You are on page 1of 441

ANALYTIC PHILOSOPHY OF RELIGION

HANDBOOK OF CONTEMPORARY PHILOSOPHY OF RELIGION


Volume 3

Series Editor

EUGENE THOMAS LONG

A list of titles in this series can be found at the end of this volume.

JAMES F. HARRIS
The College of William and Mary in Virginia, U.S.A.

Analytic Philosophy of Religion

SPRINGER-SCIENCE+BUSINESS MEDIA, B.V.

Library of Congress Cataloging-in-Publication Data is available.

ISBN 978-90-481-5983-3
ISBN 978-94-017-0719-0 (eBook)
DOI 10.1007/978-94-017-0719-0

Printed an acid1ree paper

AlI Rights Reserved


2002 Springer Science+Business Media Dordrecht
Originally published by Kluwer Academic Publishers in 2002
Softcover reprint of the hardcover 1st edition 2002
No part of this publication may be reproduced or utilized in any form or by any means,
electronic, mechanical, including photocopying, recording or by any information storage
and retrieval system, without written permission from the copyright owner.

Table of Contents

Handbook of Contemporary Philosophy of Religion: Series Outline

vi

Preface

vii

1 Introduction: The Rise of Analytic Philosophy of Religion

2 The Problem of Religious Language

28

3 The Nature of God and Arguments for the Existence of God

77

4 Religious Experience and Religious Epistemology

141

Religion and Science

194

Contemporary Challenges to Theism: Evil and Suffering

234

Contemporary Challenges to Theism: Humanism, Naturalism,


and Atheism

278

8 Religion and Ethics

330

9 The Problem of Religious Pluralism

374

10 Summary and Conclusion

413
427

Index

Handbook of Contemporary Philosophy of Religion


Series Outline

At mid-century western philosophy of religion appeared to be gasping for


breath. The logical positivists and the positivists of revelation cooperated in
challenging the legitimacy of natural theology and the philosophy of religion. Soon
after mid-century, however, new approaches to philosophy with roots in the early
twentieth century began to flourish leading to renewed interest in the philosophy of
religion. Neo-Thomists and process philosophers breathed new life into
metaphysics, and analytic and existential philosophers opened up new avenues for
philosophical reflection on the meaning and truth of God-talk. These discussions
often reflected an empirical and historical mindedness that was quietly calling into
question the classical foundations of western philosophy and philosophy of
religion. The last quarter of the century has seen the blossoming of this empirical
and historical mindedness leading to an enormous change of climate in
philosophical reflection on religion. This is an era characterized by pluralism in
human experience and diversity in philosophical method. It is an era in which
many of the traditional foundations and methods of philosophical reflection on
religion are being called into question. The Handbook of Contemporary
Philosophy of Religion explores developments in contemporary philosophy of
religion in its many forms. Volume one provides an historical map of twentieth
century western philosophy of religion and serves as an introduction to other
volumes in the Handbook. Among the early volumes will come studies of
contemporary Thomistic, analytic, comparative and process philosophy of religion.
Additional volumes are being planned covering other philosophical approaches and
selected issues. Each volume will introduce the subject in its recent and current
state and provide an analysis in light of contemporary debates and the author's own
position.
Eugene Thomas Long
Editor

VI

Preface

When Gene Long, editor of Kluwer's Handbook of Contemporary Philosophy of


Religion Series, first invited me to write the volume on Analytic Philosophy of
Religion, I accepted with great enthusiasm. My only explanation for that
enthusiasm now is that I was younger and more naive at the time. Soon after
starting work on the volume, my enthusiasm was dampened by the daunting
magnitude of the task. I began as a sprinter and quickly settled into the pace of a
long-distance runner. Although I considered myself well read in the subject, I soon
discovered that I had a great deal of research to do to be confident that I had
considered all of the major contributions to the various discussions, issues, and
problems found within analytic philosophy of religion. As I read more and more
books and articles, I realized that I had rushed into a territory already well trodden
by the angels. I am greatly impressed by the sophistication and subtlety of
philosophical argument that characterize the different debates in contemporary
analytic philosophy of religion.
This volume covers a vast amount of material. I have endeavored to provide the
fairest possible reading of different authors, and, in cases where I include my own
critical evaluations and develop my own positions, I have endeavored to provide
the strongest possible interpretations of the positions I criticize. I am greatly
indebted to several colleagues who have read different portions of the manuscript
and offered various suggestions and comments. Their influence has made this a
much better book than it would have been otherwise. My gratitude goes to Paul
Davies, Paul Draper, Laura Ekstrom, Antony Flew, Michael Gettings, Steven
Hales, George Harris, William Hasker, Earl McLane, D. Z. Phillips, Philip Quinn,
William Rowe, Charles Taliaferro, Hans von Bauer, and John Whitaker. I am
especially grateful to the advice and support of Gene Long. Of course, I remain
responsible for the final product and whatever mistakes or omissions it contains.
Finally, I must thank my secretary, Debra Wilson, without whose cheerful and able
assistance completing this book while balancing my responsibilities as a
department chair would have been impossible. My greatest fear is that I have
omitted or mis-characterized the work of some important contributor to one of the
many debates taking place within contemporary analytic philosophy of religion.
My greatest hope is that this volume will prove to be valuable to those interested in

vii

Vlll

ANALYTIC PHILOSOPHY OF RELIGION

analytic philosophy of religion in helping them to understand the sophistication


and subtlety of those debates and in helping to advance philosophical inquiry.
James Harris
College of William and Mary in Virginia
October, 2001

I. Introduction: The Rise of Analytic Philosophy of


Religion

There is little need to explain or defend the unique status which human language
has occupied and continues to occupy in most fields of human endeavor. Certainly
no other human device or invention has influenced the development of human
civilization more than human language. Language is doubly important for any
intellectual or philosophical pursuits - including, and perhaps especially,
philosophy and the philosophy of religion. Every major investigation into any of
the many areas of the philosophy of religion and everyone of the many disputes
among different schools of thought and different individual thinkers have depended
upon the sophistication and complexity as well as the subtleties and fine details
that human language allows. The same is true, of course, to a greater or lesser
extent, for all philosophical inquiry; however, the use of language to engage in any
intellectual pursuit poses unique and serious questions - both methodological and
substantive.
CONCEPTUAL ANALYSIS AND LINGUISTIC ANALYSIS
Perhaps the most fundamental and significant role that language has in any area
of inquiry is the role that it plays in the process of conceptual analysis - the
analysis and clarification of the fundamental concepts, categories, and distinctions
used in any inquiry. Bracketing and avoiding for the moment any mentalistic or
ontological questions and concerns about the nature or existence of concepts or
ideas, it is unarguable that whatever the nature of such concepts and categories, our
only philosophical access to them is through language - through the names, terms,
words, or sentences in terms of which such concepts are formulated, expressed,
and communicated. Thus, any attempts at conceptual clarity and understanding, as
well as any attempts at understanding and resolving conceptual confusions and
disagreements, depend upon language and linguistic analysis.
Much of twentieth-century analytic philosophy has focused upon linguistic
analysis for the purpose of conceptual analysis. An investigation into the nature
and status of language and the theories of meaning and reference for the words,
phrases, and sentences of which language is composed is directed at a clarification
and better understanding of the "underlying" concepts. The same has also been true
of much of twentieth-century analytic philosophy of religion. Twentieth-century
1

ANALYTIC PHILOSOPHY OF RELIGION

Anglo-American analytic philosophy is characterized by a central and essential


dependence upon language and linguistic analysis; however, in comparison with
other periods in the history of philosophy, analytic philosophy's concern with
language is best explained as the result of an increase in emphasis upon and
recognition of the importance of language.
LINGUISTIC ANALYSIS AND THE HISTORY OF PHILOSOPHY
Certainly other periods in the history of philosophy and major figures in earlier
periods in the history of philosophy have engaged in linguistic and conceptual
analysis. For example, when Aristotle undertakes what is perhaps the most
important and crucial part of his philosophical inquiry into the nature of substance
(ousia) at the beginning of Book Lambda of his Metaphysics, he engages in what
most contemporary analytic philosophers would recognize as straightforward
linguistic analysis, giving examples of how we talk and how we describe different
events and activities to clarify and drive home the distinction between substance
and predicates. Again, in On Interpretation, Aristotle develops a theory of meaning
for propositions in which he adopts the distinction between a sentence and a
proposition - a distinction that most people associate only with twentieth-century
positivism. A sentence is meaningful not by "natural means," Aristotle says, but by
"convention." And, he continues, "every sentence is not a proposition; only such
are propositions as have in them either truth or falsity. Thus a prayer is a sentence,
but is neither true nor false." (17a)
Similarly, most of the medieval scholastics were continually engaged in
conceptual analysis. As most scholars are well aware, the scholastics were
particularly concerned about language and its importance for any attempt to
describe God or to apply any of the traditional theistic predicates to God. For
example, in his Summa Theologica (l.13.5), Thomas Aquinas goes to great lengths
to develop a theory of analogy to explain how we can meaningfully attribute the
same predicates, such as "good" or "wise," to both men and God. Aquinas pays
careful attention to our use of language, and he uses the distinctions found in
language to clarify the ways in which we use language to talk about humans in
contrast with the ways in which we use language to talk about God. Again, most
contemporary analytic philosophers (while perhaps not agreeing with Aquinas'S
resulting doctrine of divine analogy) would recognize this line of inquiry as simple,
straightforward linguistic and conceptual analysis. Thus, it is easy to demonstrate
how various figures and periods in the history of philosophy and theology have
been concerned with and engaged in conceptual and linguistic analysis.
THE RISE OF LINGUISTIC ANALYSIS
In the twentieth century, the accompanying specific problems of meaning,
reference, proper names, descriptions, meaningfulness, and communication have
all become explicit parts of that general concern. There is no doubt, however, that
for much of the English-speaking world of philosophy, the first half of the
twentieth century saw the concern with problems associated with the
meaningfulness of language and with the clarification of concepts reach an
unprecedented height. The increased emphasis on linguistic and conceptual
analysis is one of those rare instances in the history of philosophy that can be

INTRODUCTION: THE RISE OF ANALYTIC PHILOSOPHY OF RELIGION

clearly and undisputedly traced both to its initial sources and to the continuing
influences that kept linguistic and conceptual analysis at the forefront of the
philosophy of religion for over a half a century. The rise of analytic philosophy
resulted from what Gilbert Ryle has called "The Revolution in Philosophy,,1 and
what Richard Rorty has called "the Linguistic Turn.,,2 The period is what has
become known as "the Age of Analysis.,,3
Although there are many similarities in methodology, interests, emphases, and
results among various philosophers who are commonly regarded as belonging
within the analytic tradition, analytic philosophy is not and has never been
monolithic. There are also widespread and significant differences among analytic
philosophers concerning their methodology, interests, emphases, and results. All in
all, analytic philosophy has been a very heterogenous "movement." Although there
are some common themes, there is also as much variety among analytic
philosophers in their fundamental philosophical commitments and positions as
there has been among idealists or realists or theologians; consequently, it is
misleading to talk about "analytic philosophy" as a single movement in philosophy
without recognizing the significant differences among analytic philosophers. This
variety has very important repercussions for the philosophy of religion - some of
which are more serious and threatening than others. As we shall see, some forms of
analytic philosophy have proven to be very sympathetic to the philosophy of
religion and have actually provided a philosophical mechanism for responding to
other more radical and hostile forms of analytic philosophy. The only way to
understand and appreciate the different ways in which the philosophy of religion
has been impacted by analytic philosophy is to examine some of the most
significant differences among some of the leading figures of what is commonly
called analytic philosophy to see how these differences impacted the philosophy of
religion.
THE LINGUISTIC TREES AND THE PHILOSOPHICAL FOREST
It is a common misunderstanding among the "uninitiated" - among those who

view analytic philosophy unsympathetically from "the outside" - that it is only


concerned with language. Some critics complain that analytic philosophy has
abandoned any semblance or pretext of "doing philosophy" in any traditional sense
by addressing the really important, traditional problems of philosophy that
determine how people live their lives. As we shall see, only in rare instances of
some few individual analytic philosophers is there some modicum of truth in this
accusation. The work of figures within the analytic tradition is representative of the
classic adage about the difference in perspective between seeing the trees and the
forest. Often, the very detailed and sometimes technical work of analytic
philosophers focuses upon specific and individual trees with little or no attention
given to the more general, sweeping, and admittedly more satisfying view of the
forest. However, in nearly every occasion, and certainly in the occasion of the
major figures in the analytic tradition, there is some interest in some philosophical
Gilbert Ryle, The Revolution in Philosophy (London: Macmillan, 1957).
Richard Rorty, The Linguistic Turn (Chicago: University of Chicago Press, 1967).
3 Morton White, The Age of Analysis (New Yark: Mentor Books, 1955).
1

ANALYTIC PHILOSOPHY OF REUGION

problem of larger scope, some aspect of which is approached by, or understood by,
or attacked by, or defended by the analysis of language and concepts. "Doing"
analytic philosophy takes both its name and its method from the sciences in which
analysis is understood as the process of understanding larger problems or
compounds in the world by dividing them into their smaller components. The idea
is to understand the whole by understanding the more manageable parts, and even
in those cases where the whole might be greater than the sum of its parts, such a
claim is recognized and understood only by a careful analysis of the parts. In
general then, analytic philosophy has little sympathy for the approach to
philosophy that is based upon a priori reasoning and that results in general
metaphysical theories that ignore the "dirty work" of careful attention to particular
details.
Generally, with only some possible rare exceptions, the concern among analytic
philosophers with linguistic and conceptual analysis has seldom been regarded as
"an end in itself'; rather, analytic philosophers have used the analysis of language
and concepts to try and "get at" philosophical issues and problems, which, in most
cases, are the same traditional philosophical problems with which philosophers
have been concerned since the time of Thales. It is true that, on the whole, certain
patterns emerge from the works of many analytic philosophers that are generally
unsympathetic to certain approaches to philosophy and to certain kinds of
philosophical theories. Such a pattern is particularly easy to identify in the case of
philosophy of religion. As a result of the influence of analytic philosophy in the
English-speaking philosophical world, philosophy of religion definitely changed its
character from what it had been in the nineteenth century. To understand this
change and how and why it took place, we need to understand a little about the
history of philosophy and how and why analytic philosophy itself developed. 4
ABSOLUTE IDEALISM AND THE RISE OF ANALYTIC PHILOSOPHY
John Stuart Mill, a giant of a philosophical figure in England in the middle of
the nineteenth century, had continued in the tradition of "British empiricism."
However, following his death in 1873, English philosophy was quickly dominated
by Neo-Hegelianism, one of the leading proponents of which was F. H. Bradley at
Oxford. According to Bradley, since the only "real" thing that exists is the
Absolute and everything else is an illusion, what appear to be individual, discrete
objects in the world - for example, my truck and my fishing rods - are really
"internally" and "organically" related into a cosmic whole.

4 There are several detailed treatments of the rise of analytic philosophy and philosophical analysis,
inclnding A. J. Ayer, Philosophy in the Twentieth Century (New York: Vintage Books, 1984); John
Passmore, A Hundred Years of Philosophy (London: Duckworth, 1957); Gilbert Ryle, ibid.; G. J.
Warnock, English Philosophy since 1900 (Oxford: Oxford University Press, 1958); Morris Weitz,
Twentieth-Century Philosophy: The Analytic Tradition (New York: Macmillan, 1966); and Morton
White, ibid. See also The Revolution in Philosophy, edited by D. F. Pears (New York: Macmillan,
1955), and J. O. Urmson, 'The History of Analysis," La Philosophie analytique (paris: Editions de
Minuit, 1962). Translated and reprinted in Rorty, ibid., pp. 294-301.

INTRODUCTION: THE RISE OF ANALYTIC PHILOSOPHY OF RELIGION

THE DOCTRINE OF INTERNAL RELATIONS


Bradley's speculative monistic and holistic view of things is built upon a theory
of internal relations according to which every genuine characteristic or predicate of
an object is "internal," that is, essential, to the nature of that object. The theory of
internal relations serves as an illustrative example and precursor of the kind of
philosophical inquiry that was to develop with the rise of linguistic analysis. Of
what does the identity of an object consist? What makes one object different from
or similar to another? From all appearances, a fishing rod is a different object from
a pickup truck, but, Bradley claims, that is just so much the worse for appearance.
In his famed major work, Appearance and Reality, Bradley insists that such
appearances are deceiving and that "underneath" and in "reality" everything is one.
The One is spiritual. The One can be known only through reason and not through
sense experience. The problem with our common understanding of the world,
according to Bradley and the other Neo-Hegelians, is that we do not understand
how everything is internally related, and what appear to us to be differences
between different objects are simply the result of our ignorance and our failure to
understand the Absolute. All predicates, characteristics, and relations are defining
of an object according to the doctrine of internal relations; so, for example, not
only what we normally regard as the individual, physical characteristics of an
object but also its location in space and time are all essential parts of what makes
an object what it is. A particularly illuminating metaphor that can be used to
explain the doctrine of internal relations is in the lines from Alfred Lord
Tennyson's poem "Flower in a Crannied Wall," in which the speaker says to the
flower,
Flower in the crannied wall,
I pluck you out of the crannies,
I hold you here, root and all, in my hand,
Little flower - but if I could understand
What you are, root and all, and all in all,
I should know what God and man is.
The suggestion in the poem is that the flower is essentially related to everything
else in the universe - spatially and temporally and causally - to the most distant
stars and to the most ancient times. If one is to understand completely the flower,
then one must understand what is responsible for the flower being the flower, that
is, the flower's connectedness and relatedness to everything else. To understand
one thing is to understand the entire universe because of the organic monism which
makes everything One. Everything is what it is because of its relation to everything
else; so, everything is essentially the same thing - the organic Whole, the Absolute.
So far as the philosophy of religion is concerned, the rise of Neo-Hegelian,
absolute idealism in England and the United States at the end of the nineteenth
century was important for at least two reasons. First, the content or substance of the
theory of absolute idealism provided a philosophical counterbalance to the rise of
science and the accompanying materialism and mechanism that was beginning to
spread across Europe and the United States in the mid-nineteenth century.
Secondly, the completely a priori method and grand, general, sweeping

ANALYTIC PHILOSOPHY OF RELIGION

metaphysics provided a philosophical counterpoint to the rigorously empirical


method and limited theories of science. Idealism protected the notion that there is a
spiritual aspect to the universe and a "higher" meaning for all existence - which
provided succor for theologians, beleaguered as they were at the end of the
nineteenth century by mechanistic and atheistic accounts of the universe and
human existence.
G. E. MOORE AND EARLY PHILOSOPHICAL ANALYSIS
The two major figures of early analytic philosophy, who reacted against
absolute idealism toward the end of the nineteenth century and who were most
responsible for starting the philosophical shift from absolute idealism to analytic
philosophy, were G. E. Moore and Bertrand Russell. Ironically, both Russell and
Moore, while undergraduates at Cambridge, had originally been under the
influence of McTaggart; however, both Moore and Russell revolted against and
attacked every aspect of absolute idealism. Whereas absolute idealism insisted
upon a spiritual monism based upon internal relations that is known only through
Pure Reason, Moore and Russell defended a pluralistic realism based upon external
relations that is known through common sense. Russell gives Moore the credit for
first arousing him from what may be called his "idealistic slumber."
He [Moore] took the lead in rebellion [against idealism], and I followed with a
sense of emancipation. Bradley argued that everything common sense believes
in is mere appearance; we reverted to the opposite extreme, and thought that
everything is real that common sense, uninfluenced by philosophy or theology,
supposes real. With a sense of escaping from prison, we allowed ourselves to
think that grass is green, that the sun and stars would exist if no one was aware
of them, and also that there is a pluralistic timeless world of Platonic ideas. The
world, which had been thin and logical, suddenly became rich and varied and
solid. s
Moore's attack upon idealism and his defense of realism and common sense was
a prolonged attack spread over several decades. His major contributions include
"The Refutation of Idealism," "A Defence of Common Sense," "Proof of an
External World," and "External and Internal Relations.,,6 Simply the titles of these
articles put them in stark juxtaposition with the basic claims of absolute idealism.
Through this series of articles and in his Principia Ethica, Moore developed the
method of philosophical investigation that was to become known as philosophical
analysis. Moore proceeds by asking what the correct analysis is of the meaning of
5 Bertrand Russell, "My Mental Development," in The Philosophy of Bertrand Russell, edited by
Paul A. Schilpp (New York: Harper and Row, 1944), p. 12.
6 G. E. Moore, "A Refutation of Idealism," Mind, Vo1.12, 1903; "External and Internal Relations,"
Proceedings of the Aristotelian Society, 1919-20; "A Defence of Common Sense," in Contemporary
British Philosophy, Second Series, edited by J. H. Muirhead (London: George Allen and Unwin, 1925);
"Proof of an External World," Proceedings of the British Academy, Vol. 25, 1939. These articles are
reprinted in many locations. "A Refutation of Idealism" and "External and Internal Relations" are
included in Moore's Philosophical Studies (Totowa, N.J.: Little, Adams, and Co., 1922). "A Refutation
of Idealism," "A Defence of Common Sense," and "Proof of the External World" are in Contemporary
and Analytic Linguistic Philosophies, edited by E. D. Klemke (New York: Prometheus Books, 1983).

INTRODUCTION: THE RISE OF ANALYTIC PHILOSOPHY OF RELIGION

certain propositions? In his Principia Ethica, 8 for example, he develops a moral


theory by analyzing the moral predicates "good" and "bad" to carefully identify the
conditions for the situations and propositions in which we would normally use one
rather than the other. The actual substance of Moore's ethical theory in Principia
Ethica - his intuitionism, his treatment of moral good as a simple, unanalyzable
predicate, and his famed "naturalistic fallacy" - is not important here. However,
his philosophical method of careful analysis of predicates and propositions was
soon to sweep the English-speaking philosophical world.
In his earliest and most influential attack on idealism, "The Refutation of
Idealism, " he undermines the idealistic claim that reality is spiritual by analyzing
the meaning of the basic claim, "Esse is percipi." Moore argues that the idealistic
argument that reality is spiritual cannot be proven since "Esse is percipi" is, in any
sense not trivially tautological, either simply false or self-contradictory.9 There are
many detailed and subtle points to Moore's analysis of the meaning of "Esse is
percipi," but one of the most telling is his insistence upon a difference between an
experience (or act of consciousness) and the object of that act of consciousness.
The two are not, as the idealists had claimed, an inseparable "organic unity." For
example, since blue and the sensation of blue are different things, then anyone who
claims that to be blue is identical to being an act of consciousness of blue makes a
self-contradictory mistake.1O Although the debate between those defending
idealism and those defending the new philosophical analysis was to go on for many
years, Moore's attacks against idealism in "Refutation of Idealism" and "External
and Internal Relations" represent a watershed after which the tide began to turn in
the world of English-speaking philosophers in favor of what we now call analytic
philosophy.
BERTRAND RUSSELL AND EARLY PHILOSOPIDCAL ANALYSIS
Although Moore might have been the leading figure in the early development of
philosophical analysis at the turn of the twentieth century, Bertrand Russell
undoubtedly has proven to be the most influential figure of the earliest period of
analytic philosophy. Russell joined Moore in his defense of realism and common
sense, but Russell had come to philosophy through mathematics. Many of his early
writings are mathematical and very technical in nature, and access to them often
requires learning certain logical and mathematical notational schemes. I will here
simply sketch the broadest implications of his work for the philosophy of religion.
Philosophers drawn to linguistic analysis as a method of philosophical inquiry
still have disagreed strongly among themselves concerning the particular kind of
language that is important to analyze or investigate. In other words, linguistic
philosophers have had significant disagreements about exactly what "language" is,
and resolving this matter has proven to be extremely difficult. If linguistic analysis
is supposed to be a tool for philosophical inquiry, then, we might say, much time
and effort have been spent simply on designing and describing the tool itself. The
work of the early Russell, as well as his later work in logical atomism, places
Particularly illustrative is Part IV of "A Defence of Common Sense," in Klemke, ibid., pp. 178ff.
G. E. Moore, Principia Ethica (Cambridge: Cambridge University Press, 1903).
9 See Moore, 'The Refutation of Idealism," in Moore, ibid., pp. 5ff., and in Klemke, ibid., pp. 123ff.
10 Moore, ibid., p. 18, Klemke, ibid., p. 130.
7
8

ANALYTIC PHILOSOPHY OF RELIGION

Russell in the group of philosophers for whom the proper kind of language for use
in philosophical inquiry is the logical language of an artificially constructed,
formal, deductive system. Russell's most monumental work, and arguably one of
the most influential works of twentieth-century philosophy, is the massive, threevolume Principia Mathematica, which he co-authored with Alfred North
Whitehead, his former math tutor at Trinity College, Cambridge. ll
Taking variables to range over classes and using the logical connectives of class
inclusion and class membership, Russell and Whitehead constructed a formal,
deductive system within which they formally reduced mathematics to logic. By
doing so, Russell saw himself as extending the careful rigor and precision of
mathematics into areas of inquiry that had been characterized by philosophical
vagueness and speculative thought. He believed that Principia Mathematica
represented a formal refutation of Kant's treatment of mathematics and his use of
the synthetic a priori. 12 By defining cardinal and ordinal numbers and arithmetic
operations in terms of logical axioms and connectives, Russell and Whitehead
"demystified" mathematics and moved it from the realm of speculative
metaphysics to the realm of straightforward logic.
MATHEMATICS AND GOD
Many people have failed to recognize these broader and more general
implications of Russell's work in mathematics and particularly Russell and
Whitehead's work in Principia Mathematica because they become either too
absorbed in or too deterred by the technical and mathematical details of the work.
Russell's theory of descriptions and Russell and Whitehead's reduction of
mathematics to logic are classic examples of the importance of keeping both the
detailed, individual perspective of the linguistic trees and the more general
perspective of the larger philosophical forest. In a very broad sense, many of the
epistemological differences between rationalists and empiricists in the history of
philosophy have arisen from the different ways in which they have treated
mathematics and mathematical knowledge. From the point of view of the
rationalists, empiricism cannot account for the necessity of mathematics by any
appeal to empirical, a posteriori reasoning. From the point of view of empiricism,
rationalism can only account for the necessity of mathematics by suspect intuition
or a priori reasoning. Russell thought that the Hegelian version of mathematics incorporated into his dialectic and adopted by Bradley and McTaggart - was just a
bunch of "muddled-headed nonsense.,,13 Russell saw in his own work the
opportunity to rescue mathematics from the "metaphysical muddles,,14 infused into
it by the rationalists, particularly by Hegel and his followers, and to provide an
accounting for mathematics completely upon empiricist grounds. Russell says,

II Alfred North Whitehead and Bertrand Russell, Principia Mathematica (Cambridge: Cambridge
University Press, 1910-13). Second edition, 1923-27.
12 Bertrand Russell, "My Mental Development," in Schilpp, ibid.
13 Ibid., p. 11.
14 Bertrand Russell, A History of Western Philosophy (New York: Simon and Schuster, 1945), p.
829.

INTRODUCTION: THE RISE OF ANALYTIC PHILOSOPHY OF RELIGION

One result of the work we have been considering is to dethrone mathematics


from the lofty place which it has occupied since Pythagoras and Plato, and to
destroy the presumption against empiricism which has been derived from it.
Mathematical knowledge, it is true, is not obtained by induction from
experience .... In this sense, mathematical knowledge is still not empirical. But
it is also not a priori knowledge about the world. It is, in fact, merely verbal
knowledge .... Thus mathematical knowledge ceases to be mysterious. 15
Even given the difficulties which were to be raised about Principia Matematica,
including Russell's own paradox and "Godel's Proof' of the incompleteness or
inconsistency of any logical system "rich" enough for the reduction of
mathematics, Russell and Whitehead still provided a major boost for empiricism by
developing a new avenue for accounting for the necessity of mathematics.
Indirectly, or perhaps even directly, this boost for the new empiricism was
important because it undermined rationalism and speculative metaphysics that
included, for many philosophers at the time, religion and theology. The close of the
nineteenth century was no different from many other periods in the history of
philosophy in that mathematics and God were thought to stand or fall on the same
a priori grounds. Principia Mathematica promised a new avenue for approaching
such issues.
RUSSELL'S THEORY OF DESCRIPTIONS
In a similar vein, Russell's famed theory of descriptions continued the push
toward a rigorous empiricism. Although at frrst glance Russell's theory of
descriptions is only a theory about language and how and when words refer, there
are other larger and very important philosophical issues at stake. At the turn of the
century, Meinong and his followers (which included Russell at one time)
postulated an ontologically rich world of different kinds of entities that were
thought to "subsist" to deal with the problem of determining how it is that we are
able to talk about things that do not exist. Even to say that the Absolute does not
exist, or that God does not exist, or that the golden mountain does not exist seems
to require that the Absolute, God, and the golden mountain "exist" in some fashion
in order for us to be able to even talk about those things. After all, if we say, "The
Absolute does not exist," on the one hand and, "The golden mountain does not
exist," on the other, then we are saying that two different "things" do not exist, are
we not? Russell's theory of descriptions is based upon a process of "expanding" or
"unpacking" proper names or descriptions into complex descriptions that exposes
and makes explicit the hidden assumptions concerning what exists in the universe.
If the assumption of existence is not met, then, Russell says, the original
proposition is false. To use Russell's famous example, if one says, "The present
king of France is bald," when there is not a present king of France, then we
apparently must face the problem of having to admit that there is something that
we are talking about when we say that it is bald even if there is no present king of
France. Russell's way of handling this problem is to analyze the original
proposition as meaning, "There exists an x, such that x is currently the King of
15

Ibid., pp. 831-32.

10

ANALYTIC PHILOSOPHY OF REUGION

France, and x is bald, and given any y if Y is currently the King of France and is
bald, then x is identical to y." Paraphrased, this means that there exists one and
only one thing such that that thing is the present king of France and is bald.
Russell's theory of descriptions provides a way of analyzing propositions and
accounting for the reference or the failure of reference of their subject terms
without the need of postulating or inferring the existence of objects that exist in
some fashion that is different from the way in which ordinary physical objects
exist. Russell's oft-repeated aphorism, "Replace inferred entities with constructed
entities," represented the re-introduction of the Law of Parsimony or Ockham's
Razor into contemporary philosophy. There is no need to resort to ontological
extremes, Russell would say, when we can account for such propositions simply by
analyzing them into their implicit meanings.
Russell's theory of descriptions represents another major step in the rise of the
"new" empiricism and philosophical analysis. His theory of descriptions has been
used to analyze general metaphysical statements, including theological statements,
with the result that sweeping and "muddle-headed" claims of existence have given
way to careful, detailed analyses of meaning that are ontologically stingy. The
resurgence of the importance of Ockham's Razor and its incorporation into
theories of ontology created a climate that was decidedly unfriendly toward the
kind of metaphysics that supported theism. Neither Russell nor Moore was ever
opposed in principle to all kinds of metaphysics. In fact, both, at different times in
their careers, indicated that a general description of the entire universe in
metaphysical terms was a legitimate philosophical enterprise. However, they both
recognized a distinction between the completely speculative and dogmatic
metaphysics of absolute idealism, which went against both human experience and
common sense, and the sort of metaphysics that explains human experience and
that is supported by common sense and experience. Although Moore never
attempted to develop such a metaphysical theory, Russell did with his theory of
logical atomism that he regarded as a logically rigorous metaphysics based upon
the necessity of mathematics and logic.
AMERICAN PRAGMATISM
Although Charles Sanders Peirce should be credited with originating the
fundamental tenets of American pragmatism, many of his writings were published
only posthumously, and the ones that were published during his lifetime were not
known widely. Certainly William James is responsible for popularizing
pragmatism and creating of pragmatism what is arguably the only uniquely
American philosophical "movement." Largely as a result of James's influence,
pragmatism supplanted idealism as the dominant form of philosophy in America
following the beginning of the twentieth century. The pragmatism that James
developed from Peirce represents a return to empiricism and a particular form of
"practical empiricism." I will focus here upon James's use of philosophical
analysis to attack idealism since his role in the United States in attacking idealism

INTRODUCTION: THE RISE OF ANALYTIC PHILOSOPHY OF RELIGION

11

and supporting empiricism and realism was analogous to the roles of Moore and
Russell in England. 16
WILLIAM JAMES AND THE ATTACK UPON IDEALISM
James became a major figure in philosophy when his Lowell lectures from 1906
were published under the title of Pragmatism a year later. Perhaps because of both
the prominence of Pragmatism and the controversy that it generated, James was
invited to deliver the Hibbert Lectures at Oxford University in 1908. James had
already published several articles attacking idealism, but his most concentrated and
sustained attack came in these Hibbert Lectures when he visited England. 17 The
original title of the lectures was "The Present Situation in Philosophy"; however,
James's title for the published version of the lectures, A Pluralistic Universe,
reveals how deeply anti-idealistic the lectures were. James is clear from the outset
that he intends to defend a revised form of empiricism that "explains wholes by
parts" instead of "parts by wholes,,18 - a method that clearly parallels the general
approach that characterizes analytic philosophy. Informally, in a letter to his
brother Henry, James revealed more of his "flashy" style by declaring that he
intended to take "the scalp of the Absolute" and put the "wretched clerical
defenders" of the Absolute on the defensive in his lectures. 19
In James's arguments against idealism in A Pluralistic Universe, we get both an
understanding of the method of philosophical analysis and an understanding of the
important stake that traditional theism had in defending idealism. What is
important in philosophy, James claims, is not simply the truth of the particular
claim or theory that is advanced but the reasons and the arguments that are
advanced in support of the claim or theory and upon which the claim or theory is
based. Rigorous philosophical argumentation, not supposition or speculation or
revelation or guesswork, is what should distinguish the claims of the philosopher. 20
In his attack upon idealism, he relies upon linguistic analysis in a manner that is
as straightforward as anything found in the early Moore or Russell. For example,
James responds to a supposed reductio ad absurdum against pluralistic empiricism
with a detailed analysis of the language used in the argument. The objection is that
if there are many different finite objects, as pluralistic empiricism claims, that exist
separate and independent of one another, then how can we ever explain how one
finite object acts upon another? This is exactly the problem Hume raised about his
own brand of pluralistic empiricism. The idealistic answer to this problem is to say
that if the many different distinct finite objects are not regarded as really distinct
from one another but simply as all joined into a single Whole, then interaction
16 Some would maintain that James's radical empiricism was closer to phenomenology than it was to
traditional British empiricism. In what follows, I emphasize James's commitment to linguistic analysis
and his attack upon absolute idealism rather than his own brand of empiricism as important in preparing
the way for twentieth-century analytic philosophy.
17 The 1908 Hibbert Lectures at Oxford University were published as A Pluralistic Universe:
Hibbert Lectures at Manchester College on the Present Situation in Philosophy (London: Longmans,
Green, and Co., 1909).
18 Ibid., pp. 7-8.
19 See The Letters of William James, edited by Henry James (Boston: little, Brown, and Co., 1926),
Volume 2, p. 303.
20 James, A Pluralistic Universe, pp. 13-14.

12

ANALYTIC PHILOSOPHY OF REUGION

between the different individual things does not have to be understood as being
"transferred" from one finite thing to another but simply as an "internal" operation
of the monistic One?! James's response to this objection is to claim that the idealist
has put forth a "purely verbal" argument. By analyzing the meanings of the words
in the argument, for example, "distinct," "separate," and 'joined," James claims
that the difficult metaphysical problem of interaction is simply a matter of a verbal
disagreement between the pluralist and the monist. 22
James's lectures in Oxford in 1908 and the publication of "The Thing and Its
Relations" in 1905 added the voice of one of the most prominent American
philosophers to the grounds well against idealism that had begun with Moore and
Russell. 23 At the beginning of the twentieth century, the philosophical landscape
was changing from one dominated by idealism and rationalistic a priori reasoning
to one dominated by empiricism and conceptual and linguistic analysis. At the
same time, the change also was from a general philosophical climate that was
friendly toward traditional theism in terms of methodology and, in some cases,
substantive content to one that was hostile toward it methodologically and
substanti vely.
Although James was very sympathetic toward what we might call "the religious
sentiment," he offered very little support for traditional theism or the approach to
the philosophy of religion that we would now call natural theology. He was
particularly opposed to the traditional arguments for the existence of God and to
rationales that provided for the traditional metaphysical attributes of the deity.z4
His Varieties of Religious Experience emphasizes the vast cultural diversity of
religious experiences found among different peoples of different cultures of the
world without offering any grounds for pre-eminence of or even preference for
traditional theism, and his pragmatism causes him to be more interested in the
"cash-value" of a religious belief than in its metaphysical truth.
THE VIENNA CIRCLE AND LOGICAL POSITIVISM
There is little doubt that the movement in twentieth century philosophy known
as logical positivism (or, less frequently, "empirical positivism") has had more
influence upon twentieth-century Anglo-American philosophy of religion than any
other single philosophical development. Logical positivism originated with the
group of philosophers, scientists, and mathematicians who formed themselves into
the famed Vienna Circle at the University of Vienna in the early 1920s. Led by
such notable figures in philosophy as Moritz Schlick, Rudolf Carnap, Herbert
Feigl, Friedrick Waismann, and others, the positivists were initially prompted by
their concerns over what they considered to be the then current deplorable state of
Ibid., pp. 54-57.
Ibid., p. 57-60. This is an argument that follows directly from James's treatment of monism and
pluralism in his Pragmatism (Indianapolis, Ind.: Hackett, 1907), pp. 61ff.
23 Although one may argue that James himself still harbored some sympathy with religious belief,
James was no defender of traditional theism or organized religion. He rejected the traditional arguments
for the existence of God, theology, and religious institutions, and, true to his radical empiricism, he
insisted upon the philosophical significance of different religious experiences, which, of course, come
in many different varieties. His influential Varieties of Religious Experience is discussed in Chapter IX.
24 See William James, The Varieties of Religion Experience (New York: New American Library,
1958), pp. 329-31.
21

22

INTRODUCTION: THE RISE OF ANALYTIC PHILOSOPHY OF RELIGION

13

philosophical inquiry. From the mid-nineteenth century through the first quarter of
the twentieth century, science had made astounding progress. Correspondingly, the
nature of scientific inquiry had also made enormous progress toward becoming
much more rigorous and controlled. The notion of precise measurement (and hence
the importance of mathematics) and the resulting clarity, precision, and rigor had
become the cornerstone of modem science. By contrast, on the continent of
Europe, philosophy was still very heavily dominated by the influence of
nineteenth-century German idealism - primarily Neo-Hegelianism. The early
positivists were dismayed by the highly speculative and widely metaphysical
nature of philosophy and were intent upon modeling philosophical inquiry upon
scientific inquiry by incorporating the clarity of concepts and rigor of reasoning
found in scientific inquiry into philosophy.
MEANING AND TRUTH
At the very basic level of the positivists' approach to the restructuring of
philosophy was the adoption of the fundamental epistemology of scientific inquiry
according to which meaning is epistemologically fundamental and
epistemologically prior to truth. Before a scientist can perform any observations or
conduct any experiment in an attempt to confirm any hypothesis, the univocal
meaning of the hypothesis must be clear to the scientist. A hypothesis must be
understood as describing, corresponding, or otherwise "attaching to" the empirical
world in a single, univocal fashion before any attempts to investigate the scientific
accuracy of the hypothesis can begin. Any data resulting from observation or
experimentation will discriminate among competing hypotheses only if the data
that are confirming of one hypothesis are not confirming of the other. In order for
confirmation to proceed, then, hypotheses must be clearly distinguished from one
another, which means that the meaning of those hypotheses must be clear and
univocal. Indeed, before a scientist can decide how to design an experiment or
know what observations to perform or which data are particularly important to
collect to test the truth of a hypothesis, the meaning of the hypothesis must be
clearly understood.
Such an understanding of the underlying epistemological priorities of the
scientific method led the early positivists to adopt the epistemological prioritizing
of meaningfulness to truth. Meaningfulness is epistemologically more fundamental
than and logically prior to the notion of truth. There were many differences among
the early positivists, and, in a genuine sense, it is misleading to lump them all
together into a single, unified group or refer to them using a single label as the use
of the appellation "positivism" suggests. However, if there is a single unifying
factor among the different early positivists, it would be the epistemological ranking
of meaning as more basic and fundamental than truth. 25

25 For a thorough history of the Vienna Circle and early positivism, see G. 1. Warnock, ibid., pp. 3034; Victor Kraft, The Vienna Circle (New York: Philosophical Library, 1953); and A. 1. Ayer, ibid., pp.

12lff.

14

ANALYTIC PHILOSOPHY OF RELIGION

THE VERIFICATION CRITERION OF MEANING


The early targets of the "attack" of the positivists upon philosophy and their
attempts to restructure philosophy upon the bedrock of scientific inquiry were the
Neo-Hegelians of the early twentieth century who carried on - in one permutation
or another - what were, in the judgments of the positivists, the fanciful and
speculative metaphysical musings of Hegel. In an attempt to find some criterion or
standard by which such highly speculative claims could be identified and
distinguished from straightforward claims of science, the positivists developed
what was to become known as the verification criterion of meaning (or,
alternatively, the verification principle or the empirical criterion of meaning). Such
a criterion or principle was intended to provide the basis for determining which
utterances are empirically meaningful and which ones are not. Contrary to a
common misunderstanding of the verification criterion of meaning, there was
never a single version of the criterion that was agreed upon by all of the original
members of the Vienna Circle. The need for such a criterion and the
epistemological ordering of meaning and truth described above was undoubtedly
something upon which the positivists would have agreed, but no single version of
the criterion ever enjoyed the approval and support of all the early positivists.
Indeed, much of the writing of the early positivists was focused upon trying to
develop a "correct" version of the verification criterion of meaning that would
command widespread assent and agreement. Some versions were regarded as too
narrow, that is, they excluded too much (usually the universal laws of science that
need to be both universal and nomological to do their work in science and are thus
not empirical in any straightforward sense); others were considered to be too
broad, that is, they excluded too little (usually because in the effort to retain the
laws of science, metaphysical claims slipped by alsO).26
AYER AND THE ANGLICIZING OF POSITIVISM
Theology was a secondary target of positivism. It was not until positivism
became anglicized that its full impact was felt in theology and the philosophy of
religion. Positivism and its focus and concerns came to dominate much of AngloAmerican philosophy for more than a quarter of a century through one of those
sometimes underappreciated accidents of history through which political affairs
shape and determine intellectual movements. The early positivists fled Vienna and
Austria in the 1930s in advance of the development and spread of Nazism and
immigrated to universities in the United Sates or Great Britain. Their influence was
immediate and widespread, and thus postwar Anglo-American philosophy became
analytic Anglo-American philosophy.
The earliest and single most important work for spreading the influence of early
positivism and the significance of the verification criterion of meaning to the
English-speaking world was Language, Truth, and Logic by A. 1. Ayer. This book
set the tone for much of the philosophical inquiry in Great Britain and the United
States in the period immediately following World War II. Ayer, of course, was not
a member of the original Vienna Circle, but his book, more than any other by far,
26 See Carl Hempel, "Problems and Changes in the Empiricist Criterion of Meaning," in Semantics
and the Philosophy of Language, edited by Leonard Linsky (Urbana: University of lllinois Press, 1952).

INTRODUCTION: THE RISE OF ANALYTIC PHILOSOPHY OF RELIGION

15

was responsible for popularizing positivism. Ayer originally developed both a


"strong" and a "weak" sense of verification for consideration for the criterion. The
strong sense of verification captures the attempts by the early positivists to develop
an adequate version of the criterion. "A proposition is said to be verifiable in the
strong sense of the term," Ayer claims, "if, and only if, its truth could be
conclusively established in experience.,,27 Ayer recognized that such a formulation
is too narrow since, as we have indicated, the general, universal laws of science
that are thought to apply to an infinite number of cases, including cases in the
future, cannot be conclusively verified by experience and are thus in the same boat
as speculative, metaphysical claims. Ayer thus falls back upon the weak sense of
verification. A proposition is verifiable in the weak sense, Ayer says, "if it is
possible for experience to render it probable.,,28
By applying this criterion to the many areas of philosophical inquiry, Ayer
determines that most of the traditional problems of philosophy are not really
legitimate areas of inquiry. "The function of philosophy" is simply the clarification
of concepts and is not concerned with adding any information to our store of
knowledge at all. Philosophy, Ayer says, is simply a division of logic. 29 Philosophy
becomes completely absorbed with language. The only meaningful statements are
tautologies and observation statements, and philosophical inquiry does not simply
begin with or use linguistic analysis or conceptual analysis, philosophy is linguistic
or conceptual analysis.
Language, Truth, and Logic was also tremendously important because Ayer
devoted a separate and complete chapter to a critique of ethics and theology. 3D
Normative statements in ethics, Ayer claimed, are not really statements with truthvalue at all and are hence neither true nor false. They are expressions of emotion or
feelings or sometimes commands to others. We cannot legitimately dispute
normative values, Ayer claims; we can only dispute the empirical facts of different
situations about which we then develop certain feelings that are noncognitive. The
same is true of aesthetics for Ayer. Emotive and noncognitive theories of ethics
were prompted by this application of the verification criterion of meaning to ethics.
Much the same is true of theology. The traditional claims of the theist - that God
exists, God is loving, and humans have immortal souls - are not genuine
propositions at all, despite their declarative form. Such claims thus assert nothing
and are consequently meaningless. 31
The impact of Language, Truth, and Logic upon English-speaking philosophers
was enormous. Although mostly derivative from the work of the early positivists,
it, along with the dissolution of the original Vienna Circle and the spread of its
members to the United States and Great Britain, ensured the immediate and
widespread adoption of the fundamental tenets of positivism by much of the
English-speaking philosophical world.

A. J. Ayer, Language, Truth, and Logic (New York: Dover Publications, 1936), p. 37.
Ibid.
29 Ibid., p. 39.
30 See ibid., Chapter VI, "Critique of Ethics and Theology."
31 Ibid., p. 117.
27

28

16

ANALYTIC PHILOSOPHY OF RELIGION

THE DECLINE OF POSITIVISM


The failures and ultimate demise of logical positivism are well documented.
Logical positivism had held forth the promise of curing philosophy of all its ills
and constructing it along the lines of science, but it was never successful in
completing this program. The problems were many. Despite all of the efforts and
attention, no one ever successfully constructed the ideal version of the verification
criterion of meaning that would clearly demarcate along the desired lines by ruling
out as meaningless metaphysics and theology while maintaining the
meaningfulness of statements of scientific laws and simple statements about the
future and the past.
The logical status of the verification criterion itself became a significant issue
that was never resolved satisfactorily. Any particular version of the verification
principle is supposed to contain the criteria for determining empirical
meaningfulness. However, the principle obviously cannot be applied successfully
to itself, that is, if any version of the verification principle is judged by its own
standards, it will be judged empirically meaningless. A moment's reflection would
indicate immediately why this must be the case. The verification principle is
intended as a completely general thesis about the nature of language, and, as such,
it must be a second-order (or metalinguistic) claim about how language is related
to the world. It cannot then include itself in the general thesis that it describes or
judge itself by the criteria that it establishes.
The responses of the positivists to the recognition of the difficulty involving the
logical status of the verification principle and the question of its own
meaningfulness were that it is a rule or principle or definition and was never
intended to be regarded as empirically meaningful. For example, by the time of the
second edition of Language, Truth, and Logic, Ayer had adopted the position that
the verification principle is a definition. 32 But if the verification principle is simply
a definition, what recommends it over other definitions? By explicitly embracing
this position, Ayer recognizes that he opens the possibility of different competing
definitions with resulting different kinds of meanings. Although he does not want
the matter of deciding among different competing definitions to be completely
arbitrary, he does not offer a compelling argument for why anyone should prefer
the verification principle over any other definition of meaning. The answer, of
course, is that the choice of definition depends upon what one wants it to do and
how one wants the landscape of meaningfulness divided. In other words, it seems
that the positivists can defend their preference for the verification principle only
because of the initial commitment to preserving science and eliminating
metaphysics, ethics, aesthetics, and theology. As we shall later see, the recognition
of the apparently arbitrary status of the verification principle opened up the
possibility for others to embrace different kinds of meaning - including poetic
meaning, ethical meaning, aesthetic meaning, and, of course, religious meaning.
THE END OF POSITIVISM
Other problems contributed to the demise of positivism. Much of the general
program of the positivists had been built upon the many developments that had
32

Ibid., p. 16.

INTRODUCTION: THE RISE OF ANALYTIC PHILOSOPHY OF RELIGION

17

taken place in symbolic logic in the latter half of the nineteenth century and at the
beginning of the twentieth century. In particular, the apparent success of Alfred
North Whitehead and Bertrand Russell in Principia Mathematica in constructing a
system that was adequate for reducing mathematics to logic held the promise that a
single logical system could be adequate for constructing a system within which a
complete description of the world might be given containing every single
meaningful proposition. Rudolf Carnap, a member of the original Vienna Circle,
attempted to construct such a system in Der Logische Aujbau der Welt (The
Logical Structure of the World). Using the logic from Whitehead and Russell's
Principia Mathematica, Carnap attempted to construct a formal, metaphysical
system within which every single phenomenon in the world - including physical,
phenomenal, psychological, and societal - could be deduced from primitive terms
describing immediate experience (sense data). In the preface to the second edition
of the Aujbau, Carnap admits his failure in this monumentally ambitious effort.
Carnap's failure was the failure of a thorough and general program of logical
positivism.
Contributing to the inflationary number of claims about different kinds of
meanings was the additional doubt that was raised concerning the uniqueness of
the logic of Principia Mathematica. Kurt Godel developed his famous "proof'
(based upon an ingenious use of mapping), which demonstrated that any formal,
logical, axiomatic system that was rich enough to reduce mathematics had to be
either inconsistent or incomplete. Any such system must either be inconsistent by
allowing within itself the proof of a theorem and its formal negation or it must
exclude from itself as provable some theorem that is known to be a theorem of the
system. With Godel's proof, the logical foundation for logical positivism is
threatened. If the system upon which all logic and mathematics are built is suspect,
then there is no foundation upon which to base the general program of logical
positivism. Positivism was faced not only with the possibility of different kinds of
meanings but with the possibility of different kinds of languages and logics as well.
ORDINARY-LANGUAGE PHILOSOPHY
A common perception is that analytic philosophy in the twentieth century has
been completely dominated by logical positivism. There is little doubt that much of
Anglo-American analytic philosophy was dominated by positivism for more than
two decades. As a philosophical movement, however, logical positivism was shortlived, and by the early 1950s its influence had waned considerably. By 1955, for
example, in the preface to one of the earliest volumes of analytic philosophy of
religion of the postpositivism era, Anthony Flew and Alasdair MacIntyre indicated
that though the label of logical positivism persisted as a movement in philosophy it
was "now defunct.,,33
The other "school" of analytic philosophy, which began to rise to a position of
dominance in analytic philosophy in midcentury, was ordinary-language
philosophy. Ordinary-language philosophy finds its origin in the later works of G.
E. Moore and Ludwig Wittgenstein of Cambridge and the works of Gilbert Ryle of
33 Anthony Flew and Alasdair Macintyre, New Essays in Philosophical Theology (London: Student
Christian Movement [SCM] Press. 1955). p. vii.

18

ANALYTIC PHILOSOPHY OF REUGION

Oxford. 34 Much of the variety among different analytic philosophers is accounted


for largely in terms of what sort of language they consider to be the legitimate
linguistic unit for philosophical use and analysis. The ordinary-language
philosophers were united by their belief that ordinary language is the appropriate
linguistic vehicle for conducting philosophical inquiry and, as such, is the
appropriate target of linguistic analysis. Interestingly, even different ordinarylanguage philosophers meant different things by "ordinary-language philosophy,,,35
but let us here take "ordinary language" just to mean nontechnical language - the
language that people might use in the "ordinary" course of the day in conversation
with others and in conducting their personal affairs. 36 Taking ordinary language in
this sense, we can say that the ordinary-language philosophers eschewed the idea
of an ideal, logically perfect language in favor of the use of ordinary language as
the proper medium for pursuing philosophical issues and for developing
philosophical theories. Furthermore, ordinary-language philosophers held that
many philosophical problems are really pseudoproblems because they have arisen
through mistakes that philosophers have made about ordinary language. Hence,
such problems can be "dissolved" through philosophical analysis by simply tracing
the pseudoproblem to its source in the misunderstanding or misinterpretation of
ordinary language.
A good example of the Oxford branch of ordinary-language analysis is Gilbert
Ryle's The Concept of Mind, in which Ryle develops a reductionist and
behavioristic theory of the nature of the mind that is based, in large part, upon the
way in which ordinary language is used to talk about the mind and behavior. In the
first chapter of this work, which has become a classic of ordinary-language
analysis, Ryle dismisses Cartesian dualism as "the Myth of the Ghost in the
Machine." According to Ryle, the Cartesians make the very fundamental logical
error of committing a "category mistake," that is, the mistake of applying concepts
or categories that meaningfully apply to one logical type to another logical type for
which the concepts or categories are logically inappropriate?7 Of course, Ryle's
attack upon Cartesian dualism is an indirect attack upon the assumed metaphysics
of much of Judaic-Christian theology since the time of Descartes. The Concept of
Mind pushes in the direction of reductionism and materialism and would be as
hostile toward theological talk of an immortal soul as it is toward metaphysical talk
of an immaterial mental substance.
With the move that occurs within ordinary-language philosophy away from the
notion of an ideal, logically perfect language, much of the philosophical
underpinning of the opposition of analytic philosophy to the philosophy of religion
begins to erode. The work of the later Wittgenstein further undermines much of the
34 For a detailed historical treatment of ordinary-language philosophy, see Passmore, ibid., Chapter
18, pp. 431-75.
35 See, for example, Gilbert Ryle, "Ordinary Language," The Philosophical Review, Vol. 62, 1953.
Reprinted in Philosophy and Ordinary Language, edited by Charles E. Caton (Urbana: University of
lllinois Press, 1963), pp. 108-127. Also included in Ordinary Language, edited by V. C. Chappell
(Englewood Cliffs, N.J.: Prentice-Hall, 1964). Also, see Norman Malcolm, "Moore and Ordinary
Language," in The Philosophy of G. E. Moore, edited by Paul Arthur Schilpp (Evanston, lll.: Open
Court, 1942). Reprinted in Chappell, ibid., pp. 5-23.
36 See Caton, ibid., pp. vi-vii.
37 Gilbert Ryle, The Concept of Mind (New York: Barnes and Noble, 1949), pp. 16ff.

IN1RODUCTION: THE RISE OF ANALYTIC PHILOSOPHY OF RELIGION

19

basis for the hostility of analytic philosophy toward philosophy of religion.


Although Moore and Russell had been the dominant figures of the early stages of
the development of philosophical analysis, Russell's influence upon the direction
of the development of analytic philosophy was far greater than Moore's in the first
four decades of the century. Russell's interest in the development of a logically
perfect language for doing philosophy had dominated not only his own work but
the work of the early Wittgenstein and the attempt to develop a theory of logical
atomism as well. With the rise of ordinary-language philosophy, the influence of
Moore overtakes the early influence of Russell.
THE LATER WITTGENSTEIN
The later Wittgenstein begins his Philosophical Investigations with a thorough
critique and rejection of the view of language that he had inherited from Frege and
Russell 38 and the development of a new understanding of language. This earlier
view of language is characterized by the belief that words are names and take their
meaning by referring to objects; propositions were regarded as combinations of
names and were thought to take their meaning by referring to states of affairs. Such
beliefs lead to a referential theory of meaning and a correspondence theory of truth
and also tend to support a foundationalist view of language according to which
language "connects" with the world at its most basic, fundamental, and primitive
level. 39 This was the view of language upon which the entire enterprise of logical
atomism was based. Although Wittgenstein was not himself a positivist, this same
view of language was the view upon which the "reductive analysis" of the logical
positivists was based.40 The fundamental underlying metaphilosophical and
metalinguistic view of this kind of language and its philosophical importance was
that the world was a certain way with a certain logical construction, and in order
for language to be capable of describing the world and asserting true and false
claims about the world, it also had to have a certain logical construction. Because
of the underlying commitment to empiricism and to the correspondence theory of
truth, neither the meaningfulness nor the truth of general metaphysical statements
could be accounted for with this view of language. This view of the nature of
language led, as we have seen above, to the rejection of metaphysics by the logical
positivists using the verification criterion of meaning. It also led to the infamous
problem of self-reference of Wittgenstein's Tractatus Logico-Philosophicus,
according to which, if the theory he develops there is really the correct way to
describe how language is related to the world, then the theory cannot really be
consistently stated.

38 This treatment of Wittgenstein's changing views concerning the nature of language is very brief
and superficial. For a more thorough treatment of his rejection of his earlier view of language and his
later view in the Philosophical Investigations, see James F. Harris, "Language, Language Garnes, and
Ostensive Definition," Synthese, Vol. 69,1986, pp. 41-49.
39 See ibid., 43-44. In terms of the language, this means that the language is ''truth-functional,'' that
is, the truth values of complex statements in the language are functions of the truth values of the most
simple statements in the language.
40 See G. P. Baker and P. M. S. Hacker, Wittgenstein: Understanding and Meaning (Oxford: Basil
Blackwell, 1980), pp. 34ff.

20

ANALYTIC PHILOSOPHY OF REUGION

ESSENTIALISM AND NONESSENTIALISM


For our consideration here, the most important feature of this view of language
that had dominated so much of Western philosophy from the time of Aristotle
onward is its essentialism. Much of earlier analytic philosophy - including
especially the logico-mathematics of Russell and the positivism of the Vienna
Circle - was based upon the understanding of language according to which each
assertive statement has a single, determinate meaning. According to different
analytic philosophers, each proposition's determinate meaning might be
determined by a unique set of truth conditions, or by the unique state of affairs in
the world that it described, or by the unique method of its verification, or by its
unique "ordinary meaning." In the Philosophical Investigations, Wittgenstein
abandons such essentialism in favor of a theory of "meaning as use," and language,
like tools, can be put to many different uses. "The function of words are as diverse
as the functions ofthese objects [tools]," he says.41 Wittgenstein also abandons the
understanding of language according to which asserting facts using declarative
sentences that are either true or false is the only philosophically important or
interesting use of language in favor of an understanding of language according to
which there are many different uses of language - all equally philosophically
important and interesting. He gives some examples that illustrate his changed
understanding of the richness of language and the myriad of uses to which it can be
put:
How many different kinds of sentence are there? ... There are countless kinds:
countless different kinds of use of what we call "symbols", "words",
"sentences" ....
Review the multiplicity of language-games in the following examples, and in
others:
Giving orders and obeying them Describing the appearance of an object, or giving its measurements Constructing an object from a description (a drawing) Reporting an event Speculating about an event Forming and testing a hypothesis Presenting the results of an experiment in tables and diagrams Making up a story; and reading it Play-acting Singing catches Guessing riddles Making a joke; telling it Solving a problem in practical arithmetic Translating from one language into another Asking, thanking, cursing, greeting, praying. 42

41 Ludwig Wittgenstein, Philosophical Investigations, translated by G. E. M. Anscombe (New York:


Macmillan, 1953), Section 11.
42 Ibid., Section 23.

INTRODUCTION: THE RISE OF ANALYTIC PHILOSOPHY OF RELIGION

21

Each of these different uses of language involves a different context - with


different speakers with different intentions and purposes and different audiences.
Each of these different contexts, including the speakers, the language, and the
gestures, constitutes what Wittgenstein called a language-game. One of the most
important features of a language-game is that it is "rule governed," which means
that certain conventions operate within the language-game with the conventional
force to determine the meanings of the different uses of language by establishing
the correct and incorrect uses of language within the language-game. Participation
in different language-games constitutes what Wittgenstein calls a "form of life."
The differences between the early and the later Wittgenstein's views concerning
language are not always so sharp as I have made them here. Several Wittgenstein
commentators have argued that the later Wittgenstein's view regarding
nonessentialism is consistent with his earlier views. For example, Rush Rhees
claims that the view of language as a calculus is still exercising an influence on the
later Wittgenstein since when one uses a calculus for something, it has a function,
and it gets its meaning from that function. 43 Also, in understanding the importance
of how Wittgenstein's analogy of the different uses of tools illustrates the different
uses oflanguage, it is important to distinguish Wittgenstein's interests in the notion
of "the use of language" from those of J. L. Austin (discussed below). Austin's
concerns are directed toward actual, specific situations of "what to say when" and
involve paying close attention to the subtle differences among different things that
result in our calling those different things by different names or in our describing
those different things using different expressions. Wittgenstein's concerns with use
might be described as "deeper." His concerns are directed toward the more
fundamental presumptions about the nature of language that allow for the very
possibility of using language in different ways.44 Wittgenstein is thus concerned
with the "grammar" (the conceptual underpinning) that allows us to use language
in different ways for different purposes (in different "language-games"). The
importance of Wittgenstein's views regarding grammar and language-games are
discussed in detail in connection with the views of D. Z. Phillips in Chapter II and
ChapterIV.
Not only is meaning and the correct use of language relative to different
language-games with their different rules, but ontology also is relative to different
language-games. In the Tractatus, the view of language that Wittgenstein adopted
was a view that he had inherited from Frege and Russell. According to this
foundationalist view, language "connects" with the world ontologically only at its
most primitive, simple level. 45 The process of analysis was intended to be one
where complex, "higher-level" uses of language were analyzed into their most
simple component elements to determine the final ontological commitments of a
language. Such a procedure was behind Russell's attempt to replace inferred
entities with constructed ones and Russell and Wittgenstein's effort to explain the
43 Rush Rhees, Wittgenstein and the Possibility of Discourse (Cambridge: Cambridge University
Press, 1998).
44 The difference between Wittgenstein and Austin is discussed at length in Stanley Cavell, The
Claim of Reason (Oxford: Oxford University Press, 1979), pp. 65ff. I am grateful to D. Z. Phillips for
bringing this point to my attention.
45 See Harris, ibid., pp. 43ff.

22

ANALYTIC PHILOSOPHY OF REUGION

entire universe from the atomic propositions in logical atomism. It was this same
view of language upon which the "reductive analysis" of Gilbert Ryle's
"Systematically Misleading Expressions" was based. 46 When Wittgenstein
abandons this view of language in the Philosophical Investigations for the view
based upon language-games, he also abandons any sort of attempt to "connect"
language to the world ontologically in the same sort of foundationalist fashion.
What sort of things our language commits us to now depends entirely upon the
context, the language-game in which the question arises. For example,
Wittgenstein suggests that if we think of how a particular composite is to be
divided up into its simple parts, the process has no definite answer outside a
particular language-game. 47 Consequently, ontology as well as meaning becomes
relative to different language-games for the later Wittgenstein.
The development of the notion of language-games in the later Wittgenstein has
proven to offer a welcomed philosophical solace for some philosophers of religion.
Whereas the strictly empirical language of the logical positivists with its
verification criterion of meaning and its strict ontology of empirical facts left no
room for meaningful claims about the existence of God, Wittgenstein's
introduction of language-games appears to leave open both questions of meaning
and ontology. According to the later Wittgenstein, we must first examine the
particular language-game within which questions about the meaning of a particular
claim or the existence of a particular object are raised and then address the
questions within that language-game.
The development of language-games obviously provides a great deal more
flexibility for the traditional theist to defend claims about the existence and nature
of God. As I discuss in Chapter II and Chapter IV, several prominent philosophers
of religion, most notably D. Z. Phillips and Paul van Buren, have developed
positions that examine the meaningfulness of the claims of theism based upon
Wittgenstein's notion of a language-game.
SPEECH ACTS AND LINGUISTIC PHILOSOPHY
The last identifiable stage in the development of twentieth-century analytic
philosophy came in the period following World War II at Oxford University.
Sometimes called "Oxford Philosophy" or the "Oxford School of OrdinaryLanguage Philosophy," this form of analytic philosophy was dominated by the
figure of John L. Austin. Trained originally in the classics, Austin brought with
him to the study of philosophy what was perhaps a unique ability to notice and
appreciate the multifarious subtleties of language. It is sometimes said of Austin
that he was interested in the study of language for its own sake, because he never
constructed any general philosophical theories and seldom made explicit the
philosophical implications of his work on language. However, the truth is that
Austin thought that general theories and even the philosophical implications of
language could only come as the result of an extremely careful and thorough study
of language that might continue for generations through the work of several

46

See Gilbert Ryle, "Systematically Misleading Expressions." Reprinted in Klemke, ibid., pp. 287-

306.
47

See Wittgenstein, ibid., Sections 47 and 48.

IN1RODUCTION: THE RISE OF ANALYTIC PHILOSOPHY OF RELIGION

23

different people. In this sense, the methodology of his linguistic philosophy


approached that of a "science oflinguistics."
Although Austin has had a significant influence upon analytic philosophy in a
number of ways, unqoubtedly, his most significant influence upon the direction of
the philosophy of religion came as a result of his development of "speech-act
theory." Like the later Wittgenstein, Austin continually insisted that philosophers
ought to be more interested in the variety of different kinds of uses of language and
ought to avoid concentrating solely upon the assertive or descriptive uses of
language - a mistake that Austin called "the descriptive fallacy." Austin's speechact theory focuses upon the particular locutions of a speaker rather than upon
"language" or "sentences" or "words" and is based upon the tripartite distinction
between locutionary acts, illocutionary acts, and perlocutionary acts. Using one of
Austin's famous examples, a particular speaker may say to a particular audience A,
"I promise that I'll be home by ten o'clock." S then performs the locutionary act of
issuing the utterance (with a definite sense and reference), the illocutionary act of
promising to be home before a certain time, and the perlocutionary act of causing
A to believe that S will be home before that time. The locutionary act is the act of
saying something, the illocutionary act is the act performed in saying something,
and the perlocutionary act is the act performed by saying something. 48
The general effect of Austin's theory of speech-acts and his linguistic
philosophy is to provide even further flexibility and greater specificity at the same
time to our understanding of linguistic analysis. Again, like the later Wittgenstein,
Austin abandons the essentialism of the early Russell and the logical positivists and
its focus upon the abstract meaning of a proposition in favor of the concrete
situation in which a speaker says something to a particular audience. Similar to the
notion of language-games, Austin's theory of speech-acts shifts the focus of
philosophical analysis to a specific situation wherein a particular speaker says
something to a particular audience, with a particular intention, a particular force,
and a particular consequence.
Speech-act theory provides a rich avenue of development for understanding
certain kinds of religious utterances that is still being explored. We might say that
Austin further loosens the tight bonds of essentialism and provides still another
different way of analyzing and understanding religious utterances (discussed in
Chapter II). In general, speech-act analysis provides the framework within which it
becomes possible to talk about and analyze "religious utterances" on a par with
"scientific utterances" or any other kind of utterance. In order to use speech-act
analysis to understand religious language, we must specify a particular utterance
issued by a particular speaker on a particular occasion under particular
circumstances to a particular audience. The requirement for such detailed
specificity may seem daunting and discouraging, but certainly speech-act analysis
is more methodologically friendly to the philosophy of religion than were earlier
forms of analytic philosophy.

48 See John L. Austin, How to Do Things with Words, edited by J. O. Urmson (Oxford: Oxford
University Press, 1962), pp. 98ff.

24

ANALYTIC PHILOSOPHY OF RELIGION

POSTANALYTIC PHILOSOPHY?
In the last two or three decades of the twentieth century, several major figures
have raised serious objections against not only the possible success of an analytical
approach to philosophical problems based upon the analysis of language and
concepts but also against the very possibility and the very coherence of such an
approach. One of the more prominent of these critics is Richard Rorty. He
contends that there is no such thing as an objective method of inquiry, no such
thing as epistemology, and no such thing as objective reality. It is not simply that
logical positivism failed to produce on its grand program, as I have described
earlier, but the entire enterprise involving the logical analysis of concepts and
language is wrong-headed and nonsensical, according to Rorty. Analytic
philosophy burned itself out and showed itself to be a dead-end approach to
philosophy. "The notion of 'logical analysis' turned upon itself and committed
slow suicide," he claims. 49 The sources for the threat of self-destruction within
analytic philosophy are many, and I cannot examine those in detail here. However,
the most influential attacks upon the fundamental distinctions and commitments of
analytic philosophy that have arisen from within the analytic tradition itself are
worthy of brief mention. Perhaps chief among these attacks has been Willard Van
Orman Quine'S rejection of the analytic-synthetic distinction and his own holism
and Thomas Kuhn's analysis of the history of science and science method as
paradigm-based. 50 According to Quine, there is not a qualitative difference
between analytic and synthetic statements or between theory and content or
between logic and fact, only a difference in degree. All statements are parts of a
"web" or "network" and occupy different positions of relative pragmatic
importance to the web; however, "no statement is immune to revision," and,
conversely, any statement may be held onto, "come what may." The decision, for
Quine, is always a pragmatic one relative to the particular network. For Kuhn,
methodology, facts, and truth are relative to a paradigm that controls and regulates
the method of inquiry and what counts as an accurate and satisfactory result of the
inquiry, and the choice of paradigms cannot be a rational one based upon reason or
evidence or data since all of these are relative to paradigms. For followers of Kuhn,
classic Western science is simply the result of adopting one paradigm, and religion
is the result of adopting another paradigm. s1 These and other critiques of analytic
philosophy and, more generally, empiricism have eroded and blurred the
fundamental distinctions between method and substance, form and content, and
logic and fact. Many philosophers, following Quine and Kuhn, have rejected the
traditional understandings of epistemology and the scientific method. The result

49 Richard Rorty, Consequences of Prag11Ultism (Minneapolis: University of Minnesota Press, 1982),


p. 227. Also see Richard Rorty, Philosophy and the Mirror of Nature (Princeton, N.J.: Princeton
University Press, 1979), pp. 323ff.
50 Willard Van Orman Quine, "Two Dogmas of Empiricism," in From a Logical Point of View (New
York: Harper and Row, 1953), and Thomas Kuhn, The Structure of Scientific Revolutions, Second
Edition (Chicago: University of Chicago Press, 1970). For a critical discussion, see James F. Hams,
Against Relativism: A Philosophical Defense of Method (LaSalle, Ill.: Open Court, 1992), Chapter 2 and
Chapter 4.
51 See, for example, Paul Feyerabend, "How to Defend Society against Science," in Scientific
Revolutions, edited by Tan Hacking (Oxford: Oxford University Press, 1981).

INTRODUCTION: THE RISE OF ANALYTIC PHILOSOPHY OF RELIGION

25

has been what many have characterized as a general and radical logical,
ontological, and epistemological relativism.
At the same time, Anglo-American analytic philosophy has been challenged by
European philosophy. Whereas analytic philosophy was identified with and the
heir to the historical tradition of British empiricism, the new, "postanalytic"
philosophy championed by Rorty and others is aligned with the post-Kantian,
European thought of Hans-Georg Gadamer, Jiirgen Habermas, and Jacques
Derrida.52 The influence of what is typically identified as "continental" philosophy
- including deconstructionism and postmodernism - has become so strong in the
last few decades of the twentieth century that the period might well be known to
future historians of intellectual thought as The Age of Relativism. 53
POSTANALYTIC PHILOSOPHY AND THE PHILOSOPHY OF RELIGION
The rise of postmodernist, postanalytic, and postempiricist approaches to
philosophy has had significant influence upon both the general philosophical
climate and the more practical aspects of academic departments of philosophy and
religion (particularly in the United States but in Great Britain as well). These
challenges to analytic philosophy have also had significant repercussions for the
philosophy of religion - both in terms of philosophical approaches to traditional
problems in the philosophy of religion and the actual constituents of academic
departments of philosophy and religion. These changes have been the object of
some interest and discussions among philosophers of religion and theologians
alike. 54 Natural theology, which has historically dominated analytic philosophy of
religion, has come under increasing attack and has been replaced by many thinkers,
who are generally regarded as belonging to the analytic tradition, by an approach to
the philosophy of religion that relies upon revelation, religious teachings or
doctrine, or religious epistemology - an approach that is more akin to theology.55
CONCLUSION
For a variety of different reasons, many people identify twentieth-century
analytic philosophy with logical positivism and its dismissive treatment of religion;
consequently, many people assume or infer that all twentieth-century analytic
philosophy has the same hostility to the philosophy of religion. I have indicated
here that twentieth-century Anglo-American analytic philosophy was not
monolithic or homogenous, and I have shown a definite evolution in the nature of
analytic philosophy. The consequences of the differences among different analytic
philosophers differ widely as far as the philosophy of religion is concerned - some
are very hostile to the philosophy of religion and some are very friendly toward it.
52 For a variety of different critical attacks upon analytic philosophy from diverse philosophical
perspectives, see After Philosophy: End or Transformation, edited by Kenneth Baynes, James Bohman,
and Thomas McCarthy (Cambridge: MIT Press, 1988), and Post-Analytic Philosophy, edited by John
Rajchman and Cornel West (New York: Columbia University Press, 1985).
53 See Harris, ibid., pp. Iff.
54 See, for example, God, Philosophy, and Academic Culture, edited by William J. Wainwright
(Atlanta, Ga.: Scholars Press, 1996).
55 Such approaches are discussed in Chapter IV. I have in mind here the work of Alvin Plantinga
involving reformed epistemology, the work of William Alston involving religious perception, and the
work of Marilyn Adams regarding evil and suffering.

26

ANALYTIC PHILOSOPHY OF REUGION

The evolution of the philosophical climate within analytic philosophy has generally
been from one of hostility toward the philosophy of religion to one that is much
more favorable. Logical positivism, the form of contemporary analytic philosophy
most hostile toward the philosophy of religion, was relatively short-lived; however,
though strict positivism itself was soon rejected by the philosophical community,
much of the methodology and many of the problems of twentieth-century analytic
philosophy of religion - as evidenced in natural theology - are the result of the
"restructuring" of theology that took place as a result of "the Revolution in
Philosophy."
In the latter half of the nineteenth century, there was little to distinguish in terms
of the methodology (and some, as we have seen, would say in terms of the content)
between idealism, which dominated much of philosophy, and theology. The rise of
contemporary analytic philosophy in the first half of the twentieth century created a
gulf between philosophy and theology that was, in some cases, the result of overt
hostility and, in other cases, the result of aggressive indifference. During this
period, in the "new" philosophical climate, philosophers and theologians no longer
competed with different theories on a general, metaphysical scale. The result was
that analytic philosophy and theology became more distinct and autonomous in
their respective pursuits than perhaps they had been at any other time in the history
of Western intellectual thought. Philosophers writing in midcentury about the
impact of analytic philosophy upon theology routinely refer to the "old" and "new"
way of addressing religious problems philosophically. 56 The rise of the philosophy
of religion as a discipline distinct from theology is, at least partially, the result of
addressing religious questions that arose from analytic philosophy. 57
Following the watershed of the decline of logical positivism, the "new" way of
doing philosophy in English-speaking countries became much more friendly
toward the philosophy of religion, and, with the threat of postanalytic philosophy,
things have progressed even further in many quarters. Many analytic philosophers
now admit that there are legitimate philosophical issues arising out of religious
beliefs and practices that can be addressed in a manner using the general
techniques of analytic philosophy. Or, alternatively, following the various attacks
upon analytic philosophy, philosophers previously associated with the analytic
tradition are willing to admit the limitations of that tradition and use different and
more "theistic friendly" methods of inquiry.
It is useful to understanding the role of a contemporary philosopher of religion
to consider the analogy of a movie critic. A movie critic critiques the product of the
filmmaker by analyzing the plot, the characters, the dialogue, the direction, the set,
and the other important aspects (maybe comparing it with other similar movies) of
a movie to form judgments about the quality and value of a movie. Ideally, the
movie critic makes his or her judgments by appealing to some sort of general
56 See, for example, the Introduction to Faith and Logic, edited by Basil Mitchell (London: George
Allen and Unwin, 1957), pp. Iff. See also Austin Farrer, "A Starting-Point for the Philosophical
Examination of Theological Belief," in Mitchell, ibid. Compare also Willem F. Zuurdeeg, An Analytical
Philosophy of Religion (New York: Abingdon Press, 1958), pp. 15ff.
57 The case could also be made that the rejection of Hegelian idealism by such nineteenth-century
existentialist thinkers as Sliiren Kierkegaard and his existentialist approach to religious questions is also
a contributing factor in the rise of the philosophy of religion as distinct from theology.

INTRODUCTION: THE RISE OF ANALYTIC PHILOSOPHY OF RELIGION

27

principles or some methodology and gives some reasons or rationale for those
judgments that then allows an audience to understand the grounds upon which the
judgments were made. This analysis provides the basis for an understanding of
why one may agree or disagree with those judgments. The philosopher of religion
critiques the work (the body of claims) of the religious believer or religious
practitioner by analyzing the plot (the religious account of things), the characters
(God, humans, and perhaps angels), the dialogue (religious language), the set
(earth, the universe, and perhaps heaven and hell), and other important aspects
(maybe comparison with other religions or alternative nonreligious accounts). To
conduct such a "review," we would normally think that it would be preferable for a
critic to be neutral - neither an advocate nor a protagonist of the film or the
religion to be critiqued. We are not as likely to trust the review of a movie critic if
we know that the critic has invested money in the production of that particular
movie or in one of its rivals. If we know that the reviewer has a vested interest in
the product, then we are likely to demand a more explicit explanation of the
rationale for the judgments and a higher standard for the methods and the
principles that are the bases for those judgments. The ideal of the analytic
philosopher of religion is to commit to an objective and neutral methodology that
involves the analysis of language and concepts in a genuine attempt to understand
and elucidate the claims and practices of religion. When the inquiry is conducted
by theists and in defense of theism, we ought to be especially cautious to scrutinize
both the method of inquiry and the results. While some might dispute whether the
methodology is objective and neutral and others might dispute whether it is
profitable, analytic philosophers have pursued such a course of inquiry, which has
resulted in a body of challenging and compelling literature addressing a core of
fundamental problems and issues. The inquiry, debate, and criticism concerning
this body of literature and this core of problems and issues comprises
contemporary analytic philosophy of religion.

II. The Problem of Religious Language

With the rise of analytic philosophy in the early part of the twentieth century
and its emphasis upon linguistic analysis, it should not be surprising that a
significant crisis developed by midcentury among both philosophers and
theologians concerning religious language. That there was such a crisis is
evidenced both by the explicit recognition of the challenge in the writings of
several leading figures in both the philosophy of religion and in theology and by
the plethora of books and articles that appeared in print in the period from
approximately midcentury until ten years or so afterward. 1 There was much
disagreement among different analytic philosophers; however, the one underlying,
common tenet upon which nearly all of them would have agreed is that language is
the one continuous thread from which the entire fabric of religion and religious
belief is woven. Some of the problems with religious language are illustrated by
the more extreme positions, for example, A. J. Ayer's claim that the language of
theology is meaningless and nonsense and Paul van Buren's claim that "the word
'God' is dead.,,2 Other concerns were prompted by the repercussions of attention to
language by the more moderate analytic philosophers and the elevation of the
importance of the analysis of language for philosophical or theological pursuits.
RELIGIOUS LANGUAGE AND RELIGIOUS KNOWLEDGE
The way in which an understanding and analysis of religious language was
viewed as fundamental to all of religion is perhaps best illustrated by the
relationship between religious language and religious knowledge. A good way to
understand the inseparable and symbiotic relationship between religious language
and religious knowledge is to see the epistemic content of religious claims as
dependent upon and derivative from an analysis of the cognitive content of those
claims, that is, the content on the basis of which those claims might be regarded as
1 A casual inspection of the copyright dates of the references in this chapter will indicate how
concentrated the interest in religious language was from the late 1940s through 1960. Although the
publication of some additional original material continued beyond that date for some thinkers, the
1960s, for the most part, saw the appearance of secondary source material in the way of books,
anthologies, and articles responding to earlier works.
2 See A. J. Ayer, Language, Truth, and Lagic (New York: Dover Publications, 1956), Chapter 6, and
Paul van Buren, The Secular Meaning o/the Gospel (New York: Macmillan, 1965), p. 103.

28

THE PROBLEM OF RELIGIOUS LANGUAGE

29

true or false. Consequently, much of the attention during this period was focused
upon issues concerning "the cognitive meaningfulness" of religious language and
the impact of these issues upon claims concerning religious knowledge. It appears
that religious language has to be cognitively meaningful in order to have epistemic
content or value. The problem of religious language and the problem of religious
knowledge came to be regarded as inextricably interwoven,3 and a reprioritizing of
issues and problems within the philosophy of religion occurred - the result of
which was that the analysis of religious language came to be regarded as basic and
fundamental to the myriad of other issues and problems. In this view, approaching
questions about claims to religious knowledge philosophically begins with an
analysis of the language that is used to affIrm, describe, or communicate those
claims, and analyzing the religious language within which those claims are
couched invariably leads to epistemological questions.
The view that understanding and analyzing religious language is basic and
fundamental to the philosophy of religion leads directly to the need for resolving
the question of the logical status of religious language, and attempting to resolve
the question of the logical status of religious language requires pursuing several
different avenues of inquiry that, at different times, parallel and intersect each
other. Is religious language assertive? Are religious beliefs "cognitively
meaningful"? If so, are the statements of these beliefs "literally true or false" or
"symbolically or metaphorically true or false"? What is the criterion (or criteria)
for determining cognitive meaningfulness? Can language have epistemic content
and not be cognitively meaningful? Are there other kinds of meaningfulness
besides cognitive meaningfulness? If so, what about the criteria for "noncognitive
meaningfulness"? But what about these questions themselves? Do they presuppose
certain rigid ways of categorizing possible approaches to religious language? In
some Wittgensteinian-based views, these very distinctions between cognitive and
noncognitive meaningfulness and between literal and metaphorical truth and falsity
are regarded as arbitrary and restrictive. In pursuing answers to these questions, as
well as the others addressed in this chapter, it is a mistake to talk about and treat
religious language monolithically. From the most radical to the most conservative,
those who have dealt with problems of religious language have admitted that
different utterances are intended to be regarded and treated in different ways. Not
all utterances are intended as statements. Different utterances by theists may be
exaltations of praise, professions of beliefs, descriptions of religious practices,
reports of religious authority or history, or factual claims about the world.
The problem of religious language in the twentieth century was a complex one
of sorting through the multifarious kinds of religious utterances and establishing
and analyzing the different logical types of religious utterances, the logical status
of the different logical types of religious utterances, and the logical relationship
between the different logical types of religious utterances and other questions in
the philosophy of religion (such as the problem of religious knowledge). For most
who have addressed these issues, the question of whether certain kinds of religious
3 See, for example, William T. Blackstone, The Problem of Religious Knowledge: The Impact of
Contemporary Philosophical Analysis on the Question of Religious Knowledge (Englewood Cliffs, N.J.:
Prentice-Hall, 1963), and Religious Language and the Problem of Religious Knowledge, edited by
Ronald E. Santoni (Bloomington: Indiana University Press, 1968).

30

ANALYTIC PHILOSOPHY OF REUGION

utterances are cognitively meaningful is at the very core of the problem of religious
language. Pursuing the issue of the cognitive meaningfulness of religious language
leads directly into the question of the proper criterion for assessing cognitive
meaningfulness. The earliest and most extreme forms of assessing cognitive
meaningfulness, advocated by the logical positivists, traded upon full or partial
direct verification and rejected all religious statements in toto and "out of hand."
Logical positivism flourished briefly and then faded (see Chapter I). However, the
spirit of positivism and the lingering, fundamental question about the cognitive
meaningfulness of religious statements did not fade away so quickly or easily.
About midcentury, several critics renewed the attack upon religious language in
more moderate guise. So far as the philosophy of religion is concerned, logical
positivism had been a wolf in wolf's clothing; the renewed attack came in the form
of a wolf in sheep's clothing.
POPPER AND FALSIFIABILITY
Since its ftrst introduction by Karl Popper, the notion of falsifiability has gained
and continued to enjoy significant currency in the philosophy of science.
Falsiftability was used by Popper as a criterion and a test for empirical and
scientific claims for settling what he calls "the problem of demarcation." Facing a
situation in which scientific theories were becoming more and more abstract primarily as a result of Einstein's introduction of the theory of relativity and
developments in quantum theory - Popper was faced with dealing with the
methodological signiftcance of scientific theories that did not lend themselves to
verification in any straightforward sense. The problem was one of distinguishing
between legitimate science and "pseudoscience." In contrast with the prevailing
view of the nature of science, which depended upon the notion of conftrmation,
Popper's answer to the problem of demarcation was that, to be empirically
signiftcant, statements of science must be falsifiable in principle. Also, in contrast
with the logical positivists' use of the verification criterion of meaning, Popper did
not develop a general theory of meaning. Through the process of attempting to
falsify a statement, we simply test individual statements "in a negative sense." The
empirical content of a statement is determined by its degree of falsifiability; thus,
the easier it is to falsify a statement (that is, the more evidence that is potentially
falisfying), the greater the degree of empirical content of the statement. 4 A
meaningful, legitimate scientific hypothesis is one on the basis of which a scientist
is able to deduce specific expectations concerning matters of fact. If, in certain,
specifiable circumstances, the predicted results are not obtained, then the
hypothesis is refuted or falsified. Any hypothesis or other claim that is theoretically
immune to any such falsification is not empirically significant, and whatever else it
might be or do, such a statement does not contain any factual information about the
world. Pseudoscientiftc claims such as those of astrology are not empirically
meaningful because they are not falsifiable.

4 For Popper's treatment of the problem of demarcation, see Karl Popper, The Logic of Scientific
Discovery (New York: Basic Books, 1959), pp. 34ff.

THE PROBLEM OF RELIGIOUS LANGUAGE

31

FLEW AND FALSIFICATION


Although the grand program of positivism proved to be a failed enterprise, its
influence was (and continues to be) significant. The positivists could never agree
on a formulation of the verification criterion of meaning that would do what they
wanted it to do;5 however, the notion of verification never completely faded away
and continued as an undercurrent of analytic philosophy. The continued dominance
of the scientific method and the glaring differences between the scientific method
and the many ways of forming and explaining religious belief invited the continued
comparison of the statements of science and religious language. One of the most
compelling variations upon the early positivists' emphasis on verification and
empirical meaning has proven to be the notion of falsification introduced by
Antony Flew. Flew's falsifiability criterion of meaning for religious statements is
an attempt to take Popper's falsifiability test for scientific claims and apply it
within the philosophy of religion. Flew's use of falsification also has proven to be
the most poignant and lasting formulation to capture and preserve the positivists'
original concern with empirical meaningfulness. Flew introduced his notion of
falsification as part of the famed "University Discussion" about religious language
in his article entitled "Theology and Falsification." This has become one of the
most commonly known, reprinted, and anthologized pieces ever produced in
analytic philosophy of religion. 6 Flew begins his discussion of falsification by
relating a parable first used by John Wisdom in his article "Gods."
Once upon a time two explorers came upon a clearing in the jungle. In the
clearing were growing many flowers and many weeds. One explorer says,
"Some gardener must tend this plot." The other disagrees, "There is no
gardener." So they pitch their tents and set a watch. No gardener is ever seen.
"But perhaps he is an invisible gardener." So they set up a barbed-wire fence.
They electrify it. They patrol with bloodhounds. (For they remember how H. G.
Wells's The Invisible Man could be both smelt and touched though he could not
be seen.) But no shrieks ever suggest that some intruder has received a shock.
No movements of the wire ever betray an invisible climber. The bloodhounds
never give cry. Yet still the Believer is not convinced. "But there is a gardener,
invisible, intangible, insensible to electric shocks, a gardener who has no scent
and makes no sound, a gardener who comes secretly to look after the garden
which he loves." At last the Skeptic despairs, "But what remains of your
original assertion? Just how does what you call an invisible, intangible, eternally
elusive gardener differ from an imaginary gardener or even from no gardener at
all?"?

5 See Carl Hempel, "Problems and Changes in the Empiricist Criterion of Meaning," in Semantics
and the Philosophy of Language, edited by Leonard linsky (Urbana: University of illinois Press, 1952).
6 Flew's paper prompted the "Discussion," which took place between Flew, R. M. Hare, and Basil
Mitchell.
7 'Theology and Falsification," in New Essays in Philosophical Theology, edited by Antony Flew
and Alasdair MacIntyre (London: SCM Press, 1955), p. 96. Cited by Flew as originally appearing in the
Proceedings of the Aristotelian Society, 1944-45. Also in Logic and Language, edited by A. G. N. Flew
(Oxford: Basil Blackwell, 1963), pp. 187- 206.

32

ANALYTIC PHILOSOPHY OF REUGION

The application of the parable of the invisible gardener to the philosophy of


religion is immediate and direct by Flew. The theist and nontheist disagree about
the existence of an omnipotent, omniscient, omnibenevolent, transcendent God
who "tends" his creation. What evidence can possibly resolve this dispute? We can
only look to evidence that would be compatible with one assertion but not the
other. In the parable, when the Believer and the Skeptic (the theist and the
nontheist) perform tests to attempt to discover evidence to support one claim or the
other, no evidence to support the theist's claim that there is a God is ever
discovered that is not equally supportive of the nontheist's claim, and no evidence
is regarded as possibly falsifying of the theist's claim. However, the Believer
continually accommodates the potentially falsifying results by qualifying the
original claim that there is a gardener by saying, "Well, there is a gardener, BUT
the gardener is invisible, intangible, etc." We can easily imagine that if more varied
and sophisticated tests were introduced with the same results, the Believer might
say, "Well, there is a gardener, BUT the gardener is completely and necessarily
undetectable by any possible means by human beings!" The original claim then
becomes meaningless, according to Flew, and indistinguishable from the
competing claim that there is no gardener at all. With the turn of what has become
an enduring phrase, Flew concluded that the original claim dies the "death by a
thousand qualifications."s
Flew's point concerning the claims of the theist is forceful and straightforward.
The theist makes such claims as "God has a plan," "God created the world," and
"God loves us as a father loves his children.,,9 If such claims are intended by the
theist as assertions - as propositions with truth value - then in order to assert
something the theist must thereby also be denying something - namely that the
negation of the original proposition is not true. A proposition that does not
preclude some state of affairs from being true, that is, a proposition that is not
falsifiable by some possible state of affairs, is thus cognitively meaningless or
cognitively empty regardless of its declarative syntactical form. According to
Flew's falsifiability criterion of meaning, for any assertion, "anything which would
count against the assertion, or which would induce the speaker to withdraw it and
to admit that it had been mistaken, must be part of (or the whole of) the meaning of
the negation of the assertion. And to know the meaning of the negation of an
assertion, is as near as makes no matter, to know the meaning of the assertion."l0
The conclusion of the application of the falsifiability criterion is obvious: if an
alleged assertion denies nothing, then it asserts nothing. As several critics have
pointed out, the state of affairs that a meaningful statement must deny must be an
empirical state of affairs (it is arguable that a standard use of "state of affairs"
designates only empirical states of affairs or empirical matters of fact).
Flew represents this position as a simple point of logic. In a footnote, he
suggests that his claim is the same, in fact, as acknowledging that to assert P is to
assert the denial of not-P (the law of double negation). However, things are surely
not so simple as this, and when the underlying complexity of Flew's position is
8 Ibid., p. 97.
Ibid.
10 Ibid., p. 98.

THE PROBLEM OF RELIGIOUS LANGUAGE

33

understood, a major difficulty arises. Given the impact that Flew's use of
falsification has had upon the ways in which religious language and religious
knowledge have been approached in the latter half of the twentieth century, this
matter deserves some lengthy and careful attention. Consider the rationale that
Flew provides to support the claim that if an alleged assertion is to assert
something, then it must also deny something:
Now to assert that such and such is the case is necessarily equivalent to denying
that such and such is not the case. Suppose then that we are in doubt as to what
someone who gives vent to an utterance is asserting, or suppose that, more
radically, we are skeptical as to whether he is really asserting anything at all,
one way of trying to understand (or perhaps it will be to expose) his utterance is
to attempt to find what he would regard as counting against, or as being
incompatible with its truth. For if the utterance is indeed an assertion, it will
necessarily be equivalent to a denial of the negation of that assertion. And
anything which would count against the assertion, or which would induce the
speaker to withdraw it and to admit that it had been mistaken, must be part of
(or the whole of) the meaning of the negation of that assertion. And to know the
meaning of the negation of an assertion, is as near as makes no matter, to know
the meaning of that assertion. And if there is nothing which a putative assertion
denies then there is nothing which it asserts either: so it is not really an
assertion. 11
One major criticism against Flew's use of falsification as a criterion of meaning
has been raised by Alvin Plantinga. 12 Plantinga understands Flew to be presenting
an argument for the claim that one way to understand an assertion is to understand
what is incompatible with that assertion. Part of the argument appears to be Flew's
claim that anything that would count against the assertion must be a part of the
meaning of the negation of that assertion. If, in fact, Flew intends this claim to be a
premise in an argument to prove his general claim that in order to assert something
a statement must deny something, then Plantinga has quite rightly raised objections
against this premise of the argument. The claim that anything that would count
against the original assertion must be a part of the meaning of the negation of that
assertion does, indeed, appear to lead to very bizarre consequences. 13 To illustrate
this point, consider the example of the assertion, "Flew exists," which Keith
Yandell uses to make Plantinga's point. 14 Given Flew's position, part of the
meaning of "Flew exists" would be the fact that "Flew was never devoured by a
dinosaur." "Flew was not executed by the Inquisition" would also count against
"Flew exists" and would thus be part of the meaning of the negation of the
negation, "Flew does not exist." Of course, the point is that there appears to be an
infinite number of different things that would count against Flew's existence and
Ibid.
See Alvin Plantinga, God and Other Minds (Ithaca, N.Y.: Cornell University Press, 1967), pp.
157ff.
13 Ibid., pp. 158-59.
14 Keith Yandell, Basic Issues in the Philosophy of Religion (Boston: Allyn and Bacon, Inc., 1971),
p.13.
11

12

34

ANALYTIC PHILOSOPHY OF REUGION

that would thus be part of the meaning of the denial of the negation of the original
assertion. Thus, it seems that we could never completely understand "Flew does
not exist" and that Flew's suggestion, as a general theory of meaning, has the
unhappy consequence that we could never completely understand the meaning of
simple, straightforward claims. 15
However, things are not as difficult for Flew as they initially appear to be.
Yandell himself recognizes that Flew can respond to this criticism by maintaining
that the countless various events, states of affairs, and facts that count against his
existence should be disjoined and not conjoined. 16 This means that a person would
not have to enumerate every single possible disjunct that may count against the
original assertion in order to understand the meaning of the negation. It seems that
a person may simply select whatever particularly relevant and poignant disjuncts
seem best suited for the job. In addition, Plantinga admits that the conclusion of
Flew's supposed argument (that if there is nothing that a putative assertion denies
then there is nothing that it asserts either) is a "truism," so the concerns about
Flew's supposed argument do not seem to be very important. It appears that his
claim treating what would count against the original assertion as "part of the
meaning" of the negation of the original assertion is the result of the lingering
influence of the verification criterion of meaning. 17 Whether such a general theory
of meaning is plausible is separable from simply treating falsifiability as a test for
cognitive meaningfulness. If we treat the above passage from Flew as a rationale or
explanation for the claim that a way of understanding an assertion is by
understanding what would count against the assertion and not as a general theory
of meaning, then our attention is refocused simply upon the issue of falsifiability as
a test for meaningfulness.
Plantinga expresses frustration at understanding exactly what Flew's challenge
to religious language is; 18 however, others - theists and nontheists alike - have
taken up the challenge very readily. The plausibility of using falsification as a
criterion for understanding the meaning of an assertion can be demonstrated by
considering a case from ordinary experience. Suppose that a young child goes to
school, where her teacher sees that she is in ill health, poorly clothed and fed, and
suffering from what seem to be obvious signs of child abuse. When the child is
asked why her father does not provide the proper medical attention and clothes and
food for her and why she has bruises and scars from abuse by the father, the child
responds, "Because my father loves me." Perhaps, one might think, the child is
denied basic necessities because the family is poor, but it is determined that the
family is quite well off. As details emerge of the child's situation and as she is
questioned further about her father's behavior and his motive for his behavior, the
child continues to profess that her father loves her. At some point, if the
investigation continues, we can imagine our response to the child. "You poor
dear," we might say. "You just don't know what 'love' means."

Ibid.
Ibid.
17 This appears to be the case if, in fact, the only thing that can falsify the original assertion is an
observation statement(s). See Plantinga, God and Other Minds, p. 162.
18 Plantinga, God and Other Minds, p. 168.
15

16

THE PROBLEM OF RELIGIOUS LANGUAGE

35

Whatever the meaning of 'love' is, loving someone must mean something
different from not loving that person. Whatever the behavior might be that is
compatible with and evidence of one person loving another, there must be some
behavior that is not compatible with or evidence of that person loving the other;
otherwise, there would be no difference between loving and not loving. If a father
can neglect and abuse his child in any way imaginable but still be said to love her,
then what would constitute not loving her? This is the basic intuition captured by
the notion of falsification.
Of course, the application of this case to the one of a transcendent God who
loves his children is obvious. Why would an omnipotent and omnibenevolent God
allow his children to suffer? Because God loves us? What might it mean to say that
God loves us under such circumstances? And what would be the difference
between God's loving us and not loving us? Falsification provides the mechanism
for understanding such a claim. What might a theist allow as falsifying the claim
that God loves us? Or, if not falsifying, what might a theist allow as possibly
counting against the claim? Are there any conceivable circumstances, we might
ask, in which the theist would withdraw the claim or be cognitively tempted, on the
basis of certain evidence, to withdraw the claim? Suppose that, at some point in the
future, every single human child born was born suffering from some terrible
incurable aliment that caused excruciating pain for every single moment of every
person's life to the extent that nothing we now regard as constituting human
happiness or human fulfillment or human satisfaction or human success were
possible? Would such a circumstance be evidence against the claim that God loves
us? Can we respond to this challenge by qualifying love so that we are talking
about "divine" love? How much can we qualify the ordinary notion of love before
we run the risk of making the claim about divine love so different from claims
about human love that it loses all significance and dies "the death of a thousand
qualifications"? Is divine love so different from human love that it loses all
resemblance to human love? How then can we understand what we mean by divine
love? Why would we then still call it love? The different ways in which one might
respond to these challenges are indicative of the different theoretical responses to
the challenge of the meaningfulness and the nature of religious language.
There have been two general categories of responses to the challenge of the
meaningfulness of religious language based upon the notion of falsifiability. I will
call these two different kinds of responses the liberal response and the conservative
response. Liberal responses to the challenge of falsification admit to some degree
the impact of the challenge upon the empirical significance of religious language
and attempt to develop other accounts of significance for religious language.
Liberal responses thus abandon the claim that religious language is empirically
significant in favor of some other kind of account of how religious language is
significant. Conservative accounts seek to protect religious language as empirically
significant by responding to Flew and the challenge of falsification by providing
some account of how religious language satisfies the condition of falsification,
thereby preserving the factual, empirical content of religious language. 19
19 This classification is inspired by several different sources, including William T. Blackstone, The
Problem of Religious Knowledge (Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1963); William Rowe,

36

ANALYTIC PHILOSOPHY OF REUGION

LIBERAL RESPONSES TO THE CHALLENGE OF FALSIFICATION


Liberal responses to the challenge of falsification take the approach of
defending religious claims by first "giving the devil his due" by admitting that the
criterion of falsification does exactly what Flew claims, that is, the criterion of
falsification demonstrates that many religious claims are not empirically
significant. Since those who defend such a method of response do not wish to
abandon religious claims altogether, the problem then becomes one of explaining
exactly what kind of meaning and what kind of value religious claims do have if
they are not empirically significant. There are many different answers to this
problem, and we can here treat only some of the major figures. 20
R. M. HARE AND BLIKS
R. M. Hare responded to Flew in the same famous "University Discussion" in

which Flew introduced the notion of falsification. In his response to Flew and the
challenge of falsification, Hare admits that on Flew's own ground, he is
"completely victorious." Hare then proposes to defend religious claims and provide
an alternative account of their meaningfulness by "shifting ground." He begins by
relating his own parable:
A certain lunatic is convinced that all dons want to murder him. His friends
introduce him to all the mildest and most respectable dons that they can find,
and after each of them has retired, they say, "You see, he doesn't really want to
murder you; he spoke to you in a most cordial manner; surely you are convinced
now?" But the lunatic replies, "Yes, but that was only his diabolical cunning;
he's really plotting against me the whole time, like the rest of them; I know it I
tell you." However many kindly dons are produced, the reaction is still the
same. 21
The lunatic's claim about dons (perhaps one that many graduate students have
shared) is not an empirical claim, Hare agrees, since it is immune to falsification.
The lunatic will adjust his belief to accommodate any possible counterevidence
that may be presented against it, thereby preventing falsification. So, Hare admits,
the lunatic's claim is not an empirical assertion with a truth value. The lunatic has
what Hare calls a "blik." A blik is a general attitude toward the world or a
significant part of it that provides a basis for how an individual relates to the world.
A blik is not an empirical claim, and it is not held on the basis of evidence or as the
result of any argumentation or demonstration. Nor do bliks provide any
explanation of the world, but they do provide the most general framework within
which reason and explanation are recognized and operate. Without a particular blik
Philosophy of Religion: An Introduction (Belmont, Calif.: Wadsworth Pnblishing Co., 1978); and
Ronald E. Santoni, Religious Language and the Problem of Religious Knowledge.
20 In addition, see 1. J. C. Smart, "Metaphysics, Logic, and Theology," in New Essays in
Philosophical Theology, edited by Antony Flew and Alasdair MacIntyre, pp. 12-27; R. F. Holland,
"Religious Discourse and Theological Discourse," Australasian Journal of Philosophy, Vol. 34, no. 3,
1956; R. B. Braithwaite, An Empiricist's View of the Nature of Religious Belief (Cambridge: Cambridge
University Press, 1955).
21 R. M. Hare, '''Theology and Falsification," in New Essays in Philosophical Theology, edited by
Antony Flew and Alasdair MacIntyre, pp. 99-100.

THE PROBLEM OF REUGIOUS LANGUAGE

37

in place, a person would not share a reasonable person's concern with evidence and
justification. 22
Everyone has bliks about the world, according to Hare. We all believe that the
steering wheel of our car will continue to tum the wheels properly and allow us to
control the car, though we may not have the knowledge of the engineering of the
mechanisms or the tolerances of the metals and other materials involved to justify
such a belief. People similarly believe that the road will continue to support the car
properly and not suddenly open up to swallow it into some great abyss.23 Most
people believe that airplanes are safe while others do not. And, many times, those
who do not believe that airplanes are safe, people who have a "fear of flying," may
well know more about aerodynamics than those who do not have such a fear. Such
an attitude or general belief seems to be, in some sense, immune to the effect of
evidence and reason.
According to Hare, people care about the subject matter of their bliks. We do
not approach the world in the cool, detached way in which the explorers in
Wisdom's original parable seem to approach the question about the existence of
the gardener. 24 The lunatic is passionate about dons, and whether one's car
continues to operate properly is a matter of great personal concern to us. Anyone
with a fear of flying who has ever had to fly in an airplane or anyone who has ever
had to fly with a family member or friend with a fear of flying knows for certain
how passionately such people care about the matter of flying! For Hare, the same is
true of the religious person. A person who has a religious blik about the world uses
that blik to structure the world and organize the things in it about which he cares
deeply in a particular way. Although Hare does not elaborate a great deal about the
way in which religious bliks operate, it is easy to see a strong similarity with
William James's notion of the pragmatic value of religious belief and the effects of
the "will to believe." Religious people "see" the world differently - "address" the
world differently - from nonreligious people. And the different bliks that people
may have about dons make everything they do and say regarding dons dependent
upon those different bliks.
There is little doubt that, for Hare, some bliks are correct and some are
incorrect, and this distinction is the target of the most serious criticism of his
notion of bliks. A comparison of bliks with Hume's natural beliefs is illuminating.
Hume, at least, had a rationale for distinguishing natural beliefs and treating them
as special cases in his general epistemology. Natural beliefs are those that human
beings cannot help holding and without which they cannot survive in the natural
world. Belief in induction is one such belief. Belief in necessary connection is
another. Such natural beliefs might certainly be theoretically open to falsification
for Hume. We cannot properly claim to know natural beliefs either a priori or a
posteriori, but this does not mean that evidence or reason are irrelevant to such
beliefs. It is simply that the evidence never reaches the level where such beliefs are
justified by the evidence. We can certainly imagine a situation in which, if the
world changed dramatically enough, or if we were transported to a different distant
Ibid., p. 101.
Ibid., pp. 100-101.
24 Ibid., p. 103.
22
23

38

ANALYTIC PHILOSOPHY OF RELIGION

world where the regUlarity and causal connections in this world were lacking, our
natural beliefs may well change. The situation with bliks is supposedly entirely
different. Bliks are supposed to be completely separated from the facts and
completely immune to revision forced by the facts. How then are we to tell a good
blik from a bad one and distinguish a correct blik from an incorrect one? In the
case of the lunatic, Hare says that his blik is "insane" and "wrong" while ours is
"sane" and "right.,,25 But surely something is seriously amiss here. It would only be
because of our knowledge of the facts of the matter, that is, only because of the
content of the blik, and the relation of a particular blik to other information that we
have about the world, that we would come to be in a position to judge one blik
"sane" and the other "insane." The distinction between "sane" and "insane" and
between "right" and "wrong" seems to be simply a thinly veiled guise for the
distinction between true and false. The reason that "we" regard the blik of the
lunatic as insane is because we regard the facts as we know them to be falsifying of
the lunatic's blik. Hare's notion of a blik may well serve to help us better focus
upon some really thorny issues about the nature of evidence and epistemological
warrant and how different people may weigh and judge such evidence and warrant
differently, but bliks do not provide a safe haven for religious beliefs in the face of
the challenge of falsification.
IAN T. RAMSEY AND MODELS AND QUALIFIERS
Ian T. Ramsey joins the ranks of the liberal respondents to the challenge of
falsification by insisting that religious situations are unique and require their own
language and their own logic. Ramsey joins with those who defend religious
language by "shifting ground," by abandoning the claim that religious claims are
straightforward empirical assertions. They do, however, retain an empirical
"grounding."
Religious situations are characterized, according to Ramsey, by odd discernment
and total commitment. "Odd discernment" means that a person recognizes a
particular situation as being highly personal and invoking a realization on the part
of that person that the situation is logically odd and that it cannot be satisfactorily
described or otherwise captured by ordinary language or ordinary logic. Ramsey
illustrates the feature of religious situations that he labels "odd discernment" by the
following story about an English High Court:
Let us recall the setting of a High Court - all very impersonal, all very formal,
quite lacking in "depth" and "vision." The name of the judge is made as suitably
abstract as possible - Mr. Justice Brown. The wigs and scarlet are meant to
conceal the fact that Mr. Justice Brown is after all a human being. If he has
rushed to the Court from riding, or mending the car, he would never think of
appearing there in riding breeches or overalls. Whatever clothes he wore to play
with the children will at least be covered with scarlet and ermine. Nor is the
argument of the Court interested in persons. We have, instead, "the Crown,"
"the accused" and "the prosecution." Here is a situation as impersonal as may
be; a mere facade of human existence. Then, one morning Mr. Justice Brown
25

Ibid., p. 100.

TIIE PROBLEM OF RELIGIOUS LANGUAGE

39

enters the Court to see as the "accused" the closest friend of his undergraduate
days; or, if we may be more melodramatic, the long-lost wife. "Eye meets eye";
astonishment; an odd word is uttered. It may be from the undergraduate friend
"Sammy:"; from the wife "Penney!", and the result is (as the papers will tell us
the next day) that the Court is "electrified." An impersonal situation has "come
alive.,,26
In such situations, Ramsey says, a discernment is made. The ordinary situation
"comes alive," and "the light dawns" or "the ice breaks." Such situations require a
special language to distinguish them from the ordinary, and since religious
situations invoke such odd discernments, they require their own special language
and logic, which will be characteristically different from the language and logic of
science or any straightforward empirical statements. Thus, in the example of the
High Court, the use of personal first names, or nicknames, is appropriate to the
highly personal situation where a continued use of the standard or normal
descriptions of "the Judge" or "the accused" would not be.
Secondly, religious situations are characterized by total commitment. An
example of such commitment might be found in a person's commitment to a
school, a college, or a nation.27 However, religious commitment is total
commitment, which means that the commitment involves all of a person's
personality, that is, the commitment is held toward every aspect of a person's life
and not just some one person or one activity; it is a commitment "suited to the job
of living.,,28
Given that religious experience is distinct from "ordinary" experience by
invoking odd discernment and total commitment, Ramsey insists that we must use
a suitable language and a suitable logic for such situations. Ordinary language and
ordinary logic will not do. In Religious Language, Ramsey suggests several
different ways in which religious language needs to be appropriately different and
logically distinct from ordinary language to do its special job. As representative of
how religious language is supposed to function differently from ordinary language,
let us consider his treatment of models and qualijiers. 29
The use of models and qualifiers explains the subtitle of Ramsey's Religious
Language - "an empirical placing of theological phrases." We begin by selecting
appropriate models based upon our ordinary experience for which we normally use
ordinary language with empirical content and truth values. Such models might
include "cause," "wise," and "good." We use such words frequently in ordinary
and familiar situations; however, in religious uses, the situations are not ordinary
and familiar. Although we might begin by wanting to use such models to capture
the situation, our ordinary language, unless suitably qualified, is unequal to the
task. To capture "logically odd" situations, for example, talking about the nature of
God or describing one's discernment of an experience of the divine, we must
qualify the original model to make the model suitably removed from its ordinary
26 Ian T. Ramsey, Religious Language (New York: Macmillan, 1957), pp. 20-21. Also see Ramsey's
"Talking about God," in Myth and Symbol, edited by F. W. Dillistone (London: SPCK, 1966).
27 Ibid., p. 34.
28 Ibid., p. 39.
29 See ibid., pp. 69ff.

40

ANALYTIC PHILOSOPHY OF REUGION

meaning and its ordinary application. To 'cause,' for example, we add the qualifier
'first' to get 'first cause.' A First Cause is so characteristically different from cause
that though we start with the empirical grounding or placing of the notion of cause,
we are led by the use of the qualifier to a new phrase that becomes "proper
currency" for the religious situation. People, books, movies, and wines may all be
good. But if we add the qualifier 'infinitely' to 'good' to get 'infinitely good,' then,
according to Ramsey, we wind up with a logically odd phrase that is "empirically
anchored" in its ordinary meaning but that is suited to talk about God because of
the work of the appropriate qualifier. 'Unity,' 'simplicity,' and 'perfection' are
qualifiers that serve a similar purpose. As in the example involving the High Court,
we begin with an ordinary situation and then qualify it until it becomes "odd."
When we add the condition of a total commitment, we have what Ramsey regards
as a religious situation.
Ramsey later insists that religious language must have "objective reference" by
appealing to empirical facts, to "what is the case.,,30 Objective reference is
determined by one's sense of being "confronted" and "acted upon.,,3! The objective
reference is thus given in the moment of discernment, and we then invoke the use
of models and qualifiers in order to try and construct ways of talking about the
objective reference of our experience.
RELIGIOUS LANGUAGE AND RELIGIOUS COMMUNITIES
The models are supposed to anchor religious language empirically to allow us to
talk meaningfully about the "objective reference" of our discernment experience,
but how are we to select the appropriate models and qualifiers to capture the
special features of the religious situation? Ramsey says that "any word may be a
'model' by means of which a characteristically religious situation can be evoked.
All words, if suitably qualified, can lead to such a situation.,,32 However, it does
not seem that the mechanism of models and qualifiers can adapt to the use of just
any word as a model. 33 Ramsey has chosen words to serve as examples of models
that are usually associated, by religious tradition, with talk of God, and it appears
that one would be dependent upon such traditions for choosing and understanding
the role of the models and the effect of the qualifiers. Ramsey's examples of
models, 'cause,' 'wise,' and 'good,' are steeped in theistic religious tradition,
practice, and church history. If we try to apply the method of models and qualifiers
to just any word, the importance of our selection of the appropriate words to serve
as models becomes clear. What about 'mud,' 'slime,' 'dung,' 'perissodactyl,' or
'forcipulata,?34 It is not clear that one could derive any religious currency from
'first dung,' 'eternal mud,' or 'infinite perissodactyl.' These phrases are certainly
"logically odd," but nothing seems to qualify them for capturing anything that is
uniquely religious in nature. We may get some currency out of 'primordial slime,'
but that would only be because of some preexisting usage. This illustrates an
Ian T. Ramsey, Models for Divine Activity (London: SCM, 1973), p. 58.
Ibid., p.61.
32 Ian T. Ramsey, Religious Language, p. 91.
33 See James F. Harris, "Models and Qualifiers," International Journal for Philosophy of Religion,
Vol. 3, no. 2, summer 1972, pp. 87ff.
34 This point and these examples follow ibid., pp. 88ff.
30
31

THE PROBLEM OF RELIGIOUS LANGUAGE

41

important point. In the example of the use of nicknames in the High Court,
ordinary experience provides the backdrop for the use of nicknames that makes
their use meaningful. There may well be something odd about the use of such
informal nicknames in a formal, courtroom setting; however, there is nothing
logically odd about the language itself. In the case of the use of models and
qualifiers to talk about God, ordinary usage does not provide us with the backdrop
to understand either the selection of the models or the effect of the qualification
that takes place. We need to first understand the religious tradition within which
certain models and qualifiers are used in order to make the proper selection.
Janet Soskice has made a similar point in a much more elaborate fashion.
Ramsey relies upon experience to provide the objective referent for the different
expressions in terms; however, no experience will justify his selection of the
models and qualifiers that he uses to capture those moments of discernment.
Soskice summarizes the point as follows:
Ramsey can elaborate his account in terms of models only by dispensing with
his empiricism, yet he can ground the disclosures in which the models
theoretically are based only by retaining it. Many of the inconsistencies of his
programme arise from this desire to incorporate metaphysical claims in an
empiricist framework. 35
According to Soskice, the only way to preserve a cognitive element for models
in religious language is by retaining some form of religious realism. The kind of
critical realism which Soskice advocates, which will allow the theist to use models
(and other nonliteral modes of speech) to talk about God, is characterized by
"experience, community and an interpretative tradition.,,36 Each religious believer
is a part of a religious community, some of whom have had religious experiences
to "ground" the references that are then made by other members of the community
within the same interpretative context. Since we are all members of different
linguistic communities - religious linguistic communities and scientific linguistic
communities alike - with different interpretative contexts, we are able to use
language to refer, not because of our own experience with the referent of whatever
referring expression we use, but because others in our linguistic community with
the same interpretative context have had experiences that "ground" the expression
with a referent and have used the same referring expression. 37 Experience of an
individual is still crucial for fixing the referent of a referring expression; however,
the real crux of Soskice's theory is social. Experience must always be interpreted
by "the categories used by a particular community of interest and within a
particular tradition of evaluation. ,,38
There is much to be said about Soskice's notions about reference. At times she
seems to be committing to a minimalist version of descriptivism, and at other
times, she seems to prefer a causal, historical view of reference. The causal theory
Janet Martin Soskice. Metaphor and Religious Language (Oxford: Clarendon Press, 1985), p. 147.
Ibid., p. 147.
37 Ibid., p. 149.
38 Ibid., p. 151. Notice that though Soskice's theory is social, her commitment to realism separates
her from the Wittgensteinians and the use of language-games.
35

36

42

ANALYTIC PHILOSOPHY OF REUGION

has some rather serious deficiencies when used within the context of religious
language. It does not seem that reference can be accounted for simply by the long
tradition of a linguistic community and an interpretative context. There is a real
tension between this claim and the realism which Soskice wants to preserve. As Ian
Barbour has pointed out, astrologers have been members of a linguistic community
with an interpretative context at least as long as theologians have, but most people
would wish to deny that the expressions that astrologers use actually refer to some
real causal connections between the positions of the planets and the stars and
human events. 39 On the causal theory of reference, when we now use a referring
expression (a proper name or a description) that someone long ago used to refer to
an individual in an "initial baptism," we have no way of identifying the referent or
even knowing whether or not the referent was an actual individual when the
referent was "fixed." To use a common example, if we account for the reference of
"God" in terms of a linguistic community within which someone long ago fixed
the reference for "God," we have no way of telling whether the referent was an
actual individual, some internalized, Freudian father figure, or some otherwise
nonexistent individual. Problems with the use of the causal theory of reference in
connection with the problem of reference in religious language are discussed at
length later in this chapter.
THE COGNITIVE STATUS OF MODELS
The need to have additional information in order to select the proper models and
the proper modes of qualification underscores an important point about the
cognitive status of models and qualifiers. The use of models and qualifiers in
religious language cannot perform a primary epistemological function, that is,
fundamental knowledge of God cannot be contained in models and qualifiers. As
we have seen, one must first know a great deal about the nature of the deity, which
is then captured by the selection of the appropriate models - a selection that is
determined (or guided) by religious tradition.
The epistemologically derivative status of models and qualifiers can further be
seen if we investigate what it means for a model to be what Ramsey calls "suitably
qualified.,,4o Not just any model can serve the purPose of capturing the religious
situation, and not just any qualification will serve that purpose either. Saying that
God is a "fake cause," for example, is not likely to capture successfully a situation
where "the light dawns" or "the ice breaks" in some characteristically religious
way. Similarly, describing God as "allegedly wise" or "formerly good" or
"fictional creator" does not develop the original models in an appropriate manner
to talk about God. We know this because of the ways in which God is described
within a particular religious tradition. Therefore, whatever value the use of models
and qualifiers might have, it seems to be subjective or social and dependent upon a
particular person's particular religious beliefs and traditions and prior information
that the person has that provides the basis for the selection of models and
qualifiers. Soskice suggests that "linguistic community and tradition" provide the
39 Ian Barbour, Religion in an Age of Science: The Gifford Lectures, 1989-91 (San Francisco, Calif.:
Harper and Row, 1990), p. 46.
40 Ramsey, Religious Language, p. 89.

mE PROBLEM OF RELIGIOUS LANGUAGE

43

context within which the cognitive value of religious language becomes possible.
Certain uses of language are initially suggested by different people in the
community, and some of these uses survive the test of time and become ingrained
while others do not. Thus, some uses of language become especially adequate for
religious experience while others do not. 41 However, even the survival of centuries
of use within a linguistic community does not preserve and guarantee the realism
which Soskice wishes to preserve. By understanding the context within which
religious language is used, we can certainly say that we understand better how the
members of that linguistic community intend to use the language, but this does
nothing to tie the language to reality as the example of astrology above
demonstrates.
ANALOGICAL AND METAPHORICAL LANGUAGE
By focusing attention upon "the logically odd" situations that are involved in the
use of religious language, Ramsey invites a careful and general evaluation of the
ability of human language to describe "religiously odd" situations in which
language is used to capture some aspect of the transcendent nature of God or some
other religious or mystical situation that is "beyond" normal human experience.
Several liberal responses to the challenge of falsification have adopted, in one form
or other, the Thomistic use of analogical language to talk about God since the
straightforward, literal uses of language seem inadequate for such purposes. While
associated historically with the period of scholasticism and since then, for the most
part, with Thomism, the use of analogical and metaphorical language found its way
into the mainstream of Protestant natural theology primarily as a result of attempts
to respond to the challenge to the meaningfulness of religious language. The
earliest, most thorough and influential accounts of analogical language during this
period and the most explicit adaptation of Thomistic thought for Protestant use are
found in the works of both A. M. Farrer and E. L. Mascall. 42 Mascall will provide
the basis of the treatment here. A complete treatment of the doctrine of analogy
with all of its theological ramifications is certainly beyond the scope of the present
investigation; so, focus here will be upon the very general use of analogy and its
implications for religious language.
Mascall explicitly laments the fact that the doctrine of analogy has found so
little use among Anglicans since he thinks that the use of analogy is fundamental
not just for natural theology but for religion in general. In order to think of or
describe the infinite in terms of the finite, human language must be qualified either
implicitly or explicitly. He claims that it is only through the use of analogy that
natural theology can avoid the extremes of a completely necessary and
unknowable, transcendent God or a completely immanent God which is
contingent. 43
Mascall's use of analogy consists in a straightforward adoption of Thomas
Aquinas's doctrine of analogy. While we can arrive at the belief in the existence of
Soskice, Metaphor and Religious Language, p. 153.
See primarily A. M. Farrer, Finite and Infinite (London, Dacre Press, 1959), and E. L. Mascall,
Existence and Analogy (London: Longmans, Green and Co., 1949). Republished by Archon Books,
1967.
43 Mascall, ibid., pp. 92-94.
41

42

44

ANALYTIC PHILOSOPHY OF REUGION

God without the use of analogy, according to Mascall, we can think of the divine
nature and attempt to describe it only through the combination of the notions of
analogy of proper proportionality and analogy of attribution. 44 Starting, as we do,
from a contingent, finite world, it is only through the use of analogy that we can
hope to capture anything meaningful about the nature of a necessary, infinite God.
Through the use of analogy of attribution, we are able to affirm that a particular
property that is found "relatively and derivatively" in finite beings belongs to God.
Properties belong to God "formerly and properly" because, since all finite creatures
are related to God as effects to a cause, then God must possess a particular attribute
in whatever way is necessary for God to produce finite creatures with that attribute.
For example, when we say that God and man are both good, since man is an effect
and God the cause, God must possess goodness in whatever capacity is necessary
in order to cause goodness in man. 45
The notion of analogy of (proper) proportionality is the result of the theological
doctrine according to which all beings - finite and infinite - possess both different
kinds of qualities and different degrees of those qualities, which are in "proper
proportionality" to their "mode of being." Borrowing vaguely upon Plotinus's
notion of a hierarchy of beings emanating from God, the notion of proper
proportionality maintains that the different beings posses different qualities with
degrees that are appropriate to their position in the hierarchy. God is perfect and
infinite so God will possess any qualities that he possesses with perfection and
infinitude. Man is imperfect and finite so man will possess any qualities that he
possess with imperfection and finitude. Using an analogy of proportionality, we
can affirm that different kinds of beings "formally" possess the same quality but in
different "modes" that are respective of their different "modes of being." Using an
analogy of proportionality, we might say of cabbages, elephants, man, and God
that they possess life in the following manner: 46
life of cabbage

life of elephant

life of man

life of God

essence of cabbage

essence of elephant

essence of man

essence of God

We would read or understand the above symbolic representation as follows: the


way in which the life of a cabbage is determined by the essence of cabbagehood as
proper to cabbages is analogous to the way in which the life of an elephant is
determined by the essence of elephanthood as proper to elephants is analogous to
the way in which the life of a man is determined by manhood as proper to men is
analogous to the way in which the life of God is determined by Godhood as proper
to God. Through this use of analogy, we can meaningfully attribute to God the
perfection of the quality that is possessed by creatures that he created. This in no
way means that finite man is able to understand anything about the essence of God.
Questions about the relationship between analogy of attribution and analogy of
proportionality and about their relative status in relation to one another have been
Ibid., 95.
Ibid., p. 102.
46 These examples come from MascalI, ibid., pp. 104ff.
44

45

TIIE PROBLEM OF RELIGIOUS LANGUAGE

45

of great concern to both the scholastics and to more recent Thomistic scholars. 47
The very lengthy and detailed disputes about the different kinds of analogy lead
into equally lengthy and detailed disputes about theology and just what it is that
man is able to know or affirm of a transcendent God. Taking his departure from E.
Gilson's commentary on Aquinas, Mascall attempts to "cut through" the tangled
theological knot that surrounds the use of analogy by emphasizing the fact that
God's essence and existence cannot be separated. Since all of our statements
affirming some quality of God also have existential import in a way in which none
of our affIrmations of qualities of fInite beings do, then analogical predication of
God is possible since it always involves (at least implicitly) claims about God's
existence. In other words, in order to protect the essence of God from being known
by fInite man, Mascall understands the different uses of analogy as always
involving claims about God's existence and not God's essence. 48 Thus, we are to
understand the above analogy of proportionality as demonstrating an analogous
relationship among different beings "in the existential order," that is, in the way in
which living things exist, and not simply a relationship among different concepts of
living things. It is the analogy of attribution that holds the two sides of the analogy
of proportionality together by affirming the dependence of the existence of fInite
beings as effects upon God as their cause. 49 Mascall uses the doctrine of analogy
not just to respond to what he sees as an artifIcial challenge to the meaningfulness
of religious language by logical positivism but as a way of explaining how
religious people have been able to talk intelligently about God through the ages. 50
Mascall's use of analogy is important historically perhaps as an example of the
introduction of Thomistic doctrine into Anglican thought in the twentieth century.
However, in terms of the fundamental issue about how human language can be
used to talk about God, it is important primarily for the problem that, for all of its
complexity and formal nature, it leaves unanswered or even unaddressed. Analogy
of proportionality is "tied together" by analogy of attribution. Analogical
predication supposedly allows us to talk meaningfully about "the existential mode"
of God, but the basic assumption of analogy of attribution is the supposed
relationship between man and other fInite, contingent creatures, who are the effects
of an infInite, necessary God, and God in the existential mode. In other words, in
order for analogical predication to work, we must assume or affIrm that the fInite
world was created by an infInite God. Once this fundamental claim is in place, we
can then use analogy to elaborate upon the ways in which God is related to features
of this fInite world. Such a use of analogy functions then within a religious or
theological context; it does not function to establish the fundamental claims that
constitute the religious or theological context.
THE EPISTEMIC STATUS OF ANALOGICAL LANGUAGE
Various critics have maintained that some sort of literal meaning or literal
knowledge is necessary for the use of analogy, but the reasons for the necessity of

47

For a brief summary of some of the disputes and the references see Mascall, ibid., pp. 113-15.

Ibid., pp. 117-19.


49 Ibid., p. 120.
50 Ibid., p. 121.
48

46

ANALYTIC PHILOSOPHY OF REUGION

the literal statements is not always clear in these criticisms. 51 For example, Thomas
McPherson insists that in order to recognize analogical uses of language we must
understand some of the terms in the analogies in their literal senses and we must
understand some literal claims about whatever the analogy is used to describe. 52
But McPherson does not make clear the reason for making such claims, and the
problem for him seems to be one of using literal uses of language to recognize
analogical uses of language. To understand how analogical and literal uses of
language might be related, it is helpful to examine how analogical uses of language
presuppose and are based upon literal uses of language. Consider the analogy of
proper proportionality, "God is to man as a shepherd is to his sheep." How does
this analogy allow a person to better understand the nature of the deity and the
deity's relationship to human beings? Let us concern ourselves only with
understanding the meaning of this analogy. The first thing that we should notice is
that in the use of any analogy there will be intended similarities and dissimilarities
between the two pairs of terms. If we begin with what we understand about the
meaning of the relationship between the pair of terms "shepherd/sheep" and then
extrapolate from that to the relationship between the pair of terms "God/man,"
what meaning do we get? Well, the answer is it all depends upon what features we
choose to focus upon in the relationship between shepherds and sheep. Leading
sheep into pastures to eat grass, shearing sheep of their wool, and raising sheep for
food for the dinner table are intended dissimilarities of the analogy and are
extraneous and irrelevant, whereas protecting sheep and being concerned about
their safety and welfare are intended similarities and are highly pertinent and
relevant. 53 In order to understand the intended meaning of the analogy, one must be
able to sort through and distinguish between the intended similarities and
dissimilarities of the analogy. In order to do this, one must already know the
intended meaning of the analogy or assume that some statements about the
supposed relationship between God and man are intended to be literally true;
otherwise, one would not be in a position to distinguish between intended
similarities and intended dissimilarities. And one cannot "read this off' from the
analogy itself. The analogy itself does not give us any clue on how to interpret it
and how to distinguish among the possible different ways of understanding the
similarities and dissimilarities between the two pairs ofterms. 54
There is also the problem of choosing an appropriate analogy. In David Hume's
Dialogues Concerning Natural Religion, Cleanthes suggests that an appropriate
analogy for capturing the relationship between God and the world is the
relationship between human beings and products of man's intelligence, such as a
51 See, for example, William T. Blackstone, "The Logical Status of God-Talk," in Religious
Language and Knowledge, edited by Robert H. Ayers and William T. Blackstone (Atbens: University of
Georgia Press, 1972), pp. 8-9, and Blackstone, The Problem of Religious Knowledge, pp. 66-67. Also
see William P. Alston, ''Tillich's Conception of Religious Symbol," in Religious Experience and Truth,
edited by Sidney Hook (New York: New York University Press, 1961), p. 17.
52 Thomas McPherson, "Assertion and Analogy," Proceedings of the Aristotelian Society, Vol. 60,
1959-60, p. 164.
53 James F. Harris, ''The Epistemic Status of Analogical Language," International Journal for
Philosophy of Religion, Vol. 1, no. 4, winter 1970, p. 213.
54 See Jerry A. Fodor and Jerrold J. Katz, The Structure of Language (Englewood Cliffs: PrenticeHall, Inc., 1964), p. 13.

THE PROBLEM OF RELIGIOUS LANGUAGE

47

house or a ship. But what makes these particular analogies appropriate to use while
others are not appropriate? In Part VI, Cleanthes protests when Philo suggests that
the analogies between an animal and its offspring or between a seed and its plant
might be more appropriate analogies. What makes one analogy appropriate and
another inappropriate is the compelling weight (on balance) of the similarities
versus the dissimilarities of the analogies. It would be difficult, for example, for
most theists to use the relationships between the pairs of terms "Nazi/ Jew" or
"slave-owner/slave" as the basis for forming appropriate analogies for explaining
the relationship between God and man. What makes an analogy appropriate or
inappropriate between different pairs of terms is not just the number of
dissimilarities versus the similarities. It is the "cumulative weight" of the
dissimilarities that makes a pair of terms unsuitable for use in an appropriate
analogy. How does one know this before selecting the terms for the analogy? It is
here that the necessity of some literally meaningful uses of language becomes
apparent. In order to select the terms for an appropriate analogy, one must first
understand something of the literal basis of the relationship that is supposed to be
captured by the use of the analogy.
ANALOGY, METAPHOR, AND RELIGIOUS PLURALISM
Interestingly, the analogical use of religious language has been seized upon by
several contemporary writers and used in the debate concerning the pluralism of
religious beliefs. The point is that the more literal understanding one has of
religious language, the more narrow and exclusive the meaning of that language is.
The increased flexibility that results from the use of analogical language and the
increased role of interpretation in the use of such language "opens things up" to
what is supposed to be a more general and inclusive understanding of the meaning
of religious language. For example, Sallie McFague has cast an entirely different
slant on the question of the meaningfulness of religious language by claiming that
religious language based upon biblical language is meaningless and irrelevant to
most people today since its social and cultural presumptions and settings are so
different from our own. 55 The meaningful use of religious language in what
McFague calls "the interpretative context" is the result of a "wide range of factors"
such as the social, cultural, and historical conditions in which the religious
believers find themselves. 56 McFague claims that while orthodox religion has
narrowly circumscribed the limits of the interpretative context in the past, the more
recent emphasis upon the pluralistic aspects of religious beliefs has resulted in the
realization that we must broaden the interpretative context to recognize the variety
of different social and cultural perspectives that influence the different ways in
which different people use religious language.
According to McFague, the additional flexibility and openness of analogical and
metaphorical language makes such language more suitable for recognizing and
allowing room for different interpretations within Christianity (and presumably
other religions as well). She says, "It is not our time and place in history that
The point is made specifically abont Judaism and Christianity.
Sallie McFague, Metaphorical Theology (Philadelphia: Fortress Press, 1982), p. 3. Also see David
Tracy, The Analogical Imagination: Christian Theology and the Culture of Pluralism (New York:
Crossroad Publishing Co., 1981).
55
56

48

ANALYTIC PHILOSOPHY OF REDGION

influences our religious language, but also our class, race, and sex;
education, and family background; our interests, prejudices, and
order to accommodate the various social, cultural, political,
perspectives, we need to move, McFague claims, toward
"metaphorical theology" whose purpose is

our nationality,
concerns.,,57 In
and historical
a completely

to understand the centrality of models in religion and the particular models in


the Christian tradition; to criticize literalized, exclusive models; to chart the
relationships among metaphors, models, and concepts; and to investigate
possibilities for transformative, revolutionary models .... What must be done in a
metaphorical theology is to open up the relationships among metaphor, model,
and concept for the purpose both of justifying dominant founding metaphors as
true but not literal and of discovering other appropriate dominant metaphors
which for cultural, political, and social reasons have been suppressed.58
While issues involving the question of religious pluralism are addressed in detail
in Chapter IX, various issues involving pluralism impact directly upon McFague
claims concerning metaphorical and analogical religious language. If all of
religious language is metaphorical, then one must ask where the specifically
religious and specifically Christian content is to be found. As we have seen above,
some factual content and some literal meaning is necessary in order for a religious
believer to choose the appropriate metaphors or other nonliteral uses of language.
A commitment to some core "factual" claims allows one to choose a variety of
different analogies or metaphors that may "speak" more meaningfully to different
people at different times, in different social and political contexts, and in different
cultures. However, this is clearly not McFague's intent. Absent any recognition or
commitment to some core "factual" claims, the metaphorical theology that
McFague suggests is indistinguishable from a collection of fairy tales. Now
different fairy tales may well have a significant amount of meaning to different
peoples in different social, political, and cultural contexts; however, there
obviously need be nothing uniquely Christian or even religious about them.
Models, analogies, and metaphors can certainly be used to tell interesting,
illuminating, poignant, and entertaining stories, but there must be something about
such stories that makes them religious or theological. If a core of factual claims
with literal meaningfulness is abandoned, then it seems analogical uses of religious
language force theology in the direction of fiction. 59
MEANING AS USE
Other liberal responses to the challenge to religious language by falsification
make use of Wittgenstein's notion of meaning as use. In his later work,
Wittgenstein abandoned the essentialism of the Tractatus, according to which the
meaning of a word must provide a single, complete, essential set of necessary and
sufficient conditions for the word. In his Philosophical Investigations, Wittgenstein
57
58

Sallie McFague, ibid.

Ibid., p. 28.

59 There are many issues here about original meaning and interpretation that are discussed at length
in Chapter IX.

THE PROBLEM OF RELIGIOUS LANGUAGE

49

maintains that we should look instead to see how a word is used by the speakers
who use it. Meaning as use is one of several Wittgensteinian notions - including
particularly the notion of a language-game - that have found their way into
discussions of religious language. I will discuss the notion of a language-game in
great detail later in this chapter and focus here simply on the notion of meaning as
use. In the Tractatus, Wittgenstein had maintained that the world must be a certain
way and that language must have a certain structure in order to describe that world.
In order for us to use language to make true propositions about the world, words
must have single, essential, univocal meanings and propositions must describe or
label univocal states of affairs. When Wittgenstein abandons such essentialism in
the Philosophical Investigations, he adopts instead the view that words might have
different meanings in different contexts when they are used for different purposes
by different language speakers. Wittgenstein's shift to meaning as use sparked
significantly different developments in the philosophy of language and in the
treatments of religious language. Wittgenstein's notion of meaning as use becomes
a dominant influence and is adopted - either implicitly or explicitly - by several
major figures who defend religious language against the challenge of falsification
by "shifting" the meaning of religious language to the way in which religious
believers use that language. Indeed, such treatments have now come to dominate
religious language.
BRAITHWAITE'S "EMPIRICIST" VIEW OF RELIGIOUS LANGUAGE
R. B. Braithwaite was one of the earliest philosophers of religion to apply some

of the changes that occur in the later Wittgenstein to an analysis of religious


language. 6o Although Braithwaite does not explicitly incorporate Wittgenstein's
notion of language-games into his theory of religious language, he does explicitly
adopt Wittgenstein's notion of "meaning as use." Since Braithwaite characterizes
his view as an empiricist's view, we have still another indication of just how much
empiricism changed at midcentury from the strict empiricism of logical positivism.
Joining others who deny that religious language can be empirically meaningful
since it is not verifiable, Braithwaite insists that religious language is still
meaningful in a way that is analogous to the way in which moral claims are
meaningful. Like moral statements, religious statements are declarations of a
person's intentions to behave a certain way and to adopt or follow a certain kind of
life. A Christian believer, for example, might use a religious utterance like "Jesus
lives" or "Jesus is the Son of God" as a way of expressing the intention of living
what the person believes to be a Christian life. Religious assertions take their
meaning not by referring to some state of affairs in the world but by expressing the
conative intentions of the speaker. Thus, Braithwaite says, "the intention of a
Christian to follow a Christian way of life ... is the criterion for the meaningfulness
of his assertions.,,61
Religious beliefs and moral beliefs are both lacking in empirical significance.
The meaning of both moral language and religious language is to be found in the
60 R. B. Braithwaite, An Empiricist's View of the Nature of Religious Belief (Cambridge: Cambridge
University Press, 1955).
61 Ibid., p. 16.

50

ANALYTIC PHILOSOPHY OF RELIGION

way in which people use such language. Braithwaite argues that people use such
language to express their intentions of committing themselves to a certain "way of
life" or "policy of behavior.,,62 Religious people use religious language to commit
themselves to a certain way of life and to declare their intentions of following a
certain set of religious teachings and practices. This use of religious expressions by
religious believers is the criterion by which the meaningfulness of such expressions
is judged. Thus, religious expressions are not assertions with truth values that are
determined in any way similar to the way in which the assertions of scientists are
verified. When religious people use religious language in ways that appear to make
what are explicit assertions - for example, in telling the creation story or in telling
the story of Adam and Eve in the Garden of Eden - Braithwaite maintains that the
meaning of the stories is psychological instead of empirical. Regardless of the
original intent, such stories (or myths or allegories) are not empirically meaningful,
but they simply make it psychologically easier for the religious person to commit
to and to maintain a particular course of action or way of life. 63 Regardless of
whether religious believers actually believe such stories to be factually true or not,
their only real importance, for Braithwaite, is to provide something of a
psychological aid to the religious person's resolution to act in a certain way.
Braithwaite thus eliminates the cognitive content of religious utterances
altogether in favor of their conative meaning. It is easy to identify a major
difficulty with this conative view of religious language. In order to "save" religious
utterances as meaningful, Braithwaite shifts ground so radically that one might
genuinely wonder what has been saved. Consider the importance of the story from
the New Testament when Jesus asked Peter, "Whom do you say that I am?" Peter's
response is, "Thou art the Christ, the Son of the Living God." Jesus says that he
will take Peter's confession as the "rock" upon which to build his church, and
centuries of Christian believers since then have taken that confession as the
foundation, the "rock" upon which the Christian faith is built. If we understand this
story as Braithwaite suggests, then Peter's confession was simply a psychological
aid to religious believers in their resolution or declaration to live a good life based
upon the teachings of Jesus. If this is true, then the foundations of both the church
and the Christian faith appear to be built upon shifting sands instead of upon a
rock. If there is no cognitive content to Peter's "confession," then it appears that it
could never have the special sense of importance that Christians have attributed to
it. Notice also that, in Braithwaite's view, the content of Peter's confession has no
bearing upon the kind of moral life to which the religious person is committed.
MACINTYRE AND ATTITUDES TOWARD LIFE
Alasdair MacIntyre adopts a tactic similar to that of Braithwaite in his defense
of the meaningfulness of religious language. MacIntyre also maintains that in order
to understand the nature of the meaningfulness and justification of religious
language, we must look to the ways in which people use such language to "accept a
certain way of living.,,64 According to MacIntyre, religious myths function
Ibid.
Ibid., p. 27.
64 Alasdair MacIntyre, "The Logical Status of Religious Belief," in Metaphysical Beliefs, edited by
Alasdair MacIntyre (London: SCM Press, 1957), p. 191.
62

63

THE PROBLEM OF RELIGIOUS LANGUAGE

51

prescriptively to provide a certain attitude toward life and to provide a framework


that will allow a religious person to commit to adopting and living by that attitude.
Religion then, for MacIntyre, does not function the way in which science does,
that is, religion does not provide us with explanatory hypotheses for why the world
exists that lend themselves to verification. In a vein that is reminiscent of Hare's
use of bliks, religious commitment on the part of the theist requires a complete,
"decisive adherence to religious beliefs" completely independent of any evidence
or reason. Religious beliefs are not held on the basis of reason or evidence but
rather on the basis of faith.65
The only ground upon which the theist holds a particular set of religious beliefs
is the "ultimate criterion of religious authority." Every process of reasoning and
justification must end at some point, and at that point, according to MacIntyre, an
appeal must be made to some authority or some fundamental criterion that is not
justified within that process of reasoning. For the religious person, that "ultimate
criterion" is religious authority based upon a person's acceptance of church history,
religious teachings, and sacred scriptures. This invulnerability of religious beliefs
to falsification is a position that MacIntyre continues to maintain, though he
recognizes that it comes with the high price tag of empirical emptiness. 66
WITTGENSTEINIAN LANGUAGE-GAMES
Neither Braithwaite nor MacIntyre mentions the notion of a language-game in
their responses to the challenge of falsification. However, the liberal responses,
which "shift ground" by resorting to the defense of a special sort of religious
meaning or a special sort of logic for religious language, implicitly rely upon a
very similar sort of philosophical insight. As I have indicated earlier, Wittgenstein
abandons the essentialism that characterized the Tractatus in favor of a theory of
meaning as use. The meaning of a word now becomes dependent upon the dynamic
and pragmatic interplay of language and language speakers. Words and sentences
might have many different meanings dependent upon the different uses to which
speakers put them. How many different uses and corresponding meanings of a
word might there be? To answer this question, Wittgenstein invites us to "think of
the tools in a tool-box: there is a hammer, pliers, a saw, a screw-driver, a rule, a
glue-pot, nails and screws. - The functions of words are as diverse as the functions
of these objects.,,67 The difference between Wittgenstein and Austin regarding
different functions of words, noted in Chapter I, is worth repeating here. Austin is
more concerned with specific situations of "what to say when" and with the subtle
differences among different actual things that cause us to call those things by
different names or to describe those things differently. Wittgenstein is more
interested in the fundamental presumptions about the nature of language that allow
for the very possibility of using language in different ways.68 Wittgenstein's
Ibid . pp. 208-09.
See his "Is Understanding Religion Compatible with Believing," in Faith and the Philosophers,
edited by John Hick (London: Macmillan. 1964). and in Rationality, edited by Brian Wilson (Oxford:
Basil Blackwell, 1970), to which the page numbers here refer, p. 75.
67 Ludwig Wittgenstein, Philosophical Investigations, translated by G. E. M. Anscombe (New York:
Macmillan, 1953), Section 11. Also compare Section 12.
6" See fn. 44, p. 21.
65

66

52

ANALYTIC PHILOSOPHY OF RELIGION

"deeper" concerns are with the conceptual underpinning (the "grammar") that
allows us to use language in different ways in different "language-games."
When we imagine the different ways in which tools might be used by a person,
we focus upon the behavior or actions of the people using the tools. Similarly,
Wittgenstein insists that the language, the speakers of the language, and the actions
of the speakers of the language - all "woven" together - constitute a "languagegame.,,69 There are, according to Wittgenstein, "countless" different kinds of
sentences and countless different meanings for words because there are countless
different uses to which language speakers may put words and sentences.
Furthermore, language-games are dynamic and changing, with new ones coming
into existence and old ones dropping from existence continually.7o "Languagegames" then are a sophisticated form of human activity using language, and we
human beings are continually developing multifarious new ways of doing this. For
example, consider how computers and the Internet have created totally new
language-games with new expressions and their accompanying actions that were
unknown and unheard of just a few years ago. By contrast, old language-games
die. For example, consider the uses of language that were associated with the use
of typewriters. There were manual and electric typewriters - a distinction that has
no meaning within the language-game of computers - and there were such
expressions as "carriage-return" - expressions that future generations will not
understand unless they engage in a historical study of the language-games of
typewriters.
For a person to learn the meaning of a word, the person must learn its role in the
language-game. The use of a word or expression in a language-game is always rule
governed, which means, according to Wittgenstein, that the use will be ordered and
patterned by the rules. The rules ensure that there are correct and incorrect and
meaningful and meaningless uses of the word. A word that can be used in any way
whatsoever is not governed by any rule and is meaningless. The analogy of playing
a game carries over very strongly to the activity of "playing" a language-game. For
games to be meaningful, rules must prescribe behavior of players in such a way
that they can act rationally. For example, rules must delineate who the players are
and distinguish them from the nonplayers and indicate what forms of behavior
(moves) are permissible and not permissible within the game, how to score points
(if there are points in the game), and other details about the playing ofthe game. If
any possible activity by any person at all is permissible in the game, then the game
becomes absurd. Meaning then is determined by the rules that govern the use of
words and expressions within a language-game. A knight in chess is a piece that
can be moved in certain ways according to the rules of chess. A piece that looks
like a knight but moves 1-3 or 3-1 instead of 1-2 or 2-1 is not a knight, and any
game in which such moves are made is not what we mean by "chess." Of course,
different language-games have different rules and hence different uses for words
and expressions and hence different meanings for words and expressions. There are
no "justifications" for language-games beyond whatever benefits result from the

69

70

Wittgenstein, Philosophical Investigations, Section 7.


Ibid., Section 23.

THE PROBLEM OF RELIGIOUS LANGUAGE

53

language-game itself. Engaging in the activity of a particular language-game is


simply something that we do - it is part of a "form of life."
THE EDGES OF LANGUAGE AND RELIGIOUS LANGUAGE
Paul van Buren and D. Z. Phillips have both developed treatments of religious
language that are based to a greater or lesser extent upon the later work of
Wittgenstein. Van Buren adapts Wittgenstein's emphasis upon meaning as use and
the conventional force of rules; Phillips constructs a full-blown adaptation of
Wittgenstein's notion of language-games. I discuss van Buren following and
Phillips in Chapter IV.
To understand van Buren's theory of religious language, we must first
understand something about his theory of language. To explain van Buren's idea of
the nature of language, it is useful to introduce a spatial metaphor and think of
language as a web that is com,}?osed of many different threads and sections that are
connected with one another. 1 Within this web, there are small sections within
which there are different uses of language - each with its own rules and its own
meanings. Following Wittgenstein, van Buren emphasizes the importance of the
limitations that exist upon our language. One limitation is set by the conventional
force of the rules that prescribe the ways in which words can be used properly;72
another limitation is the result of the fact that we cannot get "outside" language to
examine it or study it or to form theories about it. Language is like a squash court,
van Buren suggests: there is no "out.,m
Some uses of language are more centrally located within this total web than
others, and some uses of language occupy a larger space with more users than do
others. 74 The uses of language in which we engage most frequently and routinely
with our family and friends and the uses of language that we need in order to
survive and thrive in our environments occupy the central-most positions in the
giant web of human language; thus, we are most comfortable in the center of the
web and most familiar with the rules and uses of language that occur there.
To locate religious language within this "picture"of language, it is helpful to
consider what van Buren takes to be the main purpose of religious language for
religious people:
Religion, at its most serious level, I suggest, has always been marked by what
we call a sense of the mysterious. What I mean is that religious people have
spoken guardedly when they spoke of God. They have said that God is
ineffable; that he is beyond their understanding; that whatever they said of God
was not, strictly speaking, true; or that their words were inadequate to their

71 V an Buren himself suggests the analogy of a platform with planks that extend in different
directions. See Paul van Buren, The Edges afLanguage (New York: Macmillan, 1972), p. 82.
72 Ibid., p. 79.
73 Ibid., p. 82.
74 It would now seem very natural and easy to slip here into the use of Wittgenstein's notion of
language-game; however, for whatever reasons, van Buren failed to make use of this concept. He does
not make the move to regard the different uses of language as separate, autonomous language-games.

54

ANALYTIC PHILOSOPHY OF REUGION

subject .... We associate this way of speaking with the sense of mystery, wonder,
and awe.?5
In order for religious language to function it must be located at the peripheries
of the web of human language, according to van Buren, because religious language
"pushes the envelope" of what we can do with language.?6 With religious language,
we go as far out to the edges of language as we can go without falling off into
nonsense. According to van Buren, a similar thing happens with the other uses of
language that occur at the edge of language - with jokes and puns, the language of
love, poetry, and metaphysics?? These uses of language "out on the edge"
(including religious language) push language beyond its normal, routine, everyday
uses that occur in the large, central area of the web of human language. Life "on
the edge" and language "at the edge" allow human beings to capture and express
an aspect of human nature that is not captured in the "ordinary" language of
science and common sense. Just like poetry allows us to "tweak" language and use
it in what others might call "logically odd" ways in comparison with the ordinary
prose, which we might use in a business letter, religious language allows religious
believers to stretch, bend, twist, and tweak the uses of language to express and
practice their religious belief and faith.
TILLICH AND THE GROUND OF BEING
Undoubtedly, one of the most prominent and influential religious thinkers of the
twentieth century was Paul Tillich, and although Tillich was much more of a
theologian than he was a philosopher, his treatment of religious language certainly
deserves some mention here. The analogical uses of language examined above and
the general approach of the liberal responses to the challenge of falsifiability
attempt to use analogical or metaphorical language in an attempt to reach some
understanding of the divine by comparison with more ordinary empirical matters.
As we have seen, such uses of language are not regarded as literally true but only
as analogically or metaphorically true. The main point is that the analogical,
metaphorical, nonliteral meaning is predicated upon and is a function of some
"underlying" literal meaning upon which the nonliteral use of language is based;
so, analogical or metaphorical uses of language attempt to make sense of the divine
or the transcendent nature of the holy by borrowing from and building upon
ordinary experience. The same is true of Tillich's use of symbolic language to talk
about God.
Tillich takes the nonliteral use of religious language a giant step further in his
theory of religious language as "symbolic language." Although Tillich's treatment
of religious language is much too complex and interwoven with his theology to
treat in any detail here, it is perhaps enough to point out that Tillich rejects the
basic approach, which uses analogical language to try to make sense of the

Ibid., p. 43.
Ibid.
77 Ibid., pp. 101-109.

75

76

TIIE PROBLEM OF RELIGIOUS LANGUAGE

55

transcendent in terms of the immanent. 78 Despite its appearance and common use,
Tillich insists that "God" is not a proper name, that is, "God" does not refer to a
unique individual. Rather, as is now well known, he identifies God as "the Ground
of Being" - being itself. Everything that is said about God is symbolic except the
one claim that "God is being itself.,,79 Thus, God cannot be understood in terms of
or by comparison with any ordinary experience or any part or aspect of the
empirical world. We cannot hope to use religious language to better understand or
explain God, according to Tillich. A religious symbol "stands for" or "points to" its
object, but it cannot, in any way, characterize or describe that object.
They [religious symbols] must express an object that by its very nature
transcends every object in the world .... A real symbol points to an object which
can never become an object. Religious symbols represent the transcendent but
do not make the transcendent immanent. They do not make God a part of the
empirical world. 80
The criticisms, critiques, and commentaries regarding the work of Tillich are far
too numerous and voluminous to be considered here. So far as Tillich's use of the
symbolic is concerned, we might simply say that if we have no literal knowledge
of and can make no literal assertions about God other than that God is being itself,
then, it seems, we are left with no grounds for choosing or assessing different
possibly competing or even conflicting symbolic uses of language. 81 If we can
make no reasonable choice about how we talk about God and what religious
symbols we use, then it appears that anything we might choose to say is just as
good as anything else we might choose to say. William Alston puts the point
nicely: "In order to furnish descriptions from which effective identifications can be
made, the component assertions in such a [religious] system must at some point
yield implications concerning experienceable states of affairs.,,82
Tillich is certainly aware of the difficulty here, but his attempts to address the
problem are woefully inadequate. In trying to explain what it is that "God" might
refer to, he says that "it points to the ultimate level of being, to ultimate reality, to
being itself, to meaning itself. That which is the ground of being is the object to
which the religious symbol points."s3 But this is of little help in providing some
content for "God." The phrases "being itself' and "the ground of all being" are
intended to describe a God that is completely limitless. No predicates or

78 See Paul Tillich, Systematic Theology (London: Nisbet and Company, 1953). For a more focused
treatment of religious language, see his 'The Religious Symbol," in Religious Experience and Truth,
edited by Sidney Hook.
79 Tillich, Systematic Theology, p. 264.
80 Tillich, "The Religious Symbol," p. 303.
81 Several critics have made this point, including William Alston, "Tillich' s Conception of a
Religious Symbol," in Religious Experience and Truth, edited by Sidney Hook, pp. 14 and 19. Also see
William T. Blackstone, "The Status of God-Talk," in Religious Language and Knowledge, edited by
Robert H. Ayers and William T. Blackstone.
82 Alston, ibid., p. 14.
83 Paul Tillich, "Theology and Symbolism," in Religious Symbolism, edited by F. Ernest John (New
York: Harper and Brothers, 1955), p. 110.

56

ANALYTIC PHILOSOPHY OF REDGION

characterizations can be used of being itself since everything is what it is because it


participates in being itself.
Tillich's position does not admit of any of the conative attitudes or beliefs that
are the basis of Braithwaite's theory of religious language. If the ground of being
does not admit of any characterizations or predicates that allow a religious person
to understand some content for the expression, then it appears that we can have no
attitudes toward such a being at all - even, as Alston has noted, as the object of our
"ultimate concern.,,84 Tillich's characterization of God as the ground of all being
forces him into an extreme position regarding the symbolic use of religious
language, and while such a position protects religious language from the challenges
of empiricism and the test of falsification, it leaves practically nothing in the way
of content to which religious believers can attach their beliefs and by which
religious believers can identify and differentiate their beliefs. Treating religious
language as completely symbolic in Tillich's sense means that we can say nothing
that is literally or cognitively meaningful about God. Such a consequent is rather
serious for the ordinary religious believer "in the street" and for the practice of
religion. Bowman Clarke summarizes the situation as follows:
Tillich's theory of the symbolic in philosophical theology can lead him but to
two alternatives. The first admits that we cannot say in any literal sense what is
affirmed and what is denied in each symbolic assertion, and it maintains that
there are semantic depths to such assertions which we can never fathom. This
means that it is quite possible for two symbolic assertions about God to be
contradictory without our knowing it. Likewise, two people could assert the
same symbolic statement about God, meaning different things by it, and yet not
know wherein they differed, or even that they differed. This, of course is just to
admit that language is theologically useless. The other alternative recognizes
that all symbolic assertions about God ... assert but one thing, that God is beingitself, not a finite being. In either case we are left with one conclusion, that God
is the ineffable, and we would do well to retreat to holy silence. 85
There seems to be no meaningful use of 'God' left in religious language as a
result of Tillich's notion of the symbolic. If we think of the different liberal
responses that have been treated here as falling upon a continuum to the left of
center and away from the conservative defenses of the literal meaningfulness of
religious language, there appears to be very little room left on the continuum
beyond Tillich's notion of the symbolic. Only the retreat to "holy silence" falls
farther along the continuum.
THOMAS McPHERSON AND HOLY SILENCE
Thomas McPherson proposes another liberal response to the challenge of
falsifiability. Many religious beliefs are ostensibly nonsense, McPherson points

Alston, "Tillich's Conception of a Religious Symbol," p. 23.


Bowman L. Clarke, "God and the Symbolic in Tillich," Anglican Theological Review, July 1961,
pp.9-1O.
84

85

THE PROBLEM OF RELIGIOUS LANGUAGE

57

OUt. 86 For example, Christians believe that God is one person; yet he is three
persons. God is transcendent; yet God is immanent. Jesus was divine; yet he was
human. Christ dies; yet Christ lives. How are we to understand such apparently
nonsensical claims? Do they have literal meanings? Or some "deeper" - perhaps
analogical - meaning? Is there a special religious language or logic that will allow
us to make the case for a special form of religious meaning? All of these questions,
McPherson thinks, are misguided and raise illegitimate concerns about religious
beliefs; so, he offers a rather radical way out of these difficulties. Although he does
not explicitly endorse the program of the logical positivists, he does think that they
have provided a "serious contribution" to philosophy by focusing upon the fact that
when we try to say things that are unsayable, the result is nonsense. 87 McPherson
puts his point in the following manner:
There are some things that just cannot be said. As long as one tries to say them
there is trouble. But if anyone does try to say them he must take the
consequences. We ought not to try to express the inexpressible. The things that
theologians try to say (or some of them) belong to the class of things that just
cannot be said. The way out of the worry is to retreat into silence. 88
According to McPherson, if we seriously adopt the positivist way out of this
problem, it is a mistake for the theist to try and defend all religious claims as
meaningful. It would also be a mistake to resort to the elaborate (and sometimes
forced) mechanisms of "bliks" or "models and metaphors" to try and capture some
special religious meaning of religious language. Theists should just admit that
language is inadequate for saying some of the things that they want to say and then
"retreat" into a holy silence.
McPherson draws upon the work of Rudolf Otto and Ludwig Wittgenstein to
buttress his positivistic solution to the inexpressible. In The Idea of the Holy,
Rudolf Otto tries to develop Friedrich Schleiermacher's characterization of
religious experience. 89 Schleiermacher began his treatment of religious experience
by insisting that the fundamental element of religious experience is a person's
feeling of absolute dependence upon someone or something other than one's self.
This means that the self becomes metaphysically dependent, resulting in a form of
subjectivism. The existence of God, revealed in religious experience, is not known
immediately but can only be inferred on the basis of one's immediate subjective
experience. Otto insists that the fundamental element of a genuine religious
(mystical) experience (what Otto calls "the numinous experience") is a person's
immediate awareness of the holy as another - as something outside and
independent of oneself - an experience of the "Wholly Other." Such experiences
are filled with a sense of awe and mystery and cannot be adequately captured by
language because numinous experiences exceed our ability to conceptualize. This
86 Thomas McPherson, "Religion as the Inexpressible," in New Essays in Philosophical Theology,
pp.131-43.
87 Ibid., p. 142.
88 Ibid., pp. 132-33.
89 Rudolf Otto, The Idea of the Holy: An Inquiry into the Non-Rational Factor in the Idea of the
Divine and Its Relation to the Rational (Oxford: Oxford University Press. 1958).

58

ANALYTIC PHILOSOPHY OF RELIGION

basic and fundamental element of religion is an example, McPherson thinks, of


what becomes nonsense if we try to put it into words. The theist might well
consider thanking the positivists for reminding theists of this fact. There are some
things in religion - perhaps the most essential and important things - that simply
cannot be said.
McPherson compares Otto's notion of the mystical with the brief and cryptic
remarks of Wittgenstein at the end of his Tractatus Logico-Philosophicus
concerning the mystical. McPherson focuses upon passages where Wittgenstein is
concerned about whether we can make sense of asking certain kinds of questions
about the world. In the Tractatus, Wittgenstein attempted to develop a framework
within which every conceivable meaningful proposition can be expressed. It is well
known that at the end of the Tractatus, Wittgenstein admits that, given the nature
of his theory, many of the things that he has had to say to develop the theory
cannot meaningfully be said within the theory. This is the problem of an
illegitimate totality expressed in his famous ladder metaphor. There are similarly
very general questions about why the world exists, which Wittgenstein calls
"mystical." Similarly, Wittgenstein says, "The feeling of the world as a limited
whole is the mystical feeling.,,9o
The comparison of Otto with Wittgenstein is a rather tenuous one except for the
point that each admits that there are certain kinds of things that cannot be put into
language. For Otto, such a claim is the result of a rather careful and detailed
treatment of religious experience. Wittgenstein, on the other hand, although he
does indeed use the term "mystical," did not have a developed theory of the
mystical. In this context, it is clear that what Wittgenstein had in mind concerned
questions about the totality that is this world. Within the context of the Tractatus,
we can reasonably ask questions about individual facts (particular concatenations
of objects) and even about groups of such facts, such as "Is the book on the table?"
and "Why is the book on the table?" Similarly, we can and do make claims about
the truth or falsity of propositions that "picture" or label or describe such facts.
However, we cannot reasonably ask why we have the particular totality of facts
that we have that constitute this world, nor can we even say that this is all there is
and there is no more to this world. Questions then, for Wittgenstein, about why the
world exists at all or why this particular world exists rather than some other world
are questions that cannot reasonably be asked because they have no reasonable
answers. So, in the very general sense in which there are some things that cannot
be put into language, Otto and Wittgenstein agree. If "the inexpressible" is an
essential or fundamental part of "true" or "real" religion, then positivism need not
be viewed as an enemy of religion or antireligious, McPherson believes. Given
such a view of the nonrational nature of religion, "the positivistic approach" might
encourage theists to recapture and focus upon this essential aspect of religion. 91
Some critics of McPherson have unfairly interpreted his suggestion about
retreating into holy silence as having extreme results that would prevent any
religious believer from ever talking with another believer about religious beliefs,
90 See Ludwig Wittgenstein, Tractatus Logico-Philosophicus, translated by D. F. Pears and B. F.
McGuinness (London: Routledge and Kegan Paul, 1961). 6.44, 6.45, and 6.5.
91 McPherson, "Religion as the Inexpressible," p. 139.

THE PROBLEM OF RELIGIOUS LANGUAGE

59

which would eventually eliminate all religious organizations. 92 But such reactions
are extreme. McPherson, as well as Otto, recognizes that there are many different
kinds of religious beliefs and many different kinds of religious language. There are
many parts of religion that are highly "conceptualized," and these parts can
certainly be put into language and are the subject matter of books, sermons, hymns,
and even the Bible itself.93
McPherson does recognize a potential difficulty with his suggestion of "the
positivistic approach" and the resulting retreat into holy silence. Theists may, in the
end, be forced to give up too much by this approach. "[By] throwing out the water
of theology we may be also throwing out the baby of 'direct,' 'first-order' religious
assertions; and this we may well not want to do," McPherson warns his fellow
theists. Indeed, such a result seems inevitable. To the degree that the "higher"
level, more "conceptualized" parts of religion are based in some way upon a
certain kind of immediate religious experience that is nonrational and cannot be
captured by language, any sort of rational understanding, explanation, or
justification of religious belief seems to be threatened.
The distinction between "higher-order," highly conceptualized religious
language and direct, "first-order" assertions is developed by R. F. Holland into the
distinction between theology and religion. 94 Religious discourse, for Holland, is the
kind of discourse that takes place during an act of worship by the theist and is not
intended to describe God or make assertions about God. Such discourse is really
directed toward God and is more about the believer engaged in the act of worship
than it is about God. Theological discourse, in contrast, is, in McPherson's and
Otto's phrase, "highly conceptualized" and treats God as an "object of study."
Theologians are concerned with questions about the existence of God and about
what attributes God might have,95 and the discourse of theology must reflect these
concerns. Religious discourse may have some kind of meaning for the theist, but it
is not the intended cognitive, empirical (or aesthetic) meaning of theological
discourse. For Holland, the result of this distinction is that religious discourse turns
out to be so epistemologically distinct from theological discourse that it cannot
serve in any epistemological role to provide any epistemological warrant for
theological discourse. Consequently, theological discourse is left dangling by itself
- "ungrounded" in "first-order" religious experience - and religious discourse is
left as the noncognitive prayers or hymns of the worshiping theist. In Holland's
account, McPherson's fears of throwing the baby of religion out with the bathwater
of theology become realized. Perhaps reformed epistemology or a revised notion of
religious experience may provide some avenue for the theist to "ground" religious
beliefs in religious experience, but to whatever degree that might be possible, it
will not follow Otto's analysis of mystical experience.

See Blackstone, "The Status of God-Talk," p. 88.


McPherson, "Religion as the mexpressible," p. 135.
94 R. F. Holland, "Religious Discourse and Theological Discourse," The Australasian Journal of
Philosophy, Vol. 34, no. 3, 1956.
95 Ibid., p. 147.
92

93

60

ANALYTIC PHILOSOPHY OF RELIGION

THE CONSERVATIVE RESPONSE


We have seen that the liberal response involves admitting that Flew is victorious
in his use of the falsifiability criterion for cognitive meaningfulness if religious
statements are treated as having empirical or cognitive content. The liberal
responses "shift ground" by adopting a variety of different sorts of accounts of the
special logical status of religious language - ranging from the use of analogical
language and "bliks" to the use of "models and qualifiers" and symbolic language.
Whatever success such shifting ground techniques might have, it comes with what
many theists would consider is too high of a price tag. Those figures who wish to
defend religious statements by refusing to give up the claim that they are
cognitively meaningful refuse to admit that Flew's use of the falsifiability criterion
is successful. The conservative responses in defense of the meaningfulness of
religious language carry the fight to Flew on his own territory by accepting the
falsifiability criterion as a legitimate criterion for testing empirical significance and
then attempting to show how religious language meets this criterion. Various
figures have attempted to do this in various ways.
CROMBIE AND HICK AND ESCHATOLOGICAL VERIFICATION
Ian Crombie defends the cognitive meaningfulness of religious language by
introducing both the use of analogy and the use of a special kind of verification for
religious statements. I have discussed the use of analogy in some detail above, so I
will focus here upon the sense in which religious statements are verifiable for
Crombie. The use of the criterion of falsification as a test for cognitive
meaningfulness is a combination of two different kinds of demands, according to
Crombie. First, there is the "logical demand," which requires that all statements of
fact must be falsifiable unless there exists an understood and accepted "rule of
language" that logically precludes such testing by falsification (or verification).96
To understand this point, Crombie compares two different kinds of statements that
are not verifiable. If a person says, "Killing is wrong," then it will not do to put this
statement to the test of falsification by pointing to certain factual evidence to try
and disprove it. The logical status of this statement precludes any such attempts at
falsification, and if a person insists on falsification as a test for meaningfulness,
then we might say that the person has not understood the logical status of the
original claim. 97 In contrast, if a person says, "Caesar had mutton before he crossed
the Rubicon," then, even though we might say that this statement does not lend
itself to testing by falsification, there is no logical difficulty with insisting upon
such a test. Factually (or actually), such a test by falsification may simply be a
waste of time, but if a person insists upon or attempts such a test, we do not say
that the person has misunderstood the logical status of the original claim. Do
religious statements have a logical status that precludes their being put to the test of
falsification? Crombie answers this question by insisting that there is nothing
logically odd about religious statements that precludes their being tested by the
criterion of falsification for cognitive meaningfulness. 98 Thus, Crombie obviously
96 I. M. Crombie, "Theology and Falsification," in New Essays in Philosophical Theology, edited by
Antony Flew and Alasdair MacIntyre, p. 125.
97 I discuss the relationship between ethics and religion in detail in Chapter VIII.
98 'bid., pp. 125-26.

THE PROBlEM OF RELIGIOUS LANGUAGE

61

rejects the tactic of those who opt for the liberal response by shifting ground.
Religious statements are cognitive1y meaningful statements for Crombie since
there is no logical problem of their verification.
The second kind of demand that the criterion of falsification makes is what
Crombie calls a "communicational stipulation," which means that in order for a
speaker to understand fully a statement that he makes, the speaker must know what
a test of the statement would involve. If a person makes the statement above about
Caesar having mutton but has no idea what mutton is or how to go about testing for
whether Caesar had mutton before he crossed the Rubicon, then Crombie maintains
that the speaker does not understand his own statement, that is, the statement has
no "communication value" for the speaker. 99 The issue of the communication value
of religious statements is not a matter that concerns the logical status of such
statements but rather the speaker's ability to understand the claim. Crombie's
answer to the question of how religious statements fair in respect to this second
demand of falsification is a lengthy one and depends u~on whether the question is
raised within the "parable" of religion or outside of it. 10 However, for the purposes
here, the crucial point is that religious statements are not logically immune to the
test of falsification according to Crombie. If a person introduces certain evidence
that ostensibly counts against the existence of an omnipotent and omnibenevolent
God (such as sickness or death of a loved one), we do not say that the person has
misunderstood the logical status of the claims "God exists" or "God loves us,"
because the existence of pain and suffering and death is prima facie incompatible
with the existence of an all-powerful and all-loving GOd.10 1
So theists admit that the existence of some evidence might carry prima facie
weight against the existence of God, but they will not admit that such evidence
weighs conclusively against religious claims. Some critics of theism claim that the
existence of pain and suffering and death do count conclusively against the theist's
claims, but this disagreement is not presently resolvable, according to Crombie,
because we cannot get ourselves into the position to get the proper evidence to
resolve the dispute (see Chapter VI). The dispute is not unresolvable because there
is anything "logically odd" about the status of religious claims. The theist's claim
"God loves us" is comparable to the claim that Caesar had mutton before he
crossed the Rubicon. The question of verification or falsification rests upon
whether we can "get ourselves into position" to gather the proper evidence.102 The
kind of position that Crombie insists upon here is an epistemological position
where the necessary evidence to verify or falsify the claim would be available to a
possible observer. We can describe such situations by using counterfactual
hypotheticals. In the case of Caesar, we might say, "If I were to transport myself
through time so that I could observe what Caesar had for breakfast just before he
launched out across the Rubicon and into Gaul, then I would be able to tell whether
he had mutton or not." For the theist, the matter of getting oneself into the proper
epistemological position to gather the appropriate evidence to test the claim that
God loves us is a bit more complicated. First, a person must die! Then one would
Ibid., p. 125-26.
See ibid., pp. 126ff.
101 Ibid., p. 126.
102 Ibid.

99

100

62

ANALYTIC PHILOSOPHY OF RELIGION

be in a position to observe whether there is a God and if that God is loving and
forgiving. So, counterfactually, one might say something such as, "If I were dead
and experienced God then I could verify that God exists." By the use of such a test,
which has now become known as "eschatological verification," Crombie insists
that there is no logical difficulty in verifying or falsifying religious statements.
Religious statements are "verifiable in principle" (just as statements about the
distant past are); they simply are not verifiable in reality or in practice. Religious
statements are thus cognitively or empirically meaningful. 103
John Hick has joined forces with Crombie to use the notion of eschatological
verification to defend the cognitive meaningfulness of religious language. 104 Hick
acknowledges the fundamental point of using verification as a criterion for
cognitive meaningfulness. He provides a clear statement of what he understands by
the claim that a statement is empirically significant in a formulation that is
reminiscent of the early formulations of the verification criterion of meaning. Any
statement is a factual assertion and is cognitively meaningful, according to Hick,
"if and only if the state in which the universe would be if the putative assertion
could correctly be said to be true differs in some experienceable way from the state
in which the universe would be if the putative assertion could correctly be said to
be false, all aspects of the universe other than that referred to in the putative
statement being the same in either case.,,105 The question then becomes one of
whether religious statements are assertions under this definition. Hick's answer is
in the affirmative because he agrees with Crombie that the continued existence in
an afterlife would make religious claims verifiable in principle, though they would
not be falsifiable
Falsifiability and verification are commonly treated as if they are
"symmetrically related," Hick points out, that is, the empirical data that are
regarded as verifying of a particular proposition if those data are actual and
observed are also regarded as falsifying of the same proposition if those data are
not actual and are not observed. Such symmetry between verification and
falsification does hold in the cases of most propositions, but does it hold in every
case? Hick argues that it does not. Consider the following proposition: "There are
three successive sevens in the decimal determination of 1t.,,106 Determining the
decimal determination of 1t is, of course, an infinite process, and no matter how
much time is spent and no matter to what place the process is carried out, it will
always be possible to continue the process. To date, no three successive sevens
have ever turned up in the process, but at some point in the future, they might, and,
of course, if the claim that there are three successive sevens in the decimal
determination of 1t is really false, it might never be shown that it is. Therefore, this
proposition is an example of a proposition for which the symmetry of verification

Ibid.
See John Hick, Faith and Knowledge (Ithaca, N.Y.: Cornell University Press, 1957), pp. 150-62,
and "Theology and Verification," originally in Theology Today, Vol. 17, 1960. Reprinted in Religious
Language and the Problem of Religious Knowledge, edited by Ronald E. Santoni, to which the page
numbers here refer.
105 Hick, 'Theology and Verification," p. 381.
106 Ibid., p. 366.
103

104

THE PROBLEM OF RELIGIOUS LANGUAGE

63

and falsification does not hold. This claim is verifiable in principle, but it is not
falsifiable.
For Hick, "the hypothesis of continued conscious existence after bodily death"
is another example of a claim that is verifiable but not falsifiable. If one continues
to exist after one's death, then such a claim would be verified, but if one does not
continue to exist after one's death, then the claim of continued existence cannot be
falsified since there would then be no experiences or observations to falsify the
original claim. So the claim of continued existence may be false, but no one could
ever demonstrate that it is false. 107
Crombie and Hick's notion of eschatological verification has been severely
attacked by many critics. One obvious difficulty is that eschatological verification
depends essentially upon the "hypothesis of continued conscious existence after
bodily death," a hypothesis whose status appears to be as much in doubt as the
hypotheses "God exists" or "God loves us." Without some further explanation of
the nature and status of the "hypothesis of continued conscious existence," it seems
as if Hick is begging the question by appealing to one religious claim in order to
establish the empirical meaningfulness of others. Hick is aware of this difficulty
and has attempted to argue independently for the meaningfulness of the notion of
"continued personal existence," that is, independently of other religious claims.
The issue of continued personal existence is treated in Chapter VII; therefore, I
focus here only on the question of the empirical meaningfulness of such a claim in
a very brief, general manner. Much of the answer to the question about whether
claims about continued personal existence are meaningful depends to a great extent
upon what exactly it is that we mean by "continued personal existence." Flew (and
others) have argued that the notion of a continued, completely spiritual personal
existence, that is, a noncorporeal spirit, or a disembodied spirit, is nonsensical. 108
In order for the supposed continued existence to be a personal existence, the
continued existence of the same person after death who existed before death, it
seems as if it would be necessary for there to be a certain amount of continuity in
the person's experiences and memories (a condition that Hick recognizes and
incorporates into his account of continued personal existence). If, then, in this
continued, completely spiritual state of existence, a person is to have experiences
or memories of experiences that took place here on earth during the person's
lifetime as a human being, it seems that the only sense that one can make of such
claims about the possibility of such experiences or memories would require that
there be some sense organs through which such experiences are possible and a
brain through which memories are possible. Our only understanding of personal
identity seems to require some sort of body. We can temporarily bypass the issue
of whether there is a nonphysical component to personhood (for example, a soul or
a mind) by simply saying that whatever nonphysical component there may be to
personhood, a physical instantiation of whatever that might be (if anything) seems
necessary for any commonly understood notion of persons.

Ibid.
Antony Flew, "Death," in New Essays in Philosophical Theology, edited by Antony Flew and
Alasdair MacIntyre, pp. 267-72.
107

108

64

ANALYTIC PHILOSOPHY OF RELIGION

Hick is aware of and seems to accept many of these difficulties involving


completely spiritual continued personal existence, and, at least partially as a result
of these difficulties, Hick defends a Christian notion of a physical, bodily
resurrection as being cognitively meaningful. 109 Hick evokes no less of an
authority than Moritz Schlick, whom he quotes in support of his claim that
continued personal existence based upon some sort of physical, bodily resurrection
is empirically meaningful. Schlick says, "We must conclude that immortality, in
the sense defined ["survival after death" rather than "never-ending life"], should
not be regarded as a 'metaphysical problem,' but as an empirical hypothesis,
because it possesses logical verifiability. It could be verified by the following
prescription: 'Wait until you die!",llo
Others have questioned whether Hick's commitment to eschatological
verification is consistent with his commitment to religious pluralism. 111 The
commitment to eschatological verification is seen as problematic because at least
some of the many different views of the afterlife that occur in different world
religions would not allow for the kind of verification that Hick maintains is
possible. However, it appears to be plausible for Hick to accommodate his notion
of eschatological verification to his notion of religious pluralism. To do so, Hick
need only modify his parable to allow for many travelers on the spiritual path at the
same time with many different (and perhaps incompatible) views of what lies at the
end of the road. Some of those views, such as "metaphysical absolutism,"
according to which the continued existence after death is not a personal one but
one in which individuals merge and become an indistinguishable part of the One,112
would not be verifiable even in principle, but this does not mean that Hick's own
brand of continued personal existence dependent upon some sort of physical,
bodily resurrection would not be verifiable in principle. Even if metaphysical
absolutism should happen to be true, this would not falsify Hick's claim about his
version of continued personal existence since it would not result in any experiences
or observations; hence, if one accepts Hick's claim regarding the asymmetry of
verification and falsification, this would protect his claim against different possible
outcomes, based upon different religious views, regarding what happens after
death.
BASIL MITCHELL
Basil Mitchell responds to the challenge of falsification by admitting that the
theist will not allow anything to count conclusively or decisively against religious
beliefs; however, Mitchell does think that the theist is willing to allow that some
evidence does count as prima facie evidence against religious claims; otherwise,
the problem of evil would not be a problem. The existence of needless pain and
See Hick, "Theology and Verification," pp. 370ff.
Moritz Schlick, "Meaning and Verification," Philosophical Review, 1936. Also, see Hick,
"Theology and Verification," op. cit., p. 370. See also Antony Flew, "Can a Man Witness His Own
Funeral," in The Presumption of Atheism (New York: Barnes and Noble, 1976), and The Logic of
Immortality (Oxford: Backwell Publishing, 1987).
III See, for example, Michael Peterson, William Hasker, Bruce Reichenbach, and David Basinger,
Reason and Religious Belief' An Introduction to the Philosophy of Religion, Second Edition (Oxford:
Oxford University Press, 1998), p. 277, fn. 21.
112 Ibid.
109

110

THE PROBLEM OF RELIGIOUS LANGUAGE

65

suffering surely counts against the belief in the existence of an omnipotent and
omnibenevolent God, for example, but it does not conclusively falsify such a
belief. The reason for this, Mitchell thinks, is the particular attitude of the believer
toward such claims. Mitchell suggests that in contrast to Flew's Parable of the
Invisible Gardener we consider the following story:
In time of war in an occupied country, a member of the resistance meets one
night a stranger who deeply impresses him. They spend the night together in
conversation. The Stranger tells the partisan that he himself is on the side of the
resistance - indeed that he is in command of it, and urges the partisan to have
faith in him no matter what happens. The partisan is utterly convinced at that
meeting of the Stranger's sincerity and constancy and undertakes to trust him.
They never meet in conditions of intimacy again. But sometimes the Stranger
is seen helping members of the resistance, and the partisan is grateful and says
to his friends, "He is on our side."
Sometimes he is seen in the uniform of the police handing over patriots to the
occupying power. On these occasions his friends murmur against him: but the
partisan still says, "He is on our side." He still believes that, in spite of
appearances, the Stranger did not deceive him. Sometimes he asks the Stranger
for help and receives it. He is then thankful. Sometimes he asks and does not
receive it. Then he says, "The Stranger knows best." Sometimes his friends in
exasperation say, "Well, what would he have to do for you to admit that you
were wrong and that he is not on our side?" But the partisan refuses to answer.
He will not consent to put the Stranger to the test. 113
Mitchell maintains that religious utterances must be assertions to be meaningful.
Unlike the lunatic who holds the "blik" concerning all dons, the partisan will admit
that some evidence does "count against" his claim that the Stranger is "on our
side." For Hare, of course, evidence is not a factor in determining why a person
holds or does not hold a particular blik. In contrast, Mitchell's partisan does intend
to be asserting something about the Stranger that is either true or false when he
says, "He is on our side." Likewise, the theist intends to assert something about
God when the theist says, "God loves us." According to Mitchell, the reason that
the partisan (or theist) will not allow any evidence to conclusively falsify his claim
is because of his faith in the Stranger.
Just how far the partisan might be willing to go and just how long he might be
willing to persist in denying that the ostensibly falsifying evidence does not
conclusively falsify the claim that the Stranger is on his side is not possible to
determine in advance, according to Mitchell, since it will all depend upon the
strength of the commitment that the partisan feels toward the Stranger. So, it is not
possible, according to Mitchell, to determine in advance how determinedly or how
long the theist will refuse to admit falsification of claims about God while allowing
that some evidence does count against those claims. 114
113 Basil, Mitchell, ''Theology and Falsification, " in New Essays in Philosophical Theology, edited
by Antony Flew and Alasdair MacIntyre, pp.103-104.
114 Ibid., pp. 104-105.

66

ANALYTIC PHILOSOPHY OF REDGION

Whether Mitchell provides an adequate response to the challenge of falsification


depends upon exactly how we understand the reasons why a believer might accept
or reject evidence against the existence of God. If the point is that the personal
commitments and personal beliefs of an individual will affect the way in which the
individual assesses, processes, and eventually accepts or rejects evidence that may
count against some of those beliefs, then the point is obviously true. If someone
were to present me with evidence that my eldest son is a murderer and international
drug dealer connected with a Columbian cartel, my initial reaction would
obviously be based upon a prima facie disbelief of the evidence because I love my
son. There is nothing very unusual or unreasonable about such a reaction. This
simply means that there is a higher burden of proof that may exist in such
situations - a higher burden that may persist for quite some time. However, if I
take the general position that I will never accept any evidence whatsoever, no
matter what it may be, that would prove conclusively that my son is a murderer or
an evil person, then such a claim is unreasonable.
It also matters whether one's failing is epistemic or psychological. If Mitchell's
theist admits that, in principle, it is theoretically possible for some evidence to
count decisively against the existence of God, but that he personally simply cannot
bring himself psychologically to admit that such evidence ever really rises to such
a level on any particular occasion, then the theist simply admits to a psychological
failing resulting from a personal belief or faith. In such a situation, we may admit
that a religious utterance affirming the existence of God is an assertion, since it is
possible to imagine the theist going through some sort of philosophical therapy to
restore congruency to his psychological faculties. If, on the other hand, the theist's
faith leads to the epistemic decision to treat the existence of God as theoretically
immune to any possible decisive challenges of evidence, then, under such
circumstances, the claim that God exists seems to approach the level of being
tautologous and does not appear to be an assertion with cognitive content. In this
situation, the religious claim fails the test of falsification.
SPEAKING LITERALLY ABOUT GOD?
Many of those who have opted for a liberal, shifting-ground response to the
challenge posed to religious language by the notion of falsification were forced
into such an option because of the difficulty of explaining how one can use
language to make statements that are literally true or false about the transcendent
nature of God (or other transcendent aspects of religious belief). More recently in
this debate over literal versus nonliteral uses of religious language, William P.
Alston has taken on the task of attempting to remove one of the conceptual barriers
to applying certain predicates straightforwardly and literally to God.115 We have
1\5 William P. Alston, "Speaking Literally of God," in Is God GOD? edited by Axel D. Steuer and
James W. McClendon, Jf. (Nashville: Abingdon Press, 1981). Reprinted in Michael Peterson, William
Hasker, Bruce Reichenbach, and David Basinger, Philosophy of Religion: Selected Readings (Oxford:
Oxford University Press, 1991), to which the page numbers here refer. Also see his "Functionalism and
Theological Language," American Philosophical Quarterly, Vol. 22, 1985. Both "Speaking Literally of
God" and "Functionalism and Theological Language" are reprinted in Divine Nature and HUlrUln
Language: Essays in Philosophical Theology, edited by William P. Alston (Ithaca, N.Y.: Cornell
University Press, 1989).

THE PROBLEM OF RELIGIOUS LANGUAGE

67

seen above that analogical, metaphorical, or other nonliteral uses of religious


language presuppose or rely upon some underlying literal use with cognitive
meaning at some point. Alston sets out to investigate this one crucial use of
religious language where a statement is made to advance a truth claim that is held
to be "literally true" of God. Specifically, Alston wishes to investigate how one
may maintain that some "intrinsic" predicates such as "made heaven and earth"
and "spoke to Moses" can be literally attributed to God in a true assertion. 116
Although there have been different reasons why others have thought that such
straightforward, literal predication is impossible, Alston correctly identifies God's
transcendency as a major reason why many have rejected the notion of such
predication. Among the several different attributes of God that are responsible for
God's transcendent nature is incorporeality. Alston initiates an inquiry to
determine whether God's incorporeality is a conceptual block against making truth
claims in which predicates are applied literally to God, that is, the question is
whether God's incorporeality makes it logically impossible to make certain truth
claims in which predicates literally apply to GOd. 117
Alston says that he follows P. F. Strawson in choosing "personalistic predicates"
(hereafter called "P-predicates") to investigate. liS P-predicates are those, Alston
claims, that can be asserted truly only of a "personal agent," and a personal agent is
"an agent that carries out intentions, plans, or purposes, in its actions, that acts in
the light of knowledge or belief; a being whose actions express attitudes and are
guided by standards and principles; a being capable of communicating with other
such agents and entering into other forms of personal relations with them.,,119
Taking human agents as our models, we can speak literally of God using Ppredicates, Alston says, "if we can form concepts of intrinsic divine properties,,,120
and whether we can form such concepts of such properties will "depend both on
what God is like and on the content of the predicates.,,121
Alston divides P-predicates into mental or psychological predicates (Mpredicates) and action predicates (A-predicates).122 He then seeks to establish that,
according to at least some forms of "logical connectionism" (according to which
mental states of an agent are defined in terms of some "logical connection" with
the agent's behavior, demeanor, or disposition), the attribution of M-predicates to
an incorporeal entity is possible since the behavior in terms of which the particular
M-predicates are defined does not have to be overt behavior. Alston bases his
claim here upon the distinction between basic actions and nonbasic actions. 123
Alston, "Speaking literally of God," p. 366-67.
See ibid., p. 371. Alston suggests that this issue is part of a much broader project in which one
would conduct a similar inquiry regarding the other metaphysical attributes, such as infinity,
omnipotence, and eternality, in a similar manner.
118 Ibid., p. 367. Also, see P. F. Strawson, Individuals (Garden City, N.Y.: Doubleday, 1959) pp.
lOOff.
119 Alston, "Speaking Literally of God," pp. 367-68.
120 Ibid., p. 370.
121 Ibid., p. 371.
122 Ibid., p. 372. Alston's use of the designation "M-predicates" for mental predicates as a subclass
of P-predicates can easily be confusing for those familiar with Strawson's original treatment of Ppredicates in Individuals. There, M-predicates are bodily predicates.
123 Basic actions have been defmed in different ways by different people. See, for example, Arthur
Danto, "Basic Actions," American Philosophical Quarterly, Vol. 2, 1965, and Terence Peneihum,
116
117

68

ANALYTIC PHILOSOPHY OF REliGION

Basic actions are not performed by performing any other action, and other,
nonbasic actions are performed by performing a basic action. In the case of human
beings, basic actions are bodily actions; in the case of God, Alston claims, basic
actions could be mental actions or "other changes" that substitute for bodily
changes. Against critics who have rejected the claim that God, as a pure,
incorporeal spirit, could act as an agent,l24 Alston claims that since God does not
have a body, then all of God's actions may be basic actions either by being mental
actions or "other changes" in which "God could exercise direct voluntary control
over every change in the world which he influences by his activity.,,125 Hence,
Alston concludes, there is no conceptual barrier to applying M-predicates to an
incorporeal being, God. 126
In evaluating Alston's position, we must understand his intent. At one point, he
says that his purpose is to show that some forms of logical connectionism,
specifically functionalism, are compatible with the possibility of attributing Mpredicates to an incorporeal agent. But even if he is successful in doing this, it does
not make very much progress toward speaking literally about God. As Alston
himself says, being able to speak literally about God will depend both upon the
nature of God and the content of the predicates, and Alston does not provide much
help concerning possible content of predicates that may apply literally to God. If
we try and get more specific about the content of P-predicates that might be
applied literally to an incorporeal being and about how those predicates might be
literally true of such a being, several difficulties arise. The conceptual or logical
barrier against the true and literal application of P-predicates to an incorporeal
being are more difficult to remove than Alston supposes.
Alston claims to be making a modest claim and doing nothing out of the
ordinary when he suggests that P-predicates may be applied to God. Early in
building his case, he says,
I will restrict myself here to whether (some) P-predicates can be true of God in
(some of) the senses in which they are true of human beings. The only
qualification I make on that is that I shall consider a simple transformation of
certain human action predicates - "simple," in that the change does not involve
any radical conceptual innovation. 127
However, the result, that is, the claim that we can literally apply certain Ppredicates to God, is hardly a "simple transformation" of the predicates that we
normally use to describe human agents. Alston argues that the notion of a basic
action coincidentally involves bodily movement in the case of human beings since
it just happens to be a contingent fact about human beings that we can only bring
Survival and Disembodied Existence (London: Routledge and Kegan Paul, 1970), pp. 40ff. Alston
follows Penelhum in his account.
124 Including Paul Edwards, "Difficulties in the Idea of God," in The Idea of God, edited by Edward
H. Madden, Rollo Handy, and Marvin Faber (Springfield, Ill.: Charles C. Thomas, Pnblisher, 1968), p.
48, and Kai Nielsen, Contemporary Critiques of Religion (London: Macmillan, 1971), pp. 116-17.
125 Alston, "Speaking Literally of God," p. 383.
126 Ibid., p. 383-84.
127 Ibid., p. 371.

THE PROBLEM OF RELIGIOUS LANGUAGE

69

about other actions by the movements of parts of our bodies that are under our
voluntary control. Alston claims that we can conceive of agents other than human
agents that are such that other things besides their own bodies are under their
direct, voluntary control, and God may be one such agent. 128
Can we really conceive of such agents with such actions as Alston suggests? To
approach an answer to this question, we might begin by noting that Alston uses Ppredicates in a way that is crucially and fundamentally different from the way in
which Strawson originally intended them. 129 For Strawson, the concept of a person
is primitive. By saying that the concept of a person is primitive, Strawson means
that "the concept of a person is the concept of a type of entity such that both
predicates ascribing states of consciousness and predicates ascribing corporeal
characteristics ... are equally applicable to a single individual of that type.,,130
Strawson is very clear that "a necessary condition of states of consciousness [Ppredicates] being ascribed at all is that they should be ascribed to the very same
things as certain corporeal characteristics."l3l Persons, for Strawson, are
individuals of the same logical type to whom we are willing to ascribe both the
predicates ascribing states of consciousness or implying states of consciousness
(what Strawson calls "P-predicates") and the predicates ascribing corporeal
characteristics (what Strawson calls "M-predicates") 132 as we ascribe to ourselves.
There is certainly nothing wrong with Alston using the notion of "P-predicates"
in a very different way from the way in which Strawson used them. However,
Strawson has captured a very strong intuition by insisting that P-predicates and
predicates ascribing corporeal characteristics (his M-predicates) must be applied to
individuals of the same logical type. Alston suggests that we human beings can
stretch a particular concept to include features that were not present in the observed
cases from which we formed that concept, and this is certainly true. However, if
the notion of a person is stretched so far that P-predicates can reasonably be
applied to a noncorporeal entity (which would certainly be an individual of a
different logical type than ordinary persons), then a fortiori, it seems, P-predicates
could reasonably be applied to a variety of different kinds of corporeal bodies, such
as golden retrievers or trees or computers. It certainly seems more of a stretch to
apply such predicates to a noncorporeal entity than it does to apply them to
different kinds of corporeal entities. Strawson's original intuition then comes back
into play here. If we stretch the concept of a person too far, that is, beyond
individuals of the same logical type, what kind of an individual could not be a
person? What identifies individuals of the same logical type is the fact that we are
willing to ascribe both P-predicates and certain bodily predicates to those same
individuals.
If there are such things as incorporeal agents, then they certainly are of a
different logical type than human agents, and this just means that if we apply PIbid., p. 382-83.
Perhaps Alston means that he is ''following Strawson" just in naming or calling these predicates
"P-predicates"; however, the context suggests that he means that he is following Strawson in using Ppredicates as Strawson does.
130 Strawson, Individuals, pp. 97-98.
131 Ibid., p. 98.
132 Ibid., p. 100.
128

129

70

ANALYTIC PHILOSOPHY OF RELIGION

predicates to such entities it will not be in a literal manner. As Terence Penelhum


notes, even if we are willing to admit that incorporeal agents may admit of basic
actions by affecting changes in the physical world "directly," whatever the power
of agency is by which those basic actions would take place would be "positively
abnormal" compared to human agency.,,133 Even Alston admits that such a sense of
agency is "outside our ordinary experience.,,!34 Just how abnormal would such an
incorporeal agent be? Abnormal enough that most of the ways in which we
describe and treat P-predicates in the case of human beings could not be
maintained. 135 In particular, the notion that an incorporeal agent is one (which
Alston described early in his inquiry) "that carries out intentions, plans, or
purposes in its actions, that acts in the light of knowledge or belief; a being whose
actions express attitudes and are guided by standards and principles,,136 seems
impossible. For example, if all of God's actions are basic actions, and if "any
change whatsoever" can be a basic action, then we could no longer make sense of
acting upon an intention or a plan or a purpose since none of God's actions would
then be the result of some other action. If we understand every change that takes
place in the world as the result of a basic action of God, then nothing ever happens
as the result of an intention, plan, or purpose. In fact, the relevant M-predicates (in
Alston's sense of "M-predicates" - including intentions, plans, and purposes)
would not seem to apply in any literal way to God. Normally, an agent might
intend to do x without actually doing x, or an agent might intend to do x but not
intend to do x in a particular way, or an agent might intend to do x but then have to
choose which way to do x from a variety of different ways, or an agent might have
the purpose of doing x but might not yet have formulated a clear plan for a way of
doing X.137 However, if all of God's actions are basic actions, then we could make
none of these distinctions in the case of God's acting as an incorporeal agent. All
of this suggests that when we talk about incorporeal agency, such talk is not literal,
and the comparison of corporeal with incorporeal agency has not improved the
prospects for speaking literally about God.
THE PROBLEM OF REFERENCE
Certainly, for most contemporary philosophers of religion, the problem of
religious language has meant the problem of the meaningfulness of religious
language. However, there are also unique but related problems of the reference of
the proper names and definite descriptions, that is, referring expressions, used in
religious language that are separate and distinct from the question of the
meaningfulness of religious language. Several related questions concerning
reference in religious language indicate the kind of issues at stake. Do certain
religious referring expressions refer, and if they do refer, what sort of theory can
133 Terence Penelhum, Survival and Disembodied Existence, p. 42. Penelhum first suggested using
the distinction between basic and nonbasic actions to explore the possibility of a disembodied,
noncorporeal agency. See pp. 40ff.
134 Alston, "Speaking Literally of God," p. 383.
135 For a detailed treatment of the possibility of understanding God as a noncorporeal agent, see
Charles Taliaferro, Consciousness and the Mind of God (Cambridge: Cambridge University Press,

1994).
136
137

Alston, "Speaking Literally of God," p. 367-68.


See Penelhum, Survival and Disembodied Existence, p. 38.

THE PROBLEM OF RELIGIOUS LANGUAGE

71

we use to explain how such reference is successful? Is it possible to distinguish


between successful reference and reference failure in the uses of religious referring
expressions?
DESCRIPTIVIST ACCOUNTS OF REFERENCE
Philosophers have provided both descriptivist accounts and causal accounts to
explain how reference generally takes place. 138 Each kind of theory of reference
has distinct advantages and disadvantages for use with referring expressions within
religious language. According to descriptivist theories of reference, proper names
and definite descriptions successfully refer because of certain predicates or
characteristics that a particular referent possesses that are truly contained in the
name or description. Descriptivist accounts of reference thus take their departure
from empiricism, realism, and the correspondence theory of truth. Things in the
world possess certain properties or characteristics, and referring expressions are
able to refer to those things in the world by providing the descriptions of their
referents. Different individuals possess different properties, so we must use
different referring expressions to refer to different individuals. Consider the case of
proper names. If a speaker uses the name "Moses," then according to a
descriptivist account of reference, this name refers to an individual who possesses
certain properties or about whom we would say that certain descriptions are true.
For example, if I used the name 'Moses' and someone asked me, "Whom are you
talking about?" or (if the inquisitor were a philosopher), "To whom are you
referring?" I might respond by providing one or several of the many descriptions
that I take to apply truly to Moses, for example, "the baby who was hidden in the
bulrushes," "the man to whom God appeared in the burning bush," "the man who
parted the Red Sea," "the man who led the people of Israel out of Egypt," or "the
man to whom God gave the Ten Commandments.,,139 Since different individuals
possess different properties, we can explain how different referring expressions
have different referents. Using what Keith Donnellan calls "the principle of
identifying propositions," we can determine the referent of a referring expression
by identifying the individual about whom (or which) the "essential descriptions" of
the proper name or definite description are most true. 140
THE CAUSAL THEORY OF REFERENCE
In contrast, the causal theory of reference does not rely upon any predicate,
characteristics, or descriptions to explain successful reference. The individual
referred to by a referring expression, according to the causal theory of reference, is
the individual that occupies the proper place in the historical account of how a
speaker uses a certain name or description to intend to refer to some individual. To
determine the referent of a referring expression using the causal theory of
reference, a story must be provided that details the causal connection between a
particular individual and the use of a particular name or description to refer to that
138 Much of the following is drawn from James F. Harris, "The Causal Theory of Reference and
Religious Language," International lournalfor Philosophy of Religion, Vol. 29,1991, pp. 75-86.
139 For a more complete treatment, see ibid., pp. 75-76.
140 See ibid., p. 76, and Keith Donnellan, "Proper Names and Identifying Descriptions," Synthese,
Vol. 21,1970, pp. 335-58.

72

ANALYTIC PHILOSOPHY OF REUGION

individual. Causal accounts of reference thus reject the principle of identifying


propositions. In the causal theory, successful reference can take place completely
independently of any set of characteristics or properties of the referent since
successful reference depends solely upon tracing the use of the referring expression
back through its historical usage to the initial "baptism" when it was fIrst used to
refer to a specifIc individual. 141 According to Saul Kripke,
An initial "baptism" takes place. Here the object may be named by ostension, or
the reference of the name may be fIxed by a description. When the name is
"passed from link to link" the receiver of the name must, I think, intend when he
learns it to use it with the same reference as the man from whom he heard it. 142
The causal theory thus explains reference of a proper name, such as 'Moses,' by
tracing the use of the name back through its long history of use within a particular
community to an initial "baptism," when some speaker initially dubbed a particular
individual 'Moses.' When a speaker now uses the name 'Moses,' the same
individual who was the referent of the original use of the name, that is, the
individual who occupies the proper causal position in the chain of the use of the
name, is the referent - completely independently of any characteristic or property
that is commonly attributed to that individual. In the example of the proper name
'Moses,' suppose that we could somehow determine that none of the properties
that religious people commonly attribute to Moses are really true of him. 143 We
would than say such things as "Moses was never placed in the bulrushes" and "It
was not Moses to whom God appeared in the burning bush" and "Moses did not
lead the people of Israel out of Egypt" and "It was not Moses to whom God gave
the Ten Commandments." We would still be "talking about," that is, referring to,
the same individual when we said of that individual that the things we previously
thought to be true of him are not really true. The fact that a particular individual
possesses a particular property and even the fact that an individual possesses the
entire collection of properties commonly attributed to that individual are simply
contingent matters of fact.
.
To capture the intuition that underlies the causal theory of reference, we can
easily imagine a possible world very close to this one in which the same individual
whom we call "Moses" existed but in which that individual was never put in the
bulrushes. And we can imagine that in that other possible world, the individual
whom we call "Moses" was not really the person to whom God gave the Ten
Commandments, and we can do this mental experiment with the same results for
every single known attribute or property commonly attributed to Moses.

141 Such a connection is always established within a community in which other speakers recognize
and use the same referring expression. Thus, the causal theory of reference depends upon the placing of
an individual speaker within a community of speakers of the same language. See Harris, "The Causal
Theory of Reference and Religious Language," pp. 76-77, and Saul Kripke, Naming and Necessity
(Oxford: Basil Blackwell, 1972), pp. 94-98.
142 Kripke, ibid., p. 96.
143 Harris, "The Causal Theory of Reference and Religious Language," pp. 77-78.

THE PROBLEM OF RELIGIOUS LANGUAGE

73

REFERENCE AND RELIGIOUS LANGUAGE


Referring expressions within religious language that are supposed to refer to
God and other transcendent entities pose special difficulties since it is obviously
not easy to point to or identify the referents of those referring expressions within
ordinary experience. To address these special difficulties, Richard Miller has
maintained that the causal theory of reference offers some advantage for handling
problems of reference for referring expressions within religious language because
descriptivist theories tend to "undermine" historical religion whereas historical
accounts of reference reinforce the historical validity of religion. l44 Calling the
causal theory of reference "the historical theory" can be very misleading by
overemphasizing the role of "history" in the causal account. It is true that, in order
to fix the referent of a referring expression, a historical account must be given - a
story must be told - in order to arrive at the first link of the causal chain, the initial
"baptism" where the word or description was first used to refer to a specific
individual. So there is a minimal sense in which the causal theory of reference is
historical. However, the causal theory does not use or "validate" the content of any
particular historical account because none of the things said in a particular
historical tradition about a particular individual (the referent) have to be true. The
only important thing is that speakers in the community have used a certain
referring expression intending to refer to the same individual as the individual
referred to in the initial baptism when the referring expression was first used. 145 In
the example of our use of the proper name 'Moses,' according to the causal theory
of reference, the name refers to a particular individual and continues to refer to that
same individual even if none of the things traditionally and historically within the
Jewish and Christian religious communities are true. Kripke completely separates
the accuracy of any historical account used for fixing the referent of a particular
referring expression and the general, theoretical point that there must have been
some initial baptism. Kripke says, "[T]he Biblical story might have been a
complete legend, or it might have been a substantially false account of a real
person.,,146 Such a complete dismissal of the accuracy of the narrative used to fix
the referent would hardly seem to lend support to the historical validity of religion.
THE THEORY OF SPEECH-ACTS
Throughout this chapter, we have seen a continual progression and expansion in
the kind of problems and issues upon which philosophical analysis has been
brought to bear on religious language in the twentieth century. This progression
and expansion of problems and issues for religious language mirrors the changes
and developments (discussed in Chapter I) that took place within analytic
philosophy itself. The failure of logical positivism and the developments within
analytic philosophy that began with ordinary-language philosophy and the later
Wittgenstein have led to very different views about the nature of language, the
nature of religious language, and the nature of the problems that arise in religious
language. The move has been away from a very monolithic view of language and
144 Richard B. Miller, "The Reference of 'God,'" Faith and Philosophy, Vol. 3, no. 1. January 1986,
pp. 10-12.
145 Harris, "The Causal Theory of Reference and Religious Language," pp. 79-80.
146 Ibid., p. 80, and Kripke, Naming and Necessity, p. 66.

74

ANALYTIC PHILOSOPHY OF RELIGION

an essentialist view of meaning that pervaded the views of those in the early part of
the century - including Bertrand Russell, the members of the Vienna Circle, and
the early Wittgenstein - toward a much more flexible and open view of language
and a nonessentialist view of meaning. Early analytic philosophers focused solely
upon abstract propositions or assertions and the related questions about the
meaning and truth of those propositions or assertions. Given these underlying
views of the nature of language, questions were very narrowly circumscribed both
in the way in which they were raised and in the way in which they could be
addressed. The later Wittgenstein and ordinary-language philosophers broke from
this confining view of language in a way that has had a lasting effect upon both the
nature of philosophical analysis and the analysis of religious language. The result
has been a multifarious variety of approaches to analytic treatments of religious
language.
Another development in twentieth-century analytic philosophy, which occurred
contemporaneously with the later Wittgenstein and which shared many of the same
fundamental insights about the nature of language, was J. L. Austin's theory of
speech-acts. Austin also abandoned the essentialism and the emphasis upon
abstract propositions or assertions of the earlier views of language and, in fact,
abandoned as well the notion that "language" (in the abstract) can be the proper
object of philosophical analysis. Austin's theory of speech-acts replaces the
abstract, general, and monolithic notion of language with the concrete, specific,
and multifarious notion of speech-acts. A speech-act is an act wherein a speaker
issues an utterance on a particular occasion, with a particular intention, under
particular conditions, to a particular audience, with particular consequences. In
issuing an utterance, the speaker performs some act in addition to simply saying
something. For example, the speaker might make an announcement or issue a
warning or fulfill a promise. Speech-act analysis involves the analysis of speech
situations that will include the utterance itself, the status or authority of the
speaker, the intentions of the speaker, the specific conditions under which the
utterance is made, the linguistic and social conventions that govern the utterance
and the speaker, and the audience to whom the utterance is issued.
Within speech-act analysis, there is no longer the problem of religious language
to be analyzed and resolved; in fact, there will not really be religious language as
such. There will be many different religious speech-acts - utterances issued by
many different speakers with a variety of different intentions in a variety of
different circumstances to a variety of different audiences with a variety of
different consequences.
SPEECH-ACTS AND RELIGIOUS LANGUAGE
Speech-act theory has obvious applications within and implications for religious
language; however, unlike the theories of the later Wittgenstein, Austin's theory of
speech-acts has failed to have a widespread effect upon the way in which the
problems of religious language are now addressed. While the reasons for the
relative lack of influence of speech-act theory within the treatment of religious
language remain a matter of speculation, at least a contributing factor must be the
technicality and sometimes tedious detail that the analysis involves. Even given the
limited impact of Austin's notion of speech-acts, it is worth noting here some of

THE PROBLEM OF RELIGIOUS LANGUAGE

75

the attempts to apply a speech-act analysis to religious language and to suggest


some avenues for its further development. 147
The aspect of religious language to which the most obvious and immediate
application of speech-act analysis is possible involves rituals and sacraments. 148
Locutions that occur within the celebration of rituals and sacraments, such as "I
baptize this child ... " or "I absolve you ... " or "I pronounce you husband and
wife,"cannot properly be analyzed as assertions with cognitive content or truth
values. These locutions do not describe some state of affairs or take their meanings
by referring to some empirical fact. The locutions, the speech-acts, are themselves
a constitutive part of the ritual or sacrament itself. For example, part of what makes
baptism what it is requires the proper person with the proper authority in the proper
circumstances at the proper time and to the proper audience to say, "I baptize this
child .... " A similar analysis holds for other locutions that occur during rituals and
sacraments. To be absolved for one's sins simply means, in part at least, to have
the proper authority to say the words "I absolve you ... " in the proper
circumstances and at the proper time. With these examples, the importance of the
act aspect of speech-act analysis becomes apparent. In many circumstances, the
speech-acts are themselves an important, even indispensable, part of the ritual or
sacrament that is at least partially constituted by the appropriate speech-acts. 149
This analysis of the speech-acts of rituals and sacraments leaves untouched
locutions such as "God exists," which have been the focus of most of the attention
concerning questions about the meaning of religious language. Of what help might
speech-act analysis be with such locutions? Are there speech-acts in which the
speaker utters a locution with the illocutionary force of an assertion to make a
statement that is either true or false? Well, one of the acts that a speaker might
perform in issuing a speech-act is to make an assertion, and some of the instances
in which the locution "God exists" is uttered are undoubtedly instances where the
speaker intends to be issuing just such an assertion. However, also undoubtedly,
not all instances in which someone says, "God exists," will constitute an assertion.
Suppose, for example, that I am teaching my class in the philosophy of religion and
I ask the class this question: "What is the critical claim that theists make that
distinguishes theists from nontheists?,,15o A bright and promising student might
answer, "God exists." Now if we take the student's response and conduct a speechact analysis upon it, among the many things that we will conclude is that she is not
making an assertion by saying, "God exists." If there is an assertion at all, it is an
implicit one or an elliptical one that may be something such as "I assert that the
answer to your question is 'God exists"'; however, just the speech-act "God exists"
itself clearly does not have assertive force in this case of affirming the existence of
God.

147 For a detailed treatment of how speech-act analysis might be used in analyzing religious
language, see Donald D. Evans, The Logic of Self Involvement (London: SCM Press, 1963).
148 See James F. Harris, "Speech Acts and God Talk," International Journal for Philosophy of
Religion, Vol. 11, 1980, pp. 170ff.
149 Ibid., pp. 170-71.
150 We might suppose, for the sake of the example, that this is a remedial class in the philosophy of
religion.

76

ANALYTIC PHILOSOPHY OF REUGION

There are other instances where "God exists" is obviously issued with the force
of an assertion. To analyze such speech-acts, according to speech-act theory, we
must examine not just the locution; we must also look at the circumstances in
which the locution is issued, the conventions by which the speaker is able to make
the utterance, the authority of the speaker, the speaker's intentions, the position of
the speaker, the audience to which the utterance is issued, and the ways in which
the audience understands the locution to determine the force of the utterance. 151 It
is not possible to conduct such an analysis here. The traditional problems of
meaning and reference are not avoided in speech-act analysis, but they are
significantly recast. The focus of the analysis is no longer assertions taken as
assertive, declarative sentences with truth values but rather the force with which an
utterance is issued by a speaker. Instead of analyzing religious language or the
abstract, general claim "God exists," speech-act analysis will focus attention upon
a single occasion (with all of the accompanying particularities) where a specific
speaker issues the utterance "God exists." Such a pursuit leads in the direction of
extreme attention to detail and analysis of individual instances of the specific uses
of different locutions and, except for the general methodology itself, away from
general theories or conclusions.
CONCLUSION
The original attack upon the meaningfulness of religious language by the logical
positivists and the more moderate objections to religious language that followed
prompted a plethora of responses by different contemporary philosophers of
religion. There was a significant burst of interest in and concern about religious
language that dominated much of analytic philosophy of religion for a period that
began roughly in mid-twentieth century and lasted for approximately twenty-five
years. The different responses to the challenge of verification and falsification are
spread across a continuum that ranges from the radically conservative, which
attempt to preserve the literal meaningfulness of religious language, to the radically
liberal, which abandon literal significance altogether. As we have seen, both camps
have their defenders and detractors. The fundamental positions and the
fundamental issues at stake appear to have been clearly sorted through, so the last
two decades before the close of the century saw few new attempts to construct
different accounts of religious language. The issue of verification obviously does
not occupy the center stage any longer in contemporary philosophy of religion, but
the developments that have occurred in the treatment of religious language because
of the challenge of verification are now commonly assumed as part of the context
in which other issues and problems in the philosophy of religion are addressed.

151

For a more detailed treatment, see Harris, "Speech Acts and God Talk," pp. 174-79.

III. The Nature of God and Arguments for the Existence


of God

Theism is characterized by belief in a God that possesses a unique set of


characteristics or attributes. The theistic conception of God that is formed by the
combination of these attributes has been troublesome through the centuries and has
given rise to many questions and criticisms - by theists and nontheists alike - both
of the individual attributes and the collective set of attributes. These problems
persisted throughout the twentieth century and intensified in the last few decades.
Some of the issues concerning the attributes are old ones revisited in light of our
changing knowledge of the natural world as a result of the development of science.
Others are logical issues that have been given new "twists" by contemporary
scholars, quite independently of contingent matters. The coherence of the attributes
that comprise the concept of God is fundamental to theism since the viability of
theism must begin with the viability of the concept of God. Sorting through the
difficulties surrounding the traditional attributes of the theistic deity is such a
fundamental problem that Richard Swinburne devotes approximately two-thirds of
what many consider his seminal work, The Coherence of Theism, to an explanation
and defense of the coherence of these attributes. Although there are many
disagreements among theists that need to be explored, I shall take Swinburne's
view as typical of the traditional theistic concept of God: namely, "that there exists
eternally an omnipresent spirit, free, creator of the universe, omnipotent,
omniscient, perfectly good, and a source of moral obligation."!
The purpose of the first part of this chapter is to examine critically both various
new difficulties that have been raised concerning these attributes as well as various
new interpretations of old difficulties that have existed for centuries. Analytic
philosophy, with its central emphasis on conceptual and linguistic analysis, has
resulted in the availability of much finer honed tools of analysis for use by
contemporary philosophers of religion in the process of examining the concept of
God. Although, in several instances, we may agree that the arguments have been
taken to a "higher level," in the end, there is still significant disagreement
concerning the various individual predicates as well as the set of predicates that
I Richard Swinburne, The Coherence of Theism, Revised Edition (Oxford: Clarendon Press, 1993),
p.99.

77

78

ANALYTIC PHILOSOPHY OF REUGION

comprise the concept of God. Since it is impossible to treat all of the controversies
associated with the different traditional attributes of God here, I shall focus instead
on the problems and issues that appear to have dominated the literature in the
second half of the twentieth century. These are the controversies arising from
traditional theistic claims concerning God's omnipotence, omniscience, eternality,
impassibility, and necessary existence. 2
OMNIPOTENCE
Careful, conceptual analysis of the notion of omnipotence has demonstrated the
powerful effects of analytic philosophy - frequently frustrating but often clarifying
as well. The literature resulting from various attempts to define omnipotence and
the criticisms directed at those attempts is extensive, and I can do nothing more
here than suggest some of the major sources. 3 Richard LaCroix has argued that a
general definition of omnipotence is not possible, and George Mavrodes has
responded by attempting to provide such a definition. Both Joshua Hoffman and
Bruce Reichenbach have agreed that Mavrodes's definition is inadequate. Few of
the contemporary discussants in the matter of God's omnipotence, including
Mavrodes, have followed Descartes in maintaining that God can do everything,
including the logically impossible. Most have followed Aquinas in maintaining
that God can do everything that is logically possible. Thus, theism does not require
God be able to create square circles or make the same person both married and a
bachelor at the same time. I shall assume here that logical possibility establishes
the range of discourse for all actions so that we can begin to understand better the
notion of omnipotence by saying that God can do everything that can be done.
Does this amount to a limitation upon what God can do? It does not appear so. We
can certainly say, "Well, God cannot create square circles." However, if we
understand possible actions as delimited by logical possibility, then creating square
circles is not a possible action and thus is not something that God cannot do. Other
difficulties in understanding the notion of omnipotence involve actions that seem
to be prohibited by God's nature. God cannot commit suicide or sadistically torture
young children just for the fun of it given his nature. So a further limitation that
can be placed on the range of possible actions for God is that God can do
everything that is not contrary to God's nature. There are other limitations that
must be added as well. Most have thought that God cannot change the past. God
2 I do not treat here the work of Charles Hartshorne, since his view of the divine attributes is so
intimately connected with his process philosophy. See his Man's Vision of God and the Logic of Theism
(New York: Harper and Row, 1941) and The Divine Relativity: A Social Conception of God (New
Haven, Conn.: Yale University Press, 1948).
3 Several of the exchanges have taken place in Philosophical Studies. See Richard LaCroix, "The
Impossibility of Defining Omnipotence," Philosophical Studies, Vol. 32, 1977, pp. 181-90; George
Mavrodes, "Defining Omnipotence," Philosophical Studies, Vol. 32, 1977, pp. 191-202; Richard
LaCroix, "Failing to Define Omnipotence," Philosophical Studies, Vol. 35, 1979, pp. 219-22; Joshua
Hoffman, "Mavrodes on Defining Omnipotence," Philosophical Studies, Vol. 35, 1979, pp. 311-15; and
Bruce Reichenbach, "Mavrodes on Omnipotence," Philosophical Studies, Vol. 37, 1980, pp. 211-14.
Also see Anthony Kenny, The God of the Philosophers (Oxford: Clarendon Press, 1979), Chapter VII;
Stephen T. Davis, Logic and the Nature of God (Grand Rapids, Mich.: Eerdmans, 1983), Chapter 5; and
Edward R. Wierenga, The Nature of God (Ithaca, N.Y.: Cornell University Press, 1989), Chapter 1. For
Alvin Plantinga's notion of omnipotence, see the discussion of his modal version of the ontological
argument below.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

79

may be able to affect metaphysically contingent states of affairs before and while
they are happening, but once an event has happened and is in the past, even God
cannot make it not happen or change the way in which it happened. While a certain
amount of agreement has developed concerning the need for most of these
qualifications, the exact wording is important and has been the source of frequent
disagreements. 4 The question becomes one of whether or not the notion of
omnipotence, suitably qualified, is a coherent one.
THE PARADOX OF THE STONE
Even when the "restrictions" or qualifications are added to the notion of
omnipotence, many difficult conceptual issues remain. Much of the current
discussion concerning God's omnipotence has been prompted by recently
developed paradoxes that are thought by some to arise from the notion of
omnipotence, even given that God's power is understood as the ability to do
everything that is logically possible and that is consistent with God's nature. One
such paradox is "the paradox of the stone," which poses the question of whether
God can create a stone so heavy that he cannot lift it. 5 Either God can or cannot.
So, if God can create such a stone, there is at least one thing that God cannot do lift it. If God cannot create such a stone, then there is at least one thing that God
cannot do - create it. In either case, it seems impossible for God to do something
that is not obviously logically impossible and that is not obviously inconsistent
with God's nature.
The paradox of the stone has prompted a number of responses that employ
different techniques in efforts to address the apparent paradox.6 One comes from
George Mavrodes, who argues that the supposed paradox is spurious since the
notion of a stone that is too heavy for God to lift is contradictory; hence, the lifting
of such a stone involves a logically impossible state of affairs. A person might
certainly make an object that is too heavy for that person to lift, but on the
assumption that God is omnipotent, the phrase "a stone too heavy for God to lift"
means "a stone that cannot be lifted by Him whose power is sufficient for lifting
anything.,,7 Mavrodes suggests that the contradictory nature of "a stone too heavy
for God to lift" can be seen by comparing the lifting power and the creating power
of God. If one takes the paradox seriously and thinks that it forces some limitation
on God's omnipotence, one might choose to maintain the full power of God's
infinite ability to lift things and restrict God's ability to create by admitting that
God cannot create such a stone. In such a case, has one given up anything?
4 Edward. R. Wierenga provides a summary and discussion of the different versions of the different
qualifications in The Nature oJ God, p. 14-18.
5 Wade C. Savage, "The Paradox of the Stone," Philosophical Review, VoL 76, 1967.
6 In addition to the responses discussed here, see Alvin Plantinga, God and Other Minds (Ithaca,
N.Y.: Cornell University Press, 1967), pp. 168-73; Peter Geach, "Omnipotence," Philosophy, VoL 48,
1973, reprinted as Chapter 1 of Providence and Evil (Cambridge: Cambridge University Press, 1977);
and G. B. Keene, "A Simpler Solution to the Paradox of Omnipotence," Mind, VoL 69, 1960. For a
critical discussion of different responses to the paradox of the stone, see Stephen T. Davis, Logic and
the Nature oJ God (Grand Rapids, Mich.: Eerdmans, 1983), Chapter 5.
7 George Mavrodes, "Some Puzzles Concerning Omnipotence," Philosophical Review, 72, 1963, pp.
221-23. Reprinted in Philosophy oj Religion: Selected Readings, edited by Michael Peterson et aL
(Oxford: Oxford University Press, 1996), to which the page numbers here refer, p. 113.

80

ANALYTIC PHILOSOPHY OF REUGION

Mavrodes thinks not. Since God's power to lift is infinite, then God's ability to
create is still infinite also since God can still create everything that it is possible to
lift; therefore, God is not limited in lifting or creating power. 8
The main difficulty with Mavrodes's solution to the paradox of the stone is that
it proceeds on the basis of the assumption that God is omnipotent; however, as
Richard Swinburne points out, the main thrust of Savage's paradox is to call into
question the coherence of the very notion of omnipotence. 9 Mavrodes's claim that
the phrase "a stone too heavy for God to lift" is self-contradictory hinges upon
God's being omnipotent, but one can hardly use the notion of omnipotence to
resolve the paradox until the paradox has been resolved.
Whereas Mavrodes' s response to the paradox of the stone assumes that God
cannot do logically impossible things, Harry Frankfurt takes a different line in
"The Logic of Omnipotence" by arguing that a response to the paradox does not
require such an assumption.1O If we reject the principle that God cannot do
logically impossible things, then, Frankfurt points out, a solution to the paradox
comes very easily. If God can do logically impossible things, then if God can
create the stone, he can just as easily lift it, that is, doing one logically impossible
thing, he can do another. II Frankfurt's response to the paradox of the stone, which
relies upon the principle that God can do logically impossible things, gives away
too much since it amounts to abandoning logic altogether. If the main force of the
paradox is to attack the notion of omnipotence as logically incoherent, then this
"response" seems to strengthen the paradox instead of solving it by apparently
admitting that the notion of omnipotence is fundamentally paradoxical and even
contradictory.
Richard Swinburne has also responded in length to the paradox of the stone.
After considering several different formulations of the notion of omnipotence,
Swinburne finally settles upon the following (which is [D] in his scheme):
[D]: a person P is omnipotent at a time t if and only if he is able to bring about
any logically contingent state of affairs after t, the description of which does not
entail that P did not bring it about at t. 12
Swinburne relies upon [D] to defuse the paradox. When one introduces temporal
operators into the definition of omnipotence, then one can no longer talk of
omnipotence simpliciter, but one must talk of omnipotence relative to a particular
time. [D] allows logically for a person to be omnipotent at one time and not at
another, and Swinburne maintains that "in the ordinary sense of 'person,'" a person
may choose to exercise his omnipotent ability in such a way to make himself cease

Ibid., p. 114.
Richard Swinburne, The Coherence o/Theism, p. 158.
10 Harry Frankfurt, "The Logic of Omnipotence," The Philosophical Review, Vol. 73, 1964, pp. 26263. Reprinted in Philosophy 0/ Religion: An Anthology, edited by Louis Pojman (Belmont, Calif.:
Wadsworth, 1998), to which the page numbers here refer.
11 Ibid., p. 282.
12 Richard Swinburne, The Coherence o/Theism, p. 156.
8

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

81

to be omnipotent at some future time.13 So, if God chooses to exercise his ability to
make the stone at t, he will cease to be omnipotent after t. While Swinburne sees
his temporal-referenced, modified account of omnipotence as nonrestrictive, it is
questionable that [D] retains a notion of omnipotence that would be acceptable to
traditional theists. In the first place, there is no "ordinary sense of 'person'"
according to which it makes sense to talk of a person being omnipotent at all. It is
only in an extraordinary sense and only in the unique case of God that
omnipotence is even associated with the notion of personhood, which suggests that
there must be something logically unique about the property.
Understood as a power to act, perhaps upon others or oneself, omnipotence is
either an internal relation or an external relation, and given that omnipotence is
predicated of persons only in the case of God, it appears that it must be an internal
relation. However, [D] makes omnipotence an external relation because, according
to [D], God can still be God and yet give up omnipotence. [D] also threatens to
turn all of God's attributes into external relations. If God can exercise omnipotence
at t to give up omnipotence, then presumably God could exercise omnipotence at t
to give up omniscience or goodness or any of his other attributes as well. If he
could not, then he would not be omnipotent at t. Would God still be God after t if
God exercised his omnipotence at t to give up omnipotence, omniscience, and
goodness? Undoubtedly not, and whatever understanding we have of God that
prevents God from remaining God if he gives up omniscience and goodness also
prevents God from remaining God if he gives up omnipotence. 14 The question of
whether God can at one time be omnipotent and at another time not be omnipotent
focuses attention upon another set of problems with the notion of omnipotence.
THE PARADOX OF OMNIPOTENCE
Other contemporary discussion of the notion of omnipotence has been prompted
by what is commonly called "the paradox of omnipotence.,,15 The paradox of
omnipotence arises in connection with the issue of free will and the problem of evil
(which is discussed at length in Chapter VI) and was first explicitly raised by J. L.
Mackie. The paradox is put in deceivingly simple terms: can God make things that
he cannot control?16 Either an affirmative or negative answer to this question
appears to result in serious problems for maintaining the coherence of the notion of
omnipotence. Mackie distinguishes between different levels of omnipotence specifically between first-order omnipotence, which is the unlimited power to act,
and second-order omnipotence, which is the power to determine that powers to act
exist. The paradox arises because if God has second-order omnipotence, then God
13 Ibid., p. 161. For a discussion of temporal considerations in the notion of omnipotence as well as a
distinction concerning different levels of omnipotence ("first-order" and "second-order" omnipotence),
see J. L. Mackie, "Evil and Omnipotence," Mind, Vol. 64,1955, pp. 200-12.
14 There are other considerations involved here as well. For example, powers are sometimes
distinguished from attributes that are distinguished from relations. I maintain that powers and attributes
can best be analyzed in terms of relations, though I have not argued for this position here.
15 The terminology is sometimes confusing here since some scholars refer to the paradox of the stone
as the paradox of omnipotence. I use the two different desiguations here to try and maintain the
distinction between the two problems.
16 Originally in J. L. Mackie, "Evil and Omnipotence," Mind, Vol. 64, 1955. Also in J. L. Mackie,
The Miracle o/Theism (Oxford: The Clarendon Press, 1982), pp. 160ff.

82

ANALYTIC PHILOSOPHY OF REliGION

could determine to limit his own powers and to give powers to other creatures to
act in ways that he could not control. It may seem that it is logically impossible for
there to be beings that an omnipotent being cannot control, but it seems equally
logically impossible for an omnipotent being not to be able to create such beings
using his second-order omnipotence. I?
Mackie uses this paradox to argue that the free will defense is the only plausible
defense for the theist to account for the problem of evil. Otherwise, all evil and all
sin are traced back to God, but if the free will defense forces an affirmative answer
to the paradox of omnipotence, this means that there are creatures whom an
omnipotent God cannot control. I8 On the other hand, Anthony Kenny thinks that
the answer to Mackie's question must be in the negative. His treatment of the
paradox of omnipotence is similar to Mavrodes's treatment of the paradox of the
stone. Whether God's creating things that he cannot control is something that is
logically possible depends upon how the notion of omnipotence is defined. Kenny
apparently thinks that if omnipotence is defined suitably, then such an act on the
part of God will result in no limitation on God's omnipotence; if the act would
result in some limitation, then the act is logically impossible and senseless. 19
Kenny's own attempt at a definition of omnipotence is a reformulation of the one
given by Aquinas and can be developed into Plantinga's notion of maximal
greatness, discussed below: "A being is omnipotent if it has every power which it
is logically possible to possess.,,20 Such a treatment of omnipotence requires
quantifying powers and raises issues that can only be resolved within modal logic
(specifically, the use of contrary-to-fact subjunctive hypotheticals). However, the
advantage of treating omnipotence in terms of the sum of all logically possible
powers instead of the ability to perform all logically possible acts is significant
since it appears obvious that powers are not exhausted by the actual effects of
exercising those powers. Still, as Kenny recognizes, such a treatment must include
the claim that the powers that it is logically possible for God to possess must be
consistent with God's nature. 21 God still cannot possess the power to commit
suicide or to torture human beings just for the fun of it.
OMNISCIENCE
The claim that God is omniscient is traditionally taken to mean that God knows
everything - perfectly, and for all eternity. However, it is now widely agreed
among both theists and nontheists that this general claim requires some
qualification. For example, if we take the claim of omniscience to involve
propositional knowledge, then it appears obvious that God can only know true
17 There has been some discussion of whether this is a genuine paradox or simply a difficult issue
that forces a certain response from a theist depending upon whether or not God may have second-order
omnipotence without ever exercising it. See Mackie, The Miracle of Theism, p. 160.
18 Ibid., pp. 161-62. Alvin Plantinga's free will defense and Mackie's criticisms are discussed at
length in Chapter VI.
19 Anthony Kenny, The God of the Philosophers, pp. 94-95.
20 Ibid., p. 96.
21 Ibid., pp. 97-98. There are other difficulties that cause further qualifications as well. Although
God may possess the power to beget a son, he cannot possess the power to beget my son, and although
God might possess the power to write a book, he cannot possess the power to write the book that I
write.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

83

propositions. This is simply part of what we mean by "knowing." Consider claims


about the past. Even God cannot know that Germany won the Second W orId War
or that George McGovern was elected president of the United States in 1972 or that
Maggie Thatcher is still the prime minister of England at the beginning of the
twenty-first century since these claims are not true. The past is what it is, and if it
is plausible to claim that God knows everything about the past, then this claim
must be taken to mean that God knows everything about the past as it is, that is,
that God knows every true thing about the past.
If it is granted that God's knowledge of propositions is restricted to true
propositions, as an obvious result of the present analysis of what it means to know,
then this may be seen as no real restriction at all since God still knows everything
that is knowable. Can we then characterize omniscience as knowing all true
propositions, or are there other possible restrictions on God's knowledge that may
be in order? Let us reserve momentarily considerations about God's knowledge
concerning propositions about the future, the truth of which may be still in doubt.
What about God's knowledge concerning true propositions about the present and
past that we already know to be true? Consider what we might call "tensed
propositions" or propositions with "tensed facts.'.22 Norman Kretzmann has
identified certain propositions that apparently can only be known to a single person
at a particular time, for example, the case of Jones who is in the hospital suffering
from amnesia. Other people might have identified Jones and know that Jones is in
the hospital while Jones who does not know that he is Jones still knows that he is
in the hospital. 23 And I can know that I finished writing the preceding section of
this chapter yesterday, but I can only know it for today. Some true indexical
propositions are apparently indexed to the knower while others are indexed to time
and cannot be known by just anyone or "for all eternity.',24
Given the foregoing considerations and others concerning different kinds of true
propositions that it seems implausible to claim that God knows "perfectly and for
all eternity," the commonly accepted position regarding omniscience has come to
parallel that concerning omnipotence: namely, that God knows perfectly and for all
eternity everything that it is logically possible for God to know. Such a
qualification need not be understood as a restriction on God's knowledge, and the
point can be put positively. If God knows everything that it is logically possible for
him to know, then there is nothing that it is logically possible for God to know that
he does not know.
OMNISCIENCE AND FREE WILL
More interesting and more perplexing problems develop from consideration of
God's possible knowledge of the future. If God knows everything about the future
- perfectly and for all eternity - then what are the ramifications of such divine
22 See Richard Swinburne, "Tensed Facts," American Philosophical Quarterly, Vol. 27, 1990, pp.
117-30.
23 Nonnan Kretzmann, "Omniscience and Immutability," Journal of Philosophy, Vol. 63, 1966.
24 For responses to Kretzmann, see Nector-Neri Castaneda, "Omniscience and Indexical Reference,"
Journal of Philosophy, Vol. 64, 1967, pp. 203-210, and Swinburne, The Coherence of Theism, 167-72.
Castaneda argues that there is a propositional content to indexicals that can be known by other people,
but Swinburne denies this.

84

ANALYTIC PHILOSOPHY OF REUGION

foreknowledge for human free will? Normal, contingent events are such that we
take them to be the kinds of events that may happen or may fail to happen. For
example, I am now writing this chapter of this book, but I might have failed to
write it had I been hit by a truck or contracted some terminal disease earlier.
However, if God knew from all eternity that I would be writing this chapter of this
book on this very day, and if God knows only true propositions, then it has been
true for all eternity that I would be doing what I am doing at this exact moment.
And if it has been true for all eternity that I would be doing what I am doing at this
exact moment, then it appears that I could not have done otherwise by choosing to
go fishing rather than writing this chapter in this book at this exact moment on this
exact day. Human free will thus appears to be inconsistent with perfect divine
foreknowledge.
PIKE'S ARGUMENT
This argument has been put very formally and convincingly by Nelson Pike, and
his way of formulating the matter has prompted a plethora of responses and created
the framework within which most of the critical discussion has taken place. It is
thus worthwhile to consider carefully Pike's argument:
1.

"God existed at t1" entails "If Jones did X at t2, God believed at t1 that Jones
would do X at t2."

2.

"God believes X" entails "X is true."

3. It is not within one's power at a given time to do something having a


description that is logically contradictory.
4. It is not within one's power at a given time to do something that would
bring it about that someone who held a certain belief at a time before the
time in question did not hold that belief before the time in question.
5. It is not within one's power at a given time to do something that would
bring it about that a person who existed at an earlier time did not exist at
that earlier time.
6. If God existed at t1 and if God believed at t1 that Jones would do X at t2,
then if it was within Jones's power at t2 to refrain from doing X, then (1) it
was within Jones's power at h to do something that would have brought it
about that God held a false belief at t], or (2) it was within Jones's power at
t2 to do something that would have brought it about that God did not hold
the belief he held at t], or (3) it was within Jones's power at t2 to do
something that would have brought it about that any person who believed at
t1 that Jones would do X at t2 (one of whom was, by hypothesis, God) held a
false belief and thus was not God - that is, that God (who by hypothesis
existed at t1) did not exist at t1.
7.

Alternative 1 in the consequent of item 6 is false (from 2 and 3).

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

8.

Alternative 2 in the consequent of item 6 is false (from 4).

9.

Alternative 3 in the consequent of item 6 is false (from 5).

85

10. Therefore, if God existed at t 1 and if God believed at t 1 that Jones would do
X at t2, then it was not within Jones's power at t2 to refrain from doing X
(from 6 through 9).
11. Therefore, if God existed at t" and if Jones did X at t2, it was not within
Jones's power at t2 to refrain from doing X (from 1 and 1O)?5
Several people have responded to this argument in various ways. This is a
dilemma in which human free will is as important an ingredient as is divine
foreknowledge; therefore, the dilemma may be dissolved depending upon the
position one takes regarding free will. One might choose to give up the notion of
human free will altogether, in which case the dilemma of God's foreknowledge is
obviously not a dilemma involving free will any longer; however, just as
obviously, such a denial raises severe difficulties involving other attributes of God
and other theistic beliefs about the purpose of human life and the possibility of an
eternal life based on reward or punishment. Similarly, one might take a
compatabilist position regarding free will and determinism, in which case the fact
that a person's action x is causally determined is compatible with that person being
free in respect to x and with that person being morally responsible for x. 26
Alternatively, one might place some limitations upon our understanding of
omniscience and simply deny that knowledge of the future is possible, as Richard
Swinburne does. 27 Swinburne maintains that while the future is under God's
control, God does not have knowledge concerning the future. He defines his more
limited account of omniscience as follows: "A person P is omniscient at a time t if
and only if he knows every true proposition about t or an earlier time and every
true proposition about a time later than t which is true of logical necessity or which
he has overriding reason to make true, which it is logically possible that he
entertains then.,,28 Swinburne's comments help to clarify this limitation. Since God
is omnipotent, he might choose to change the universe in radical ways (for
example, changing the natural laws of the universe) or even destroy it; however, if
God is free, then he will not know in advance all of the ways in which he might
affect the future. Thus, to paraphrase Swinburne, what God knows, he does not
control (that is, the past), and what God controls, he does not know (that is, the
future). Swinburne qualifies what God may know about the future in cases where
God has an "overriding reason" for determining or controlling a future event;
25 Nelson Pike, "Divine Omniscience and Voluntary Action," Philosophical Review, Vol. 74, 1965,
pp. 27-46. Reprinted in Philosophy of Religion: Selected Readings, edited by Michael Peterson, William
Hasker, Bruce Reichenbach, and David Bassinger (Oxford: Oxford University Press, 1996) to which the
page numbers here refer, pp.117 -18.
26 See Anthony Kenny, "Divine Foreknowledge and Human Freedom," in Aquinas: A Collection of
Critical Essays, edited by Anthony Kenny (Garden City, N.Y.: Anchor Books, 1969), pp. 255-270, as
well as The God of the Philosophers, Chapter V.
27 Richard Swinburne, The Coherence of Theism, pp. 179-83.
28 Ibid., pp. 180-81.

86

ANALYTIC PHILOSOPHY OF REUGION

however, this loophole is certainly big enough to generate such a large number of
exceptions as to undermine the entire line of argument. Surely, one might think,
God has a really good reason for acting as he does anytime God acts. What is the
alternative? That God acts capriciously? God must always have good reasons for
acting as he does; the debate will then tum to how much of a good reason must a
good reason be to become an "overriding" reason. Any response then appears to
become arbitrary and ineffective.
The intractable form of the dilemma between divine foreknowledge and free
will assumes an incompatibilist position regarding free will and human actions,
that is, it requires the incompatibility of determinism and moral responsibility and
assumes that free will exists. Among free-willists, there has been near-universal
agreement that for a person S to exercise free will and to have moral responsibility
with respect to some action x at tJ, it must have been possible for S to do other than
x at tl' This principle - that a person is morally responsible for x only if that person
could have done other than x - has become known as the principle of alternative
possibilities (PAP). Though PAP has great intuitive appeal, and though it has been
a mainstay of incompatabilists since the time of Aristotle, PAP has recently been
assailed by Harry Frankfurt, and its status is now much more uncertain than it has
been in the past. 29 While the debate continues among moral philosophers and
metaphysicians concerning the status of PAP, most attempts to resolve the
dilemma of divine foreknowledge and human free will assume that PAP is
preserved. 3o
OCKHAMIST SOLUTIONS TO THE DILEMMA OF DIVINE
FOREKNOWLEDGE
Most of the responses to the dilemma of divine foreknowledge and most of the
critical discussion has centered on the possibility of a general Ockhamist solution
that denies that God's foreknowledge is a part of the "accidental necessity" of the
past. 31 Since it is impossible to treat all of the different responses to Pike in detail, I
will focus upon just the most important considerations here. 32 Further, detailed
treatment of these problems must involve an examination of time and particularly
29 See Harry Frankfurt, "Alternative Possibilities and Moral Responsibility," Journal of Philosophy,
Vol. 66, 1969, pp. 829-39. For a discussion of Frankfurt's attack on PAP and various responses as well
as a defense of PAP, see Laura Waddell Ekstrom, Free Will: A Philosophical Defense (Boulder, Colo.:
Westview Press, 1999).
30 See Linda Zagzebski, "Foreknowledge and Human Freedom," in A Companion to Philosophy of
Religion, edited by Philip L. Quinn and Charles Taliaferro (Oxford: Blackwell Publishers, 1997), pp.
294-96. Zagzebski accepts Frankfurt's rejection of PAP but argues that PAP is not the important issue
for the problem of divine foreknowledge.
31 Marilyn Adams, "Is the Existence of God a Hard Fact?" Philosophical Review, Vol. 76, 1967;
Joshua Hoffman, "Pike on Possible Worlds, Divine Foreknowledge, and Human Freedom,"
Philosophical Review, Vol. 88, 1979; Anthony Kenny, The God of the Philosophers, pp. 51-87; John
Martin Fischer, "Freedom and Foreknowledge," Philosophical Review, Vol. 92, 1983; Joshua Hoffman
and Gary Rosenkrantz, "Hard and Soft Facts," Philosophical Review, Vol. 93, 1984; William P. Alston,
"Divine Foreknowledge and Alternative Conceptions of Human Freedom," International Journal for
Philosophy of Religion, Vol. 18, no. 1, 1985.
32 For a thorough discussion of William of Ockham and the contemporary Ockhamists, see Linda
Trinkaus Zagzebski, The Dilemma of Freedom and Foreknowledge (Oxford: Oxford University Press,
1991), Chapter 3.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

87

the past, analyses of various notions of necessity, and a theory for sUbjunctive
conditionals and counterfactuals.
William of Ockham attempted to resolve the dilemma of God's foreknowledge
by borrowing and adapting Aristotle's distinction between potency and actuality.
At whatever time a particular determinate truth x becomes true, until that moment
in time, there was some potency for not-x. Though the past is necessary and x is
determinedly true for any moment in time, since the past still allows for potency
for not-x until the moment that x becomes true, the past does not necessitate the
truth of x. Similarly, because there is potency for not-x until the moment that it
becomes true, there is also potency for God's divine foreknowledge of x to be false
- even though God has divine foreknowledge of x for all time and it is
deterrninately true that God foreknew x for all time. Thus, Ockham denies that the
past determinate truth of God's foreknowledge of future contingent events is part
of the "accidental necessity of the past" that exerts some kind of determination
upon those future events. 33 Linda Zagzebski summarizes the point: "So even
though the truth of a future contingent proposition cannot change and the state of
affairs expressed by the proposition will obtain, there is still potency for the
opposite until such time as it becomes actual, and necessity goes with actuality
[that is, the loss of potency results in necessity].,,34 So, for Ockham, God's divine
foreknowledge is compatible with the contingency of future events and with free
will.
HARD AND SOFT FACTS
Many of the attempts by contemporary Ockhamists to exclude God's divine
foreknowledge, one way or another, from the accidental necessity of the past make
use of the distinction between "hard" and "soft" facts that was first introduced by
Nelson Pike in his response to John Turk Saunders. The distinction was
popularized by Marilyn Adams and is now commonly associated with her work. 35
Adams maintains that the dilemma of divine foreknowledge is really a dilemma for
human free will only if God's knowledge of future contingencies is a "hard fact" in
the past, and she denies that this is the case. Some facts in the past extend into the
future (for example, the fact that exists now and has existed for some time in the
past that I will finish writing this book in December), other facts are completely
about the past. She defines a hard fact as follows: "Statement p expresses a 'hard'
fact about a time t =df. p is not at least in part about any time future relative to t ."
And "a Statement p is at least in part about some time t" is defined so that "the
happening or not happening, the actuality or nonactuality of something at t is a
necessary condition of the truth of p.,,36 Thus, facts that are at least partially about
some future time are "soft facts." Only hard facts are completely about the past and
genuinely in the past and thus a part of the accidental necessity of the past. So, for
33 This summary is a synthesis of the treatments found in Zagzebski, ibid., pp. 68-70, and Marilyn
Adams, William Ockham (Notre Dame, Ind.: University of Notre Dame Press, 1987), Chapter 27.
34 Zagzebski, The Dilemma of Freedom and Foreknowledge, p. 18.
35 Nelson Pike, "Of God and Freedom: A Rejoinder," Philosophical Review, Vol. 75, 1966, pp. 36979; John Turk Saunders, "Of God and Freedom," Philosophical Review, Vol. 75, 1966, pp. 219-225;
and Marilyn Adams, "Is the Existence of God a Hard Fact?"
36 Marilyn Adams, ibid., pp. 493 and 494.

88

ANALYTIC PHILOSOPHY OF RELIGION

example, "I will finish writing this book in December" is obviously a proposition
about December - some months in the future at the moment - and therefore is a
soft fact, though the fact that I will finish writing the book in December exists now
and has also existed before this moment. The crucial point is that God's
foreknowledge is not considered as a hard fact by Adams, and thus, God's
knowledge is not a part of the accidental necessity of the past. If God foreknows at
t) that I will finish writing this book in December, then the truth of the proposition
that constitutes God's foreknowledge, "Harris will finish the book in December,"
obviously depends upon something that will happen some months in the future,
after tJ, and is a soft fact by her account.
Adams thus follows Ockham by arguing that since God's foreknowledge is a
soft and not a hard fact (in fact, God's existence is a soft fact for Adams since that
existence is eternal), God's knowledge is not a part of the accidental necessity of
the past. Since soft facts are not simply and completely about the past but "extend"
into the future, it is still possible to affect the outcome of such facts without raising
the controversial matter of "changing the past" (Pike's premise 4). William Rowe
gives a very illustrative and persuasive example which illustrates the difference
between hard and soft facts. Consider the difference between
fact, In 1941 Japan attacked Pearl Harbor.
fact2 In 1941 a war begins between Japan and the United States that lasts five
years. 37
As a reader considers these two facts sometime after the year 2000, both facts
are simply and completely about the past; however, what if we imagine
considering these facts during the year 1943? Considered in 1943, fact) will be
simply and completely about the past while fact2 will still be about the present and
the future until 1945. The importance of this distinction for the debate concerning
divine foreknowledge and human freedom is that, until 1945, it would still be
within the power of different individuals to do something to alter the fact - either
by extending the war or by ending it sooner. Perhaps decisions made by various
people on either side could have altered the outcomes of various battles, such as
the pivotal naval battle of Midway, in such a way to alter fact2. Likewise, though
God might have foreknown from all eternity that Harris will finish writing this
book by December, various things that I do might affect the outcome of that fact
without affecting something about the past.
While some have embraced the notion of hard facts,38 others have attacked and
ultimately denied the distinction between hard and soft facts. John Martin Fischer,
for example, has argued that Adams's way of treating the distinction collapses all

37 William L. Rowe, Philosophy of Religion: An Introduction (Belmont, Calif.: Wadsworth, 1978),


p.l63. [Though the war between Japan and the United States ended in 1945 and lasted four years, I use
Rowe's original example to be true to his text.]
38 For example, Alvin Plantinga, "On Oekham's Way Out," Faith and Philosophy, Vol. 3, no. 3,
July 1986, pp. 235-69.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

89

facts into soft facts,39 and Zagzebski has raised the objection that the distinction
between hard and soft facts is simply a nominal, artificial one rather than a
genuine, metaphysical one. 40 William Hasker has argued convincingly that the
crucial facts about God's prior beliefs are hard instead of soft facts. 41 1 cannot
examine these criticisms and responses in detail here, but 1 will briefly add an
additional criticism of Adams's treatment of the distinction between hard and soft
facts where it concerns God's beliefs and knowledge. Consider God's beliefs
concerning facti and fact2 at times before 1941. To describe such beliefs, we would
simply say, "God believed before 1941 that Japan would bomb Pearl Harbor in
1941," and, "God believed before 1941 that a war would begin between Japan and
the United States that would last for five years." Call these two facts about God's
knowledge facti and fact2. respectively, and notice that facti and fact 2 are facts
about facti and fact2. respectively. Following common usage, we can say that the
propositions expressing facti and fact2 are extensional, and the propositions
expressing face and fact2 are intensional. It is well known that the truth of a
proposition about an intensional claim is independent of the truth of whatever
constituent propositions it may contain; for example, the truth of "Mary believes
that p" is independent of the truth or falsity of p. Something of the same nature
appears to characterize hard and soft facts in an intensional context. There appear
to be no difficulties in regarding facti as a hard fact, but why is fact 2 not a hard fact
also by parallel reasoninf42 Even if fact2 is taken noncontroversially as a soft fact,
that does not make fact a soft fact. To reason otherwise turns not just all facts
about God into soft facts but all beliefs and claims to knowledge about the future
by human beings into soft facts as well. Consider a moral and legal defense that
takes the form "I can't be blamed for killing Jones at t2 because although 1 knew at
tl that pulling the trigger would fire the gun and kill Jones, at tl the gun had not
fired." Denying that beliefs of the type in fad and fact 2 are hard facts undermines
our common understanding of intensional states and human freedom and moral
responsibility. Therefore, claims regarding whether God's intensional states of
belief and knowledge are hard or soft facts seem to be independent of whether the
content of those intensional states are hard or soft facts. If God believed before
1941 that a war would begin between Japan and the United States that would last
for five years (fact2), then it seems that such a war had to occur (fact2) and that no
human action would prevent or alter the truth of fact2.
Other responses to the dilemma of divine foreknowledge that have attempted to
preserve both God's omnipotence and human free will (and PAP) have followed
Molina, using his suggestion of "middle knowledge," that is, God's perfect
knowledge of what every human free agent would freely choose in every possible
39 John Martin Fischer, "Introduction: God and Freedom," iu God, Foreknowledge, and Freedom,
edited by John Martin Fischer (Stanford, Calif.: Stanford University Press, 1989), pp. 35-36. Also see
Fischer's "Freedom and Foreknowledge" (Chapter 4) in the same volume, originally published in The
Philosophical Review, VoL 92, no. 1, 1983, pp. 67-79. For a response to Fischer, see Joshua Hoffman
and Gary Rosenkrantz, "Hard and Soft Facts," The Philosophical Review, VoL 93, no. 3, 1984, pp. 41934, which is reprinted also in God, Foreknowledge, and Freedom, Chapter 7.
40 Zabzebski, The Dilemma of Freedom and Foreknowledge, pp. 74-76.
41 William Hasker, God, Time, and Knowledge (Ithaca, N.Y.: Cornell University Press, 1989),
pp.75-95.
42 I am here assuming that not all facts about God are soft simply because God is eternal.

90

ANALYTIC PHILOSOPHY OF REliGION

situation. Middle knowledge is very special since it is contingent knowledge about


contingent states of affairs in the world, and God supposedly has middle
knowledge before creation and uses middle knowledge in his creation of the
universe. Alvin Plantinga and Alfred Freddoso have both defended a solution to
the dilemma of divine foreknowledge using Molina's notion of middle
knowledge. 43 Plantinga's lengthy response to Pike attacks Pike's premise 6 as
being nonexhaustive of the different ways in which a person may refrain from
doing an action. 44
Contemporary treatments of middle knowledge become involved in the analysis
of subjunctive conditionals, counterfactual conditionals, and possible worlds. A
defender of middle knowledge must defend "counterfactuals of freedom," where
presumably agents are free to make certain counterfactuals about their actions true.
(For example, if I were offered the opportunity to ride on the space shuttle, I would
go.) Criticisms of middle knowledge cover a wide spectrum. 45 Especially
noteworthy is William Hasker's detailed and convincing argument for the rejection
of middle knowledge. Hasker argues that there are no true counterfactuals of
freedom since no genuinely free human agent has it within his power to make a
counterfactual of freedom true. 46
Finally, others, trying to reconcile and preserve both divine foreknowledge and
human free will, have followed Boethius by denying the usual asymmetry between
the past and the future involving God's knowledge. For example, Eleonore Stump
and Norman Kretzmann maintain that God's knowledge is timeless - outside of
time altogether - and they deny that an eternal being can have foreknowledge at
al1. 47 On this account, an eternal God's knowledge of an event is simultaneous with
that event so that the way of constructing the dilemma of divine foreknowledge in
terms of God's knowing in the past some true proposition about the future would
not be permitted. 48 The relationship of God's knowledge to time leads to a
discussion of the attribute of eternality.
ETERNALITY
The transcendent otherness of God removes God in various ways from this
world and the natural laws and metaphysical features of this world. One of the
most perplexing issues involving God's transcendent nature is how God is related
to time. One of the difficulties involved in sorting through the different positions
43 See Alvin Plantinga, The Nature of Necessity (Oxford: Oxford University Press, 1974), Chapter 9,
and Alfred Freddoso's Introduction in On Divine Foreknowledge: Part N of the Concordia, Luis de
Molina (Ithaca, N.Y.: Cornell University Press, 1988).
44 Alvin Plantinga, God, Freedom, and Evil (New York: Harper & Row, 1974), pp. 66-73.
45 For a critical discussion of these criticisms, see Zabzebski, The Dilemma of Freedom and
Foreknowledge, pp. 14lff.
46 William Hasker, God, Time, and Knowledge, pp. 39-52. For a lengthy response to Hasker and a
defense of Plantinga, see Edward R. Wierenga, The Nature of God, pp. 150-160. Anthony Kenny also
attacks Plantinga's notion of middle knowledge in The God of the Philosophers, pp. 67ff.
47 Eleonore Stump and Norman Kretzmann, "Eternity," Journal of Philosophy, Vol. 78, 1981.
Reprinted in The Concept of God, edited by Thomas V. Morris (Oxford: Oxford University Press,
1987), p. 247.
48 Others have rejected this way of resolving the dilemma. Kenny calls the timelessness of God
"incoherent" while Swinburne denies that timelessness is consistent with other theistic beliefs about
God.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

91

on this issue is that different people have used terms such as "eternal," "timeless,"
and "everlasting" in different ways with different meanings. Considering the
different meanings for "eternal" is helpful for identifying and distinguishing the
way in which scholars have divided on the issue of God and time. According to the
Oxford English Dictionary, the word 'eternal' has two similar but different
meanings: "infinite in past and future duration; without beginning or end; that
which has always existed and always will exist" and "not conditioned by time; not
subject to time relations." Most medieval scholars - including Anselm, Augustine,
Aquinas, and especially Boethius - took God's relationship to time to be
characterized by the second definition given of "eternal," according to which God
is "not conditioned by time; not subject to time relations." God is regarded as thus
"outside" of time and "timeless," and, as such, God does not have a beginning or
end or a history or a biography.
ET-SlMULTANEITY
Two contemporary defenders of a view of God's timelessness, which is based
on Boethius, are Eleonore Stump and Norman Kretzmann. Stump and Kretzmann
pay particular attention to the problem of simultaneity that has plagued the
Boethius view, according to which all time exists "all at once" for God. Several
critics have thought that if God is timeless and if all time exists at the same time,
then this raises the severe problem that all events in human history occur
simultaneously. The result, to paraphrase Anthony Kenny, is that Nero fiddles and
Rome burns while I am writing this book.49 Stump and Kretzmann attempt to
address such criticism by developing their notion of ET-simultaneity - a notion
that is derived from the lesson learned from Einstein's theory of relativity that
simultaneity must be specified relative to a frame of reference. Stump and
Kretzmann take the normal temporal order of things and God's divine, atemporal
order of things as requiring references to two distinct frames of reference for
specifying simultaneity. They define ET-simultaneity as follows:
Let "x" and "y" range over entities and events. Then:
(ET) for every x and every y, x and y are ET -simultaneous if
(i) either x is eternal and y is temporal, or vice versa; and
(ii) for some observer, A, in the unique eternal reference frame, x and y
are both present - i.e., either x is eternally present and y is observed as
temporally present, or vice versa; and
(iii) for some observer, B, in one of the infinitely many temporal reference
frames, x and y are both present - i.e., either x is observed as eternally
present and y is temporally present, or vice versa. 50
Such an understanding of the nature of God seems the best way to protect the
integrity of God's other metaphysical attributes. If God is "unconditioned" by time,
49

39.

Anthony Kenny, The God of the Philosophers (Oxford: Oxford University Press, 1979), pp. 38-

50 Eleonore Stump and Norman Kretzmann, "Eternity," Journal of Philosophy, VoL 78, 1981, pp.
429-58. The pages numbers here refer to the reprint in The Concept of God, edited by Thomas V.
Morris (Oxford: Oxford University Press, 1987), pp. 231.

92

ANALYTIC PHILOSOPHY OF RELIGION

God does not change by growing old or weak; God's immutability, simplicity, and
omnipotence are thus preserved. 51 If God is timeless, then God's experience is also
unconditioned by time; so God does not deliberate before coming to know or
acting. Another significant advantage of viewing God as outside and independent
of time is that such a view is taken by many to resolve the dilemma between divine
foreknowledge and human free will that arises because of a linear understanding of
time and the asymmetry of the past to the present and the future, that is, God
supposedly knows in the past what one will do now or in the future. If God is
timeless and God's knowledge is timeless, then the dilemma between divine
foreknowledge and free will cannot be constructed in its usual way, that is., God's
past knowledge does not conflict with future actions by some agent. But these
advantages come at a significant price. As Stump and Kretzmann acknowledge,
God cannot really be said to foreknow anything. According to the definition of ETsimultaneity, no event occurs after God's awareness of that event relative to God's
frame of reference, and "future" events are future only relative to our temporal
frame of reference. 52
Several critics have raised different objections to the notion of ET-simultaneity,
including Zagzebski 53 and Hasker. In addition, others have objected that the very
notion of timelessness, which Stump and Kretzmann are trying to explain with
their definition of ET-simultaneity, is itself incoherent. Hasker, who defends his
own notion of timelessness, specifically asks,
What are we to make of the clause "x and yare both present" when one of the
relata is temporal and the other is eternal? For an entity to be temporal, after
all, just is for that entity's existence to be spread out in a temporal sequence but in eternity, nothing exists in temporal sequence, how can a temporal "y" be
present in eternity? Again, to be eternal means to exist in a "total present"
without temporal sequence, so how can an eternal "x" be present in time ?54
It is possible to gain some conceptual purchase on the notion of timelessness
and ET -simultaneity by considering some of the physics and psychology that
underlie human experience. 55 The word "present," for example, is commonly
understood as the cutting "knife-edge" of the future as it becomes past, and it is
commonly agreed that human experience is only of the present - the "here and
now." Some events, which are yet to become actual, are "out there," in the future,
and we cannot experience them until they "get here," to the present. Such a view of
human experience is naive and misleading. There is no support for the notion of an
unextended, absolute present in empirical psychology; neither is it supported by
philosophical analysis. In attempting to explain the smooth "flow" of human
51 For a critical discussion of the relationship of timelessness to these attributes, see Nelson Pike,
God and Timelessness (New York: Schocken Books. 1970), Chapter 3, and Richard Swinburne, The
Coherence of Theism (Oxford: Clarendon Press, 1993), pp. 219ff.
52 Stump and Kretzmann. ibid., p. 247.
53 Zagzebski, The Dilemma of Freedom and Foreknowledge, p. 41ff.
54 Hasker, God, Time, and Knowledge, p. 164.
55 Much of the following is drawn from James F. Harris, "An Empirical Understanding of
Eternality," International Journalfor Philosophy of Religion, Vol. 22, 1987, pp. 165-83.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

93

experience, William James used the notion of an extended present moment - the
specious present (a notion he borrowed from Charles Peirce). Human experience is
continuous and smoothly flowing because our experience is not of discrete instants
of time but of extended periods of time. Each experience involves what James
called "the echo of the objects just past" and "the foretaste of those just to
arrive.,,56 The specious present is not the unextended "knife edge" of the absolute
present moment but more like the gentle peak of an arc between future and past.
Consider, for example, how we experience a musical melody. We do not
experience each note separately and distinctly from the others but a "flow" of the
notes as they are related to one another. Exactly how far the specious present
moment is extended along the curve of the arc is a function of both human
anatomy and the different circumstances involving the particular experience.
Psychological experiments indicate that most people experience a specious present
in a range of from one half second to six to twelve seconds. Various drugs and
conditions such as fatigue and illness can affect one's experience of the specious
present (which may explain the reported mystical experiences of some ascetics).57
The important point is that there is no theoretical limitation on the specious
present, and if it is easy to imagine other species of animals or intelligent aliens
with greatly expanded specious present moments, then it is possible to imagine a
divine being with an infinitely extended specious present moment. Such an
understanding provides one way of providing some plausible conceptual
understanding of how temporal events might be eternally related. The specious
present also provided an empirical grounding for Samuel Taylor Coleridge's poetic
waxing about divine timelessness in his Omniania:
It is not impossible that to some infinitely superior being the whole universe

may be as one plain, the distant between planet and planet being only as the
pores in a grain of sand, and the spaces between system and system no greater
than the intervals between one grain and the grain adjacent.
If similarly long extended periods of time are thought of as compressed into one
eternally extended specious present, then all time may be understood as
experienced as "now" by God.

TIMELESSNESS, DIVINE PROVIDENCE, AND DIVINE PASSIBILITY


Even if it is possible to provide some conceptual explanation of the notion of
timelessness, what does such an understanding imply for other common theistic
beliefs about the nature of God? Several critics have objected that attributing
timelessness to God undermines the common belief that God can be and has been
involved in human affairs and human history. This objection has been put in
different forms. For example, Nelson Pike has objected that if God is timeless, then
God cannot be a person since the notion of a timeless person is incoherent. 58
William James, Principles of Psychology (New York: Henry Holt, 1893), Vol. IV, p. 606.
See Harris, "An Empirical Understanding of Eternality," for more detailed discussion of the limits
of the specious present, pp. 177ff.
58 Nelson Pike, God and Timelessness, Chapter 7.
56
57

94

ANALYTIC PHILOSOPHY OF RELIGION

Leaving aside for the moment the question of whether God is a person, let us
consider other objections raised against regarding God as timeless on the grounds
that divine providence and divine passibility are both undermined. For example,
Richard Swinburne claims that attributing timelessness to God "seems to contain
an inner incoherence and also to be incompatible with most things which theists
ever wish to say about God."s9 Perhaps foremost in those things that theists may
want to say about God is that God is both involved in and affected by human
affairs. If God is "outside" time, then it seems normal temporal predicates cannot
meaningfully be applied to God, and God's immutability must be understood in a
strong sense according to which God is unchanged by any temporal events. If God
is unchanged by any temporal events, then it becomes increasingly conceptually
difficult to make sense of divine agency (whatever form it might take) and of
divine passibility. Using the distinction between internal and external relations is
helpful in understanding the problem of the effect of timelessness on God and the
distinction between "real" change and "accidental" change. 6o Internal relations are
those that are essential to the actual nature of a thing while external relations are
those that are nonessential. For me simply to observe the flower in the crannied
wall, for example, does not change the essential nature of the flower, but for me to
cut the flower and kill it does; thus, observing the flower is an external relation
while killing it is an internal one. 61 Attributing timelessness to God appears to
make all of God's relations external ones.
If all of God's relations are external, how does one make sense of the act of
creation and specific accounts of such actions as appearing to Moses or Paul? And
what of God's reaction to the suffering of an innocent child? About God's actions,
one might say that since God is timeless and immutable in a strong sense that God
has always intended to do whatever it is that he does so he is unaffected by what it
is that he does (though human beings may obviously be affected). Whatever sense
we may then make of speaking of "reasons" for God's actions, they would be
independent of human beings and the affairs of human beings. Thus, when God
appeared to Moses or to Paul, God did not change, but Moses and Paul did. 62 Such
a view, though, raises serious questions about traditional theistic beliefs. For
example, it would mean that God had intended from all eternity to deliver the Jews
from captivity in Egypt, quite independently of anything that anyone - including
Moses and Pharaoh - did or said. For Christians, it seems impossible to make sense
of God's acting to redeem the world by sending Jesus Christ to earth if God had
intended to do so for all eter~ity re~'1ardless ?f w~at ~nyone said or did or w~at
happened throughout human hIstory. If God s bemg tImeless means that creatIOn
is not an internal relation of God for Muslims, and if the deliverance from Egypt is
c

Swinburne, The Coherence of Theism, p. 228.


Using this distinction is more helpful, I think, than using "the Cambridge Criterion" introduced by
Peter Geach, "What Actually Exists," Proceedings of the Aristotelian Society, Supplemental Volume,
Vol. 42,1968, pp. 7-16.
61 It should be noted that several have criticized and rejected this distinction.
62 For a discussion of the notion that God has eternal intentions, see Swinburne. The Coherence of
Theism, pp. 221-22.
63 Dun Scotus notwithstanding. Although Scotus held that the incarnation was intended by God from
all eternity and not as a result of original sin, this certainly has never been and is not now the dominant
view.
59
60

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

95

not an internal relation of God for Jews, and if the redemption of the world through
Jesus Christ is not an internal relation of God for Christians, then such a view of
God hardly seems to capture what theists have meant by "God."
GOD AS EVERLASTING
Such problems have led many scholars to reject the claim that God is timeless
and "outside" time in favor of an understanding of God as everlasting - an
understanding of God based upon the first definition of "eternal" from the OED,
given above. According to this view, God is everlasting - infinitely extended in
time without beginning or end - but definitely and essentially in time.
Considerations concerning ET -simultaneity and an infinite specious present
moment notwithstanding, making sense of divine providence and the more
fundamental notion of divine agency when God is regarded as timeless and outside
of time seems to be an intractable problem and a telling one against such a view.
Primarily for this reason, several contemporary thinkers have defended the claim
that God is eternal in the sense of being everlasting (infinitely extended in time),
including Richard Swinburne, Nicholas Wolterstorff, and Stephen Davis. 64
Swinburne and Davis argue that the notion of timelessness is incoherent, and
Wolterstorff and Davis argue that a timeless God cannot be the Christian God. In
making the latter point, W olterstorff focuses upon the role of God as Redeemer and
argues that God as Redeemer must change. 6s Davies focuses upon the role of God
as Creator. 66 In both roles, it appears that one must be able to apply temporal
predicates meaningfully to God, and, as I have argued above, if God is timeless,
such roles and such predicates cannot be internal predicates of God. Swinburne,
Wolterstorff, and Davis certainly seem correct so far as traditional Christian
theology is concerned. For God to be the Creator or the Redeemer requires certain
temporal predicates to be essential, internal predicates of God.67
DIVINE P ASSIBILITY AND IMPASSIBILITY
The question of whether God has both internal and external relations also serves
to focus the discussion of God's passibility or impassibility. Exactly how are we to
understand God's reaction to human suffering and animal suffering? Given God's
omniscience, God must be aware of human and animal suffering, but is God
affected in any way by this knowledge? Does God suffer as a result of human and
animal suffering? In terms of the history of Christian theology, the neo-Platonist
64 Richard Swinburne, The Coherence of Theism, pp. 217ff; Nicholas Wolterstorff, "God
Everlasting," God and the Good: Essays in Honor of Henry Stob, edited by C. Orlebeke and L. Smedes
(Grand Rapids. Mich.: Eerdmans. 1975), pp. 181-203. Reprinted in Philosophy of Religion: Selected
Readings, edited by Michael Peterson et aI., pp. 125-33; Stephen T. Davis, "Temporal Eternity," from
Logic and the Nature of God (Grand Rapids. Mich.: Eerdmans, 1983). Reprinted in Philosophy of
Religion: An Anthology, edited by Louis Pojman (Belmont, Calif.: Wadsworth, 1998), pp. 235-42. Page
references to both Wolterstorff and Davis refer to the reprinted versions.
65 Wolterstorff, "God Everlasting," pp. 125-26.
66 Davis. "Temporal Eternity," pp. 237ff.
67 If this is true, one might wonder why Aquinas, Boethius, and others have taken God to be
timeless. To this question I have no definite answer, but Swinburne's suggestion that attributing
timelessness to God in Christian theology resulted from neo-Platonism seems very plausible. See
Swinburne, The Coherence of Theism, p. 225.

96

ANALYTIC PHILOSOPHY OF REUGION

view that God does not change and is not affected by any contingent events or
circumstances, including events and circumstances that result in suffering, has been
the dominant and orthodox position. 68 Impassibility appears to be most compatible
with God's other metaphysical attributes - immutability, perfection, simplicity, and
eternality.69 The difficulty comes with the apparent incompatibility of impassibility
with divine love. For Christians specifically, the current view has certainly shifted
from neo-Platonism to a much more personal view of God. In order of priority, the
dominant current view regards divine love as outranking impassibility and the
other metaphysical attributes. As Richard Creel explains, "It is conceivable that
one could affirm that God is love and deny that God is impassible yet remain a
Christian, but it is inconceivable that one could affirm that God is impassible and
deny that God is love, yet remain a Christian.,,70 Jung Young Lee puts the matter in
terms of a conflict between God's nature and God's attributes: "The divine nature
governs the quality of divine attribute," he says; thus, "if Agape is the divine nature
and passibility is a divine attribute, the latter must conform to the former.,,71
Central then to the debate of whether God is passible or impassible must be an
understanding of Divine Love. I will return to a further consideration of this matter
momentarily.
It will be helpful to adopt some definition of impassibility. There have been
many different definitions provided; however, they are all variations on a common
theme.72 For sake of discussion, I will use the definition offered by Richard Creel,
which captures the most fundamental elements of impassibility:
To say that God is impassible is to say that God cannot [my emphasis] be
affected by anything .... Hence, it is not sufficient to say merely that God has
not been or never will be affected by anything. Impassibility means it is
logically impossible for God to be affected by anything. 73
Implicit in Creel's definition and Lee's characterization of impassibility is the
distinction that I have used earlier between internal and external relations.
Undoubtedly, God is related to human and animal suffering in some manner. The
question is what is the nature of that relationship. For defenders of God's
impassibility, including Creel, God cannot be related to human suffering as an
internal relation since God's essential nature would then change. If it is logically
impossible for God to be affected by suffering, then God's relation to suffering
68 For a critical discussion of the history of impassibility (which focuses on Augustine and the
Stoics) and its relation to the other divine attributes, see Nicholas Wolterstorff, "Suffering Love," in
Philosophy and the Christian Faith, edited by Thomas V. Morris (Notre Dame, Ind.: University of
Notre Dame Press, 1988), pp. 196-237.
69 See Richard Creel, "Immutability and Impassibility," in A Companion to Philosophy of Religion,
edited by Philip L. Quinn and Charles Taliaferro (Oxford: Blackwell Publishers, 1997), pp. 313ff, for a
short discussion of impassibility and the other attributes of God. Also see, Nicholas Wolterstorff,
"Suffering Love," pp. 217ff.
70 Richard Creel, Divine Impassibility (Cambridge: Cambridge University Press, 1986), p. 2.
71 lung Young Lee, God Suffers for Us: A Systematic Inquiry into a Concept of Divine Passibility
(The Hague: Martinus Nijhoff, 1974), p. 7. Lee is also quoted by Creel, ibid.
72 For a detailed discussion of different definitions and a critical survey of the literature, see Richard
Creel, Divine Impassibility, pp. 3ff.
73 Richard Creel, "Immutability and Impassibility," p. 313.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

97

must be an external relation. This represents the classical neo-Platonist, Stoical


view of God. Otherwise, human affairs and human actions would have the power
to causally produce changes in God's nature.
LAMENT FOR A SON AND REDEMPTIVE LOVE
On the other hand, Wolterstorff takes God's relation to suffering to be an
internal one. For him, the view that God is, by nature, impassible and that "God
remains blissfully unperturbed while humanity drowns in misery" is a "grotesque"
one. 74 Protecting the perfect, immutable, neo-Platonist and Stoical view of God
comes at too high a price for Wolterstorff. While he agrees that God does not
suffer pathos or the "negative" emotions, W olterstorff still maintains that God
should be understood as suffering, since he loves his creatures and delights in their
well-being but suffers in their suffering. 75 The view that God's benevolence is
simply God's eternal disposition to do good is thus unsatisfactory for Wolterstorff.
In such a view, God's redemptive love and the act of redemption is not seen as the
result of God's concern for and reaction to the human condition of sin and
suffering but simply as part of God's nature.
Redemptive love, Wolterstorff argues, must be the result of God's rejoicing and
suffering with his creatures. Drawing upon the very personal experience of the
death of his own son, he poignantly develops the image of a God who suffers even
as humans suffer when the "bonds of love" are severed. 76 Suffering the
catastrophic death of a child, any loving parent's greatest fear, gives one some
insight into Jesus' love for us and God's love for his Son. "Through our tears we
see the tears of God," he says, and "through the prism of my tears I have seen a
suffering God.'.77 God's act of redemption is thus understood as a result of his
being affected by human sin and suffering and not as a result of his perfect,
immutable, benevolent nature. W olterstorff s view of a suffering God thus
reinforces the claim that God must be regarded as both everlasting and changing in
order to be the Redeemer that is central to Christian theology. If God is unaffected
by either the sinful condition of man or the suffering and death of Jesus Christ,
then, for Wolterstorff, the notion of redemption is hardly coherent.
If God is affected by human suffering, there still remain difficult issues in
determining in just what respect God may be affected. As Richard Creel has
pointed out, it makes some significant difference whether it is God's nature, God's
will, God's actions, God's feelings, or God's knowledge that is being affected?8
The argument for God's passibility is most strong concerning God's feelings and
74 Nicholas Wolterstorff, "Suffering Love," p. 211. Charles Taliaferro also develops a theory of
passibility in Consciousness and the Mind of God (Cambridge: Cambridge University Press, 1994), pp.
315-33. I do not discuss his theory of passibility here, but I do discuss his theory of integrated dualism
in some detail below.
75 Ibid., pp. 214-15 and pp. 224ff.
76 Nicholas Wolterstorff, "Suffering Love," in Philosophy and the Christian Faith, edited by
Thomas V. Morris, pp. 196-237, and Lament for a Son (Grand Rapids, Mich.: Eerdmans Publishing Co.,
1987).
77 Wolterstorff, Lamentfora Son, pp. 80 and 81.
78 Richard Creel, Divine Impassibility, Chapters 2 and 7. Also see Kelly James Clark, "Hold Not
Thy Peace at My Tears," in Our Knowledge of God, edited by Kelly James Clark (Dordrecht,
Netherlands: Kluwer Academic Publishers, 1992), pp. 173ff.

98

ANALYTIC PHILOSOPHY OF REUGION

God's actions. For Christians, if God is "unmoved" by the great calamities of the
human condition and if God is unfazed by a parent's lament at the soul-rending,
agonizing death of a child and if God's acts of redemption, including the act of
sending his son to die, are simply the result of God's perfect, preprogrammed
nature, then it seems implausible to retain any of the meaning in the claim that God
is Love. Such a treatment of God's passibility trades heavily, however, upon an
anthropomorphic attribution of human emotions and actions to God and leads to an
inquiry into the difficulties of understanding the notion of divine agency.
GOD AS PERSONAL AGENT
Although the attributes of omnipotence, omniscience, etemality, and necessary
existence have been more widely discussed in the literature - both historically and
recently - perhaps the greatest conundrum in understanding the nature of the deity
within theism is formed by the set of attributes that are associated with regarding
God as a personal agent. Even if the puzzles involving the other metaphysical
attributes normally attributed to God are resolved, it seems that the nature of God
must be explained in some way and in some terms that are at least similar to the
various characteristics that are commonly attributed to human agents. Thus, God is
said to be loving or caring and to have awareness and knowledge and reasons for
performing actions having relationships. These "personal" characteristics
distinguish the God of theism from the Prime Mover of Aristotle, the metaphysical
Absolute of Hegel, and the One of Hinduism. 79
The difficulties with unraveling the conundrum of personal agency involve the
attributes of incorporeality, omnipotence, consciousness, omnipresence, and
omnificence and are far too complex to be dealt with in anything but the most
summary fashion here. I will raise problems instead of offering solutions and will
simply recognize the efforts of others who have attempted to provide answers.
DIVINE AGENCY, INTENTIONS, AND PURPOSES
Richard Swinburne's treatment of divine agency is representative of a classical
theistic position. 80 In the role of creator and sustainer of the universe, God acts
upon the universe and carries out various intentions and purposes and has various
other states of consciousness. To explain how God acts, Swinburne adopts Arthur
Danto's notion of basic actions (discussed in Chapter II). Basic actions are actions
that an agent simply does but does not do in virtue of doing something else, and
mediated actions are actions that an agent does in virtue of performing some other
action. 8! Thus, for a human agent, moving one's arm is a basic action while, by
79 Many critics have objected that the claims about the personhood of God which have become so
thoroughly ingrained in traditional theism are the confused result of a complete misunderstanding of the
nature of persons originating in Cartesian dualism. Such criticism is found in Kai Nielsen, "God,
Disembodied Existence and Incoherence," Sophia, Vol. 26, no. 3, 1987, and Richard Rorty, "Mind as
Ineffable," in Mind in Nature, edited by Richard Q. Elvee (San Francisco: Macmillan, 1973). The most
thorough discussion of the matter of dualism and its effect upon the traditional theistic conception of
God is found in Charles Taliaferro, Consciousness and the Mind of God (Cambridge: Cambridge
University Press, 1994).
80 Richard Swinburne, The Coherence of Theism, Chapter 8.
8t Ibid., pp. 104ff. and p. 135. Derived from Arthur Danto, ''Basic Actions," American Philosophical
Quarterly, Vol. 2, 1965, pp. 141-48.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

99

moving my arm, I might send a signal to someone. For God, the creation of natural
laws may be understood as a basic action and the consequences or happenings that
follow from those laws may be understood as mediated actions through which God
sustains the universe.
There are many difficult issues involved in such a description of divine agency
that lead into metaphysical theory and the philosophy of mind and cannot be dealt
with in any detail here. I shall only suggest the direction in which the inquiry
should proceed and some of the problems that need to be resolved. Perhaps the
most fundamental issue is the question of the nature of the dependence of
consciousness and various conscious states, including intentions and purposes, on a
physical body and, specifically, a brain. The "dependence argument" - the
argument that consciousness and personal identity depend upon the physical body
- continues to be a major difficulty for the theist, both in terms of explaining God
as a noncorporeal, divine agent and in terms of explaining immortality in terms of
a continued disembodied existence of an individual (discussed in Chapter VII).
Swinburne dismisses this problem much too quickly and summarily, saying that we
are sometimes willing to attribute intentions and agency in circumstances where
we do not understand the physical. In the case of Uri Geller, the man who claimed
to have the power of telekinesis and who supposedly used it to bend spoons on
television shows, we may, Swinburne suggests, accept his power to bend the spoon
(at a distance, by a mental act) and his intention of doing so, although we cannot
supply an explanation of the physical forces involved, in order to accept his agency
as a possible explanation for the bent spoons. Since the correlation of the mind
with the brain is still an "unestablished scientific hypothesis," "we may," he
concludes, "explain a man's thoughts by his intentions without us having to believe
that certain brain states and natural laws have to hold in order for a man to think or
have intentions."s2 The dependence argument is discussed in some detail in
Chapter VII in conjunction with the issue of immortality, so I will make only a few
brief comments here concerning the impact of Swinburne's comments on the
notion of divine agency.
There are many questions about how consciousness and such mental states as
intentions and purposes can be meaningfully ascribed to a noncorporeal divine
spirit. First, it is very tenuous to pin nonphysical agency on an argument from
ignorance, that is, in effect saying we do not know that it does not work (as in the
case of Geller). In fact, very few serious-minded people believed that Geller's
mysterious powers were genuine, and his appearances had much more of the air of
a circus sideshow where the audience knows that it is being conned but just does
not know how. Perhaps it is meaningful to speak of intentions and purposes where
there is no brain, but it is a much greater stretch of credulity and understanding to
speak of purposes and intentions where there is nothing physical at all. s3
Philosophers of mind now debate whether computers can think and have intentions
and purposes, but, in this case, microchips have simply replaced brains in the
debate, and a functional analysis of the process of thinking may easily attribute
Swinburne. ibid. p. 141.
Swinburne argues that such a notion is meaningful. See ibid., pp. 106ff. This issue is discussed in
the context of the issue of immortality in Chapter VII.
H2

83

100

ANALYTIC PHILOSOPHY OF RELIGION

thought to a microchip. Since the commonly accepted mind-body dualism does not
work very well in this regard, some have struggled with alternative explanations
for the physical setting for embodiment of divine agency in order to make sense of
that notion (see the discussions of Grace Jantzen and Charles Taliaferro below).
The age-old problem of literal descriptions versus analogical uses of language and
the attempt to maintain something of the distinction between the transcendent and
immanent natures of God confound the issue of divine agency as much as any
other issue concerning God's nature. On one side of the razor-sharp edge of a
meaningful theistic notion of divine agency is pantheism, and on the other side is
atheism. Attempts to balance on this thin edge have resulted in some very strained
positions that seem to slide in one direction or the other. For example, Swinburne
says that although God has no body, we must understand God as embodied in the
world in such a manner that God has no particular perspective or "point of view"
but is "dispersed" throughout the world in such a manner that he is "able to move
any part of the world directly.,,84 All actions are thus basic actions for God, and
everything that transpires in the natural universe - everything from supernovas
exploding to planets orbiting the sun to rain falling to rivers flowing to people
writing books - can be understood as direct actions by God.
Many problems are generated by this view. Does this view make sense? Is it
meaningful to ascribe actions and intentions and purposes to God under such a
view? One obvious problem is that even if such a view of divine embodiment does
make sense, it cannot account for ascribing actions based on intentions and
purposes to God before such embodiment; therefore, it cannot account for the
actions of God as Creator, though creation is regarded as a basic action by
Swinbume. 85 This view thus provides no framework and no point of reference to
understand creation as a basic action. Understanding actions based on intentions
and purposes requires a context within which reasonable predications can be made
concerning regularities of events based upon logical and natural laws so that an
agent can choose a course of action that is judged likely to bring about the intended
result, but it hardly makes sense to talk in such a manner concerning the actions of
God if all of God's actions are basic actions.
Can God be reasonably said to have intentions given the other divine attributes
that make up the nature of God? If so, what does it mean to say that God has
intentions? William Alston holds that we should think of God as "literally a
personal agent," but then he is quick to add that God does not have intentions in
the same way that human beings do and that intentions in the case of God are
something "radically different" from what they are in the case of human beings. 86
The question is - given how "radically different" divine intentions are - whether
there is enough residual similarity remaining to the way in which we ascribe
human intentions to give cognitive content to the notion of divine intentions. Just
how different divine intentions are is illustrated by Alston's example: From "God
Ibid., p. 105.
Ibid., p. 106.
86 William P. Alston, "God's Action in the World," in Evolution and Creation, edited by Ernan
McMullin (Notre Dame, Ind.: University of Notre Dame Press, 1986). Reprinted in William P. Alston,
Divine Nature and Human Language: Essays in Philosophical Theology (Ithaca, N.Y.: Cornell
University Press, 1989), to which the page numbers here refer, p. 198.
84

85

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

101

intends x" and "God is omnipotent," we can infer "x will come to be." However,
no such inference is possible with human agents: From "Jones intends x" and
"Jones is not omnipotent," we certainly cannot infer "x will come to be." For a
human agent to intend x means that x mayor may not come to be. In fact, it seems
an essential part of our understanding of human intentions that there are various
circumstances and various courses of actions available to the agent - some of
which are better or worse than others for bringing about the intention and some of
which will thwart the intention completely. A human agent must adjust her actions
to the circumstances in order to bring about the desired action or state of affairs.
Speaking this way about the actions of an omnipotent being hardly seems logically
appropriate. If God is perfect and omnipotent, whatever God does will bring about
his intention. Thus, since God could never fail to bring about an intention, the
following logical point is raised. What difference would it ever make whether God
had an intention or not? Similarly, what difference would it make whether God had
one intention rather than another? Also, human agents cannot intend things that are
not in their control. I cannot intend to fly to Mars, or for the sun to rise tomorrow,
or for another agent to perform some action. But presumably there are no natural
limitations on what God can intend. The lack of any such limitation leads Alston to
endorse a version of "omnidetermination" (or what is more generally called
omnificience), that is., that every single event in the universe is an action of God. 87
Alston considers this possibility on the basis of both causal determinism and
libertarianism. He concludes that omnificience follows logically and
straightforwardly from determinism. Libertarianism still allows for omnificience in
a qualified way, since God could control human free will and limit its scope in
order to bring about his intentions and prevent it from threatening things about the
world that he values by providing limitations and counterbalances to free will in
the natural world. 88
But what does this position mean for free will? If free will is understood as
requiring the principle of alternative possibilities (PAP), the intuitive principle that
a person is responsible for an action only if she could have done otherwise, then
Alston's modified version of omnificience makes the notion of free will based on
PAP vulnerable to the ultimate Frankfurt -type counterexample. Harry Frankfurt's
attack upon PAP is based upon the following kind of case: Suppose that Black is a
mad scientist who has devised a thought-and-will-control device that unerringly
makes another person think, will, desire, intend, decide, choose, and ultimately act
in a way in which he can control. Black wants Jones to perform a certain action x,
and he uses his device to monitor Jones's thought processes, which he can also do
unerringly. Black does nothing so long as Jones is thinking and deciding "on his
own" to do x; however, if Jones thinks, wills, desires, intends, or chooses anything
that leads him to other than x, Black will use his "little black box" to make Jones
do X. 89 Such cases have become known as "counterfactual intervener" cases, and
the metaphysical implications of such cases are very significant. Frankfurt claims
that this type of case demonstrates that PAP is false, since, although Jones is
Ibid., pp. 199-207. Also see J. L. Mackie, The Miracle of Theism, pp. 16lff.
Alston, "God's Action in the World," p. 207.
89 Harry Frankfurt, "Alternative Possibilities and Moral Responsibility," Journal of Philosophy, Vol.
66, 1969. pp. 829-39.
87
88

102

ANALYTIC PHILOSOPHY OF RELIGION

morally responsible for doing x since he did it on his own, we cannot reasonably
say that he could have done otherwise. 90 Omnificience, even modified or qualified
omnificience, turns the mad scientist into an omnipotent (if benevolent) agent
whose actions seem to make human free will and moral responsibility impossible
or, at least, questionable. It is at least as difficult to reconcile human freedom with
omnificience and middle knowledge as it is to reconcile it with divine
foreknowledge.
POST-THEISTIC TREATMENTS OF GOD
Problems with the traditional notion of divine agency have led some to abandon
the classical theistic view of God in favor of post-theistic conceptions of God.
Perhaps foremost among such figures is Charles Hartshorne, who abandons
traditional theistic notions of agency altogether through his process view of God
and his panpsychism, which is not a form of traditional theism at all. Others perhaps most notably Tillich - have abandoned the notion of person in favor of the
notion of Being. 91 Grace Jantzen has abandoned the classical view of God as
incorporeal, while trying to stay within the broad confines of theism, in favor of a
holistic pantheism where God and the world are an undivided whole and God is
regarded as "embodied" in the world in the same way in which people are
embodied in their bodies. 92 At the same time, she tries to avoid a straightforward
pantheism by arguing that God is also transcendent from the world in a similar way
that people "transcend" their bodies by not being simply identical with their
bodies, as reductive materialism maintains. Jantzen attempts to strike some middle
ground between Cartesian dualism and eliminative materialism; the result appears
to be some form of epiphenomenalism where whatever is incorporeal in God's
nature is the result of a function that is dependent upon the world. Such a view
counterbalances the pull of the transcendent nature of God away from the world
and into metaphysical abstraction and reinserts God into the world. The
immanence of God seems to be obviously necessary for God to be a personal,
caring, loving, and redemptive God, for which God needs to be understood as
somehow "in the world." Her view also helps with what we might call the
Christian "paradox of incarnation": if God is incorporeal, how can God be
incarnated in Jesus Christ; how can the incorporeal be corporeal? To attempt to
resolve such a paradox without simply abandoning it, as Luther did, as one of the
great mysteries of God, something must be given up, and Jantzen gives up the
incorporeality of God. 93 This, however, is a steep price to pay and hardly seems

90 Whether Frankfurt is right about the effect of this type of case on PAP and whether moral
responsibility depends upon PAP have been the subject of some debate in the literature on free will.
Peter van Inwagen, for example, agrees with Frankfurt that these cases show that PAP is false in An
Essay on Free Will (Oxford: Clarendon Press, 1983), pp. 179-80, while Laura Ekstrom disagrees in
Free Will: A Philosophical Study (Boulder, Colo.: Westview Press, 1999), Chapter 6.
91 One fairly recent such example is Gary Legenhausen, "Is God a Person?" Religious Studies, VoL
22,1986,pp.307-23.
92 Grace Jantzen, God's World, God's Body (Philadelphia, Pa.: The Westminster Press, 1984),
Chapter 5.
93 For a critical (but ultimately negative) assessment of the different theological supports for
incorporeality, see Jantzen, God's World, God's Body, pp. 105-130. See also David Paulsen, "Must God

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

103

compatible with traditional theism, which is why I have identified this view, along
with Hartshorne's view, as post-theistic. 94
GOD'S AGENCY RECONSIDERED
Charles Taliaferro has formulated a general theory of divine agency in terms of
what he calls "integrated theism," which is intended to preserve much of the
traditional view of God by preserving both the immanent and transcendent aspects
of divine existence. 95 More than any other known figure, Taliaferro attempts to
explain God's agency by developing his own sophisticated theory of persons.
Based upon what he calls an "integrated dualism," Taliaferro's view of persons is a
holistic one that emphasizes the integral wholeness of the embodiment of the
mental in the physical. While metaphysically dualistic, integrated dualism denies
the extremes of that dualism, particularly the extreme privacy and metaphysical
independence of the mental. In the case of human beings, personhood involves an
embodiment that is a continuing process of varying degrees of attachment and
detachment with the physical world. For a healthy existence, the extreme and
artificial bifurcation that is often associated with dualism needs to be avoided.
Thus, one should not think of one's body as "merely" a physical object being
directed by a mental substance, a mind; rather, deliberative actions are "psychophysical" events in which the mind and the body are integrated together
holistically.96
In the case of God, Taliaferro's integrated dualism leads to a view of God as
embodied spirit in the world. The view of God that results from relying upon
Cartesian dualism emphasizes the metaphysically transcendent nature of God
where God is seen as an incorporeal substance directing things "from above" or
from "outside" the corporeal world. On the basis of integrated dualism, just as
persons are integrated wholes of mind and body, God is integrated with the world
in such a way that allows interaction between the two without introducing the same
metaphysical and theological problems that arise from Cartesian dualism.
Taliaferro's integrated dualism thus avoids the extreme identification of God
with the world embraced by Jantzen and provides a plausible conceptual model for
avoiding the age-old metaphysical issues about the interaction of mental and
physical substances. Integrated dualism, for example, offers a much more
promising way of understanding God's relation to the world, and particularly with
human beings, than a view of God, such as Richard Swinburne's, which treats God
as a completely noncorporeal person. 97 The result of a view such as Swinburne's is
that God becomes such a metaphysically distinct and disparate kind of
noncorporeal entity that it becomes increasingly more and more conceptually
difficult to imagine what it would be like for God to be "like" human persons at all.
This conceptual gulf gives rise to the kind of objection raised by Patrick Sherry,
Be Incorporeal?" Faith and Philosophy, Vol. 6, no. I, 1989, pp. 76-87. For a rebuttal to Jantzen and
Paulsen, see Charles Taliaferro, Consciousness and the Mind of God, pp. 256-71.
94 For criticism of Jantzen on this score, see Taliaferro, ibid., pp. 249ff., especially his note on p. 249
for further references.
95 Charles Taliaferro, Consciousness and the Mind of God.
96 Ibid., pp. 233-34.
97 Richard Swinburne, The Coherence of Theism, pp. IOIff.

104

ANALYTIC PHILOSOPHY OF RELIGION

who objects that such a view of God's nature turns God into some "superFrankenstein" freak that bears only a frightening and ghoulish resemblance to
human persons. 98 Taliaferro is in a better position to respond to this kind of
objection using his integrated dualism since the relationship between God and the
world is much more symbiotic, involving a psycho-physical "connection" where
the two are "profoundly interwoven.,,99
But many unanswered questions remain about how Taliaferro's treatment of
God's personhood affects our understanding of the nature of God and other theistic
beliefs. For example, as he admits, personal immortality seems to be best explained
by a theory of dualism, and there seems to be much more difficulty in
understanding how a continued spiritual, disembodied existence is possible on the
basis of integrated dualism than on Cartesian dualism. lOo And there is a pervasive
vagueness concerning exactly what the notion of integrated means. Consider the
following classical formula for understanding the relationship between God and the
world:
The world minus God = zero.
God minus the world = God.
Taliaferro agrees with John Macquarrie that the second part of this formula
would, if true, represents a complete devaluing of all creation. "The loss of the
world would involve an incalculable loss" for God, Taliaferro claims, since there is
an "affective identity of God and the world."lol But would God still be God if God
lost the world? Taliaferro's integrated dualism and his commitment to a strong
theory of passibility raises other questions about the status of the necessary
existence of God. Is the "integrated" relationship that God has with the world an
internal or external relation? If God's relationship with the world is internal to the
nature of God, even in terms of God's "affective identity," then traditional
understandings of God as necessarily existing and of God as creator must be
revised, since necessary existence as traditionally understood does not seem to
allow the kind of close, integrated, symbiotic relationship between God and the
world defended by Taliaferro. 102
NECESSARY EXISTENCE AND THE ONTOLOGICAL ARGUMENT
Perhaps one of the most distinguishing features of the concept of God in
traditional theism is the claim that God exists necessarily, completely
independently of anything else in the universe. Contingently existing things depend
upon certain states of affairs or conditions or features of the universe holding true
or obtaining and thus exist or do not exist depending upon those states of affairs,
98 Patrick Sherry, "Spirits, Saints, and Immortality (Albany, N.Y.: State University of New York
Press, 1984), p. 13. For a lengthy response to Sherry's objection, see Taliaferro, ibid., pp. 264ff.
99 Taliaferro, Consciousness and the Mind of God, p. 266.
100 This issue is discussed at length in Chapter VII. Using some kind of replica body theory might
help integrated dualism answer this question.
101 Taliaferro, Consciousness and the Mind of God, pp. 335-36.
102 Taliaferro's integrated dualism seems somewhat akin to Alfred North Whitehead's consequent
nature of God - a kinship that can only be noted here.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

105

conditions, or features. On the other hand, God's necessary existence is understood


as conditional upon nothing else, and thus it is impossible for God not to exist if
God exists necessarily. Necessary existence is the only attribute of God that gives
rise, all by itself, both to its own a priori argument for the existence of God and to
its own disproof for the existence of God. The reason for this is that the ontological
argument maintains that, given a clear understanding of the concept of God (and
specifically God's necessary existence), God's existence cannot consistently be
denied. Necessary existence and the ontological argument are inextricably
interwoven, and, after almost two hundred years of disfavor and neglect, the
ontological argument enjoyed a significant revival of renewed interest on the part
of philosophers of religion in the second half of the twentieth century.
KANT'S CRITICISM OF THE ONTOLOGICAL ARGUMENT
The primary reason for the earlier decline of interest in the ontological argument
and its loss of favor was Kant's famous attack upon the argument that has come to
be summarized as "existence is not a predicate." Anselm's version of the argument
is based implicitly upon the claim that existence is a part of perfection in his claim
that existing in the understanding and in reality is more perfect than existing
simply in the understanding alone. At the same time, Descartes's version in his
Meditations is based explicitly upon the claim that existence is a perfection and
that, in the case of God, existence cannot be separated from essence. Most
philosophers and theologians alike took Kant's criticism to be a devastating one,
and thus, for approximately two hundred years, the ontological argument fell into
relative obscurity.
The last few decades of the twentieth century saw a resurgence of interest in the
ontological argument on at least two different scores. First, some figures responded
to Kant's criticism in an effort to undermine his attack upon the argument, and
second, different versions of the ontological argument, the so-called "modal"
versions ofthe argument, which arguably avoid the part of Anselm's argument that
"treats existence as a predicate," began to emerge. The second of these forms of
defense of the ontological argument has been the most significant development;
thus, I will treat the debate concerning whether existence is a predicate more
summarily and focus most of the attention here upon the new, modal versions of
the ontological argument.
Alvin Plantinga's response to Kant's criticism of the ontological argument
serves as one of the more complete and effective retorts to the attack based upon
the claim that "existence is not a predicate."I03 As Plantinga observes, to refute
Anslem's argument, it will not do to simply quote Kant's maxim, "existence is not
a predicate." One must be able to show where the argument goes wrong by
identifying some premise in the argument that makes a mistake that can be clearly
described in Kant's terms.
Though suspicions may abound, to nail down the exact mistake, according to
Plantinga, is "profoundly difficult."I04 The part of Anselm's version of the
ontological argument in his Proslogion that most closely parallels Descartes's
103
104

Alvin Plantinga. God and Other Minds (Ithaca, N.Y.: Cornell University Press, 1967), pp. 29ff.
Ibid., p. 27.

106

ANALYTIC PHILOSOPHY OF RELIGION

claim that existence cannot be separated from essence is the part that makes use of
a reductio that takes the following form: Suppose that that most perfect being
exists in the understanding alone. Then that most perfect being can be thought of as
existing both in the understanding and in reality, and this would be more perfect.
Hence, there would then be something more perfect than that most perfect being,
and since this is absurd and impossible, then that most perfect being cannot exist in
the understanding alone. Since this part of the argument is based upon the claim
that a being which exists in reality as well as in the understanding is "greater" than
one that does not exist in reality, Kant objected that Anselm and Descartes make
the logical mistake of treating existence as a defining predicate of the concept of
God. But does Anselm's version of the ontological argument rest upon making
such a claim? And exactly what does it mean to say that existence cannot be a
predicate?105
One of the more difficult and important issues is whether Kant is claiming that
no propositions that ascribe existence are necessarily true or whether he is claiming
that no such propositions are analytic. Kant introduces the distinction between a
judgment being analytic and its being a priori, and this distinction is crucial to the
whole program of his Critique of Pure Reason since he wants to defend synthetic a
priori judgments; however, his claims about "God exists" seem to confuse the two
kinds of judgments. If the claim is that "God exists" cannot be analytic, then
arguably, this is a telling criticism of Descartes but not Anselm. As Plantinga
correctly points out, perhaps Anselm's proof requires that "God exists" be
necessarily true, but it does not require that it be analytically true. 106 As the two
distinctions have normally been understood since Kant introduced them, analytic
and synthetic are kinds of statements, and the differences are drawn in syntactical
and semantical terms; a priori and a posteriori are kinds of knowledge or ways of
verifying knowledge, and the differences are drawn in epistemological terms.
Putting Kant's criticism in terms of predication seems to lead in the direction of
interpreting his objection as having to do with treating "God exists" as analytic;
however, more sophisticated versions of the ontological argument clearly do not
trade upon treating "God exists" in such a manner. What then is the force of Kant's
objection if we understand the main thrust of his criticism to be that existential
claims cannot be necessarily true, that is, a priori?
CAN EXISTENCE BE A PREDICATE?
Kant's main reason for claiming that no existential proposItIOn can be
necessarily true is that in denying such a proposition we do not contradict
ourselves. Whereas Kant treats "God is omnipotent" as necessary, he thinks that
"God exists" is not necessary since, if we deny that God exists, "we reject the thing
itself with all its predicates.,,107 J. L. Mackie explains Kant's claim that "'Being' is
105 It should be noted that Kant raises a number of different objections to the ontological argument,
of which "existence is not a predicate" is just one. For a full discussion of Kant's different objections,
see Plantinga, ibid., pp. 29ff., and J. L. Mackie, The Miracle of Theism, pp. 43ff.

Plantinga, God and Other Minds, pp. 30-31.


hnmanuel Kant, Critique of Pure Reason, translated by Norman Kemp Smith (London:
Macmillan, 1963).
106

107

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

107

obviously not a real predicate" in terms of the function of the existential quantifier.
While in "God exists," 'exists' is a grammatical predicate, as 'green' is in "This
tree is green," "exists" has a different logical function that is captured by the use of
the existential quantifier. Whereas "The tree is green" or "God is omnipotent"
describes a supposed tree or a supposed God to which the predicates 'green' and
'omnipotent' apply respectively, "God exists" offers no description of God,
according to Mackie, "but merely as it were puts the items [God] into the
picture.,,108 The function of the existential quantifier is to make the ontological
claim that the predicate or description is instantiated in some existing individual. If
we interpret existence as simply a value captured by the existential quantifier, then
existence is not something that belongs to the individual in question at all; and
Descartes' claim that God is the one case that is an exception to the normal logical
distinction between essence and existence can hardly be maintained. 109 The use of
the quantifier does serve to answer the objection of Plantinga's that Kant allowed
for existential propositions in mathematics that are necessarily true. "There is a
prime number between seventeen and twenty" and "The successor of six is greater
than five," for example, are both necessarily true existential statements that can be
accounted for in terms of the use of the existential quantifier. 110 Such an account,
following Quine, turns existence into "being the value of a variable" and separates
the question of ontology - the question of what "really" exists - from the question
of the ontological commitment of a particular language. 1ll Kant could easily say
that the language of mathematics commits one to the existence of prime numbers
without saying that prime numbers exist or subsist in some Meinongian fashion.
However, the treatment of existence as simply the instantiation of predicates
captured by the existential quantifier is highly controversial. Mackie himself
maintains that existing is something that individuals do (like "being green") and
admits that further discussion is needed to decide whether the notion of existence
that is at stake can be captured by the use of the existential quantifier. I 12 A more
thorough treatment of analyzing existence in terms of the existential quantifier is
offered by Graham Oppy.1l3
Kant's maxim at least serves to put us on notice that it is not possible simply to
define God into existence, a caution that is more appropriate for Descartes than for
Anselm. If the ontological argument boils down to saying simply that existence is
part of the definition of God and hence means that it is impossible for something
that is defined as existing not to exist, then surely it is fallacious since we could
then define all sorts of things into existence. Mackie's example involves defining a
Martian into existence by defining the term "Remartian" (short for "real Martian")
108 Mackie, The Miracle of Theism, p. 46. Mackie follows W. C. Kneale on this point. See W. C.
Kneale, "Is Existence a Predicate?" in Readings in Philosophical Analysis, edited by Herbert Feigl and
Wilfrid Sellars (New York: Appleton-Century-Croft, 1949).
1119 Mackie, ibid., pp. 46-47.
110 Plantinga, God and Other Minds, p. 32.
111 Some would say, following Quine, that the ontological question of what "really" exists is
nonsensical and is replaced by the question of the ontological commitment of a particular language.
112 Ibid., p. 47.
113 Graham Oppy, Ontological Arguments and Belief in God (Cambridge: Cambridge University
Press, 1995), pp. 33ff. This volume contains the most complete annotated bibliography on the
ontological argument available, pp. 200ff.

108

ANALYTIC PHILOSOPHY OF RELIGION

to include existence as part of its meaning; Remartians would then exist, and it
would be contradictory to deny that they doy4 Mackie objects on the basis of this
point that existence cannot be a perfection of God as Descartes claims.
Oppy's suggestion of adding a definitional operator is a useful one. He is
certainly right that the argument "By definition, God exists, therefore God exists"
is fallacious. But "By definition, God exists, therefore God exists according to this
definition" is trivially valid. ll5 If we apply such an operator to Descartes' claim
that God's existence cannot be separated from God's essence, then we get,
according to this essence of God, God exists. It is then left to the reader to agree or
disagree with Descartes's account of the essence of God.
Analyzing the meaning of the claim that existence is not a predicate and
exposing the definitional force of some versions of the ontological argument are
issues that have occurred frequently in the contemporary literature. However, such
concerns leave untouched Anselm's more interesting version of the argument that
arguably raises neither problem. Modern, modal versions of the ontological
argument are attempts to capture the fundamental insight of Anselm's original
argument in ways that explicitly avoid any question about predication or
definition. 116
HARTSHORNE'S MODAL ONTOLOGICAL ARGUMENT
There are now several different versions of modern, modal ontological
arguments - too many to allow for careful treatment of them all here.l17 Charles
Hartshorne and Norman Malcolm are primarily responsible for renewing interest in
the ontological argument and for introducing modal versions of the argument into
the contemporary debate. 118 I will focus attention on their arguments and on Alvin
Platinga's somewhat later version because these are both the better known versions
of the modal argument and because they are the more interesting and challenging.
Both Hartshorne and Malcolm claim that Anselm actually presents two different
versions of the ontological argument. The version most frequently discussed
114 Mackie, The Miracle of Theism, pp. 42-43. This is a modem-day version of Gaunilo's objection
that a perfect island would necessarily have to exist.
1150ppy, Ontological Arguments and Belief in God, pp. 48-49.
116 For further discussion of the issue of whether existence can be a predicate, see the articles by
Bertrand Russell and Jerome Shaffer in The Many-Faced Argument, edited by John Hick and Arthur
McGill (London: Macmillan, 1968), pp. 219-45.
111 For example, a version of the ontological argument that is now receiving some attention is the
one that was found in the unpublished papers of Kurt GOdel. Giidel's original version appears in his
Collected Works, Volume 3, edited by S. Feferman et al. (Oxford: Oxford University Press, 1995). For
critical comments, see C. Anthony Anderson, "Some Emendations of Giidel's Ontological Argument,"
Faith and Philosophy, Vol. 7, 1990, pp. 291-303; Graham Oppy, "Giidelian Ontological Arguments,"
Analysis, Vol. 56, 1996, pp. 226-30; and C. Anthony Anderson and Michael Gettings, "GOdel's
Ontological Proof Revisited," in Lecture Notes in Logic 6: Godel '96, edited by Petr Hajek (New York:
Springer-Verlag, 1996), pp. 167-72.
118 Charles Hartshorne, '''The Necessarily Existent," in Man's Vision of God (New York: Harper and
Row, 1941), pp. 229-341. Page numbers refer to reprint in The Ontological Argument, edited by Alvin
Plantinga (Garden City, N.Y.: Anchor Books, 1965); The Logic of Perfection (LaSalle, lll.: Open Court,
1962); and Anselm's Discovery (LaSalle, Ill.: Open Court, 1965). Norman Malcolm, "Anselm's
Ontological Arguments," Philosophical Review, Vol. 69, no. 1, January 1960. Reprinted in Malcolm's
Knowledge and Certainty (Englewood, Cliffs, N.J.: Prentice-Hall, 1963), to which the page numbers
here refer.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

109

appears in Chapter 2 of the Pros log ion while the second "modal" version, which
both Hartshorne and Malcolm take as their point of departure for their modal
arguments, appears in Chapter 3. 119 Here is Anselm's hitherto neglected modal
argument as quoted by Malcolm:
If it [that than which nothing greater can be conceived] can be conceived at all it
must exist. For no one who denies or doubts the existence of a being a greater
than which is inconceivable, denies or doubts that if it did exist its nonexistence, either in reality or in the understanding, would be impossible. For
otherwise it would not be a being a greater than which cannot be conceived. But
as to whatever can be conceived but does not exist: if it were to exist its nonexistence either in reality or in the understanding would be possible. Therefore,
if a being a greater than which cannot be conceived, can even be conceived, it
must exist.

Even though his version of the modal ontological argument is not so well known
among analytic philosophers because of his Whiteheadean orientation and his
abandonment of traditional theism in favor of process theism, Hartshorne perhaps
deserves more credit than Malcolm for resurrecting interest in the ontological
argument and developing modal versions of that argument. Early versions of
Hartshorne's modal argument appeared as early as 1941, but I shall focus here
upon the later, more mature versions. In this second version of his ontological
argument, Hartshorne insists that it is clear that Anselm's major insight, upon
which the ontological argument is based, is that necessary existence is a superior
manner of existence to ordinary, contingent existence and that ordinary, contingent
existence is a defect. Both Hume and Kant misunderstood the point by focusing
upon whether what exists is greater than what does not exist. According to
Hartshorne, however, Anselm's point is that what exists and cannot not exist is
greater than that which exists and can not exist. 120 This way of putting the issue
reveals just how irrelevant the claim that existence is not a predicate is to the
modal interpretation of the argument.
This way of putting the ontological argument makes clear that it is not simply
nonexistence that is incompatible with perfection but even the possibility of
nonexistence. What is critical for Anselm's argument is the claim that if perfection
is actualized, it is necessarily actualized. Contingent existence is as much an
imperfection as is nonexistence, and both are logically incompatible with
perfection. We thus arrive at what Hartshorne identifies as "Anselm's Principle": if
perfection exists, it exists necessarily. In terms of God's existence, this means if
God's existence is possible, God necessarily exists. Hartshorne's modal argument
thus proceeds on the assumption that the existence of God, that is, the existence of
a perfect being, is not logically impossible. If God does not exist, then there is no
119 Whether Anslem intended to present two different arguments in Chapter 3 or whether he intended
to be exploring the attributes of the being whose existence he had established in Chapter 2 is an
intriguing question of historical exegesis. However, the main substantive issue is not whether the modal
version of the argument can be attributed to Anselm but whether a sound modal version of the argument
can be fonnulated.
120 Hartshorne, The Logic of Peifection, p. 58.

110

ANALYTIC PHILOSOPHY OF REliGION

possibility that God exists; if God does not exist, then the claim that God does exist
must be logically absurd. The "merely possible," contingent existence of God must
be inconceivable. The modal argument does not prove that God exists because it
does not and cannot exclude the possibility of the logically impossibility that God
exists. 121 Strictly speaking, then, Hartshorne's modal argument does not purport to
prove that God exists; it purports to prove that if it is possible for God to exist, then
God must necessarily exist. If it is possible for God to exist, then the concept of
God is conceivable, that is, not logically impossible. Thus, as Hartshorne
summarizes the modal proof, "if 'God' stands for something conceivable, it stands
for something actual.,,122
THE LOGICAL STRUCTURE OF HARTSHORNE'S MODAL ARGUMENT
Using the standard modal symbolism, Hartshorne's interpretation of Anselm's
ontological argument may be formalized in the following manner: 123
Where "N" stands for "it is necessary (logically true) that"
"(~x)Px" stands for "a perfect being exists"
"-" stands for "it is not true that"
"V" stands for "or"
and "p~ q" stands for "p strictly implies q" or "N -(p and -q)"
1. (~x)Px ~ N (~x)Px
2. N(~x) Px V - N(~x) Px
3. - N(~x) Px ~ N - N(~x) Px
4. N(~x) Px V N - N(~x) Px
5. N -N(~x) Px ~ N-(~x) Px
6. N(~x) Px V N-(~x) Px
7. -N-(~x)Px
8. N(~x) Px
9. N(~x) Px ~ (~x) Px
10. (~x) Px

Anselm's principle
Law of excluded middle
Necessity of modal status
From 2 and 3
Modal form of modus tollens from I
From 4 and 5
Assumption (perfection is not impossible)
From 6 and 7
Modal axiom
From 8 and 9

The conclusion is that a perfect being exists, the modal equivalent of Anselm's
"that than which nothing greater can be conceived" exists. Most of the steps in this
proof follow from standard S5 modal logic. This observation reveals a detail that is
sometimes ignored concerning the modal versions of the ontological argument. If
"necessary" is taken to mean "logically necessary," then every modal argument
must occur within or using some formal logical system or language within which
the modal operators are defined and within which formal derivations (inferences)
are justified. 124 Thus, "necessary" will always be defined relative to a particular
language, as will the conclusion, "a perfect being exists." If any of the standard
Hartshorne, "The Necessarily Existent," p. 124.
Ibid., p. 135.
123 See Hartshorne, The Logic of Perfection, pp. 50-51.
124 John Hick has objected that Anselm intended "necessary" to mean "ontological necessity" rather
than "logical necessity." See his "A Critique of the 'Second Argument,'" in John Hick and Arthus C.
McGill, The Many-Faced Argument (New York: Macmillan, 1967).
121

122

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

111

modal systems are used, T, S4, or S5, then one embraces those systems by
embracing the proof. Since those systems are provably consistent, additional care
must be taken to make sure that the additional assumptions (discussed below) do
not make those systems inconsistent. 125
Steps (1) and (7) must bear the greatest scrutiny and require the most
explanation and justification. It may be true that "the Devil is in the details," but, in
the case of the modal ontological argument, if God is in the argument, God is in the
details. We must therefore examine the details of steps (1) and (7) closely.
The assumptions in steps (1) and (7) occur in one form or another in the
different modal proofs. In Hartshorne's case, step (l) represents the recasting of the
nature of divine perfection in a way that is intended to avoid the issue of whether
existence is a predicate altogether. What step (1) asserts is that "merely possible"
existence, that is, contingent existence, is logically incompatible with the notion of
divine perfection. Hartshorne describes Anselm's insight as the realization that it is
impossible that divine perfection is capable of being actualized and also capable of
not being actualized. Step (1) embodies the claim that if perfection is actualized,
then it must necessarily be actualized, and if the notion of perfection is logically
possible, that is, if it involves no logical contradiction, then it is capable of being
actualized. 126
IS MODAL PERFECTION A PREDICATE?
Given the assumption in step I, the argument no longer turns on whether
nonexistence is a defect or on whether existence can legitimately be considered a
predicate. The modal argument turns on whether contingent existence is a defect whether the mere possibility of nonexistence is a defect - and whether the mere
possibility of nonexistence can legitimately be treated as a predicate. Is modality
with respect to existence a legitimate predicate? This question focuses upon the
modal counterpart to Kant's claim that existence is not a predicate. Hartshorne
thinks that even the critics of the argument recognize and treat modality of
existence as a predicate. His argument is that standard treatments of predication
that treat predicates as contingently satisfied, that is, as actualized or nonactualized,
are based upon the conceivability of both the existence of a thing to which the
predicate applies and the nonexistence of that thing. Hartshorne concludes that
contingency is thus regularly regarded as a predicate of predicates, and if
contingency is recognized as a predicate, so must be noncontingency.127
The way in which the rationale for the claim that necessary existence is a
legitimate predicate must be described - "the contingent existence of predicates is
a predicate" - suggests that there is a logical problem here. Indeed, Hartshorne
considers the objection that the argument is based upon a logical-type confusionone of the most serious objections to the argument. 128 If existence is of a different
logical type than any predicate, then any universal, abstracted form of a predicate
cannot logically imply that that universal is instantiated. Thus, the objection goes,
125 For an introduction to these modal systems, see G. E. Hughes and M. J. Creswell, An
Introduction to Modal Logic (London: Methuen and Company, 1968).
126 Hartshorne, The Logic of Perfection, p. 60.
127 Ibid., 52.
128 Ibid., p. 46.

112

ANALYTIC PHILOSOPHY OF REUGION

just as "blueness" cannot imply that any particular blue thing exists, "perfection"
cannot imply that any particular perfect thing exists. The problem is this: If
modality of existence rather than existence per se is a predicate, as this modal
version of the ontological argument claims, then when the modality is necessary
instead of possible, the modality includes the actual existence and strictly implies
the actual existence as evidenced in step (9). If necessary existence includes actual
existence, how can a necessarily existent being still be perfect, that is, of a "higher"
logical type than merely contingently existent beings, if that necessary existence
implies existence per se? Is it not the case that perfection is "contaminated" if it
includes actual existence? And if necessary existence includes or implies actual
existence, even if the modality of existence is a legitimate predicate, how can
necessary existence be regarded as a "better" predicate appropriate for perfection
whereas contingent existence is not?
These questions point to powerful, telling criticisms on which the entire fate of
Hartshorne's modal version of the ontological argument rests. 129 He finds a
response only in his "neo-classical," process view of the nature of God, the details
of which are beyond the scope of this volume. The general direction of the
response involves distinguishing two different aspects or levels of existence, one of
which is more necessary and abstract and one of which is more contingent and
concrete. 130 As Hartshorne uses this distinction, that a person will exist tomorrow
is more abstract than that a person will exist with certain specific predicates at a
specific time, for example, wearing sandals and a blue shirt, walking down a road,
whistling a specific tune, and hearing a blue jay call. Thus, there is the more
abstract kind of existence, which involves no specified instances of predicates but
simply some instance of some predicate or some other instance of some other
predicate, and there is the more concrete kind of existence, which involves some
specific instance of some specific predicate. Hartshorne treats the former, more
abstract kind of existence as internal to perfection and the latter, more concrete
kind of existence as external to perfection. l3l This "process view of God," which
admits change and limitations into the nature of God, is viewed by traditional
theists as also admitting imperfections into the nature of God.
MALCOLM'S MODAL VERSION OF THE ONTOLOGICAL ARGUMENT
Norman Malcolm also thinks that there are two different arguments to be found
in Anselm's Proslogion and that the second argument, found in Chapter 3 and
quoted above, is a modal argument. Malcolm's interpretation of the main part of
Anselm's "second" argument, which turns the argument into a modal argument, is
the same as Hartshorne's. In this proof, it is not existence that is a part of
perfection, but necessary existence. The proof does not tum upon whether God
would still be "that than which nothing greater can be conceived" if God did not
exist but rather whether God would still be "that than which nothing greater can be
conceived" if it were logically possible for God not to exist. Malcolm explains the
differences between Anselm's first and second proofs in the following way: "His
129
130

See ibid., p. 63.


Ibid., pp. 63ff.

131 Hartshorne thinks that admitting potentiality in God for future concrete states actually strengthens
the case for regarding God's existence as necessary. See ibid., pp. 63-68.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

113

[Anselm's] first ontological proof uses the principle that a thing is greater if it
exists than if it does not exist. His second proof employs the different principle that
a thing is greater if it necessarily exists than if it does not necessarily exist.,,132 This
modal proof treats the modality of existence, necessary existence, as a predicate of
God and claims that it is logically necessary that God exists and that it is logically
impossible for God not to exist. If it is logically impossible for God not to exist,
then not only can nonexistence not be reasonably predicated of God (as the first
proof supposedly establishes), but contingent existence cannot be reasonably
predicated of God either (as the second proof supposedly establishes). 133 Here then
is Malcolm's summary of the modal proof:
If God, a being a greater than which cannot be conceived, does not exist then He
cannot come into existence. For if He did He would either have been caused to
come into existence or have happened to come into existence, and in either case
He would be a limited being, which by our conception of Him He is not. Since
He cannot come into existence, if He does not exist His existence is impossible.
If He does exist He cannot have come into existence (for the reasons given), nor
can He cease to exist, for nothing could cause Him to cease to exist nor could it
just happen that He ceased to exist. So if God exists His existence is necessary.
Thus God's existence is either impossible or necessary. It can be the former
only if the concept of such a being is self-contradictory or in some way logically
absurd. Assuming that this is not so, if follows that He necessarily exists. 134

Since Hartshorne provides a formalization of his modal version of the


ontological argument, and Malcolm does not, I will not formalize Malcolm's
argument here. The argument lends itself to varying formalizations. 135
IS NECESSARY EXISTENCE A PREDICATE?
In the above argument, Malcolm, like Hartshorne, treats necessary existence but
not contingent existence as a predicate. While "the notion of contingent existence
or of contingent nonexistence cannot have any application to God," he explains,
"necessary existence is a property of God in the same sense that necessary
omnipotence and necessary omniscience are His properties.,,136 Whereas
Hartshorne himself raised and attempted to respond to the logical-type objection to
treating necessary existence as a predicate, Malcolm's critics have done the job of
raising objections for him. One main issue centers upon whether necessary
existence and contingent existence are different kinds of existence or whether
Norman Malcolm, "Anselm's Ontological Arguments," p. 146.
As with Hartshorne, there is significant controversy over Malcolm's interpretation of Anselm.
See, for example, Clement Dore, Theism (Dordrecht, Netherlands: D. Reidel, 1984), pp. 144ff.
134 Ibid., pp. 149-50.
135 Richard Gale takes Malcolm to be offering two different versions of Anselm's "second" proof,
but it seems more likely that Malcolm intended his comments to be simply elaborations of a single
modal proof. At any rate, the differences are not great. For Gale's formalizations of Malcolm's modal
proof(s), see his On the Nature and Existence of God (Cambridge: Cambridge University Press, 1991),
pp. 209ff. Compare Keith E. Yandell, Basic Issues in the Philosophy of Religion (Boston: Allyn and
Bacon, Inc., 1971), pp. 10 Iff.
136 Norman Malcolm, "Anselm's Ontological Arguments," pp. 149 and 150.
132
133

114

ANALYTIC PHILOSOPHY OF REUGION

necessary existence and contingent existence are simply two very different things.
In the first case, we have existence simpliciter, which is suitably qualified by
"necessary" and "contingent," and in this case, if necessary existence is a
predicate, then so must be existence simpliciter since if anything exists necessarily,
it necessarily exists simpliciter. But if necessary existence is a predicate, then so
must be contingent existence since anything that exists contingently also must exist
simpliciter. In the latter case, if necessary existence and contingent existence are
simply two completely different things, and if necessary existence is to be regarded
as a predicate but contingent existence is not, then it is difficult to see what the two
have in common. Necessary existence is explained by both Hartshorne and
Malcolm in comparison and contrast with contingent existence, but if necessary
existence is a predicate while contingent existence is not, then the two are so
different in nature that calling both "existences" seems inappropriate. 137
IS NECESSARY EXISTENCE REDUCIBLE TO EXISTENCE
SIMPLICITER?
Malcolm claims that for God to be God, God must be unlimited, and
unlimitedness must mean that God is not dependent upon anything else for
existence; hence, existence for God cannot be mere existence, that is, contingent
existence that God mayor may not have, but must be necessary existence. He
explains this claim by using a comparison with omnipotence. God is not simply
omnipotent, but God is necessarily omnipotent. For God to be necessarily
omnipotent means that omnipotence is somehow (borrowing a Kantian spacial
metaphor) "contained in" the concept of God. This view of the nature of the
designation 'God' means that 'God' cannot be a logically proper name in the
Russellian sense, picking out a bare particular, but must be a disguised description
or at least have a meaning that is captured in a "cluster" of core predicates in the
manner suggested by John Searle. Thus, for Malcolm, for x to be God means
necessarily for x to be omnipotent and to exist. Perhaps, in more familiar terms,
'God is omnipotent' is analytic, as is 'God exists,' in the same way as 'A bachelor
is male' and 'A bachelor is unmarried' are analytic. But if God possesses the
predicate 'omnipotence' necessarily, does this mean that there is a separate and
distinct predicate 'omnipotence' out there in the world to be predicated of
individuals? And if God possesses the predicate 'necessary existence,' does this
mean that there is a separate and distinct predicate 'necessary existence' out there
in the world to be predicated of individuals? If a bachelor is necessarily male, does
this mean that a bachelor has a different property of maleness than a married male
does? The answers to all of these questions must be in the negative. If God's
necessary existence is to be understood in such a manner as to make 'God exists'
logically true, then 'necessary' must be understood as the kind of necessity that
designates a mode of predication, a way in which a predicate is related to the
individual, and not a different kind of predicate. This mode of predication
137 For criticisms of Malcolm on this point, see W. Baumer, "Anselm, Truth, and Necessary Being,"
Philosophy, VoL 37, 1962, pp. 257-58, and Terrence Peneihum, "Some Recent Discussions of the
Traditional Proofs - The Ontological Argument," in Religion and Rationality (New York: Random
House, 1971), pp. 365-72. For discussion of other responses, see Oppy, Ontological Arguments and
Beliefin God, pp. 211-12.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

115

described by 'necessary' is captured in modal logic by use of the modal operator.


By being necessarily male, a bachelor does not have a different kind of predicate
than do males that are not bachelors, the bachelor simply possesses the predicate
'male' in a different manner, in a different mode, than do other nonbachelor males.
The ascription of a predicate e by use of the particular mode of predication
captured by 'necessary' is thus parasitic upon the ascription of e simpliciter, that is,
for an individual to be necessarily e means for that individual to be e in a particular
manner. If this understanding of 'necessary' is correct, then 'necessary existence'
is not a separate and distinct property possessed by God but simply a mode in
which God possesses mere existence. In this sense, 'necessary existence' is not a
separate predicate at all of existent beings or objects, and if 'necessary' is
understood as a predicate at all, it is a second-order predicate of the proposition
'God exists.' Thus, Malcolm's reading of Anselm's second, modal ontological
argument reduces to the first, and we are once again back to the issue of whether
existence simplicter can be a predicate.
THE POSSIBILITY OF MODAL PERFECTION
The modal versions of the ontological argument considered above proceed from
the assumption that God's existence is possible. The arguments proceed on the
basis of the claim that if the existence of a being than which nothing greater can be
conceived is possible, then the existence of that being is necessary. In step (7) of
Hartshorne's proof, this assumption is described as it is not the case that the
existence of divine perfection is impossible. In both Hartshorne and Malcolm's
interpretations of Anselm, we should replace "that than which nothing greater can
be conceived" with "that than which nothing greater can be conceived and which
cannot be conceived not to exist." Let us call such a being a modally perfect being,
and let us call the possibility that such a being exists the possibility of modal
perfection (PMP).138 Granting PMP means that 'God exists' becomes logically
true; however, if PMP is denied, then 'God exists' becomes contradictory, and the
existence of God becomes logically impossible. Hartshorne thought that PMP is
established by the cosmological argument, that is, he thought that while the
cosmological argument does not establish the existence of God, it does establish
the possibility of God's existence. 139 Others, as we shall see, have disputed whether
divine perfection is possible.
As Clement Dore recognizes, the modal versions of the ontological argument
turn the fundamental question concerning God's existence from "Does God exist?"
to "Is God's existence logically possible?"l40 The only way in which there could be
a negative answer to this modal question and a rejection of PMP is for the concept
of a modally perfect being to be contradictory, and much of the debate concerning
the modal versions of the ontological argument focuses upon this issue. For his
part, although he does not think that a demonstration that the concept of a modally
138 While modality may apply to other attributes of God, I take "modally perfect being" as shorthand
for perfection solely in terms of existence.
139 For a detailed treatment of the relationship between the ontological and cosmological arguments
for Hartshorne, see H. G. Hubbeling, "Hartshorne and the Ontological Argument," in The Philosophy of
Charles Hartshorne, edited by Lewis Edwin Halrn (LaSalle, TIl.: Open Court, 1991), pp. 367ff.
140 Clement Dore, Theism, Chapter 6.

116

ANALYTIC PHILOSOPHY OF REUGION

perfect being is not contradictory is possible, Malcolm does not feel that there is
any legitimate burden of proof for a defender of the modal argument to provide
such a demonstration of the consistency of the concept of a modally perfect
being.141 Others, however, have taken up the issue of the logical consistency of a
modally perfect being, which Malcolm avoids. Malcolm assumes that in the
absence of some demonstration that the concept of a modally perfect being is
contradictory, we can justifiably assume that it is logically consistent and thus
assume PMP; however, as Dore points out, in the absence of a demonstration that
the idea of a modally perfect being is consistent, it is equally epistemically
acceptable to conclude that the concept of a modally perfect being is contradictory.
Thus, if the defender of the modal versions of the ontological argument intends for
the modal permutation of the argument to be more persuasive than the original
nonmodal version, then some defense of PMP is in order. Otherwise, the
assumption of PMP will appear as nothing but question begging to nontheists. 142
Paul Henle resurrects a Gaunilo-type objection to PMP by proposing the
possibility of the existence of a being named Nee, which is a being that exists
necessarily but which is otherwise limited in various ways.143 Nee is neither
omniscient nor omnipotent, but if Nee exists necessarily, then such a being would
represent a limitation upon God's omnipotence since God could not destroy such a
being. It is easy to extend the argument for Nee to populate the world with a whole
host of modally perfectly existent but otherwise limited beings. The existence of
Nee seems to follow from the same assumption of PMP that Malcolm and
Hartshorne assume in their modal proofs. While the possibility of a modally
perfect but otherwise limited being might be acceptable (or even preferable for
Hartshorne), such a being is not acceptable to the traditional theist. As Anselm
claimed, God is the only modally perfect being, and thus, if PMP is granted, it
leads to the existence of one and only one necessarily existent being. What reason
may there be for thinking that God and only God exists necessarily? Henle argues
it is because 'God' is treated as a proper name. He claims that if the question of the
logical possibility of God and Nee is posed as a general question about the
possibility of the existence of any such beings in general without the use of proper
names, there would be no inclination to conclude that such beings exist. If
Malcolm reconstructed his modal argument in a general form to consider whether
necessarily existent beings exist, and if the assumption of the possibility of the
existence of such beings were put in a general form without naming a specific
being, then, Henle thinks, the modal argument would lose its appeal completely.
As he concludes, "If then there was any inclination to believe in the existence of
Nee [or God], it must have been merely because of the use of the proper name.
Since the use of proper names implies or presupposes existence, it is easy to slide
into the belief that what has been named exists, and this is particularly easy when
the entity is described in terms of existence."l44 Malcolm's modal version of the

Norman Malcolm, "Anselm's Ontological Arguments," p. 159.


Dore, Theism, pp. 70ff.
143 Paul Henle, "Uses of the Ontological Argument," Philosophical Review, Vol. 70, 1961, pp. 102ff.
144 Ibid., p. 106.
141

142

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

117

ontological argument thus prejudices the case in favor of a necessarily existing


being, according to Henle, by the use of the proper name "God." 145
The effect of Henle's argument is easily seen in Dore's defense of the logical
consistency of the notion of a modally perfectly existent being, which occurs in the
context of his consideration of the argument for the logical coherence of a modally
perfect being based upon religious experience: people have experienced God, and
experience of the lo~ically impossible is not possible, therefore God is not
logically impossible. I 6 In his defense of the claim that people have, in fact,
actually experienced God, Dore relies upon an appeal to the fact that a large
number of "sane and sober people" have reported such experiences (the reliability
of such reports is the subject of the next chapter). The main point here concerns
how Dore compares the putative reports of the experience of God with the putative
reports of other kinds of logically impossible things such as square circles. If a
person sincerely reports experiencing such a thing as a square circle, then we reject
such a report because such experiences are not possible since they would be
experiences of the logically impossible. But Dore does not treat experiences of
God as experiences of the logically impossible, as he does the reported experiences
of square circles. Why? In contrasting putative experiences of God with putative
experiences of square circles, Dore always uses the designation 'God' as a proper
name while, at the same time, he describes square circles as a class or by the use of
an indefinite description, 'a square circle.' The difference lends apparent substance
to a subtle form of begging the question, since the use of the proper name 'God'
inclines one to posit an individual named by the proper name, in this case a
modally perfect individual,147 while the use of an indefinite description does not.
THE MODAL DISPROOF OF THE EXISTENCE OF GOD
Perhaps the best known case of refusing to grant the PMP required by
Hartshorne and Malcolm's versions of the modal ontological argument results in J.
N. Findlay's notorious modal disproof of the existence of GOd. 148 Findlay begins
by arguing that to explain the nature of the object that is worthy of religious
worship and an "adequate object of religious attitudes" we must presume that God
has an unsurpassable superiority to human beings. He says,
The worthy object of our worship can never be a thing that merely happens to
exist, nor one on which all other objects merely happen to depend. The true
object of religious reverence must be one, merely, to which no actual
independent realities stand opposed; it must be one to which such opposition is
145 For monotheists, it is crucial then to argue that there is one and only one necessarily existent
being, a part of the argument that is frequently neglected. For a generalized version of Henle's point,
see R. Kane, "The Modal Ontological Argument," Mind, Vol. 93, 1984, pp. 336-50. Kane shows that it
is easy to use the modal argument to populate the universe with an infinite number of necessarily
existent but slightly less than perfect beings. See Oppy, Ontological Arguments and Belief in God, pp.
171ff., for discussion of this development.
146 Dore, Theism, p. 74.
147 Ibid., pp. 74-75.
148 J. N. Findlay, "Can God's Existence Be Disproved," in New Essays in Philosophical Theology,
edited by Antony Flew and Alasdair MacIntyre (London: SCM Press, 1955), pp. 47-75. This article is
widely reprinted in different anthologies.

118

ANALYTIC PHILOSOPHY OF REUGION

totally inconceivable. God mustn't merely cover the territory of the actual, but
also, with equal comprehensiveness, the territory of the possible. And not only
must the existence of other things be unthinkable without him, but his own nonexistence must be wholly unthinkable in any circumstances. 149
Nor, we might add, can God possess any of the other attributes that make him
unsurpassably superior to human beings in an accidental or contingent manner. As
I have already noted in the above discussion of Hartshorne and Malcolm, their
modal arguments turn on the claim that 'God exists' must be necessarily (or
logically) true. The consequence of such a claim is that the existence of God, for
Findlay, becomes "either senseless or impossible.,,15o Based upon what he calls a
"contemporary outlook," necessary truths can only make claims about the possible
instance of different predicates, they can never make claims about the actual
instantiations of predicates. This means that the necessity of propositions is
explained in terms of the linguistic conventions of the language in which the
propositions occur. The result is that different formulations of the modal argument
turn out to be valid relative to different languages in which they are formulated.
From consideration of the assumption that God's existence cannot be contingent,
Findlay thus arrives at a formal denial of PMP because he treats the concept of an
existent being who exists necessarily as logically absurd or logically impossible.1 51
It is helpful to recognize in Findlay's argument both a theological claim and a
philosophical claim. The theological claim is that God must exist necessarily to be
a worthy object of religious attitudes, and the philosophical claim, the result of "the
contemporary outlook," is that no necessarily true proposition can be an existential
proposition. To reply to Findlay's modal disproof of the existence of God, the
theist must deny one or the other of these two claims. John Hick denies the
theological claim that the proposition "God exists" must be necessarily true for
God to be a worthy object of worship.152 Alvin Plantinga denies the philosophical
claim that no necessarily true statement can be an existential one. As he notes, even
Kant seemed to treat existential statements of mathematics (for example, there is
one prime number between one and five) as synthetic, a priori, and hence
necessary.153 Plantinga's appeal to Kant notwithstanding, his acceptance of the
theological claim and his rejection of the philosophical claim lead him in a new

Ibid., p. 52.
Ibid., p. 54.
151 For a detailed and formalized treatment of Findlay's disproof, see Bowman Clarke, "Modal
Disproofs and Proofs for God," Southern Journal of Philosophy, fall 1971, pp. 247-58. As Clarke notes,
Findlay's objection amounts to a modal extension of tbe Hume-Kant objection to Anselm's version of
tbe ontological argument.
152 See John Hick, "A Critique oftbe 'Second Argument,'" p. 343.
153 See Alvin Plantinga, God and Other Minds, p. 174. Plantinga is correct tbat statements of
matbematics are a priori for Kant, but tbey are not logically true in tbe sense that Malcolm and
Hartshorne claim tbat "God exists" is logically true. While Kant's distinction between analytic and
syntbetic statements does not map neatly onto contemporary uses of "necessarily true" and "logically
true," it seems tbat tbe modal versions of tbe ontological argument discussed so far would have us
understand "God exists" as analytic a priori in tbe sense of analytic tbat depends upon essential
predication. For further discussion, see Jerome Shaffer, "Existence, Predication, and tbe Ontological
Argument," in The Many-Faced Argument, edited by John Hick and Artbur McGill, pp. 244-45.
149

150

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

119

direction and to a significantly different version of the modal argument that does
not rely upon the claim that necessary existence is a predicate of God.
PLANTINGA'S VERSION OF THE MODAL ARGUMENT
Plantinga's modal argument relies upon a completely different interpretation of
PMP that does not treat necessary existence as a predicate and that is informed by
possible world theory. According to Plantinga, even if the versions of the modal
argument defended by Hartshorne and Malcolm are successful and necessary
existence is a perfection, their fatal flaw is that they can only establish that such a
perfect being exists in some possible world. But our particular, actual world may
not be the world in which that perfection exists. 154 Plantinga embraces the
theological claim made by Findlay by noting that the great-making qualities that
are relevant to making God a proper object of worship include not only those
actual properties that God has in this world but also those properties that God has
in other possible worlds. After all, a theist does not simply believe that God is the
most perfect being in this particular world and that possibly in some other world
God is less than perfect. 155 Plantinga thus begins his version of the modal proof by
a reexamination of the notion of perfection and by introducing the notions of
maximal excellence and maximal greatness. A being G has maximal excellence
entails that G has certain properties (the usual properties of omnipotence,
omniscience, and so forth) in a given world W. A being G has maximal greatness
entails that G has maximal excellence in every possible world. Plantinga's version
of PMP (PPMP) is the assumption that it is possible that maximal greatness is
exemplified, that is, it is possible that there is a being G in some possible world W
that has maximal excellence in every possible world. Assuming S5, if G has
maximal greatness in some world W, then 'G has maximal greatness in W would
be necessarily true in W; however, since necessary truth cannot vary from world to
world, then if 'G has maximal greatness in W is necessarily true in any possible
world, it is necessarily true in all possible worlds. Therefore, there must be a being
G that has maximal greatness (or, in another formulation, unsurpassable greatness)
in every possible world, including the actual world in which we live, and 'G has
maximal greatness' is necessarily true.
Plantinga's version of the modal argument depends upon the assumption that
PPMP is true. Plantinga himself thinks that this assumption is true and that his
modal version of the ontological argument is thus sound; however, he does not
think that it is possible to prove that PPMP is true. At the same time, since he does
not think that it is possible to prove that PPMP is false, he thinks that, minimally,
he has shown that it is rational for the theist to accept the modal ontological
argument since it is rational to accept PPMP. Plantinga thus maintains that his
modal proof establishes theism as a rational belief even if no conclusive proof can
154 Alvin Plantinga, The Nature of Necessity (Oxford: Clarendon Press, 1974), p. 213. This claim of
Plantinga's is ambiguous. What Hartshorne and Malcolm have shown is that a being necessarily exists
(exists in all possible worlds) and that this same being has perfection in at least one possible world. So
we should understand Plantinga's claim to mean that while a necessarily existing being exists in this
actual world, this world may not be the one in which it has perfection. The following description of
Plantinga's argument follows pp. 214-217.
155 Ibid., pp. 213-14.

120

ANALYTIC PHILOSOPHY OF REDGION

be offered for the existence of God, a position that parallels his free will defense to
the problem of evil (discussed in Chapter VI). Whether one accepts Plantinga's
argument comes down to whether one accepts PPMP, and whether one thinks that
it is rational to accept his modal argument comes down to whether one thinks that
it is rational to accept PPMP. 156
IS MAXIMAL GREATNESS (UNSURPASSABLE GREATNESS)
POSSIBLE?
Plantinga's modal version of the ontological argument and his attempt to
reconstruct Hartshorne and Malcolm's version of the argument relies upon the
assumption that it is possible that maximal greatness is exemplified in some
possible world W, that is, PPMP. Several objections have been raised against the
property of maximal greatness that cast doubt on Plantinga's conclusion that it is at
least rational for a person to accept PPMP. I will consider here only two of the
more serious and telling kinds of objections to Plantinga's property of maximal
greatness and the resulting effects upon the rationality of the acceptance of PPMP.
Difficulty arises when, as Michael Tooley shows, Plantinga's treatment of the
predicate maximal greatness is generalized for all predicates as follows: for any
predicate P and for any individual x, x is maximally P if and only if x exists in all
possible worlds and P occurs in all possible worlds. Given that predicates other
than greatness can be exemplified to a maximal degree, then contradictory results
follow. For example, x can be a maximal universal solvent, capable of dissolving
everything in every possible world, while y can be maximally insoluble, incapable
of being dissolved by anything. It is possible to construct ontological arguments,
paralleling Plantinga's for both x and y, but of course x and y cannot both exist. 15?
Tooley argues further that Plantinga's argument can be used to establish the
exemplification of predicates in all possible worlds that are already known in this
world to be empirically false and hence incapable of being exemplified in all
possible worldS. 158 Without an effective retort to this criticism, it does not seem
rational to accept PPMP.
A final consideration, which amounts to an apparent reductio of PPMP and
Plantinga's version of the ontological argument, is that it is possible to use the
same technique used by Plantinga to establish that predicates are exemplified that
appear to be explicitly incompatible with the exemplification of maximal
greatness. For example, Tooley argues that it is possible to use Plantinga's
argument to establish the existence of the Devil, that is, the exemplification of
maximal evil, in which case we get the demonological argument. If objections are
raised by the theist to the possibility of the exemplification of such a predicate
(based upon claims about the logical compatibility of omnipotence and evil), there
are other troubling predicates that do not raise this issue. Perhaps most plausibly
and convincingly, Richard Gale argues that the property of being a morally
unjustified evil in any possible world is logically incompatible with the

156
157

424.
158

See ibid., pp. 220-21.


Michael Tooley, "Plantinga's Defence of the Ontological Argument," Mind, Vol. 90, 1981, p.

Ibid., p. 425-26.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

121

exemplification of unsurpassable greatness in every possible world. 159 As Gale


notes, for those theists who accept PPMP and Plantinga's version of the modal
argument, the problem of evil becomes a modal problem of evil. The normal way
of addressing the evidential problem of evil is to consider the various claims made
concerning the existence of evil in the world and then construct a plausible
theodicy in response to those claims. But if the existence of a morally unjustified
evil is logically incompatible with the existence of unsurpassable greatness, then
morally unjustified evil cannot exist in any possible world, and the notion of a
morally unjustified evil must consequently be logically contradictory. The crucial
question changes from the empirical question of whether morally unjustified evil
actually exists in this world to the modal question of whether morally unjustified
evil is logically possible. 16o William Rowe's example of the fawn that dies a
horrible, excruciating death during a forest fire for no good reason (discussed in
detail in Chapter VI) thus must be logically impossible given the existence of
un surpassable greatness. The result for Plantinga's claim that it is epistemically
rational for the theist to believe PPMP is to pit the relative plausibility of PPMP
(the possible existence of maximal greatness) against the possibility of the
suffering fawn since, if this argument is correct, the possibility of the two
conjointly is not logically possible. While we must rely upon intuitions to assess
relative possibilities, the suffering fawn appears to be imminently more plausible
as evidenced by the attention given by nontheists and theists alike to discussion of
the case as a part of the long-lived evidential problem of evil.
Plantinga himself recognizes that there are other predicates whose possibility is
not compatible with maximal greatness, including near-maximality (exemplified in
any being if and only if it does not exist in every possible world but has a degree of
greatness not exceeded by any other being in any world in which it exists) and nomaximality (exemplified in any being which is such that there is no maximally
great being), but he thinks that maximal greatness is a more plausible predicate.
Plantinga does not claim that his modal version of the ontological argument is an
argument in natural theology that will appeal, on the basis of reason, to all rational
human beings. Sane and rational people, he claims, might consider PPMP and
reject it in favor of no-maximality. He seems to think that the choice between
PPMP and other predicates that are incompatible with maximal greatness is a "toss
up" and that it is hence rational for the theist to accept PPMP on the grounds that
he has shown that PPMP is not logically contradictory. However, if the arguments
offered here are sound, then maximal greatness is not as likely to be exemplified as
other predicates with which it is incompatible, and thus, if this is true, then the
belief in PPMP is not rationally justified.
One final objection to Plantinga's modal argument concerns the actual modal
logic of his argument. As J. L. Mackie notes,161 Plantinga is committed to a system
of modal logic that does not allow for iterated modalities, the "stacking" of
modalities where one may have such an expression as "It is possible that it is
necessary that it is possible that p." There are several modal systems that deny the

159

160
161

Richard Gale, On the Nature and Existence o/God, pp. 227-28.


Ibid., p. 229.
J. L. Mackie, The Miracle o/Theism. pp. 57ff.

122

ANALYTIC PHILOSOPHY OF RELIGION

legitimacy of iterated modalities, including S5, whose consistency is widely


accepted and not at issue here (more accurately, in other systems iterated
modalities reduce to noniterated ones in systems such as S5). However, Plantinga's
treatment of logical possibility raises issues that are quite different from those
found in the use of such systems as S5. His use of "world-indexed properties"
makes the properties that an individual x possesses in one possible world W
dependent upon the properties that x possesses in another possible world W.
Plantinga asks us to consider the property of being snub-nosed and to suppose that
it is true of Socrates in this world that he was snub-nosed. If we call this world W,
then, in some possible world w", in which Socrates does not have the property of
being snub-nosed, he still has the property of being snub-nosed in W. Thus, the
properties that Socrates has in w" depend upon the properties that he has in W. 162
The introduction of such world-indexed properties makes the features of a
particular possible world dependent upon the features of other possible worlds.
Mackie points out that such properties make it impossible for Plantinga to have the
same notion of logical possibility as is found in S5, the modal system that he
embraces. According to S5, in which possible worlds are independent of each
other, a noncontradiction is explained in terms of logical possibility that is, in turn,
explained in terms of at least one possible world in which that possibility is
actualized. However, if the features of different possible worlds are not logically
independent of one another, then the actualization of a noncontradictory property
in one possible world may well be logically incompatible with the actualization of
a different noncontradictory property in another possible world. 163 World-indexed
properties thus appear to undermine the kind of logical possibility that Plantinga
needs for the claim that maximal greatness is logically possible.
THE COSMOLOGICAL ARGUMENT
As we have seen, the different modal versions of the ontological argument
depend upon some form of the assumption that the existence of a necessary being
is possible, either in the form of PMP (the existence of modal perfection is
possible) assumed by Hartshorne and Malcolm or in the from of PPMP (the
exemplification of maximal greatness is possible) assumed by Plantinga. While
Plantinga seems content with simply claiming that belief in this assumption is at
least as rational as belief in its denial, Hartshorne and others have maintained that
positive support for the assumption comes from the cosmological argument. 164 For
Hartshorne, the ontological cannot do the job of proving the existence of a
necessary being by itself since it needs the support from the cosmological
argument for legitimizing the assumption of PMP. This connection with the
ontological argument is a major reason (along with the developments in science
discussed below and in Chapter V) for the resurgence of interest in the
cosmological argument in the second half of the twentieth century. In the case of
the cosmological argument, the major scientific development influencing the
debate is the big bang theory of modern physics and astronomy.
162

For Plantinga's discussion of world-indexed properties, see The Nature of Necessity, pp. 62ff.

163

J. L. Mackie, The Miracle of Theism, p. 60.

164 Kant, of course, argued just the converse, namely, that the cosmological argument depends upon
the ontological argument.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

123

The phrase "the cosmological argument" is something of a misnomer since the


cosmological argument is not a single argument; there are several distinct forms of
the cosmological argument. 165 The plurality of the different arguments is
recognized by most contemporary thinkers. Even Aquinas treated the first three of
his "Five Ways" as different arguments, though these are different versions of the
cosmological argument. William Rowe calls the cosmological argument "a family
of arguments," and Richard Gale and J. L. Mackie both examine the argument in
chapters entitled "Cosmological Arguments.,,166 I cannot examine all the different
versions of the cosmological arguments here, so I will take the cosmological
argument in the singular to be an abstraction from a family of arguments that share
the common characteristic of attempting to prove the existence of God by
appealing to the existence of certain facts, events, objects, or phenomena in the
world and then using some principle or law involving causality to establish an
original cause of those facts, events, objects, or phenomena.
Certain criticisms and limitations of the traditional forms of the cosmological
argument have now become commonplace. For example, many critics have
observed that the cosmological argument, even if it is sound, does not establish that
the first cause must have the traditional attributes of the God of theism. Separate
arguments or addenda to the main cosmological arguments are necessary to
establish that the first cause is a personal being with the attributes usually
attributed to God and to establish that the first cause - even if it did exist at some
point in the past - continues to exist now. Standard objections also include the way
in which the possibility of an infinite regress of causes is so summarily addressed
and dismissed in traditional versions of the argument. In the famed debate between
Bertrand Russell and Frederick Copleston, Russell raises what has now become a
rather standard objection that the cosmological argument commits a form of the
part-whole fallacy by arguing that it is a mistake to attribute a cause to the
universe, taken as a whole, on the grounds that causes are attributed to individual
events and objects in the universe. 167 The main thrust of Russell's point is a logical
one, an example of what we might now call, following Gilbert Ryle, a "category
mistake." It is as logically inappropriate, Russell claims, to argue that the universe
must have a cause from the claim that individual events must have causes as it is to
argue that the human race must have a mother because all individual human beings
have mothers. 168 Different versions both of these objections and various responses
to them are common in the contemporary literature. For example, William Rowe
responds to Russell that his objection is not a valid one because not all proponents
165 For various forms of the cosmological argument, see William L. Craig, The Cosmological
Argument from Plato to Liebniz (London: Macmillan, 1980).
166 William L. Rowe, The Cosmological Argument (Princeton, N.J.: Princeton University Press,
1975), p. 7; Richard Gale, On the Nature and Existence of God, Chapter 7; J. L. Mackie, The Miracle of
Theism, Chapter 5.
167 Bertrand Russell and F. C. Copleston, "The Existence of God - A Debate," a 1948 program of
the British Broadcasting Corporation, reprinted in A Modern Introduction to Philosophy, edited by Paul
Edwards and Arthur Pap (New York: The Free Press. 1957), p. 473. Edwards's introduction to this
section (pp. 372ff.) is particularly illuminating. The debate is also reprinted in The Existence of God,
edited by John Hick (New York: Macmillan, 1964).
168 This is essentially the same point made by Ronald W. Hepburn, Christianity and Paradox
(London: Watts Publishing, 1958), pp. 167-68.

124

ANALYTIC PHILOSOPHY OF REUGION

of the cosmological argument argue to the existence of a single cause of the


universe on the basis of claims about individual causes. 169 Some defenders of the
argument (for example, Rowe cites Samuel Clarke) simply rely upon a strong form
of the principle of sufficient reason to do the job.
THE PRINCIPLE OF SUFFICIENT REASON
The "principle or law involving causality" upon which the cosmological
argument relies is generally described as some form of the principle of sufficient
reason (PSR), and much controversy has continued to surround the use of this
principle in the contemporary literature. Much of the current discussion of the
cosmological argument focuses upon the PSR since, if it is granted, the argument
appears to proceed deductively and be sound. There are different forms of the PSR,
and formulations of the cosmological argument differ depending upon which form
of PSR is used. 170 The form of the PSR that is most commonly incorporated into
newer formulations of the cosmological argument appears to be a "strong form"
that requires that everything that exists or occurs must have a causal explanation of
its existence or occurrence, and this explanation must include a sufficient reason
why that particular thing exists or why that particular event occurs rather than
something else or rather than nothing at all. 171 If this principle is applied to the
universe, and if everything in the universe exists contingently, then PSR requires
that there be some causal explanation of the existence of the universe that explains
why the universe exists as it does instead of nothing at all, and this causal
explanation must be noncontingent, that is, a necessarily existing being. Unless the
cause for the contingent universe or for an infinitely regressing contingent chain of
causal explanations is located in something "outside" the universe, then, it is
argued, a complete explanation is not provided. Hence, the necessity for the
existence of a necessarily existent being. 172
The basis of Russell's objection that the cosmological argument commits a partwhole fallacy is Hume's claim that if the causes of each of the members of a
collection are provided, then there is no need to give a further explanation of the
collection itself. So, Hume says in Part IX of his Dialogues Concerning Natural
Religion, if the particular cause of twenty particles is given, then it is unreasonable
to ask for the cause of the whole group of particles. Rowe, no true defender of the
cosmological argument or theism, defends the PSR against this objection by
arguing that this principle of Hume's is plausible enough for finite collections but
169 See William L. Rowe, ''Two Criticisms of the Cosmological Argument," The Monist, VoL 54, no.
3, 1970. Reprinted in Philosophy of Religion: Selected Readings, Third Edition, edited by William L.
Rowe and William J. Wainwright (Fort Worth, Tex.: Harcourt Brace College Publishers, 1973). Page
numbers refer to this reprint.
170 The most thorough discussion of PSR and examples of how different versions of the
cosmological argument rely upon different versions of PSR are found in William L. Rowe, "The
Cosmological Argument and the Principle of Sufficient Reason," Man and World, VoL 1, no. 2, 1968.
Reprinted in William L. Rowe, The Cosmological Argument, pp. 60-114.
171 Compare J. L. Mackie, The Miracle of Theism, p. 82, and Rowe, ibid., 145.
172 Bruce Reichenbach explicitly states the form of the PSR that he thinks the cosmological
argument requires as one that says that all contingent beings require a sufficient explanation for their
existence. See Bruce Reichenbach, The Cosmological Argument: A Reassessment (Springfield, ill.:
Charles C. Thomas, 1972), p. 68.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

125

is false for any infinite collection in which the causal explanation for members of
the collection is given in terms of other members of that collection. 173 Rowe argues
that if we consider the collection of men, M, and if we assume that M has an
infinite number of members and if we explain the existence of each member of M
by saying that each member is begotten by some other member, then, even though
we have explained the existence of each member of M, we have not provided an
explanation for M itself. To do so, he claims, we must explain why M has any
members at all.174 To explain why a particular collection exists, following the PSR,
we need to explain why it has the particular members it has rather than none at all.
In the case of an infinite set in which the existence of the members of that set are
explained in terms of other members of that set, Rowe maintains that this cannot be
done. Thus, Rowe thinks, although there may be other problems with the
cosmological argument, the PSR is immune to this particular objection.
Ironically, Rowe himself has provided the framework within which a retort to
his defense of the PSR is possible. 175 The particular example used by Rowe of the
class of men and the explanation in terms of begetting is plausible enough since
begetting is a nonreflexive relation. Eventually, it seems, we need some further
explanation, "outside" the class of men, in terms of amino acids, genetics, and
perhaps complex conditions conducive to life to explain why the class of men has
any members at all. However, the limiting case for the cosmological argument is
the collection of all contingent things. If we consider the collection of all
contingent things, that is, the entire universe, then Rowe's defense of the PSR
becomes unconvincing. Even if we agree that to explain why this universe exists,
we must explain why we have this particular collection of contingent things
existing rather than no collection at all, it is not clear at all that we would be forced
to go "outside" the universe of contingent beings to provide such an explanation.
And a fortiori, it seems that it would be impossible to do so. Suppose we provide
an explanation in terms of the following cosmological principle:
CP: A necessary being, N, created the contingent universe, U.
CP describes the creation of a contingent universe, an infinite collection of
contingent things and events, by a necessarily existing being. If, on the one hand,
this act of creation is itself a contingent act, then it is a part of the infinite
collection of contingent events for which an explanation must be provided, and it
cannot itself be that explanation. If, on the other hand, this act of creation is a
necessary act, then it appears that the universe must exist necessarily, and, as
Leibniz claimed, there are no contingent things or events, an equally unattractive
alternative. 176

William L. Rowe, ''Two Criticisms of the Cosmological Argument," pp. 151ff.


Ibid., 152.
175 See William L. Rowe, The Cosmological Argument, pp. 100-101.
176 There are other difficulties as well with Rowe's defense of the PSR on this score. See Richard
Gale, On the Existence and Nature of God, pp. 259ff.
173

174

126

ANALYTIC PHILOSOPHY OF RELIGION

SWINBURNE'S INDUCTIVE COSMOLOGICAL ARGUMENT


Although traditional versions of the cosmological arguments are a posteriori
because they all depend upon some empirical claim about the universe, the
reasoning process in the arguments is deductive. The role of some version of the
PSR as a premise in such arguments is crucial. While some defenders of the
cosmological argument, such as Bruce Reichenbach, have claimed that the PSR is
synthetic a priori and "a basic and intuitive principle of human reasoning"
comparable to the law of noncontradiction, other critics of the argument, such as J.
L. Mackie, have objected that there are no grounds for supporting the PSR and no
reason to believe that its demand could ever be met, even in principle. 177 While
agreement on the deductive cosmological argument based upon the PSR as a
necessary truth might not be forthcoming, Richard Swinburne has argued that it is
possible to provide a probabilistic confirmation of the PSR and hence of theism.
The result is his inductive cosmological argument. 178
Swinburne explains that empirical data e may confirm a hypothesis h by making
that hypothesis more probable than it would be without e. In this case, Swinburne
says, the result is a good "C-inductive argument." Whatever judgment we may
make regarding the probability of h, that judgment may be changed in the face of
additional evidence not yet gathered or in the face of future experience. When we
consider the probability of h on the basis of certain background knowledge that we
already possess and then calculate the effect of considering e, the probability of a
hypothesis h is increased if e is more likely to occur on the basis of the background
knowledge and h than it is on the basis of the background knowledge alone. Thus,
where k stands for the background knowledge and P(x/y) stands for the probability
of x on the basis of y, P(elkandh) > P(elk). Swinburne admits that P(elkandh) is
not high, but he claims that, whatever the probability of P(elkandh), it is higher
than the probability of P( elk). 179 Thus, on this account, the probability of the
existence of the physical universe on the basis of h, the existence of a personal
creator, along with our store of scientific knowledge of the universe is greater than
the probability of the existence of the universe on the basis of our store of scientific
knowledge about the universe alone. According to Swinburne, the additional
probability of this "personal explanation" means that we have a good C-inductive
argument for the existence of God. 1SO There have been several objections raised to
this inductive version of the cosmological argument. 181 Here, I will simply take
Swinburne's account at face value and evaluate the claim that P(elkandh) > P(elk).
While talking about probability on such a grand, sweeping, general scale is
difficult, it is clear that we need a much clearer idea of what is involved when we
talk about the probability of P( elk), which leads into a consideration ofthe current
best scientific theories concerning the origin of the universe.
177 Bruce Reichenbach, The Cosmological Argument: A Reassessment, p. 69, and J. L. Mackie, The
Miracle of Theism, p. 85. Given Kant's attack on the ontological argument, it is ironic that Reichenbach
likens the PSR to the principle of causality and then gives a very Kantian defense of it.
178 Richard Swinburne, The Existence of God (Oxford: Oxford University Press, 1979), Chapter 7.
179 Ibid., p. 130.
180 Swinburne also develops an inductive version of the teleological argument for the existence of
God. These probabilistic arguments, along with miracles and religious experience, are parts of
Swinburne's "cumulative case" for the existence of God.
181 See, for example, J. L. Mackie, The Miracle of Theism, pp. 95ff.

THE NATURE OF GOD AND ARGUMENTS FOR TIlE EXISTENCE OF GOD

127

THE BIG BANG


Using Einstein's general theory of relativity, cosmologists now set the
beginning of the universe with the initial "big bang" at something close to fifteen
billion years ago. 182 The big bang theory came about as the result of the solutions
to the field equations in Einstein's general theory of relativity developed by the
Russian mathematician Alexander Friedman. When Einstein applied his field
equations to the universe with general relativity, the resulting prediction was either
an expanding or contracting universe (observations of the famed redshifts caused
by the Doppler effect confirmed an expanding universe). Since Einstein could not
bring himself to accept this conclusion, he introduced what he called "the
cosmological constant," which inserted a rather ad hoc force of repulsion into the
universe to balance the force of gravity and provide a static universe. 183 Friedman's
solutions to Einstein's field equations made the general theory of relativity
compatible with an expanding universe and led to the Friedman and Lemaitre
theory that such an expanding universe originated from a state of enormous
density. While there are still different models of the big bang, observations of the
relative velocities with which more distant galaxies compared with closer galaxies
are receding from us (described by Hubble's Law) confirm some form of an
expanding universe.
The version of the big bang theory that is now most widely accepted by
physicists is the one developed by the American physicist George Gamow, which
accounts for the initial explosion in terms of nuclear processes. The original big
bang took place in a unity of enormous temperatures and colossal density, which
contained all of the matter in the universe, which then exploded and expanded at a
tremendous velocity.184 From the initial explosion, the universe expanded and
cooled. 18s For the first several hundred thousand years, the universe was simply a
hot and opaque cloud of radiation and matter. As matter began to localize, the
182 There are many scientific descriptions available of the big bang. The description here is a
synthesis of the most commonly accepted current theories and draws primarily upon a popular college
textbook for astronomy and cosmology, George Abell, David Morrison, and Sidney Wolff, Exploring
the Universe, Fifth Edition, (Philadelphia, Pa.: Saunders College Publishing, 1987), Chapter 37, pp.
656-76. For a very lay-accessible account, see also James S. Trefil, The Moment of Creation (New
York: Macmillan, 1983), Chapter 1. For descriptions of the big bang employed by philosophers of
religion, see Quentin Smith, 'The Uncaused Beginning of the Universe," and "Atheism, Theism, and
Big Bang Cosmology," in Theism, Atheism, and Big Bang Cosmology, William Lang Craig and Quentin
Smith (Oxford: Clarendon Press, 1993), and Kenneth Nelson, "Evolution and the Argument from
Design," Religious Studies, Vol. 14, 1978, pp. 427-28.
183 There are many explanations of this development for the layperson. A very popular recent source
is Brain Greene, The Elegant Universe (New York: Vintage Books, 1999), pp. 82ff. Einstein later was
to say that introducing the ad hoc adjustment to relativity theory to ensure a static universe was the
biggest mistake of his life.
184 For a more detailed scientific description of the big bang, see Abell et al., Exploring the
Universe, pp. 665ff. The version discussed here is what physicists call the "standard model" of the big
bang.
185 An interesting point here involves the problem of the perspective of any theoretical observer that
epistemologists (but probably not physicists) will find interesting. The common conception of the big
bang where one imagines a bright flash of light in distant space will not work. Since all of the mass of
the universe was contained in the original singularity, there was no "outside" to the big bang. The big
bang occupied all of existing space and would have to be imagined more like a sudden bright flash of
the entire visual field.

128

ANALYTIC PHILOSOPHY OF RELIGION

universe became transparent. Gradually, the temperature dropped to a level where


stable atoms of hydrogen and helium formed, and a billion years or so after the big
bang, stars and galaxies began to stabilize as a result of the gravitational forces
pulling the atoms of helium and hydrogen together. In the interiors of the stars,
nuclear reactions reheated the cosmic particles, and the first heavy elements were
formed. The most widely accepted understanding of how our own solar system
began is based upon the theory first suggested by Immanuel Kant and Simon de
Laplace. The Kant-Laplace model is known to physicists as the nebular
hypothesis .186 According to this model, our solar system is considered to be the
result - some 4.7 billions years ago - of a giant solar nebular that rotated around a
central concentration of tremendously hot matter. The center of the nebula became
our sun, and the planets are the result of concentrations of matter that were "spun
off' by the nebula. The heat and ultraviolet rays from the sun acted upon the gases
present on earth to form amino acids, and from there life eventually evolved over
some billions of years to the form in which we now know it.
THE BIG BANG AND THE COSMOLOGICAL ARGUMENT
The significant theological implications of the big bang arise from the period
immediately "prior to" and immediately "following" the big bang itself. Even
making sense of talking about any time prior to the big bang is troublesome since
physicists agree that time itself began with the big bang; however, the focus of
attention here will be the singularity within which the big bang occurred. The
rather direct and simple way in which theists might respond to the big bang is
simply to ask, "Well, what caused the big bang? Why cannot God be understood as
the first cause of the big bang?"
Although the initial singularity from which the big bang arose is described
differently by different models of the big bang theory, all agree that this singularity
was one in which the normal physical and cosmological laws and features
commonly attributed to the universe did not exist. Normally, scientific descriptions
of the big bang are thought to be unable to penetrate what is referred to as Planck
time (which is 10 -43 after the big bang).187 Before this point in time, normal
gravitation breaks down into quantum mechanics. At the turn of the twenty-first
century, scientists are still in search of the grand unified theory that will allow
them to understand and explain beyond Planck's time to the moment of the big
bang itself. Whether further penetration by scientific explanation is even
theoretically possible is very debatable. 188 Various scholars have described the
singularity from which the big bang occurred, as well as the initial moments, as
completely chaotic and lawless and thus impenetrable - even theoretically - by
scientific explanation. The limits of the ability of science and scientific theory to
penetrate to the exact moment of the big bang is embodied in Stephen Hawking's
Abell et a!., Exploring the Universe. p. 247.
Planck time is named after Max Planck, who is credited with a major role in developing quantum
theory.
188 Some physicists think that superstring theory is a candidate for a theory that will unify general
relativity and quantum theory by piercing the Planck limit on the snpermicroscopic level. However,
even superstring theory is not proffered as a way of piercing the Planck limit in regards to the big bang.
See Brian Greene, The Elegant Universe.
186
IS7

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

129

principle of ignorance. According to the principle of ignorance, the singularity


from which the big bang developed is a condition under which the normal,
classical concepts of space and time theoretically "break down," as do all of the
other laws of physics. 189 Such a breakdown in all known natural laws means that
the possibility of all scientific explanation as well as all scientific prediction breaks
down as well. In other words, no prediction could even theoretically be made about
what kind of universe the big bang would produce or even that it would produce a
universe rather than an infinite, chaotic nonuniverse. Such a situation is one that
"would thus emit all configurations of particles with equal probability."I90 Paul
Davies describes the singularity of the big bang as "the instantaneous suspension of
physical laws, the sudden abrupt flash of lawlessness that allowed something to
come out of nothing.,,191
GOD AND THE GAP IN THE BIG BANG
The inability of scientific explanations to penetrate the big bang singularity
appears to leave a great cosmic "gap," which, of course, the theist fills with God as
the creator. Davies's description of the big bang as allowing for something coming
from nothing without any possible scientific explanation sets the stage for a theistic
explanation to "fill the gap" with the use of a creator God as the cause of the big
bang. Both theists and physicists have seen the big bang theory as leaving open
such an opportunity for a theistic explanation. 192 Thus, so far as Richard
Swinburne's inductive cosmological argument discussed earlier in this chapter is
concerned, it seems that the existence of the explanatory gap for the universe in big
bang theory at least allows for the possibility of adding some additional hypothesis
involving the existence of God that would increase our ability to explain the
universe.
THE KALAM COSMOLOGICAL ARGUMENT
Perhaps the best known and most clearly formulated version of the cosmological
argument that incorporates the fundamental concepts of big bang theory is found in
the work of William Lane Craig. Craig seizes upon the fact that before the big bang
nothing can be said to exist and that all matter and energy as well as space-time
itself began with the big bang to construct an updated version of what is called the
kaliim cosmological argument for the existence of GOd. 193 The kaliim cosmological
argument avoids questions about an infinite regress of causes by explicitly
189 Stephen Hawking, "Breakdown of Predictability in Gravitational Collapse," Physical Review,
Vol. D14, 1976, p. 2460. Quoted by Quentin Smith in "Atheism, Theism, and Big Bang Cosmology," in
Theism, Atheism, and Big Bang Cosmology, p. 198.
190 Ibid.
191 Paul Davies, The Edge of Infinity (New York: Simon and Schuster, 1981), p. 161.
192 For example, Stephen Hawking thinks that his oscillating view of the universe is preferable to the
standard big bang model - in part, at least - because it has no beginning and thus, in contrast to the
standard big bang model, does not leave a gap for the beginning of space-time, which allows an appeal
to God. See Stephen Hawking, A Brief History of Time (New York, Bantam Books, 1988).
193 See William Lane Craig, The Kaliim Cosmological Argument (New York: Harper and Row,
1979), and "The Caused Beginning of the Universe," in Theism, Atheism, and Big Bang Cosmology.
Craig borrows the name "kaIam" for his version of the cosmological from the movement in Islam
comparable to natural theology in the Western, Judeo-Christian tradition.

l30

ANALYTIC PHILOSOPHY OF RELIGION

claiming that the universe began to exist at some point in the past. An "actual
infinite" in time is impossible, and past time is thus explicitly regarded as finite.
Craig argues that if we grant the premises,
1. Whatever begins to exist has a cause,
and
2. The universe began to exist,
then we must conclude that
C. The universe has a cause.
This certainly is a valid argument, and Craig thinks that the cause in the
conclusion (C) can only be God. 194 A personal God is the only plausible solution to
a puzzle that he thinks exists in the big bang theory. He claims that if the cause of
the big bang is located in the singularity itself, then it must have existed "eternally"
without a beginning and with no change at all, and then, at some finite point,
approximately fifteen billion years ago, it caused the universe to come into
existence. Craig thinks that such a scenario is "exceedingly odd," since such a
cosmic cause, in his words,
cannot have any beginning of its existence nor any prior cause. Nor can there
have been any changes in this cause, either in its nature or operations, prior to
the beginning of the universe. It just exists changelessly without any beginning,
and a finite time ago it brought the universe into existence .... The cause is in
some sense eternal and yet the effect which it produced is not eternal, but began
to exist a finite time ago. How can this be? If the necessary and sufficient
conditions for the production of the effect are eternal, then why isn't the effect
eternal? How can all the causal conditions sufficient for the production of the
effect be changelessly existent and yet the effect not also be existent along with
the cause? How can the cause exist without the effect?195
The kind of cause that Craig attributes to the creator is "agent causation" and is
the result of free will, whereby God can introduce new conditions into the universe
that were not present earlier and thereby produce new effects. 196 Craig adds to this
argument support that he musters from the anthropic principle and claims that "the
incredibly complex and delicately balanced nexus of initial conditions [for the big
bang] necessary for intelligent life" can only be explained as the product of
intelligent design by a "personal Creator." The existence of God is thus, for Craig,
"confirmed by science."
OBJECTIONS TO THE KALAM COSMOLOGICAL ARGUMENT
Paul Draper has objected rather convincingly that Craig's argument involves an
equivocation on what it means for a thing to begin to exist. 197 In one sense, a thing
begins to exist within time, and in a second sense, a thing begins to exist within or
194 William Lane Craig, "Scientific Confirmation of the Cosmological Argument," in Philosophy of
Religion: An Anthology, edited by Louis P. Pojman, pp. 38ff.
195 Ibid., p. 38.
196 Ibid., p. 39.
197 Paul Draper, "A Critique of the Kalam Cosmological Argument," in Louis P. Pojman, Philosophy
of Religion: An Anthology, edited by Louis P. Pojman, pp. 42-47.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

131

with time. Craig claims in his first premise that everything that begins to exist has a
cause, and he offers empirical, scientific evidence to buttress this claim. However,
experience can only provide support for the claim that things that begin to exist
within time have a cause; so it seems that Craig must be using "begin to exist" in
the first sense in premise one. In premise two, however, if "begins to exist" means
"begins to exist within time," then all of the evidence to which Craig appeals
would serve to undermine the truth of this second premise and hence the soundness
of his argument since, as we have seen, according to big bang theory time began to
exist with the big bang. Since time did not exist before the big bang, the universe
did not come into existence within time. Quentin Smith has also objected that
rigorous scientific investigation on the subatomic level has provided us with
empirical evidence that quantum acausality does occur with the uncaused
emergence of energy or particles (virtual particles),198 which undermines the truth
of the first premise even if "begins to exist" is restricted to "begins to exist in
time."
In addition to these objections, Craig's version of the kaliim cosmological
argument faces problems that arise from his use of big bang theory to buttress his
argument. Once we use the big bang theory to trace the beginning of the universe
back to the big bang singularity, the most compelling and important point is that, at
that point, "all bets are off." No natural laws - no ordinary notions of space or time
or laws of physics - can be applied. The singularity is inherently chaotic and
unpredictable in the most fundamental sense, so Craig'S claim that he finds the
possibility of an "eternal" cause existing before the big bang without its effects
"exceedingly odd" has no force at all. Everything and nothing we might say about
the big bang singularity would be exceedingly odd; therefore, it should come as no
surprise and it should not seem odd that the classical notions of cause and effect
lead to odd consequences when applied to the big bang singularity. That is exactly
what we should expect. But what is true of the classical notions of space and time
and cause and effect must also be true of the most fundamental metaphysical
notions. Appealing to a kind of Kantian transcendental argument to make a priori,
metaphysical claims about the nature of the singularity will not work. There is
absolutely no reason that an ordered universe must arise from the big bang
singUlarity and especially no necessity that a universe ordered in the exact way in
which this particular universe is ordered must arise. Whatever speculation one may
be led to make about the source of the singularity or the source of the universe
developing from that singularity cannot extend beyond our understanding of cause
that is derived from the known universe. Since we cannot make sense of applying
any of the classical laws or notions of science or metaphysics to the big bang
singularity, then a fortiori we cannot make sense of applying such notions as
"agent causation," as Craig attempts, since any understanding we might have of
agent causation is dependent upon those known laws. Craig explains his notion of
agent causation with the following analogy:

198 Quentin Smith, "The Uncaused Beginning of the Universe," in Theism, Atheism, and Big Bang
Cosrrwlogy, p. 123. For a full discussion concerning the caused vs. uncaused - the theistic vs. the
atheistic - interpretations of the big bang theory, which space here does not permit, see the exchange
between Craig and Smith in Theism, Atheism, and Big Bang Cosmology.

132

ANALYTIC PHILOSOPHY OF RELIGION

For example, a man sitting from eternity could will to stand up; thus a temporal
effect arises from an eternally existing agent. Similarly, a finite time ago a
Creator endowed with free will could have willed to bring the world into being
at that moment. 199
Of course whatever force there is to the analogy of a sitting man rising exists
only in this universe with our present understanding of its natural laws and the
classical notions of space and time. Certainly, applying the notion of agency
causation to the big bang is more exceedingly odd than simply adopting the notion
of an uncaused big bang. 2O If a theist embraces enough of the big bang theory to
reach the big bang singularity, then he cannot suddenly abandon everything that
science tells us about that singularity to introduce arbitrarily the notion of an agent
acting causally in a fashion that resembles anything like our ordinary
understanding of agent causation.
The same result applies to Richard Swinburne's inductive cosmological
argument discussed earlier in this chapter. Swinburne's claim is that the probability
of the evidence that we have about the universe on the basis of our scientific
knowledge plus the hypothesis of God is greater than the probability of that
evidence on the basis of our scientific knowledge alone, that is, P( e!kandh) >
P( elk)?OI For Swinburne, it is less unlikely that God would exist uncaused than it is
that a universe would exist uncaused. But what this somewhat lengthy
consideration of the best available scientific explanation of the universe reveals is
that talking about the ultimate cause of the universe is not possible in any
meaningful sense. If all our scientific laws and concepts break down in the original
singularity before the big bang, then, as I have claimed, a fortiori, our
understanding of personal agency breaks down as well. Swinburne's claim that a
"personal explanation"adds to the probability of the universe is false because it is
impossible to assess the probability of the existence of the known evidence about
our universe on the basis of our scientific knowledge plus the existence of a
personal creator. In fact, in a very circuitous way, and in a way he never intended,
we seem to have confirmed Bertrand Russell's point from his 1948 BBC debate
with Frederick Copleston - namely, that to apply the notion of cause to the entire
universe is a logical absurdity.
THE ATHEISTIC KALAM COSMOLOGICAL ARGUMENT
In contrast with Craig's theistic interpretation of the big bang, Quentin Smith is
responsible for developing what is essentially an atheistic version of the kaIam
cosmological argument based upon big bang theory. Smith argues that
1. The big bang singularity is the earliest state of the universe.

Craig, "Scientific Confirmation of the Cosmological Argument," p. 39.


For a more detailed treatment of the notion of an uncaused big bang, see Quentin Smith, "The
Uncaused Beginning of the Universe," pp. I 25ff.
201 See above and Richard Swinburne, The Existence 01 God, pp. 130-32.
199

200

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

133

2. The big bang singularity must be inanimate given the extreme conditions of
infinite density and temperature.
3. The big bang singularity is fundamentally chaotic and lawless and does not
guarantee an animate universe will result (Hawking's principle of ignorance).
4. An animate universe is better than an inanimate universe (a theistic
theological principle).
5. If God exists, then God created the earliest state of the universe.
Therefore, God did not create the earliest state of the universe and God does not
exist.
Steps (1) through (3) are results of big bang cosmology embraced by both Craig
and Smith. Steps (4) and (5) are results of what Smith takes to be orthodox
Christian theology.
There appear to be several responses that a theist may make to this argument.
First, clearly this is not a "knock down" atheistic argument since it is based on big
bang cosmology, and there are other alternatives to the big bang as an explanation
for the beginning of the universe. A theist may reject the big bang theory altogether
and embrace some other view, such as Hawking's oscillating view, but whether
theism would fare any better under other cosmologies is doubtful. For example,
Hawking's oscillating universe has no beginning, and there is thus no role for God
as creator. Alternatively, the theist is best off attacking premise (4). Even if a theist
denies that there is a best possible world, there must be, as Smith notes, some
possible worlds that are "better" than others. Minimally, to be consistent with the
orthodox notion of God as a "personal" creator who values and loves human
beings, a universe with human beings (or some being very similar) must be
regarded as preferable to one without such beings. 202 Given that the original big
bang singularity provided no guarantee that an animate universe with human
beings would evolve, a theist might argue that God miraculously intervened along
the way to ensure that the universe would be animate. Now perhaps there is no
restriction that God create as efficiently as possible, but certainly the suggestion
that God intervened to ensure that the universe would be animate is forced,
artificial, and ad hoc. In any case, such a claim would be a completely theological
one and quite independent of big bang cosmology.
THE ANTHROPIC PRINCIPLE AND THE TELEOLOGICAL ARGUMENT
Recent discussion of the classical teleological argument for the existence of God
is examined in some detail in Chapter VI in connection with the treatment of the
problem of evil. I will examine here only the recent, more novel developments that
have generated discussions of a different nature concerning the teleological

202 See Smith, "The Uncaused Beginning of the Universe," pp. 201-202. Compare Richard
Swinburne, The Existence of God, pp. 145-51.

134

ANALYTIC PHILOSOPHY OF RELIGION

argument and that have given rise to different kinds of issues. 203 Perhaps the most
significant such development has been the attention given to what is now called the
anthropic principle (AP). Although it is frequently regarded as a very recent
discovery by scientists,204 the general features of the universe that are captured in
AP have arguably been explicitly recognized by scientists for over a hundred years.
For example, Alfred Russel Wallace, who, along with Charles Darwin, developed
evolutionary theory, marveled at the way in which the vast and complex inorganic
universe had to be precisely arranged in minute detail in order for organic life to
develop.z5 Certainly, however, the recent developments in cosmology involving
big bang theory (examined above) have increased our scientific knowledge about
the initial conditions of the universe enormously and provided significant
additional evidence to support the anthropic principle. On the one hand, then,
modern science has seriously undermined traditional arguments for the existence of
God, as we have seen above in the case of the cosmological argument; however, on
the other hand, theists have seized upon AP as a way of reformulating the
teleological argument in the face of our increased scientific knowledge of the
umverse.
AP,WAP,SAP
AP was first explicitly formulated by Brandon Carter in 1974?06 As with the
cosmological argument, there is no single version of this principle. Following
Carter, it is now common to distinguish a weak version (W AP) and a strong
version (SAP) of AP, and WAP and SAP are also formulated differently by
different figures. The fundamental characteristic notion that is focused upon by AP
is the extremely low antecedent probability that sophisticated conscious life could
ever exist in the universe given what scientists now know about the initial
conditions that gave rise to the universe. Slight, even relatively minuscule,
variations in any number of different, seemingly unrelated conditions, features, or
properties of the universe would have presumably meant that conscious life would
never have existed and that furthermore, a fortiori, no conscious experience of any
part or aspect of the universe would ever have taken place. Thus, almost ironically,
even given the extremely low initial probability that intelligent, carbon-based life
would exist, W AP says that any features of the universe that we human beings may
observe or discover (no matter how extraordinary or improbable they may seem)
are exactly the features that the universe must have in order for observers to exist

203 Parallel issues involve the cosmological argument as welL See John D. Barrow and Frank J.
Tipler, The Anthropic Cosmological Principle (Oxford: The Clarendon Press, 1986), especially section
2.9, pp. 103ff.
204 The coining of the phrase "the anthropic principle" is usually attributed to Brandon Carter in
1974.
205 Alfred Russel Wallace, Man's Place in the Universe (New York: McClure, Phillips and Co.,
1903).
206 See Brandon Carter, "The Anthropic Principle and Large Number Coincidences," in
Confrontation of Cosmological Theories with Observation, edited by M. S. Longair (Dordecht,
Netherlands: D. Reidel Publishers, 1974).

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

135

and to experience and discover anything,z07 According to W AP, given the wide
band of initial probabilities, the one set of conditions that allows for conscious life
and experience is regarded as infinitesimally small; therefore, the only explanation
for why the universe is as we observe it to be is that this is the only way it could be
in order for us to exist and for us to be able to observe it!208 But W AP offers no
explanation for why this extremely unlikely set of conditions came to exist. SAP
says that the eventual development of life in the universe is not simply a possible
development but that intelligent life had to develop at some point in the life of the
universe, and thus the universe must have the conditions, features, and properties
that allow intelligent, conscious life to develop.209
AP AND DESIGN
It is SAP that is of more interest and more significant consequence for the
teleological argument, and it is SAP that theists have used to buttress the

teleological argument. If the initial conditions of the universe were such that
eventually intelligent, carbon-based life had to develop, what is the best
explanation for this "fine-tuning" of the universe? Theists who base the
teleological argument upon the features of the universe identified by the anthropic
principle argue that the existence of God and the creation of the universe by God
provide a more reasonable and probable explanation of what is regarded as
"design" than explanations based upon simple scientific explanations or pure
chance. Such arguments are probabilistic arguments to the best explanation. M. A.
Corey's reasoning concerning the AP is typical of theists who defend the
teleological argument based on AP. Since we can never hope to establish the
original cause of the universe indubitably, the best we can do is to assess the
relative probability of different possible causes. Given the choice between the
plausibility of simply "stunning coincidences" on the one hand and creation by
God on the other to explain the existence of the very narrow band of initial
conditions of the universe that led to the existence of intelligent life identified in
AP, Corey maintains that the theistic account is the more reasonable and plausible
account. 210
Somewhat ironically, the probabilistic teleological argument based upon AP is
not structured in the same way as the classic form of the teleological argument
using analogy of proper proportionality. However, different theists still frequently
rely upon the use of analogy to justify the choice of the probability of creation over
2m John D. Barrow and Frank J. Tipler, The Anthropic Cosmological Argument (Oxford: Oxford
University Press, 1986), pp. 16ff. This book provides the most complete account of the antbropic
principle available.
208 For example, it is claimed that if the rate of expansion of the universe had been reduced by one
part in a million million, the universe would have collapsed upon itself. For a discussion of this and
other such conditions, see John Leslie, "Anthropic Principle, World Ensemble, Design," American
Philosophical Quarterly, Vol. 19, no. 2, 1982, pp. 141ff. See also Barrow and Tipler, ibid., pp. 288ff.
209 Barrow and Tipler, ibid., p. 2l.
210 M. A. Corey, God and the New Cosmology: The Anthropic Design Argument (Lanham, Md.:
Rowman and Littlefield Publishers, 1993), pp. 205-15. Corey gives significant weight to the fact that
some scientists, such as Stephen Hawking and Paul Davies, allow for some notion of principle of design
to account for the initial conditions of the universe. Compare Corey's restrained comments about
scientists with the exaggerated comments of William Craig, '''Theism and Physical Cosmology," in A
Companion to Philosophy of Religion, edited by Philip L. Quinn and Charles Taliaferro, p. 422.

136

ANALYTIC PHILOSOPHY OF RELIGION

other nontheistic explanations of these initial conditions of the universe.


Representative of the kind of analogies used by theists to support creative design
over chance is the one proposed by Richard Swinburne:
Suppose that a madman kidnaps a victim and shuts him in a room with a cardshuffling machine. The machine shuffles ten packs of cards simultaneously and
then draws a card from each pack and exhibits simultaneously the ten cards. The
kidnapper tells the victim that he will shortly set the machine to work and it will
exhibit its first draw, but that unless the draw consists of an ace of hearts from
each pack, the machine will simultaneously set off an explosion which will kill
the victim, in consequence of which he will not see which cards the machine
drew. The machine is then set to work, and to the amazement and relief of the
victim the machine exhibits an ace of hearts drawn from each pack. The victim
thinks that this extraordinary fact needs an explanation in terms of the machine
having been rigged in some way .... There is indeed something extraordinary in
need of explanation in ten aces of hearts being drawn.211
While it is easy to agree with Swinburne that there is something extraordinary in
this example that needs explanation, it does not seem as if this is really an
appropriate analogy to use to illustrate the fundamental issue raised by AP. The
victim's experience of the ten aces of hearts being drawn and the resulting
amazement and incredulity are the result of other experiences that the victim has
had and other knowledge that the victim possesses about probability, decks of
cards, and so forth, such as the number of cards in the deck and the number of aces
of hearts. To make the analogy appropriate for illustrating the main point of AP,
we need to imagine a completely naive and virgin perceiver (but a rational one, I
suppose) who has no previous experience and no knowledge of cards or
probabilities. Exactly how to adjust Swinburne's example to make it appropriate to
illustrate AP is not clear, but perhaps if we imagine that the victim suffers amnesia
after he is kidnaped and before he is put in the room and if the victim remembers
nothing about the physical world - including probability, cards, rooms, machines,
bombs, or even other people - then when the kidnapper makes his threat and the
machine produces ten aces of hearts, the victim will not be surprised at all. The fact
that the victim perceives anything, including a machine producing ten aces of
hearts, would not be surprising at all. It simply is what is perceived. To the victim,
the fact that he experiences something would not seem to be extraordinary and in
need of explanation?12
There are many different naturalistic explanations that physicists use to explain
the existence of the initial conditions of the universe and how and why those
conditions came to be so friendly and conducive to intelligent life. The basic
approach involves an attempt to explain why this actual world came to be selected
from an infinite number of different possible worlds that might have become actual
but which did not. In these different possible but nonactual worlds, the initial
Richard Swinburne, The Existence a/God, p. 138.
For further discussion of the probability of the set of initial conditions of the universe, see John
Earrnan, "The SAP Also Rises: A Critical Examination of the Anthropic Principle," American
Philosophical Quarterly. Vol. 24, no. 4,1987, pp. 309.
211

212

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

137

conditions and the fundamental constants would be different, so the question


becomes why this particular actual universe came to be actual from a "world
ensemble" of different universes.213 World ensembles are explained in different
ways. Appealing to the indeterministic results of quantum theory and the
possibility of the "branching" of the universe into different possibilities as a result
of unstable and nonmechanistic subatomic particles is one popular way. Both
Brandon Carter and Paul Davies adopt such a position regarding world ensemble.
This view is the result of accepting what is called the Copenhagen interpretation of
quantum mechanics, according to which only this actual world really exists while
other possible worlds are worlds in what Wittgenstein once called "logical space"
that might have existed but never did. 214 Theists find such naturalistic explanations
implausible and point to a number of scientists who have claimed that the existence
of God and creation by a personal being are needed as a principle to explain why
this actual world came to be "selected" from the world ensemble. Otherwise, they
claim, human existence is simply the result of a coincidence or an "accident" of
staggering and literally cosmic proportions.
AP has been taken by some theists to reverse the effects of Copernican
astronomy and the resulting Copernican principle, which removed any reason for
regarding Earth or human beings as occupying any position of special privilege in
the universe. AP appears to place human beings back at the very center of things
and to reestablish man's special status with God. To such an interpretation, we
should add a couple of important cautionary notes: In the first place, the
appellation "anthropic principle" is something of a misnomer since this principle
does not attribute special status to homo sapiens per se. There is nothing in the
cosmological version of AP that requires the existence of human life in the
universe - just intelligent, carbon-based, living observers. Whether such life may
even be regarded as humanoid depends upon just how far one is willing to stretch
the notion of humanoid. Certainly, AP is compatible with the existence of what is
frequently called "lower forms of life" or of intelligent creatures that bear no
anatomical resemblance or similarity in terms of physical appearance at all to
human beings. 215 As Corey observes, the cosmological evidence for AP supports at
most a biocentric principle. The claim for the special status of human beings, as he
observes, "involves an anthropocentric generalization from the scientific evidence
... [based] on a variety of philosophical and theological intuitions about humanity's
importance in the overall universal order.,,216 So extending the biocentric principle
213 The notion of a world ensemble is now common and was first used by Brandon Carter, '"The
Anthropic Principle and Large Number Coincidences," and G. Steigman, "Confrontation of Antimatter
Cosmologies with Observational Data," also in Confrontation of Cosmological Theories with
Observational Data, edited by M. S. Longair.
2[4 There are many discussions of the competing interpretations of quantum mechanics. For a very
accessible account, see Paul Davies, Other Worlds (New York: Simon and Schuster, 1980), pp. 145ff.
Also see M. A. Corey, God and the New Cosmology, pp. 175ff. For a discussion of the less accepted
many-worlds interpretation of Hugh Everett, according to which observers select the actual world in
which they live by their presence from many actual worlds, see The Many-Worlds Interpretation of
Quantum Mechanics, edited by Bryce S. DeWitt and Neill Graham (Princeton, N.J.: Princeton
University Press, 1973).
215 Corey recognizes that intelligent observers might be whales or dolphins or "human-like"
creatures on other planets. M. A. Corey, God and the New Cosmology, p. 7.
216 Ibid.

138

ANALYTIC PHILOSOPHY OF REUGION

to the anthropic principle is stretching the scientific evidence. And much of the
scientific evidence that supports AP would still be true even if carbon-based
intelligent life had never existed - for example, the delicate balance between the
mass of the universe and the force of the big bang that allowed for galaxies, stars,
and plants to form (though, of course, we would never have known this had life not
existed).
The second cautionary note that should be added is one concerning the status
and significance of AP. While theists are eager to seize upon AP and the scientific
evidence that is used to support it as evidence of intelligent design in the universe,
scientists and philosophers are certainly not all agreed about the formulation of or
the importance of the principle. John Earman calls W AP "nothing but a tautology"
and claims that different forms of AP are grounded more in philosophy than in
science.217 Other variations on AP stretch credulity far beyond what is
scientifically defensible. For example, the final anthropic principle (FAP) claims
that after carbon-based intelligent life has come into existence in the universe (as it
obviously has), it will never end, and the participatory anthropic principle (PAP)
claims that observers are necessary to bring the universe into existence. 218 Some
scientists regard AP as a metaphysical or cosmological principle rather than a
scientific one, and Earman calls PAP and FAP "empty teases" and "unbridled and
muddled speculation.,,219 In his review of Barrow and Tippler's book, Martin
Gardner called the FAP the CRAP (completely ridiculous anthropic principle)?20
What one should gather from all of this is that AP and all of its variations and
permutations are highly speculative and controversial in the scientific community,
as is the use of argument to the best explanation in this case. Does this mean that
appealing to a theistic explanation of the "fine-tuning" of the universe involving
design and personal creation is more probable and more rational than a completely
naturalistic explanation based upon the speculative and controversial versions of
AP, WAP, SAP, FAP, or PAP? Well, it is not clear that putting the question in this
way is really appropriate since scientists who appeal to AP do not use it in
anything like the normal sense of an explanation. In fact, scientific recognition of
AP (and its variations) is an admission that a normal scientific explanation is not
possible and should not be expected about the initial conditions of the universe. 221
Scientists who support AP may easily agree with Swinburne that the existence of
the universe and the particular order that we find in the universe are "too big" to be
217 John Eannan, "The SAP Also Rises: A Critical Examination of the Anthropic Principle," pp. 308
and 313.
218 FAP was first fonnulated by Barrow and Tipler, The Anthropic Cosmological Principle, p. 23.
Their Ultimate Observer (p. 471) is reminiscent of Berkeley's use of God as the Eternal Perceiver in his
idealism. As a corollary to the effect of the observer in quantum physics, PAP was first suggested by
John Wheeler, "Genesis and Observership," in Foundational Problems in the Special Sciences, edited
by R. E. Butts and 1. Hintikka (Dordrecht, Netherlands: Reidel, 1977). For discussions ofFAP and PAP,
see Eannan, ibid., pp. 312ff., and Corey, ibid., pp. 3-4, 167-68, and 185-88. Ignored in much of the
discussion is Wheeler's disclaimer that PAP is "too frail" to withstand criticism and Barrow and
Tipler's disclaimer that FAP and SAP are "quite speculative."
219 John Eannan, ibid., p. 313.
220 Martin Gardner, "WAP, SAP, FAP, and PAP," New York Review of Books, Vol. 33, no. 8, 1986,
pp.22-25.
221 See Eannan, "The SAP Also Rises: A Critical Examination of the Anthropic Principle,"
pp.
309 and 3 13 ff. , for a discussion of anthropic explanation.

THE NATURE OF GOD AND ARGUMENTS FOR THE EXISTENCE OF GOD

139

explained by science. 222 As we have seen earlier in the case of big bang theory,
scientific explanation does, in fact, come to an end, and further inquiry turns into
metaphysics and perhaps theology.z23 So concluding that a theistic explanation is
more rational and probable than AP seems to be an exaggeration and a
misstatement of the case. Rather, the theist may conclude that the theistic
explanation of the universe in terms of creation by a personal deity is more rational
and probable than no explanation or explanation in terms of simply chance, but
then this claim can be made quite independently of AP. The theist may thus take
the additional cosmological evidence that supports AP as simply additional
evidence of design. What this discussion does show is something similar to the
claim made by Alvin Plantinga regarding the free will defense to the problem of
evil (discussed in Chapter VI): the theistic position is, at least, a plausible one and
not unreasonable or irrational.
THE CUMULATIVE CASE ARGUMENT
What are we to conclude from this examination of the effect of new
developments during the twentieth century upon the traditional arguments for the
existence of God? On the one hand, various recent scientific developments have
been taken by some to undermine those traditional arguments; on the other hand,
those same various scientific developments have been taken by others to buttress
those arguments. In this critical discussion, I have tried to show that neither of
these conclusions is conclusively justified and that the force of the traditional
arguments supplemented with the recent developments in science is still a matter of
much controversy requiring further inquiry. There is still much work to be done
involving the traditional arguments by natural theologians. It does appear that the
theists can claim the modest victory that results from theism "holding its own" as a
viable alternative explanation of the existence of the universe to a completely
naturalistic, scientific explanation. As I have shown in the case of the big bang and
the cosmological argument and in the case of the anthropic principle and the
teleological argument, scientific explanations of the universe eventually become as
speculative and as metaphysical as theism. In this respect, Richard Swinburne is
correct that strictly scientific explanations of the universe come to an end, and it is
where science ends that theism finds a niche that has thus far withstood new
scientific discoveries and the development of new scientific theories about the
universe.
A second result of this consideration of recent developments involving the
traditional arguments for the existence of God is a reinforcement of the notion
made explicit by Kant and Hartshorne that the arguments are interrelated in rather
crucial ways. The ontological argument (even the modem modal versions)
proceeds only on the assumption in some form that the existence of God is
Richard Swinburne, The Existence of God, p. 138.
There is also the additional problem raised by Richard Gale of whether one can reasonably use
argument to the best explanation to make a prediction about the past based upon evidence from that
past's own future. But there is a goose and gander problem here. If the use of argument to the best
explanation is problematic for AP, then it is as well for a theistic explanation of the universe. See
Richard Gale, "The Anthropic Principle," Scientific American, Vol. 245, 1981, pp. 157. Also see, M. A.
Corey, God and the New Cosmology, pp. 209ff.
222
223

140

ANALYTIC PHILOSOPHY OF REUGION

possible. The cosmological argument in light of big bang theory just might
demonstrate that even if this assumption cannot be proven to be true at least it is
not an unreasonable assumption for the theist to make since science cannot offer
any explanation of the original singularity from which the universe came. The
teleological argument with the anthropic considerations can then be taken to
demonstrate that some intelligent design by a personal creator is at least as
plausible an explanation of the "fine-tuning" of the universe for intelligent life as
cosmic coincidence. While this may appear to be a modest advance for theism, it is
still an advance that appears to give theism an equal "place at the table" with
science, and, in this respect, there is little modest about it at all.
This unified view of the arguments for the existence of God supports what
Richard Swinburne has called the cumulative case for the existence of God. No one
of the arguments for God is decisive, as we have seen, but gradually, bit by bit, the
case for the existence of God is strengthened, Swinburne claims, by each
consideration. Other factors not yet considered - particularly religious experience playa significant role in Swinburne's cumulative case argument. We now turn our
attention to the problems of religious experience and religious epistemology in the
next chapter.

IV. Religious Experience and Religious Epistemology

The twentieth century saw significant new developments in and significant new
directions for the treatments of religious experience and religious epistemology.
The philosophical discussions in these areas have now taken on significant new
dimensions. While natural theology, based upon evidentialism, dominated AngloAmerican analytic philosophy in the early part of the twentieth century, antievidentialism - either in the various forms of religious experience, Reformed
epistemology, or fideism - became very dominant in the last few decades of the
century.
THE JUSTIFIED TRUE BELIEF THEORY OF KNOWLEDGE AND
EVIDENTIALISM
In contemporary epistemology, an evidentialist view regarding knowledge is
reflected in the justified true belief theory of knowledge (JTB), which has
dominated Western epistemology. According to the JTB theory of knowledge, a
person S knows that p if and only if
i. P is true,
ii. S believes that p, and
iii. S is justified in believing that p.
Conditions (i) through (iii) are regarded as both necessary and sufficient for S's
knowing that p. Condition (iii) has been explained in different ways by different
people, but, perhaps most commonly, epistemic justification has been explicated in
terms of "adequate evidence."! That we must provide some explanation of (iii)
seems obvious if we are to rule out epistemic luck. If S believes p on certain
evidential grounds that turn out to be false, but then p turns out fortuitously and
coincidentally to be true for other reasons than the ones for which S believes p to
be true, then we are not willing to grant that S knows that p. This is the lesson of
Edmund
Gettier's
famed
counterexamples. 2
These
counterexamples
notwithstanding, probably no other feature has so thoroughly characterized or
1 See, for example, Roderick Chisholm, Perceiving: A Philosophical Study (Ithaca, N.Y.: Cornell
University Press, 1957).
2 Edmund Gettier, "Is Justified True Belief Knowledge?" Analysis, Vol. 23, 1963. Reprinted in The
Theory of Knowledge: Classical and Contemporary Readings, edited by Louis P. Pojman (New York:
Wadsworth, 1999), pp. 142-43.

141

142

ANALYTIC PHILOSOPHY OF RELIGION

dominated Western philosophical epistemology than the fundamental position that


underlies the justified true belief theory of knowledge: justified beliefs are
epistemically better than nonjustified ones, knowledge must be based upon
justified beliefs, and justification requires adequate evidence.
There appear to be two ways in which evidentialist theists might respond to the
general challenge of JTB theory in accounting for knowledge of God. First, the
theist may respond by confronting evidentialism directly and rejecting the
fundamental claim that beliefs must be justified on the basis of evidence. Such a
response amounts to either denying condition (iii) in the definition of the JTB
theory of knowledge or explicating (iii) in some terms and on some grounds other
than "adequate evidence." One obvious way of attempting to maintain the
epistemic value of a belief while denying that the belief is justified in terms of
adequate evidence is to maintain that the belief in question is a "foundational" or
"basic" belief. Foundationalist epistemologists, such as Alvin Plantinga, take this
approach (as I examine below). Secondly, a theist may respond by claiming that
the beliefs of the theist are justified in the same way and on the same grounds as
other beliefs. William Alston develops such a position with his theory of
perceiving God (also examined below). This sort of response amounts to retaining
(iii) in terms of "adequate evidence" and then attempting to demonstrate how
religious beliefs are justified along the same lines as are other kinds beliefs and
using the same kind of evidence as is used to justify other kinds of beliefs. 3
THE JAMES AND CLIFFORD DEBATE
The dispute between evidentialists and nonevidentialists is nicely framed by the
dispute between William Clifford and William James early in the twentieth
century. In his famed "The Ethics of Belief," Clifford argues that there is a
normative aspect to knowing and believing that prohibits belief in the face of
insufficient evidence. Clifford tells a story of what appears to be a compelling case
of a situation in which, he claims, we are inclined to invoke a normative standard
to prohibit a person from believing something in the absence of adequate evidence.
A shipowner sends a ship full of emigrants out to sea even though he knows that
the ship is old, was not well built originally, and is often in need of repairs. He
considers overhauling the ship before it sails but then convinces himself that it will
make the trip safely, so he sends the ship off, and it sinks with all aboard. The
shipbuilder saves the expense of overhauling the ship and collects the insurance
money. The shipowner refuses to investigate the doubts he initially entertained,
and even if we concede, for the sake of the argument, that the shipowner is honest
and that he comes to believe sincerely that the ship is seaworthy, Clifford insists
that the shipowner "had no right to believe on such evidence as was before him."4
According to Clifford, proper belief cannot be the result of inadequate evidence or
3 Most commentators usually include Alston in the camp of the Reformed epistemologists; however,
it often appears as if Alston is not rejecting the need for episternic justification of religious beliefs
entirely but is arguing rather that the perception of God provides such episternic justification. Reformed
epistemology, in contrast, takes the belief that God exists as basic and warranted though not inferred
from or based upon other beliefs.
4 William K. Clifford, ''The Ethics of Belief," from Lectures and Essays, 1879. Reprinted in The
Theory of Knowledge: Classical and Contemporary Readings, edited by Louis P. Pojman, p. 551.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

143

the result of "suppressing doubts and avoiding investigation." The net result, he
concludes in his famous dictum, is that "it is wrong, always, everywhere, and for
anyone, to believe anything upon insufficient evidence.',5
For Clifford, on occasions when the evidence is not adequate for belief, it is
better to suspend judgment than it is to believe on the basis of inadequate evidence
and risk error. But what constitutes "adequate evidence" for Clifford? In
contemporary terminology, his position is a strong form of naturalistic
evidentialism. We are not entitled to hold any belief in the absence of such
evidence and should suspend judgment concerning such a belief. Clifford
generalizes his view about the nature of beliefs on the grounds that the actions that
are predicated upon inadequately justified beliefs often affect other people. This is
certainly true in the example of the shipowner; however, Clifford easily and glibly
concludes that no belief "however seemingly trivial the belief' and no believer
"however obscure" is really "insignificant or without its effect on the fate of
mankind.',6 But surely some beliefs are more important than others, and some
believers are more important than others. Undeniably, Hitler's belief that the Jews
are an inferior race had a significant impact upon the fate of mankind. However, an
isolated hermit's belief that eating a banana for breakfast each morning will help
promote a long, healthy life certainly seems to be insignificant and lacking in any
consequences concerning the fate of mankind.
WILLIAM JAMES AND "THE WILL TO BELIEVE"
James concedes to Clifford that religious beliefs are not supported by adequate
evidence of the kind Clifford demands; however, at the same time, James does not
wish to relegate all religious beliefs to faith. He attempts to preserve some sort of
notion of epistemic justification for religious beliefs, so he sets for himself the task
of redefining condition (iii) of the JTB theory of knowledge. In his lectures on
Pragmatism, James makes the now well-known distinction between what he calls
"the tender-minded" and "the tough-minded" approaches to problems in
philosophy.7 The tender-minded, the rationalists, are regarded as sentimental, softheaded, and religious while the tough-minded, the empiricists, are regarded as
callous, brutal, and nonreligious. The naturalists, of course, are to be counted in the
camp of the empiricists. The rationalists offer a philosophical theory within which
religious beliefs are welcomed, but it is not empirical. The empiricists offer a
philosophical theory that is strictly empirical but within which religious beliefs are
not welcomed. James tries to stake out a "middle ground" by "stretching" the
narrow confines of the beliefs that are permitted on strictly naturalistic grounds.
James is an empiricist in his general philosophical method, but he cannot bring
himself to follow his empiricism wherever it may lead, that is, to a strict
naturalistic ontology that admits nothing beyond the natural world. In "The Will to
Believe," he maintains that a person can be justified by the passions and the will in
believing, in certain circumstances, where the evidence available on strictly
empirical grounds is inadequate to justify the belief. He says,
Ibid., p. 554.
Ibid.
7 William James, Pragmatism (New York: New American Library, 1907), pp. 22ff.
5

144

ANALYTIC PHILOSOPHY OF REUGION

Our passional nature not only lawfully may, but must, decide an option between
propositions, whenever it is a genuine option that cannot by its nature be
decided on intellectual grounds; for to say, under such circumstances, "Do not
decide, but leave the question open," is itself a passional decision - just like
deciding yes or no - and is attended with the same risk of losing the truth. 8
The "genuine option" that James describes is a choice between two competing
hypotheses (in this case, the choice between the religious hypothesis and the
nonreligious hypothesis) that James further characterizes as forced, living, and
momentous. 9 A forced option is one in which the choice between the two
competing hypotheses is exhaustive. A person must choose one of the two, and the
possibility of a third hypothesis or of avoiding the choice altogether is not
available. A living option is a choice between two competing hypotheses when,
given the circumstances and the nature of the person making the choice, both of the
hypotheses are genuinely possible choices for the person to make. Finally, a
momentous option is a choice between two competing hypotheses where the
choice involves a matter of some vital importance with significant consequences.
While embracing a general empiricism, James thus seeks to avoid what he
regards as the restrictive implications of relying only upon empirical observations
and sense experience as the sources of the evidence for beliefs. Restricting the
notion of adequate evidence to empirical sense experience in condition (iii) of the
JTB leads to a thoroughgoing naturalism. Whereas Clifford exhorts a person to
avoid error above all else, James regards "the chase for truth as paramount."l0 A
general will never win any battles, James claims, if he advises his troops to avoid
the risk of being wounded above all else. To win battles - either military ones or
epistemic ones - requires taking some risks, according to James, and the fact that
we might make mistakes in holding certain beliefs is not that serious a problem. 11
According to James, only in situations involving "scientific questions" are we
justified in suspending judgment when we have inadequate evidence, since, in such
situations, we are simply "recorders" of an objective truth. Such scientific
questions are not forced upon us, nor, James claims, do they represent living or
momentous options. 12 For James, naturalism must be circumscribed to our
knowledge of the natural world.
James thinks that his understanding of how the roles of a person's "passional
nature" and the will may justify belief in certain situations represents a significant
improvement over the blatantly self-interested pragmatism of Pascal's wager.
Pascal argues that since the known evidence equally supports God's existence and
God's nonexistence, a reasonable person should choose to believe that God exists
since the benefits of the positive belief are much greater. There is much that may
be said about Pascal's wager, but here I will simply note that James denies that a
8 William James, '''The Will to Believe," in The Theory of Knowledge: Classical and Contemporary
Readings, edited by Louis Pojman, p. 558.
9 Ibid., p. 555.
10 Ibid., p. 558.

11
12

Ibid., p. 559.
Ibid., p. 559.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

145

person can calculate such probabilities and the potential benefits and losses in the
artificial manner that Pascal suggests and then convince oneself to sincerely
believe, on such bases, that God exists.!3 James's point is similar to, and
undoubtedly drawn from, the point made by Charles Peirce in his attack upon the
general, methodological skepticism of Descartes. Peirce insisted that doubt must be
genuine and that a person just cannot bring himself to doubt the kinds of things that
Descartes claimed he can doubt. Similarly, James claims, belief must be genuine.
A person just cannot make himself sincerely believe something as the result of the
kind of artificial and selfish calculation that Pascal's wager requires.!4 The wager
thus is not a genuine option. With genuine options, the will and our passions justify
belief in the absence of adequate evidence, and the will to believe thus is James's
way of redefining condition (iii) above in the JTB theory of knowledge to extend
the notion of 'justification" to include nonepistemic considerations. IS
CONTEMPORARY RESPONSES TO THE CHALLENGE OF
EVIDENTIALISM
In terms of more recent responses to the JTB theory, contemporary evidentialist
positions rely upon natural theology and particularly the arguments for the
existence of God. 16 Much of this volume, as well as most of the problems and
issues discussed herein, are related in one way or another to the challenge of
evidentialism to various theistic beliefs - including the problem of trying to get
clear about the concept of God, the problem of the meaningfulness of claims made
about God, the problem of the epistemic force of arguments for the existence of
God, the problem of the counterevidential force of the problem of evil, questions
about the comparative reasonableness of religion and science, questions about the
comparative reasonableness of atheism and theism, questions about the source of
religious ethics, and questions about the comparative reasonableness of different
claims made by different religious believers.
The contemporary evidentialist approach that requires basing theistic beliefs on
reason and evidence is found amongst both theists and nontheists. For the most
part, this challenge can be seen as a continuation of the age-old debate concerning
the respective roles of faith and reason in religion and as a continuation of the
Clifford-James debate concerning the role of evidence in terms of religious beliefs.
The evidential challenge from nonbelievers - atheists and humanists - who reject
the various attempts to establish the reasonableness of belief in the existence of
13 Ibid., p. 556. There are many discussions that compare Pascal and James. For example, see J. L.
Mackie, The Miracle of Theism (Oxford: Clarendon Press, 1982), pp. 200-203.
14 Pascal may respond that if a person behaves religiously (for example, by praying or using holy
water), then religious belief will gradually and eventually follow, but such a response seems very
optimistic and belies what is known of human psychology. James maintains that there are, of course,
other difficulties with Pascal's wager. For example, obviously God may punish a person for believing
on such grounds. See James, ibid., 556.
15 It should be noted that James follows Peirce in holding that knowledge is essentially fallible.
Thus, any knowledge claims based upon religious beliefs must also be regarded as tentative and fallible.
16 However, Nicholas Wolterstorff disagrees that evidentialism is tied so closely to natural theology.
See his "The Migration of the Theistic Arguments: From Natural Theology to Evidential Apologetics,"
in Rationality, Religious Belief, and Moral Commitment, edited by Robert Audi and William J.
Wainwright (Ithaca, N.Y.: Cornell University Press, 1986), pp. 38ff. Wolterstorff attributes
evidentialism to the philosophical influence of the Enlightenment.

146

ANALYTIC PHILOSOPHY OF REUGION

God based upon evidence and argumentation using reason is well represented by
Antony Flew. Here is how he puts the challenge:
If it is to be established that there is a God, then we have to have good grounds
for believing that this is indeed so. Until and unless some such grounds are
produced we have literally no reason at all for believing ... . It must be up to
them [theists] ... to bring forward sufficient reasons to warrant their claim that ...
there is a GOd. 17

There is little doubt that this kind of evidentialist challenge and attempts to
respond to it have dominated much of contemporary analytic philosophy of
religion. Specifically, the attempts to provide support for beliefs concerning the
existence of God by the use of arguments for God's existence are perhaps the most
obvious examples of theistic responses to this challenge. I have treated the novel
developments in these different arguments during the twentieth century in the
previous chapter.
RELIGIOUS EXPERIENCE AND THE CUMULATIVE CASE ARGUMENT
Aside from the developments in the traditional arguments for the existence of
God and the accompanying kinds of concerns that are the purview of natural
theology, questions about the need for justification of religious beliefs have taken
on new dimensions and raised new issues and problems. In the second half of the
twentieth century, much of religious epistemology was dominated by different
attempts to respond to the challenge of evidentialism by circumventing the
challenge, defusing it, or treating it as a pseudoproblem. Instead of responding
directly to the evidentialist challenge, as natural theology does, these responses
deny its force against theistic beliefs, that is, they deny that we must have reasons
or evidence for theistic beliefs or that such beliefs are in need of epistemic
justification in the manner represented by the JTB theory of knowledge and
evidentialism. Currently, religious epistemology and concerns about the
reasonableness and rationality of religious belief have broadened far beyond the
traditional responses to the evidentialist challenge, using the arguments from
natural theology. I will discuss four such responses in this chapter - the inductive,
"cumulative case argument"; the perception of God; Reformed epistemology; and
fideism. These responses can be measured and positioned in respect to one another
in terms of the respective distances between them and the evidentialist approach of
natural theology. Thus, these current schools of thought in religious epistemology
might be understood as responses to the evidentialist challenge as it has been taken
up from within theism by theists who reject natural theology.
THE CUMULATIVE CASE ARGUMENT
While the inductive, cumulative case argument does not represent anything like
the departure from evidentialism that Reformed epistemology and fideism do, this
17 Antony Flew,"The Presumption of Atheism," in The Presumption of Atheism and Other
Philosophical Essays on God, Freedom, and Immortality, Antony Flew (New York: Barnes and Noble
Books, 1976), pp. 22-23.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

147

approach does begin to weaken the demand made upon human reason as that
demand was previously represented by the arguments for the existence of God in
natural theology. The cumulative case argument does not defend any single one of
the arguments or any single kind of evidence as deductive proof or conclusive
evidence for the existence of God. The sharpest contrast here can be made with a
theist who defends one of the arguments for the existence of God as valid and
sound and thus as containing adequate conclusive evidence for the claim that God
exists. Defenders of the cumulative case argument do not rely upon a single
argument or a single source of evidence for support of the claim that God exists,
nor do they maintain that the existence of God is ever established completely
conclusively. Rather, the cumulative case for the existence of God builds
gradually, incrementally, and inductively in terms of probability until it becomes
more probable that theism is true than that it is not (to a reasonable person).
Though individual arguments and considerations for God's existence may
individually have flaws and be less than convincing, these arguments and
considerations are seen as collectively interfacing with one another in such a
fashion that they collectively support rational theistic belief whereas taken
individually they do not. The effect is something like a woven fabric rather than a
single thread of argument. 18 Although the cumulative case argument is not an antievidentialist approach (after all, it is still an argument), still it is a much weaker
form of evidentialism than is traditionally associated with the arguments in natural
theology. Basil Mitchell develops such a cumulative case; however, the position is
now most closely identified with the later work of Richard Swinburne, who
maintains that, in addition to the traditional arguments for the existence of God,
evidence to make belief in God's existence reasonable comes from a consideration
of religious experience and miracles. 19
Much of the recent literature has been devoted to efforts at defining religious
experience, distinguishing it from mystical experience, and drawing out the
similarities and dissimilarities between religious experience and sense
experience?O I cannot address all of these issues here. Since I take the major
18 Thus, the cumulative case argument is not simply an accumulation of different arguments. For a
helpful discussion that illustrates the point nicely, see Caroline Franks Davis, The Evidential Force of
Religious Experience (Oxford: Clarendon Press, 1989), pp. 108ff.
19 Basil Mitchell, The Justification of Religious Belief(Oxford: Oxford University Press, 1981), and
Richard Swinburne, The Existence of God (Oxford: Clarendon Press, 1989).
20 The following are some of the primary contemporary sources that treat the distinction between
religious and mystical experiences and also address the comparison of religious experience with sense
experience: George Mavrodes, Belief in God: The Epistemology of Religious Experience (New York:
Random House, 1970); C. B. Martin, Religious Belief (Ithaca, NY: Cornell University Press, 1959);
Michael Martin, Atheism: A Philosophical Study (Philadelphia: Temple University Press, 1990); Rudolf
Otto, The Idea of the Holy (Oxford: Oxford University Press, 1958); Wayne Proudfoot, Religious
Experience (Berkeley: University of California Press, 1985); William Rowe, "Religious Experience and
the Principle of Credulity," International Journal for the Philosophy of Religion, Vol. 13, 1982, pp. 8592; Ninian Smart, The Religious Experience (London: Macmillan, 1991); T. Stace, Mysticism and
Philosophy (New York: Macmillan, 1960); William Wainwright, Mysticism (Madison, Wis.: University
of Wisconsin Press, 1981); William Wainwright, "Mysticism and Sense Perception," Religious Studies,
Vol. 9, 1973. Much of this article appears as Chapter 3 of his Mysticism and is reprinted in
Contemporary Philosophy of Religion, edited by Steven Cahn and David Shatz (Oxford: University of
Oxford Press, 1982), pp. 123-45; Keith Yandell, The Epistemology of Religious Experience
(Cambridge: Cambridge University Press, 1993). Also see the discussion of William Alston's

148

ANALYTIC PHILOSOPHY OF REUGION

controversy in the last part of the twentieth century to revolve around the attempts
to explicate religious experience in terms that are analogous to sense experience, I
will focus mainly on this issue. Questions about the strength and justification of the
claim that there is an analogy between religious experience and sense experience
focus attention on one of the main areas of contention in the controversy between
those who defend religious experience as epistemically efficacious in supporting
the claim that God exists and those who do not.
THE CUMULATIVE CASE ARGUMENT AND RELIGIOUS EXPERIENCE
Swinburne offers the following definition of a religious experience: a religious
experience is "an experience which seems (epistemically) to the subject to be an
experience of God (either of his just being there, or doing or bringing about
something) or of some other supernatural thing.,,21 Two aspects of this definition
bear closer inspection: First, this definition focuses onlr upon what are usually
called numinous experiences, following Rudolf Otto,2 since it is only such
experiences that provide possible evidence of the existence of some supernatural
being external to the subject having the experience. In other words, the person
having the experience takes the experience to be of a supernatural being. Thus, this
definition leaves out such experiences as a person being simply overcome with
feelings of love and peace. It is now customary to classify such experiences as
mystical. Secondly, it is important to notice that this is a definition that is given in
terms of what we may call an "internal" description of a person's experience.
Swinburne does not describe a religious experience as one in which a person says,
"I saw God," but rather one in which a person says, "I seemed to have seen God."
In the first description, since the report cannot be true unless God exists, there is
less likely to be agreement about the truth of such descriptions; in the latter
description, there is likely to be considerably more generosity in accepting such an
"internal" report from a subject. If one insists upon a definition of religious
experiences simply in terms of reports of experiences that are descriptions of
externally existing entities, then there is not likely to be much fruitful discussion
concerning religious experience since a religious experience would then be defined
as veridical. By relying upon an internal account of what an experience "seems" or
"appears" to a person to be, we can start with some agreement that religious
experiences do occur and then pursue the argument to see what comes of such a
modest beginning. 23 At the same time, religious experiences have a noetic quality,
that is, they are experiences that seem to be experiences of an object or a being
(unlike, for example, the experience of pain).
Perceiving God below and references cited there. The source that provided the framework for much of
the discussion in the twentieth century was William James, The Varieties of Religious Experience (New
York: Longmans, Green, and Co., 1902).
21 Ibid., p. 246. This definition is followed closely by William Alston, discussed below, in his theory
of the perception of God.
22 See Rudolf Otto, The Idea of the Holy (New York: Oxford University Press, 1958).
23 Richard Swinburne, The Existence of God, p. 245-46. Alston uses the same theory of appearing
and follows a very similar tactic in his perception of God, discussed below. Interestingly, William
Rowe, who opposes claims concerning the cognitive content of religious experience, first suggests this
position in "Religious Experience and the Principle of Credulity," International Journal for the
Philosophy of Religion, Vol. 13, 1982, pp. 85-92.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

149

Like others who have written about religious experiences, Swinburne


distinguishes different kinds of religious experiences.24 I will follow Swinburne in
considering only those cases of religious experiences where the purported
experience is of God. If one begins, as Swinburne does, with an internal account of
religious experience, according to which it seems to a subject S that some x is
presented, then the problem becomes one of providing an account where such an
experience provides evidence for some claim beyond the claim to the experience
itself. Certainly, if S has an experience E wherein he is appeared to godly by x,
then (assuming that S is honest and serious in his report), we can say that E
occurred, but what can we say beyond this? What sort of evidence does E provide
for saying anything else about S or x?
THE PRINCIPLE OF CREDULITY
Swinburne proposes that it is normally rational to accept honest and sincere
reports from subjects about their experiences of certain objects. In other words, in
the absence of special considerations or known circumstances where we may
expect that something has gone wrong, it is rational to accept a subject S's report
when S reports that he has been appeared to by some object x. C. D. Broad first
explicitly formulated such a presumption in favor of accepting the credulity of a
person's report of his experience and applied that presumption to religious
experience to argue that reports of religious experiences should be given prima
facie acceptance in the absence of some reason or circumstance that would make
us doubt such reports. 25 Swinburne adopts Broad's approach for defending
religious experience and formulates what he calls the principle of credulity (POC):
In the absence of special considerations, all religious experiences ought to be
taken by their subjects as genuine, and hence as substantial grounds for belief in
the existence of their apparent object - God. 26
Like the cumulative case argument, POC is now most closely identified with
Swinburne. His defense of POC is based upon the comparison he draws between
the way in which POC is used in normal sense perception and the way in which it
is used in religious experience. 27 Swinburne gives attention to two ways in which
others have thought that the POC may be restricted. His account of these
restrictions and his replies are important for considerations raised later in this
chapter. One argument for restricting the use of POC derives from the claim that
the POC requires inductive justification and that such justification is not possible in
the case of religious experience. This argument claims that the fact that x appears
24 Ibid., pp. 249. Also, see Otto, ibid., and William Wainwright, Mysticism (Madison, Wis.:
University of Wisconsin Press, 1981).
25 C. D. Broad, "Arguments for the Existence of God, II," The Journal of Theological Studies, Vol.
40,1939.
26 Richard Swinburne, The Existence of God, p. 154.
21 William Wainwright has drawn the comparison between religious experience and sense
experience earlier in a similar fashion. See William Wainwright, Mysticism, Chapter 3. Much of this
chapter is drawn from his "Mysticism and Sense Perception," Religious Studies, Vol. 9, 1973, pp. 25778. The article is reprinted in Contemporary Philosophy of Religion, edited by Steven Cahn and David
Shatz.

150

ANALYTIC PHILOSOPHY OF REUGION

to S is good prima facie grounds for taking x to exist only if x' s appearing to S has
proven to be good grounds for taking x to exist in the past, that is, x was actually
determined to exist in the past on the occasions when x appeared to S to exist. We
have such good inductive grounds in the case of using POC with sense perception
and the existence of physical objects; however, arguably, we do not have good
inductive grounds for using poe in the case of religious experience and the
existence of God. Swinburne's retort to this objection is that even in ordinary cases
of physical objects a subject does not remember all the previous cases of being
appeared to by those physical objects in a way that would be necessary to perform
an inductive argument; therefore, the use of POC must be "a basic principle not
further justifiable.,,28
The second objection to the POC attempts to restrict the use of POC to familiar
sensible qualities in ordinary cases. Thus, if we take ourselves as being appeared to
redly, we are justified in assuming that there is something red that is doing the
appearing since red is a common property of sensible objects. But if we take
ourselves as being appeared to in a more sophisticated and less usual sense, by
Russian caviar, Italian leather, or Scottish wool, then POC does not provide good
grounds for believing that Russian caviar, Italian leather, or Scottish wool exists
without the additional help of inductive justification. The recognition or
identification of such items requires some past experience with the items in
question and a developed ability to make fine discriminations to avoid mistakes.
Swinburne replies to this objection to the POC by claiming that it is dependent
upon drawing a distinction between what is given in "real"experience and what is
the result of an interpretation of experience. No such distinction, he maintains, can
legitimately be drawn. 29
Unless these two objections are met, they give rise to four "special
considerations" that would prohibit our granting credulity to the claim that x exists
when it seems to S that x is present in a religious experience. Anyone of these
special considerations can serve as a potential defeater of POe. I will simply state
these potential defeaters summarily:3o i) the conditions under which x seems to be
present to S have been unreliable in the past; ii) S's reports concerning x or other
experiences have been unreliable in the past; iii) other evidence makes it very
improbable that x is present; and iv) other evidence makes it highly probable that x
is not the cause of it seeming to S that x is present, whether x exists or not. In the
absence of any of these four defeaters, S is justified in assuming that things really
are as they appear to him to be. I shall return to consideration of these potential
defeaters of POC and the two objections to POC that give rise to them
momentarily.
THE PRINCIPLE OF TESTIMONY
Finally, to the POC Swinburne adds a further important principle of rationality,
the principle of testimony (POT): normally, in the absence of any special
circumstances to the contrary, we are justified in believing that what other people
Richard Swinburne, The Existence of God, pp. 255-56.
One person who raises such an objection is Roderick Chisholm. See ibid., pp. 257-58.
30 Ibid., pp. 260-64.
28

29

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

151

tell us about their experiences actually happened, or, alternatively, we are justified
in believing that what other people tell us they experienced as occurring probably
did occur as they claim. 3 !
With POC and POT in place, Swinburne's argument from religious experience
is simple and straightforward. Given the number of reports of different people of
religious experiences of God - some but certainly not all of which might be
defeated by (i) through ( iv) - it is more probable that God exists than that God
does not exist, that is, given the reports of God's seeming to appear to a number of
different subjects having religious experiences, it is more probable that things are
as they have been reported than it is that they are not; thus it is more probable that
God does exist and did appear in the way in which he seemed to appear than that
God does not exist.
THE PRINCIPLE OF INCREDULITY
The two objections considered by Swinburne directly address the issue of
whether the POC is justified in the case of religious experience. 32 Consider the first
objection against the POC - namely, that the POC requires inductive justification.
A special instance of this objection involves situations where a subject S is
appeared to by x but has no previous experience with x's although S does have
experience with the properties that define x. Swinburne's example involves the
appearance of a centaur to S: we may say that, although S has no experience with
centaurs, he does have experience with the properties that define a centaur (that is,
having the body and legs of a horse and having the head, torso, and arms of a man,
and so forth). Thus, one may seem to be appeared to by a centaur, even though one
has never experienced a centaur, by being appeared to by a concatenation of
properties that one normally associates with centaurs. The same kind of analysis
applies to experiences of God, according to Swinburne, since 'God' is also defined
in terms of familiar properties such as "'person' without a 'body' who is unlimited
in his 'power,' 'knowledge,' and 'freedom. ",33 One experiences God by
experiencing the concatenation of predicates or powers that one normally
associates with God.
Swinburne's treatment of this case is very pivotal in determining whether his
cumulative case argument using religious experience is viable and whether the
POC can withstand this challenge to its use with religious experience. There are
several reasons for doubting that Swinburne's defense is successful. The process of
concatenation of predicates is not so simple and straightforward as Swinburne
suggests. One may experience one thing as having the body and legs of a horse and
another thing having the head, torso, and arms of a man without being able to
combine the two sets of predicates for a single thing. For example, one may
experience one figure as having three and only three angles and another figure as
having four sides but be unable to project the two predicates for a single figure.
31 Ibid., pp. 271-72. Swinburne thinks that there are normally no special circumstances that would
legitimately make one doubt the veracity of a person's report of religious experiences.
32 It is the application of POC to religious experience rather than POC itself as a general
epistemological principle that is most at issue. Rowe emphasizes this point in "Religious Experience
and the Principle of Credulity," ibid., pp. 90-91.
11 Richard Swinburne, The Existence of God, p. 256.

152

ANALYTIC PHILOSOPHY OF RELIGION

There is at least one further difficulty that arises here. It is significant that
Swinburne, unlike Alston (discussed later in this chapter), is explicitly committed
to a descriptivist account of reference wherein the referent of a referring expression
(namely, 'God,' in this case) is determined in terms of the possession of certain
essential properties by the referent (see Chapter II). On such an account of
reference, successful reference is completed when a referent is found that does
indeed possess the properties in question. In the case of 'God,' this means for some
subject S to be appeared to godly, S must be appeared to through the exhibition of
certain properties or the exercise of certain powers in a godly fashion. But could
we reasonably expect that any S would ever be in a position to determine that it is
God exhibiting such properties or exercising such powers? Following Aquinas, let
us say that God exhibits each property and exercises each power to the maximal
degree possible. Could S recognize that the degree to which the property or the
power that is presented by x is being presented to the maximal degree possible for
that property or power? To do so, we must give S the uncanny ability to distinguish
between the exhibition of degrees of properties or powers that are far beyond
anything normally experienced by S or present in the world and that perhaps are
very near the maximal degree possible but still not the maximal degree possible.
While successful reference may still succeed in such circumstances, S would never
be in a position to determine when it succeeds and when it does not, which
seriously undermines the POCo
This seems to be the problem noted by J. L. Mackie when he says, "Nothing in
an experience as such could reveal a creator of the world, or omnipotence, or
omniscience, or perfect goodness, or eternity, or even that there is just one god.,,34
Siding with William James, Mackie concludes that the best that religious
experiences can reveal is "the existence of some greater friendly power, whose
precise identity and character are left wholly indeterminate.,,35 Imagine, for
example, that S is appeared to by some being that is very similar to Nee, suggested
by Paul Henle and discussed in the previous chapter. 36 The being that appears to S
is nearly Nee, that is, Nnee. Nnee possesses every power and exhibits every
property to the maximal degree minus one and, we could even add, exists in every
possible world except one. It is implausible to imagine that even an S who
presumably has a great deal of experience with such matters and who is steeped in
the tradition, scriptures, and the like could manage such a fine discrimination that
would enable S to distinguish between Nnee and Nee. This consideration also
mitigates decisively against the proposal made by William Wainwright and others
that agreement amongst subjects who are regarded as having experience with such
matters is an important test for evaluating religious perceptual claims. 37 The most
we can say is that when x appears to a subject S and exhibits a property or power to
a certain level or degree that is beyond S' s ability to understand or is beyond the
level or degree that the property or power has ever been experienced by S as
J. L. Mackie, The Miracle of Theism (Oxford: Clarendon Press, 1982), p. 182.
Ibid., p. 183. It may very well be that even granting that such experience can determine that the
power (being?) is friendly rather than unfriendly (God rather than Satan) is granting too much. See the
discussion of William Alston's appeal to direct reference later in this chapter.
36 Paul Henle, "Uses of the Ontological Argument," Philosophical Review, VoL 70,1961, pp. 102ff.
37 William Wainwright, Mysticism, pp. 88ff.
34
35

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

153

exhibited in this world, S is inclined to identify x as God. However, on the basis of


simply being appeared to by x through the exhibition of such properties or the
exercise of such powers, it is not reasonable to attribute to S the power of
determining that x is God who is seeming to appear rather than a near-god, an
intelligent alien from some distant galaxy, or Satan. Thus, though S might
successfully refer to God when being appeared to by x, S would never be in a
position to know whether the reference is successful, nor, when the reference is
successful, would S be in a position to know on the basis of which property or
power the reference was successful.
Swinburne's reply to this first objection to poe belies his reply to the second.
This objection claims that we can trust only a report of familiar sensible qualities
by a subject S while remaining skeptical of reports of S's being appeared to by
more unusual and sophisticated things. Surely God is the most unusual and
sophisticated thing in the universe, and for S to make the kind of identification that
Swinburne suggests that a subject can make in his reply to the first objection to the
poe on the basis of the experienced properties or powers, S must make an
inference from the experience of being appeared to by those properties or powers
to the claim that it is God doing the appearing. Even if we introduce a principle of
charity and allow S the ability to discriminate the exhibition of unusual properties
in a way that we have just seen is very questionable, the process would still require
that S provide some "interpretation" of the experience - an inference or conclusion
based upon the presumed association of the properties in question - as God. It
appears that something like this process must take place in order for the theist to
explain how it is that people such as Jim Jones and David Koresh make mistakes
(assuming that these are honest, sincere mistakes). They "take" their being
appeared to godly incorrectly.38
These considerations make it difficult to understand how one could even make
sense of the four potential defeaters of poe, (i) through (iv), that Swinburne
introduces. The way Swinburne discusses the potential defeaters, it is some third
party, some audience A, to whom S has reported his experience that makes the
judgment about the effect of the potential defeater. But for A to consider whether
any of the potential defeaters (i) through (iv) actually defeat S's claim, A must
assume that S has successfully managed to identify his seeming to be appeared to
by x as seeming to be appeared to by God. I have argued that Swinburne has not
argued convincingly that this is possible; the poe thus degenerates into the
principle of incredulity. The inviolate nature of seeming originated with sense data
theory, but in the case of sense data theory, the seemings were the contents of
immediate sense experience. "I seem to see red" is worlds apart from "I seem to
see God" in terms of its character as the report of an immediate and inviolate sense
experience. We have detailed accounts of the physiological and perceptual
mechanisms involved in seeing red. We have no such account of the mechanisms
involved in the process of seeing God.

38 It is worth noting that William Alston attempts to avoid this problem by adopting a direct realism
- a move that raises other problems in identifying God as the object of an experience. See the discussion
below.

154

ANALYTIC PHILOSOPHY OF RELIGION

It does not seem that any rational subject S could ever be in an epistemic
position to be confident of an internal justification of claiming to be appeared to by
God. Suppose, for example, Yahweh appears to Moses as x and parts the Red Sea,
and Moses then identifies his seeming to be appeared to by some x that parts the
Red Sea as being appeared to by Yahweh. Parting the Red Sea is not an act that
requires maximal power. It does not, for example, require as much power as
creating or destroying the universe; therefore, some lesser being might have parted
the Red Sea - a being that is less than omnipotent. Moses has managed a
successful identification of x as Yahweh, but completely unbeknownst to himself,
and if Moses does not have the internal justification relative to his own epistemic
situation to be confident that x is Yahweh, then no audience A would be justified,
using POC, in attributing veracity to Moses's claim that Yahweh appeared to him
as x. Indeed, it seems that any audience A should be suspicious of any such claim
and remain incredulous.

THE PRINCIPLE OF CREDULITY CONTINUED


Aside from questions about what experiences there are that would permit S to
determine whether an appearance of x to S is really the appearance of God, there is
the additional problem of specifying general criteria that would allow any subject S
to adjudicate between apparently veridical but delusive appearances of x and
apparently veridical and genuine appearances of x. Attributing cognitive content to
any religious experience requires such successful demarcation amongst different
experiences. This becomes the crucial issue for the POC and requires detailed
analysis. For example, William Rowe has challenged the POC on the grounds that
the POC requires that there be ways of distinguishing such experiences, but he
thinks that there is no way of specifying the conditions under which a particular
subject can be assured of experiencing God. Without the specification of such
conditions, Rowe argues, the POC is not justified and S's seeming to experience
God cannot be taken as prima facie evidence that God exists. 39 Rowe's point can
be put clearly in terms of other problems discussed in this volume. Given
naturalistic explanations of religion, such as those of Sigmund Freud and Emile
Durkheim discussed in Chapter VII, there must be some way to distinguish
between S' s seeming to be appeared to by God when, in fact, S is being appeared
to by God and S's seeming to be appeared to by God when, in fact, S is being
appeared to by his suppressed father figure or by society personified. In terms of
the discussion of Swinburne above, we can put the point in the following way:
when S uses the appellation 'God' to refer to the entity that seems to be appearing
through the exhibition of certain predicates or the exercise of certain powers, what
are the conditions under which S can determine that he "got it right" and that the
entity seeming to appear to him is indeed God ?40 In one case, S is being appeared
to by God, and, in the other case, S is not being appeared to by God. Unless Scan
determine which situation she is in, then surely the evidential efficacy of POC is
39 William Rowe, "Religious Experience and the Principle of Credulity." pp. 90ff. Also see David
Conway, "Mavrodes, Martin, and the Verification of Religious Experience," International Journal for
the Philosophy of Religion, Vol. 2, no. 3, 1971.
40 The problem of determining correctly the reference of "God" during a religious experience is
treated in detail below in the discussion of William Alston's Perceiving God.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

155

reduced. Competing, possible explanations of religious experiences in terms of the


social sciences or in terms of neuroscience reduce the chances of supplying some
rationale to buttress the epistemic force of such experiences, and, as Rowe argues,
the existence of evil is a potentially rebutting condition since the existence of evil
is inconsistent, he thinks, with the existence of GOd. 41 The force of Rowe's
objections to POC is to propose that there is always reasonable doubt regarding
Swinburne's defeating conditions (i) and (ii) and whether or not those conditions
have been addressed in a way that justifies attributing cognitive status to a religious
experience.
Some theists have defended religious experience from such criticisms based
upon the impossibility of reducing religious experience to other kinds of
experience by claiming that an experience of the sacred or the holy is irreducible
and hence sui generis - not derived from or reducible to any other kind of
experience. S's report of seeming to be appeared to by God means that for S, his
experience is an experience of God, and no explanation of that experience in terms
of ordinary phenomena will be adequate to account for this aspect of S' s
experience. To be sure, a person's being appeared to by God is an intentional
experience with an object, and phenomenologically for S, that object is God; no
reduction of the object of that experience in terms of S's suppressed father figure
or in terms of society or in terms of any other object other than God will be
adequate for S. The separate issue, however, that determines how S's claim will be
treated in terms of its cognitive content or epistemic efficacy is the problem of
determining how an objective audience assesses S's claim. Wayne Proudfoot
explains the difference between S's first-person account of his experience and that
of an objective audience in terms of the difference between descriptive reduction
and explanatory reduction. 42 There is a first-person privilege with normative force
that we must recognize for a person describing his own immediate experience.
Thus, if Jones is in the forest alone and then later reports being frightened by a
bear, it will not do as an explanation of Jones's fear to later have someone reveal to
him that what he thought was a bear at the time was simply a fallen tree. Jones's
fear was fear of a bear - not of a tree. Jones would certainly not admit to being
frightened by a tree nor would an objective audience describe his experience this
way. A reductive description of Jones's fear will not do. However, when it comes
to an explanation of Jones's fear, someone could offer the reductive explanation
that Jones saw a tree, thought it was a bear, and was thus frightened, and for an
explanation of S's experience to be adequate, the explanation need not be one that
is acceptable to S.43 The entire force of Swinburne's POC as well as Rowe's
criticisms of that principle have to do with the issue of the explanation of S's
experience; thus, even if we admit that the phenomenological object of S's being
appeared to is God or the sacred or the holy, there is still the matter of determining
how an objective audience is to explain S's experience.

41

For a response to Rowe, see Keith Yandell, The Epistenwlogy of Religious Experience, pp. 229-

30.
42 Wayne Proudfoot, Religious Experience (Berkeley: University of California Press, 1985), pp.
196ff.
43 This analysis is based upon an example used by Proudfoot, ibid., pp. 192-93.

156

ANALYTIC PHILOSOPHY OF REUGION

Others, such as C. B. Martin, have argued in a similar fashion that, in the case of
sense experience, there are "tests and independent procedures" for determining
whether a particular claim based upon that sense experience is true or not, but in
the case of religious experiences there are no such tests or procedures. 44 Therefore,
Martin concludes, the object of a religious experience must be subjective and
private, unlike the object of sense experience. Typical of responses to Martin's
claim that there are no tests or procedures for testing the veracity of a person's
claim to religious experience is the one made by Wainwright. He insists that
Martin is simply wrong that there are no tests and procedures for religious
experiences that are similar to the ones used for sense experiences. There are tests
and procedures in religious communities that can be used to assess any S' s claim to
a religious experience e, according to Wainwright. He identifies six criteria that are
used within the Christian community (particularly Catholicism). These include 1)
the consequences of e for S; 2) the consequences of e for others; 3) the "depth, the
profundity, and the sweetness" of S's report of e; 4) the agreement or disagreement
of S's report of e with orthodox beliefs; 5) the similarity of e with other accepted
accounts of other religious experiences within the religious community; and 6) the
pronouncement of the final authority in the religious community regarding e. 45
Wainwright argues that there are significant similarities in the way in which these
criteria are employed in the respective cases of sense experience and religious
experience. At the same time, he recognizes that there are significant dissimilarities
- the most significant of which is the fact that in the case of sense experience the
agreement or disagreement of others and the success or failure of predictions based
upon e are crucial while these considerations are not important in the case of
mystical (religious) experience. 46 Wainwright suggests that this difference may be
accounted for in terms of the differences between the kinds of objects involved in
the different cases. On the one hand, ordinary physical objects are publicly and
readily accessible given their nature, so it stands to reason that we would expect
others to experience them as we do given similar circumstances. On the other hand,
God is radically different from ordinary physical objects - not publicly or readily
accessible. It thus stands to reason that we would not necessarily expect others to
experience God as we do - especially if one of the conditions of one's experience
of God is God's grace, so that God may appear to some while not to others even in
very similar circumstances. 47 But with this response, Wainwright dismantles much
of the case for the similarity between religious experience and sense experience
that he has tried to build. Aside from the rather ad hoc nature of this response, that
is, God is just unique and so different from anything else that we cannot expect the
same rules to apply, Wainwright seems to undermine the other tests and procedures
that he has offered to validate religious experiences, since surely God's unique
nature would override any of the other conditions he has specified for the other
tests and procedures to take place. Thus, any disparity in the account of any failure

C. B. Martin, Religious Belief, Chapter 5.


William Wainwright, ibid., p. 86-87. It is worth noting that these criteria appear to be codified in
William Alston's notion of a Christian mystical doxastic practice (CMP), discussed below.
46 Ibid., pp. 93ff.
47 Ibid.
44

45

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

157

of an experience to meet any of the other tests and procedures could be accounted
for in terms of God's unique nature.
Numerous other challenges, which I cannot address here, have also been raised
against claims for the cognitive status of religious experience and the cumulative
case argument. 48 Theists who appeal to the use of religious experience in a
cumulative case argument must also acknowledge that the same cumulative case
strategy can also be used to oppose the theists' claims regarding religious
experience and the existence of God. In other words, no single objection to the
claims regarding religious experience and its use in the cumulative case argument
to support the existence of God need be regarded as decisive, but when all of the
objections are considered together - including those raised here and others in the
literature - with each compensating for the weaknesses of the others, it could be
argued that it becomes more probable than not that reported religious experiences
are not experiences of God. While I know of no such cumulative case argument
being used against religious experience per se, a similar strategy has been used
effectively by Paul Draper in his probability version of the argument from evil
against the existence of God (discussed in Chapter VI).
THE RIGHT TO BELIEVE
A different reaction to the various problems raised concerning the POC is
offered by John Hick. Hick recognizes and admits not only that the universe is
"religiously ambiguous" but that the universe generally is ambiguous. 49 The
various problems raised above concerning the criteria for distinguishing veridical
from delusory experiences leave us in an uncertain and ambiguous epistemic state;
however, taking his departure from William James's "The Will to Believe," Hick
insists that the religious person has the right to believe what he wants to believe in
the absence of defeating experiences or evidence. The world is an ambiguous place
that must be interpreted. It may be interpreted either religiously or naturalistically,
and there is no definite way of confirming which interpretation is correct. With
either interpretation, one runs the risk of being wrong, but this risk is unavoidable.
The religious person is as rationally justified in trusting the reports of the major
figures in a particular religious tradition and his own religious experience as those
who choose a naturalistic interpretation are in trusting their own sources of
authority and experience. There is always the theoretical possibility, according to
Hick, that all of religion and all of religious experience is illusory, but it would be
irrational for the religious person to give up his religious belief on the basis of what
he regards as such a thin theoretical possibility. The same is true of nontheistic
ways of "seeing" the universe. Thus, the religious person is as rational in choosing
to experience the universe religiously as is the nontheist. 5o

48 See Caroline Franks Davis, The Evidential Force of Religious Experience (Oxford: Clarendon
Press, 1989), Chapter 5.
49 John Hick, An Interpretation of Religion: Human Responses to the Transcendent (New Haven,
Conn.: Yale University Press, 1989). See the excerpt reprinted in Contemporary Perspectives on
Religious Epistemology, edited by R. Douglas Geivett and Brendan Sweetman (Oxford: Oxford
University Press, 1992), p. 316.
50 Ibid., pp. 316-17.

158

ANALYTIC PHILOSOPHY OF REUGION

Hick rests his claim for the religious interpretation of the world upon the more
general epistemological claim that all experience is really a form of "experiencingas.,,51 Hick bases his notion of experiencing-as on Wittgenstein's notion of seeingas that is used to explain visually ambiguous situations such as the Necker cube,
lastrow's duck-rabbit, and Kohler's goblet-faces. 52 Since cases of visual perception
are sometimes ambiguous and can be "seen" in different ways, sometimes as a
duck and sometimes as a rabbit, for example, Hick concludes that the universe can
therefore be experienced in different ways, either religiously or nonreligiously.
According to Hick, in the case of familiar objects, such as a fork on a table, we still
experience such things as what they are, and we call such a process "identifying"
or "recognizing" the fork. Obviously, most members of modern Western European
or American society would easily recognize a fork as a fork whereas a Stone Age
savage would not. 53 All conscious experiencing is thus, for Hick, a form of
experiencing-as, and the religious person who experiences the world religiously is
simply interpreting his sense perception of the world in the same way that is done
with all experience. Sense perception itself involves a subjective element of
interpretation, so religious interpretation is no more subjective and no less
objective than ordinary sense experience.
Hick's notion of experiencing-as is another attempt to construct an account of
religious experience using the analogy of sense experience. However, as Hick
himself recognizes, the claim that one can sense the presence of God is "far
removed" from the claim that one can sense the presence of a fork or a rabbit. 54 It
is so far removed that it appears to be impossible to ever bridge the gap between
the two. It is one thing to claim, as Wittgenstein does, that sense experience
involves some element of interpretation or "seeing-as," and it is quite another thing
to claim that "religious seeing" is simply another element of "seeing-as." It is not
clear even that visual "seeing-as" has obvious counterparts with the other senses,
such as "smelling-as" and "touching-as." In such cases, the subjective,
interpretative element seems clearly less grounded in the sense experience itself. If
Hick's "experiencing-as" extends beyond sense experience to other forms of
experiencing, then the interpretative element will be even less grounded in or tied
to the sense experience. If one person sees a figure as a duck and another person
sees the same figure as a rabbit, it is possible to compare their two ways of seeing
the figure and to get each to see the other's way of seeing by pointing to various
sensible details of the figure. However, if one person sees a burning bush and
another person sees a supernatural presence, it will not be possible to compare their
two ways of seeing and to get each to see the other's way of seeing by pointing to
certain sensible features of the bush. In religious experiencing-as (as opposed to
sensible experiencing-as), the subjective element of the experiencing-as obviously
dominates. The comparison of religious experience to ordinary sense perception is
51 John Hick, "Religious Faith as Experiencing-As," in Classical arui Contemporary Readings in the
Philosophy of Religion, edited by John Hick (Englewood Cliffs, N.J.: Prentice-Hall, 1964, 1990), p.
409.
52 See Ludwig Wittgenstein, Philosophical Investigations, translated by G. E. M. Anscombe (New
York: Macmillan, 1953), Part II, Section xi.
53 John Hick, "Religious Faith as Experiencing-As," p. 409.
54 Ibid., p. 412

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

159

a problem that dominates several different accounts of religious experience. I turn


to a much fuller consideration of this comparison in the treatment of William
Alston's account of perceiving God.
PERCEIVING GOD
William Alston follows in the tradition of those who have defended religious
beliefs by basing them upon religious experience and by explicating religious
experience along lines that are closely analogous with sense perception. Alston
provides one of the most elaborate and detailed contemporary attempts to establish
the reliability of reports of religious experiences. We can understand his point of
contact with sense experience as an attack on attempts to explicate or restrict
condition (iii) of the JTB theory of knowledge to empirical knowledge gained
through sense experience, which he thinks is incapable of meeting any serious
challenge by skepticism without resorting to circularity. Evidentialism takes as its
benchmark our empirical knowledge of the physical world, which, in turn, is based
upon our perception of the physical world. Thus, the problem of our perception of
the physical world and epistemological questions about the reliability of the
knowledge that we claim to have about that world have occupied epistemologists
for centuries. Attempts to respond to skepticism and defend sense perception
simply on the grounds of sense experience itself and to certify it as providing the
"adequate evidence" required by condition (iii) of the JTB theory have all failed,
Alston claims, since all such defenses and apologies are circular. The only way in
which beliefs that are based directly upon sense perception can be defended or
justified at all is by appealing to their coherence with other "background" beliefs.
The justification of perceptual beliefs is then derived from the reliability of the
practice that generates them. The justification of the background beliefs, in turn, is
provided by the doxastic practices that provide the socially approved criteria for
assessing perceptual claims and for distinguishing well-formed, veridical
perceptual beliefs from noureliable ones.
To provide epistemic justification for our beliefs that are based upon a direct
experience of the external, physical world, we must rely upon what Alston calls a
"perceptual belief producing system" composed of the perceptual beliefs, the
background beliefs, and socially established doxastic practices that provide the
standards for forming beliefs in a normal, acceptable manner. 55 Beliefs that are
based upon direct sense experience cannot be justified on their own and by their
own bootstraps, that is, without the socially grounded doxastic elements introduced
by the perceptual belief-producing system that provide the criteria for determining
when and how and under what conditions it is reasonable to accept such beliefs.
While Alston sometimes is regarded as belonging in the camp of Reformed
epistemology, it seems more accurate to see him as attempting to occupy a
"middle" ground between evidentialism and Reformed epistemology. He does not
include himself in the Reformed group, although he does acknowledge some
sympathy with Alvin Plantinga's claim that belief in God is properly basic
55 William P. Alston, Perceiving God (Ithaca, N.Y.: Cornell University Press, 1991), p. 100. For a
precis of this book, see William P. Alston, "Precis of Perceiving God," Philosophy and
Phenomenological Research, Vol. 54, no. 4,1994.

160

ANALYTIC PHILOSOPHY OF RELIGION

(discussed below). He distances himself from Plantinga, however, by criticizing


Plantinga's Reformed approach as too narrow because of its exclusive focus on an
"internal" defense of theism, that is, a defense from within the doxastic practices of
theism. Alston intends to provide some "external" defense of those doxastic
practices that will appeal to those outside the doxastic practices of theism. 56 He
also expresses some sympathy with the cumulative case argument since he thinks
that any particular basis for religious belief cannot, by itself, resolve the doubts that
are raised against it.57 He also appears to be ambivalent about natural theology and
arguments for the existence of God. He is not reluctant to appeal to such
arguments, in principle, to buttress his position regarding religious perception or to
allow a theist to appeal to such arguments to defend his beliefs;58 however, he
seems to think that the arguments are simply too abstract to do the job properly. 59
Like Swinburne, Alston also focuses upon the kind of experience during which
something is given or something appears to the person having the experience. It is
this presentational quality of the experience that causes Alston to call such
experiences perceptual. Thus, perceptions of God are experiences in which the
person having the experience perceives God as being of some nature or performing
some action. Examples of such experiences may include the experience of God
"comforting, strengthening, guiding, communicating a message, sustaining the
subject in being - or to the effect that God has some (allegedly) perceivable
property - goodness, power, lovingness." Such "perceptions" of God give rise to
and provide justification for what Alston calls "M-beliefs" (Manifestation
beliefs).60 According to the theory of appearing, which he relies upon for his
analysis of sense perception, the presentational quality of an experience - the
phenomenological experience of a subject being presented with or given an object
of experience - is fundamental, unanalyzable, and, most importantly, precognitive
(not dependent upon beliefs) and thus a mode ofperception. 61
Here is the crux of Alston's claim: The justification of beliefs about the external
physical world relying upon sense perception (SP) and M-beliefs about God
relying upon perception of God both require the help of doxastic structures and
practices. The Christian mystical perceptual practice (CMP) provides a justification
for M-beliefs about the existence of God in a way that is closely analogous to the
way in which SP provides us with a justification for beliefs about the external,
physical world. CMP provides the doxastic framework for regulating or
standardizing ways of forming M-beliefs based upon the direct perception of God
within what he calls "the main stream" of Christianity.62 M-beliefs that are based
upon reports of direct mystical acquaintance with some object of that experience
are reasonable to accept if they result from the direct perceptual experience and the
doxastic practices that exist in the community of religious believers.
"Normal"sense perceptions and perceptions of God are thus the results of different
William P. Alston, Perceiving God, p. 197.
Ibid., pp. 306-307.
58 Ibid., p. 270.
59 Ibid., p. 144.
60 Ibid., p. 1.
61 Ibid., p. 55. Swinburne treats a subject being appeared to similarly. See above.
62 Ibid., p. 193.
56

57

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

161

doxastic practices - each of which, according to Alston, must be accepted at face


value, given prima facie acceptance, unless there are reasons for disqualifying a
particular experience or a practice. 63 To grant such prima facie acceptance or
justification to beliefs based upon sense perception and its doxastic system, SP, but
not to M-beliefs and CMP amounts to evoking a double standard and requiring
more in the way of epistemic justification for M-beliefs than we do for beliefs
regarding the external world.
It is usual to point out that Alston has not made the case that M-beliefs are true
but only that they are prima facie justified, but such a disclaimer ignores the wolf
lurking under the sheep's clothing. If M-beliefs are prima facie justified for a
subject S holding the beliefs, this means nothing more than that S is justified in
regarding them to be true. And if an observer, K, regards S's M-beliefs to be
justified, then K is justified in regarding S's beliefs to be true, and so on until the
entire community of believers is justified in believing S' s M -beliefs to be justified;
the question of truth lurks close to the surface of prima facie justification. Thus, if
Alston is right, the perception of God that occurs in ways that are approved and
sanctioned by CMP provides prima facie justification for M-beliefs about the
existence and nature of God.
Like Swinburne, Alston recognizes possible defeaters for CMP. Specifically,
naturalistic explanations for the alleged perception of God, contradictions in the
output of CMP, and conflicts between the outputs of CMP with the outputs of
science and the integration of the outputs of CMP into a naturalistic metaphysics.
However, Alston thinks that none of these possible defeaters of CMP is successful.
In each case, he provides a possible way of reconciling the defeater with CMP. In
the case of naturalistic explanations, God may be lurking somewhere further in the
background than naturalistic accounts can reach; in the case of contradictions, each
doxastic practice, including those for sense experience, tolerate some degree of
contradiction; and in the case of conflicts with science, Alston thinks that theism is
basically consistent with science. 64
COMPETING DOXASTIC PRACTICES
Many of the same problems that have been identified earlier regarding
Swinburne's use of religious experience in his cumulative case argument can be
raised against Alston's account of perceiving God as well. However, I will not
reconsider how those same objections pertain to Alston's position but will
concentrate instead on what Alston regards as the most serious problem with his
account of religious experience and its grounding in CMP. The various religious
beliefs in the world and the different doxastic practices that give rise to those
beliefs appear to be in serious, logical conflict with one another, and different
attempts to resolve or dissipate this conflict have not been successful.
Alston considers John Hick's treatment of religious experience and Hick's way
of defusing the apparent differences between different religious beliefs by using
the Kantian distinction between the noumenal Real and its phenomenal

63

Ibid., p. 153.

64

See ibid., Chapter 6, pp. 226ff.

162

ANALYTIC PHU,oSOPHY OF REUGION

manifestations, but he eventually dismisses this approach. 65 Alston's conclusion is


that the conflicts between different and rival socially established doxastic practices
cannot be explained away or vitiated, and this unresolved, logical conflict is the
greatest challenge to his account of religious perception and CMP. Such conflict
amongst rival doxastic practices poses the possibility that it is impossible to give
an external justification of CMP (or any other religious doxastic practice), that is,
whatever justification there is must be internal to CMP. This challenge is a serious
threat to the rationality of accepting CMP since, if the challenge of religious
diversity is not met, it would only be from the epistemic viewpoint of the
practitioner within CMP that one could find good reasons for accepting CMP over
rival sets of doxastic practices for religious beliefs. However, within rival sets of
religious beliefs with different accounts of religious experiences, practitioners are
all in the same epistemic position, so practitioners of other religious practices
would have exactly the same epistemic justification as do members of CMP. 66 If an
irreducible religious pluralism forces the practitioner of CMP to turn inward to
CMP itself for whatever justification there might be, then Alston's enterprise of
providing some external justification for CMP is threatened. The alternative
emphasizes the most fundamental aspect of Alston's account. Alston's grounding
of religious belief based upon religious perception in socially established doxastic
practices means that whatever questions of justification concerning M-beliefs may
arise are removed to the level of those practices. 67 Here, according to Alston,
epistemic preference must be given to the long-standing, firmly established social
practices over what he calls "idiosyncratic" doxastic practices. Since both beliefs
about the physical world and rival M-beliefs are functions of their respective
socially established doxastic practices, the question becomes one of determining
which doxastic practices are reliable in producing beliefs that have prima facie
justification. Alston's answer to this question is given in terms of "practical
rationality." We are rationally justified (in a practical sense) in relying upon
doxastic practices that regulate and standardize the ways in which we form beliefs
when those ways of forming our beliefs "are established in our society and ...
firmly embedded in our psyches." Besides, he claims, "there are no alternatives
that commend themselves to rational reflection as superior," that is, the only way
in which we can form beliefs rationally is by relying upon the ways in which we
normally do exactly that. 68 Reliable doxastic practices must be "established," not
idiosyncratic, since it is reasonable to think that deeply rooted doxastic practices
that have persisted for some period of time would not have done so were they not
reliable. We can assume, therefore, that long-lived, well-established doxastic
practices deserve the benefit of the doubt by granting that the beliefs produced by
them are prima facie justified, but we do not grant the same benefit to beliefs
resulting from eccentric, idiosyncratic practices. "When a doxastic practice has
persisted over a number of generations, it has earned the right to be considered
seriously," he says. On the other hand, "newcomers will have to prove
Ibid., pp. 264-65. See Chapter IX for a discussion of Hick.
See ibid., pp. 266-70.
67 It should be noted that religious experience is one of several different sources of justification of
religious belief for Alston.
68 Ibid., p. 168.
65

66

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

163

themselves.,,69 Fulfilling desired social goals and satisfying the promises for a
richer spiritual life for individuals over a long period of time are marks in favor of
CMP. 70
This response to the challenge posed to CMP by religious diversity signals a
change in tenor of Alston's account of religious experience and CMP. His entire
enterprise is one that proceeds by comparing religious perception with sense
perception and M -beliefs with beliefs based upon sense perception. Alston then
admits that the kind and degree of evidence that it is possible to garner for religious
doxastic practices falls far short of the kind and degree of evidence that it is
possible to garner for sense experience and science. 71 In the face of what Alston
regards as the most serious problem for his account of religious experience, the
challenge to CMP posed by religious diversity, the Christian believer must make
judgments about the relative benefits of CMP over competing doxastic practices,
but such judgments involve results or goals that are decidedly subjective and
internal to CMP, such as a richer spiritual life. In the end, it seems that the believer
must trust his faith and what his faith tells him about the nature of God in order to
justify CMP over other competing accounts of religious experience; otherwise,
there seem to be no grounds for a believer preferring or valuing the benefits of
CMP over the benefits of other doxastic practices. 72
INTERNAL JUSTIFICATION OF CMP?
Several additional objections and criticisms have been raised against Alston's
account of religious experience as religious perception. 73 Many of these criticisms,
which I will not pursue here, are similar to those raised above against Swinburne's
use of religious experience. I will concentrate here instead upon the additional
problems raised by Alston's grounding of religious experience within socially
established doxastic practices. There is the initial issue of explaining the difference
between well-established and idiosyncratic practices in a way that is not ad hoc and
question-begging, and this must be done within a context where the ultimate
criteria for assessing the advantages and disadvantages of different kinds of
doxastic practices are internal to the different doxastic practices. Whereas the
practitioner of CMP may require that a well-established practice be one that has
endured many generations and attracted thousands of converts, practitioners in
other doxastic practices may have very different criteria for winnowing the good
from the bad doxastic practices. Indeed, one can easily imagine that one may
object, as for example a Marxist may, that a doxastic practice's being socially well
established and long-lived requires the political underpinning of oppressive
governments, such as those committed to capitalism, and that this actually counts
against the legitimacy of the beliefs produced by such systems. Instead of the
Ibid., p. 170.
Ibid., p. 276.
71 Ibid., p. 277.
72 Ibid. For further discussion of Alston and the problem of religious pluralism, see Chapter IX.
73 Several critical issues are raised by the respondents in the Alston Symposium in Philosophy and
Phenomenological Research, Vol. 54, no. 4, 1994, including Richard Gale, "Why Alston's Mystical
69

70

Doxastic Practice Is Subjective," George Pappas, "Perception and Mystical Experience," and Robert
Adams, "Religious Disagreements and Doxastic Practices." See also William Hasker, "On Justifying
the Christian Practice," The New Scholasticism, Vol. 60,1986.

164

ANALYTIC PHILOSOPHY OF REUGION

criterion of spiritual enrichment of individuals in CMP to which Alston appeals,


Marxists might plausibly appeal to correcting the economic oppression of the
working class. Adherents to different doxastic practices - theistic and non-theistic
- will not only have different practices producing different beliefs but also
different criteria for assessing the efficacy of those practices and those beliefs.
Appealing to the specific criteria that favor CMP commits Alston to the kind of
religious chauvinism and religious imperialism that he wants to avoid. 74
There is the additional, independent matter of determining how CMP comes to
be regarded as well established. In the beginning, as with all new doxastic practices
one may suppose, Christianity was not socially well established. In fact, according
to even biblical accounts, it was regarded as rather idiosyncratic at the time - just
another "new-fangled" cult. Just how did Christianity become well established and
how did it accumulate the authority that present-day Christians now attribute to its
early claims about religious experiences? The answer cannot be because of the
success of the social goals or because of the fulfillment of the promises of a
spiritual life for its early converts. Consider the account of the appearance of Jesus
to the two men on the road to Emmaus on the very day of his resurrection, or his
appearance to his disciples a few days later, or the appearance of the Holy Spirit on
the Day of Pentecost. These different perceptions of the sacred were all
idiosyncratic at the time of their occurrence, but to explain how CMP came to be
socially well established and to survive for more than two millennia, it seems as if
one must give enormous weight to these early, idiosyncratic experiences that really
occurred outside the venue of a doxastic practice at all. Socially well established
doxastic practices thus must, to some degree, grow out of and be a function of
earlier idiosyncratic practices. 75 But, if this is true, then the problem created by
religious diversity is greatly magnified. In deciding the reliability of different
claims regarding religious experiences, Alston must consider not only the religious
beliefs of the different well-established doxastic practices but the individual,
idiosyncratic ones as well- many of whom are regarded as charlatans or kooks.
If the justification for what is regarded as socially well established CMP is
primarily internal, and if the early experiences that are now regarded as part of the
legitimizing force for CMP were originally idiosyncratic, then one way of
explaining how early, idiosyncratic CMP came to be socially well established CMP
is through the acceptance of some final religious authority, as specified in William
Wainwright's sixth criterion for assessing religious experience (see the above
discussion of Swinburne's cumulative case argument). The suggestion here is that
appeal to religious authority must be much more responsible for the codification
necessary for the doxastic practices of CMP to become socially well established
74 For a more constructive suggestion and optimistic outlook for Alston's defense of CMP, see the
following two articles by Philip L. Quinn: "And Thinner Theologies: Hick and Alston on Religious
Diversity," International Journal for Philosophy of Religion, Vol. 38, 1995, pp. 145-64. Reprinted in
The Philosophical Challenge of Religious Diversity, edited by Philip L. Qninn and Kevin Meeker
(Oxford: Oxford University Press, 2000), especially pp. 235ff; and "Religious Diversity and Religious
Toleration," International Journal for Philosophy of Religion, Vol. 49, 2001, pp. 1-24. Qninn's
treatment of Alston is discussed in Chapter IX.
75 Robert Adams suggests that more emphasis should be placed on the individual's belief rather than
the social nature of the doxastic practice, but he does not develop this point. See Robert Adams,
"Religious Disagreements and Doxastic Practices," p. 886.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

165

than Alston recognizes. Consider just one such well-known example of religious
authority at work. Joan of Arc was burned at the stake as a heretic in 1431 at least
in part because she claimed that the voices she heard were from God. In 1920, Joan
was canonized by Pope Benedict XV and became Saint Joan. Arguably, politically,
in terms of how the French regarded Joan because of her attack on the English at
Orleans during the Hundred Years War, she was not in need of canonization.
However, in terms of her position within the Roman Catholic Church and in terms
of how her report of her idiosyncratic perceptions of God came to be regarded as
part of the socially well established CMP, the annulment of her sentence and her
canonization were crucial. 76 This case and the role of authority in the process of
turning idiosyncratic experiences and beliefs into accepted parts of the socially
well established CMP suggest that M-beliefs are not simply and directly justified,
as Alston claims. M-beliefs are supposed to be prima facie reliable, since they are
apparent, direct perceptions of something, but idiosyncratic M-beliefs do not wear
their legitimacy on their sleeves. M-beliefs must be sanctioned by CMP and await
further approval by and adoption into the tradition that comprises CMP. Whether
there is a comparable authority at work in the case of sense experience is not
obvious. At least there is no comparable explicit secular counterpart to
ecclesiastical authority.77 One way of describing the process whereby idiosyncratic
M-beliefs become beliefs accepted by CMP is that what begin as introspective and
subjective experiences then take on the same epistemic status as perceptual
experiences; however, surely this process would not affect the phenomenological
analysis of the experiences for the person having the experiences. 78
THE METAPHYSICAL REQUIREMENT
Part of Alston's answer to the problem of how S can be sure that x is God when
x appears to S as God involves abandoning a classical descriptivist account of
reference, such as the one that, as I have shown above, Swinburne embraces. For
Alston, for S to recognize an object x as x when it appears during a perception, it is
not necessary for S to perceive any essential properties of x. According to Alston,
it is sufficient for S to recognize x as x for x simply to present itself as possessing
properties that "are sufficiently indicative of (are a reliable guide to) the object's
being" what it is. 79 But this answer will not do as a general account of how S
comes to be able to identify x as God since it requires S to have a store of
knowledge of which properties are "sufficiently indicative" and "reliable" guides
for identifying God and which ones are not. For S to develop such an ability to
discriminate amongst different properties in terms of their success or lack thereof
in identifying God seems to require S to have repeated experiences in the past of

76 Arguably, equally for Roman Catholics and non-Catholic Christians alike since the legend of St.
Joan is now so thoroughly embedded in Western folklore, Other saints have reported not being so sure
of the source of their religious experiences at the time that they were having them. See St. Teresa,
Interior Castle (London: Thomas Baker Publishing, 1930).
77 For Alston's comparison of SP and CMP on this score, see Perceiving God, Chapter 3 and pp.
250ff.
78 For further criticism of Alston on the grounds that nonsensory perception of God is subjective, see
Richard Gale, "Why Alston's Mystical Doxastic Practice Is Subjective," pp. 869-75.
79 William Alston, Perceiving God, pp. 96-97.

166

ANALYTIC PHILOSOPHY OF REDGION

successfully and unsuccessfully identifying x as God when x has appeared as


possessing certain properties. so
Modeled on sense perception, any religious experience must satisfy what
Richard Gale has called the "metaphysical requirement."Sl In the case of sense
experience, this requirement demands that any S's experience of an any object x
must take place within a spacial-temporal dimension D that allows S to individuate
the objects of experience and thus individuate x from other objects in D. The
existence of D is also necessary to specify the conditions under which x is able to
appear to S and the position that S occupies within D relative to x. Using D, we can
also specify S's position relative to other observers and the relationships between x
and other observers. In order not to beg the question against Alston, since Mexperiences are supposed to be nonsensory perceptual experiences, the
metaphysical requirement can be specified in a more general way to the effect that
the object x of S's experience must occupy some dimension MD (a "mystical
dimension," although not necessarily a spatial-temporal one). MD is necessary in
order for S to identify x and individuate x from other objects in MD. Using MD,
we can understand the conditions under which x is able to appear to S in MD, S's
position within MD, S's relationship to x within MD, as well as S's relationship to
other observers within MD and their relationship to x. It is easy to explain the
metaphysical requirement for M-experiences using the jargon of sense-data theory.
During a perceptual experience, whether sensory or non sensory, sense data must
appear to S within a wider, sensory field within which the sense data move or
change and are related to other sense data within the field and to S. If S's entire
sense field is filled with the same sense datum, such as a bright, glowing red, then
S can make no sensory or perceptual discriminations about the experience other
than, in the classic sense-data example, "red, here now." Clearly, however, during
M-experiences, S is supposed to have finer and more detailed powers of sensory
awareness, without which the experience would never be regarded as mystical or
religious or a case of being appeared to by God.
Alston's response to this objection does not adequately address the seriousness
of the problem. He simply repeats his claim that we do not have to reply upon
essential characteristics in identifying God and that we use "background
knowledge" to differentiate those characteristics that are reliable and sufficiently
indicative of God's presence from those that are not. The force of the objection,
however, is that the accumulation of such background knowledge is impossible
unless, during those earlier experiences in which some S develops the background
knowledge, it is possible to differentiate x from other objects that appear to S in
MD and to identify God as x instead of identifying God as y or z or some other
object that appears in a nonsensory fashion to S within MD. To do this
successfully, or even to judge whether he has done it successfully or not, S must
differentiate x from y and z and other objects that appear in a nonsensory fashion
80 While Alston distinguishes between S's being justified and S's knowing that he is justified in
believing that he is being appeared to by God, still unless there is some indication of what might be the
criterion for determining whether S is right or wrong in believing that the properties that x presents in
appearing to S are sufficient for identifying God, then there is no reason to maintain that it is reasonable
to believe that S is being appeared to by God.
81 Richard Gale, "Why Alston's Mystical Doxastic Practice Is Subjective," pp. 871ff.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

167

to S within MD. Such differentiation can only take place within a mystical
dimension, according to which it becomes possible to describe the position of x
within that dimension, the relative positions of other objects within that dimension,
the position of S within that dimension, the relative position of S to x within that
dimension, the position of other observers within that dimension, and the relative
positions of other observers to x within that dimension. For example, suppose that
God appears with the heavenly hosts or simply with the archangel Gabriel, then S
must be able to distinguish God from the heavenly hosts and from Gabriel within
MD. M-experiences clearly do not satisfy the metaphysical requirement, and
Alston apparently does not conceptualize M-experiences taking place within a
mystical dimension in the same way in which sensory perceptions take place
within a spatial-temporal dimension. To understand nonsensory perception in the
model of ordinary sensory perception, we must understand how this metaphysical
requirement could be satisfied. If Alston denies that M-experiences require a
dimension in the same way that sense experiences do, then the proper response is
that he has insisted upon drawing the parallel and the comparison between Mexperiences and sense experiences as closely as possible; therefore, to abandon or
deny the parallel structure of such experiences at this point would seem to
undermine the entire enterprise.
DIRECT REFERENCE AND RELIGIOUS PERCEPTION
One plausible way of explaining how one is able to identify God during a
nonsensory experience of God is by abandoning the use of a descriptivisitic
account of reference during such times. Perhaps for this reason Alston has joined
forces with Richard Miller in claiming that the causal theory of reference,
championed by Saul Kripke and Keith Donnellan, offers some advantages in
explaining how reference takes place during religious experience. 82 This theory
emphasizes the fundamental nature of what Alston calls "direct reference," that is,
reference that takes place when a speaker uses a referring expression to refer to an
object x of immediate experience without using predicates or characteristics of x.
According to Alston, religious believers refer to God in a primary sense when they
used the name 'God' during worship and prayer to refer to the object of their direct
religious experience. Using the causal theory of reference "anchors" the referent
for 'God' in direct experience without the need for identifying or essential
predicates, and other people within the same religious community can then trace
their use of 'God' to the same initial baptism of the use of the word. 83
Alston recognizes that there are several controversial aspects to his claims
involving direct experience of God and direct reference to God as an object of such
experience. I will explore just the major ones of those difficulties. In the case of
direct reference, it is crucial that the speaker be able to successfully refer to a
particular individual without saying anything that is descriptively true of that
individual. The use of any referring expression in such a circumstance amounts to
what Keith Donnellan calls a referential use of a definite description (in contrast
Richard B. Miller, "The Reference of 'God,'" Faith and Philosophy, Vol. 3, no. 1, 1986.
William P. Alston, "Referring to God," International Journalfor Philosophy of Religion, Vol. 24,
no. 3, November 1988, pp. 10-12. Reprinted in Divine Nature and Human Language: Essays in
Philosophical Theology, William P. Alston (Ithaca, N.Y.: Cornell University Press, 1989).
82

83

168

ANALYTIC PHILOSOPHY OF REUGION

with an attributive use). 84 In the case of the referential use of a definite description,
a speaker's audience is able to identify the intended referent of the referring
expression even if what the speaker uses the referring expression to state about the
individual is not true. It is plausible to understand how such "direct reference" can
take place when the referent being referred to is an object of public experience.
Donnellan gives clear and convincing examples of the referential uses of definite
descriptions that have become well known. If a speaker sitting in a courtroom nods
at the accused and says to an audience, "Smith's murderer is insane," then the
speaker's audience is able to identify the individual about whom the person is
speaking even if it later turns out that Smith was not really murdered at all and the
accused is innocent. Such a situation requires that the referent be an object of
public experience where both speaker and audience can identify the same referent
for the referring expression; otherwise, one would never be in a position to
determine that the audience and the speaker were both identifying the same
individual, that is, referring to the same individual. 85 When a theist uses direct
reference to refer to God when the object of immediate experience is a public
object, such as a burning bush or a beautiful sunset, then the referent of 'God'
becomes tied too closely to the physical world. Given the transcendent nature of
God, if a person identifies the bush or the sunset with God, then a theist must say
that the person has gotten it wrong. The referent of 'God' is not the physical object
that is publicly observable.
If the referent of a referring expression used to refer to an object of direct
experience is not a publicly observable individual, then other problems arise. I will
summarize the main point by simply saying that there would then be no
Wittgensteinian criterion of correctness to determine correct and incorrect
reference for such a referent. 86 To use Alston's example, Satan may "represent"
himself to a person and give that person evil "messages." In such a situation, using
direct reference, the person could not identify the object of his direct experience as
Satan rather than God. In the first place, by using the causal theory of reference, no
description or content can be used to identify the referent, so the content of the
"evil" messages cannot be used to rule out the possibility that this may really be
God appearing to him. All the speaker is left with is the "bare" individual of his
experience, and if this is an essentially private object of experience, then the person
is left with no way of determining whether a particular judgment about that
individual would be correct or not. It may be possible to "fix" the referent for a
referring expression upon a single occasion of direct experience, but it would be
impossible ever to determine the referent or ever to determine that a future referent
of a later use of the referring expression was the same individual. Lacking any
criterion of correctness, there would thus be no way for the theist to determine
whether the object of an immediate experience is the same individual that he has
earlier referred to when he has used the name 'God' or whether the object of this
84 See Keith Donnellan, "Reference and Definite Descriptions," in Readings in the Philosophy of
Language, edited by Jay Rosenberg and Charles Travis (Englewood Cliffs, N.J.: Prentice-Hall), p. 198,
and James F. Harris, 'The Causal Theory of Reference and Religious Language," International Journal
for Philosophy of Religion, Vol. 29, 1991, pp. 82-83.
85 Harris, ibid., p. 83.
86 Ibid., pp. 83-84.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

169

immediate experience is the individual that he has earlier referred to when he used
the name 'Satan.' 87 Alston also claims special benefits for the causal theory of
reference that may contribute to and support ecumenism and perhaps address the
problem that religious pluralism poses for CMP. Such benefits seem initially
plausible since, as we have seen, fixing the referent for a referring expression is
completely independent of the accuracy of any content of the descriptions used to
refer to that individual. By abandoning any claim for the validity of a specific
historical content, the causal theory of reference creates the possibility of a single
referent for the many different and disparate referring expressions used by different
religious peoples in different religious traditions, but this is, at best, a very remote
possibility. This would be as far as the benefit of the causal theory would extend,
however, because there would be an enormous gap between any theoretical benefit
and any practical benefit. To gain any practical benefit, we would have to trace
each causal chain of different referring expressions used in each religious tradition
back to each one's initial "baptism" and then somehow demonstrate that the
individual referred to in each case is really the same individual! While there is the
remote theoretical possibility that a single referent for different referring
expressions was fixed at their initial uses, the practical possibility of identifying the
same referent for all of the many different religious referring expressions seems so
improbable that it is not clear how one would even go about attempting it. 88
Miller and Alston suggest that the causal theory may be a way of referring that
has some benefits for dealing with some of the special difficulties of treating
problems of reference within religious language. However, it seems that the theory
of reference that is preferable for use within the context of religious language is a
minimal descriptivist theory of reference that is called the cluster theory. Consider
Wittgenstein's treatment of how the name "Moses" might refer using the cluster
theory of reference compared to the earlier treatment by Kripke using the causal
theory of reference:
If one says "Moses did not exist," this may mean various things. It may mean:

the Israelites did not have a single leader when they withdrew from Egypt - or:
their leader was not called Moses - or: there cannot have been anyone who
accomplished all that the Bible relates of Moses - or: etc. etc. - We may say
following Russell: the name "Moses" can be defined by means of various
descriptions. For example, as "the man who led the Israelites through the
wilderness," "the man who lived at that time and place and was then called
'Moses,'" "the man who as a child was taken out of the Nile by Pharaoh's
daughter" and so on ....
But when I make a statement about Moses, - am I always ready to substitute
some one of those descriptions for "Moses"? I shall perhaps say: By "Moses" I
understand the man who did what the Bible relates of Moses, or at any rate a
good deal of it. 89

See Ludwig Wittgenstein, Philosophical Investigations, Section 79.


James Harris, "The Causal Theory of Reference and Religious Language," pp. 84-85.
89 Ludwig Wittgenstein, Philosophical Investigations, Section 79.
87

88

170

ANALYTIC PHILOSOPHY OF REUGION

This view ties the reference of the name 'Moses' in some minimal but necessary
way to the set of descriptions that are normally used to explain the meaning of the
name. It does not insist that it is necessary for anyone of those descriptions to be
true but rather relies upon some minimal (and unspecified) subset of those
descriptions being true. 90 This cluster theory of reference for proper names, which
depends to some extent upon the accuracy of the biblical account of Moses,
certainly offers a more plausible account for satisfying Miller's concern about the
"historical validity of religion" than does the causal theory of reference.
ANTI-EVIDENTIALISM AND REFORMED EPISTEMOLOGY
In the second part of the twentieth century, one of the strongest attacks on
classical evidentialism came from Alvin Plantinga, who explicitly addresses the
evidentialist objection to theistic belief raised by William Clifford (discussed
above) that theistic belief is uureasonable and unjustified since there is not
sufficient evidence to justify such a belief.91 In general, evidentialism maintains
that "the strength of one's belief ought always to be proportional to the strength of
the evidence for that belief.'.92 Nicholas Wolterstorff summarizes nicely how the
evidentialist challenge to religious belief has been made:
The challenge can be seen as consisting of two contentions: It was insisted, in
the first place, that it would be wrong for a person to accept Christianity, or any
other form of theism, unless it was rational for him to do so. And it was
insisted, secondly, that it is not rational for a person to do so unless he holds his
religious convictions on the basis of other beliefs of his which give to those
convictions adequate evidential support. No religion is acceptable unless
rational, and no religion is rational unless supported by evidence. 93
Plantinga aims his attack at classical epistemological foundationalism, perhaps
the dominant form of evidentialism since the Enlightenment, according to which
all knowledge is constructed upon basic beliefs that provide the epistemological
foundation for other beliefs that are justifiably inferred from those basic beliefs. He
introduces the notion of a noetic structure, which is the set of beliefs a person
holds and the relationships amongst those beliefs. One's noetic structure is divided
into those beliefs that are basic (those that are held on the basis of no evidence) and
those that are not (those that are held on the basis of evidence).94 While the
See John Searle, "Proper Names," Mind, Vol. 67,1958.
Plantinga has developed his position in a number of different writings, including the following:
"Is Belief in God Rational?" in Rationality and Religious Belief, edited by C. F. Delaney (Notre Dame,
Ind.: University of Notre Dame Press, 1979); "Is Belief in God Properly Basic," Nous, Vol. 15, 1981,
pp. 41-51; "Rationality and Religious Belief," in Contemporary Philosophy of Religion, edited by
Steven Cahn and David Shatz (Oxford: Oxford University Press, 1982). Parts of this article come from
two previous articles: "Reason and Belief in God," in Faith and Rationality, edited by Alvin Plantinga
and Nicholas Wolterstorff (Notre Dame, Ind.: University of Notre Dame Press, 1983); and "Episternic
Justification," Nous, Vol. 20, 1986, pp. 3-18. See especially "Rationality and Religious Belief," pp. 25562.
92 Alvin Plantinga, "Rationality and Religious Belief," p. 24.
93 Nicholas Wolterstorff, "Introduction," in Faith and Rationality, edited by Alvin Plantinga and
Nicholas Wolterstorff (Notre Dame, Ind.: University of Notre Dame Press, 1983), p. 6.
94 Alvin Plantinga, "Rationality and Religious Belief," pp. 48ff.
90
91

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

171

determination of what counts as properly basic beliefs has varied, in terms of


classical foundationalism, the standards are very high, and thus very few beliefs are
regarded as properly basic beliefs. For the classical foundationalists, including
Descartes, only those beliefs that are intuitively self-evident or incorrigible are
properly basic. Others have maintained, including radical empiricists of the early
twentieth century, that only the reports of the content of immediate sense-data
experience can constitute properly basic beliefs. In developing his position, which
has now become known as Reformed epistemology, Plantinga objects that the
restrictions placed upon properly basic beliefs by classical foundationalism are
arbitrary, self-serving, and too limited. Classical foundationalism is thus no more
than "a bit of intellectual imperialism.,,95 Plantinga wants to eliminate the
intellectual imperialism by abandoning the overly restrictive criterion of proper
basicality used by classical foundationalism and broadening the category of
properly basic beliefs.
Reformed epistemologists, who follow John Calvin in this regard, maintain that
natural theology has made the mistake of responding to the evidentialist objection
to theistic belief by trying to provide the evidence and arguments for theistic belief
demanded by the evidentialists. This is seen as a misguided and inappropriate
response to the evidentialist challenge. 96 Plantinga maintains that belief in God
does not have to be based upon evidence or argumentation, but that it is
nonetheless completely rational and reasonable to hold a theistic belief in the
existence of God since such a belief can be properly basic. Establishing a belief as
a properly basic belief does not mean that it is automatically true or that its truth is
independent of any evidence whatsoever, but it does mean that its reasonableness
or rationality is independent of propositional evidence or a process of reasoning
whereby the existence of God is inferred from some other belief. Proper basicality
establishes the epistemic status of a belief and not its truth or falsity. For Plantinga,
if a person does not have good reasons for the belief in the existence of God or the
belief in the existence of God is not supported by evidence or arguments, it does
not follow that the belief is therefore groundless or irrational or unreasonable.
Reformed epistemology thus derives from a Reformed account of foundationalism
and a broader version of foundationalism that includes belief in the existence of
God amongst the properly basic beliefs. This is the basic platform of Reformed
epistemology. For the Reformed platform to be plausible epistemologically, it must
not be simply arbitrary or established by fiat. In the case of classical
foundationalism, there are criteria or a set of conditions for determining properly
basic beliefs. What are the criteria or the conditions of proper basicality upon
which the Reformed platform rests?

Ibid., p. 270.
Alvin Plantinga, "The Reformed Objection to Natural Theology," Christian Sclwlar's Review,
Vol. 11, no. 3, 1982. Page numbers refer to reprint in Philosophy of Religion: Selected Readings, edited
by Michael Peterson, William Rasker, Bruce Reichenbach, and David Basinger (Oxford: Oxford
University Press, 1996); Nicholas Wolterstorff, "The Migration of the Theistic Arguments: From
Natural Theology to Evidentialist Apologetics," in Rationality, Religious Belief, and Moral
Commitment, edited by Robert Audi and William J. Wainwright (Ithaca, N.Y.: Cornell University Press,
1986); and "The Reformed Tradition," in A Companion to the Philosophy of Religion, edited by Philip
L. Quinn and Charles Taliaferro (Oxford: Blackwell Publishers, 1997), pp. 165-70.
95

96

172

ANALYTIC PHILOSOPHY OF REUGION

Plantinga does not begin his rationale for Reformed epistemology by offering or
defending a criterion for or a set of criteria for determining the extended class of
properly basic beliefs. Plantinga insists that we must arrive at any such criteria
inductively by considering examples of possible candidates for properly basic
beliefs, deciding the proper basicality of those candidates, and then abstracting the
necessary and sufficient conditions that explain their basicality. The criteria must
be constructed "from below," that is, from the ground up rather than from the top
down, by considering different examples, and Christians are as free to choose their
examples as nontheists. There is no reason to insist that everyone agree on the
examples. "The Christian community is responsible to its set of examples" and not
to those of anyone else. 97
So Plantinga does not offer a criterion or criteria for properly basic beliefs;
instead, he defends the Reformed platform by making use of what Terence
Penelhum has dubbed the parity argument. 98 The parity argument is used to attempt
to minimize the epistemological difference between theistic beliefs based upon
faith and secular beliefs based upon science or common sense by claiming that
many of the latter beliefs cannot be justified solely on the basis of evidence or
human reason. In the present context, the parity argument proceeds by granting to
the classical foundationalist that some beliefs are properly basic and then
attempting to draw close parallels between those admitted properly basic beliefs
and the belief in the existence of God so that theistic belief becomes included in
that accepted group of properly basic beliefs. Support for the Reformed platform is
derived if these parallels can be drawn closely enough. Plantinga begins this
process by focusing upon such commonly held beliefs as the following: 99
i. I see a tree,
ii. I had breakfast this morning,
and
iii. That person is angry.
These are examples, Plantinga claims, of beliefs that are not considered
groundless even though they are properly basic. Having a certain experience, such
as seeing a tree or being appeared to treely, may be grounds for (i) and may justify
(i), but the experience is not adequate evidence for (i) that we then use to argue for
or from which we infer the truth of (i). With each of these properly basic beliefs,
"there is some circumstance or condition that confers justification; there is a
circumstance that serves as the ground of justification."lOo
Plantinga claims that the same sort of analysis holds for a theistic belief in God
and that thus, according to the parity argument, belief in God should be treated

Alvin Plantinga, "Rationality and Religious Belief," pp. 76-77.


Terence Penelhum, God and Skepticism (Dordrecht, Netherlands: D. Reidel, 1983), p. 30 and pp.
146-47. Penelhum groups Plantinga with Pascal and Kierkegaard as an evangelical fideist - a label that
Plantinga disavows (see his "Reason and Belief in God," p. 90). The question of fideism aside,
Penelhum's parity argument does capture Plantinga's general epistemological strategy of treating belief
in God in the same manner as secular, common-sense beliefs, so I will use this designation for
Plantinga's approach. I discuss the issue of fideism below.
99 Alvin Plantinga, "Rationality and Religious Belief," p. 271.
100 Ibid.
97

98

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

173

similarly as other properly basic beliefs. 101 Refonned epistemology thus regards
belief in God as a properly basic belief without admitting that it is groundless or
unjustified. With (i) through (iii), there are conditions under which each is regarded
as properly basic. What are the conditions under which belief in the existence of
God is properly basic? In answering this question, Plantinga is very Calvinistic.
God has created human beings with a special propensity to believe in his existence,
a sensus divinitatis, and a special disposition to fonn this belief when it is triggered
by various conditions that cause human beings to form the belief that God exists.
Although Plantinga disavows any attempt to provide a full treatment of these
different conditions, they may include such things as "guilt, gratitude, danger, a
sense of God's presence, a sense that he speaks, perceptions of various parts of the
universe.,,102 Under such conditions, the sensus divinitatis is stimulated, and the
person then is likely to form such beliefs as God is speaking to me, or God has
created all of this, or God disapproves of what I have done, or God forgives me, or
God is to be thanked and praised. All of which mean, loosely speaking, that the
person believes that God exists. 103 As I have noted earlier, saying that such beliefs
are properly basic does not mean that they are thereby determined to be true; it
simply means that it is rational for the person to hold the beliefs and to regard them
as true.
THE GREAT PUMPKIN OBJECTION
Perhaps the most obvious and serious set of objections to Refonned
epistemology and treating belief in God as properly basic is the result of what
Terence Penelhum calls the "deep unease" that it generates. Plantinga's approach
seems to leave the door wide open for different people to make different claims
about properly basic propositions so that "anything goes" in a way that leads to
"dogmatic assertions" and a "potential chaos of clashing convictions."I04 Plantinga
anticipates this objection by raising the possibility of a person believing in the

!O1 Some take the import of Reformed epistemology to be positive while others take it to be purely
negative, that is, some take Reformed epistemology to build a positive case for the Reformed platform
while others take it to simply shift the burden of proof to those who would oppose the Reformed
platform. See especially Richard Grigg, "Theism and Proper Basicality: A Response to Plantinga,"
International Journal for Philosophy of Religion, Vol. 14, 1983, pp. 123-27. My own reading of
Plantinga is that he attempts to build a positive case for the Reformed platform, but this reading is not
uncontroversial. See, for example, Mark McLeod, ''The Analogy Argument for the Proper Basicality of
Belief in God," International Journal for Philosophy of Religion, Vol. 21, 1987, pp. 3-20.
102 Alvin Plantinga, "Rationality and Religious Belief," ibid., p. 273.
!O3 Ibid., p. 273-74.
104 Terence Penelhum, Reason and Religious Faith (Boulder, Colo.: Westview Press, 1995), p. 95.
The fact that Plantinga himself anticipates this objection and tries to respond to it has not prevented his
critics from pressing the issue, and several have, including David Basinger, "Reformed Epistemology
and Hick's Religious Pluralism," in Philosophy of Religion, edited by Michael Peterson et al. (Oxford:
Oxford University Press, 1996), pp. 336-46; Jay M. Van Hook, "Knowledge, Belief, and Reformed
Epistemology," The Reformed Journal, Vol. 32, 1981, p. 16; and William Alston, "Plantinga's
Epistemology of Religious Belief," in Alvin Plantinga, edited by J. E. Tomberlin and Peter van Inwagen
(Dordrecht, Among: D. Reidel, 1985), pp. 300-301. For the effects of Plantinga's position on relativism
and religious pluralism, see William Lad Sessions, "Plantinga's Box," Faith and Philosophy, Vol. 8, no.
1,1991, pp. 51-66.

174

ANALYTle PHILOSOPHY OF RELIGION

return of the Great Pumpkin on Halloween. 105 Since the Reformed epistemologist
maintains that belief in the existence of God is properly basic, then will he not be
forced to admit that belief in the Great Pumpkin as well as other "bizarre
aberrations" can be properly basic also, "thus throwing wide the gates to
irrationalism and superstition,,?l06
Plantinga's response to this objection is not very satisfying. He argues that,
simply because the Reformed epistemologist rejects the criterion of classical
foundationalism for proper basicality, he is not therefore committed to accepting
just any belief as properly basic. But the issue is not whether the Reformed
epistemologist himself would be forced to accept belief in the Great Pumpkin but
rather whether the Reformed epistemologist would be forced to admit that those
who accept such a belief are as entitled to it as a properly basic belief as theists
who accept belief in God as properly basic. Since rationality for Plantinga is a
combination of the logical, cohesive relationship among a person's beliefs along
with the process by which and the conditions under which they are formed, then it
is obviously possible that different people may develop logical, cohesive beliefs
about the Great Pumpkin, UFOs, space aliens, or other "bizarre aberrations."
To explain the difference between properly basic beliefs and other beliefs,
Plantinga relies upon his inductive, "from the ground up" approach, which starts
with the accumulation of a set of beliefs that are "obviously" properly basic and
then proceeds to an inferred or an extrapolated criterion for proper basicality from
that set of agreed-upon properly basic beliefs. In the case of the Christian's
identification of belief in the existence of God as properly basic, the difference is
to be accounted for in terms of the natural tendency implanted by God in all men to
believe in the existence of God, the sensus divinitatis. Presumably, there is no such
natural tendency implanted in us to believe in the Great Pumpkin or UFOs and so
forth. 107
Given this kind of response, it is difficult to see how Plantinga and other
Reformed epistemologists can avoid either relativism, subjectivism, irrationalism,
or superstition on the one hand or arbitrariness, question begging, or circularity on
the other hand. If the way out of this problem for the Reformed epistemologist
requires relying upon John Calvin's notion of a sensus divinitatis, a natural
tendency to "see God," then this seems to be a classic bootstrap method of building
the case for the Reformed platform. If the criterion for proper basicality is to be
developed inductively by examining a lot of different cases of properly basic
beliefs, then how are we to decide which beliefs are properly basic and whether
belief in God is to be counted among them? The answer to this question appears to
be that only those believers who are properly exercising their natural tendency to
see God can determine that belief in God is properly basic. Others - both
nonbelievers and believers whose sensus divinitatis is malfunctioning because of
sin - may not think that belief in God is properly basic, but that is just because they
are defective in some way because they are sinful.
105 An example taken from the cartoon strip "Peanuts" and, ironically, an example I've used in class
for over twenty years to illustrate Hume's point of how the evidence for the argument from design
underdetermines the appropriate analogy to be used.
106 Alvin Plantinga, "The Reformed Objection to Natural Theology," p. 318.
107 Ibid., pp. 319-20.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

175

Plantinga is left without a criterion of proper basicality where any disagreement


about the examples can be dealt with in a very effective but nonarbitrary manner.
Plantinga says that the Christian community is free to choose and use its own set of
examples of what it takes to be properly basic beliefs. Now, leaving aside the
matter of the unlikelihood of ever being able to identify a single Christian
community, the issue of how one could ever arrive at an agreed-upon set of
properly basic beliefs is a troubling one. Not only would nonbelievers disagree
with believers, but obviously believers would disagree among themselves. The
significant feature of Plantinga's examples from common sense, (i) through (iii)
above, is that they are claims about which we would find near universal agreement
- lunatics and eccentric philosophers aside. lo8 It is such near universal agreement
that lends rationality (but not truth) to these beliefs. But there is no such near
universal agreement about the Reformed platform, that is, the expansion of
properly basic beliefs to include belief in the existence of God. In fact, as Gary
Gutting notes, there is the disagreement of substantial numbers of people who
would be regarded as epistemic peers of those who do hold such beliefs. 109 The
whole history of natural theology - from Aquinas to the present day - argues
against the Reformed platform, and to dismiss this history as completely
misguided, Reformed epistemologists must dismiss natural theologians as well as
epistemically defective and sinful.
Plantinga and fellow Reformed epistemologists must maintain that natural
theologians have not only been wrong about the epistemic status of belief in the
existence of God, but their mistake must be explained as the result of a defective
epistemic mechanism. If belief in the existence of God is properly basic, as the
Reformed platform claims, then failure to "see" or admit or accept the basicality of
this belief because of a defective sensus divinitatis corrupted by sin is not only
wrong but irrational as well. Dismissal of such a long-standing and dominant
tradition within Christianity on these grounds is arbitrary and question-begging.
Proper basicality turns out to be the epistemic high ground, since resolving
disagreement about the status of the belief that God exists depends upon
recognizing the correct set of beliefs as properly basic. However, an account of
how the proper set of properly basic beliefs is determined depends upon accepting
the theological claims regarding the sensus divinitatis and its import. But obviously
these claims themselves are not properly basic. Acceptance of such a controversial
theological claim must either be the result of an evidential-styled argument or
proof or the result of a religious doctrine that follows from accepting one particular
religious tradition or religious authority rather than another. We see at work here
something very similar to the need of a religious authority that we saw at work
above in William Alston's treatment of doxastic practices. Plantinga's "bottom up"
technique for establishing the criteria for proper basicality by inductively
examining certain selected examples of properly basic beliefs requires a similar
kind of authority to explain why some people just "see" that certain beliefs are
properly basic while others do not.
108 See Gary Gutting, Religious Belief and Skepticism (Notre Dame, Ind.: University of Notre Dame
Press, 1982), p. 83.
109 Ibid.

176

ANALYTIC PHILOSOPHY OF REUGION

REASON WITHIN THE BOUNDS OF RELIGION


The charge that Reformed epistemology depends upon embracing fundamental
theological beliefs that provide the core of religious epistemology and eventually
determine what is rational or reasonable is not a charge that would greatly concern
Reformed epistemologists. In Reason within the Bounds of Religion, for example,
Nicholas Wolterstorff explicitly embraces the view that fundamental "control
beliefs" function to make Christians both reject certain kinds of theories because
they do not comport well with the Christian's control beliefs and also to devise
certain kinds of theories because they do comport well with those same control
beliefs. 110 W olterstorff cites several examples of how he thinks control beliefs
function in the natural sciences and social sciences, but his main aim is to make
explicit how he thinks control beliefs should function in the case of the philosophy
of religion and, in particular, in the case of Christian philosophers. He claims that
"the religious beliefs of the Christian scholar ought to function as control beliefs
within his devising and weighing of theories," and furthermore, having such
religious beliefs "functioning as control beliefs is absolutely central to the work of
the Christian scholar."lll Wolterstorff captures the theme that runs through much
of this chapter from Swinburne to Alston to Plantinga - the rejection of classical
natural theology (to one degree or other) and the acceptance of a more faith-based
religious epistemology. Whether there is any epistemology left in Reformed
epistemology given this explicit recognition of the fundamental role of faith or
whether Reformed epistemology simply resolves into theology are issues that need
to be addressed. If the epistemology in Reformed epistemology is derivative from
and dependent upon Reformed theological claims that, in turn, are derivative from
and dependent upon religious faith, then "Reformed" in "Reformed epistemology"
does not modify or characterize a kind of epistemology but rather vitiates the
substantive noun in the way that "counterfeit" vitiates the substantive noun in
"counterfeit currency."
EVANGELICAL FIDEISM?
It is perhaps because of these considerations that Terence Penelhum treats
Plantinga's Reformed epistemology as a form of evangelical fideism and groups
him with earlier figures such as Pascal and Kierkegaard. ll2 Fideism comes in
different forms and in different strengths, and there is a wide gap between
Plantinga and Kierkegaard, that is, since nothing in Reformed epistemology places
belief in God beyond the scope of reason, it is far removed from Kierkegaardian
fideism. As Penelhum characterizes evangelical fideism, it is a modest appellation
that depends upon simply divorcing belief in God from reason and evidence and
from attempts to prove the existence of God using the traditional arguments found
in natural theology. In other words, evangelical fideism sounds very much like the
anti-evidentialist stance and the rejection of natural theology that characterize
110 Nicholas Wolterstorff, Reason within the Bounds of Religion (Grand Rapids, Mich.: William B.
Eerdmans, 1976), p. 64.
111 Ibid., p. 66. It should be noted that Wolterstorff intends that control beliefs should function
generally this way for the Christian scholar in his weighing of such scientific theories as Copernican
astronomy and Darwinian evolutionary theory. See ibid., p. 77.
112 Terence Penelhum, God and Skepticism, pp. 146ff.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

177

Reformed epistemology. Still, Plantinga is intent upon denying that Reformed


epistemology is a form of fideism. He says,
The Reformed epistemologist is not a fideist at all with respect to belief in God.
He does not hold that there is any conflict between faith and reason here, and he
does not even hold that we cannot attain this fundamental truth by reason; he
holds, instead, that it is among the deliverances of reason .
... On the Reformed view ... the deliverances of reason include the existence
of God just as much as perceptual beliefs, self-evident truths, memory beliefs,
and the like. It is not that theist and nontheist agree as to what reason delivers,
the theist then going on to accept the existence of God by faith; there is, instead,
disagreement in the first place as to what are the deliverances of reason.J13
There may be additional religious beliefs for which the theist must reply upon
faith, according to Plantinga, but not the belief that God exists. 114 However, if the
expansion of properly basic beliefs to include the belief that God exists depends
upon acceptance of the Reformed belief that God has endowed human beings with
a natural tendency to believe in him under certain conditions, then this belief in the
sensus divinitatis must come from somewhere. From where does it come? The
answer must be that such beliefs come from a particular community of religious
believers and that, furthermore, different communities of believers disagree about
these beliefs. This is why Stephen Wykstra calls religious beliefs "evidenceessential," since he claims that they are dependent upon evidence that is assumed
or believed to be available to the community of believers (although the evidence
need not necessarily be available to each individual member of the community).115
This is the view incorporated into William Alston's requirement of a system of
doxastic practices, discussed above, for assessing the claims of individuals to
religious experience; however, this is clearly not the view of the Reformed
platform. Without an appeal to some assumed evidence from a community of
believers or from religious instruction or from holy scripture to support such
beliefs and some epistemological defense of the doxastic practices of such a
community, the Reformed platform does indeed seem to rest ultimately and simply
upon faith. ll6
This conclusion is confirmed by Plantinga's latest work, Warranted Christian
Belief, the last of his trilogy of books addressing the epistemology of religious
beliefs. 117 I must treat Plantinga's argument in this book in a summary fashion
here, but his argument itself is somewhat a culmination and summary of his earlier
Alvin Plantinga, "Reason and Belief in God," p. 90.
For further comparison of Penelhum and Plantinga on this point and discussion of different forms
of the parity argument, see Richard Askew, "On Fideism and Alvin Plantinga," International Journal
for Philosophy of Religion, Vol. 23, 1988, pp. 3-16.
115 Stephen Wykstra, "Review of Faith and Rationality," Faith and Philosophy, Vol. 3, 1986, p. 209.
116 For further discussion of this point, see Robert Pargetter, "Experience, Proper Basicality, and
Belief in God," in Contemporary Perspectives on Religious Epistemology, edited by R. Douglas Geivett
and Brendan Sweetman (Oxford: Oxford University Press, 1992), pp. 150-67.
117 Alvin Plantinga, Warranted Christian Belief (Oxford: Oxford University Press, 2000). The earlier
volumes were Warrant and Proper Function (Oxford: Oxford University Press, 1993) and Warrant:
The Current Debate (Oxford: Oxford University Press, 1993).
113

114

178

ANALYTIC PHILOSOPHY OF REliGION

work. He focuses upon what he calls de jure objections to Christian belief, that is,
objections that Christian belief is "unjustifiable" or "rationally unjustified" or
"irrational." His main defense against such objections to the rationality of Christian
belief is to claim that they all rest upon, or assume at some point, that some de
facto objections, objections to the truth of Christian beliefs, are true. Plantinga
argues that there is no viable de jure objection that does not presuppose that
Christian beliefs are false. ll8 The most threatening de jure objection is that
Christian belief is unwarranted because it is false or very probably false. Therefore,
the main battle is to be waged over de facto objections, because if Christian belief
is false, then Plantinga agrees that it very probably has no warrant, but if Christian
belief is true, then, he claims, it very probably does have warrant. 119 While
Plantinga does not defend the truth of Christian belief, he thinks that establishing
its proper basicality is enough to ensure its warrant. He continues to rely upon a
properly functioning sensus divinitatis to produce theistic beliefs that are properly
basic and sufficiently warranted for knowledge. 120 The very low probability that
cognitive capacities arose completely naturalistically, without a divine origin and
plan, would not justify treating them as reliable. Thus, only if the origin of our
cognitive capacities is divine can we really trust them. Both the way in which we
make judgments about warrant and what we regard as properly basic are a part of
the "proper function" of our cognitive faculties, and the most probable account of
such proper function is one that rests upon divine origin and design. 12I As we have
seen, however, people obviously disagree about which beliefs are properly basic
and which ones are not. Thus, the cognitive faculties of human beings must
function properly at some times and improperly at others. Plantinga's whole case
thus depends upon identifying and explaining the proper function of a human
cognitive faculty that, in turn, produces properly basic beliefs. It is at this point that
Reformed epistemology turns into Reformed theology, because a properly
functioning cognitive faculty is explained in terms of the divinely orchestrated
sensus divinitatis that is not corrupted or damaged by sin. According to Plantinga,
an improperly functioning sensus divinitatis can be restored to its proper function
only by ''faith and the concomitant work of the Holy Spirit in one's heart.,,122 So, in
fact, Plantinga's Reformed epistemology ultimately rests upon faith and Reformed
theology and thus bears a close resemblance to earlier forms of fideism.
WITTGENSTEINIAN FIDEISM
In its most extreme form at least, fideism represents perhaps the most
exaggerated anti-evidentialist position for the theist. For example, in the case of
Sj:'Sren Kierkegaard, religious belief must be the result of faith that extends beyond
available evidence or rational argument. Compelling evidence or logically sound
arguments for the existence of God would make true religious faith impossible, so
the religious believer must make the "leap of faith" to believe in the face of the
lack of evidence or good reasons for doing so or even in the face of suspicions and
Alvin Plantinga, Warranted Christian Belief, pp. viii-x.
Ibid., pp. 186-88.
120 Ibid., p. 186.
121 I take this to be one of the main points of Warrant and Proper Function.
122 Alvin Plantinga, Warranted Christian Belief, p. 186. Also see Chapter 8.
118

119

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

179

doubts that would discourage one from believing.123 For the true believer, the
"proof' is an "inward proof' of faith.
In the second half of the twentieth century, fideism came to be identified with
the later work of Ludwig Wittgenstein, and thus the appellation "Wittgensteinian
Fideism" has come into common usage - a designation embraced by some
Wittgensteinians and rejected by others. Whereas Kierkegaardian fideism depends
upon the recognition of a conflict between religious faith and reason, "Wittgenstein
fideism" must be understood as a milder form of fideism where religious faith is
independent of reason - at least independent of some universal notion of reason
that is "outside" religious belief itself.124 It must also be said at the outset that
Wittgenstein himself did not explicitly develop a position of fideism, but then the
later Wittgenstein did not explicitly develop many positions that are now regarded
as Wittgensteinian. Thus, its has been left to his followers to construct the position
by extrapolating from various loosely connected Wittgensteinian themes regarding
language and language-games. 125 It must also be said that there is a rather wide
range of interpretations of Wittgenstein on this score by different Wittgensteinians.
It follows that there continues to be a fair amount of controversy about the different
interpretations of Wittgenstein and about just how much of what is taken as
constituting Wittgensteinian fideism can be appropriately attributed to Wittgenstein
himself.
Wittgenstenian fideism is developed from the following different theses based
upon different remarks made by Wittgenstein in his later writings: the purpose of
philosophy is simply to describe and understand language-games. "Philosophy may
in no way interfere with the actual use of language; it can in the end only describe
it .... It leaves everything as it is." Thus philosophy is not supposed to evaluate,
justify, or criticize language-games. This is true because "every sentence in our
language 'is in order as it is.,,,126 He emphasizes that "philosophy simply puts
everything before us, and neither explains nor deduces anything.,,]27 Thus, the most
fundamental and constitutive elements of religious belief and the religious form of
life cannot be justified and cannot reasonably be expected to be justified. What
philosophy puts before us and describes is the grammar of words - the pattern of
usage that words have within language-games; thus, the description of the grammar
of a word is a description of the way in which it is used in a language-game.
Whatever explanation there is occurs within a language-game and may amount to
something significantly different from language-game to language-game,
depending upon the language-game within which explanation takes place. There is
\23 There are different interpretations of Kierkegaard on this point. See, for example, C. Stephen
Evans, "Kierkegaard and Plantinga on Belief in God: Subjectivity as the Ground of Properly Basic
Religious Beliefs," Faith and Philosophy, Vol. 5, no. 1, 1988, pp. 25-39, and ''The Epistemological
Significance of Transformative Religious Experiences: A Kierkegaardian Exploration," Faith and
Philosophy, Vol. 8, no. 2, 1991, pp. 180-92; and Marilyn Gaye Piety, "Kierkegaard on Rationality,"
Faith and Philosophy, Vol. 10, no. 3, 1993, pp. 365-79.
\24 For a discussion of different forms of fideism, see Terence Penelhum, "Fideism," in A
Companion to Philosophy of Religion, edited by Philip L. Quinn and Charles Taliaferro (Oxford:
Blackwell Publishers, 1997), pp. 376-82.
125 Discussion ofWittgenstein's notion oflanguage-games can also be found in Chapter II.
126 Ludwig Wittgenstein, Philosophical Investigations, Sections 98 and 124.
127 Ibid., Section 126.

180

ANALYTIC PHILOSOPHY OF RELIGION

no explanation of language-games themselves or the forms of life to which they


give rise. All explanation comes to an end, and wherever that end may be, the end
comes within and is determined by the limits of the language-game. 128 There is no
single or universal Archimedean point from which different languages-games can
be understood or evaluated since different languages games and the forms of life
that they constitute can only be understood and evaluated from within the
particular language-games. 129
THE GROUNDLESSNESS OF BELIEF
Two of the earliest followers of Wittgenstein to begin to piece these different
themes together into a thesis regarding religious belief were Peter Winch and
Norman Ma1colm.130 In the context of a reply to Kant's objection to the ontological
argument on the grounds that existence is not a predicate (see Chapter III),
Malcolm develops his form of language-game fideism by claiming that it is only
the religious person who has suffered a "storm in the soul" and experienced the
resulting "guilt that is beyond all measure" (presumably by having sinned) and
then enjoyed the "forgiveness that is beyond all measure" who can really
understand the Christian concept of God. From such experiences comes the
concept of "a forgiveness that is beyond all measure." Thus, one must understand
the concept of God from "inside" the language-game and not from "outside" it. 131
Nearly two decades later, Malcolm developed his language-game perspective on
religious beliefs further by relying more heavily upon Wittgenstein's On Certainty
after this volume was published. There are many fundamental beliefs, Wittgenstein
maintained, for which we do not have adequate evidence but which we accept
unconditionally in spite of their "groundlessness." Malcolm considers the common
belief that physical objects exist and continue to exist when no one is experiencing
them. When a physical object "goes missing," such as when a person cannot find
his car keys, we never assume that the keys just vanished, that is, suddenly just
ceased to exist. We assume that what Wittgenstein calls "the principle of the
128 Norman Malcolm develops this point regarding the limit of explanation more fully in
Wittgenstein: A Religious Point of View? (Ithaca, N.Y.: Cornell University Press, 1994), Chapter 6.
Also see Peter Winch's comments in the same volume in his discussion of Malcolm's essay, pp. 103ff.
Winch makes the important point that it was the particular kind of explanation that depends upon a
theory of meaning that Wittgenstein opposed (according to which the meaning of a word is given in
terms of necessary and sufficient conditions, that is, essentialism). The important thing is not simply
that explanation comes to an end but that confusions about the nature of language and meaning make
philosophers ignore or resist this fact.
129 For a more complete listing of additional tenets of Wittgensteinian fideism, see Kai Nielsen,
"Wittgensteinian Fideism," Philosophy, Vol. 42, no. 161, 1967, pp. 192-93.
130 Peter Winch, The Idea of a Social Science and Its Relation to Philosophy (London: Routledge
and Kegan Paul,1958), and "Understanding a Primitive Society," American Philosophical Quarterly,
Vol. 1, 1964, pp. 307-25. Page numbers refer to the reprint in Rationality, edited by Bryan R. Wilson
(Oxford: Basil Blackwell, 1970). Norman Malcolm, "Anselm's Ontological Arguments," Philosophical
Review, Vol. 69, no. 1, January 1960 (page numbers refer to the reprint in Malcolm's Knowledge and
Certainty [Englewood, Cliffs, N.J.: Prentice-Hall, 1%3]), and 'The Groundlessness of Belief," in
Reason and Religion, edited by Stuart Brown (Ithaca, N.Y.: Cornell University Press, 1977), which is
reprinted in Contemporary Perspectives on Religious Epistemology, edited by R. Douglas Geivett and
Brendan Sweetman (Oxford: Oxford University Press, 1992), pp. 92-103, to which the page numbers
here refer.
131 Malcolm, "Anselm's Ontological Arguments," p. 160.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

181

continuity of nature" holds, but what of some different society that does not make
this assumption? If people in a different society believe that items literally vanish,
cease to exist, when they "go missing" and then are instantly reconstituted when
they are found, do the people in this society have less evidence for their belief than
we do in our belief that such things do not happen? 132 The point of this story is that
inquiry that depends upon belief, reason, evidence, arguments, or justification
occurs with a "system" that provides the fundamental groundless beliefs according
to which inquiry takes place. "To be sure there is justification; but justification
comes to an end,,,133 Wittgenstein says. If we think of justification as a kind of
testing, then "whenever we test anything, we are already presupposing something
that is not tested." So, he asks rhetorically, "Doesn't testing come to an end?,,134 To
even question things, to have doubts and then attempt to resolve those doubts
through testing, requires that we believe that "some propositions are exempt from
doubt, are as it were the hinges on which those [doubts] turn.,,135
Asking for or expecting justification of everything, that is, in this context,
asking for or expecting justification of language-games, is part of what
Wittgenstein called the "general disease of thinking" and what Malcolm calls "the
pathology of philosophy.,,136 No rule, signpost, or directive can guide us
completely in our actions. There must come a point where the guidance and the
reasons and the justifications for our actions come to an end, and then we just go
on. In WiUgenstein's famed example, a signpost cannot tell us how to interpret the
signpost. Malcolm's conclusion from this is that "grounds come to an end.
Answers to How-do-we-know? questions come to an end. We must speak, act, live,
without evidence. This is so not just on the fringes of life and language, but at the
center of most regularized activities.,,13?
The claim that justification and explanation come to an end at some point is not
a particularly radical claim and has been commonly recognized throughout the
history of epistemology. However, Malcolm develops this claim into a much
stronger one. We cannot and should not, according to Malcolm, be expected to
justify our language-games since justification is a process that takes place within
language-games that are themselves unjustified or self-justified. Malcolm
summarizes his point very succinctly:
Within a language-game there is justification and lack of justification, evidence
and proof, mistakes and groundless opinions, good and bad reasoning, correct
measurements and incorrect ones. One cannot properly apply these terms to a
language-game itself. It may, however, be said to be "groundless" not in the
sense of a groundless opinion, but in the sense that we act it, we live it. We can
say, "This is what we do. This is how we are."J38

See Malcolm, "The Groundlessness of Belief," pp. 92-94.


Ludwig Wittgenstein, On Certainty (New York: Harper and Row, 1969), 192.
134 Ibid., 163 and 164.
135 Ibid., 341.
136 Malcolm, "The Groundlessness of Belief," p. 96-98.
137 Wittgenstein, Philosophical Investigations, Section 292, and Malcolm, ibid., p. 97.
138 Malcolm, "The Groundlessness of Belief," p. 98. Also see Norman Malcolm, Wittgenstein: A
Religious Point of View (Ithaca, N.Y.: Cornell University Press, 1994), Chapter 6.
132

133

182

ANALYTIC PHILOSOPHY OF RELIGION

This much stronger view of language-games embodies the more controversial


claims that language-games are autonomous in the sense that the criteria for
justification, evidence, reality, meaningfulness, rationality, and truth are all relative
to and internal to the individual language-games. 139 I will address these issues
momentarily, but first, I will examine the similar kind of positions developed by
Peter Winch and D. Z. Phillips.
RELIGION AND UNDERSTANDING A PRIMITIVE SOCIETY
The Idea of a Social Science and Its Relation to Philosophy, by Peter Winch,
has proven to be a work of major importance for several different reasons. 140
Although Winch's main focus in this short book is the nature of the social sciences
and the problem of understanding a society that is different from one's own, Winch
develops an interpretation of Wittgenstein's language-games that has major
implications not only for religious belief but, more generally, for epistemology as
well. Winch argues that the social sciences and the social institutions and practices
that are the objects of study in the social sciences are fundamentally different from
the natural sciences and the natural world. Social institutions are conventional,
rule-governed phenomena that must be grounded in and understood in terms of a
social context. Claims about human social life must be understood as relative to
and "filtered" through a subjective, social perspective - including religious belief
as well as such fundamental notions as human rationality. Winch argues his point
using the much-discussed example of the Africian Azande's use of magic, first
introduced by the research of E. E. Evans-Pritchard. 141 The Azande engage in a
common practice that Evans-Pritchard and Winch identify as a practice of
witchcraft - the belief that they can control the course of natural events by relying
upon religious oracles and various "magical" means. According to Winch, the
Azande rely upon the oracles and magic to regulate their everyday lives in much
the same way as those in modern, Western societies rely upon clocks or
mathematical calculations to regulate their lives. 142 The Azande thus have a view
of things that is apparently at serious odds with the modern scientific view of
things. When the question is raised about the intelligibility of the Azande' s view of
139 See Patrick Sherry. "Is Religion a 'POnTI of Life'?" American Philosophical Quarterly, Vol. 9,
no. 2, 1972, especially pp. 164ff. Sherry distinguishes what he calls the relativist/conventionalist and the
pragmatic/empiricist interpretations of Wittgenstein. Winch and Phillips fall into the more radical
relativist!conventionalist group. Further criticism of the relativist/conventionalist interpretation of
Wittgenstein and defense of universal epistemological notions such as truth and reason are found in
Patrick Sherry, Religion, Truth and Language Games (London: Macmillan, 1977), and William P.
Alston, "The Christian Language-Game," in The Autonomy of Religious Belief (Notre Dame, Ind.:
University of Notre Dame Press, 1981), pp. 128-62. As I discuss below, the degree of autonomy
enjoyed by language-games and the nature of the relationships among different language-games are
complex and difficult issues. For further discussion of Malcolm's claims of the groundlessness of
religious belief, see the exchange from the 1975 symposium among Malcolm, Colin Lyas, and Basil
Mitchell in Reason and Religion, edited by Stuart C. Brown (Ithaca, N.Y.: Cornell University Press,
1977).
140 Peter Winch, The Idea of a Social Science and Its Relation to Philosophy. The same ideas are
further developed in his "Understanding a Primitive Society," pp. 78-111.
141 E. E. Evans-Pritchard, Witchcraft, Oracles and Magic arrwng the Azande (Oxford: Oxford
University Press, 1937).
142 Peter Winch, "Understanding a Primitive Society," pp. 10-11.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

183

the world, Winch insists that we must specify the context and the framework in
which the question is being raised. Their view of the world makes perfectly good
sense to the Azande, and we must not characterize science as dealing with
"objective reality" or attribute a status to modem science so that it is understood as
"a paradigm against which to measure the intellectual respectability of other modes
of discourse.,,143 According to Winch, the notions of intelligibility and rationality
must be re1ativized to the language-game in which they occur. Thus, so far as
rationality is concerned, there are "our standards and theirs," and, in any discussion
of rationality, we must specify "whose concept of rationality is being alluded to"
since something is rational to a person "only in terms of his understanding of what
is and is not rational." 144
For Winch, what is true of the concept of rationality is true a fortiori of the
concept of God:
God's reality is certainly independent of what any [particular] man may care to
think, but what that reality amounts to can only be seen from the religious
tradition in which the concept of God is used .... The point is that it is within the
religious use of language that the conception of God's reality has its place. 145
Thus, questions concerning the intelligibility of the Azande's use of magic or
the intelligibility of theists' understanding of the reality of God must be addressed
within the context created by the respective language-games. To be fair to Winch,
it should be noted that he does not say that the reality of God is determined by the
language-game, but what is real is reflected in the language-game. One's
understanding of what is real as well as the difference between what is real and
what is not real are all functions of the language_game. 146 Reality as well as logic
are results of the way in which words are used in the language-game, the grammar
of the language-game, and thus are best understood as social constructs. 147 Winch
arrives at a position based upon his interpretation of Wittgenstein's use of
language-games where language-games are seen as logically and conceptually
autonomous.
Like Wittgenstein, Winch did not develop his positions into a full-blown theory
about religious belief, but others who have come after Winch have taken his claims
regarding the Azande and rationality as the fundamental insight upon which a
Wittgensteinian view of religion is to be constructed. Has Winch made his case
that there are different standards of rationality relative to different languagegames? Well, to submit to an irresistible pun, it just all depends upon what is
meant by "rationality." First of all, if beliefs are understood as the things upon
which we predicate our actions, then, on one level - from a modem, scientific
perspective - one would say that to rely upon oracles to predict the weather instead
of a meteorologist and weather satellites is "irrational." This sense of rationality is
one that appears to be obviously transcultural and arguable across language-games.
Ibid., pp. 80 and 81.
Ibid., p. 97.
145 Ibid., pp. 81-82.
146 Ibid., p. 82.
147 Compare Winch, The Idea of a Social Science, p. 100.
143

144

184

ANALYTIC PHILOSOPHY OF REUGION

The deeper sense of irrational that Winch needs for his case is one in which,
according to him, the Azande tolerate contradictions in their beliefs in a way that is
inconsistent with a Western, European view. How might this happen if beliefs are
understood as what are necessary for directing rational, intentional behavior? An
Azande A wants to bring about a certain desired result k. A is given an oracle, and
the oracle says to A, "Do j in order to bring about k." A must then reason
something like the following: "I want k to happen. The oracle says in order to bring
about k, I must do j. I believe the oracle, therefore, I will do j." But this looks like
good, old-fashioned, Western European logic at work. If A reasoned, "I want k to
happen. The oracle says in order to bring about k, I must do j. I believe the oracle,
therefore, I will not do j but h instead," then that would be irrational, that is, bad
logic, but it would arguably be irrational for the Azande as well as Western
Europeans. 148 These difficulties with Winch's extrapolation of the grammar of
language-games to logic and rationality notwithstanding, his interpretation of
Wittgenstein that results in the conceptual and logical autonomy of languagegames has opened the door for those who defend a fideistic interpretation of
Wittgenstein.
THE GRAMMAR OF RELIGIOUS LANGUAGE
The division between conservatives and liberals used in Chapter II to structure
different responses to the challenge to the meaningfulness of religious language
raised by the verification criterion of meaning does not allow for a very
satisfactory way of accommodating those who would wish to deny that very way
of categorizing responses to the challenge. While most of those who give what I
have labeled "liberal" responses still worry about trying to secure a metaphysical
referent for God through the elaborate use of metaphorical or analogical language,
Wittgensteinians insist that the disputes between realists and nonrealists and
between cognitivists and noncognitivists are to be dissolved as illegitimate disputes
and as nonissues. Thus D. Z. Phillips (and, to a lesser extent, R. B. Braithwaite and
Paul van Buren, discussed in Chapter II) do not fit comfortably into the scheme
used there. 149 As Michael Coughlan notes,150 Wittgenstein himself had
comparatively little to say about religious belief, and, we might add, even less to
say about religious language. To develop a Wittgensteinian approach to religious
language, one must proceed based upon what Wittgenstein said about language and
language-games in general and upon what he said about religious belief and then
extrapolate to religious language. This is what Phillips has done, and he is most
responsible for developing a detailed application of Wittgenstein's notion of
language-games to religious language.
Along with other features that distinguish them, language-games are
distinguished by the fact that each has its own "grammar." As noted already, the
148 I have argued this point more fully in James F. Harris, Against Relativism (LaSalle, TIL: Open
Court, 1992), pp. 1OOff.
149 Phillips distinguishes his view from that of Braithwaite and calls Braithwaite's views "confused."
See his Religion without Explanation (Oxford: Basil Blackwell, 1976), pp. 140ff.
150 Michael J. Coughlan, "Wittgenstein, Language, and Religious Belief," in God in Language,
edited by Robert P. Schar1emann and Gilbert E. M. Ogutu (New York: Paragon House Publishers,
1987), p. 149.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

185

uses of language that occur within a language-game are governed by the


conventions and rules within that language-game. These rules and conventions
determine the grammar of the language by determining the correct and incorrect
uses of words that can occur within the language-game. Each language-game is
autonomous in determining its own grammar, which means that the grammar of
religious language is unique, that is, it is used in ways that are determined simply
by the conventions of the language-game.
It is possible to synthesize a picture of Phillips's general view of the nature of
religious language by drawing upon a number of his writings. 151 To understand the
grammar of religious language, we must understand the nature of religious beliefs.
To understand religious beliefs, we must look to the language-game of religious
beliefs, where we find that the unique grammar of the language-game controls the
uses of language within the language-game. 152 Following Wittgenstein, Phillips
insists that it is not the purpose of the philosopher (or for present purposes, we may
say the philosopher of religion) to develop technical or new systems of language
for capturing or expressing religious beliefs and practices. Philosophers should not
try and impose preconceived or armchair notions of what language must be like or
should be like in order to be used in certain ways. We simply "leave everything as
it is" and look to see how the language is used in the language-game. 153
According to Phillips, when we look at the grammar of the language that is used
by a religious believer when that person affirms the existence of God or says that
God loves us, we find that the believer is not really advancing a factual
hypothesis. 154 Such expressions are ways of conveying a commitment to the way in
which one lives one's life or the way in which one intends to live one's life. For
Phillips, when the religious believer says that God exists, this does not amount to
saying that some additional entity exists in the universe. 15S Rather, it is a claim

151 For example, see D. Z. Phillips, The Concept of Prayer (Oxford: Basil Blackwell, 1965); Religion
without Explanation (Oxford: Basil Blackwell, 1976); and Faith and Philosophical Inquiry (London:
Routledge and Kegan Paul, 1970). Also see "Religious Beliefs and Language-Games," in The
Philosophy of Wittgenstein: Aesthetics, Ethics, and Religion, edited by John Canfield (New York:
Garland, 1986), and in Faith and Philosophical Inquiry as Chapter V.
152 Throughout this discussion of language-game fideism, I use the expressions "the religious
language-game" and "the theistic language-game." Determining the boundaries and varieties of
language-games is not an issue I can address here. W. D. Hudson uses the expression "the theistic
language-game" and calls Christianity a language-game in Ludwig Wittgenstein: The Bearing of His
Philosophy upon Religious Belief (Richmond, Va.: John Knox Press, 1968), pp. 69 and 70. As I have
noted above, Malcolm describes religion and science as two different language-games in "The
Groundlessness of Belief," ibid., p. 100, and specifically includes nontheistic Buddhism as part of the
religious language-game, ibid., p. 101. Phillips speaks of "religious language-games" (in the plural). For
further discussion of this point, see R. H. Bell, "Wittgenstein and Descriptive Theology," Religious
Studies, Oct. 1969, and Patrick Sherry, "Is Religion a Form of Life," American Philosophical
Quarterly, Vol. 9, no. 2, 1972. There is obviously not a single language-game of religion or a single
language-game of theism and, hence, no single granunar of religious language. It may be helpful to
think of a religion or theism as aggregates of different language-games. Religion and science would
then differ not by each being a different discrete language-game but by each being different aggregates
of different language-games, some of which may be common to the two different aggregates
153 D. Z. Phillips, Religion without Explanation, p. 41.
154 Ibid., pp. 102ff.
155 However, as Phillips has indicated in correspondence, ''no additional 'entity'" does not
necessarily mean "no addition at all." Certainly, when the theist affirms God's existence, he is affirming

186

ANALYTIC PHILOSOPHY OF REUGION

about the way in which things in the universe exist. This is the source of
Wittgenstein's famed comment that when a religious person says that God exists
and an atheist says that God does not exist, they are not really contradicting each
other. 156 Two people can contradict each other, Wittgenstein claims, only if they
speak the same language and share the same logic and understanding. By this,
Wittgenstein means that two people can contradict each other only if they share a
common understanding of the criteria by which their disagreement may be
resolved. To use an example from Phillips, if one person says that handling the ball
is a foul and another person says that handling the ball is not a foul, do they
contradict one another?157 The answer is, of course, "it just all depends." If one
person happens to be talking about European football and the other about
American baseball, then they are talking "past one another," that is, what each says
seems to be completely umelated to what the other says. To contradict one another,
the two need to be within the context of the same game where the same rules
apply. The framework of the language-game and its conventions provide the
grammar of the language-game, within which agreement or disagreement becomes
possible. In the case of the existence of God, Phillips explains, the logic of the
word 'God' is unique since we ought not to take 'God' to be a proper name of an
individual. In the language-game of religious belief that Phillips plays, God is not
simply one more being among a whole universe of other beings, all of which can
be judged to exist or to not exist according to the same criteria. 158 God is the kind
of thing that, within the language-game, it is impossible to think of as not existing;
however, this is not a claim that can be generalized to the religious language-game
since, as we have seen, there is no single religious language-game.
Such a Wittgensteinian view of the nature of religious language also means that
the whole notion of verification, the discussion of which has occupied much of
Chapter II, is misplaced in the case of religious language. Wittgenstein might say
that those who suggest verification or falsifiability as a test for the meaningfulness
of religious language are using the criterion from one language-game (science) and
trying to impose it upon another language-game (religion) where it does not belong
and where it does not make sense. Michael Coughlan suggests an analogy to
explain how two different people may view the world differently as a result of
different language-games. 159 We may think of the different ways in which two
different people may view the universe and suggest that instead of seeing the
cosmos in black and white, the religious believer sees it in color. In this analysis,
the universe is the same for the believer and nonbeliever alike in an ontological
sense, but religious belief adds a certain dimension or quality to the way in which
the religious believer regards the universe and his place in it that gives both the
universe and the person's place in it a greater depth and meaning.

something "additional" about the world. Compare what Wittgenstein said about pain: it is not a
"something," but it is not a "nothing" either.
156 See Ludwig Wittgenstein, Lectures and Conversations on Aesthetics, Psychology and Religious
Belief, edited by Cyril Barrett (Oxford: Basil Blackwell, 1966), pp. 53-72.
157 D. Z. Phillips, "Religious Beliefs and Language-Games," p. 223.
158 See Phillips, ibid., and The Concept of Prayer, Chapter I.
159 Michael Coughlan, "Wittgenstein, Language, and Religious Beliefs," p. 159.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

187

To understand a religious utterance, according to Phillips, we must look at its


place in the language-game and at its relationships with the other utterances made
by and the practices engaged in by the people who "play" that language-game. l60
We will find no explanation or justification for why people say what they do or
why the grammar of the language is what it is because "it makes no sense to ask
such questions," Phillips says.161 In a particular language- game, the grammar is
what it is, language is used the way in which it is used, and people talk the way that
they talk - "the language-game is played" simply because that is the way things
are. To paraphrase Wittgenstein and Harry Truman, justification must stop
somewhere, and it stops here.
IS RELIGIOUS LANGUAGE "JUST" A LANGUAGE-GAME?
There have been many objections to Wittgensteinian fideism and the view of
religious language developed by Phillips,162 and I will not attempt to address all or
even a large number of those objections here. I will however focus upon the major
difficulties that have been raised and the major issues generated by the treatment of
religious language as a language-game. A major issue for Phillips (as well as for
Braithwaite and van Buren, discussed in Chapter II) involves how his philosophical
account of religious language is related to how religious people actually use
religious language and what religious people actually intend to be doing when they
use religious language. According to Wittgenstein, we are supposed to look at the
language-game and leave everything as it is, but these Wittgensteinian approaches
to religious language have all the appearances of a revisionist account of religious
language. When Wittgenstein says that we are to look at the language-game, he
means that we are to look at the grammar of the language-game, the underlying
conceptual mapwork of the language-game that makes it possible to say the things
that are said and to do the things that are done in the language-game. The grammar
itself is independent of what the people who use the language-game actually say
and do, and the grammar cannot be examined empirically by watching people
using a language. As John Whittaker has observed, a Wittgensteinian examination
of the grammar of a language-game has nothing to do with popUlarity surveys or
the common understanding of the man in the street. 163 If, for example, some people
think that human rights are quasi-physical ingredients in the constitution of a
human being, their believing it does not make it so. They would be just wrong. The
"grammar" of human rights requires something quite different. The correct
grammar of a language-game is "laid down" (established) in tradition and
representative and authoritative literature and oral history. It is this underlying,
160

See D. Z. Phillips, "Religious Beliefs and Language-Games," p. 111.

161

Ibid., p. 119.

162 See, for example, R. W. Hepburn, "From World to God," Mind, Vol. 72, 1963; Kai Nielson,
"Wittgensteinian Fideism," Philosophy, July 1967, reprinted in John Canfield, The Philosophy of
Wittgenstein (New York: Garland, 1986), and Kai Nielsen, "A Critique of Wittgensteinian Fideism," in
The Autonomy of Religious Belief A Critical Inquiry, edited by Frederick Crosson (Notre Dame, Ind.:
University of Notre Dame Press, 1981). This last piece also appears in Philosophy of Religion: Selected
Readings, Third Edition, edited by William L. Rowe and William J. Wainwright (Fort Worth, Tex.:
Harcourt Brace College Publishers, 1998).
163 In correspondence with the author. What this position ignores is that claims about the grammar of
a language-game must rely upon someone's intuitions.

188

ANALYTIC PHILOSOPHY OF REUGION

"deep" grammar that is supposed to be "left as it is" and not the surface, popular
usage. The "surface grammar" can even easily obscure the "depth grammar"; thus,
the philosophical account of religious language may disagree with the reports of
ordinary theists. However, the distinction between surface and depth grammar
notwithstanding, a philosophical account of religious language ought to map onto
common religious practices. Claims about depth grammar seem to rely upon
something like conceptual intuitions, and conflicting claims about depth grammar
cannot all be equally compatible with the tradition and the authoritative literature
and oral history of a particular language-game.
In order for the philosopher to "leave things as they are" when he examines the
depth grammar of religious language, he must make sure that he is not imposing a
revisionist interpretation upon the conceptual map work. How does one insure this?
In one manner or another, Braithwaite, van Buren, and Phillips all talk about what
religious believers are really doing when they use religious language, but it is not
clear how they can guarantee that their accounts are not the result of imposing their
own interpretations and their own intuitions upon what the religious person is
doing.
WHO CONTRADICTS WHOM?
Consider the dispute, for example, between the creationists and the evolutionists
(discussed in Chapter V). Do their different accounts of the origin of life contradict
one another? Well, in good Wittgensteinian fashion, the answer may be "it just all
depends." A good Wittgensteinian may insist that in such cases, contradictions are
not there waiting to be discovered; a contradiction occurs "when it arises." Peter
Winch describes in some detail one possible scenario where we may analyze the
possible conflict between the language-game based in Genesis and the one based in
Darwin's Origin of Species. IM Winch asks us to imagine a devout religious man
before the rise of Darwinian evolutionary theory who was steeped in the story of
creation as told in Genesis with all of the associated religious beliefs and practices.
Such a person would simply have a very different approach to any question
concerning the origin of life from "a scientifically minded" person steeped in the
scientific method and Darwinian evolutionary theory. The "natural order" that
Darwin both assumed and described was understandable only within a context that
depended upon the scientific method and scientific practices. 165 So, in such a
situation, we may agree that the two were "talking past" one another and did not
really contradict one another. But what of the current dispute between creationists
and evolutionists? Imagine a different situation, at the beginning of the twenty-first
century, in which the devote religious believer is steeped not only in the creation
story as told in Genesis but in the scientific method as well. This person accepts
science's accounting of the natural order of things and the scientific method as
providing explanations for most natural phenomena. However, when it come to
providing an account of the origin of the human species, the religious person says
that he believes that God literally created Adam and Eve as described in Genesis
and that human beings did not evolve. It certainly seems that in this case the
164
165

Peter Winch, Trying to Make Sense (Oxford: Basil Blackwell, 1987), pp. 134ff.
Ibid., pp. 135-36.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

189

modern creationist is disagreeing with the evolutionists about facts concerning the
natural world. Suppose now that such a creationist is told that he is simply playing
a language-game with a particular grammar, according to which his claim amounts
to a commitment to a certain way of life or a poetic way of saying that he has an
optimistic view of life and of the place of human beings in the universe. At the
same time, he is told that science is the language-game that tells us how things
actually occur in the natural world, including the origin of human life. It certainly
sounds as if such a treatment amounts to forcing an interpretation and a particular
understanding of the nature of the grammar of the language on the religious
believer just as much as the verificationists have been accused of forcing a
particular interpretation upon religious language. And it is very plausible to
imagine the modern creationist responding, "No, I really mean that God created
human beings in our present form and that we did not evolve. Evolution is just
wrong!"
Phillips's claim that in the language-game of theism the claim "God exists" does
not amount to affirming the existence of an additional entity in the universe is
obviously at odds with the usual interpretations of the tradition and the
authoritative literature. Trying to restore something of a realist account to religious
belief, William Alston complains that, in Phillips's account, the things that a
particular group of theists, such as Christians, say about God are simply about the
attitudes and beliefs of those playing the religious language-game. However,
certainly many Christians intend to be making "realist" claims about the existence
of God and God's nature, that is, they intend to claim that God and God's nature
"are as they are independently of us and our religious activities, our religious
language-games, our religious form of life, or anything else having to do with the
ways we think, conceptually structure experience of the world, or react to
things.,,166 Phillips's language-game analysis of religious belief, whether it is called
"fideism" or not, does contrast starkly with more traditional theistic accounts, and
Alston appeals to a realist account of truth to draw the distinction between the two
as sharply as possible. 167 However, Phillips might respond here that the "facts" of a
realist's account are themselves suspect, open to interpretation, and simply another
function of the grammar of the language-game.
With his account of religious language-games, Phillips is perilously close to
slipping into a de facto atheism. According to his understanding of the grammar of
"God exists," this is not a claim about "the existence of an object," and neither are
questions about possible different meanings of the word 'God' questions about
whether different uses refer to the same individual. 168 He refers to Wittgenstein's
claim in On Certainty that the existence of God "plays an entirely different role to

166 William P. Alston, "Taking the Curse Off Language-Games: A Realist Account of Doxastic
Practices," in Philosophy and the Grammar of Religious Belief, edited by Timothy Tessin and Mario
von der Ruhr (London: St. Martin's Press, 1995), pp. 24-25.
167 Ibid., pp. 30ff. For discussion of the differences between Phillips and Alston. see M. Jamie
Ferreira, "Universal Criteria and the Autonomy of Religious Belief," International Journal for
Philosophy of Religion. Vol. 15, 1984. pp. 3-12.
168 D. Z. Phillips, Religion without Explanation, p.174. Phillips explicitly anticipates the charge of
defending a "disguised atheism" on pp. 149ff.

190

ANALYTIC PHILOSOPHY OF RELIGION

that of the existence of any person or object I ever heard of.,,169 But theologians
throughout the ages have maintained that God's existence is entirely different from
the existence of persons and other objects, though they have not denied that God is
a being of some sort.
Denying that religious believers use 'God' to refer to God or use 'God exists' to
affirm the existence of God makes it difficult to make sense of Phillips's
description of the behavior of religious people. He says, for example, that during
worship, "the praising and glorifying does not refer to some object called God.
Rather, the expression of such praise and glory is what we call the worship of
God."170 Presumably, the grammar of religious belief accounts for the unique ways
in which words and phrases from common usage are used with the religious
language-games. But, according to the Oxford English Dictionary, the normal uses
of 'praise' and 'glory' as verbs are transitive uses that must take objects - unlike,
for example, 'rejoice.' Praising and glorifying are actions directed at some object
whereas rejoicing is an intransitive action. One can rejoice without rejoicing
anything, but one cannot praise or glorify without praising or glorifying something.
To praise God is thus "to attach value to," "to esteem," "to laud," or "to extol the
glorious attributes of' God. To glorify God is "to invest with glory," "to extol,
honor, magnify with praise," or "to advance the glory of' God. Is the object taken
by these verbs simply a "surface," grammatical object? Then how would one
understand the actions of praising and glorifying if there is no understood object to
which these actions are directed? If one's understanding of the grammar of
religious belief permits such an aberrant use of 'praise' and 'glorify,' then why use
these words at all? If 'praise' and 'glorify' do not retain something of their normal
meaning, use, and grammar - including taking an object - then, it seems, we could
easily replace 'praise' and 'glorify' with any other words, such as 'abhor' and
'loathe.' Then 'I abhor God' could just be what we mean by 'the worship of God,'
but clearly this locution is seriously infelicitous.
THE RIGHT PLACE TO STOP?
Many of the criticisms raised against Phillips's version of language-game
analysis of religious language revolve around the fundamental issue of how far to
go and where to stop in our analysis, interpretation, and explanation of religious
language and who is to conduct the analysis, interpretation, and explanation. On
Wittgensteinian grounds, the confusion resulting from "the bewitchment" of
language means that we cannot simply point to "the facts" in our analysis since
"the facts" are part of the confusion. For Wittgenstein, determining just how far
one goes and where one stops in the analysis of language-games is a fundamental
issue in combating the confusion since there is a natural or philosophic urge to go
beyond the proper stopping point. Overanalyzing just adds to the confusion. All
justification and explanation must come to an end, and language-games are
"complete" in such a manner that philosophical analysis is not going to improve
them in some way, nor is it appropriate to attempt to do so. Some critics have
169 Wittgenstein, Lectures and Conversations on Aesthetics, Psychology and Religious Belief, p. 59.
Quoted by Phillips in Religion without Explanation, pp. 174-75.
170 Phillips, ibid., p. 149.

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

191

complained that when this "full stop" to the analysis of language-games is invoked
for religious language, an artificial limitation is placed on philosophical criticism
and debate. l7l Phillips's answer to this criticism is that language-games are not
immune to criticism, but the kind of criticism that is possible and important does
not rely upon detailed linguistic distinctions between language-games and practices
or rituals or modes of discourse. Rather, legitimate criticism takes the form of
examining how the language-game operates in human life. 172
Who is in the best position to conduct such an examination of how a particular
language-game operates in human life? Phillips claims that it is part of the logic of
'God' within ordinary theistic belief that it is impossible or inconceivable for God
not to exist and that the grammar of 'God' within the language-game makes it
meaningless to even suggest that God does not exist. This example serves to focus
attention on the difficulty of agreeing on who is to evaluate the grammar of 'God'
and how the evaluation is to take place. Surely not all practicing theists would
recognize this as a description of how the grammar of 'God' actually operates.
Theists regularly endure "trials of faith" (that is, circumstances that make a
believer question and doubt God's existence or God's goodness, such as the death
of a child). The agonizing doubt that frequently accompanies such experiences
suggests that not all believers would agree with Phillips's description of the
grammar of 'God.' Undoubtedly, many theists will insist that when they say, "God
exists," they do mean to affirm the existence of an individual, personal being.
When they weaken in their faith, they doubt the existence of such a being. 173 How
do we tell if such a reaction is confused or a result of the proper grammar of
'God'? The more general issue is how we are to resolve differences of opinion
about the grammar of 'God.'
Wittgenstein seems to suggest that the participants themselves are in the best
position to understand their language-game. However, as I have shown, there is
obvious disagreement even among the participants; of what would such an
understanding consist? Such understanding must consist, it seems, of the ability to
use the different words, expressions, and sentences properly according to the
grammar of the language-game - the ability to "carryon" - to make the next
"move" in the game. Can one have such an understanding and not be a practicing
participant? Former participants come to mind easily as a group that may be said to
understand the language-game as well as current participants. If former participants
can understand a language-game, what about astute observers? Certainly, there
may be some differences on the experiential level between the way in which a
participant and a nonparticipant experience the grammar of a language-game, but
does this mean that there must be a difference in the understanding of the
171 For a full discussion of such criticisms and Phillips's reply, see his Belief, Change and Forms of
Life (Atlantic Highlands, N.J.: Humanities Press International, 1986), Chapter 2.
l72lbid., p. 33. See pp. 26-33 for Phillips's complete reply. Also see D. Z. Phillips, "Belief, Change
and Forms of Life: The Confusions of Externalism and Internalism" in The Autonomy of Religious
Belief, edited by Frederick J. Crossen (Notre Dame, Ind.: University of Notre Dame Press, 1981), for
detailed textual references for his views regarding the possibility of criticizing religious beliefs,
especially pp. 85-90.
173 While Phillips admits that many theists regard their claims as true, the kind of truth the theist is
interested in is the kind, he insists, that means the claims are ones that the theist can "live by" and
"digest."

192

ANALYTIC PHILOSOPHY OF REUGION

grammar? Given the obvious disagreements concerning conceptual intuitions about


the grammar of religious language, a language-game-based approach to religious
language must provide a better mechanism for determining when the confusion has
been unraveled and when we are on the path from error to truth.
The final point here concerns the nature of language-games themselves. It is
important to keep in mind that one should not think of any religious languagegame as completely separate and autonomous. Here, the analogy of a game is
misleading. The notion of a single, separate language-game hardly even makes
sense for Wittgenstein since the notion of language-game is itself the result of a
family resemblance. Language-games may be distinct from one another in some
ways but similar in other ways, and every person is a participant in a great number
of different language-games that intersect, overlap, and crisscross to constitute a
form of life. A form of life is not like a jigsaw puzzle, where different shaped
language-games are individual pieces of the puzzle that fit together, against one
another, to form the whole picture. To diagram a form of life properly, one would
need a three-dimensional overlay to show how language-games connect, intersect,
and overlap. Wittgenstein suggests the analogy of spinning a thread by twisting
many different individual fibers together to illustrate the "family resemblances" of
games; the only thing that runs the full length of the thread is the "continuous overlapping of those fibers.,,174 Such relationships among language-games indicate that
we cannot separate, identify, and describe a single language-game by itself; rather,
we must examine different language-games and how they are related within a
particular form of life.
However, if a form of life is analogous to the thread of woven fibers of
language-games, then the question arises, "What makes a religious language-game
(or religious language-games) so special?" As Wittgenstein notes, language-games
"come and go" and become "obsolete and forgotten." If language-games are
simply "given" instead of being based upon some empirical facts about human
nature or the natural world or some "deeper," metaphysical facts, then it seems that
there is nothing in particular to commend playing the religious-theistic-Ianguagegame. What commends the language-game(s) of theism over the language-game of
Santa Claus, or the language-game of witches, or the language-game of UFOs and
intelligent aliens, or simply the language-game of atheistic naturalism? If
philosophy simply leaves everything the way it is and describes the grammar of the
language-game, then there can be no claims for the importance of playing one
language-game rather than another (expect internal to the language-game itself).175
To account for the presumed special status of religious language-game(s), one must
appeal to some descriptive claims about human nature or make some sort of claim
about the consequences and benefits of the religious language-game(s) - claims
that are not a part of the religious language-game(s). Theists want to defend theism
by providing an answer to the question "Why ought one play this game?"
Presumably, Phillips would not admit that this is a legitimate question that
deserves or allows a legitimate response. However, if others wish to try and
Ludwig Wittgenstein, Philosophical Investigations, Section 67.
This line of criticism has been urged by others: See W. D. Hudson. Ludwig Wittgenstein: The
Bearing of His Philosophy upon Religious Belief, pp. 68ff; Patrick Sherry, "Is Religion a 'Form of
Life'?" pp. 165; and Kai Nielsen, "Wittgensteinian Fidesim," p. 209.
174

115

RELIGIOUS EXPERIENCE AND RELIGIOUS EPISTEMOLOGY

193

explain why the language-game of theism is played, some appeal may be made to
anthropological evidence in an attempt to support the claim that there is a universal
religious urging or universal religious nature of man, but this is a long way from
establishing theism. If one does rely upon any sort of evidential claims about
human nature or about the benefits of theism to explain why "the language-game is
played," then this move would undermine the apparent autonomous nature of the
language-game. Without grounding the theistic language-game in some such
claims, however, the Wittgensteinian fideist is left in a position of being unable to
recommend or commend the language-game( s) of theism to himself or others. 176

176 Further discussions of Wittgensteinian language-games and religious beliefs are found in Richard
Bell, "Theology as Grammar: Is God an Object of Understanding," Religious Studies, Vol. II, 1975, pp.
307-17; Alan Brunton, "A Model for the Religious Philosophy of D. Z. Phillips," Analysis, Vol. 31,
1970-71, pp. 43-48; Dallas M. High, "Belief, Falsification, and Wittgenstein," International Journalfor
Philosophy of Religion, Vol. 3, 1972, pp. 240-50; James Kellenberger, 'The Language-Game View of
Religion and Religious Certainty," Canadian Journal of Philosophy, Vol. 2, no. 2, 1972, pp. 255-75;
and Fergus Kerr, Theology after Wittgenstein (Oxford: Basil Blackwell, 1986).

V. Religion and Science!

Arguably, in the course of Western intellectual history, the rise of Christian


theology and the development of modem science stand alone as the most
influential and widespread among all the different ideas, theories, and
developments since the beginning of recorded history.2 When we think of the many
fundamental ways in which Christianity and science have been responsible for
shaping and molding and otherwise influencing the fundamental concepts, values,
and structure of modem Western societies, it is difficult to argue with such an
assessment. It is not surprising then that religion and science are frequently found
to have been in conflict for control of the hearts and minds of men and women. In
modem times, the scientific revolution stands alone in terms of both the breadth
and depth of the social changes that have accompanied it. It is not a matter of
simple hyperbole that the scientific revolution is called a revolution. Its changes
have proven to be global; in contrast, the Protestant Reformation, as revolutionary
as it might have proven to be, was a "domestic affair" among people in Western
European countries. 3
By examining two different historical "case studies" of major pivotal periods of
conflict between religion and science, we can better understand the current
relationship between religion and science. In the history of the relationship
between religion and science, there are two periods that are most significant: one is
the rise of modem science in the seventeenth century and the resulting conflict
between Copernicus's heliocentric astronomy and the Roman Catholic Church, and
the second is the development of Charles Darwin's theory of evolution in the late
nineteenth century and the resulting conflict between evolutionists and
fundamental Protestants. The conflict between Copernican astronomy and the
Roman Catholic Church is one whose general nature is known to most people who
have studied either the history of the development of Christianity or the history of
1 Attention in this volume to the relationship between religion and science is not confined to this
chapter. Readers should also examine Chapters III and VII for further discussion of the interplay
between religion and science in the twentieth century.
2 See, for example, Herbert Butterfield, The Origins of Modern Science, Second Edition (New York:
Macmillan, 1957), pp. 175-90. Also see Alfred North Whitehead, Science and the Modern World (New
York: Free Press, 1925), p. 2.
3 See Whitehead, ibid., pp. 1-2.

194

RELIGION AND SCIENCE

195

the development of modern science. The story is worth retelling, however,


because, as the saying goes, the devil is in the details, or in this case, the science is
in the details. These brief excursions into the history of the conflict between
religion and science will provide the basis for understanding the relationship
between religion and science in the twentieth century.
GALILEO GALILEI AND COPERNICAN ASTRONOMY
Contrary to popular belief, Galileo Galilei did not have a hand in actually
formulating Copernican astronomy. He was, however, its most influential
supporter and popularizer, and he did provide the first experimental evidence to
support the new heliocentric view of the solar system. At the time, the new
heliocentric cosmology was in direct conflict with the accepted and dominant
Ptolemaic geocentric astronomy. In terms of explaining and predicting
astronomical phenomena, such as the position of the planets, Copernican theory
was immediately and imminently more successful than Ptolemaic theory.
Committed to the Aristotelian notions of the "perfect" circular motion of heavenly
bodies and their uniform speed, Ptolemaic astronomers had to appeal to the
introduction of increasingly ad hoc and more complicated notions such as
epicycles and deferents to explain the irregular motions of the planets. Of course,
Copernican astronomy also struck at the heart of accepted and dominant Christian
theology and its anthropocentric view of the unique importance of human beings.
Several scholastics, including William of Ockham in the fourteenth century (most
remembered now for wielding "Ockham's Razor," the law of parsimony), had
maintained that it must be possible for God to create many different worlds and
perhaps even many better worlds than this one given that God is omnipotent. 4
However, by the early seventeenth century, Christian theology was dominated by
the anthropocentrism according to which this world and human beings were seen
as occupying a unique and central part of God's perfect creation.
GALILEO AND THE CHURCH
The conflict between Galileo and the Roman Catholic Church covered a wide
range of issues and occurred on several different levels. Most obvious perhaps is
the matter of the content of the Ptolemaic theory versus the content of the
Copernican theories. 5 One theory says that the sun moves around the earth, and the
other says that the earth moves around the sun. This difference was certainly an
important issue, but it was not the most important or fundamental one. In the spirit
of trying to avoid a confrontation, the Church had taken the stand that the new
Copernican theory could be used as an "instrumental aid" for making calculations
easier in astronomy if it was not represented as being actually true about the world.
Andreas Osiander had written an introduction to Copernicus's On the Revolutions
of the Celestial Spheres in which he had maintained that the heliocentric theory
made the calculations of the astronomers easier as they predicted the position of
the planets (after all, they would not have to deal with the epicycles and deferents
4 See John Hedley Brooke, Science and Religion: Some Historical Perspectives (Cambridge:
Cambridge University Press, 1991), p. 62.
5 See James F. Harris, Against Relativism: A Philosophical Defense of Method (LaSalle, TIL: Open
Court, 1992), pp. 16ff.

196

ANALYTIC PHILOSOPHY OF RELIGION

in their calculations). The Church was willing to allow Galileo and others to use
the Copernican theory and even to teach it to others so long as they did not
represent it as being the correct and true account of the real world, but Galileo
refused this "compromise."
The main source of the dispute was focused upon three fundamental and closely
related issues. 6 The first issue, which was more fundamental and important than the
content of the two different theories, was the matter of the status of scientific
theories. Galileo was not willing to allow the difference between the two theories
to be captured by simply calling the Copernican theory a "supposition" or a
convenient "instrumental aid" for purposes of calculations. Galileo thought that
scientific theories have observable consequences that can be deduced from the
theories, and scientific theories thus tell us something about the world or predict
something about actual, natural phenomena. Scientific theories are not "loose,"
"floating" theories that are unattached or unrelated to the "real" world of our sense
experiences.
This view of the status of scientific theories follows from the second source of
dispute between Galileo and the Roman Catholic Church. A scientific view of the
universe includes the view that there are general laws and principles in terms of
which individual and particular natural phenomena can be understood, explained,
classified, and predicted. Furthermore, this view of the relationship between
general laws and principles on the one hand and facts or states of affairs on the
other is a universal and general view about the nature of the universe. Science does
not provide us with a sporadic or episodic view of the nature of things. Alfred
North Whitehead has described this modern, scientific point of view by claiming
that it involves "a vehement and passionate interest in the relation of general
principles to irreducible and stubborn facts.,,7 These general principles were
understood as mathematical in nature. In the seventeenth century - beginning with
Copernicus through Galileo to Newton - mathematics was no longer viewed as a
system of quaint games that mathematicians played just among themselves with no
relevant consequences for the real world. In the seventeenth century, Copernicus,
Galileo, and Newton used mathematics to formulate the abstract and universal laws
and principles of science to demonstrate through precise formal reasoning how
observable consequences would follow from these abstract, general laws and
principles. 8
The third area of major dispute between Galileo and the Roman Catholic
Church, and perhaps the most fundamental and important, involved the importance
of and the attention paid to those "irreducible and stubborn facts." The "scientific
method" requires that theories be confirmed by data resulting from carefully
controlled, regulated, documented, and repeated scientific experiments and
observations. Galileo arrived at his commitment to the Copernican theory over the
Ptolemaic theory not just because of the simpler mathematics involved. His many
different observations with his improved version of the telescope provided the

Ibid., pp.17-22.
Whitehead, Science and the Modern World, p. 3.
8 Ibid., pp. 54-55.

RELIGION AND SCIENCE

197

"irreducible and stubborn facts" as the bedrock for his claim that the Copernican
theory was the correct one to be used to describe our solar system.
THE CHURCH VERSUS SCIENCE
This fundamental and irreconcilable difference about the proper method for
arriving at claims about the natural world and the proper kind of epistemic warrant
for such claims lies at the very heart of the difference between Galileo and the
Roman Catholic Church. As I have described this difference elsewhere,
On the one hand, there were the appeals to papal authority, church history, and
Biblical accounts of cosmology, the authority of the history of philosophy and
the evidence from "common sense." On the other hand, there were his own
"scientific" observations and his use of scientific reasoning. 9
Galileo took the observations that he made with his telescope to be confirming
evidence of the Copernican theory. 10 It is on the basis of such scientific evidence
rather than papal or biblical authority that scientific theories about the natural
world should be judged, he insisted, and in 1610, he published Sidereus Nunicus
(The Starry Messenger), in which he reported his astronomical observations.
Galileo made his position about the relative importance of the Bible and science
for attending to understanding and explaining matters of natural phenomena quite
clear. II In such matters, the scientist is held to the higher standard of the
"inexorable and immutable" laws of nature and, of course, the "irreducible and
stubborn" facts. Such reasoning is what lies behind the often quoted remark of
Galileo's that "the Bible tells us how to go to heaven, not how the heavens go."
GALILEO AND THE INQUISITION
Following the publication of The Starry Messenger, the conflict between Galileo
and the Church intensified. In 1616, the Roman Catholic Church branded
Copernican heliocentric astronomy as "false and altogether contrary to the Holy
Scripture," On the Revolutions of the Celestial Spheres was placed on the Index of
banned books, and the teaching of Copernican theory in any form was prohibited. 12
In 1624, Galileo received a license from the Church to write a book comparing
Ptolemaic and Copernican theories on the condition that he not take a position
favoring one or the other in the book. When Galileo published his Dialogue on the
Two World Systems in 1632, the book clearly made a strong, convincing case on
Harris, Against Relativism, p. 19.
Galileo observed that the surface of the moon was rough with mountains, much like the surface of
the earth, and not a "perfect sphere" as Aristotle (and Ptolemy) had claimed. He also discovered the
"stars" of Jupiter, the smaller, orbiting bodies of Jupiter that proved to be the moons of Jupiter, and he
correctly predicted their orbits and positioning. He was the first to observe sunspots and, perhaps most
importantly, the phases of Venus caused by the sun's reflected light as the planet orbits the sun - a
direct piece of evidence that was predicted by the Copernican theory.
11 See Galileo Galilei, "Letter to Madame Christine of Lorraine, Grand Duchess of Tuscany,
Concerning the Use of Biblical Quotation in Matters of Science," in Men of Physics: GaZileo GaliZei,
His Life and His Works, translated by Raymond 1. Seeger (Oxford: Pergamon Press, 1966), p. 271.
Also, see Harris, Against Relativism, p. 20.
12 See Jerome Langford, Galileo, Science, and the Church (New York: Desclee, 1966), pp. 97-98.
9

10

198

ANALYTIC PHILOSOPHY OF RELIGION

the basis of scientific evidence for the Copernican heliocentric theory. Galileo was
put on trial for heresy by the Inquisition in 1633, made to sign a "confession" for
his sins, and, finally, sentenced to house arrest in his home near Florence, where he
remained until his death in 1642, the same year in which Isaac Newton was born.
GALILEO AND SCIENCE
For understanding the contemporary, twentieth-century relationship between
science and religion, the most interesting and important aspect of the conflict
between Galileo and the Roman Catholic Church is the struggle for authority over
matters having to do with the physical world and natural phenomena. Galileo
insisted that the Bible and the Church have no authority over such matters and
should not seek to assert authority over such matters. The Church, of course, took
the position that theology trumps science.
A very significant element for analyzing and evaluating the lessons to be
learned from this conflict for the twentieth century is the fact that we now know
that the "best scientific evidence" that Galileo presented in support of the
Copernican theory in his Dialogue on the Two World Systems in 1632 was either
very weak, misinterpreted, or just plain wrong.13 Scientists now know that the
connections Galileo drew between sunspots and the earth's tides and the
Copernican theory were not justified. In a sense, by today's standards, Galileo was
a champion of bad science.
The important point, however, is that bad science is still science because of the
method of investigation and the method used for arriving at certain claims about
the nature of the physical world. The scientific method does not guarantee that its
use will not result in mistakes; on the contrary, there are opportunities for various
kinds of mistakes at every stage of a scientific inquiry. However, the scientific
method is the only method that includes as a part of the method itself a mechanism
for identifying and correcting its mistakes. 14 Since science is an ongoing, public
method conducted by members of a scientific community, any individual's
mistakes are always open to the scrutiny of the entire community over an indefinite
period of time. Indeed, the only reason that we now know about the mistakes that
Galileo made is because of the work of other scientists.
We should therefore remember Galileo as a champion of the scientific method
and of science generally rather than simply as a defender of Copernican astronomy.
Galileo saw, perhaps even more clearly than Copernicus, the importance of
accounting for individual, particular natural phenomena in terms of a general
theory based upon abstract mathematical laws and principles. He destroyed the old
"two world view," according to which the heavens were separate and distinct from
the causal and natural explanations that apply to natural phenomena on earth. In
this sense, he paved the way for Newtonian mechanics and for the possibility of a
truly general cosmology. Galileo insisted - upon the pain and threat of possible
excommunication, torture, and even death - that in matters pertaining to the natural
world, there is no higher authority than science based upon careful empirical
13 See William R. Shea, Galileo's Intellectual Revolution (New York: Science History Publications,
1972), pp. 174ff.; William Wallace, Galileo and His Sources (Princeton: Princeton University Press,
1984), pp. 308ff.; and Harris,Against Relativism, pp. 20ff.
14 See Harris, Against Relativism, Chapter VII.

RELIGION AND SCIENCE

199

observation and the application of abstract, general laws and principles. Gali1eo's
stance in defense of science and the scientific method marked a pivotal point in the
new direction that was to come in the separation of religious authority and
scientific authority in matters having to do with the physical world and natural
phenomena. Galileo earned the distinction of becoming perhaps the first martyr of
science, and given the developments in astronomy since the seventeenth century,
his prosecution by the Church for his scientific beliefs and teachings has since
served as a symbol of the folly that can result when religion and science conflict.
DARWIN AND THE THEORY OF EVOLUTION
Charles Darwin's publication of the Origin of Species in 1859 created an
equally confrontational situation between religion and science. Because of the
many different changes that had taken place in organized religion since the
seventeenth century, the conflict generated by Darwin's theory of evolution by
natural selection pitted science against Protestantism - the Church of England in
Great Britain and fundamentalist Protestantism in the United States. Although it
was easy to ignore the differences between the biblical accounts and science
theories about abstract matters such as astronomy and mechanics, biology brought
the differences too close to home to ignore. Darwin's theory of evolution, if true,
seemed to explicitly undercut any literal interpretation of the Bible account of the
creation story in Genesis, and the notion that the guiding and determining factor in
evolution is natural selection undercut any literal interpretation of a divine
presence or divine involvement in human history. Because evolutionary theory
based on natural selection was so revolutionary and challenged the fundamental
understanding of human nature and human history, it created more furor and
provoked more of a confrontation with religion than any other development in
science since the time of Galileo.
Darwin's voyage around the world on the H. M. S. Beagle provided him with
the empirical data that were to provide the basis for his work. He observed and
documented the geographical isolation of different species of animals. For
example, in the Galapagos Islands, completely different species of life existed on
different islands that were only a few miles apart. Upon his return to England,
Darwin and other experts were able to study the fossils and the anatomies of the
specimens he had collected. According to the theory of speciation, supported by
both Aristotelianism and Christian theology, different natural kinds with different
essences should be suited to different natural environments, so one should expect
that very similar species would be found in very similar environments and very
different species would be found in very different environments. The results of
Darwin's research conclusively confirmed the general and sweeping reality that the
overwhelming explanation for the locations of relatively similar and different
species has to do with the relative geographical proximity or isolation among the
different species rather than the environmental similarities or differences of the
locations in which they are found.
DARWIN AND NATURAL SELECTION
The dominant and received biological view of the immutability and
unchangeability of the species (defended by Christian theology at the time) could

200

ANALYTIC PHILOSOPHY OF REUGION

not be made to accommodate or explain the wealth of empirical data that Darwin
had collected. Once again, it was evident how the commitment to the "irreducible
and stubborn facts," the bedrock of scientific investigation, is fundamental to the
scientific process. Darwin began to try and explain the natural data he had recorded
and the natural phenomena he had observed by searching for the applicable general
abstract laws and principles. The revolutionary aspect of Darwin's answer was not
just the theory of evolution but evolution by natural selection. Others had
suggested the idea of evolution earlier, primarily the French biologist Jean
Lamarck in his Zoological Philosophy in 1809, but Lamarck never abandoned the
notion of a hierarchy of beings embedded in Christian theology, and he did not
provide an explanation for a mechanism to explain how and why changes might
take place in species over time except in terms of some divinely created plan
according to which different species became more and more organically "perfect"
over long periods of time.
Whatever foreshadowing of Darwin's theory of evolution there might have
been, it caused no great stir and no direct conflict between science and religion.
Darwin's explanation of the changes in terms of adaptations that were beneficial
for the survival of the organism being selected by nature was revolutionary and
changed everything that was to follow. IS Most commentators agree that it was
Darwin's notion of natural selection rather than the bare notion of evolution that
was so revolutionary. Nothing in the life sciences has been the same since Darwin,
and much in philosophy and theology has changed also as a result of the joint
notions of evolution and natural selection. Darwin's theory of evolution (which one
should always understand as including natural selection) has unified all of biology
- from genetics and DNA to comparative anatomy and speciation to fossils and the
history of life on this planet. Probably more controversy has surrounded Darwinian
evolutionary theory - both in science and religion - than any other scientific theory
that has ever been developed simply because Darwinian evolutionary theory, more
than any other scientific theory, offers some answer to the question of why there is
life and why human beings exist as we do. So perhaps more than any other
scientific theory, evolution pitches its tent in the camp of religion.
Because of the broad and general implications of evolution, it had more
philosophical and theological import than any other single scientific theory ever
developed. Darwin's notion is so powerful that Daniel Dennett has said if there
were a prize for the single best idea that anyone has ever had, he would give it to
Darwin. 16 The notion of evolution itself became threatening to both organized
religion and theology because it abandoned the notion of the immutability of the
species based upon divinely established essentialism to account for the unique
nature of each species. However, it was the notion of natural selection, according
to which it is nature and not God that is responsible for shaping and determining
the nature of the different species, that threatened any common understanding of
how God might have a hand in the workings of nature. It is evolution by natural
15 Actually, others (including Patrick Matthew, a Scottish naturalist in the early nineteenth century)
had suggested the notion of natural selection as a bare idea, but no one had developed any sort of
general, cohesive theory incorporating the notion of natural selection.
16 Daniel Dennett, Darwin's Dangerous Idea: Evolution and the Meanings of Life (New York:
Simon and Schuster, 1995), pp. 20-21.

RELIGION AND SCIENCE

201

selection, and not just the notion of evolution per se, that is mainly responsible for
the twentieth-century conflict between religion and the biological sciences. 17
EVOLUTION, NATURAL SELECTION, AND RELIGION
One of the most extensive and detailed treatments of the controversy created in
religion in the twentieth century by the development of Darwin's evolutionary
theory is Ian Barbour's Issues in Science and Religion. 18 Barbour indicates four
areas of conflict between Darwinian evolutionary theory and religion: 19 1) The
challenge to the argument from design, 2) The challenge to human dignity, 3) The
challenge of evolutionary ethics, and 4) The challenge to holy scripture. Of these
different areas of conflict, the first two are, by far, the most significant. By the late
nineteenth century, the literal interpretation of the Bible did not occupy the central
position in mainstream, organized religion that it had in the seventeenth century.
The Roman Catholic Church and the Church of England had both abandoned literal
interpretations of the Bible because of the continued development of science and
the parallel development of "higher criticism," scholarly interpretations and
analyses of the scriptures, and the recognition of different "levels of truth" allowed
for nonliteral interpretations of biblical texts that might conflict with accepted
scientific claims. 2o For example, the Copernican heliocentric astronomy that had
caused such a furor in the seventeenth century was, by the mid-nineteenth century,
commonly accepted and accommodated by nearly all recognized forms of
organized religion.
Given that the immutability (and the essentialism) of the species is undermined
by evolution, the avenue of addressing moral issues by appealing to a uniquely
human moral sense or moral capacity was no longer available, so the unique
position of human beings in the universe (Barbour's "human dignity") was
undermined. While Darwin himself thought that evolutionary theory could
accommodate normative ethics, others (for example, Thomas Huxley) did not.
Issues about the foundation and justification and the general problem of the
relationship between normative moral theory and religious belief and the appeal to
divine authority are addressed in Chapter VIII.
EVOLUTION AND DESIGN
What Barbour calls the challenge to religion in terms of claims about the design
of the universe and the challenge in terms of how to understand human dignity
serve to illustrate the revolutionary aspects of evolutionary theory. These areas of
conflict also serve to illustrate how the relationship between science and religion
changed from the seventeenth to the twentieth century. In the case of the conflict
between Galileo and the Roman Catholic Church, the main conflict was over the
epistemological status of Copernican theory and the nature of the scientific
method. In the case of Darwin and evolutionary theory, the conflict between
17 See Charles Darwin, On the Origin of Species by Means of Natural Selection (London: Murray,
1859), p. 127. Quoted in Dennett, ibid., pp. 41-42.
18 Ian Barbour, Issues in Science and Religion (Englewood Cliffs, N.J.: Prentice-Hall, 1966), pp. 80114.
19 Ibid., pp. 89-98.
20 Ibid., p. 100.

202

ANALYTIC PHILOSOPHY OF REDGION

science and religion arose because of the specific content of the theory - not
because of the scientific method. Indeed, one source of opposition to evolutionary
theory has been that it is not based upon good scientific method and evidence. 21
Religious belief in design in the universe and the teleological argument based
upon such claims have proven to be remarkably resilient to philosophical attack.
For example, in his attack on the argument from design in his Dialogues
Concerning Natural Religion,22 Hume raises many interesting and difficult
problems for the argument from design (see Chapter III), including the problem of
evil (see Chapter VI); however, one of the most puzzling and difficult issues in
Hume's treatment of the argument from design is that at the conclusion of the
Dialogues (Part XII), after Philo's scathing attack upon Cleanthes's use of the
argument from design, Philo concludes by suddenly agreeing with Cleanthes and
accepting the notion of a divine intelligence that is responsible for the design in the
universe!
It is significant to compare in a certain respect the ways in which Philo reacts to
the cosmological and teleological arguments. When Hume has Philo, Cleanthes,
and Demea examine the cosmological argument (Part IX), he has Cleanthes
explicitly suggest that the cause for a particular "system" may very well originate
within that system in order to undermine the necessity of a "First Cause." If we
trace a succession of causes and effects and explain each single "link" in the causal
"chain," then there is nothing more to be said about the cause of the chain as a
whole rather than simply the combination of the individual causes that are a part of
the chain. 23 Thus, we do not have to "go outside" the system to a First Cause to
explain the cause of the system itself. The interesting and important point of
comparison with the argument from design is that Hume cannot bring himself to
entertain seriously the possibility that the design of a system might originate
completely within a system. 24
In Part VIII, Hume does have Philo suggest that the design and order in the
world are necessary to preserve the world. This suggestion is contrary to the notion
of design used in most forms of the argument from design, in which design is
regarded as an additional lUXury that the world enjoys. This change in view sees
design as a necessity for the world's existence. Without the detailed and precise
adjustments of "means to ends" in an animal, it soon perishes, and so too the world
would perish without the regularity and order we observe in it. 25 Here, Hume is
perhaps closer to recognizing what is the revolutionary aspect of Darwin's notion
of natural selection than at any other place. For Darwin, design is not a luxury; it is
a necessity. Instead of rhapsodizing in awe about the design in nature, we should
expect it. Such design is a natural part of the world that we now experience, since
it has evolved by natural selection.

For example, see Dennett, Darwin's Dangerous Idea, Part n, pp. 149-331.
David Hume, Dialogues Concerning Natural Religion, edited with an introduction by Richard
Popkin (Indianapolis, Ind.: Hackett, 1980).
23 Ibid., Part IX, p. 56. This point is discussed more thoroughly in Chapter Ill.
24 Dennett calls this Hume's "close encounter" with the notion of evolution. See Dennett, Darwin's
Dangerous Idea, pp. 28ff.
25 Hume, Dialogues Concerning Natural Religion, Part Vll, pp. 51-52.
21

22

RELIGION AND SCIENCE

203

Most of Hume's attacks are aimed at the inference by the use of analogy from
observed, mechanical design to the existence of an intelligent designer, but if this
sort of inference is not justified and yet the design is still there, how is it to be
explained? Hume could not bring himself to admit that the apparent design in the
uni verse is the result of mere chance, but he had no other explanation without
using God. Darwin provides an alternative explanation to mere chance for the
existence of design in the natural world. Chance operates within the general,
universal laws of evolution and natural selection that provide the order and
regUlarity. Design in the natural world can thus be explained as arising from within
the natural order of things without the necessity of appealing to some divine
intelligence outside the system. Some of Hume' s strongest attacks are aimed at the
evidential part of the argument from design, particularly the existence of evil.
Darwin reinstates and strengthens the evidential claims of design in nature, but he
then completely undermines the argument part of the argument from design, that
is, the inference to the existence of a theistic or deistic divine designer, by
providing a scientific explanation for design that occurs completely within the
system of nature. 26 The account of the existence of design by appealing to
Darwinian evolutionary theory is an argument to the best, most rational
explanation that parallels Hume's argument concerning miracles. Given that this
apparent design exists, what is the best, most plausible explanation for it? Maybe
sheer chance does not fair very well in comparison with God, but evolutionary
theory does.
At the time of Darwin, perhaps the most popular and well-known version of the
argument from design was traced to the argument originated by William Paley at
the very beginning of the nineteenth century in his Natural Theology. In an
example that has now become famous and that is included in most discussions of
the argument from design, Paley appeals to our intuitions to have us recognize an
intrinsic difference between a rock, which one might find in an open field, and a
watch one might find in a drawer. The evidence of design in the organic world,
among living organisms, is far greater, Paley claims, than the design of the watch.
With his famous example of the human eye, he argues that the world of living
creatures is rich with an infinite variety of different mechanisms that are so highly
designed and specialized for the various purposes of the different organisms that
we cannot help but admit the existence of a designer who, according to the
principle of proper proportionality, must be divine and infinite to account for the
infinitude of design in the natural world. The only plausible explanation for such
obvious, overwhelming design seemed to be something outside the natural world
itself. Thus, the explanation of design in the universe has traditionally been the
exclusive purview of philosophers and theologians, and naturalists (or natural
philosophers) had taken their only clue for explaining such design from those
philosophers and theologians.
After Darwin, it was no longer necessary to "go outside the system," to go
beyond the natural world itself, to explain such design, and design in the natural
26 This way of putting the contrast between Hume and Darwin has evolved through discussions with
several colleagues and through much reading. Credit to a particular source that is now lost to both
memory and my notes may well be in order.

204

ANALYTIC PHILOSOPHY OF REUGION

world can be understood as a natural and expected part of the way in which things
operate in the natural world rather than as a lUxury or the result of the providential
influence of a beneficent divine creator. The way was thus cleared in the second
part of the nineteenth century for the rise of deism and the view of God as the
absent watchmaker.
EVOLUTION, NATURAL SELECTION, AND ANTHROPOCENTRlCITY
Darwin's rejection of the immutability of the species and the essentialism upon
which claims about such immutability were based was a second aspect of Darwin's
theory of evolution that resulted in direct conflict with accepted religious belief.
Christian theology in the mid-nineteenth century was completely compatible with
the attempt by natural science to find the hidden essences in nature. To this extent,
Christian theology, which emphasized the uniqueness of human beings by
appealing to an eternal soul, a divine spark, or a special moral sense, was
compatible in spirit with Aristotle, who accounted for the uniqueness of human
beings by appealing to a particular intellectual ability or a particular telos. The
classification of natural kinds by naturalists in the nineteenth century was based
upon the notion of scala naturae, or "the great chain of being," derived from
Christian theology. According to this classification, human beings, of course,
occupy the highest position in the hierarchy of beings and objects in the physical
world because human beings were thought to be the only creatures in the natural
world that are "made in God's image" and capable of some sort of interaction with
God and a "personal" relationship with God.
In the great chain of being's hierarchical account of natural kinds, human beings
are placed at the very top with the rest of the natural world stretching out
metaphorically in a long continuous chain through the other living organisms
(animals and plants) to inorganic material objects (rocks and the like). Human
beings were thought "to have dominion over nature" because of our special status,
and the rest of nature was thought to exist for the benefit and use of human beings.
Humans then have intrinsic value to God whereas every other part of nature has
utilitarian value depending upon its value for human beings and the use to which
humans can put it. The basic claim of speciation, that individual species are distinct
natural kinds resulting from certain essential characteristics or traits, was thus the
dominant view in both secular philosophy and natural theology in the midnineteenth century.
This anthropocentric view of the position of human beings in nature is undercut
in one fell swoop by Darwin. In evolutionary theory, human beings are simply one
species among many, and the explanation for how human beings exist as we do
with the particular attributes that we have is the same kind of explanation that is
given for each species in turn. Whatever special and unique properties or attributes
that human beings may have are the result of favorable properties on the basis of
which our ancestors interacted with their environment. Such properties (traits)
were then selected for by natural selection and passed on to us through countless
generations. The view of human beings embraced by twentieth-century biology,
which religion had to accommodate, is a view according to which there is no
"essence" to human beings (or any other species, for that matter). Instead of being
creatures that are placed in nature by God, human beings (and all other creatures)

RELIGION AND SCIENCE

205

are creatures that are made by nature. Human beings are thus knocked from their
lofty pedestal and wind up (literally, if one goes back far enough) in the mud and
slime with all the other species.
Darwin's theory of evolution also undercut anthropomorphism based upon a
Cartesian denial of sensations and feelings in animals. The long-standing solution
to the question of why animals suffer (since their suffering can be neither the result
of Original Sin nor redemptive) was simply to deny that they do suffer. Such a
view of the nature of animals and the resulting special status of human beings is
impossible given evolutionary theory. We human beings have inherited both our
physiological characteristics and our physical behaviors from animals. Animals are
responsible for the kind of creatures that we human beings are, so it becomes
wildly implausible, without some ad hoc intervention, to claim that human beings
are qualitatively different from other animals because we have a mind and
sensations and feelings while other animals do not. As James Rachels says,
Once we see the other animals as our kin, we have little choice but to see their
condition as analogous to our own. Darwin stressed that, in an important sense,
their nervous systems, their behaviors, their cries are our nervous systems, our
behaviors, and our cries, with only a little modification. They are our common
property because we inherited them from the same ancestors. 27
Much controversy has been generated by Darwin's treatment of the place of
human beings in his theory of evolution, and there have been several attempts to
preserve the special status of human beings by both scientists and theologians
alike. 28 Pursuing the differences between Darwin and other biologists and
theologians would be an interesting and valuable exercise in the history of the
conflict between religion and science, but it would take us too far afield from our
present concerns. These somewhat cursory examinations of the conflicts generated
by Galileo' s defense of Copernican astronomy and Darwin's development of the
theory of evolution are valuable because they demonstrate a significant change in
the nature of the conflicts between science and religion over a period of
approximately three hundred years.
In the case of Gali1eo, at a time early in the history of modem science, the
conflict between religion and science was one generated mainly by a conflict
concerning the competing authority of the methods of religion and science. Galileo
adopted an objective method based upon careful, empirical observations. Papal
authorities refused to look through Galileo' s telescope and consider his empirical
evidence since they regarded the telescope as an "instrument of the devil" that
confused and distorted the senses instead of enhancing them. The main conflict
between religion and Darwin's theory of evolution did not involve a comparable
27 James Rachels, Created from Animals: The Moral Implications of Darwinism (Oxford: Oxford
University Press, 1990), p. 131. Rachels focuses upon the result of the loss of man's special place in the
universe upon our understanding of moral theory.
28 For example, A. RWailace tried to preserve a special place for human beings because of our
"high faculties." See Barbour, Issues in Science and Religion, pp. 92-93, and Dennett, Darwin's
Dangerous Idea, pp. 66-67. For a detailed and lengthy account of the differences between Darwin and
Wallace, see Adrian Desmond and James Moore, Darwin (London: Michael Joseph Publishing, 1991).

206

ANALYTIC PHILOSOPHY OF REUGION

and direct challenge to the method of science itself. There were (and still are)
claims that evolution itself is not good science, but the most significant and
dramatic conflicts arose because of the content of Darwin's theory - not because of
his method of searching for general and universal natural laws to explain empirical
facts.
By the end of the nineteenth century, the use of the scientific method and the
development of modem science had become so highly developed and so firmly
ensconced in human attempts to understand and explain natural phenomena that
very little conflict between religion and science continued to be generated by
disputes concerning the method of scientific inquiry - except in the case of extreme
fundamentalists. Later in this chapter, I will show how both the method of religion
and the method of science have undergone intense scrutiny and significant changes
in the twentieth century. Much the same can also be said for the content of both
science and religion. Because of these changes in both method and content, the
earlier, internecine conflict between religion and science, illustrated by the cases of
Galileo and Darwin, has been significantly vitiated. In some cases, philosophers of
science would say that religion has become more like science; in other cases,
theologians would say that science has become more like religion.
CONFLICT OF METHOD: SCIENTIFIC CREATIONISM
Except for the remaining Protestant fundamentalists, who still cling to a literal
interpretation of the Bible, conflicts between religion and science in the twentieth
century have not been generated by disagreements about method so much as by
disagreements about the relative domains of authority of religion and science and
the purposes to which each is put. Fundamentalist opposition to evolution in the
United States resulted in the Scopes Trial of 1925, which pitted the renowned
Presbyterian William Jennings Bryan for the prosecution against equally renowned
trial lawyer Clarence Darrow for the defense. The now famous "monkey trial"
involved the case of John T. Scopes, a high school science teacher in Dayton,
Tennessee, who was accused of violating a state statute against teaching evolution
in public schools. This case, celebrated though it was, did not prove to be
philosophically interesting at all. The highly visible and public trial did not provide
a venue for a discussion of the merits of either evolution or the literal interpretation
of the Bible. The trial did serve to illustrate the considerable political influence of
Protestant fundamentalism in the southern United States in the early part of the
twentieth century, but underlying the political aspect of the trial was the
fundamentalist opposition, based upon a literal interpretation of the Bible, to
evolution.
The conflict between fundamentalism and evolutionary theory parallels the
dispute between Galileo and the Roman Catholic Church in that the main source of
the dispute stems from claims about competing methods of inquiry. The authority
of the Bible, which is still regarded by fundamentalists as the literal Word of God,
is pitted against "monkey science.,,29

29 An interesting issue that is seldom brought to light is that there are two different and conflicting
stories of creation in Genesis.

RELIGION AND SCIENCE

207

SCIENTIFIC CREATIONISM AND EVOLUTION


Far more interesting and challenging than the opposition to evolution based
explicitly upon a literal interpretation of the Bible is the response of many
conservative religious groups based upon "scientific creationism" or simply
"creationism." Among the dozens of different writers who have defended scientific
creationism is Henry Morris?O Morris and other scientific creationists claim that
their attack upon evolution does not rely solely upon a literal interpretation of the
Bible or the attempt to undermine the scientific method. On the contrary, the
scientific creationists claim that their attack upon evolution is because it is bad
science, that is, the creationists use their understanding of the scientific method to
claim that creationism is better science than evolution! This represents a significant
change in tactic on the part of the religious believers in opposition to science and
specific scientific theories. It is a tactic based upon the "if you can't beat them, join
them" mentality. Given the predominantly scientific age of the twentieth century,
opposition to evolution based explicitly upon literal readings of the Genesis stories
of creation lacks much of the persuasive force it might have had in the nineteenth
century. Scientific creationists ostensibly adopt the scientific method and then try
and show that evolution is hoisted on its own petard when judged as a scientific
theory and that "real" science or "good" science bears out the Biblical account of
creation.
IS SCIENTIFIC CREATIONISM SCIENTIFIC?
There are many issues to be explored in the examination of evolution theory
itself and in the comparison between evolution and creationism, and I cannot
attempt to address all of those issues here?! A very brief examination of the basic
contentions of the creationists reveals the basic grounds for the dispute about
creation and how the theory of evolution has affected the current relationship
between science and religion. Ironically, so far as the history of conflict between
religion and science is concerned, when evolution is disparaged as bad science by
the scientific creationists and compared with other models of "good" science, it is
sometimes Galileo's defense of Copernican astronomy with which evolution is
compared. For example, creationist David Watson says,
So here is the difference between Darwin and Galileo: Galileo set a
demonstrable fact against a few words of Bible poetry which the Church at that
time had understood in an obviously naive way; Darwin set an unprovable

30 See Henry Morris, The Twilight of Evolution (Grand Rapids, Mich.: Baker Book House, 1963);
The Remarkable Birth of Planet Earth (San Diego: Creation-Life Publishers, 1972); and The Troubled
Waters of Evolution (San Diego: Creation-Life Publishers, 1974). Also see Scientific Creationism,
edited by Henry Morris (San Diego: Creation-Life Publishers, 1974). and D. C. C. Watson, The Great
Brain Robbery (Chicago: Moody Press, 1976).
31 Douglas Futuyma, Science on Trial: The Case for Evolution (New York: Pantheon Books, 1982);
Langdon Gilkey, Creationism on Trial: Evolution and God at Little Rock (San Francisco: Harper &
Row, 1985); Phillip Kitcher, Abusing Science (Cambridge: MIT Press, 1982); and Michael Ruse,
Darwinism Defended (London: Addison-Wesley, 1982).

208

ANALYTIC PHILOSOPHY OF REliGION

theory against eleven chapters of straightforward history which cannot be


reinterpreted in any satisfactory way.32

There are several things about this passage from Watson that are important.
First, he significantly minimizes the conflict between Galileo and the Roman
Catholic Church and seriously distorts the significance of that disagreement. As we
have seen, Copernican astronomy undermined not just a few words of "poetry" in
the Bible but a fundamental theological view of the design of the universe, the
nature of human beings, and the place of human beings in the natural universe.
Copernican theory also undermined the prevailing cosmology, which incorporated
the "two world" view of the universe and was the result of the Thomistic
combination of Aristotelian metaphysics and Christian theology.
The claim that Galileo was defending a simple, "demonstrable fact" while
Darwin was defending an "unprovable theory" is an oversimplified and equally
distorting way of characterizing the difference between Galileo and Darwin. The
"fact" that the earth revolves around the sun is demonstrable only by assuming a
rather sophisticated sense of demonstration. To the unaided "naked eye" (and
hence to common sense), it is obvious that the sun is the one that does the moving,
and so far as the "facts" go, when Galileo produced what he took to be the most
dependable, compelling facts to support Copernican theory, he got many of them
wrong! As we have seen, the dispute between Galileo and the Roman Catholic
Church was not simply over a single "demonstrable fact." The dispute cut very
deeply and hinged upon Galileo's use of the scientific method and his insistence
upon its general authority in matters having to do with the natural world. So,
Galileo was defending a scientific theory and the scientific method as much as
Darwin was.
The "eleven chapters of straightforward history" upon which Watson relies in
his attack upon Darwin is good, old-fashioned biblical literalism again. Once we
begin to examine closely the scientific part of "scientific creationism," the facade
of anything resembling science begins to crumble. Creationists claim to defend
creation over evolution because of the "scientific facts," but the only source of
such "facts" turns out to be the Bible. For example, Henry Morris argues that the
facts support the claim that the universe is really very young (so young that
evolution could never have taken place as Darwin describes), but it turns out,
according to Morris, that "the only proper method" for ascertaining the age of the
universe is by relying upon "biblical data alone.,,33
A good example of how science is abused (to borrow Phillip Kitcher's phrase) is
the way in which the "scientific creationists" play fast and loose with their use of
scientific laws and principles in putting them to quasi-scientific uses. For example,
creationists have borrowed the second law of thermodynamics, the law of entropy,
in their attack upon evolution. The law of entropy says that any closed system
tends to break down over time and lose energy, that is, any closed system moves
from a state of greater order to one of lesser order and from one of more energy to
one of less energy. Creationists use this scientific law to argue that the universe
32
33

D. C. C.Watson, The Great Brain Robbery, p. 46. Cited in Kitcher, Abusing Science, p. 31.
Henry Morris, The Twilight of Evolution, p.77.

RELIGION AND SCIENCE

209

must be wearing down and not, as evolution claims, evolving to a condition of


greater order. But, of course, as several critics have pointed out, the law of entropy
applies only to closed systems. Life on earth is not a closed system (since we
receive new energy from the sun), so the law of entropy has nothing to do with the
question of whether evolution is "true" or not. 34
Several critics have done a very thorough and effective job of debunkinf the
myth of the claim that creationists have made that creationism is scientific. 3 The
clash between evolutionists and fundamentalists in the nineteenth century was an
explicit clash between science and religion - between a scientific explanation for
the origin of life and human beings and a theological explanation. Scientific
creationism appears to move the disagreement to one about what the proper
scientific explanation is, that is, the clash appears to be within science about
whether evolution or creationism is the best scientific account of the origin of life
and human beings. However, this description of the conflict is distorting since, as
we have seen, there is a great deal of creationism in scientific creationism and not
much in the way of science. What little science there is in creationism is bad
science filled with half-truths, misapplications of scientific laws and principles, and
misrepresentation of scientific facts and theories. As Phillip Kitcher says, "The
road to Creationism is paved with bad philosophy. However, the engines that
transport us down that road are fueled by bits of science, variously chopped,
twisted, crushed, mangled, and blended.,,36 The pretenses of scientific creationism
to be scientific create the appearance of moving religion and science closer
together, at least in terms of methodology, when, in reality, we are left with the
same conflict between religion and science that is explicit in the clash between
evolutionists and fundamentalists. The major source of the conflict still concerns
the proper method for investigating and describing nature and natural phenomena,
and, in this sense, the current conflict between creationists and evolutionists, initial
appearances aside, is the same conflict as the one between Galileo and the Roman
Catholic Church.
DARWINIAN EVOLUTION AS SCIENCE
One of the many ways in which different theories differ is in their respective
scopes, and it is here that there is a small kernel of truth in Watson's comparison
between Galileo and Darwin (although Watson does not make the comparison on
this basis). The enormous magnitude of Darwin's theory of evolution based upon
natural selection was hinted at earlier. Darwinian evolution has proven to be an
extremely broad, "unifying" theory for the biological sciences and for some of the
physical sciences such as geology and astronomy as well. In order to evaluate the
criticism of evolution raised by the creationists, we must make some careful
distinctions about just what form of evolutionary theory is under consideration.

34 See Scientific Creationism, edited by Henry Morris, Chapter 3, "Uphill or Downhill"; Ted S.
Clements, Science vs. Religion (Buffalo, N.Y.: Prometheus Books, 1990), pp. 159-60; and Michael
Ruse, Darwinism Defended, p. 296.
35 See Ted S. Clements, ibid., pp. 158-67; Michael Ruse, ibid., pp. 285-302 and Chapter 14; and
Phillip Kitcher, Abusing Science, Chapter 4.
36 Phillip Kitcher, ibid., p. 82.

210

ANALYTIC PHILOSOPHY OF REUGION

Taken very broadly as a general theory of the slow, incremental development in


the natural world, evolution has been adopted by both geology and astronomy to
explain the gradual development of the universe. Geologists now commonly
identify "older" and "younger" geological formations, and astrophysicists
commonly use the notion of evolution to describe the development of the universe
itself. Astronomers now identify the relative ages of various heavenly bodies,
including stars and galaxies, and stars are described as "evolving" through different
stages - from interstellar matter to the "main sequence" stage of stars to red giants
to red dwarfs to black dwarfs. In the life cycles of stars, which span hundreds of
millions of years, we find an evolutionary explanation of the origin and evolution
of the universe itself. Our solar system, for example, is said to have evolved from a
rapidly rotating solar nebula that contained relatively small aggregates of matter
that then grew by attracting solar debris by gravitation, through a process called
"accretion," to form the planets.
Similar evolutionary explanations are given by geologists for both the
geological and chemical composition of Earth itself. These gradual, incremental
evolutionary processes are quite different from, and, on any literal level, ostensibly
incompatible with, the account of creation given in Genesis. Since it is impossible
to address all of the many issues that arise with treating evolution as a general
theory and the various conflicts that arise with religion, I will concentrate here
upon the status of the more narrow form of Darwinian biological evolution based
upon natural selection for natural species and examine the different points of
dispute between biological evolution and religion.
IS EVOLUTIONARY THEORY BAD SCIENCE?
One of the first things that we should note is that original Darwinism based
upon natural selection is incomplete. Some creationists have objected that
Darwin's theory of natural selection was 'just" another theory since it was not
based upon any factual knowledge about how traits were inherited. Darwin never
managed to determine a mechanism for the way in which the desired traits favored
by natural selection are preserved and passed on to succeeding generations of a
species. The only operative model to explain the inheritance of different traits from
the parents by the offspring that Darwin had available at the time was one that
adopted the notion that the inherited trait was a result of a "blending" (an
averaging) of the traits of the two respective parents. As Darwin's critics pointed
out, such a notion of inheritance seems to actually work against the notion of
natural selection, since any particular trait that may offer some distinct advantage
for the organism in terms of survival will be reduced in strength, diluted, or vitiated
by a factor of one-half each time that it is passed on to a member of the next
generation.
It remained for Gregor Mendel to provide the explanation of the kind of
inheritance in which a particular trait may be passed on in undiluted and
uncompromised form from the parent to an offspring. What has now developed
into modern genetics supplies the missing explanation for Darwinian evolution
whereby the inheritance of different traits is explained in terms of the result of

RELIGION AND SCIENCE

211

dominant and recessive genes?? The addition of Mendelian genetics to the original
Darwinian notion of evolution constitutes what is usually referred to as "NeoDarwinism."
It is difficult to understand how it can be taken as a serious criticism of
Darwin's evolution based on natural selection that it is not really scientific since he
had no adequate theory of trait inheritance when he formulated the theory of
evolution by natural selection. It may be true that Darwin's original theory was
incomplete, but exactly how much can one reasonably expect of one man in one
lifetime?
Instead of a criticism of original Darwinian theory, it should be seen as a
strength of the general theory that Mendel's theory of genetics, which provides the
story of how traits get inherited, was developed independently, and it should be
taken as a sign of Darwin's genius that he was able to formulate the theory of
evolution based upon natural selection without a clear notion of how the traits were
passed from one generation to the next.. Mendelian genetics yields specific
empirical predictions that have been verified countless times over the past hundred
years by countless students of biology. So modern genetics is definitely falsifiable
as a scientific theory, but the point that is missed by critics of evolution is that the
particular theory of genetics developed by Mendel was, in fact, predicted by the
theory of evolution based upon natural selection. In order for Darwin's account of
the development of the different species to be correct, the traits selected for by
natural selection must be passed on to the offspring in an unadulterated and
undiluted form. The dependence of evolution by natural selection upon such an
understanding of genetics provides a simple and straightforward mechanism for
falsifying the theory. If fifty or a hundred years into the future, geneticists decide,
on the basis of empirical data from experiments, that they have been mistaken all
of this time about the nature of trait inheritance, and if they come up with a
significantly different theory of genetics from the current theory of genetics based
upon Mendelian genetics (according to which traits are passed from parents to
offspring in an undiluted form as the result of a dominant gene), then that would be
falisifying evidence against evolution by natural selection. This is just one example
of how evolution through natural selection has direct empirical consequences that
may be used to test the theory and that make the theory falsifiable, meaningful, and
scientific. 38
THE SCIENTIFIC METHOD: REVOLUTIONS IN SCIENCE
Undoubtedly, the most significant and influential development in the philosophy
of science in the second half of the twentieth century is the view of the history and
nature of science developed by Thomas Kuhn in The Structure of Scientific
37 For details of the importance of the shift from a "blending" theory of inheritance to Mendel's
theory of genetics, see ibid., pp. 9ff.
38 For other objections to evolution by natural selection and their resolution, see Michael Ruse,
Darwinism Defended, pp. SSand Another major challenge to Darwin's theory of evolution by natural
selection comes from Michael J. Behe, who maintains that evolution by natural selection cannot explain
the irreducible complexity of the biochemical processes that take place on the cellular and molecular
levels of living organisms. His major work, Darwin's Black Box (New York: The Free Press, 1996), is
discussed in Chapter VII.

212

ANALYTIC PHILOSOPHY OF REDGION

Revolutions. 39 Kuhn's account of the nature of science has not only given rise to its
own body of literature in the philosophy of science, but it has also prompted a
number of responses that have traced the importance and the impact of his view of
science upon the relationship between science and religion.
The classical view of science claims that science uses an objective methodology
based upon a careful empirical observation of facts or collection of data and
general or universal natural laws to explain and predict natural phenomena. The
key characterizations of such a classical view of science is that science is objective
and the facts and the data are real. As we saw earlier in this chapter with the case
of Galileo, the basic commitment is to the "irreducible and stubborn facts." Kuhn's
view of the nature of science challenges both of these claims made by the classical
view. Science, Kuhn claims, is paradigm based. 40 Within different fields of
science, according to Kuhn, a certain model for conducting scientific inquiry
gradually begins to dominate the field until it finally wins wide or near-universal
acceptance by the scientists practicing in that particular field. This "paradigm"
prescribes the way in which scientific inquiry is conducted in that field of science,
including the methods that are acceptable, the kinds of questions and answers that
are legitimate, the kind of instrumentation that is used, and even the kind of facts
that are said to exist.
Once such a paradigm becomes established within a particular field of scientific
study, what Kuhn calls "normal" science is conducted by scientists to show how
the paradigm that is in place is successful. The period of "normal science" is
characterized by Kuhn as a period of "puzzle-solving" and is characterized by the
efforts of the practicing scientists in the field, who are committed to the same
paradigm, to continually refine and articulate the prevailing paradigm in greater
and greater detail. 41 For example, in modern medicine one paradigm about diseases
is the view that attributes the cause of diseases to bacteria and viruses, and many
scientists are engaged in research pursuing in greater and greater variety and detail
how different diseases and different bacteria and viruses are causally related.
Another paradigm within modern medicine is based upon the connection between
heredity and disease and the way in which a person's genes may trigger certain
diseases, and other scientists are similarly engaged in research pursuing how
certain diseases and genes may be causally related. According to Kuhn's view of
the nature of science, with each paradigm the theories, research techniques,
methods, instruments, and so forth used by the different scientists will be
determined by the different paradigms to which they are committed. Even what the
scientists take to be "objective" facts are relative to the paradigm that is being
used.
Perhaps the most controversial aspect of Kuhn's view of science is his claim
that paradigms are "incommensurable.,,42 Since the vocabulary, techniques, facts,
and even the criteria for the analysis and evaluation of paradigms are all internal to
the paradigms themselves, one cannot compare paradigms or objectively evaluate
39 Thomas S. Kuhn, The Structure of Scientific Revolutions (Chicago: University of Chicago Press,
1970).
40 Ibid., p. 79.
41 Ibid., pp. lOff.
42 Ibid., pp. 103.

REUGION AND SCIENCE

213

different paradigms on the basis of facts or data. For Kuhn, the process of replacing
one paradigm with another is a "revolutionary shift" that takes place by a sudden
and complete switch from one way of looking at things to a different way of
looking at things. Thus, for Kuhn, a paradigm change is like a Gestalt shift in
perception. 43 The process of adopting a paradigm is not one in which a person
weighs the evidence, compares different paradigms, and then chooses the one that
best fits the data. Kuhn describes the change as an event that is not based upon
reason or evidence but that is very much like a "conversion experience" that "can
only be made on faith.,,44 Such a change of paradigms is a revolution in science.
There has been significant controversy within the philosophy of science over these
claims of Kuhn's and even controversy over what Kuhn himself meant by such
claims. I cannot pursue here the controversies within the philosophy of science
generated by Kuhn's theory of paradigm-based science, so I will focus here simply
upon the impact that this view has had on the relationship between science and
religion. 45
THE SOCIOLOGY OF SCIENCE AND RELIGION
According to the traditional view of science, science inquiry is an intrinsically
rational method whereby theories are tested and confirmed or disconfirmed by the
"hard," "irreducible and stubborn" facts of reality. According to Kuhn's paradigmbased theory of science, facts are no longer completely objective and neutral;
rather, they are shaped by theories that are themselves shaped by paradigms. Since
the process by which paradigms become paradigms is not a matter of science or
even reasoning, according to followers of Kuhn, more attention must be given to
ways in which historical, cultural, and sociological influences have shaped both the
history and methodology of science. To understand science, some critics have
claimed, we need to shift our attention from "science" in the abstract to the
scientific community and to scientists, that is, we need to understand scientists
better and the broader cultural influences that affect the way scientists conduct
science. Claims about the magnitude and importance of such influences are spread
across a broad spectrum, and the more extreme claims that have been made, which
have grown out of a Kuhnian framework regarding the cultural and historical
relativism of science, far exceed anything Kuhn himself ever suggested. The
"postmodern," radical relativists claim that all knowledge is really sociology and
"socially constructed" while more moderate critics defend a core of rationality and
epistemic notions such as truth and reason. 46 The core issues center around the
Ibid., p. 120.
Ibid., p. 151. Some philosophers of science have interpreted this claim very radically, maintaining
that the commitment to a new paradigm must be "irrational" since it is made independently of
confirming evidence and in opposition to the evidence that supports the existing paradigm. Such a
radical view gives rise to what has become known as "the irrationality thesis." See Larry Laudan,
Progress and Its Problems (Berkeley: University of California Press, 1977), pp. 3ff. Other critics have
responded that such extreme interpretations are "gross distortions." See Richard Bernstein, Beyond
Objectivism and Relativism (Philadelphia: University of Pennsylvania Press, 1985), p. 24.
45 For further discussion of Kuhn's notion of paradigm-based science, see James F. Harris, Against
Relativism, pp. 79-85.
46 Since the focus in this volume is upon the analytic tradition, I will not pursue the details of the
postmodernists' attack upon science because of considerations of space. For more details of
43

44

214

ANALYTIC PHILOSOPHY OF REDGION

fundamental question of to what extent sociological, cultural, and historical biases


have influenced the way in which scientists have constructed scientific theories and
performed scientific research. It is one thing to suggest that culture influences the
way the priorities of science are determined, and it is quite another thing to suggest
that the methodology of science is culturally determined or culturally biased. 47 If
science is "totally"socially constructed, then we ought to expect such social
features as race, sex, age, class, and culture to influence the way in which it gets
constructed. Classical descriptions of the methods of science and religion treat
science as "objective" and religion as "subjective." In the eyes of many critics, the
effect of the attacks upon the objectivity of the scientific method from both the
neo-Kuhnians and the postmodernists has been to move science and religion
toward one another in terms of method.
Critiques of religion have progressed along lines that are very similar to those
that have been used to critique science. For example, both liberation theology and
feminist theology are based upon critiques of traditional Judeo-Christian theology
that claim both mainstream science and theology are the products of a white, male,
privileged class. Critics object that both Western science and the dominant,
traditional Judeo-Christian theology were "constructed" by white males who were
socially, politically, and economically privileged and that the structure and content
of both scientific theories and theological theories reflect the racial, sexual, class,
and cultural biases of those white males. 48 Liberation critiques of science and
religion alike insist that both Western science and Christian theology have both
oppressed and suppressed the poor and minorities. Feminist critiques of both
Western science and Christian theology alike insist that both are patriarchal and
have oppressed women.
The different issues related to the nature of science and scientific reasoning are
much too complex to address here. There is still a great deal of dispute and
disagreement among philosophers of science concerning the extent and importance
of such issues for science, and practicing scientists have certainly not "converted"
to Kuhn's paradigm-based view en masse. The primary intent here has been to
suggest how this attempt to undercut the traditional view of the intrinsic rationality
of science (and the corresponding notion of a critical realism and its claims for an

postmodernism and science, see David Hull, Science as a Process: An Evolutionary Account of the
Social and Conceptual Development of Science (Chicago: University of Chicago Press, 1988); Barry
Barnes, Interests and the Growth of Knowledge (London: Routledge and Kegan Paul, 1977); David
Bloor, Knowledge and Social Imagery (London: Routledge and Kegan Paul, 1966); Helen E. Longino,
Science as Social Knowledge: Values and Objectivity in Scientific Inquiry (Princeton, N.J.: Princeton
University Press, 1990).
47 See Harris, ibid., pp.l75ff.
48 See, for example, James H. Cone, God of the Oppressed (New York: Seabury Press, 1975);
Gustavo Gutierrez, A Theology of Liberation (Maryknoll, N.Y.: Orbis Books, 1973); Evelyn Fox Keller,
Reflections on Gender and Science (New Haven, Conn.: Yale University Press, 1985); Sandra Harding,
The Science Question in Feminism (Ithaca, N.Y.: Cornell University Press, 1986); Letty Russell,
Feminist Interpretations of the Bible (Philadelphia: Westminster Press, 1985); and Rosemary Radford
Ruether, New WomanINew Earth (New York: Seabury Press, 1975).

RELIGION AND SCIENCE

215

objectively and independently existing reality) has narrowed the gap between
science and religion in the eyes of many theologians and even some scientists. 49
PAUL FEYERABEND AND THE RELIGION OF SCIENCE
Before Thomas Kuhn came along with his theory of "scientific revolutions,"
when conflicts between science and religion arose the most common responses
from religion were either to rely upon the literal interpretation of the Bible, to
attack the scientific claims as disputed or "bad" science, or to suggest that science
and religion simply have different domains of authority. However, on the basis of
the more radical interpretations of Kuhn, some critics have completely rejected the
scientific method as the privileged method for investigating natural phenomena.
These critics have claimed that science has become "just another religion" and has
no special cognitive or epistemological status. Paul Feyerabend, for example, is a
radical neo-Kuhnian who thinks that Kuhn was too vague and tentative and did not
go far enough in undermining the traditional notion of science. Feyerabend
explicitly draws the implications of this view for the relationship between science
and religion, and he attacks the hitherto sacrosanct scientific method and its claim
to an objective and rational investigation of reality:

The idea of a fixed method, or of a fixed theory of rationality, rests on too naive
a view of man and his social surroundings. To those who look at the rich
material provided by history, and who are not intent on impoverishing it in order
to please their lower instincts, their craving for intellectual security in the form
of clarity, precision, "objectivity," "truth," it will become clear that there is only
one principle that can be defended under all circumstances and in all stages of
human development. It is the principle of anything goes. 50
Feyerabend's radical relativism means that there is no real difference in terms of
the methodology or the intrinsic rationality between science and religion. He
explicitly endorses the view that science is "just another ideology," and he insists
that society must be protected against science since science is given a number of
unfair advantages (such as government protection and funding) in its influence
upon modern society. Ideologies are like "fairytales" for Feyerabend without any
sort of objective or rational grounding, and, like other ideologies before it,
according to Feyerabend, science has now become oppressive and dogmatic. 51
Science and religion should be viewed as competing ideologies for the minds of
human beings, and they should be placed on the same footing for that competition
by imposing a separation of science and the state just as we have imposed a
separation of religion and the state. 52

49 For a Kuhnian approach and a detailed comparison to the different ways in which paradigms
function in science and religion, see Ian Barbour, Religion in an Age of Science: The Gifford Lectures,
1989-91 (San Francisco: Harper and Row, 1990), pp. 51-58.
50 Paul Feyerabend, Against Method (New York: Verso, 1975), p. 19.
" See Paul Feyerabend, "How to Defend Society against Science," in Scientific Revolutions, edited
by Ian Hacking (Oxford: Oxford University Press, 1981), pp. 156ff.
52 Ibid., p. 162.

216

ANALYTIC PHILOSOPHY OF REDGION

In many ways, Feyerabend's critique of science and the results that follow are
similar to the claims made about the relationship between religion and science by
some theists. "Religion is just another way of looking at the world," they may say,
or "religion and science are just different ways of interpreting experience"
(compare the discussion of John Hick in Chapter IV). Such a view is also similar to
that of those philosophers who adopt the later Wittgenstein's notion of languagegames. According to this view, science and religion are just different languagegames with different players and different rules. Some people choose to play one
language-game while others choose to play the other, and there is no good
epistemological reason to prefer playing one game to the other. The
Wittgensteinian approaches of Peter Winch and D. Z. Phillips (see Chapter IV) to
religious language and religious belief, based upon the notion of a language-game,
encourages the view that religion and science are different language-games with
different grammars, which means that the different people engaged in the different
language-games are engaged in completely different pursuits with using different
language with different conventions. 53
Feyerabend's position may be very appealing initially to religious thinkers who
see the relationship between religion and science in modem society as one of
conflict since it appears to "level the playing field" by putting religion and science
on an equal footing. However, Feyerabend's "solution" to the conflict between
science and religion is a double-edged sword since it cuts against religion as much
as it does against science. Feyerabend's criticisms are aimed against science and
the scientific method, but his conclusions apply to religion as well as science since
religion would become "just another ideology" - a fairy tale - with no basis in fact.
Once the nose of the camel of radical relativism gets under the tent, there is no
keeping the rest of the animal out! Scientists, priests, astrologers, and magicians
would be all cut from the same cloth, and none would be grounded in reality any
more than any other. No religious believer would be able to argue against
competing views. The distinction between science and pseudoscience would be
lost as well as the distinction between religion and pseudoreligion. In a world of
such intellectual anarchism where "all bets are off' and "anything goes," might
makes right, and we are back to the Inquisition and the conflict between Galileo
and the Roman Catholic Church where the side with the political, social, and
military power decides what is right.
MUST WE DESTROY RELIGION IN ORDER TO SAVE IT?
The complete separation of science and religion based upon the characterization
of each as simply different ideologies (or different language-games) is not a
position that will hold much appeal to many theists. On the one hand, such a
position seems to elevate religion to equal status with science. On the other hand,
the status of both religion and science turns out to be one that is not very appealing.
Kai Nielsen issues a serious warning to theists:

53

See R. H. Bell, "Wittgenstein and Descriptive Theology," Religious Studies, Vol. 5, 1969, pp. Sff.

RELIGION AND SCIENCE

217

Conceptual relativism, if indeed it is a correct account of what it makes sense to


say and of how in the end we settle ultimate questions, may save religion from
empiricists and "scientific realist" attacks, through establishing that religious
discourses are conceptually autonomous and have their own distinctive order
and rationales with their own distinctive conceptions of reality, rationality,
knowledge, evidence intelligibility and the like. But such a "salvation" may
pave the way for its [religion's] ultimate destruction or utter transformation. 54
Nielsen's warning about "saving" religion by appealing to the extreme
relativism of conceptual relativism is reminiscent of the infamous quote that came
out of the Vietnam War: "We had to destroy the town in order to save it," an
officer was reported to have said. Many theists may well feel similarly about
"saving" religion from science by destroying what they take to be fundamental and
essential to their religious belief. In Feyerabend's world, not only is it impossible
to make any epistemological distinction between religion and science, it is also
impossible to make any such distinction between religion and astrology or between
religion and witchcraft. Such an assessment is not very likely to have much appeal
to either scientists or theists. As A. R. Peacocke notes, scientists could not carryon
in their daily work of scientific research unless they maintained some view such as
a critical or skeptical realism. Neither, it seems, could most religious believers
carryon with the regular routines and religious practices of their "life of faith"
unless they too maintained that their religious beliefs have something to do with an
objective reality. 55 These warnings and reservations are repeated in Chapter II with
the discussion of Wittgensteinian language-games and again in Chapter IX with the
discussion of the notion of religious pluralism.
There is also the additional problem, which we cannot pursue in any detail here,
that Feyerabend's position is helplessly inconsistent and self-refuting, as are the
claims of all radical relativists. 56 He claims, for example, that his radical
"deconstruction" of science is rooted in history and a proper understanding of the
human situation. He thinks that the view that science has developed a fixed method
of inquiry and achieved some special epistemological status is the result of a view
of human beings that is "too naive" and that we can understand this by studying
human history. But, of course, any understanding of human nature and the human
situation must be the result of some sort of scientific study unless that view is
simply completely arbitrary. How would we ever determine that the view of human
beings and the human situation offered by the social sciences is "too naive"except
by doing more social science? Similarly, it is true that any historical study of the
human situation may be "impoverished" by those conducting the study for selfserving reasons, but such a study may also be "embellished" by others conducing
the study for other self-serving reasons. How would we ever determine any of this
except by doing more careful and rigorous historical studies'? To the extent that
Kai Nielsen, Contemporary Critiques of Religion (New York: Macmillan, 1971), p. 103.
A. R. Peacocke, The Sciences and Theology in the Twentieth Century (Notre Dame, Ind.:
University of Notre Dame Press, 1981), pp. xi-xii. Also see the discussion of Wittgensteinian languagegames in Chapter IV.
56 For a more detailed and lengthy attack on the notion of radical relativism, see James F. Harris,
Against Relativism.
54
55

218

ANALYTIC PHILOSOPHY OF RELIGION

Feyerabend is able to offer a reasoned and considered view about why we should
abandon any traditional understanding of human rationality and science, he
undermines the very intellectual anarchism that he advocates. Since he bases his
view upon arbitrary and unfounded claims about the proper way to engage in any
serious study of the human situation, he begs the question.
THE SAPIR-WHORF HYPOTHESIS AND LINGUISTIC RELATIVITY
Claims about the differences between religion and science have also focused
upon the different languages of each.s7 As I have shown in Chapters I and II in
some detail, in the twentieth century, analytic philosophy focused a great deal of
attention upon the importance of language for addressing philosophical problems.
One source of such emphasis upon language comes from linguistics and the social
sciences. Additional doubt about the existence of a single, objective, independent
reality and science's ability to formulate theories to describe and predict
phenomena in such a reality comes from the work of the linguist Edward Sapir and
the anthropologist Benjamin Whorf. One source of some of the radical claims now
made about epistemological relativity involving such basic epistemological notions
as truth and reason and evidence and warrant are based upon linguistic relativity. If
there is a single reality that different human beings in different cultures experience
and then describe in their different languages, then we should expect some rough
agreement among those different languages in terms of the content and substance
of what they say about the world after we have made whatever translations are
necessary and after we have come to whatever understandings are necessary to be
able to compare the different languages.
The Sapir-Whorf hypothesis maintains that language is not simply a vehicle or a
conduit through which experience of reality gets expressed. Rather, Whorf thinks
that human language actually shapes and molds and, to a significant degree,
actually formulates and constitutes what people experience in their different
perceptions of reality. Whorf based his "hypothesis" upon his many years of
studying the languages of different Native Americans - particularly Hopi - in the
second quarter of the twentieth century. He concluded that language is a
"background" upon which experience is built. Through the empirical results of
cross-cultural, anthropological linguistic studies, Whorf claimed, we can determine
that "the background linguistic system (in other words, the grammar) of each
language is not merely a reproducing instrument for voicing ideas, but rather is
itself the shaper of ideas, the program and guide for the individual's mental
activity, for his analysis of impressions, for his synthesis of his mental stock in
trade."s8
It follows from the Sapir-Whorf hypothesis that the influence of language
causes different peoples who speak different kinds of natural languages to organize
and divide the world differently. "We cut nature up," Whorf contends, "organize it
57 See, for example, Ted S. Clements, Science vs. Religion, Chapter 4; Ian Barbour, Religion in an
Age of Science: The Gifford Lectures, 1989-91, pp. 13ff; and Ian Barbour, Issues in Science and
Religion, pp. 121ff.
58 Benjamin Whorf, "Science and Linguistics," Technological Review, Vol. 42, no. 6, 1940.
Reprinted in Language, Thought,and Reality: Selected Writings of Benjamin Whorf, edited by John B.
Carroll (Cambridge: MIT Press, 1956), to which the page numbers here refer, p. 212.

RELIGION AND SCIENCE

219

into concepts, and ascribe significances as we do, largely because we are parties to
an agreement to organize it in this way - an agreement that holds throughout our
speech community and is codified in the patterns of our language."s9 While many
might not deny the claim regarding the influence of language upon human thought
in some areas, the claim that modern science is neither objective nor impartial but
is simply the particularized result of the unique structures of Western modes of
thought resulting from the unique grammar of Indo-European languages has
proven to be both the most significant and the most contentious of results of the
Sapir-Whorf hypothesis. 60 This claim suggests that even the most fundamental
concepts of science, such as time and matter, are not universal but are conceived in
very different ways by people who speak non-Indo-European languages. The Hopi
language, for example, does not have tenses of verbs for expressing anything
resembling the future, present, and past. 61 Thus, according to the Sapir-Whorf
hypothesis, the Hopi's view of the universe is very different from the one with
which "Western" science is concerned. And according to the linguistic relativity
that supposedly follows from the Sapir-Whorf hypothesis, since we are all
determined by the grammar of our native language, "Western" science is simply a
certain way of regarding the world, which is the result of the way in which the
grammar of Indo-European languages filters and organizes experience.
According to the Sapir-Whorf hypothesis, the special, privileged position that
has been accorded science is not justified because the empirical observations upon
which science is based are never completely objective and independent of the
influence of the cognitive and linguistic categories, which, in turn, are never
independent of cultural and societal influences. Such a result of the Sapir-Whorf
hypothesis seems to place science on much more of an equal footing with religion.
However, the other difficulties of the Sapir-Whorf hypothesis notwithstanding,
embracing the hypothesis proves to be another double-edged sword for the theist.
If Whorf is correct, it is not just science that is influenced by the cultural
differences reflected in the different grammars of different natural languages. All
of a person's experiences and cognitions would be similarly influenced including, of course, and perhaps especially, a person's claim to religious
experiences and knowledge concerning religious matters. In other words, the SapirWhorf hypothesis cuts against not only any special, universal, objective,
epistemological status for science, it also cuts against any such claims for religion.
Thus, the hypothesis, if it is sustained and confirmed by the empirical data, would
seem to lead to a radical religious pluralism and would seem to preclude any sort
of comparison or evaluation of different religions. Religious beliefs would thus
become relative to and dependent upon the different culturally determined
grammars of different languages. Different peoples would have different religious
experiences just as people supposedly have different experiences of the world, and,
arguably, it would be impossible to establish any universal or objective viewpoint
Ibid., p. 213.
See ibid., pp. 214 and 216. Also see "Language, Mind, and Reality," in Language, Thought, and
Reality, edited by John B. Carroll, p. 246.
61 Within the Hopi language, the notion of time is infused with the metaphysical distinction between
manifested and manifesting. See Whorf, "An American Indian Model of the Universe," in Language,
Thought, and Reality, p. 59.
59

60

220

ANALYTIC PHILOSOPHY OF RELIGION

that would not be influenced by language in order to compare or evaluate any such
claims. No religious believer could claim any underpinning or support "in reality"
for any particular set of religious beliefs. Whatever support there may be would
simply be the fact that the grammar of one's language provided the categories that
then determined the nature of the religious claims. 62
There are many areas of dispute concerning the Sapir-Whorf hypothesis, but I
will mention only two here very briefly. First, there is much dispute over the
empirical evidence that Whorf and other linguists have offered in support of the
hypothesis. The conflicts among the different natural languages seem to be much
less significant than Whorf claimed. A single example will suffice here: In what
has now become a famous (or infamous) example, Whorf identified the Eskimo
language and what he claimed was the variety of different words the Eskimos use
for snow to illustrate how different languages used by different peoples lead to
different peoples experiencing the world differently and "cutting up" the world
differently by imposing different linguistic categories upon the world. 63 The
exaggerated claims that have been made regarding the different words that the
Eskimos mayor may not use for snow have now become a matter of such notoriety
that the incident is now commonly referred to as the "Great Eskimo Vocabulary
Hoax.,,64
Within linguistics, another major source of criticism concerning the SapirWhorf hypothesis is the work that has been done in generative linguistics or
transformational linguistics by Norm Chomsky. Chomsky has argued for a
distinction between what he calls surface and "deep" grammar and has further
maintained that though different natural languages differ in terms of their
superficial surface grammars, they share a common and universal "deep structure."
This notion of "Cartesian linguistics," as it has come to be known, maintains that
there is an innate linguistic ability that is universal among all humans beings and
that is reflected in the shared deep structures of different natural languages.
Currently, different varieties of Cartesian linguistics are very popular in structural
linguistics. It is beyond the scope of the present project to investigate Cartesian
linguistics here, but if Chomsky and the other structural linguistics are correct, then
Whorf's claims regarding linguistic relativity are undermined at the deep level of
language. Although Chomsky's Cartesian linguistics is still hotly debated,
linguistics does not now give nearly as much credence to the Sapir-Whorf
hypothesis as it did initially; however, the effects of the hypothesis are still widely
felt even if it is not now held in as high regard by linguists as it once was.
THE CHANGING CONTENT OF SCIENCE AND RELIGION
Although it is always dangerous to make such sweeping generalizations, it is
still fairly safe to say that the actual descriptive content of the theories of different
branches of science, that is, the different ways in which scientists actually describe
the world, have changed more in the twentieth century than in any other period in
62

Although, as I discuss in Chapter IX, John Hick argues that there is a single, underlying noumenal

Real.
Benjamin Whorf, "Science and Linguistics," in wnguage, Thought, and Reality, p. 216.
See The Great Eskimo Vocabulary Hoax and Other Irreverent Essays on the Study of wnguage,
edited by Geoffrey K. Pullum (Chicago: University of Chicago Press, 1991).
63

64

RELIGION AND SCIENCE

221

the history of science. Certainly, as we have seen earlier in this chapter, the
seventeenth century with Copernicus, Galileo, and Newton was responsible for
initiating the scientific age; however, except for Copernican astronomy, the
developments of early science, by and large, did not challenge common sense. In
fact, we may say about Newtonian mechanics that it is a formal, scientifically
encapsulated form of a common-sense view of the world. In the twentieth century,
major developments and changes in the way in which the universe is described by
physics - from subatomic physics to astrophysics - challenged both older, classical
scientific views of the universe and common sense. 65 These changes have also had
a significant impact upon the relationship between science and religion. It is
impossible to treat here all, or even a significant number, of the important
developments in physics in the twentieth century, so I will focus upon only
quantum theory and the changes in the ways in which we regard the nature of
things for which quantum theory is responsible. 66 The resulting view of the
subatomic world is one that has proven to be another source of new areas of
conflict between science and religion.
First the science: One aspect of quantum theory that is basic and fundamental
and that is responsible for the rejection of much of classical Newtonian mechanics
is the "dual aspect" that seems necessary to explain phenomena exhibited by
electrons. At the turn of the twentieth century, physicists were divided over
whether electrons were particles or waves with different experimental results
supporting each kind of theory. However, the work of Niels Bohr demonstrated
that there is a reducible duality of electrons according to which electrons are both
particles and waves. 67 In addition, the location of electrons can be determined only
within a particular range determined by a statistical probability. The resulting view
of the subatomic world is literally "mind-boggling." Given quantum theory, we can
no longer diagram or otherwise depict in any univocal fashion what the inside of an
atom is like, and at its most elementary levels, matter is not material. As Ian
Barbour summarizes the impact of quantum theory on our understanding of the
most fundamental "building blocks" of the world,
The atom is inaccessible to direct observation and unimaginable in terms of
sensory qualities; it cannot even be described coherently in terms of classical
concepts such as space, time, and causality. The behavior of the very small is
radically different than that of everyday objects. We can describe by statistical
equations what happens in experiments, but we cannot ascribe familiar classical
attributes consistently to the inhabitants of the atomic world. 68

65 For a detailed treatment of the impact of these changes in physics upon religion, see Ian Barbour,
Religion in an Age of Science: The Gifford Lectures. 1989-91, Chapters 4 and 5.
66 For very insightful and illuminating treatments of relativity theory and "the big bang theory," see
Ian Barbour, ibid., pp. 108ff. and pp. l25ff. I discuss big bang theory and the anthropic principle in
Chapter III in the treatments of the cosmological and teleological arguments for the existence of God.
67 One experiment that yields such results is the well-known "split-screen electron interference test,"
in which a stream of electrons is passed through a screen with two slits in it. The results of this
experiment support both a particle and a wave theory of electrons. For a full description of this
experiment, see Ian Barbour, ibid., pp. 96-97, and James F. Harris, Against Relativism, pp. 30-3 L
68 Ian Barbour, Religion in an Age of Science: The Gifford Lectures. 1989-91, p. 97.

222

ANALYTIC PHILOSOPHY OF RELIGION

In addition to the wave-particle duality of electrons, Werner Heisenberg


determined that there is a theoretical limitation about what we can know about
electrons. Specifically, we can determine the position of an electron, we can
determine the momentum of an electron, but we cannot determine both the location
and momentum of an electron at the same time because by measuring the one we
influence the other. 69 Heisenberg devised his famous principle of uncertainty to
explain this theoretical limitation upon the information that we can expect to ever
have about the subatomic world.
THE PRINCIPLE OF COMPLEMENT ARlTY
There are several different ways in which the very different world of quantum
theory has impacted upon classical scientific views with the net result of
undermining the classical view of matter as basically materialistic and
deterministic. Just as examples, I will concentrate here upon only those that are
illustrated by the principle of complementarity and the principle of indeterminacy.
The principle of complementarity: When Bohr first began to try and understand
the consequences of his own quantum theory, along with Heisenberg's notion of
uncertainty, the only way in which he could make sense of it was by suggesting
that the subatomic phenomena are actually partially constituted by the method used
to detect and measure them. The measuring devices and the experimental
arrangement under which the measurement takes place actually affect the "reality"
that is being measured. This view means that on the subatomic level the actions of
the scientists are not simply ones of observing an objective reality but are actually
part of making that reality what it is. Bohr introduced the principle of
complementarity, which says that a subatomic system has no condition or features
independently of a particular experimental arrangement or a particular mode of
measurement. The old, classical distinction between subjective, observing subjects
and objective, observed facts cannot be maintained on the subatomic level. In a
very real sense, scientists influence the observed "facts" of subatomic particles
simply by observing them?O It turns out that the irreducible and stubborn facts are
neither irreducible nor stubborn on the subatomic level. Different things are what
they are as a result of the interaction of the experiencing subject and the
experienced object. According to the principle of complementarity, a thing on the
subatomic level must always be defined in relation to the mode of investigation
according to which it is observed and described. Neither the particle description
nor the wave description is a "true" description of the subatomic world, but
together, the two different descriptions complement each other to give us a
complete description.
Obviously, the principle of complementarity may hold great appeal to the
religious believer if it is applied more generally beyond the subatomic level, and
both scientists and religious believers alike have suggested that such a wider
60 For a very clear and simple explanation of how and why this is true, see John Hedley Brooke,
Science and Religion: Some Historical Perspectives (Cambridge: Cambridge University Press. 1991), p.
328.
70 There are many accounts of this phenomenon. For a particularly accessible account in a
fascinating book about superstring theory, see Brian Greene, The Elegant Universe (New York: Vintage
Books, 1999), pp. I 12ff.

RELIGION AND SCIENCE

223

application of the notion of complementarity is possible.71 One may suggest, to


borrow from one of the examples discussed at length earlier in this chapter, that
one may give both an evolutionary account and a creationist's account of the origin
of man and that both would be equally legitimate since there would then be two
different phenomena, each of which would be relative to its own account.

TAOISM AND PHYSICS


The principle of complementarity seems to lie behind the comparisons that
Fritjof Capra draws between quantum theory and Taoism.72 Capra draws several
parallels between both quantum theory and relativity theory and the various
religious traditions of the East. One such comparison involves "the unity of all
things." One of the essential characteristics of Eastern thought, Capra emphasizes,
is a form of metaphysical monism according to which everything is One and
everything is connected and interdependent as "inseparable parts of the cosmic
whole.,,73 This "unity" of all things is also a recurring theme, Capra claims, of the
"new" physics, that is, quantum theory and relativity theory, since the new science
has revealed "that the constituents of matter and the basic phenomena involving
them are all interconnected, interrelated and interdependent; that they cannot be
understood as isolated entities, but only as integrated parts of the whole.,,74 A
second major point of comparison between modem physics and Eastern thought
for Capra is the role that human beings play in the construction of reality. The new
physics reveals that all of our concepts are fundamentally limited and do not really
capture "features of reality" but are simply "constructs of the mind" that then are
projected onto reality.75
Critics have objected that Capra emphasizes the similarities between modern
physics and Eastern religions by focusing upon a few key concepts or phrases that
are used very differently within their respective contexts. It is true, for example,
that Bohr had insisted upon a "holistic" approach where quantum phenomena are
treated as part of a system that includes the experimental design. However, exactly
how this claim would "translate" out of the context of physics is not clear at all.
The landscape in which Capra maps modern physics onto Eastern thought is
painted with a very broad brush. For example, as Ian Barbour has observed, the
phrase "the unity of things" has very different meanings in physics and Eastern
religions. In the latter case, the unity is a metaphysical, undifferentiated unity
whereas, in the case of modem physics, the unity is highly structured and
differentiated by various principles and laws. 76 There are also significant
differences in the ways in which space and time are viewed by modern physics and
Eastern religions. Barbour is certainly correct to point out that although the notion

71 See ibid., pp. 331ff., and Ian Barbour, Religion in an Age of Science: The Gifford
Lectures,
1989-91, pp. 114ff.
72 Fritjof Capra, The Tao of Physics (Boulder, Colo.: Shambhala Publications, 1975).
73 Ibid., p. 130. This characterization is very similar to the absolute monism that dominated much of
the English-speaking philosophical world in the late nineteenth century (see Chapter I).
74 Ibid., p. 131.
75 Ibid., pp. 161-63.
76 Ian Barbour, Religion in an Age of Science, p. 119.

224

ANALYTIC PHILOSOPHY OF RELIGION

of space within modern physics is no longer the same notion as the absolute space
of Newtonian mechanics, it is not illusory, as much of Eastern thought maintains. 77
The principle of complementarity was intended to allow competing and
mutually exclusive descriptions of the same phenomena within the same
experimental context and "on the same level." Two different descriptions from the
two different sources could easily be experimentally tied to the same phenomena,
and then the results could collectively "complement" each other to give a complete
picture of the phenomena. It is impossible to determine in the comparisons of
modern physics with Eastern religions if the two are really even trying to describe
the same phenomena. Also, how could we come to regard two such disparate
"methods" as providing complementary information about the same phenomena
when one claim is based on a theory supported by highly controlled and
circumscribed scientific experiments while the other is a highly speculative and
noncircumscribed metaphysics based on a mystical form of meditation?
An even more fundamental and serious problem results from extrapolating from
the way in which an observer influences the observed phenomena on the subatomic
level to the way in which an observer may influence the observed phenomena on
the level of room-sized objects. The "weirdness" that results from quantum
mechanics is a weirdness that emerges when we try and understand subatomic
level phenomena, that is, it is only beyond a certain supermicroscopic level that
quantum effects occur. And that level is unimaginably small. Planck's constant,
which governs the limits of quantum effects, sets the Planck limit of such effects at
10-33 centimeter (or a millionth of a billionth of a billionth of a billionth of a
centimeter).78 There is then no reasonable way to extrapolate from the quantum
eflects of electrons to the way in which events occur in the physical world in which
human beings live.
THE HEISENBERG PRINCIPLE OF UNCERTAINTY
According to the Heisenberg principle of uncertainty (also known as the
Heisenberg indeterminacy thesis), there are serious theoretical limitations upon
what we can ever hope to know about the subatomic world. Since it is impossible
to ever determine both the momentum and the location of an electron at any given
moment, the complete description of any given quantum state must be made in
terms of probabilities. On the basis of any description of any quantum state, the
best we can do is to make a probabilistic prediction about some description of
some future quantum state, for example, if we measure and determine the location
of an electron, then we can only predict with some statistical probability where the
electron will be when we are able to locate it. In other words, since the description
of any quantum state is necessarily given in terms of statistical probabilities, then
any existing quantum state must be just one of many different possible quantum
77 For a more detailed comparison of modem physics and Eastern religions on space and time, see
Ian Barbour, ibid., pp. 119ff.
78 This figure comes from Brian Greene, The Elegant Universe, p. 130. Pay attention to his footnote
7. A great advantage of Greene's book is the analogies he uses to help lay readers conceptualize the
extremes of quantum mechanics, relativity. and string theory. In this case, he suggests that if an atom is
magnified to the size of the known universe, the Planck length (within the atom) would grow to the size
of a tree.

RELIGION AND SCIENCE

225

states that might have existed given the past states that preceded it. Heisenberg
thought, at least late in his life, that this uncertainty was objective and
metaphysical. It is not simply that certain matters are uncertain because of human
ignorance or because of the present limitations of scientific inquiry, but nature
itself is "open" and the future is undetermined. 79 Heisenberg thought that this
metaphysical indeterminacy meant that potentiality and chance and novelty
actually exist in nature. Some theists have seized upon this opening for a way of
explaining how God can be understood as active in the world. For example,
Willard Pollard maintains that the indeterminacy that exists on the subatomic level
provides the opportunity for God to influence the world in a "natural way" without violating any natural laws. He suggests that Providence can be understood
as God being active in the selection of one quantum state over another possible
one. 80
Other theists have also attempted to incorporate the fundamental aspects of
quantum theory into a theistic understanding of the world by extrapolating from the
indeterminacy of the subatomic world to the macro world of human beings. For
example, Christopher Mooney claims that since the world is a unity and since God
would not allow one set of laws and principles to operate on the microscopic,
subatomic world and a different set on the macrolevel of human beings, then we
have to assume that the same sort of uncertainty exists for human beings as does
for quanta. 81
The macroworld is also dominated by "irregular behavior," Mooney declares,
and regular behavior such as planetary motion is the exception instead of the rule.
In support of this claim, Mooney cites the effect of chaos theory, which, he says,
has "placed complex systems beyond the grip of any mechanistic determinism.,,82
The indeterminacy in the macro world and the inherent potentiality in nature
reaches "full flower" with human freedom with which human beings can
"determine future goals" and "initiate responsible moral choice." In other words,
"there would be a certain degree of indeterminacy in human freedom that is not
unlike that of the quantum world.,,83
Mooney's move from the indeterminacy of the subatomic level of quantum
theory to the level of human freedom represents an extension of scientific
principles beyond the limits of their reasonable application. In the first place, the
effects of quantum theory, including indeterminacy, are confined to the
supermicroscopic world by Planck's limit. Also, indeterminacy on the quantum
level is sheer chance. It is simply the lack of determinism. The chance is explained
in terms of random statistical probability. Whatever the human freedom is that
allows human beings to "determine future goals" and "initiate responsible moral

79 For a thorough discussion of uncertainty and religion, see Ian Barbour, Religion in an Age of
Science: The Gifford Lectures, 1989-91, pp. lO1f.
80 See Willard Pollard, Chance and Providence (New York: Charles Scribner's Sons, 1958).
81 Christopher F. Mooney, Jr., Theology and Scientific Knowledge (Notre Dame, Ind.: University of
Notre Dame Press, 1996), pp. 99-100.
82 Ibid., pp. 100-01.
83 Ibid., p. 102.

226

ANALYTIC PHILOSOPHY OF REUGION

choice," it must be understood as involving something more than just sheer


randomness. 84
Others have made the rather implausible leap from the indeterminacy of
quantum theory on the subatomic level to chaos theory on the macrolevel. Perhaps
there is indeterminacy in the macro world, but quantum theory cannot justify such a
claim, and chaos theory certainly does not "place complex systems beyond the grip
of any mechanistic determinism." On the contrary, properly understood, chaos
theory is based completely on the overdeterminism of complex systems. It is only
because of the intricacy and complexity of the causal connections of "nonlinear"
systems, such as those found in weather patterns and the flow of liquids, that we
are unable to predict their future states. It is not that the weather is undetermined so
that it becomes metaphysically irregular; it is rather that the weather is so
completely determined in such intricate ways that we can never hope to decipher
all of the variables and all of the countless permutations that may result. The only
indeterminacy that results from chaos theory is epistemological indeterminacy.85
While the changing content of modem science may offer opportunities for
"closing the gap" between science and religion, theists should avoid highly
speculative interpretations and widely divergent extrapolations. Undoubtedly, the
scientific view of the universe has changed significantly in the twentieth century.
In terms of the very broad metaphysical differences that may translate into
different theological views of the universe and man, we can say that the world is no
longer materialistic. Energy and process have replaced matter as the fundamental
metaphysical "stuff' of the universe. There is inherent randomness and chance in
the world on the subatomic level, and the world is now regarded as an integrated
and interrelated whole rather than simply a collection of discrete, individual
objects. The world of modem physics is, in some ways, a more friendly world for
religion, and the ways of addressing the nature of God, human nature, and the
nature of the relationship between God and human beings have changed. All of this
suggests that if we are not to put new wine in old wine skins, we need to provide
some new theological wine skins for the new scientific wine.
THE INTEGRATION OF RELIGION AND SCIENCE
Among the different ways in which scientists and theologians have regarded the
relationship between science and religion, the attempt at the "integration" of the
two is one of the more interesting and promising. I will follow Ian Barbour here by
distinguishing three different ways in which the integration of science and religion
has been envisioned and attempted by different thinkers. 86 First, natural theology,
which is responsible for the arguments for the existence of God, borrows from
84 Ian Barbour makes a similar point by claiming that human freedom emerges at a certain level of
complexity. See Ian Barbour, Religion in an Age of Science: The Gifford Lectures, 1989-91, p. 117.
85 One may argue, as does William Alston, that there may be additional, nonnaturalistic causes that
must also be counted, but this seems like an argument from ignorance to the effect that there may be
other things about the universe that we do not know and that we cannot understand but for which we
must make allowance. For an interdisciplinary approach to many of the issues raised by chaos theory for
the philosophy of religion, see Chaos and Complexity: Scientific Perspectives on Divine Action, edited
by Robert John Russell, Nancey Murphy, and Arthur R. Peacocke (Vatican City State: Vatican
Observatory Publications, 1995).
86 Ian Barbour, Religion in an Age of Science: The Gifford Lectures, 1989-91, pp.23-24.

RELIGION AND SCIENCE

227

science the results of scientific inquiry for use in its arguments and theories. For
example, contemporary versions of both the teleological argument from design and
the cosmological argument from causality are considerably more sophisticated than
earlier versions because of the evidence that science has supplied about the nature
of the universe, such as the use of the anthropic principle (see Chapter III).
Secondly, a theology of nature is constructed from traditional sources for theology
such as sacred scriptures, church history, and religious experiences; however, the
resulting theology is a dynamic one that is constantly being changed, adjusted, and
reinterpreted by being "brought up against" and applied to the world as revealed by
scientific inquiry. Thirdly, integration has been attempted by a few thinkers on a
much more fundamental level through a "systematic synthesis" through which both
modern science and religion are reinterpreted in terms of a more general
metaphysics. I turn now to an examination of the attempts at integration of science
and religion by those who have attempted to construct a theology and a general
metaphysics that are based upon the findings of modern science.
A THEOLOGY OF NATURE
Arthur Peacocke has attempted to integrate religion and science by treating the
two as simply two different but "interacting approaches to the same reality.,,87 In
order to consider such an integration, Peacocke rejects the radical, "strong"
program of sociology that would reduce all knowledge to socially constructed
knowledge in favor of a form of critical realism. As I have indicated earlier in this
chapter, the relativism that results from radical relativism leaves science and
religion as two competing and unrelated ideologies. Peacocke also rejects the "neoorthodoxy" of Karl Barth, who insisted that theology is a completely separate
enterprise from science. For Barth, theology is completely autonomous and
depends upon the revelation of God through Jesus Christ rather than any empirical
knowledge based upon this world. In fact, for neo-orthodoxy, empirical
knowledge, based upon scientific inquiry, actually impedes one's search for
knowledge of God, which must thus come from a personal encounter with God. In
contrast, by developing a "theology of nature," Peacocke maintains that traditional
theology can be made more meaningful and theology can be shown to be related to
the world in which we human beings live by incorporating the information about
human beings and the world in which we live that is the result of the investigations
by the natural sciences. 88 Peacocke describes his theology of nature as follows:
Since the aim of the critical-realist theology is to articulate intellectually and to
formulate, by means of metaphor and model, experiences of God, then it
behooves such a theology to take seriously the critical-realist perspective of the
sciences on the natural, including the human, world. For on that theology's own
presuppositions, God himself has given the world the kind of being it has and it
must be in some respects, to be ascertained, revelatory of God's nature and
87 A. R. Peacocke, The Sciences and Theology in the Twentieth Century (Notre Dame, Ind.:
University of Notre Dame Press, 1981), p. xiv. Peacocke argues for his rejection of the sociology of
knowledge and his adoption of a form of critical realism in his Imitations of Reality (Notre Dame, Ind.:
University of Notre Dame, 1984), pp. 19ff.
88 A. R. Peacocke, The Sciences and Theology in the Twentieth Century, p. xiii.

228

ANALYTlC PHILOSOPHY OF RELIGION

purposes. So theology should seek to be at least consonant with scientific


perspecti ves on the natural world. 89
Because of the changes that took place in science in the twentieth century,
science and religion now find themselves, according to Peacocke, "inextricably
interlocked with each other" in what is really an enterprise that is common to both
science and religion - the attempt to provide an explanation for the intelligibility of
the universe and its meaning (including the nature and meaning of human
existence).9o In the end, he thinks both science and religion as well as humanity in
general will all be the better for the integration of science and religion. 91
Peacocke carries out the program for his theology of nature across a wide
spectrum of different theological issues, and it is only possible here to sketch the
very general outline of how he tries to reinterpret the various themes of Christian
theology within a scientific view of the world. In order for theology to be
meaningful in a scientific age to people for whom the scientific point of view is
basically and thoroughly ingrained, Peacocke maintains that theology must take
place in the world described by science and speak to the world described by
science.92 For example, in order to understand the "classical" theory of creation,
the notion that God created the universe at a particular moment of time must be
abandoned. In order to accommodate the understanding of the universe that
modern science has provided as one that is continually producing "new emergent
forms of matter,,,93 any understanding of God's relationship with the world that is
"static" must be abandoned. Peacocke thinks that modem science has demonstrated
that the universe is dynamic and that things in the universe are still being "worked
out." Thus, God's relationship with the world must be understood as a complex
relation where God sustains and maintains the world through a continuing activity
and interaction with the world. Such a view moves our understanding of God away
from the transcendent and toward the immanent nature of God. 94 God must create
and sustain the universe through the universe and as a part of the universe - not as
a distinct, separate, transcendent being. "There is no room for any deus ex machina
or 'God of the gaps,'" Peacocke concludes. 95
Any theological account of human nature must also take into consideration what
modern science has revealed about human beings. Human beings are a part of the
physical world, but we are also products of an evolutionary process that has
produced new "levels" of organization with emergent properties that are not
reducible to strictly physical-chemical processes. Thus, on the one hand, a human
being is a physical creature "describable in terms of the physics and chemistry of
his constituent atoms and molecules," but on the other hand, human beings have
Ibid., p. 21.
Ibid., p. 5-6. Peacocke explains that he is really concerned with the natural sciences as opposed to
the social sciences and with theology as opposed to religion.
91 Philip Clayton urges a similar approach. See his God and Contemporary Science (Grand Rapids,
Mich.: William B. Eerdman's, 1997).
92 One should keep in mind here that since Peacocke is a critical realist, the description of the world
provided by science cannot be taken strictly literally.
93 Peacocke, Science and the Christian Experiment (London: Oxford University Press, 1971), p. 123.
94 Ibid., Chapter 5.
95 Ibid., p. 169.
89

90

RELIGION AND SCIENCE

229

features that have been emphasized by theology, including self-consciousness,


intelligence, and the capacity for a relationship with God. These features are
explained as the emergent properties that are produced at a certain level of
complexity in man's evolutionary development.% So the theological view of man
can be interpreted as an understanding of the product of the slow, continuing
process of evolutionary creation. Even the diVinity of Christ, Jesus as God
Incarnate, according to Peacocke, takes on new meaning if Christ, God as man, is
understood as the end point, the ultimate end, of the process of "cosmic evolution"
that is the expression of God's will for human kind. 97 The example of Christ, the
divine become human, serves as an example of what we all may become - the
human become divine - as the result of the continuing evolution of mankind.
Peacocke's theology of nature may well prove to be the next step in a natural
progression for natural theology in the future. In fact, the difference between the
two is already blurring. Traditionally, natural theology has borrowed from the
empirical sciences, in terms of both method and content, to gather the evidence
upon which to construct rational arguments for the existence of God. As I have
shown in Chapter III, newer forms of those arguments have borrowed heavily upon
developments in the sciences. For example, the anthropic principle has
significantly influenced recent versions of the teleological argument. In the next
chapter, I will show how the incorporation of probability theory has been
responsible for a different form of the evidential problem of evil. The key to the
difference between natural theology and a theology of nature is the extent to which
theists are willing to reformulate the content of theological claims on the basis of
new scientific discoveries and theories. The impact of evolution by natural
selection is a good place to see where scientific theory has caused changes in the
nature of the claim about creation and design in the universe. For Peacocke and
other theists who are willing to adjust their theology in the face of the
developments in modern science, it is now common to explain God's responsibility
for design in terms of the very general natural laws through which evolution takes
place instead of through the immutable essences of different species created in an
initial, single act of creation by God. A theology of nature holds the promise of
"the best