as, in Hellas, we find Helios and the
deities of the wind, Hephaistos the fire-god.
The distinction between the religions of the Hellenes and “the
barbarians,” which Aristophanes defines as the difference between
the worship of ideal divine personages, such as Zeus, Apollo, and
Demeter, and the direct worship of elementary powers, such as sun
and moon, is not borne out by modern research. Where we find sun-
worship or moon-worship in the East, it does not appear to have
been directed immediately to the thing itself regarded as a living or
animate body, but to a personal god of the sun or the moon—Bel,
Shamash, or Sin. We can only distinguish the Greek from the
Oriental in respect of Nature-religion by the lesser degree of
devotion that the Hellene showed to it. Only those of his divinities
whose names connoted nothing in the material or natural world,
could develop into free moral personalities, and dominate the
religious imagination of the people. Nowhere, for instance, had
Helios any high position in the Greek world except at Rhodes, where
we must reckon with pre-Hellenic, Minoan, and later with Semitic
influences. Therefore, when, shortly before and after the beginning
of the Graeco-Roman period, a wave of sun-worship welled from the
East over the West, it may have brought with it religious ideas of
high spirituality and ethical purity, yet by the race-consciousness of
the Hellenes it must have been judged to be a regress towards a
barbaric past.
The instinct of the Greek in his creation of divine forms shows
always a bias towards the personal and the individual, an aversion to
the amorphous and vague, and herein we may contrast him with the
Persian and Egyptian. A certain minor phenomenon in these
religions will illustrate and attest this. All of them admitted by the side
of the high personal deities certain subordinate personages less
sharply conceived, divine emanations, as we may sometimes call
them, or personifications of moral or abstract ideas. Plutarch
specially mentions the Persian worship of Truth, Goodwill, Law-
abidingness, Wisdom, emanations of Ahura-Mazda, which in the
light of the sacred books we may, perhaps, interpret as the Fravashis
or Soul-powers of the High God; and in certain Egyptian myths and
religious records we hear of a personification of Truth, whose statue
is described by the same writer. But at least in the Persian system
we may suspect that such divine beings had little concrete
personality, but, rather, were conceived vaguely as daimoniac forces,
special activities of divine force in the invisible world. Now the Greek
of the period when we really know him seems to have been mentally
unable to allow his consciousness of these things or these forces to
remain just at that point. Once, no doubt, it was after this fashion that
his ancestors dimly imagined Eros, or the half-personal Curse-power
Ἀρά; but he himself could only cherish Eros under the finished and
concrete form of a beautiful personal god, and the curse was only
vitalised for him when it took on the form of the personal Erinys. This
topic is a fruitful one, and I hope to develop it on a later occasion.
It suggests what is now the next matter I wish to touch on—the
comparison of the Mediterranean religions in respect of their
anthropomorphism. Philosophically, the term might be censured as
failing to distinguish any special type of religion; for we should all
admit that man can only envisage the unseen world in forms
intelligible to his own mind and reflecting his own mental structure.
But, apart from this truism, we find that religions differ essentially and
vitally according as this anthropomorphism is vague and indefinite or
sharply defined and dominating; according as they picture the
divinity as the exact though idealised counterpart of man, and
construct the divine society purely on the lines of the human, or
refrain from doing this either through weakness and obscurity of
imagination or in deference to a different and perhaps more elevated
law of the religious intellect. Now, of the Hellenic religion no feature
is salient as its anthropomorphism, and throughout its whole
development and career the anthropomorphic principle has been
more dominating and imperious than it has ever been found to be in
other religions.11.1 At what remote period in the evolution of the
Hellenic mind this principle began in force, what were the influences
that fostered and strengthened it, in what various ways it shaped the
religious history of the Hellenic people, are questions that I may be
able to treat more in detail in the future. But there are two important
phenomena that I will indicate now, which we must associate with it,
and which afford us an illuminating point of view from which we may
contrast the Greek world and the Oriental. In the first place, the
anthropomorphic principle, combining with an artistic faculty the
highest that the world has known, produced in Greece a unique form
of idolatry; and, in the second place, in consequence chiefly of this
idolatry, the purely Hellenic religion remained almost incapable of
that which we call mysticism.
Now, much remains still to be thought out, especially for those
interested in Mediterranean culture, concerning the influence of
idolatry on religion; and not only the history, but the psychology of
religion, must note and estimate the influence of religious art. It may
well be that the primitive Greeks, like the primitive Roman, the early
Teuton, and Indo-Iranian stocks, were non-idolatrous, and this
appears to have been true to some extent of the Minoan culture.
Nevertheless, the Mediterranean area has from time immemorial
been the centre of the fabric and the worship of the eikon and the
idol. The impulse may have come from the East or from Egypt to the
Hellene; he in his turn imparted it to the Indian Aryans, as we now
know, and in great measure at least to the Roman, just as the
Assyrian-Babylonian temple-worship imparted it to the Persian.
Nowhere, we may well believe, has the influence of idolatry been so
strong upon the religious temperament as it was upon that of the
Hellenes; for to it they owed works of the type that may be called the
human-divine, which surpass any other art-achievement of man.
I can here only indicate briefly its main effects. It intensified the
perception of the real personal god as a material fact. It increased
polytheism by multiplying the separate figures of worship, often,
perhaps, without intention. It assisted the imagination to discard what
was uncouth and terrifying in the Hellenic religion, and was at once
the effect and the cause of the attachment of the Hellenic mind
towards mild and gracious types of godhead. The aniconic emblem
and uncouth fetich-formed figures were here and there retained,
because of vague ideas about luck or for superstitious fetichistic
reasons; but the beautiful idol was cherished because it could
arouse the enthusiastic affection of a sensitive people, and could
bring them to the very presence of a friendly divine person. The
saying that the Olympian deities died of their own loveliness means
a wrong interpretation of the facts and the people. But for a beautiful
idolatry, Hellenic polytheism would have passed away some
centuries before it did, the deities fading into alien types or becoming
fused one with the other. Nor was its force and influence exhausted
by the introduction of Christianity, for it shaped the destinies of the
Greek Church, and threw down a victorious challenge to the
iconoclastic Emperors.
If now we were to look across the Mediterranean, and could
survey the religious monuments of Persia, Assyria and Babylonia,
Phoenicia, and the Hittite people, we should find a general
acceptance of the anthropomorphic idea. The high personal deities
are represented mainly in human form, but the art is not able to
interpret the polytheistic beliefs with skilfully differentiated types. In
Chaldaic and Assyrian art one type of countenance is used for
various divinities, and this such as might inspire awe rather than
affection. And the anthropomorphism is unstable. Often animal traits
appear in parts of the divine figure. Nergal has a lion’s head; even
the warrior Marduk is invoked in the mystic incantations as “Black
Bull of the Deep, Lion of the dark house.”14.1 In fact, over a large part
of anterior Asia, anthropomorphism and theriomorphism exist side by
side in religious concept and religious art. We may say the same of
Egypt, but here theriomorphism is the dominating factor.
As regards the explanation of this phenomenon, many questions
are involved which are outside my present province. I would only
express my growing conviction that these two distinct modes of
representing the divine personage to the worshipper are not
necessarily prior and posterior, the one to the other, in the evolution
of religion. They can easily, and frequently do, coexist. The vaguely
conceived deity shifts his shape, and the same people may imagine
him mainly as a glorified man of human volition and action, and yet
think of him as temporarily incarnate in an animal, and embody his
type for purposes of worship or religious art in animal forms.
I would further indicate here what I cannot prove in detail—that
theriomorphism lends itself to mysticism, while the anthropomorphic
idolatry of Greece was strongly in opposition to it. The mystic
theosophy that pervaded later paganism, and from which early
Christianity could not escape, originated, as Reitzenstein has well
shown, mainly in Egypt, and it arose partly, I think, in connection with
the hieratic and allegorical interpretation of the theriomorphic idol.
There was nothing mystic about the Zeus of Pheidias, so far as the
form of the god was concerned. The forms were entirely adequate to
the expression of the physical, moral, and spiritual nature of the god.
The god was just that, and there was nothing behind, and, as the
ancient enthusiast avers, “having once seen him thus, you could not
imagine him otherwise.” But when a divinity to whom high religious
conceptions have already come to attach is presented, as it might be
in Egyptian religious art, with the head of a jackal or an ape, the
feeling is certain to arise sooner or later in the mind of the
worshipper that the sense-form is inadequate to the idea. Then his
troubled questioning will receive a mystic answer, and the animal
type of godhead will be given an esoteric interpretation.
Plutarch, in the De Iside et Osiride,15.1 is one of our witnesses. He
finds a profounder significance for theosophy in the beetle, the asp,
and the weasel than in the most beautiful anthropomorphic work of
bronze or marble. He here turns his back on his ancestors, and goes
over to the sect of the Egyptian mystic.
But the most curious testimony to my thesis is borne by an
inscription on an Egyptian lamp—an invocation of the God Thoth: “O
Father of Light, O Word (λόγος) that orderest day and night, come
show thyself to me. O God of Gods, in thy ape-form enter.”15.2 Here
the association of so mystic a concept as the “Logos,” the divine
Reason, an emanation of God with the form of an ape, is striking
enough, and suggests to us many reflections on the contrast
between the Egyptian theriomorphism and the human idolatry of the
Greek. The Hermes of Praxiteles was too stubborn a fact before the
people’s eyes to fade or to soar into the high vagueness of the
“Logos,” too stable in his beautiful humanity to sink into the ape.
But before leaving this subject I would point out a phenomenon in
the Hellenic world that shows the working of the same principle. The
Orphic god Dionysos-Sabazios-Zagreus was πολύμορφος, a shape-
shifter, conceived now as bull, now as serpent, now as man, and the
Orphic sects were penetrated with a mystic theosophy; and, again,
they were a foreign element embedded in Greek society and religion.
While we were dealing with the subject of anthropomorphism, we
should consider also the question of sex, for a religion that gives
predominance to the god is certain to differ in some essential
respects from one in which a goddess is supreme. Now, although the
conception of an All-Father was a recognised belief in every Greek
community, and theoretically Zeus was admitted to be the highest
god, yet we may believe Athena counted more than he for the
Athenians, and Hera more for the Argives. And we have evidence of
the passionate devotion of many urban and village communities to
the mother Demeter and her daughter Kore, to whom the greatest
mysteries of Greece, full of the promise of posthumous salvation,
were consecrated. Also, in the adjacent lands of earlier culture we
mark the same phenomenon. In Egyptian religion we have the
commanding figure of Isis, who, though by no means supreme in the
earlier period, seems to dominate the latter age of this polytheism. In
the Assyrian-Babylonian Pantheon, though the male deity is at the
head, Ishtar appears as his compeer, or as inferior only to Asshur.
Coming westward towards Asia Minor, we seem to see the goddess
overshadowing the god. On the great Hittite monument at Boghaz-
Keui, in Cappadocia, skilfully interpreted by Dr. Frazer, we observe a
great goddess with her son coequal with the Father-God. In the
lands adjacent to the coast a Mother-Goddess, sometimes also
imagined as virgin, Kybele of Phrygia, Ma of Cappadocia, Hipta of
Lydia, Astarte of Askalon, Artemis of Ephesos who was probably a
blend of Hellenic and Oriental cult-ideas, appears to have been
dominant from an immemorial antiquity; and Sir Arthur Evans has
discovered the same mysterious feminine power pre-eminent in the
Minoan religion. We may even affirm that she has ruled a great part
of the Mediterranean down to the present day.
The various questions suggested by this predominance of
goddess-worship are fascinating and subtle. The sociological one—
how far it is to be connected with a system of counting descent
through the female, with a matrilinear society—I have partly
discussed elsewhere.17.1 I may later be able to enter on the question
that is of more interest for the psychology of religion—the effect of
such worships on the religious sentiment. Here I can merely point to
the phenomenon as a natural and logical product of the principle of
anthropomorphism, but would call attention to the fact that in the
East it sometimes developed into a form that, from the
anthropomorphic point of view, must be called morbid and
subversive of this principle; for the rivalry of divine sex was here and
there solved by the fusion of the two natures in the divinity, and we
find a bisexual type—a male Astarte, a bearded Ishtar.17.2 The
healthy-minded anthropomorphism of the Hellene rejects this
Oriental extravagance.
If we now could consider in detail the various moral conceptions
attached to the high State divinities of Greece and the East, we
should be struck with a general similarity in the point of view of the
various culture-stocks. The higher deities, on the whole, are ethical
beings who favour the righteous and punish transgressors; and the
worship of Greece falls here into line with the Hebraic conceptions of
a god of righteousness. But in one important particular Hellenic
thought markedly differs from Oriental, especially the Persian. In the
people’s religion throughout Hellas the deities are, on the whole,
worshipped as beneficent, as doing good to their worshippers, so
long as these do not offend or sin against them. The apparent
exceptions are no real exceptions. Ares may have been regarded as
an evil god by the poet or the philosopher, but we cannot discover
that this was ever the view of the people who cared to establish his
cult. The Erinyes are vindictive; nevertheless, they are moral, and
the struggle between them and Apollo in the Aeschylean drama is
only the contest between a more barbaric and a more civilised
morality. In the list of Greek divine titles and appellatives, only one or
two at most can be given a significance of evil.
Doubtless, beneath the bright anthropomorphic religion lurked a
fear of ghosts and evil spirits, and the later days of Hellenic
paganism were somewhat clouded with demonology. But the
average Greek protected himself sufficiently by purification and easy
conventional magic. He did not brood on the principle of evil or
personify it as a great cosmic power, and therefore he would not
naturally evolve a system of religious dualism, though the germs
from which this might grow may be found in Orphic tradition and
doctrine. Contrast this with the evidence from Egypt, Assyria, and
Persia. The Egyptian and Assyrian records bear strong impress of
the prominence and power of the belief in evil spirits. The high gods
of Assyria were continually being invoked and implored by the
worshipper to save him from the demons, and one of these, Ira, a
demon of pestilence, seems to have received actual worship; and
much of Egyptian private ritual was protective magic against them.
But nowhere did the power of evil assume such grand proportions
as in the old Mazdean creed of Persia, and the dualism between the
good and evil principle became here the foundation of a great
religion that spread its influence wide through the West. The religion
had its prophet, Zarathustra, in whose historic reality we ought not to
doubt. In his system the faithful Mazdean is called upon to play his
part in the struggle between Ahura-Mazda and Angra-Mainyu, and
this struggle continues through the ages till in the final cataclysm the
Daevas, or evil demons, will be overthrown. We note here that this
faith includes the idea of a final Judgment, so familiar to Judaic and
Christian thought, but scarcely to be found in the native Hellenic
religion. Further, it should be observed that the Mazdean dualism
between good and evil has nothing in common with the Platonic
antithesis between mind and sense, or St. Paul’s between spirit and
flesh, or with the hatred of the body that is expressed in Buddhism.
The good Mazdean might regard his body as good and pure, and
therefore he escaped, as by a different way did the Greek, from the
tyranny of a morbid asceticism. Only he developed the doctrine of
purity into a code more burdensome than can be found, I think,
elsewhere. The ideas of ritual-purity on which he framed this code
are found broadcast through the East and in Egypt, and appear in
the Hellenic religion also. The Greek, however, did not allow himself
to be oppressed by his own cathartic system, but turned it to
excellent service in the domain of law, as I have tried to show
elsewhere.20.1
Generally, as regards the association of religion and morality, we
find this to be always intimate in the more developed races, but our
statistics are insufficient for us to determine with certainty the
comparative strength of the religious sanction of morals in the
ancient societies of the Mediterranean. The ethical-religious force of
the Zarathustrian faith seems to approach that of the Hebraic. We
should judge it to be stronger, at least, than any that was exercised
in Hellas, for Hellas, outside the Orphic sects, had neither sacred
books of universal recognition nor a prophet. Yet all Hellenic morality
was protected by religion, and the Delphic oracle, which occasionally
was able to play the part of the father-confessor, encouraged a high
standard of conduct—as high as the average found elsewhere in the
ancient world. We may note, however, one lacuna in the Hellenic
code: neither Greek ethics, on the whole, nor Greek religion,
emphasised or exalted or deified the virtue of truth; but we hear of a
goddess of Truth in Egypt, and it becomes a cardinal tenet and a
divine force in the Zarathustrian ideal.
Again, in all ancient societies religion is closely interwoven with
political, legal, and social institutions, and its influence on these
concerns the history of the evolution of society and law. It is only in
modern society, or in a few most ideal creeds at periods of great
exaltation, that a severance is made between Cæsar and God. Save
Buddhism, the religions of the ancient societies of the East and of
Egypt were all in a sense political. Darius regards himself as
specially protected by Ahura-Mazda, and we are told by
Herodotus21.1 that in the private Persian’s prayers no separate
personal benefits were besought, but only the welfare of the King
and the Persian community. The gods of Assyria inspire counsel and
order the campaign; Shamash, the sun-god, is “Just Ruler” and the
“Lord of Law”; and Ninib is styled the god “who lays for ever the
foundation-stone of the State,” and who, like Zeus ὅριος, and the
Latin Terminus, “protects the boundaries of the cornfield.” The Syrian
goddess of Bambyke, Kybele of Phrygia, Astarte of Askalon, all wear
the mural crown, the badge of the city goddess. But I doubt if our
materials are as yet rich enough to inform us in what precise way
religion played a constructive part in the oldest civilisations,—
namely, those of Assyria and Egypt. We may observe that the code
of Hammurabi, our oldest legal document, is curiously secular and in
many respects modern.
The question can be most fruitfully pursued in the study of the
Greek societies; for no other religion of which we have any record
was so political as the Hellenic, not even, as I should judge, the
Roman, to which it bears the closest resemblance in this respect.
The very origin of the πόλις, the city-state, was often religious; for
the name or title of the deity often gave a name to the city, and the
temple was in this case probably the centre of the earliest residence.
In the organised and complex Greek societies, every institution of
the State—the assembly, the council, the law-courts, the agrarian
economy, all the regulations of the family and clan—were
consecrated and safeguarded by the supervision of some deity.
Often he or she was worshipped as in a literal sense the State
ancestor, and in one of the temples might be found burning the
perpetual fire which symbolised the permanence of the city’s life.
And in Greece we find a unique phenomenon, which, though small,
is of great significance—the deity might here and there be made to
take the office and title of a civic magistrate. For instance, Apollo
was στεφανηφόρος in Asia Minor cities, and in the later days of
Sparta, as the recent excavations have shown, the ghost of
Lykourgos was elected as the chief inspector of the education of the
young.
To the superficial observer, then, the Greek civic society might
appear a theocracy. But such a view would imply ignorance of the
average character of the ancient Greek world. There can be no
theocracy where there is no theocrat. In Asia Minor the priest might
be a great political power, but in Greece this was never so. Here the
political, secular, utilitarian interest dominates the religion. The high
divinities become politicians, and immersed in secular affairs, and
even take sides in the party strife, as some of the religious titles
attest. Thus Greek religion escaped morbidity and insanity,
becoming genial and human, and compensating by its adaptability to
the common needs of social life for what it lacked of mystery and
aloofness. Therefore, also, in Greek invocations and hymns we do
not often hear the echo of that sublimity that resounds in the Iranian,
Assyrian, and still more in the Hebrew liturgies.
Another interesting point of comparison is the relation of religion to
the arts and sciences. Their association may be said to have been
more intimate in Hellenism than it has been found to be in any other
creed. We can estimate what music and the drama owed to Apollo
and Dionysos, and how the life of the philosopher, artist, and poet
was considered consecrated to certain divinities. We hear of the
Delphic oracle encouraging philosophic pursuits. The name
“Museum” is a landmark in the religious history of education, and we
know that the temple of Asklepios in Kos was the cradle of the
school of modern medicine. The records of the other religions of this
area show glimpses of the same association, and more extended
research may throw further light on it. The Babylonian gods Nebo
and Ea were divinities of wisdom and the arts, and to the former,
who was the inventor of writing, the library of Ashurbanapal was
consecrated. Chaldean astronomy was evolved from their astrology,
which was itself a religious system. But demonology was stronger in
Assyria, Persia, and Egypt than in Hellas, and demonology is the foe
of science. In the Zend-Avesta the priestly medicine-man, who heals
by spell and exorcism, is ranked higher than the scientific
practitioner. A chapter might be written on the negative advantages
of Greek religion, and none was of greater moment than this—that it
had no sacred books or authoritative religious cosmogony to oppose
to the dawn and the development of scientific inquiry. Asklepios had
been a practitioner in the method of thaumaturgic cures, but he
accepted Hippokrates genially when the time came.
As regards the relation between Greek philosophy and Greek
religion, something may remain to be discovered by any scholar who
is equally familiar with both. It would be absurd to attempt to
summarise the facts in a few phrases here. I wish merely to indicate
the absence in pure Hellenic speculation of any elaborated system of
theosophy, such as the late Egyptian “gnosis,” till we come to Neo-
Platonism, when the Greek intellect is no longer pure. We discover
also a vacuum in the religious mind and nomenclature of the earlier
Greek: he had neither the concept nor any name to express the
concept of what we call “faith,” the intellectual acceptance and
confessional affirmation of certain dogmas concerning the divinity;
and in this respect he differed essentially not only from the Christian,
but also from the Iranian and Buddhistic votary.
A great part of the study of ancient religion is a study of ritual, and
it is interesting to survey the Mediterranean area, so as to discern
similarities or divergencies in the forms of religious service.
Everywhere we observe the blood-sacrifice of animals, and very
frequently the harmless offering of fruits and cereals, and now one,
now the other, in Greece as elsewhere, was regarded as the more
pious. The former is of the higher interest, for certain ideas which
have been constructive of higher religions—our own, for example—
have grown out of it. At first sight the animal oblation seems
everywhere much the same in character and significance. The
sacrificial ritual of Leviticus does not differ in any essential trait from
that which commended itself to the Greeks and the other peoples of
these lands. Certain animals are everywhere offered, at times as a
free and cheerful gift, at other times as an atonement to expiate sin
and to deprecate wrath. Certain other animals are tabooed, for
reasons that may repay searching out.
In most regions we have evidence of the practice of human
sacrifice, either as an established system or as an occasional
expedient. The motives that prompted it present an important and
intricate question to the modern inquirer. The two nations that grew
to abhor it and to protest against it were the Hebrew and the Greek,
though the latter did not wholly escape the taint of it; for he had
inherited the practice from his ancestral past, and he found it
indigenous in the lands he conquered. Repellent as the rite may be,
it much concerns the study of the religions of the cultured races.
Now, an interesting theory concerning sacrifice was expounded
and brought into prominent discussion by Professor Robertson Smith
in his Religion of the Semites, and in an earlier article in the
Encyclopædia Britannica—namely, that a certain type of ancient
sacrifice was a mystic sacramental communion, the worshipper
partaking of some sacred food or drink in which the spirit of the deity
was temporarily lodged. This mystic act, of which there is no clear
trace in the Old Testament, is reported from Egypt,25.1 and it appears
to have been part of the Attis ritual of Phrygia. We find doubtful
traces of it in the Eleusinian and Samothracian mysteries; also a
glimpse of it here and there in the public religion of Hellas. But it is
best attested as a potent force in the Dionysiac worship, especially in
a certain savage ritual that we may call Thracian, but also in the
refined and Hellenised service as well.
I cannot dwell here on the various aspects of this problem. The
Hellenic statistics and their significance I have partly collected and
estimated in a paper published some years ago.25.2 The application
of the sacramental idea to the explanation of the Eleusinian
mysteries, ingeniously attempted by Dr. Jevons, I have discussed in
the third volume of my Cults of the Greek States; and the Dionysiac
communion-service is considered at length in the fifth.
The attractiveness of the mystic appeal of the Sacrament appears
to have increased in the later days of paganism, especially in its
period of struggle with Christianity. That strangest rite of the expiring
polytheism, the ταυροβόλιον, or the baptism in bull’s blood, in the
worship of Kybele, has been successfully traced back by M. Cumont
to the worship of the Babylonian Anaitis. The sacramental concept
was the stronghold of Mithraism, but can hardly be regarded as part
of its heritage from Persia, for it does not seem to have been familiar
to the Iranian religion nor to the Vedic Indian. In fact, the religious
history of no other Aryan race discloses it with clearness, save that
of the Thraco-Phrygian and Hellenic. Was it, then, a special product
of ancient “Mediterranean” religious thought? It would be important
to know, and Crete may one day be able to tell us, whether King
Minos took the sacrament. Meantime, I would urge upon those who
are studying this phenomenon in the various religions the necessity
of precise definition, so as to distinguish the different grades of the
sacramental concept, for loose statements are somewhat rife about
it.
Apart from the ritual of the altar, there is another mode of attaining
mystic union with the divinity—namely, by means of a sacred
marriage or simulated corporeal union. This is suggested by the
initiation formulæ of the mysteries of Attis-Kybele. The cult of Kybele
was connected with that of the Minoan goddess, and the strange
legend of Pasiphae and the bull-god lends itself naturally to this
interpretation. The Hellenic religion also presents us with a few
examples of the holy marriage of the human bride with the god, the
most notable being the annual ceremony of the union of the
“Queen,” the wife of the King Archon, at Athens, with Dionysos. And
in the mysteries of later paganism, as well as in certain forms and
symbolism of early Christianity, Professor Dieterich has traced the
surviving influence of this rite.
Among all the phenomena of ritual, none are more interesting or in
their effects more momentous than the rites that are associated with
the dogma of the death of the divinity. That the high gods are
naturally mortal and liable to death is an idea that is certainly rare,
though it may be found in Egyptian and old Teutonic mythology; but
the dogma of the annual or periodic death and resurrection of the
divinity has been, and is, enacted in much peasant ritual, and
worked for the purposes of agrarian magic in Europe and elsewhere.
More rarely we find the belief attached to the mystic forms and faith
of some advanced religion, and it is specially in the Mediterranean
area where it appears in a high stage of development. It is a salient
feature of the Egyptian worship of Isis; of the Sumerian-Babylonian
ritual, in which the dead Thammuz was bewailed, and which
penetrated Syria and other parts of Asia Minor; of the worship of
Attis and Adonis in Phrygia and the Lebanon; and of certain shrines
of the Oriental Aphrodite. It is associated often with orgiastic sorrow
and ecstatic joy, and with the belief in human immortality of which
the resurrection of the deity is the symbol and the efficacious means.
This idea and this ritual appears to have been alien to the native
Hellenic religion. The Hellenic gods and goddesses do not die and
rise again.
Only in one Aryan nation of antiquity, so far as I am aware, was
the idea clear and operative—the Thraco-Phrygian, in the religion of
Dionysos-Sabazios. This alien cult, when transplanted into Greece,
retained still some savagery in the rite that enacted the death of the
god; but in the Orphic sects the ritual idea was developed into a
doctrine of posthumous salvation, from which the later pre-Christian
world drew spiritual comfort and some fertile moral conceptions. This
Thracian-Dionysiac influence in Hellas, though chastened and
sobered by the sanity of the national temperament, initiated the
Hellene into a certain spiritual mood that was not naturally evoked by
the native religion; for it brought into his polytheism a higher
measure of enthusiasm, a more ecstatic spirit of self-abandonment,
than it possessed by its own traditional bent. Many civilised religions
appear to have passed beyond the phase of orgiastic fervour. It
emerges in the old Egyptian ritual, and most powerfully in the religion
of Phrygia and of certain districts of Syria; but it seems to have been
alien to the higher Semitic and the Iranian religions, as it was to the
native Hellenic.
I have only been able here, without argument or detailed
exposition, to present a short summary of the more striking
phenomena in the religious systems of our spiritual ancestors. Many
of the problems I have stated still invite further research, which may
considerably modify our theories. I claim that the subject possesses
a masterful interest both in its own right and for the light it sheds on
ancient philosophy, ancient art, and ancient institutions. And it ought
in the future to attract more and more the devotion of some of our
post-graduate students. Much remains to be done even for the
Hellenic and Roman religions, still more for those of Egypt and
Assyria. Here, in our University of Oxford, under whose auspices the
Sacred Books of the East were translated, and where the equipment
for the study is at least equal to that of any other centre of learning,
this appeal ought not to be made in vain.
CHAPTER II.
Statement of the Problem and the Evidence.
The subject I have chosen for this course may appear over-
ambitious; and the attempt to pass critical judgment upon the facts
that arise in this wide comparative survey may be thought
premature. For not only is the area vast, but large tracts of it are still
unexplored, while certain regions have yielded materials that are
ample and promising, but of which the true interpretation has not yet
been found. We have, for instance, abundant evidence flowing in
with ever-increasing volume of the Sumerian-Babylonian religion, but
only a portion of the cuneiform texts has as yet been authoritatively
translated and made available for the service of Comparative
Religion. The Hittite monuments are witnesses of primary value
concerning Hittite religion: but the Hittite script may reveal much
more that is vital to our view of it, and without the help of that script
we are not sure of the exact interpretation of those religious
monuments; but though we have recently heard certain encouraging
expressions of hope, the master-word has not yet been found that
can open the door to this buried treasure of knowledge. And again,
the attempt to gauge accurately the relation and the indebtedness of
Greek religion to that of the near East cannot be wholly successful,
until we know more of the Minoan-Mycenaean religion; and our hope
hangs here partly on the discovery of more monuments, but mainly, I
am convinced, on the decipherment of the mysterious Minoan
writing, to which great achievement Sir Arthur Evans’ recent work on
the Scripta Minoa is a valuable contribution. Therefore the time is
certainly not yet ripe for a final and authoritative pronouncement on
the great questions that I am venturing upon in this course.
But even the early premature attempts to solve a problem may
contribute something to the ultimate satisfying solution. And often in
the middle of our investigations, when new evidence continues to
pour in, there comes a moment when it is desirable to look around
and take stock, so to speak, to consider whether we can draw some
general conclusions with safety, or in what direction the facts appear
at this stage to be pointing. In regard to the religions of anterior Asia
and South-Eastern Europe, and the question of their relationships,
this is now, I feel, a seasonable thing to do—all the more because
the Asiatic region has been mainly explored by specialists, who have
worked, as was profitable and right, each in his special province,
without having the time or perhaps the training to achieve a
comparative survey of the whole. We know also by long experience
the peculiar dangers to which specialists are prone; in their
enthusiastic devotion to their own domain, they are apt to believe
that it supplies them with the master-key whereby to unlock many
other secret places of human history. This hope, soon to prove an
illusion, was regnant when the interpretation of the Sacred Vedic
Books was first accomplished. And now certain scholars, who are
distinguished specialists in Assyriology, are putting forward a similar
claim for Babylon, and are championing the view that the Sumerian-
Assyrian religion and culture played a dominating part in the
evolution of the Mediterranean civilisation, and that therefore much
of the religious beliefs and practices of the early Greeks and other
European stocks must be traced back to Mesopotamia as their
fountain-head.31.1 This will be encouraging to that distinguished
writer on Greek religion, Dr. O. Gruppe, who almost a generation ago
proclaimed in his Griechische Mythologie the dogma of the
emanation of all religion from a single centre, and the dependence of
Greece upon the near East.
Now there ought to be no prejudice a priori against such a theory,
which stands on a different footing from what I may call the Vedic
fallacy: and it is childish to allow to the Aryan, or any other racial
bias, any malignant influence in these difficult discussions. Those
who have worked for years upon the marvellously rich records of
Mesopotamian culture, whether at first hand or, like myself, at
second hand, cannot fail to receive the deepest impression of its
imperial grandeur and its forceful vitality, and of its intensity of
thought and purpose in the sphere of religion. Naturally, they may
feel, such spiritual power must have radiated influence far and wide
over the adjacent lands; and no one could maintain that South-
Eastern Europe was too remote to have been touched, perhaps
penetrated, by it. For we know that, under certain conditions, the
race-barrier falls down before the march of a conquering and
dominating religion. And now, in the new light of a wider historical
survey, instead of saying, as once was said, “What is more its own
than a people’s gods?” we may rather ask, “What is less its own than
a people’s gods?” always, however, remembering that race-tradition,
inherited instinctive feeling and thought, is very strong in these
matters, and that a people will, often unconsciously, cling to its
ancestral modes of religious consciousness and expression, while it
will freely borrow alien forms, names, and ritual.
The inquiry indicated by the title of these lectures is naturally
twofold; it may be applied either to the earlier or the later periods of
the Hellenic and Hellenic-Roman history. The question concerning
the later period, though much critical research is needed for its clear
solution, is far simpler and more hopeful: for the evidence is
immeasurably fuller and more precise, and historical dates and
landmarks are there to help. The history of the invasion of the West
by Mithraism has been masterfully stated by Cumont; the general
influence of the Anatolian religions upon Graeco-Roman society is
presented and estimated by the same writer in his Religious
Orientales; by Toutain, in his Les Cultes Paiens dans l’Empire
Romain; by our own scholar, Samuel Dill, in Roman Society in the
last Century of the Western Empire; and more summarily by
Salomon Reinach in his Orpheus. Therefore I am not going to
pursue the inquiry at this end, although I may have to notice and use
some of the later evidence. I am going to raise the question
concerning the very origins of the Hellenic religious system, so as to
test the recently proclaimed dogma of certain Assyriologists, and to
determine, if possible, whether the Orient played any formative part
in the organic development of Greek religion. For this is just the
question which, I venture to maintain, has never yet been critically
explored. From what I have said at the beginning, it is obvious that I
cannot promise final and proved results. It will be gain enough if we
can dimly discern something behind the veil that shrouds the origins
of things, can reach to something that has the air of a reasoned
scientific hypothesis, and still more if we can indicate the paths along
which one day light may come.
We may then begin at once with stating more clearly what are the
necessary conditions for a successful solution of the problem. First,
we must accomplish a thorough exploration of the religions of the
Anatolian and Mesopotamian lands; secondly, we must explore the
Minoan-Mycenaean religion, and estimate the strength of its
influence on the later period; thirdly, we must be able to decide what
beliefs and practices the Hellenes brought with them from the North;
lastly, before we can hope for any precision in our results, we must
be able to answer with some degree of accuracy a burning
chronological question: What was the date of the arrival through the
Balkans from the North of those Aryan-speaking tribes that by
mingling with the Southerners formed the Hellenic people of history?
For only then shall we be able to test the whole question, by
considering the position of the Eastern powers at this momentous
epoch. The third of these inquiries, concerning the aboriginal
religious ideas of the earliest Aryan Hellenes, is perhaps the most
troublesome of all. I may venture upon it at a later occasion, but it is
far too difficult and extensive to combine it with the others in a short
treatise. Nor can I do more than touch lightly on the Minoan-
Mycenaean period; for I wish to devote the greater part of these
lectures to the comparative survey of Greece, Anatolia, and
Mesopotamia, as this task has never yet been critically performed.
Something like an attempt was made by Tiele in his Histoire des
anciennes Religions, but when that book was written much of the
most important evidence had not yet come in.
But before beginning the exploration of any large area, whether for
the purposes of Comparative Religion, archaeology, or anthropology,
we must possess or acquire certain data of ethnography and secular
history. We must, for instance, face the chronological question that I
mentioned just above, before we can estimate the formative
influences at work in the earliest phases of Hellenic development.
Recent archaeological evidence, which I cannot here discuss,
renders us valuable aid at this critical point of our inquiry. We can no
longer relegate the earliest Hellenic invasion of Greece to a very
remote period of Mediterranean history. The arguments from the
Minoan culture, combined with the still more striking evidence, of
which the value is not yet fully appreciated, obtained by the recent
excavations of the British School on the soil of Thessaly,34.1 point to
the conviction that this, the epoch-making event of the world’s
secular and spiritual life, occurred not much earlier than 1500 B.C.
On this hypothesis, our quest becomes less vague. We can consider
what influences were likely to be radiating from the East upon the
opposite shores of the Aegean during those few centuries, in which
the Hellenic tribes were passing from barbarism to culture, and the
religious beliefs and ritual were developing into that comparatively
advanced and complex form of polytheism which is presented about
1000 B.C. in the Homeric poems. By this date we may assert that
the Hellenic spirit had evolved certain definite traits and had acquired
a certain autonomous power. While continuing always to be quickly
responsive to alien influence, it would not henceforth admit the alien
product with the submissive and infantine docility of barbarians In
fact, when we compare the Homeric religion with that of the fifth
century, we feel that in this particular sphere of the social and
spiritual life the Hellene in many essentials had already come to his
own in the Homeric period. Therefore, in trying to track the earliest
streams of influences that moulded his religious consciousness,
what was operative before the tenth century is of more primary
importance than what was at work upon him afterwards.
Now the most recent researches into Mesopotamian history
establish with certainty the conclusion that there was no direct
political contact possible between the powers in the valley of the
Euphrates and the western shores of the Aegean, in the second
millennium B.C. It is true that the first Tiglath-Pileser, near the end of
the twelfth century, extended the Assyrian arms to the shores of
South-Eastern Asia, to Cilicia and Phoenicia35.1; but there does not
seem to have been any permanent Assyrian or Babylonian
settlement on this littoral. The city of Sinope in the north, which, as
the legend attests and the name that must be derived from the
Assyrian god Sin indicates, was originally an Assyrian colony, was
probably of later foundation, and geographically too remote to count
for the present inquiry. In fact, between the nascent Hellas and the
great world of Mesopotamia, there were powerful and possibly
independent strata of cultures interposing. We have to reckon first
with the great Hittite Kingdom, which included Cappadocia and

Ebffiledoc - 993download PDF Neurocognitive Mechanisms Explaining Biological Cognition Gualtiero Piccinini Ebook Full Chapter

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ebffiledoc - 993download PDF Neurocognitive Mechanisms Explaining Biological Cognition Gualtiero Piccinini Ebook Full Chapter

Uploaded by

Copyright:

Available Formats

Neurocognitive Mechanisms:

Explaining Biological Cognition

Molecular Mechanisms of Autonomy in Biological Systems

Physical computation : a mechanistic account Piccinini

Yeast Cell Surface Engineering Biological Mechanisms

Biological and Bio inspired Nanomaterials Properties

Explaining Health Across the Sciences Jonathan Sholl

Explaining Imagination 1st Edition Peter Langland

Explaining Health Across the Sciences 1st Edition

Coulson and Richardson’s Chemical Engineering, Fourth

My undergraduate studies at the University of Turin, during the early 1990s,

digital, might be picking up on something that digital computation, analog

Chapter 10 is a substantially revised descendant of Piccinini 2007c.

1.1 Possible relations between higher-level properties and lower-level properties 26

This book defends a neurocomputational theory of cognition grounded in a

By cognitive capacities, I mean capacities such as perception, memory, reason-

out to be neural computation over neural representations. Neural computation is

1.1 Parts, Wholes, and their Properties

This book argues that cognition is explained by multilevel neurocognitive

¹ In mereology—the formal study of the part–whole relation—it is convenient to deﬁne “part” so

1.2 Causal Exclusion

Again, different people disagree about the relationship between properties of

do people. Many ﬁnd it personally insulting to be reduced to something else—it

Many advocates of the special sciences—especially psychology—defend some

their properties change. Objects that organize together in speciﬁc physical

You might also like