Book Review

Gödel’s Theorem:
An Incomplete Guide
to Its Use and Abuse
Reviewed by Panu Raatikainen

of the Notices was dedicated to Gödel and con-

Gödel’s Theorem: An Incomplete Guide to Its tains many informative articles on Gödel and on
Use and Abuse the incompleteness result in particular by some of
Torkel Franzén the leading experts in mathematical logic.
A K Peters, Wellesley, MA The incompleteness theorem is discussed in
$24.95, paperback, 2005 countless popular science books, and several are
182 pages, ISBN 1-56881-238-8 even devoted to Gödel’s result. Unfortunately,
such books typically show more enthusiasm than
Apparently no mathematical theorem has aroused competence and tend to be loaded with inadequa-
as much interest outside mathematics as Kurt cies and errors.1 There is thus still demand for a
Gödel’s celebrated incompleteness result pub- knowledgeable and reliable exposition of the in-
lished in 1931. It is invoked not only by mathe- completeness result. Torkel Franzén aims to fulfill
maticians, logicians, and philosophers but also this need with his book, and he succeeds outstand-
by physicists, theologians, literary critics, archi- ingly. He not only provides the reader with adequate
understanding of the content of Gödel’s theorem
tects, and others. Some eminent physicists have
and how it is proved but also evaluates critically
interpreted it as showing that “the theory of every-
and thoroughly many applications and misapplica-
thing” demanded by other physicists is impossible
tions of the theorem and corrects various common
to achieve. It is sometimes claimed to prove the
misconceptions.
existence of God or of free will, the necessary
In addition to obvious nonsense, there are among
incompleteness of the Bible or of the U.S. Constitu-
the nonmathematical ideas inspired by Gödel’s the-
tion, or the impossibility of genuine knowledge in
orem many that by no means represent postmod-
mathematics—just to mention a few of the many
ernist excesses, but rather come to mind natural-
alleged applications (see also [9]).
ly to many people with very different backgrounds
Gödel is unquestionably among the greatest
when they think about the theorem. It is especial-
mathematicians of our times, and he made many ly such naturally occurring misunderstandings that
important contributions to mathematical logic Franzén intends to correct.
and other fields. But it is undoubtedly his incom- The book does not pay much attention to Gödel’s
pleteness result that made his reputation. In the life and other scientific achievements—only three
year 2006 the one-hundredth anniversary of the pages are devoted to them (bits of history are also
birth of Gödel was celebrated all over the world given along the way). Gödel’s strange person and his
with various conferences. The April 2006 issue eventful life are certainly interesting and deserve
attention, but fortunately there already exists an ex-
Panu Raatikainen is an Academy Research Fellow in the cellent biography, Logical Dilemmas: The Life and
Academy of Finland and a docent of theoretical philos-
ophy at the University of Helsinki. His email address is 1
For some apt critiques, see the following reviews in the
[email protected]. Notices : [1], [3], [5].

380 Notices of the AMS Volume 54, Number 3

Work of Kurt Gödel by John W. Dawson ([2]; see [1] even in ZFC. Proving them would thus require a
for a review). formal system that incorporates methods going be-
yond even ZFC. There is thus a sense in which such
The Incompleteness Theorems truths are not provable using today’s “ordinary”
In order to understand Gödel’s theorem, one must mathematical methods and axioms or cannot be
first explain the key concepts occurring in it: “for- proved in a way that mathematicians would today
mal system”, “consistency”, and “completeness”. regard as unproblematic and conclusive.
Very roughly, a formal system is a system of axioms Gödel’s second theorem concerns the limits of
equipped with rules of reasoning which allow one consistency proofs:
to generate new theorems. The set of axioms must Second incompleteness theorem. For any consis-
be finite or at least decidable; i.e., there must be tent system F within which a certain amount of el-
an algorithm that enables one to mechanically de- ementary arithmetic can be carried out, the consis-
cide whether a given statement is an axiom or not tency of F cannot be proved in F itself.
(otherwise, one might stipulate, e.g., taking all true It is important to note that this result, like the
statements of arithmetic as axioms; such a theory first incompleteness theorem, is a theorem about
is trivially complete but highly abstract and totally formal provability (which is always relative to some
useless in practice). formal system). It does not say anything about
A formal system is consistent if there is no state- whether, for a particular theory T, the statement “T
ment for which the statement itself and its negation is consistent” can be proved in the sense of being
are both derivable in the system. Only consistent shown to be true by a conclusive argument or by an
systems are interesting in this context, for it is an argument acceptable by mathematicians. For many
elementary fact of logic that in an inconsistent for- theories, this is perfectly possible.
mal system every statement is derivable, and con- Franzén describes in some detail but very infor-
sequently such a system is trivially complete. And mally the ideas of the proofs of the incompleteness
a formal system is complete if for every statement theorems. Later, he also explains, again quite infor-
of the language of the system, either the statement mally, the basic notions and results of the theory of
or its negation can be derived (i.e., proved) in the computability, essential for proper understanding
system. of the incompleteness results. Franzén also clari-
Gödel proved two different though related in- fies the relation of the incompleteness theorem to
completeness theorems, usually called the first another result of Gödel which is often misleadingly
incompleteness theorem and the second incom- called “the completeness theorem” and to the ex-
pleteness theorem. “Gödel’s theorem” is some- istence of so-called nonstandard models. The book
times used to refer to the conjunction of these ends with an appendix that gives a slightly more
two and sometimes to either—usually the first— formal yet still easily understandable explanation
separately. Accommodating an improvement due of the incompleteness theorems. In all these cases,
to J. Barkley Rosser in 1936, the first theorem can be Franzén has done an admirable job. These sections
stated as follows: provide an excellent ground for evaluating various
First incompleteness theorem. Any consistent alleged consequences of Gödel’s theorem, to which
formal system F within which a certain amount of we now turn.
elementary arithmetic can be carried out is incom-
plete; there are statements of the language of F Antimechanism, Faith, and Skepticism
which can neither be proved nor disproved in F. There is a popular view according to which Gödel’s
A common misunderstanding is to interpret theorem shows that the human mind cannot be any
Gödel’s first theorem as showing that there are sort of computing machine but infinitely surpasses
truths that cannot be proved. This is, as Franzén any machine. The alleged justification goes like
points out, incorrect, for the incompleteness the- this: For any formal system, which can be viewed
orem does not deal with unprovability in any as a computing machine generating theorems,
absolute sense, but only unprovability in some Gödel’s proof exhibits an unprovable sentence
particular axiom system. And for any statement S (often called the Gödel’s sentence of the system).
unprovable in a particular formal system F, there We humans can know the truth of this sentence,
are trivially other formal systems in which S is whereas the formal system or its corresponding
provable. On the other hand, there is the extremely machine cannot. There is thus—so the argument
powerful standard axiom system of set theory (the goes—something noncomputable about human
so-called Zermelo-Fraenkel set theory, which is thinking, perhaps even some irreducibly spiritual,
denoted as ZF, or, with the axiom of choice, ZFC), nonmaterial component of the human mind. Such
which is more than sufficient for the derivation of antimechanist conclusions have been drawn from
all ordinary mathematics. Now there are, by Gödel’s Gödel’s theorem, for example, by a philosopher,
theorem, arithmetical truths that are not provable J. R. Lucas [4], and more recently by a distinguished

March 2007 Notices of the AMS 381

 mathematical physicist, Roger Penrose [6], [7], and of mind and TOE. But there are some more specif-
these conclusions seem to enjoy some popularity. ically theological appeals to the incompleteness
The idea is apparently quite natural and attractive, theorem. Some of these are simply preposterous,
for it gets reinvented again and again. and others at best are based on analogies. Some-
Nevertheless, such conclusions are not justi- times it is suggested that Gödel’s theorem shows
fied on the basis of the incompleteness theorem. that the only possible way of avowing an unprov-
Franzén explains clearly why this is so: in general, able truth is faith. But, first, Gödel did not exhibit
we have no idea whether or not the Gödel sentence any absolutely unprovable truths, only relative
of an arbitrary system is true. What we can know is ones; and, second, if we have, on the basis of math-
only that the Gödel sentence of a system is true if ematical reasoning, absolutely no idea whether a
and only if the system is consistent, and this much given highly complex formal system is consistent
is provable in the system itself. But in general we or not, it is quite unclear how Christian faith (or
have no way of seeing whether a given system is anything else) could help.
consistent or not. Later in the book Franzén ex-
Gödel’s theorem is often thought to support
plores in some detail variants and ramifications of
some form of skepticism with regard to mathemat-
the Gödelian antimechanist argument and shows
ics: it is contended that we cannot, strictly speaking,
them all wanting.
prove anything or that the consistency of our fun-
Franzén then moves on to discuss various
damental theories (such as ZFC) is shown to be
attempts to apply Gödel’s theorem outside mathe-
matics. It has been claimed that the incompleteness doubtful. Franzén argues against such claims that
theorem demonstrates the incompleteness of the nothing in Gödel’s theorem in any way contradicts
Bible, the U.S. Constitution, and Ayn Rand’s phi- the view that we have absolutely certain knowledge
losophy of objectivism. He points out that such about the truth of the axioms of the system and,
suggestions ignore the essential condition that the consequently, of their consistency. We don’t need
system must be capable of formalizing a certain Gödel’s theorem to tell us that we must adopt some
amount of arithmetic. None of the mentioned “sys- basic principles without proof. If we have no doubts
tems” have anything to do with arithmetic. Even about the consistency of, say, ZFC, there is nothing
worse, they are nothing like a formal system: they in the second incompleteness theorem to give rise
do not have an exactly specified formal language, to any such doubts. And if we do have doubts about
a set of axioms, or rules of inference. Therefore, the consistency of ZFC, we have no reason to believe
Gödel’s theorem simply is not applicable in such that a consistency proof of ZFC given in ZFC itself
contexts. would do anything to remove those doubts.
More reasonable have been attempts to apply Franzén also devotes a brief chapter to the vari-
the incompleteness theorem to physics. The hypo- ants of incompleteness results arising from the
thetical “theory of everything” (TOE) is sometimes so-called Algorithmic Information Theory, or the
taken to be an ideal of theoretical physics. However, theory of Kolmogorov complexity, and especially
such eminent physicists as Freeman Dyson and the various philosophical interpretations of these
Stephen Hawking have invoked Gödel’s theorem to results by Gregory Chaitin (one of the founders of
suggest that there is no such theory of everything this theory). For example, Chaitin claims that his
to be had. Now it seems more reasonable to assume results not only explain Gödel’s incompleteness
that a formalization of theoretical physics would theorem but also are the ultimate, or the strongest
be the subject of the incompleteness theorem by possible, incompleteness results. Franzén first
incorporating an arithmetical component. Never- explains these results and then shows that such
theless, Franzén adds, Gödel’s theorem tells us only claims are in no way justified by mathematical facts
that there is an incompleteness in the arithmetical
(see also [8]).
component of the theory. Whether a physical the-
ory is complete when considered as a description
Concluding Remarks
of the physical world is not something that the
incompleteness theorem tells us anything about. This is a marvelous book. It is both highly compe-
Franzén also discusses various theological con- tent and yet enjoyably readable. At some points
clusions drawn from Gödel’s theorem. Abstracts there are even glimpses of humor, as when Franzén
from the Bibliography of Christianity and Mathe- declares in the preface: “For any remaining in-
matics declare, for example, that Gödel’s theorem stances of incompleteness or inconsistency in
demonstrates that physicists will never be able to the book, I consider myself entirely blameless,
formulate a theory of physical reality that is final since after all, Gödel proved that any book on the
or that the human mind is more than just a logical incompleteness theorem must be incomplete or
machine. Such theological appeals to Gödel’s theo- inconsistent. Well, maybe not” (p. ix). At last there
rem only recycle the above-discussed and deficient is available a book that one can wholehearted-
Gödelian arguments against the mechanist theory ly recommend for anyone interested in Gödel’s

382 Notices of the AMS Volume 54, Number 3

incompleteness theorem—one of the most excit-
ing and wide-ranging achievements of scientific
thought ever.

References
[1] Martin Davis, Review of Logical Dilemmas: The
Life and Work of Kurt Gödel (by John W. Dawson
Jr.) and Gödel: A Life of Logic (by John L. Casti
and Werner DePauli), Notices of the AMS 48 (2001),
807–813.
[2] John W. Dawson Jr., Logical Dilemmas: The Life
and Work of Kurt Gödel, A K Peters, Wellesley, MA,
1997.
[3] Juliette Kennedy, Review of Incompleteness: The
Proof and Paradox of Kurt Gödel (by Rebecca
Goldstein), Notices of the AMS 53 (2006), 448–455.
[4] J. R. Lucas, Minds, machines, and Gödel, Philosophy
36 (1962), 112–137.
[5] Timothy McCarthy, Review of Gödel’s Proof (by
Ernest Nagel and James R. Newman), Notices of the
AMS 51 (2004), 333–337.
[6] Roger Penrose, The Emperor’s New Mind: Con-
cerning Computers, Minds, and the Laws of Physics,
Oxford University Press, New York, 1989.
[7] , Shadows of the Mind: A Search for the
Missing Science of Consciousness, Oxford University
Press, New York, 1994.
[8] Panu Raatikainen, Review of The Unknowable
and Exploring Randomness (by G. Chaitin), Notices
of the AMS 48 (2001), 992–996.
[9] , On the philosophical relevance of Gödel’s
incompleteness theorems, Revue Internationale de
Philosophie 59 (2005), 513–534.

March 2007 Notices of the AMS 383