subscribe to our mailing list:
Anthropic Reasoning and the Contemporary Design Argument in
Astrophysics: A Reply to Robert Klee
Mark A. Walker
University of Toronto
15 Devonshire Place
Milan M. Ćirković
Astronomical Observatory Belgrade
Volgina 7, 11160 Belgrade
Serbia and Montenegro
Posted February 24, 2004
The Significance of Fine Tuning: Observation Selection Effects vs. the design Argument
- Introductory: Pythagoras, Klee, and Anthropic Fine-tunings
- Getting the Numbers Right
Abstract. In a recent study of astrophysical "fine-tunings" (or "coincidences"), Robert Klee critically
assesses the support that such astrophysical evidence might be thought to lend to the design argument
(i.e., the argument that our universe has been designed by some deity). Klee argues that a proper
assessment indicates that the universe is not as "fine-tuned" as advertised by proponents of the design
arguments. We argue (i) that Klee's assessment of the data is, to a certain extent, problematic; and (ii)
even if Klee's assessment of the data is correct, it provides a necessary but not a sufficient response to
the design argument. However, an adequate skeptical rejoinder to the design argument can be made by
appealing to the anthropic principle.
The validity and meaning of the so-called 'fine-tunings', underlying the anthropic principle, has
always been highly controversial.  In a recent interesting critical study, colourfully entitled "The Revenge
of Pythagoras: How a Mathematical Sharp Practice Undermines the Contemporary design Argument in
Astrophysical Cosmology," Robert Klee criticizes both the mathematics and physics of (putative) fine-tunings,
in order to show that the teleological conclusions drawn from them are ill-founded (Klee 2002).
Klee's primary target is a contemporary design argument that invites us to believe that the best
explanation for the apparent fine-tuning of certain astrophysical numbers is that our universe was
designed by a deity. Clearly there are at least two possible means to criticize the theist's argument: one
could challenge the significance of the astrophysical numbers, or one could challenge the numbers
themselves. Klee takes the latter (uncommon) approach. While the ultimate target of Klee's criticism is
the design argument, often the proximate target is anthropic reasoning and the anthropic principle. The term 'anthropic principle' comes from Carter (1974), and refers to the idea that the values of fundamental
constants of our universe are (purportedly) highly constrained by the contingent fact that human
observers exist: what might otherwise seem to be possible values for the fundamental constants can be
ruled out because they are inconsistent with (say) the existence of stars that are stable for billions of
years, or with the existence of supernovas. The 'anthropic thinking' here is that both stable stars and
supernovas are necessary for the evolution of intelligent life: the former as a home and energy source for
life to evolve, the latter to produce the heavier elements in the periodic table necessary for many life
processes. Klee considers almost exclusively theistic applications of anthropic reasoning; such reasoning
invites us to infer abductively from the fact that our universe has such a precise set of fundamental
constants to the conclusion that our universe has been designed by a deity. So, in criticizing anthropic
reasoning, Klee understands himself to be undercutting the argument from design.
While we are sympathetic to the conclusion that Klee draws, namely, skepticism about the
soundness of the design argument, we believe that his argument is deficient. The paper divides into two
parts. First, we critically assess Klee's attempt to show that the astrophysical numbers are not as finely
tuned as has often been alleged; we argue that Klee's corrective is itself not without errors. Second, we
show that, even if Klee's reassessment of the astrophysical data is correct, nevertheless; this in itself does
not undercut the argument from design, since Klee's argument is open to a plausible theistic rebuttal.
Specifically, we find that Klee's argument provides a necessary but not sufficient response to the design
argument. To shore up skepticism about the design argument, we show that the atheist may appeal to
Klee's major complaint against the contemporary design argument in astrophysics is that it
employs what he terms a 'mathematical sharp practice'. The term 'sharp practice' is borrowed from the
legal tradition: "…a practice that, while technically not illegal, is nevertheless shrewdly selfadvantageous
to the point of being seriously misleading" (p. 333). Among the sins covered by the idea of
mathematically sharp practices are: imprecision in quoting dimensionless ratios of fundamental constants
and cosmological parameters, and stretching of the meanings of terms "of the same order of magnitude"
or "of similar order of magnitude". Klee finds examples of sharp practice throughout the history of
anthropic thinking, from Weyl and Eddington to the monograph of Barrow and Tipler (1986).
In most general terms, surely Klee is correct to condemn mathematical sharp practice -- wherever
it may occur. Thus, he is correct that, if the design argument is to be considered sound, then it is
important that the numbers that are cited in support of "fine-tuning" (or coincidences) are trustworthy,
and in no way simply the result of some mathematical sharp practice. While we believe that there are
many pieces of his analysis that are error-free, and perhaps should have been said long ago; nevertheless,
we have some general and specific worries about Klee's treatment of the astrophysical data. Indeed, we
believe there is room to wonder whether Klee himself is guilty of mathematical sharp practice.
A very general concern arises from Klee's treatment of the astrophysical data in Table 1-- the
centrepiece of his article. Here he places in the same column (i) values which are known to one part in a
thousand or better (like proton mass), (ii) values which of necessity vary within one order of magnitude
(like human length and mass), and (iii) those which -- although determined well-enough -- span several
orders of magnitude due to the different physics involved (like the planetary masses), and, finally, (iv)
those which are plainly unknown, remaining uncertain to within more than an order of magnitude (like
galactic masses). From a basic high-school (or undergraduate) course in statistics and data analysis, one
may learn that unweighted averages of highly uneven data are worthless. However, Klee does not
hesitate to quote five decimal places without any noticeable doubt in worthiness of these results, just as
schoolchildren often use their calculators, without realizing that uncertainties cut your decimal
calculation at (say) the first decimal figure. "Oh, the cheapness of mindless calculation in this modern
age!" -- the exclamation of Sir Roger Penrose (1991, p. 25) comes appropriately to mind. So one may
wonder whether Klee himself is guilty of "cooking" numbers (to invoke Klee's own phrase) in order to
suit his arguments, that is, indulging in "sharp practice"!
To secure this point, let us consider Klee's discussion of galactic mass and size in Table 1 as
instructive. Selecting this example seems more than fair, since Klee chooses as main targets of his
criticisms astrophysical cosmologists. Therefore, it seems most appropriate to discuss those entries in
Table 1 that most directly touch upon our astrophysical knowledge, (as opposed to other measures cited
in the table less directly related, e.g., the size and mass of elementary particles or human beings). Let us
begin with a very general point: if one wishes to draw conclusions dealing with the history and
philosophy of science, one would do very well to first ascertain the fine details in each particular
discipline that rarely find their way into textbooks and other polished presentations, for here it is where
one will often find "awful truths". One of the "awful truths" of astrophysical science is that our
knowledge of the mass and size of galaxies is highly uncertain: these uncertainties are of an order of
magnitude or greater. Even in our present state of ignorance, however, it is clear that most of the matter
in any galaxy is located in the dark halo, and that everything visible -- technically speaking, what is
contained within the Holmberg radius of a galaxy, typically about 20 kpc (kiloparsecs, 1 kpc = 3.09 x
1019 m) -- is just the proverbial "tip of the iceberg", never more than about 20% of the size and 10% of
the mass of the whole (and often much less). It is a daunting observational task to determine these
parameters exactly, and it is not disgraceful to admit that we still do not know the mass of even our own
galaxy (the Milky Way) -- not to mention other galaxies -- with any great deal of precision. Thus, even if
there is a claimed anthropic "prediction" of values of these quantities, it seems clear that we cannot at
this stage justly assess the value of such a prediction.
To see this, consider that in Table 1, Klee uses the value of 1041 kg to represent the mass of the
Milky Way, which is about 5 x 1010 M (Solar masses, 1 M = 1.989 x 1030 kg), which is very far off the
mark, being certainly too small. We shall quote only a handful out of a host of measurements performed
in the last decade, in order to show the complexity of the problem. By considering the Milky Way's outer
satellites, Kulessa and Lynden-Bell (1992) obtained a mass of around 1.3 x 1012 M within 230 kpc. The inner part of the Galaxy is accessible to measurements via the high-velocity stars; using this technique, García Cole et al. (1999) estimate the mass within 45 kpc to be 5 x 1011 M, still an order of magnitude larger than Klee's value. Kochanek (1996), using several methods concordantly, obtains similarly (4.9 ± 1.1) x 1011 M out to the galactocentric distance of 50 kpc. Finally, Wilkinson (2000) suggests 1.9 x 1012 M with the scalelength of 170 kpc.
Similar data is available for other giant spiral galaxies comparable to the Milky Way in
luminosity and other properties. The best studies so far have been performed by Dennis Zaritsky and his
coworkers, on the basis of observations of satellite galaxies. Zaritsky and White (1994) claim the average
mass of a spiral galaxy to be 1.5-2.6 x 1012 M within 200 kpc. In continuation of that study, Zaritsky et
al. (1997) confirm the average mass to lie "in excess of 2 x 1012 M" with strong dynamical indications
that dark haloes extend to more than 400 kpc! In any case, Klee's fiducial radius of about 2 kpc (about 6 x
1019 m) is flabbergastingly wrong (being, for instance, more than 4 times smaller than our own distance
from the Galactic centre!). Thus, one of the biggest inaccuracies of "anthropic" predictions in Table 1 is
simply a consequence of the two orders of magnitude error on Klee's part.  Carr and Rees can be
forgiven an error like this, since most of this modern galactic astronomy was completely unknown in the
late 1970s; but should not be overlooked by Klee, who writes in 2002, well after what, in effect, has been
a revolution in our understanding of the nature of galaxies. As Earman (1995, p. 268) cogently notes:
"...it needs repeating that philosophy of science quickly becomes sterile when it loses contact with what
is going on in science."
The same imprecision may be applied to quantities such as stellar mass and length, since (if we
forsake vulgar heliocentrism!) these quantities are not unmistakably or fairly represented by their Solar
values, which Klee uses to check the accuracies of the anthropic predictions. This is not to say that the
values thus obtained would be very different from the value Klee uses (Solar values); and it is even
probable that the discrepancy with the anthropic predictions would be even larger in these cases (since,
for instance, the average stellar mass is obviously lower than 1 M, no matter which stellar mass function
turns out to be the best). But these empirical findings of astrophysics must be taken ino account if one
purports to seriously discuss astrophysical quantities. To not do so would usually bias us towards (and
not against) our privileged position in the universe!
The desired net result of Klee's entire numerical exercise concerning the purported fine tuning of
the galactic mass is to somehow show that the "enthusiasts" of the anthropic principle are less precise
than its detractors. The reality is, however, that neither is very precise. The considered judgment here,
contra some design argument proponents and Klee, is that, to invoke another phrase from the legal
tradition, "the jury is still out".
Our second case for further scrutiny is Klee's interpretation of the (in)famous story about finetunings
associated with carbon-12. While this presents us with a whole nest of interesting and
controversial issues, we shall, for the sake of brevity, confine ourselves to a single point: Klee's use of
contemporary research into the anthropic significance of the nuclear resonances in light elements. Klee
concludes this section (6) of his paper with the following remark: "The quantitative details are thorny and
depend on the specific model of stellar structure used, but the width of the window of compatibility
appears to be wide enough to undermine any claim that we have a case of fine tuning on our hands"
(347). Klee attempts to support this contention by citing empirical research -- but herein lies the problem:
this is a sizeable subfield of anthropic research, and there have been several important references dealing
with it (Livio et al. 1989; Pochet et al. 1991; Oberhummer, Csoto, and Schlattl 2000; Hogan 2000).
Unfortunately, Klee cites just one of these modern studies, and the oldest one at that (Livio et al. 1989),
whose results (as Klee correctly sees) could be called critical toward the conventional anthropic view of
these nuclear resonances. This selective approach creates an impression that modern research has
somehow disproved Hoyle, Barrow and Tipler and others who perceive something important in these
data, and in need of an explanation. This impression is misleading. Other studies (and subsequent to
Livio et al.) report different results, and much more in the spirit of the idea that reality is fine tuned. For
example, Oberhummer et al. (2000) emphasize that (p. 90)
...a change of more than 0.5% in the strength of the strong interaction or more than
4% change in the strength of the Coulomb force would destroy either nearly all C or
all O in every star. ... Therefore, for the above cases the creation of carbon-based life
in our universe would be strongly disfavoured. The anthropically allowed strengths of
the strong and electromagnetic forces also constrain the Higgs vacuum expectation
value... and yield tighter constrains on the quark masses than do the constrains from
light nuclei... Therefore, the results of this work are relevant not only for the anthropic
cosmological principle..., but also for the mathematical design of fundamental
elementary particle theories.
That Klee fails to discuss or at least cite this (and other recent studies) seems to be a case of being
shrewdly self-advantageous with the data to the point of being seriously misleading, that is, of engaging
in sharp practice.
Of course, one may always claim that 0.5% or 4% are a far cry from "fantastical precision"; but
the very point (as we shall argue) is that obtaining such "fantastic precision" is not necessary to vindicate
the anthropic principle. Oberhummer et al. are certainly conscious that 4% is not "fantastic precision" by
every standard when they invoke anthropic considerations, but they are not seeking such precision;
rather, they seek -- along the entire scientific, notably physical tradition -- something infinitely more
important: an explanation. That is, what, if anything, explains the fact that our universe has values that
fall within such a narrow range compatible with life? To this question we now turn.
As noted above, we believe that there are many pieces of Klee's analysis that are error-free, and
perhaps should have been said long ago. Furthermore, we agree with the general spirit of Klee's
proposal to tackle the astrophysical numbers "head-on". However, we believe we have shown that it is
not clear that Klee himself has demonstrated his case against anthropic reasoning with respect to the
astrophysical data. Indeed, we believe that in several places Klee may be hoisted on his own petard: the
sharp practice. In this section we propose to examine Klee's thesis from a different angle. Specifically,
let us grant that Klee has shown that some of the astrophysical numbers are not of such "fantastic
precision" as has been advertised by some proponents of anthropic reasoning and ask: Does it follow that
the design argument is refuted? We will construct a reply on behalf of the theist to show that Klee's
result does not refute the design argument.
We should pause here to note that it might be thought that there is a certain irony in the fact we must construct (almost from scratch) a rejoinder on behalf of the theist, (even though it is a position that
we ultimately reject), given the intense scrutiny that anthropic reasoning has received in some quarters.
Surely, it might be thought, the issue must have already been discussed. The explanation here is actually
quite simple: the intelligent design agenda is actually a fringe in anthropic thinking. So it is with a wry
smile that one should read the opening statement of Klee's abstract: "Recent developments in
astrophysical cosmology have revived support for the design argument among a growing clique of
astrophysicists." (p. 331) However, in the subsequent discourse there is not a single astrophysicist which
is unequivocally cited as supporting the design argument (with the possible exception of the late Sir Fred
Hoyle) , not to mention any "clique," and especially "growing clique". One would expect the author to
support his allegations with more specific names and statements. As astrophysicists associated in a rather
vague manner with the design argument, we find only the names of Davies, Barrow, Press, Lightman,
and Tipler.  All in all, the "growing clique" is rather thin, so it is no surprise that the burden falls to us of
formulating what we shall call the 'sophisticated theist's' response.
The unsophisticated theistic design argument is bedazzled by whole numbers much in the
manner of the Ancient Pythagoreans.The sophisticated theist has a different take on both the
astrophysical and theological data. We might imagine, then, the sophisticated theist arguing thus:
Unlike some theists, we are more than happy to grant that Klee is correct that the finely tuned
numbers are not as finely tuned numerically as has been alleged. However, an appropriate
appreciation of the relevant theology shows that Klee's assessment of the data does not support
an anti-design argument conclusion. Indeed, we think a case can be made that this lack of
precision should in fact be expected. Consider that God could have made the universe such that
it is overwhelmingly obvious that there was a designer at work. For instance, God could have
engraved or marked every stone, every animal, every plant, etc., with the word 'GodTM'. We can
easily imagine scientific investigation eliminating any possibility of hoax, e.g., if radioactive
dating revealed that the engraving on many rocks predates the evolution of life on this planet
then we would have pretty solid evidence of a designer at work in our universe. Furthermore, we
can imagine any number of ways of mathematically ordering the universe. Think about a very
simple example, the Prime Number Universe: God could have made a rather gauche universe
where every solar system is based on prime numbers, e.g., the number of planets in a solar
system is always a prime number, and every planet has a prime number of moons where planets
with larger orbits have a larger prime number of moons. To get a fix on this, imagine what God
would have to do to make our solar system into a member of the Prime Number Universe: He
might destroy the last two planets so that we have a prime number of seven in our solar system.
Furthermore, he would have to add and subtract a few moons to yield the following pattern:
Mercury would have two moons, Venus three, Earth five, Mars seven, Jupiter eleven, Saturn
thirteen, Uranus seventeen. If modern astronomy (starting with Galileo) had discovered that our
solar system had this prime number configuration it would be an obvious conjecture to think that
this was the result of the work of some Grand Designer. Clearly then God has in his powers the
ability to flaunt the fact that He designed the universe. But while this sort of obviously designed
universe is within God's ability to make, God has moral reasons for not making the universe in
this way. Specifically, God did not design the universe this way because He wanted to make the
universe, where possible, ambiguous between a secular and theistic interpretation; so that we
could freely choose to embrace Him or not, and that we may freely will to perform good or bad
acts. As Murray argues (1999, p. 243)
...it seems clear that fully robust and morally significant free will cannot be exercised by
someone who is compelled by another in the context of a threat. ...[I]f God does not
remain "hidden" to a certain extent, at least some of the free creatures He creates would
be in the condition of being compelled in the context of a threat and, as a result, such
creatures could not exercise their freedom in this robust, morally significant manner.
A universe that exhibited overwhelming design qualities would not allow any free choice in the
matter of whether to believe in God, or whether to perform good or bad acts in the absence of a
threat. So, it would have been improper for God, to make fine tunings "fantastically precise" -- to
use Klee's phrase -- because then He would not have remained sufficiently "hidden". Thus, it is
to somewhat oversimplified to claim that the absence of "fantastical precision" is an argument
against the theist's assessment of the astrophysical data. The fact that the relevant numbers are
pretty close to the conceivable optimum, although not exact, supports God's mission to create a
somewhat ambiguous universe: God wants to lay clues to His existence but He does not want to
make the evidence so overwhelming that it forces everyone to believe in Him. (On the other hand,
He did not want to make the evidence against His existence overwhelming either). The fact that
God has laid clues can be seen from the fact that it is pretty clear to all in the debate that the
astrophysical numbers cannot be thought to be purely random.  As Oberhummer et al. (2000, p.
90) argue ...a change of more than 0.5% in the strength of the strong interaction or more than
4% change in the strength of the Coulomb force would destroy either nearly all C or all O in
every star. To this we can add parameters of our best physical theories (like the GUTs or the
standard Big Bang models) are fine tuned (although perhaps not with "fantastic precision") in
the sense that the universe would have been inhospitable to life if their values had been slightly
different, e.g., no life seems imaginable if the universe had contained only a highly diluted
hydrogen gas, or if it had recollapsed before the temperature anywhere had dropped below
10,000 K, which corresponds to early expansion speeds different from the actual for only (plus
or minus) one part in 1010. Likewise, the universe containing only black holes (and thus
extremely inhospitable for anything remotely similar to life) is statistically more probable from
the observed configuration of matter for a factor containing a stupendous double exponential
(e.g., Penrose 1989). If the values of our universe had been set randomly we would almost
certainly not have existed. So what possible explanation is there for the fact that the universe just
happens to fall into this narrow range of values allowing for the development of life? It seems
that the atheist cannot explain this other than by saying that it is a fortuitous fact about the
universe. A much more plausible inference is that the universe has been created by a Designer,
which explains the fine tuned (but "not-so-fantastically tuned") numbers we observe. That is, the
numbers are tuned sufficiently to provide clues to God's existence, but not too fantastically
precise as to make it overwhelmingly obvious that God is the designer of the universe.
The sophisticated theist's rebuttal here appears to leave naturalists with a rather unappealing
dilemma: either we accept the theist's explanation for the anthropic data -- that is, the data that says that
the values for the fundamental constants of our universe fall into the very narrow range that is compatible
with the existence of life -- or we offer no explanation for this. The way out of this apparent dilemma is to
realize that appeals to anthropic reasoning need not be theistic. As we noted above, theistic
appropriations of anthropic reasoning constitute a minority: anthropic reasoning makes perfectly good
sense from a naturalistic view. There are a significant number of scientists (and philosophers) who
recognize validity and significance of the anthropic principle -- but they do not see anything even
remotely similar to the design argument in it! Unfortunately, Klee does not discuss such naturalistic
appeals to anthropic reasoning. Now it may be thought that Klee's exculpation here is that for exegetical
reasons he focused on only theistic versions of anthropic reasoning, because his ultimate target is the
design argument. However, the reply on behalf of the sophisticated theist shows that his argument does
not provide sufficient support for his skepticism about the design argument.
It is arguable that the core of the anthropic reasoning, as noticed by many physicists (like J.
Richard Gott, Fred C. Adams, Andrei Linde, etc., in addition to those which Klee mentions), is in fact
anti-teleological. It is easy to appreciate how such a conclusion might be reached. Naturalistic anthropic
reasoning accepts the apparent fine-tuning of our cosmological domain as a puzzling empirical fact in
need of an explanation. As we have said, the design hypothesis provides one possible (but not necessarily
plausible) explanation. But anthropic reasoning, together with other advances in physics, provides an
alternative explanation: if there is an ensemble of universes -- also called the multiverse -- with a suitably
varying range of properties (and current quantum cosmological models, like Linde's chaotic inflation
(e.g. Linde, 1986, 1990), do suggest that that is the case), then one would naturally expect some of these
universes to be just right for life. Even if such universes are in a small minority, because of an
"observation selection effect" -- the life-less universes, where the constants are "wrong" contain no
observers and cannot be observed -- such a theory can predict that we should observe precisely what we
do in fact observe: a universe that appears to be reasonably fine tuned. This does not happen because the
underlying theory is fine tuned, or incorporates anything a priori improbable (which is considered by
some to be a hallmark of the design hypothesis), but because an observation selection effect guarantees
that we observe a very atypical part of the whole of physical existence. In other words, embedding of
atypical (or "finely tuned") region of the entire reality into the larger whole, gives a perfect antiteleological
and naturalistic explanation of our (a priori improbable) observations (this was emphasized
by the great late Robert Nozick as his "principle of fecundity").  Physicists and astronomers have been
familiar with the observational selection effect for a long time, some aspects of it (e.g., Malmquist bias in
astronomy ) being the subject of studious and detailed mathematical modeling.
The question arises, if scientists have been aware of observational selection effects for so long,
why has it suffered relative neglect in the philosophy of science? While not directly addressing this issue,
Klee suggests a possible reason when he writes (p. 352):
I believe it is a mistake to see any virtue in trying to construct formal Bayesian
probability arguments in the present context. To his credit, Leslie does not do so. He
formulates his argument as an abductive one in which the mathematical data on the
allegedly fantastically narrow intervals of variance are not used in a technically formal
Klee here is referring to Leslie's theistic argument in Universes (Leslie, 1989). According to Klee, the
problem with trying to formulate something stronger than Leslie's inference to the best explanation
argument is that there is an insurmountable problem arising from issues involving measurement scales.
While we cannot explore the issue here (and Klee does not pursue it either) the basic complaint is that the
interval of variance compatible with hydrocarbon life can be made to seem larger or smaller depending
on one's measurement scale. So, according to Klee, because of this, there is no hope of quantifying the
argument with greater precision, and thus, nothing stronger than an argument to the best explanation can
be made. This in turn suggests that there is a limited amount that philosophers and scientists can do to
clarify and quantify the argument. Whether this is ultimately so remains to be seen. An antidote for such
pessimism can be found in Bostrom (2002) (which is not discussed by Klee). It points out immediately
that cosmology is fundamentally incomplete without taking into account the necessary "anthropic bias":
conditions we observe in fundamental physics, as well as in the universe at large, are atypical, and
require an explanation. Furthermore, Bostrom has constructed a Bayesian argument precisely of the sort
Klee says there is no virtue in formulating. If this line of argument can be maintained, then not only is
Bostrom's approach to anthropic coincidences superior to Leslie's, but it offers a chance of obtaining
solutions to many serious problems in the philosophy of cosmology and physics. The central piece of this
approach is the Observation Equation (Bostrom, 2002, p. 173), which subsumes seemingly vague
assumptions and observational selection criteria in full mathematical rigour. Moreover, the last chapter of
Bostrom's monograph gives an example how the very same Bayesian apparatus can be applied to both
astrophysical fine tunings and to such issues as traffic planning and the origin of the thermodynamical
arrow of time. In any event, if something along the lines sketched by Bostrom is correct, then it points the
way philosophers of science might pursue the question of anthropic coincidences in a naturalistic and
To summarize, if the design hypothesis was the only way of explaining the astrophysical fine
tuning, then by an inference to the best explanation, one could think that fine tuning lent support to that
hypothesis. But in fact, naturalistic anthropic reasoning shows that there is another viable explanation for
the puzzling phenomenon of fine tuning (the multiverse) grounded in useful and much-investigated
physical ideas (namely those of inflation and quantum cosmology). It thus decreases the amount of
support that fine tuning could be thought to give to the design hypothesis. Therefore, this line of
anthropic reasoning is, in its effect, anti-teleological. Whether this view should ultimately triumph is a
large topic that cannot be explored further here. The point we want to make simply is that there is a
potential rival explanatory theory, hence, we can reject the sophisticated theist's claim that the design
argument is the only viable explanation for the apparent fine tuning of the visible universe.
Not only does the naturalistic version of anthropic reasoning force a re-evaluation of the design
argument, it also forces a re-evaluation of many objections to anthropic reasoning in Klee's article. For
instance, in the section entitled "The recalcitrant sloppiness of crud" we read, as a response to the Press
and Lightman (1983) study: "But how can the details be considered abhorrent in this context? design
arguments are all about details [italics in the original]" (p. 345). Well, design arguments might be about
details, but this is irrelevant in the present case, since Press and Lightman did not intend to promote any
design argument! And -- in contradistinction to the design argument -- observation selection effects are
not all about details. They are just about probability measures, and any interval is as good as any other in
obtaining weights in the parameter space. It is only the interpretation of these weights that decides
whether we should be flabbergasted or just indifferent toward one value or another. The simple truth
about "anthropic" fine tunings is that some are rather precise, others not so. There appears to be a great
spread in this "finesse" of coincidences, further research should give us a better idea which ones require
an appeal to an anthropic explanation.
We have argued, on the one hand, that Klee's claim that the astrophysical data does not support
the idea of "fantastic precision" in itself is insufficient to support his skepticism about the design
argument; and on the other hand, that a proper appreciation of naturalistic anthropic reasoning can be
used to undermine the putative explanatory power of the design hypothesis. But it may be wondered at
this point how we can claim (as we did above) that Klee's argument is necessary for refuting the design
argument. For if, as we argued, the theist and the atheist both have an explanation for "less than
fantastically precise fine-tunings", then how can Klee's argument add anything to the undermining of the
To answer this, we need attend first to an unfortunate confusion, on the part of both admirers and
detractors of anthropic reasoning, that the idea that the Eddingtonian numerology is somehow essential
for understanding the anthropic principle. One of the virtues of the historical part of Klee's paper is that
one can see how these ideas developed independently. As Klee describes the history, Eddington's
numerology takes historical precedence. The appeal to numerology turns on the idea that the
fundamental constants of our universe are in some sort of "mysterious" or "occult" numerological
relationships.  The anthropic principle asserts that the fine tuning of our universe for life depends on the
fundamental constants of our universe falling into a rather narrow band. It is arguable whether Klee
always keeps this distinction in mind, but then again, even the monograph of Barrow and Tipler (1986) --
what Klee describes as the "Bible" of anthropic reasoning -- is not entirely innocent either in keeping
these ideas distinct. However, that these ideas are logically independent of one another should be
obvious. The fact that different values of the fundamental constants and cosmological parameters would
lead to a lifeless universe has nothing to do with alleged simple arithmetical relationships between those
constants and parameters. Anthropic thinking follows from the former, but not from the latter. To run
them together is to fall prey to an irony of historical development, namely, the accident that the anthropic
principle was brought to light when, in 1961, Dicke demolished Dirac's large-number hypothesis.
Obviously this should not lead anyone to think that there is a necessary relationship between the two.
Surely we can imagine a universe that exhibits all sorts of Eddington numerology, but is inhospitable to
life (imagine radiation levels too high to allow life to develop). Conversely, we can imagine a universe
where the physical parameters are finely tuned for life, but the parameters themselves do not exhibit any
Eddington numerology. Thus, any critique of Eddingtonian numerology leaves the anthropic principle
unscathed; to claim otherwise is a non sequitur.
While these ideas are distinct, we can imagine possible universes where they are both true. Let us
think of these as 'gratuitously fined tuned universes'. Gratuitously fined tuned universes are a proper
subset of finely tuned universes: they are finely tuned for life and exhibit mathematical "orderliness"
above and beyond what is required for life. As an example of a gratuitously fine tuned universe, think
again the idea of the Prime Number Universe. Recall that this is a universe much like our own with the
exception that every solar system has been arranged to exhibit a certain amount of mathematical
orderliness: there is always a prime number of planets and every planet has a prime number of moons.
This universe is fine tuned in the sense that the fundamental constants fall into a narrow band which are
hospitable to life, but the universe also exhibits excessive mathematical orderliness: there is no reason to
suppose that, if this universe is like ours, then having a prime number of planets in every solar system
and every planet having a prime number of moons is in any way necessary for life. The naturalist, we
believe, must concede that a gratuitously fine tuned universe is one where the design argument may have
greater explanatory power than its rivals. For now we are imaging two things requiring explanation: that
the fundamental constants of the universe fall into a narrow band or parameter space necessary for life,
and that the universe exhibits mathematical orderliness above and beyond the requirements for life. A
naturalistic appeal to anthropic reasoning can explain the former, but not the latter. On the other hand, the
theist could appeal to the aesthetic of the designer to explain the excessive mathematical orderliness of
such a universe, and so would seem to offer a better explanation. 
So, the question of whether the design argument has superior explanatory power resolves to this:
Is our universe gratuitously finely tuned? The answer, we believe, is no, and here is where we can invoke
Klee's results. Klee should be read as showing that the astrophysical data does not support the idea that
we live in a gratuitously fine tuned universe. Klee is convincing in his rejection of Eddington
numerology, and that the astrophysical data does not reveal some sort of "fantastic precision". We have
argued that at least some of the data supports the idea that the universe exhibits fine tuning (section 2.2
and 3.1), but we do not think that there is evidence that the universe is gratuitously fine tuned.
It should be emphasized that we take this to be an empirical matter: it may turn out to be true that
the universe is gratuitously fine tuned. Imagine, for example, that advances in astronomy eventually
reveal that every solar system except our own in the visible universe is as described in the Prime Number
Universe. Presumably we might attribute the first such distant solar system with this pattern to mere
randomness. But surely after finding that solar system after solar system exhibited a prime number
pattern it might be reasonable to suppose that our universe was designed, and that the designer wanted to
remain hidden until we had technically advanced to the point where we could make observations of
distant solar systems. Alternatively, perhaps advances in astrophysics will show that the fundamental
constants have some sort of hidden numeric relation that cannot be explain simply as a requirement for
the development for life. Surely it would be dogmatic to rule-out a priori such possibilities -- but it would
have to be evidence of this sort to support the hypothesis that the universe is gratuitously fine tuned.
Our major complaint against Klee is that he fails to see that a successful defence of a naturalistic
worldview in cosmology may well be dependent upon appreciation of the anthropic observation-selection
effects. If we wish to refute sophisticated design arguments, we need to show the way to reconcile the
observed non-randomness with the a priori negligible size of the life-permissible subset of the parameter
space. Fortunately, modern quantum cosmology, coupled with the anthropic reasoning, offers a perfectly
sound avenue along which further research in both physics and philosophy may find the solution.
The case against the contemporary design argument in astrophysics might be summarized by
offering the theist a dilemma: either the theist maintains that the universe is gratuitously fine tuned or the
theist does not. If the former, then the theist must adduce evidence that our universe exhibits more than
the fine tuning necessary for life, in addition, the data must support the idea that the universe is designed
with some concern for mathematical orderliness. But Klee does a good job of summarizing evidence
against this hypothesis. If the theist takes the latter horn of the dilemma, then any explanatory advantage
evaporates, if we allow that atheists too have an explanation for the apparent fine-tuning of our universe
(as was discussed in 3.2 above).
We wish to conclude by emphasizing once again that fine tuning is completely divorced from
Eddingtonian numerology and other claims that the universe exhibits gratuitous mathematical ordering.
There is no logical or physical (especially not causal) relationship between the two, except the historical
one. Any numerology is a counsel of despair, particularly today, when we may be on the verge of the
completely unified theory of the dynamics of matter in nature. To criticize the anthropic principle on the
basis of weaknesses of naive numerological musings of the 1930s is about as appropriate as criticizing
the concepts of modern chemistry on the basis of the flaws of the Ancient theory of matter consisting of
earth, air, fire, and water. The anthropic principle has been perhaps overly criticized in recent decades,
and often on flimsy enough grounds. It is, probably, a high time and good opportunity (offered, among
other things, by some of the recent physical and cosmological studies quoted above) to pause, to rethink,
and try to reach a new and deeper level of discussion.
Barrow, J. D. and Tipler, F. J. 1986, The Anthropic Cosmological Principle (Oxford University Press,
Bostrom, N. 2002, Anthropic Bias: Observation Selection Effects in Science and Philosophy (Routledge,
Carter, B. 1974, "Large Number Coincidences and the Anthropic Principle in Cosmology", in M. S.
Logair (ed.) Confrontation of Cosmological Theories with Observational Data, Reidel. 291-8.
Ćirković, M. M. and Bostrom, N. 2000, "Cosmological Constant and the Final Anthropic Hypothesis,"
Astrophysics and Space Science 274, 675-687.
Earman, J. 1995, "Recent work on time travel," in Time's Arrow Today, ed. by Steven Savitt (Cambridge
University Press, Cambridge), pp. 268-310.
García Cole, A., Schuster, W. J., Parrao, L., and Moreno, E. 1999, "The mass of the Milky Way from a
uvby – β survey of high-velocity stars," Revista Mexicana de Astronomía y Astrofísica 35, 111-
Garriga, J., Mukhanov, V. F., Olum, K. D. and Vilenkin, A. 2000, "Eternal inflation, black holes, and the
future of civilizations," Int. J. Theor. Phys. 39, 1887-1900.
Hogan, C. J. 2000, "Why the Universe is just so?" Rev. Mod. Phys. 72, 1149-1161.
Hoyle, F. 1982, "The Universe – Past and Present Reflections," Annual Reviews of Astronomy and
Astrophysics 20, 1-35.
Hoyle, F. 1994, Home Is Where the Wind Blows (University Science Books, Mill Valley).
Hoyle, F. and Wickramasinghe, N. 1999, "The Universe and Life: Deductions from The Weak Anthropic
Principle," Astrophysics and Space Science 268, 89-102.
Klee, R. 2002, "The Revenge of Pythagoras: How a Mathematical Sharp Practice Undermines the
Contemporary design Argument in Astrophysical Cosmology," Brit. J. Phil. Sci. 53, 331-354.
Kochanek, C. S. 1996, "The mass of the Milky Way," Astrophysical Journal 457, 228-243.
Kolakowski, L. 1986, "Emperor Kennedy Legend. A New Anthropological Debate," Salmagundi nR 72,
Kragh, H. 1996, Cosmology and Controversy (Princeton University Press, Princeton).
Kulessa, A. S. and Lynden-Bell, D. 1992, "The mass of the Milky Way Galaxy," Monthly Notices of the
Royal Astronomical Society 255, 105-118.
Leslie, J. 1989, Universes, (Routledge, London).
Linde, A. D. 1986, "Eternally existing self-reproducing chaotic inflationary universe," Phys. Lett. B 175,
Linde, A. D. 1990, Inflation and Quantum Cosmology (Academic Press, San Diego).
Livio, M., Hollowell, D., Weiss, A., and Truran, J. 1989, "The Anthropic Significance of an Excited
State of 12C," Nature 340, 281-284.
Miller, G. E. and Scalo, J. M. 1979, "The initial mass function and stellar birthrate in the solar
neighbourhood," Astrophysical Journal Supplement 41, 513-547.
Murray, J. M. 1999, "Coercion and the Hiddenness of God", in Philosophy of Religion: The Big
Quesitons, ed. E. Stump and M. J. Murray, (Blackwell, Oxford), pp. 241-250.
Nozick, R. 1981, Philosophical Explanations (Harvard University Press, Cambridge).
Oberhummer, H., Csoto, A., and Schlattl, H. 2000, "Stellar Production Rates of Carbon and Its
Abundance in the Universe," Science 289, 88-90.
Penrose, R. 1991, The Emperor's New Mind (Penguin Books, New York).
Pochet, T., Pearson, J. M., Beaudet, G., and Reeves, H. 1991, "The binding of light nuclei, and the
anthropic principle," Astronomy & Astrophysics 243, 1-4.
Press, W. H. and Lightman, A. 1983, "Dependence of Macrophysical Phenomena on the Values of the
Fundamental Constants," Philosophical Transactions of the Royal Society of London, Series A,
310, pp. 323-336.
Smart, J. J. C. 1989, Our Place in the Universe: A Metaphysical Discussion (Blackwell, Oxford).
Tegmark, M. 1998, "Is 'the Theory of Everything' Merely the Ultimate Ensemble Theory?" Ann. Phys.
Wilkinson, M. 2000, "The Dark Matter Halos of Galaxies: Masses and Lensing Properties," Observatory
Zaritsky, D. and White, S. D. M. 1994, "The massive halos of spiral galaxies," Astrophysical Journal
Zaritsky, D., Smith., R., Frenk, C., and White, S. D. M. 1997, "More Satellites of Spiral Galaxies,"
Astrophysical Journal 478, 39-48.
 On occasion we will use the singular 'anthropic principle' although we acknowledge that this glosses over the fact that there are a number of different anthropic principles. See Chapter 3 of Bostrom (2002) for some discussion.
 Besides being unclear whence the values of astronomical values (planetary, stellar, galactic masses and radii, etc.) in Klee's article come. He goes into detail in quoting the source for the physical constants, but we are left in total darkness as to the source of astronomical values, where his errors are perhaps at least as numerous as of the
 Klee attempts to link the late Sir Fred Hoyle with the use of anthropic reasoning in support of the design argument. This may not be such a difficult task, since Hoyle's personality is so complex and multifaceted, and his
erudite ideas and interests so numerous. Thus, it is not hard to fetch something for everybody in his huge opus.
Interestingly enough, during the "great battle" of twentieth-century cosmology, Hoyle always wore the atheist's
suit, and has attacked his Big Bang opponents for invoking a miraculous and unknowable "creation event" at the
initial singularity. It seems that he did actually change his views subsequently, allowing for the theistic
interpretation of the empirical findings of physics and astronomy, specifically he invoked a form of the design
argument similar to the one Klee describes; but he did not attempt to hide it. Thus, he wrote in the most visible of
all places, Annual Reviews of Astronomy and Astrophysics (Hoyle 1982): "Suppose you were a superintellect
working through possibilities in polymer chemistry. Would you not be astonished that polymers based on the
carbon atom turned out in your calculations to have the remarkable properties of the enzymes and other
biomolecules? Would you not be bowled over in surprise to find that a living cell was a feasible construct? Would
you not say to yourself, in whatever language supercalculating intellects use, "Some supercalculating intellect must
have designed the properties of the carbon atom, otherwise the chance of my finding such an atom through the blind
forces of nature would be less than 1 part in 1040000." Of course you would, and if you were a sensible superintellect
you would conclude that the carbon atom is a fix."
 Carr and Rees are, as mentioned above, explicitly exculpated, and the position of Klee on Carter, Dicke, and Wheeler remains unclear.
 The naturalistic explanations at this point usually invoke a deeper structure contained, presumably, within the currently-sought "Theory of Everything", which could be shown to be mathematically or logically necessary. We
cannot enter this fascinating topic here.
 Nozick (1981). It is not beyond reason that such an understanding is present, for instance, in the passage quoted above from the summary of Oberhummer et al. (2000) article; note the locution "in our universe".
 The difference between the average absolute magnitudes of stars (or galaxies or any other similar sources) in magnitude-limited and distance-limited samples, discovered in 1920. by K. G. Malmquist.
 The naturalistic explanations at this point usually invoke a deeper structure contained, presumably, within the currently-sought "Theory of Everything", which could be shown to be mathematically or logically necessary. We
cannot enter this fascinating topic here.
 Of course the sophisticated theist response above would lead us to ask why God allowed aesthetic
considerations -- the mathematical ordering of the solar systems -- to trump the moral imperative to remain