===== PAGE 1 =====

ik JOURNAL OF LIBERTARIAN STUDIES
VOLUME 19, NO. 1 (WINTER 2005): 79-95

REJOINDER TO CAPLAN
ON BAYESIAN ECONOMICS

WALTER BLOCK

To THE BEST OF my knowledge, there were three Methodenstreits (or
debates in the literature over the proper method to use in econom-
ics) and a handful of “mini-Methodenstreits” in which Austrians took
part.

The first Methodenstreit was the most famous, and established
the Austrian School of economics. It took place in the latter two
decades of the nineteenth century, with the historicists Gustav von
Schmoller, Lujo Brentano, Adolf Wagner, and Werner Sombart on
one side, and the Austrians Carl Menger, Eugen von Boshm-Bawerk,
Friedrich von Wieser, and Ludwig von Mises on the other.1

The second started in the mid-twentieth century with the publi-
cation of Friedman's classic article on methodology, followed by a
plethora of others reacting to it. But here, unlike the first, where the
two sides were evenly matched at least in terms of numbers, only a
few Austrians participated.2

Walter Block is Harold E. Wirth Eminent Scholar Endowed Chair and Pro-
fessor of Economics, Loyola University New Orleans. The author wishes to
acknowledge the help given to him by his Loyola University New Orleans
colleague Bill Barnett. The usual caveats apply. (wblock@loyno.edu.)

1See Ludwig von Mises, The Historical Setting of the Austrian School of Eco-
nomics (Auburn, Ala.: Ludwig von Mises Institute, 1984); “Methodenstreit,”
www. wikipedia.org; and Walter Block, “Reply to Caplan on Austrian Eco-
nomic Methodology” (unpublished manuscript; department of economics,
Loyola University).

2Milton Friedman, “The Methodology of Positive Economics,” in Essays in
Positive Economics (Chicago: University of Chicago Press, 1953), pp. 3-43. A
conference on this seminal article took place in 2003, but no Austrians were
invited. See http://www?2.eur.nl/fw/philecon/Friedman53.html for a list
of participants. The Austrian perspective included Glenn Fox, “Economics
as Prediction,” Chap. 5 in Reason and Reality in the Methodologies of Econom-
ics: An Introduction (Cheltenham, U.K.: Edward Elgar, 1997); and Samuel
Bostaph, “Epistemological Foundations of Methodological Conflict in Eco-
nomics: The Case of the Nineteenth Century Methodenstreit” (Ph.D. diss.,
Southern Illinois University at Carbondale, 1976), pp. 119-39.

79


===== PAGE 2 =====

80 — JOURNAL OF LIBERTARIAN STUDIES 19, NO. 1 (WINTER 2005)

There were also what might be regarded as several mini-Metho-
denstreits, consisting of only two or three entries; a neoclassical argu-
ment and an Austrian reply with no follow up, or an exchange
between only two participants.3

We are now in the midst of the third such Methodenstreit, which
began with an article by Bryan Caplan, and has subsequently
included articles by William Barnett, Walter Block, Guido Hiils-
mann, and Frank van Dun, as well as replies by Caplan.# Note that
this Methodenstreit is almost the opposite of the second in terms of
the proportion of participation of Austrians and neoclassicals. That
one, launched by Friedman, included only two Austrians, along

30ne such example involved Robert Nozick and Walter Block. See Robert
Nozick, “On Austrian Methodology,” Synthese 36 (1977): 353-92; and Walter
Block, “On Robert Nozick’s ‘On Austrian Methodology,” Inquiry 23, no. 4
(Fall 1980): 397-444. For a second example, see G.J. Schuller, Review of
“Human Action by Ludwig von Mises,” American Economic Review 40, no. 3
(June 1950); Murray N. Rothbard, “Mises’s ‘Human Action”: Comment,”
American Economic Review 41, no. 1 (March 1951); G.J. Schuller, “Rejoinder,”
American Economic Review 41, no. 1 (March 1951); and Murray N. Rothbard,
“Praxeology: Reply to Mr. Schuller,” American Economic Review 41, no. 4
(December 1951).

For more on the debates between Austrians and their intellectual oppo-
nents, see Walter Block, Christopher Westley, and Alex Padilla, “The Case
for Refereed Journals Devoted to Austrian Economics Revisited (Are there
any objective criteria for determining truth in economics?),” (unpublished
manuscript; department of economics, Loyola University).

4Bryan Caplan, “The Austrian Search for Realistic Foundations,” Southern
Economic Journal 65, no. 4 (April 1999): 823-38.

Austrian responses to Caplan include William Barnett, II, “Contra
Caplan” (unpublished manuscript; department of economics, Loyola Uni-
versity New Orleans); Walter Block, “Austrian Theorizing, Recalling the
Foundations: Reply to Caplan,” Quarterly Journal of Austrian Economics 2, no.
4 (Winter 1999): 21-39; Walter Block, “Realism: Austrian vs. Neoclassical
Economics, Reply to Caplan,” Quarterly Journal of Austrian Economics 6, no.
3 (Fall 2003): 63-76; Block, “Reply to Caplan on Austrian Economic Method-
ology”; Jorg Guido Hiilsmann, “Economic Science and Neoclassicism,”
Quarterly Journal of Austrian Economics 2, no. 4 (Winter 1999): 3-20; and
Frank van Dun, “Bayesianism and Austrian Apriorism,” http://
allserv.rug.ac.be/-frvandun/Texts / Articles /Bayesianism-Austrian.pdf.

Caplan, in turn, has offered replies to his critics. See Bryan Caplan, “Prob-
ability, Common Sense, and Realism: A Reply to Hiilsmann and Block,”
Quarterly Journal of Austrian Economics 2, no. 4 (Summer 2001): 69-86; and
Bryan Caplan, “Probability and the Synthetic A Priori: A Reply to Block,”
Quarterly Journal of Austrian Economics 6, no. 3 (Fall 2003): 77-83.


===== PAGE 3 =====

REJOINDER TO CAPLAN ON BAYESIAN ECONOMICS — 81

with a myriad of neoclassicals. The present one consists of Caplan,
the only neoclassical, and four Austrians.

In the present article, it is my goal to critically comment on
Caplan’s most recent argument. The next section is devoted to Prob-
ability and Science, followed by a section on Synthetic a priori and
Common Sense, and then a section on Caplan’s Bayesian Cure.

PROBABILITY AND SCIENCE

I fear that a serious misunderstanding has sprung up between
Caplan and myself on the relationship between probability and the
synthetic a priori. I relish the opportunity to try to set matters
straight.

Throughout his most recent article, Caplan focuses upon my
supposed “concession” regarding synthetic a priori statements and
probabilities. He writes: “It is a major concession for Block to admit
that synthetic a priori propositions can have a low probability and
empirical propositions can have a high probability.” Caplan also
labors under the false assumption that I maintain that empirical
claims are “unscientific.”>

In my view, the only sense in which the truth of synthetic a pri-
ori statements can be uncertain is in the psychological, not the logical,
sense. Caplan, unfortunately, interprets me in precisely the wrong
way on this matter.

Previously, I said that we could be less certain of synthetic a
priori statements that involved long chains of reasoning, or convo-
luted deductions, and more sure of those that involved more direct
logic, or simpler mathematical operations. For example, we can in
this sense be more sure that the square root of 9 is 3 than that the
square root of 2,187,441 is 1,479. Or, to use my example, the claim
that a triangle has 3 sides is relatively simple, that it has 180
degrees in all three of its angles is moderately difficult to under-
stand, and the Pythagorean Theorem is more complex yet.” But
here, we are limited to discussing psychological states of people
being confronted with these various truths. In terms of their pure
logic, they are all of a piece; that is, to deny any one of these con-
tentions is to commit an internal contradiction, a strict no-no in
logic.

5Caplan, “Probability and the Synthetic A Priori,” p. 77.
6See Block, “Realism.”

7Ibid., pp. 71, 72. To be clear, I assume that the base of these numbers is ten,
and I posit Euclidian, not Riemannian, Lobachevskian, or some other type
of geometry.


===== PAGE 4 =====

82 — JOURNAL OF LIBERTARIAN STUDIES 19, NO. 1 (WINTER 2005)

These claims are not testable; they are not falsifiable. There is no
possible state of the world that would render them false. Yet, Caplan
claims that since some of the more complex statements in this cate-
gory can be psychologically confusing, their logical status has some-
how been converted into mere empirical claims. Surely, this is a cat-
egory mistake on his part.

On the other hand, based on my “concession” that, in the empir-
ical world, too, there can be more or less certain claims with high
and low probability of being correct, he concludes that we can be
more confident of highly secure contingent claims than about diffi-
cult to see but necessary ones, for example, that we can rely more
heavily on the statement “Elephants are heavy creatures” than on
the Pythagorean theorem. However, say what you will about the
former, its denial involves no self-contradiction, while to assert the
contrary of the latter most certainly does. This is a category mistake
on stilts.

I am tempted to say that I regret my Table 2 since it has sown so
much confusion.8 But, upon reconsideration, I do not, for it opens up
avenues of exploration that otherwise would not have arisen. Using
it shows that Caplan does not understand methodology.? Testing the
Pythagorean theorem indeed. Why did I construct this table? Solely
as an attempt to bridge what I saw as a yawning gap between
Caplan’s argument and my own. Caplan continually brought prob-
ability into play,10 a concept I saw as beside the point.

Very well, I decided, if Caplan cannot argue in the absence of
probability, let us by all means discuss it, but in a way that shows its
irrelevance to the issues under contention. This is what I meant
when I wrote “we can accommodate his concerns” in my presenta-
tion of Table 2.11 In other words, I was utilizing the “arguendo”
method in an attempt at a reductio ad absurdum. 1 was saying, in
effect, that the entire notion of probability leads us astray from the
discussion involving the distinction between contingent and neces-
sary truths. However, I tried to make clear, if you must discuss it,
here is a way to do so that need not deflect us from our main debate.

8Block, “Realism,” p. 71. In Caplan, “Probability and the Synthetic A Priori,”
all page number citations to Block, “Realism,” are incorrect, apparently due
to an editing error. In the present article, whenever I cite Caplan’s article
with reference to my prior article, I substitute the correct pagination.

9Van Dun, “Bayesianism and Austrian A Priorism,” has similar misunder-
standings over this table.

10See Caplan, “Probability, Common Sense, and Realism.”
Block, “Realism,” p. 71.


===== PAGE 5 =====

REJOINDER TO CAPLAN ON BAYESIAN ECONOMICS — 83

Rothbard uses the arguendo gambit against Blum and Kalven,
who write: “An equitable apportioning of sacrifice requires inflict-
ing equal hurt on each taxpayer.”12 Rothbard responds:

To a thoroughgoing individualist and libertarian, this basic goal of
the sacrifice theorists reveals the utter absurdity of their position.
Instead of worrying about what constitutes “equal hurt,” why
inflict any hurt at all? Why tolerate an institution that represents
only pain and injury, and then try to find some sort of “equitable
means” of spreading it around? The entire concept of “just and
equal” in suffering is an absurdity. Setting aside this point for the
moment. . 13

Rothbard then continues with his analysis of the writings of
these two authors, ignoring the devastating criticism he has just lev-
eled against them. Does this mean Rothbard has changed his mind,
and now accepts the validity of equal hurt? Does this mean, to utilize
Caplan’s terminology, that Rothbard has “conceded” to Blum and
Kalven the propriety of hurting people, so long as the hurt is spread
around equally, or equitably? If you believe this, then you are ready
to seriously entertain the notion that I “conceded” to Caplan that we
can be logically more secure in believing well established empirical
findings in economics than we can regarding less certain synthetic a
priori statements, “less certain” only because they involve long
chains of reasoning. This is psychological, not logical, uncertainty.

I claim that Gary Becker and other supposed neoclassicals are
really praxeologists under the skin, based on their reaction to the
Card-Krueger “finding” that minimum wage legislation has very
different effects than previously supposed.l4 Caplan explains the
Becker et al. reaction not in terms of praxeology, but on the basis of
Bayesian probability theory.15 I reject this analysis on the ground
that it is based on yet another confusion between logic and psychol-
ogy. Yes, for all I know, there may be any number of economists such
as Becker, livid with the Card-Krueger article, and determined to

12walter J. Blum, and Harry J. Kalven, Jr., “The Uneasy Case for Progressive
Taxation,” University of Chicago Law Review 19 (1952): 417-520.

13Murray N. Rothbard, “The Uneasy Case for Degressive Taxation,” Quar-
terly Journal of Austrian Economics 4, no. 1 (Spring 2001): 52; emphasis added.

14Gee David Card and Alan B. Krueger, “Minimum Wages and Employ-
ment: A Case Study of the Fast-Food Industry in New Jersey and Pennsyl-
vania,” American Economic Review 84, no. 4 (September 1994): 772-93; cf.
Gary Becker, “It’s Simple: Hike the Minimum Wage, and You Put People
Out of Work,” Business Week Magazine (March 6, 1995).

15Caplan, “Probability and the Synthetic A Priori.”


===== PAGE 6 =====

84 — JOURNAL OF LIBERTARIAN STUDIES 19, NO. 1 (WINTER 2005)

show its flaws, who are motivated by something along the lines of
Caplan’s Bayesian analysis. But that is mere psychology. I am making
an entirely different point. As I see it, I am not motivation-monger-
ing. I am claiming, instead, that the critics, supposedly neoclassical,
of the Card-Krueger argument logically could not take the stance
they did without a healthy dose of praxeological insight.

On the Bayesian explanation, in contrast, the fact that minimum
wage legislation leads to unemployment for unskilled workers is so
well established that neoclassicists approach any deviation from this
finding with great skepticism. But how, on this interpretation, did
any such knowledge get established initially on the Bayesian
account? It logically could not have been so established based on
this system, because it requires a previous body of knowledge on the
basis of which to examine new claims. The response, “it’s turtles all
the way down” seems to be appropriate here.

With this introduction to Caplanian methodology, we are ready
to consider chapter and verse. What follows is an analysis of
Caplan’s statements. I shall quote him extensively, and then
respond. I am led to this mode of presentation to forestall future mis-
understandings between us. My hope is that if we cannot in the end
agree, then we should at least achieve real disagreement, as opposed
to presently passing each other as ships in the night.

Caplan writes:

It is a major concession for Block to admit that synthetic a priori
propositions can have a low probability and empirical proposi-
tions can have a high probability.16

It is no “concession” at all. It is merely evidence of an attempt to
try to communicate more clearly with Caplan. Yes, “Elephants are
heavy” has a high probability of being correct at any given time, and
we are more likely to make mistakes in long involved sets of math-
ematical proofs, but these statements occur in two entirely separate
universes of discourse, the former empirical, the latter a priori.
Caplan is wrong to draw from this the implication he does. The “low
probability” attached to a specific synthetic a priori statement is not
with regard to its truth or its being correct, but rather relevant to the
odds of arriving at it; e.g., it is true that, modulo ten, the square root
of 2,187,441 is 1,479, but the odds that most people could figure this
out on their own is, I suspect, small, especially given the large num-
ber of innumerate people in the world. There are certain truths for
which the probability is small that people will ferret them out; never-
theless they are indubitable truths.

Caplan writes:

lé1bid., p. 77.


===== PAGE 7 =====

REJOINDER TO CAPLAN ON BAYESIAN ECONOMICS — 85

[1}f synthetic a priori claims, like empirical ones, vary in probabil-

ity, then Block has no reason to single out the latter as unscien-

tific.17

But I do no such thing. Indeed, a word search of the electronic
version of my article does not reveal that I even used the word
“unscientific,”18 let alone asserted that empirical claims cannot pass
muster under this criterion. Indeed, my view is the very opposite. As
I see it, empirical claims most certainly can be scientific, in the phys-
ical sciences. It is only in fields like economics, mathematics, and
logic, but not in the physical sciences, that we can do better, e.g.,
attain synthetic a priori knowledge.

Caplan writes:

Take for example “Most socialists would be offended by Defending
the Undefendable.” It is hardly true a priori, yet it is plainly true. How
then can Block label such a claim scientifically empty? Indeed, how
can he call it scientifically inferior to a low-probability synthetic a
priori claim?19

There are more mistakes in this short quotation than you can
shake a stick at. I do not at all label the claim that socialists would be
offended by my book Defending the Undefendable “scientifically
empty.” Rather, this empirical claim seems very likely. It is not “sci-
entifically” empty; rather, it is praxeologically empty, since its denial
involves no self-contradiction. This claim about the sensibilities of
socialists is a high-probability empirical statement. We cannot prop-
erly say it is scientifically inferior to a “low-probability” (again, only
in the psychological sense) synthetic a priori claim. If the purpose of
science is to discern truth, then every synthetic a priori statement is
superior to every empirical statement. But, “you can’t add apples
and oranges.” A synthetic a priori statement is true (if not, then it is
not truly a synthetic a priori statement); however, an empirical state-
ment is never true in this sense. Therefore, one really can’t compare
them regarding scientific value.

Caplan writes:

[Block] has already foreclosed one route: equating “scientific
knowledge” with certainty. After all, he admits that many synthetic

171bid., p. 77.

181t would help to achieve real disagreement, and thereby promote rational
discussion, if Caplan would explicitly cite me and quote me when he attrib-
utes specific claims to me.

19Caplan, “Probability and the Synthetic A Priori,” p. 78, discussing Walter
Block, Defending the Undefendable (New York: Fox and Wilkes, 1991); empha-
sis in the original.


===== PAGE 8 =====

86 — JOURNAL OF LIBERTARIAN STUDIES 19, NO. 1 (WINTER 2005)

a priori claims—like empirical ones—fall short of certainty. If “Most
socialists would be offended by Defending the Undefendable” is
unscientific due to lack of certainty, so are claims about the minimum
wage.20

But the lack of certainty between empirical claims such as those
regarding my book, and synthetic a priori ones concerning the mini-
mum wage, occupy entirely different philosophical categories. The
former are uncertain due to our lack of information about the world.
We think it likely that most socialists would be shocked and
offended by Defending the Undefendable, but heck, there might be a
few of them out there with a sense of humor, or who may even enjoy
a brutal attack on their position.

The uncertainty concerning the minimum wage or the
Pythagorean theorem is entirely of a different type. Here, we are not
concerned only with a state of the world, but also with necessary con-
ditions which cannot be denied but on pain of self-contradiction.

The ways we prove these two contentions are also very different.
We can examine socialists and their feelings, or elephants and their
weight, but we can never establish knowledge that is certain in the
logical sense. After all, the next socialist down the pike, or the next
elephant, can be an exception to the rule. Not so for synthetic a pri-
ori statements, not at all.

Caplan writes:

By admitting that the empirical can be more probable than the syn-
thetic a priori, my critic has distanced himself from the weakest
parts of his first reply.2!

Not a bit of it. I fear I shall have to continue to wallow in my
“weakness.” Yes, “elephants are heavy” is an empirical claim that
most would accept. In that sense, it is a very likely empirical claim.
On the other hand, never in a million years would the average per-
son be able to prove Godel’s Theorem. In the psychological sense,
this insight is unlikely, or improbable. But this is truly comparing
apples and oranges. It is as if Caplan is saying that the best weight
lifter on the planet can bench press triple his own weight; therefore,
such a person should be a great golfer or chess player. The one has
nothing to do with the other.

Caplan writes:

20Caplan, “Probability and the Synthetic A Priori,” p. 79; emphasis in the
original.

211bid.


===== PAGE 9 =====

REJOINDER TO CAPLAN ON BAYESIAN ECONOMICS — 87

Many of Block’s examples of the synthetic a priori are poorly cho-
sen. Demand curves . . . do not necessarily slope downward. Roth-

bard’s own discussion of backward-bending curves admits as
much.22

Unhappily, Caplan leaves out an important part of my statement,
which reads as follows: “Demand curves are downward sloping
(e.g., when forced to give up one unit of a stock of a good, the eco-
nomic actor will give up the least valuable of them).” Would Caplan
be so rash as to claim that when forced to give up one unit of a stock
of a good, the economic actor will give up other than the least valu-
able of them? If not, he then necessarily acquiesces with my con-
tention. Of course, if income is allowed to vary as we move along a
demand curve, it will not necessarily be downward sloping
throughout its entire length. We all know about backward-bending
demand curves and Giffen goods. This dispute is not about that.
Rather, it concerns creating straw men and then attacking them.

SYNTHETIC A PRIORI AND COMMON SENSE

In a continuation of his claim that my synthetic a priori examples are
“poorly chosen,” Caplan writes:

Nor is it necessarily true that raising the minimum wage reduces
employment, ceteris paribus. It does not have to do so under
monopsony.

This ought to put paid to the notion, if ever entertained, that
Caplan is some sort of subterranean Austrian, at least of the Roth-
bardian variety. After all, the concepts of monopsony and monopoly,
when applied to the free-enterprise system, are dead from the neck
up. The only legitimate denotations of these words are when the
government grants an exclusive privilege, whether to a marketing
board buyer (monopsony) or to a seller (monopoly). This is not the
time nor place to go into the neoclassical fallacies concerning these
concepts. Suffice it to say that no one has ever succeeded in demon-
strating a perfectly competitive price or quantity with which to con-
trast the so-called market monopoly or monopsony price or quan-
tity. Indeed, the entire concept of free-market monopoly or

22Ibid. Caplan’s refusal to directly quote me is apparently not personal;
here, he fails to cite Rothbard. I always tell my students to document any-
thing about which any reasonable person might take issue. Evidently, rules
of documentation are not uniform. Why should the reader have to go
through the trouble of finding Rothbard’s actual discussion of this issue?

231bid., p. 79.


===== PAGE 10 =====

88 — JOURNAL OF LIBERTARIAN STUDIES 19, NO. 1 (WINTER 2005)

monopsony is a veritable contradiction in terms. The entire model is
logically incoherent.24

Caplan, further continuing his claim that my cases in point of
synthetic a priori are misconceived, writes:

Or to take a more fanciful counterexample, imagine that employ-
ers have been hypnotized to believe that the minimum wage
declines every time Congress increases it. Then raising the mini-
mum wage would increase employment instead of decreasing it, at
least until bankruptcy set in.2

Unfortunately for Caplan, that last bit, “at least until bankruptcy
set in,” gives away the entire point: Such craziness cannot last in the
long run, even given these rather heroic assumptions. The piper must
be paid: Whenever the minimum wage rises, no matter what anyone
thinks, employment will eventually have to be reduced. But no one
ever claimed any more for the minimum wage than that. Certainly,
no competent economist ever asserted that this legislation would
immediately create unemployment. In the very next microsecond,
right after a rise in the mandated level but before human action can
possibly react to the change, employment would not fall at all. It
takes purposive choice on the part of entrepreneurs for any such
result to ensue. Nor are we interested, praxeologically, in how long
this market process will take, only that it necessarily must occur. It
would appear that Caplan, despite his own best efforts to the con-
trary, agrees with this Austrian insight. He, in effect, demonstrates
this with his “at least until bankruptcy set in” concession.

Caplan “quibbles” with two other of my synthetic a priori exam-
ples. He states:

My objections to “There are mutual benefits to trade in the ex ante
sense,” and “Man acts (to create a world more to his liking in the
future than one which would arise but for his action),” are only
quibbles. To be strictly true, these assertions would have to allow
for indifference: “There are mutual nonlosses to trade in the ex ante
sense,” and “Man acts (to create a world as or more to his liking in
the future than one which would arise but for his action).”26

24For a critique of neoclassical monopoly theory, see William Anderson,
Walter Block, Thomas J. DiLorenzo, Ilana Mercer, Leon Snyman, and
Christopher Westley, “The Microsoft Corporation in Collision with Antitrust
Law,” Journal of Social, Political, and Economic Studies 26, no. 1 (Winter 2001):
287-302; Thomas J. DiLorenzo, “The Myth of Natural Monopoly,” Review of
Austrian Economics 9, no. 2 (1997): 43-58; and Murray N. Rothbard, Man,
Economy, and State (Los Angeles: Nash, 1970).

25Caplan, “Probability and the Synthetic A Priori,” p. 79.
26Tbid., p. 79.


===== PAGE 11 =====

REJOINDER TO CAPLAN ON BAYESIAN ECONOMICS — 89

I don't call these “quibbles” at all. Rather, my intellectual oppo-
nent is re-arguing the issue of indifference, without citing, let alone
mentioning, my previous objections to his views. So let me again ask
my learned friend: Why would anyone bestir himself to act if he
only expected to make a non-loss; e.g., not improve his subsequent
position, after the act, even in the ex ante sense? Why, pray tell,
would man act (to create a world as much to his liking in the future
as one which would arise but for his action)? Why not do something
else? Why not just turn over and go back to sleep, rather than waste
energy achieving a state of affairs that is no better than the one he
already has? Not only does praxeology cry out for an answer to these
questions, so does common sense.

Caplan writes:

[Olnce you admit, as Block does, that synthetic a priori claims vary
in probability, then even the synthetic a priori can be empirically
tested. . . . Itis true, as Austrians have long insisted, that there is no
point empirically testing synthetic a priori claims known with cer-
tainty. Empirically testing less solid synthetic a priori propositions,
in contrast, can be fruitful 27

There is so much wrong here it is hard to know where to start.
First of all, there is no such thing as a “less solid” synthetic a priori
proposition, and a “more solid” synthetic a priori proposition. All
synthetic a priori propositions are equally “solid” from a logical per-
spective. From a psychological viewpoint, there are some that are
harder to grasp, and others that are easier, but, to the extent that a
synthetic a priori proposition is a synthetic a priori proposition, its
“solidity” from a logical point of view is equal. It may be harder to
understand that 1,479 is the square root of 2,187,441 than that the
square root of 9 is 3, but these two claims are equally valid or
“solid.”

Second, not only is it not “fruitful” to empirically test “less
solid” synthetic a priori propositions, it is not even possible. Indeed,
it is logically incoherent. Austrians do not reject empirical testing per
se. Econometrics has a role to play with regard to economic state-
ments that are not axiomatic, for example, concerning the magni-
tude of a particular demand elasticity. Nor, even, is it denied among
most praxeologists that regression analysis can be properly used
with regard to axioms, i.e., synthetic a priori statements. Here,
though, the role is not to use empirical research to test the axioms;
rather, if anything, it is to have axioms test the empirical research.
Or, better yet, econometric analysis can illustrate the underlying
economic realities, as with, for example, minimum wages or rent

271bid., p. 80.


===== PAGE 12 =====

90 — JOURNAL OF LIBERTARIAN STUDIES 19, NO. 1 (WINTER 2005)

controls. Thus, I am claiming that Caplan has failed to fully reckon
with Mises’s elaboration upon these distinctions.?8

Caplan, however, offers perhaps the strangest analysis ever
made in the history of methodology:

Imagine that we were uncertain about the Pythagorean Theorem.
We could spend more time checking out the steps of the proof. On
the other hand, we might measure a wide variety of triangles to
verify that c2 = a2 + b2. Both efforts would rationally tend to increase
our confidence in the theorem, even though neither is foolproof. If we
found a clear empirical counterexample, we would conclude that
either the premises or the reasoning behind the theorem were
incorrect, despite our inability to pinpoint them. . . . If less-than-
certain empirical evidence contradicts an absolutely certain syn-
thetic a priori claim, then, as the Austrians advise, we should ignore
the empirics. The opposite holds if indubitable empirics contradict
a less-than-certain synthetic a priori claim.2?

Let us consider the errors here one by one. First, Caplan sug-
gests that we measure a wide variety of triangles to verify that c2 =
a? + b2. Suppose we find a triangle for which this is not true. But this
means it is not a right triangle at all (obviously, the triangle does not
contain a 90° angle)! The point is, we already have a criterion for a
shape being a right triangle: namely, that 2 = a2 + b2. Let us make it
even simpler. Triangles have three (straight) sides, and the internal
angles add up to 180 degrees. How do we “test” this, pray tell? I can
just see Caplan sending his graduate students out into “the field” to
capture some wild triangles, corral them, see if they have three
straight sides, and measure their internal angles. But shapes will be
picked on the basis of having these characteristics in the first place!
If this is a test, it is a fudged “test.”

Second, what, precisely, can be meant by the phrase “indu-
bitable empirics”? History, according to some wag, is “just one
damn thing after another.” So is it with empirical research. This dis-
cipline can never establish cause and effect, only, at best, constant
conjunction or correlation. There was a lot of economic freedom
(e.g., very little government regulation of the economy) in the U.S.
in the eighteenth century, and the people were relatively poor. There
is much less economic freedom in the U.S. in the twenty-first cen-
tury, and the people are far wealthier, at least in terms of material
goods. This, I suppose, is an “indubitable” empirical finding, if any-
thing is, but it hardly establishes that government regulation

28See Ludwig von Mises, Theory and History (New Haven, Conn.: Yale Uni-
versity Press, 1957).

29Caplan, “Probability and the Synthetic A Priori,” p. 80; emphasis in origi-
nal.


===== PAGE 13 =====

REJOINDER TO CAPLAN ON BAYESIAN ECONOMICS — 91

enriches a country. This is because empirical economics, or, better
yet, “indubitable empirics,” cannot distinguish whether we are now
richer because of or despite government interference with economic
liberty.

Third, there being no such thing as “indubitable empirics” in
economics, and no such thing as anything less than “an absolutely
certain synthetic a priori claim,” at least in the logical albeit not in the
psychological sense, there can never be a case for rejecting any syn-
thetic a priori proposition in favor of an empirical one, no matter
how well established the latter, and complicated the former. Perhaps
the real problem is with some specific synthetic a priori statements.
If Caplan thinks he knows of some that are refuted by empirical evi-
dence, then he should be able to point out the logical misstep(s) or
faulty premise(s) that misled praxeologists into thinking such state-
ments to be synthetic a priori statements when in fact they involved
logical contradictions. This he has not done. This he has not even
tried to do. This he has not recognized the need for, if his viewpoint
is to even approach coherency.

BAYESIAN CURE

How, according to Caplan, are we best to explain the outraged reac-
tion of the neoclassicals to the Card-Krueger claim concerning the
employment effects of minimum wages? We are to utilize Bayesian
insights. Caplan writes:

The Bayesian model states, in essence, that rational people do not
revise their whole worldview every time a new data point
emerges. Rather, they marginally update their initial views as facts
comes [sic] in.30

But this appears to be a very poor explanation on its very face.
Even Caplan concedes: “Why then, as Block insightfully asks, do
economists react so violently to empirical evidence against the con-
ventional view of the minimum wage’s effect?”31 If they were really
Bayesians, e.g., acting according to this model, these economists
would marginally shift their perspective on the minimum wage law.
They might be slightly surprised—this was just one new “data
point”—but they would certainly not be angry. However, econo-
mists were affronted, as well they should be, by the Card-Krueger
“findings,” as even Caplan admits.

Caplan writes: “Becker . . . had a ‘strong prior’: he was almost,
but not quite, sure of the standard conclusion.”32 Caplan makes it

301bid., p. 81.
311bid., p. 80; emphasis added.
321bid., p. 81.


===== PAGE 14 =====

92 — JOURNAL OF LIBERTARIAN STUDIES 19, NO. 1 (WINTER 2005)

sound as if Becker, who, presumably, had no strong feelings about
the elasticity of bananas in Ohio in 1955, but thought it was -1, now
confronts a new “data point” indicating it was really -1.1. Does
Becker come out of the starting blocks with fire in his eye over this
new “data point”? He does not. He calmly, and very slightly, adjusts
his presumptions on this matter.

But that is not at all what Becker and his colleagues did when
these “two Princeton economists” rubbed the profession’s collective
noses in their nonsensical claim. Rather, they were breathing fire on
this matter, again, as Caplan himself acknowledges: “economists
react so violently to empirical evidence against the conventional view
of the minimum wage’s effect.”33

According to Caplan’s numerical example, Becker, upon reading
Card-Krueger, reduces the probability he places on the typical econ-
omists’ view of minimum wage employment effects from 98 percent
to 95.5 percent. However, Becker reading Card-Krueger was more
like waving a red flag in front of the bull, as even Caplan notes.34 It
is very difficult to reconcile this with a slight movement in favor of
that pernicious and fallacious position.

Caplan writes: “Bayes’s Rule shows that we should automati-
cally be more skeptical of empirical evidence in favor of an improb-
able conclusion.”35 However, if this were all there were to the mat-
ter, how would we ever initially establish truth in economics? The
point is, to say the best possible for Bayesianism, we need some
independent criterion of truth before it sets in, and this the system
itself cannot furnish.

Caplan lists three possible alternative explanations for the feroc-
ity of the reaction to the Card-Krueger analysis on the part of the
neoclassicals:

Explanation #1: Neoclassical economists are covert praxeologists,
though they will not admit it;

Explanation #2: Neoclassical economists are intellectually dishon-
est dogmatists posing as empirical scientists; [or]

Explanation #3: Neoclassical economists are Bayesians with some
strong priors.30

Let us focus on 1 and 3, the only two possible explanations that
interest me here. Caplan correctly attributes belief in the first to me,

33Ibid., pp. 81, 80; emphasis added.
341bid., p. 82.

351bid., p. 82.

361bid., pp. 80-81.


===== PAGE 15 =====

REJOINDER TO CAPLAN ON BAYESIAN ECONOMICS — 93

and in the third to himself. The problem I have with his treatment of
these two alternatives is that he believes implicitly that they are
incompatible with one another. Caplan’s view is that if he can prove
explanation 3 correct, he disproves, or at least disparages, explana-
tion 1. However, I cannot see my way clear to agreeing with this. It
seems to me that the two alternatives are compatible with one
another. Both can be true, together. Neither contradicts the other.

Now, I have given reason just above to indicate why I disagree
with explanation 3. But let us assume, arguendo, that 1 accept?”
Caplan’s Bayesian explanation. Why does that prove that explana-
tion 1 is false? The reason that explanations 1 and 3 are not neces-
sarily contrary to one another is that they occupy different philo-
sophical categories. Explanation 1 is a claim that neoclassicals are
really praxeologists, a logical category. Explanation 3 is a claim that
neoclassicals form their beliefs on the basis of Bayesian considera-
tions, and belongs to the realm of psychology. Caplan, in order to
undermine explanation 1, will have to do better than merely to suc-
cessfully defend explanation 3—which he has not done anyway. He
will have to come to grips, directly, with explanation 1, something he
has not even attempted to do, let alone accomplished.

Basic economic theory tells us, with a single voice, that rent con-
trol causes housing shortages. However, several of my econometric
regression results failed to bear this out. Caplan offers a Bayesian
explanation, according to which these

anomalous results were almost three times as likely to be flawed as
his regular results. His dissertation committee justifiably greeted
the former with extra skepticism.38

But why? Without economic theory to guide us, it is as likely that
rent control causes housing shortages as that it solves them. True,
most empirical results support the shortage hypothesis, but truth
can never be demonstrated based on majority vote. Moreover,
according to Caplan’s own analysis, new “data points” are sup-
posed to shift received opinion in their direction.3? E.g., he is on
record as stating that, due to Card-Krueger, Becker's trust in the
view that minimum wages cause unemployment shifted from 98
percent to 95.5 percent. Why wasn’t my dissertation committee

37Lest there be any misunderstanding here, I reiterate my rejection of expla-
nation 3. I am only assuming this proposition to be correct for the sake of
argument in an attempt at a reductio. This should not be interpreted as a
“concession” to Caplan’s Bayesian analysis.

38Caplan, “Probability and the Synthetic A Priori,” p. 83.
3Nbid., p. 82.


===== PAGE 16 =====

94 — JOURNAL OF LIBERTARIAN STUDIES 19, NO. 1 (WINTER 2005)

similarly moved in my direction by my “findings” that rent control
solves housing shortages? They should have been, on Caplan’s
Bayesian account. However, let me assure you, I was there, and this
was not at all their reaction. Rather than being slightly moved in my
direction, they were derisive and, if truth be told, dismissive and
even disappointed. That is, the fix was in. We all knew the correct
results. The incorrect ones were fudged away.

CONCLUSION

Caplan has launched the third Austrian Methodenstreit. But, as in the
other two cases, he has not succeeded in overturning the Austrian
edifice. As they say in boxing, he has not so much as laid a glove on
it. His publications in this vein, along with the critiques they have
engendered, have instead served to strengthen praxeology. Caplan
simplistically confuses logic and psychology, and his Bayesian
account is of no account.

BIBLIOGRAPHY

Anderson, William, Walter Block, Thomas J. DiLorenzo, Ilana Mercer, Leon
Snyman, and Christopher Westley. “The Microsoft Corporation in Col-
lision with Antitrust Law.” Journal of Social, Political, and Economic Stud-
ies 26, no. 1 (Winter 2001).

Barnett, William, II. “Contra Caplan.” Unpublished manuscript. Depart-
ment of Economics. Loyola University New Orleans.

Becker, Gary. “It’s Simple: Hike the Minimum Wage, and You Put People
Out of Work.” Business Week Magazine (March 6, 1995).

Block, Walter. “Austrian Theorizing, Recalling the Foundations: Reply to
Caplan.” Quarterly Journal of Austrian Economics 2, no. 4 (Winter 1999).

——. Defending the Undefendable. New York: Fox and Wilkes, 1991.

——. “On Robert Nozick’s ‘On Austrian Methodology.” Inquiry 23, no. 4
(Fall 1980).

——. “Realism: Austrian vs. Neoclassical Economics, Reply to Caplan.”
Quarterly Journal of Austrian Economics 6, no. 3 (Fall 2003).

——. “Reply to Caplan on Austrian Economic Methodology.” Unpublished
manuscript. Department of Economics, Loyola University New
Orleans.

Block, Walter, Christopher Westley, and Alex Padilla. “The Case for Refereed
Journals Devoted to Austrian Economics Revisited (Are there any
objective criteria for determining truth in economics?)” Unpublished
manuscript. Department of Economics, Loyola University New
Orleans.

Blum, Walter J., and Harry J. Kalven, Jr. “The Uneasy Case for Progressive
Taxation.” University of Chicago Law Review 19 (1952).

Bostaph, Samuel. “Epistemological Foundations of Methodological Conflict
in Economics: The Case of the Nineteenth Century Methodenstreit.”
Ph.D. diss., Southern Illinois University at Carbondale, 1976.


===== PAGE 17 =====

REJOINDER TO CAPLAN ON BAYESIAN ECONOMICS — 95

Caplan, Bryan. “The Austrian Search for Realistic Foundations.” Southern
Economic Journal 65, no. 4 (April 1999).

——. “Probability and the Synthetic A Priori: A Reply to Block.” Quarterly
Journal of Austrian Economics 6, no. 3 (Fall 2003).

—. “Probability, Common Sense, and Realism: A Reply to Hiilsmann and
Block.” Quarterly Journal of Austrian Economics 2, no. 4 (Summer 2001).

Card, David, and Alan B. Krueger. “Minimum Wages and Employment: A
Case Study of the Fast-Food Industry in New Jersey and Pennsylva-
nia.” American Economic Review 84, no. 4 (September 1994).

DiLorenzo, Thomas J. “The Myth of Natural Monopoly.” Review of Austrian
Economics 9, no. 2 (1997).

Fox, Glenn. “Economics as Prediction.” Chap. 5 in Reason and Reality in the
Methodologies of Economics: An Introduction. Cheltenham, U.K.: Edward
Elgar, 1997.

Friedman, Milton. “The Methodology of Positive Economics.” Essays in Pos-
itive Economics. Chicago: University of Chicago Press, 1953.

Hiilsmann, Jorg Guido. “Economic Science and Neoclassicism.” Quarterly
Journal of Austrian Economics 2, no. 4 (Winter 1999).

“Methodenstreit.” www. wikipedia.org.

Mises, Ludwig von. Theory and History. New Haven, Conn.: Yale University
Press, 1957.

——. The Historical Setting of the Austrian School of Economics. Auburn, Ala.:
Ludwig von Mises Institute, 1984.

Nozick, Robert. “On Austrian Methodology.” Synthese 36 (1977).

Rothbard Murray N. Man, Economy, and State. Los Angeles: Nash, 1970.

——. “Mises’s ‘Human Action”: Comment.” American Economic Review 41,
no. 1 (March 1951).

—. “Praxeology: Reply to Mr. Schuller.” American Economic Review 41, no.
4 (December 1951).

——. “The Uneasy Case for Degressive Taxation.” Quarterly Journal of Aus-
trian Economics 4, no. 1 (Spring 2001).

Schuller, G.J. “Rejoinder.” American Economic Review 41, no. 1 (March 1951).

——. “Review of Human Action by Ludwig von Mises.” American Economic
Review 40, no. 3 (June 1950).

Van Dun, Frank. “Bayesianism and Austrian Apriorism.” http://allserv.
rug.ac.be/~frvandun/Texts/ Articles /Bayesianism-Austrian.pdf.