The anthropic principle, also known as the "observation selection effect", is the hypothesis, first proposed in 1957 by Robert Dicke, that the range of possible observations that could be made about the universe is limited by the fact that observations could happen only in a universe capable of developing intelligent life. Proponents of the anthropic principle argue that it explains why the universe has the age and the fundamental physical constants necessary to accommodate conscious life, since if either had been different, no one would have been around to make observations. Anthropic reasoning is often used to deal with the idea that the universe seems to be finely tuned for the existence of life.

There are many different formulations of the anthropic principle. Philosopher Nick Bostrom counts them at thirty, but the underlying principles can be divided into "weak" and "strong" forms, depending on the types of cosmological claims they entail. The weak anthropic principle (WAP), as defined by Brandon Carter, states that the universe's ostensible fine tuning is the result of selection bias (specifically survivorship bias). Most such arguments draw upon some notion of the multiverse for there to be a statistical population of universes from which to select. However, a single vast universe is sufficient for most forms of the WAP that do not specifically deal with fine tuning. Carter distinguished the WAP from the strong anthropic principle (SAP), which considers the universe in some sense compelled to eventually have conscious and sapient life emerge within it. A form of the latter known as the participatory anthropic principle, articulated by John Archibald Wheeler, suggests on the basis of quantum mechanics that the universe, as a condition of its existence, must be observed, thus implying one or more observers. Stronger yet is the final anthropic principle (FAP), proposed by John D. Barrow and Frank Tipler, which views the universe's structure as expressible by bits of information in such a way that information processing is inevitable and eternal.

Source: Anthrophic principle - Wikipedia

In other words, when theists assert that the extraordinary fine-tuning of the fundamental constants of the universe, facilitating life, demands a theistic explanation, proponents of the anthropic principle often counter that such fine-tuning is unsurprising—after all, we were bound to exist within a universe capable of sustaining life, otherwise we wouldn't have been here to contemplate it.

Is the Anthropic Principle's rebuttal to the fine-tuning argument sound?

I argue it is not. Allow me to elucidate through an analogy.

The Sniper Firing Squad Analogy

Imagine a scenario where a criminal, facing the death penalty, is placed in the center of a vast arena, surrounded by 10,000 skilled snipers, each armed with a high-quality rifle boasting a 99% accuracy rate. Just before firing, each sniper meticulously ensures their equipment is in optimal condition.

If we presume each sniper operates independently, the likelihood of all 10,000 missing their target can be calculated as 0.01 ^ 10,000 = (1/100) ^ 10,000 = 1 / 10^20,000. This equates to a minuscule 1 preceded by 20,000 zeros in decimal notation.

As all 10,000 snipers take aim and fire upon command, the criminal, anticipating his demise, is astounded to find himself unscathed, with all bullets narrowly missing their mark, hitting nearby points on the ground around him.

In disbelief, he exclaims, "How is this possible? I should be dead. This must have been by design. Someone must have intervened or planned this."

In response, an advocate of the Anthropic Principle in the audience interjects, "Why the astonishment? Why seek a deeper explanation? It's simply because you exist in the universe where the snipers happened to miss. Otherwise, you wouldn't be here to pose the question."

Is this line of reasoning valid? If not, does it not undermine the objection posed by the Anthropic Principle?

  • Comments have been moved to chat; please do not continue the discussion here. Before posting a comment below this one, please review the purposes of comments. Comments that do not request clarification or suggest improvements usually belong as an answer, on Philosophy Meta, or in Philosophy Chat. Comments continuing discussion may be removed.
    – Philip Klöcking
    Commented Apr 2 at 11:51
  • The anthropic argument makes sense only if there is a multiplicity of universes. If indeed there are billions of universes with various sets of parameters, then it would make sense that at least one or more of them is able to generate life, and we would ne necessarily be in this one. But if there's only one universe then the argument is irrelevant. Same thing in your analogy, it assumes there are multiple universes, which is not demonstrated.
    – armand
    Commented May 25 at 6:48

12 Answers 12


Rather than replying to the proposed counter-analogy, I'm going to go back to the original objection:

When theists assert that the extraordinary fine-tuning of the fundamental constants of the universe, facilitating life, demands a theistic explanation, proponents of the anthropic principle often counter that such fine-tuning is unsurprising—after all, we were bound to exist within a universe capable of sustaining life, otherwise we wouldn't have been here to contemplate it.

The problem here is that the objection presupposes that Design is not the cause. Another Answer conveniently listed the proposed alternatives. Let's consider them:

  1. Multiverse: This is a completely philosophical construct for which we have no physical evidence. Its only possible justification is for the sake of rejecting a Design argument.
  2. Unknown Physics: This is a "nature-of-the-gaps" argument. Again, its only justification is "we don't like the Design conclusion".
  3. Other Deities: This isn't actually an objection, since it still supposes Design.

The commonality of all of the above is notable: "we'll accept any explanation for fine-tuning as long as it isn't the Christian God".

Arguments such as "improbable events happen all the time" aren't applicable, as most such events lack specificity. The likelihood of a specific arrangement of a deck of cards is low, but simply dealing those cards and shouting "look, a miracle!" doesn't work, because the likelihood that a random sorting of a deck of cards will result in some arrangement of a deck of cards is always 1.0. If, on the other hand, I were to write out an arrangement, then ask you to shuffle the deck and deal, and you deal the arrangement I wrote out, any reasonable person would call shenanigans (i.e. would declare that the outcome was subject to design and not chance).

This is called specificity, and refers to the inverse of the ratio of possible outcomes possessing some property as compared to all possible outcomes. If you inspect claims that an unlikely event can be attributed to chance, you will almost always find one of two things; either the prior specificity was very low, or the probabilistic resources were very high. For instance, if, in, the previous example, you were allowed to shuffle and deal the cards many (many) times before matching the arrangement I wrote down, it becomes more likely that you are able to do so by chance.

Scientists tell us that other possible values of physical constants would not only fail to produce life as we know it, but are almost certain to fail to produce any life. Naysayers claim that the probability of life happening is high, but the onus is on them to demonstrate either that the space of possible constants is high (that is, the circumstances resulting in life have low specificity) or that the probabilistic resources are high (e.g. multiverse theory). Note that https://christianity.stackexchange.com/questions/94341 is also relevant.

There is another problem, however; life is astonishingly diverse and has (according to belief in Common Descent) managed improbable events not once, but many, many times. (The eye, for instance, has supposedly "evolved" some dozens of times.) If we assume sufficient probabilistic resources, the chances of some life existing may be very high, but the chances of such diversity existing are still very low. To give an alternate take on the sniper analogy, life existing can be likened to a blindfolded sniper standing on the peak of a high mountain hitting a target miles away. If we suppose the sniper fires an infinite (or a very high, anyway) number of shots, then the probability of a hit approaches 1. Now let us say that intelligent life requires that the sniper makes ten shots, and that he stops once he does (equivalent to intelligent life reaching the point of pondering the fine-tuning question). The probability of such a universe existing, given infinite probabilistic resources, is 1, but our universe is one in which the sniper made all ten shots in a row. Yes, given infinite probabilistic resources, the probability that such a universe exists is still 1, but the probability that it happens to be this universe is very low. Thus, multiverse not only makes a philosophical assumption about probabilistic resources, it also assumes that we just happen to be in a particular such universe.

Let's look at some other objections:

Does a designer actually make things more likely?

Clearly, the answer is "yes". The probability of a random process producing an intelligible sequence of words is... low. The probability of an intelligent entity working toward the purpose of effective communication doing so is significantly higher.

Fine tuning has not been and can not be verified, only claimed.

Strictly speaking, since we lack the ability to create additional universes, this is true. What it ignores, however, is that we are able to model universes, and those models tell us that a universe with different constants would not produce life. The claims of fine-tuning have been demonstrated. The onus is now on opponents to refute those claims or demonstrate that their own position is viable. To my knowledge, this has not been done.

  • 5
    "The claims of fine-tuning have been demonstrated." On the contrary, they have only ever been theorized about. No demonstration has ever been made that the fundamental values that we call constants are in fact variable. Until you do that, fine tuning is nothing more than a thought experiment.
    – Corey
    Commented Apr 1 at 23:21
  • 1
    This is the only worthwhile material on this page.
    – Fattie
    Commented Apr 2 at 11:47
  • 1
    "The problem here is that the objection presupposes that Design is not the cause." Actually, it merely proposes an alternative explanation. The proposition that there are possible universes in which consciousness is not possible does not depend on rejecting the divine creation of this one. On the other hand, it would be begging the question to say, for example, that such universes are not possible because God would not create them.
    – sdenham
    Commented Apr 2 at 12:31
  • 1
    @Corey -- The claim of fine tuning has been PARTIALLY demonstrated. The Standard Model of QM assumes that the value of the constants are free variables. And the SM is immensely successful. And it predicts a CC (it is something like 10^250 off on magnitude!, but it does predict a non-zero CC). Spontaneous symmetry breaking is also predicted by current physics: mdpi.com/2073-8994/16/1/13 pnas.org/doi/10.1073/….
    – Dcleve
    Commented Apr 2 at 16:09
  • 1
    Also, inflation, which is a key feature in the BB model, assumes varying properties to the SM. Add the recently confirmed change in CC over the history of our universe, and most of the assumptions behind Fine Tuning have significant support. That is why few actual physicists question Apparent Fine Tuning today. The Multiverse plus Anthropic alternative, despite the lack of evidence for a Multiverse, is now widely held by physicists for this reason.
    – Dcleve
    Commented Apr 2 at 16:10

To me, a more accurate "sniper analogy" for fine-tuning would be:

You run between 2 points of cover in a heavy firefight. There was probably hundred bullets hitting somewhat near you without hitting you. You assume that every single one of those bullets were shot by each of 100 different well-trained seasoned snipers specifically aiming at you with no significant distractions, such that their accuracy would be extremely high. You then conclude that you generate an invisible bullet-deflecting shield around you, because "the" alternative of them all missing is just too unlikely.

To break it down more concretely:

  • We have no idea how likely it is for the physical constants to be what they are, we just know that they are what they are. Theists tend to make some bold assertions about how likely it is, but that's typically nothing but speculation.

  • We have no idea how likely God's existence is, which we need to know to consider it as an alternative. Even if you aren't going down the route of strict probabilistic reasoning, contrivance and explanatory power still comes into play.

    Related answer where I detail this argument more concretely and mathematically.

  • God isn't the only conceivable alternative.

    • The multiverse is another alternative.
    • Or we can suppose there's some as-yet-unknown law of physics. This is distinct from the first point, of us just having no idea how likely it is, rather than it seeming unlikely and having a law of physics make that more likely.
    • Or we can suppose any other conceivable deity or set of deities did it. One might respond that whichever deity or deities did it can be called "God", but that's a long way from attributing to that deity all the things that "God" supposedly did according to whichever religion a theist is trying to argue for.

It's reasonable to suppose that some very unlikely explanation may not be the case, and that there could be some more likely explanation. But that's a long way from concluding that some other specific explanation is in fact the correct one (also, very unlikely things happen all the time).

Related: a more detailed response of mine to fine tuning.


The fact that the odds, P, of the criminal being missed was one preceded by 20,000 zeros does not make it more likely that some other cause was at play. In particular, it does not mean that the probability of some other cause was 1-P, or 0.999... with 20,000 nines. If you think otherwise, you are confusing the odds of something happening given that it happens randomly, with the odds that the cause is random.

As an example, consider a lottery in which the chance of winning the jackpot- assuming it is won at random- is one in a million. If you win, the chances were 0.000001. That does not mean that the odds of you winning in another way- by cheating, for example- were 0.999999. The odds that you won by cheating are not 1-P, where P is the probability of winning if the lottery is won fairly, so decreasing P does not make cheating more likely.

  • 3
    This is wrong. Winning the lottery does make it more likely that you were cheating (compared to if you lost the lottery), and the more unlikely the win, the greater the chances of cheating given that you won. It does not mean the chances of cheating were 1-P. The chances of successfully cheating a lottery are very small too, one in many millions. But we may suppose the chances of cheating a lottery are relatively independent of the chances of winning, so that if the chances of winning fall below that threshold, it becomes more likely that the winner cheated than that they won fairly.
    – causative
    Commented Mar 31 at 22:28
  • 6
    In the extreme case, if there is a chance of exactly 0 of winning the lottery fairly (because it's rigged), but someone appeared with a winning ticket, then the winner definitely cheated.
    – causative
    Commented Mar 31 at 22:32
  • 1
    @causative no, you are misreading what I have said. Suppose the only way to cheat a given lottery is to guess a ten-digit administrator password at a single try. The odds of you winning by cheating (assuming you try to cheat) are 0.0000000001. Whether the chances of winning by chance are one in ten or one in a million or one in a trillion does not affect your chance of winning by cheating. Commented Apr 1 at 7:20
  • Your chance of winning by cheating depends only on the length of the administrator password. If the chance of my winning the lottery by chance was one in a trillion trillion trillion, you cannot say it was more likely that I cheated, if the administrator password had fifty digits say. In other words, the odds of me winning by chance is independent of the odds of me winning by cheating, and knowledge of the one does not give you information about the other. Commented Apr 1 at 7:24
  • You can argue that the Universe turning out as it has is vastly improbable. However, to say whether the alternative is more likely, you need to model the odds of the alternative. Commented Apr 1 at 7:26

You are correct that in the sniper analogy, the extreme improbability of all the snipers missing by chance does make some other explanation considerably more likely. Practically certain, given the very low probability of them all missing by chance.

Before I go on, let me mention that the chance of a deity existing with the properties described in any specific religious tradition is also an extreme case of fine tuning. Much more extreme than any scientific theory, due to the inherent complexity of a thinking, acting being. That said:

The Anthropic Principle doesn't really justify any particular hypothesis about the universe's creation. The Anthropic Principle is only post-hoc. If a prior hypothesis is complex (has a long minimum description length), then that hypothesis is a priori unlikely, and the Anthropic Principle does not make it more likely.

The prior probability of a hypothesis is exponentially lower as the minimum description length grows.

If a physical theory requires certain constants to be fine-tuned to particular values to obtain the universe as we see it, then every bit in those values does add to the minimum description length of the theory. And it's not enough simply to get the bits in the right range to permit intelligent life; if a theory demands a constant be a specific measurable value, then every single bit in that value to the finest precision we can measure, must be counted in the description length of the theory.

Solomonoff's theory of inductive inference is the ultimate word on how we should reason about hypotheses about the universe's creation. Roughly speaking (simplifying a bit), if M is the minimum description length of a hypothesis, the prior probability of that hypothesis is (approximately) Ak^(-M) for some base k and constant A. Then, you simply check, for each of the (infinite) possible hypotheses, whether that hypothesis exactly matches observations. If it doesn't match, you cross it out. Then, you sum up the probability mass of all the hypotheses not crossed out, to get a normalizing constant Z. The posterior probability of a hypothesis is then Ak^(-M) / Z. And roughly speaking, in practice, the hypothesis with minimum M wins and gets a probability near 1, and all other hypotheses lose and get probabilities near 0.

So, it all comes down to whether M_T, the minimum description length for a scientific theory T, is shorter or longer than M_G, the minimum description length for a thinking deity. If M_T requires fine-tuned constants, then that hurts it in comparison to M_G. But M_G is likely very, very long; how long would be a computer program that would let you simulate a human being? And it's not enough to specify just any intelligent being; M_G has to be a specification of an intelligent being that would produce the exact universe we observe. So if M_T is still under a few kilobytes, then M_T probably still wins, by a landslide.

  • 2
    @ScottRowe There isn't just one God Hypothesis - there are an infinite number of them. Each would be a computer program that exactly generates a complete universe, and that includes as part of its code a specification for a God entity that does the rest. Some of them (an infinite number, actually) do exactly match observations, because they fine-tuned the properties of the God entity so it does generate exactly what we observe. So it's not a question of whether the God hypothesis matches observations, but about how long the shortest God hypothesis that matches observations is.
    – causative
    Commented Mar 31 at 23:35
  • Has there ever been a theistic system of thought in which it's expected that the Deity can be described with an algorithm?
    – adam.baker
    Commented Apr 1 at 6:11
  • @adam.baker Many such. Gottfried Leibniz, for example, modelled God as a utilitarian optimising agent seeking the fittest of all possible, non-contradictory universes, and who would never make arbitrary decisions (the principle of sufficient reason).
    – wizzwizz4
    Commented Apr 1 at 19:51
  • Solomonoff's crackpot discussion (it doesn't rise to a "theory") is no better or worse than other crackpot discussion about the (various utterly different) meanings of the "anthropic principle" (indeed, it's nothing more than adding some equations to "Occam's razor", philosophy's most completely valueless concept). Even Richard Dawkins (for goodness sake) firmly believes that the reality you are experiencing as we speak could be a simulation, and that's (just one) explanation that washes away such explanations, much as pre copernican explanations of planetary movement were washed away.
    – Fattie
    Commented Apr 2 at 11:55

Is this line of reasoning valid? If not, does it not undermine the objection posed by the Anthropic Principle?

Not particularly, no.

Let’s get the kneejerk reaction out of the way first: If you caricature proponents of the anthropic principle as saying

Why the astonishment? Why seek a deeper explanation?

you should consider to check your own verbal sniper rifle. It is missing the mark - many proponents of the anthropic principle are astonished by the universe and do seek ever deeper explanations.

The other thing the scenario is missing is that it only aims at the target you want to be dead. But the elephant in the room is this: Does a designer actually make things more likely? The argument is one of probability, after all, not of possibility (the Anthropic Caricature probably chuckles at this point, realising the mark was missed again).

The situation: There are 10,000 skilled snipers, each armed with a high-quality rifle boasting a 99% accuracy rate. Just before firing, each sniper meticulously ensures their equipment is in optimal condition. So there are lots of highly experienced people with the intention to kill and who have ruled out tampering.
How, pray tell, could someone "design" this situation to go the way it allegedly did? How likely is it that 10,000 weapons were sabotaged so that trained experts could not spot it yet still hit the mark almost perfectly? That 10,000 people were bribed, threatened, indoctrinated (or whatever the Anti-Anthropic Caricature would do) without fault or notice? That someone would go through all the trouble instead of, say, bribing a judge and letting the poor guy just walk free?

Unless we beg the real question and say a wizard designed it, a designer is, pardon my French, not any more likely.


I see two problems with the "sniper" rebuttal.

Firstly, the Anthropic Principle supposes that there is vast range of possible universes, and then says that the one of those which we observe is one in which we're capable of living to observe it. But your scenario supposes that there is one convict sentenced to death by firing squad. Suppose instead that your scenario plays out 1020,000 times with that many different convicts. Then it is not so remarkable that some convict survives his ordeal; and the one who survives is of course the one who gets to experience being alive afterwards.

Secondly, the Anthropic Principle supposes that a universe which cannot support life cannot be observed at all. But in your scenario, it is quite possible for the convict to observe himself being shot and suffering before he dies, if that would be the outcome, and it would be quite possible for everyone else there to observe his death. It is only the "being dead" part of that outcome which the convict himself cannot experience. So in your scenario it is surprising that the convict survives, because the other outcome is observable (and much more probable).

  • Comments have been moved to chat; please do not continue the discussion here. Before posting a comment below this one, please review the purposes of comments. Comments that do not request clarification or suggest improvements usually belong as an answer, on Philosophy Meta, or in Philosophy Chat. Comments continuing discussion may be removed.
    – Philip Klöcking
    Commented Apr 2 at 10:18
  • "the Anthropic Principle supposes that there is vast range of possible universes" Not at all. That is one (totally crackpot) thread of discussion that comes up when the anthropic principle is discussed.
    – Fattie
    Commented Apr 2 at 11:56
  • @Fattie Note that "possible" doesn't imply they exist. Quoting Wikipedia: "Proponents of the anthropic principle argue that it explains why the universe has the age and the fundamental physical constants necessary to accommodate conscious life, since if either had been different, no one would have been around to make observations." So it's supposed that the fundamental constants could possibly be different, i.e. that universes with other constants would be possible. If one rejects that, then there is nothing to explain: our universe has these constants because others would be impossible.
    – kaya3
    Commented Apr 2 at 16:26
  • @kaya3 not sure what to say. (The fact that someone happened to type that on Wikipedia is no more or less interesting than any comment typed here.) As discussed endlessly in the removed comments, "anthropic principle" is used for many ("Bostrom says there are 30") wildly different, profoundly different in paradigm, concepts; "multiverse talk" is heard, in some subset of those, in some subset of the theories about each of those.
    – Fattie
    Commented Apr 2 at 16:40
  • @Fattie The removed comments like this one? As I said, the idea that different physical constants are possible doesn't entail that a multiverse exists. You are telling me something I have said myself.
    – kaya3
    Commented Apr 2 at 16:50

Physicists do say things like if the cosmological constant were changed 1 part in 10^10 or perhaps 10^100 there'd be no life/macroscopic objects. But that doesn't imply there is 1/10^10 or 1/10^100 probability the constant took that value. It could be that there is such a (low) probability, but it's not implied by the fine-tuning. After all, if god made it that way the probability is 1. This is why the sniper analogy doesn't change anything. Do the snipers have 1 / 10^20,000 probability to miss or not? You can't be certain about the probability then vacillate after the event. If you determined the probability to be 1 / 10^20,000, and lived, you're just that lucky. That's just logic.

When applied to the universe, it helps localize ourselves/macroscopic objects amongst a vast potential of possibilities. In the sniper analogy, it serves the same purpose, contextualizing your survival.


Is the Anthropic Principle's rebuttal to the fine-tuning argument sound?

In short, yes it is.

Take a deck of cards and suffle it as much as you like. When you've sufficiently randomized the deck, deal 10 cards face up.

What are the odds that you dealt exactly those cards in that order?

100%. Because you just did it.

And that's what the Anthropic Principle fundamentally expresses. Not that there was some agent choosing the sequence of cards, but that past probabilities are always 100%. It doesn't matter what the odds were prior to the start of the process, after the fact the other possibilities (in this particular world, for those of you who prefer the Many Worlds hypotheses) have been eliminated.

Is this line of reasoning valid? If not, does it not undermine the objection posed by the Anthropic Principle?

Your blah-blah about some crazy hypothetical does nothing to support anything. It's just a setup for a slander against opponents to fine tuning. Since nothing in your example is even remotely related to fine tuning other than the presence of made-up scary numbers, it's doesn't even relate to the Anthropic Principle. Your parody of the objection raised by your straw man audience member is just absurd.

Of course the Anthropic Principle is just an observation. If you want an actual objection to fine tuning there are plenty of them.

For instance: fine tuning has not been and can not be verified, only claimed.

Fine Tuning proposes that there is a continuum of possible values for some set of the fundamental values, but offers no substantiation for this point other than "some physicists think so." We have no ability to examine anything other than our own universe, so we actually cannot validate this point. We have no way to investigate what if anything is the reason for these values being what they are, so any assumptions drawn from their variability are suspect at best.

We do know that there are some fundamental constants that are truly constant. For instance, there is no possible universe where the value of Pi - the ratio of circumference to diameter of a Euclidean circle - is other than what we observe it to be. The value of Pi is necessarily 3.1415926535897... and so on. The same is true for some other constants. The existence of true constants, values which are invariant among all possible version of reality, is a simple fact.

The variability of the fundamental values of our universe is not a simple fact, it's a hypothesis that cannot be tested. It's an interesting thought experiment, nothing more.

And without a solid proof that these fundamental values can in fact vary, any and all hypotheses that derive from that variability - I'm talking about fine tuning here, just in case you missed it - are equally unverifiable.

Fine tuning requires no further refutation.

  • This answer comes off as overly emotional and resentful. Commented Apr 1 at 17:48
  • @RussellMcMahon I specifically wrote "Euclidean circle" to exclude warped space.
    – Corey
    Commented Apr 1 at 22:17
  • @JustSomeOldMan The question came off as condescending and overly idiotic. What's your point? Tone policing doesn't address my point. Care to offer an actual objection?
    – Corey
    Commented Apr 1 at 22:19
  • @ac15 You may be right BUT in a a reality where the geometry is different what do the series expansion's "mean". I imagine (but do not know) that one could conceive of realities where the expansions were if anything just mathematical trivia - "Imagine that there was a universe where ...". That does not of course make them "wrong" - just pointless for practical purposes. Perhaps :-). Commented Apr 2 at 12:52
  • The assumptions behind the Standard Model of QM hold that the SM values are free variables for our universe's physics. There are additionally multiple other free variables that apply, such as when and to what degree symmetry breaking created a positive baryon count for our universe: mdpi.com/2073-8994/16/1/13. Physicists think that Fine Tuning is real, because their successful models say our universe is wildly unlikely to be able to support life. You can reject expert opinion of course, but that is to reject rationality and naturalism.
    – Dcleve
    Commented Apr 2 at 15:38

If we accept that

  • universe=fine tuned for life = TRUE
  • and that this implies it was designed or engineered = TRUE

Then isn't it also true that...

Evidence of a design phase having taken place billions of years ago, resulting in septillions of planets, some with life...

... is... evidence of a design phase having taken place billions of years ago, resulting in septillions of planets, some with life.

And not more than that?

  • Yes and no. Yes, our universe being fine tuned for life, does not support religious views of the designer. However, the no is it does allow us to spell out designer hypotheses, and then test for consequences from them. Our universe, for instance, is "tuned" for life, but far from "optimized" for life. Omni-creator-Gods can reasonably be expected to be able to optimize to achieve their goals, so this is a contrary observation for any Omni-God. What creator would fail to optimize? Perhaps a student still learning? Or a dysfunctional committee?
    – Dcleve
    Commented Apr 2 at 15:45

The Strong Antrophic Principle (which the question is not about) says that the universe was configured just so, to enable human life; i.e. it includes intent.

The Weak Antrophic Principle (WAP) does not include intent, and makes no statement on probabilities, but only states that we of course live in an universe supporting life, because if it would not support life (or to be more exact, an observer), we could not be here.

Your sniper example has been hand-crafted to specifically lead to the end result that was wished, by fine-tuning the probabilities just so.

So, TLDR, the one has nothing to do with the other.

To speak to your example again: it tunes the probabilities so the observed outcome was very, very unlikely, and also openly makes very clear statements about the actual probabilities, by saying something about 99% accuracy, describing the practice of the shooting and so on. From that it infers that there must be intent. And it is correct! There was some intent. Maybe all the shooters had secret orders to miss the target. This would be a very simple explanation, it would explain everything perfectly well, it would not require mystical reasons and so on and forth. Heck, one could even come up with valid reasons for this when writing a book or movie - maybe the whole operation was designed to scare the victim out of their mind, and play some awful mind game for whatever reason. Weirder stories have been written.

But this has nothing to do with the WAP. As mentioned, the WAP makes no assumption about probabilities of life, nor of whether slightly tweaking any of the universal constants make life impossible (maybe our style of life would be - but who says that other forms of life would not be possible), nor how many universes there might be (serially, one after each other, or multiverses). There are modern arguments (roughly following Darwinism) which explain very convincingly how the fundamental concepts of evolution do not only apply to fully formed life, but also to early states where non-biological chemicals form themselves to more complex structures and so on and forth (The Selfish Gene is a relatively old, but still fascinating read, and the current edition has many footnotes mentioning related, later works). So all in all it could perfectly well be that the probability for life in an universe, even with different universal constants, would be high. Again, the WAP does not require any statement about that. This is just to explain that the two arguments have little to do with each other.

More reading in the SEP, in the page on Fine-Tuning.


Short Answer

The Anthropic Principle is not a rebuttal to Fine Tuning, it is an alternate explanation for why it is observed.

Many philosophical thinkers, including many of the answers posted here, come up with invalid rationalizations to reject fine tuning, by resorting to various versions of "things are as they are". Note this attitude is to embrace the "unexplained brute fact" leg of the Munchausen Trilemma, AND to explicitly abandon the core principle of naturalism -- which is to seek for an explanation for why things are as they are.

Your sniper analogy is aimed at those rationalizations, not at the Anthropic principle, hence it is not a refutation of Anthropic thinking at all.

Note, to get the multiple events that are needed for the Anthropic principle to operate upon, the Anthropic alternative to theistic creation is generally assumed to require a Multiverse. A noteworthy exception is Lee Smolin, who absolutely rejects multiverses. Smolin agrees our universe appears to be Fine Tuned, and his non-theistic Anthropic alternative is to invoke Deep Time, and postulate our universe has a near-infinite history of random draw Big Bangs.


Here is an outstanding description of how the Multiverse could work, under the presumption of String Theory. In my review, I tried to compare the assumptions involved to those of theism: https://www.amazon.com/gp/customer-reviews/R3JVQDAK1408BR?ref=pf_ov_at_pdctrvw_srp

Here is Smolin's takedown of String theory. https://www.amazon.com/gp/customer-reviews/R6JY5GBLAV2BV?ref=pf_ov_at_pdctrvw_srp

And here is Smolin trying to articulate an alternative that might work better than string theory. https://www.amazon.com/gp/customer-reviews/R16VWWZ5I5SC8Q?ref=pf_ov_at_pdctrvw_srp


Your objection seems to be just that 1020000 is a large number, and you don't believe that there are that many worlds.

I think you have to concede that there are that many possible worlds (barring determinism). Any time someone generates 20,000 random digits, which happens all the time, there are 1020000 − 1 possible worlds where they got a different result. So your objection amounts to saying that 1020000 is a plausible number of possible worlds, but not a plausible number of actual worlds. It can't be that that much stuff exists; where would you put it?

The seeming largeness of 1020000 is based on comparison with everyday experience. But everyday experience is, well, anthropically selected. Competition for scarce resources is probably necessary to drive the evolution of high intelligence, so any evolved intelligent life will find itself in a place with scarce resources. It doesn't follow that scarcity, at that level or at any other level, is a fundamental attribute of reality. On the contrary, the universe shows every sign of not having been made on a budget, or at least not on a budget whose size is comprehensible to human beings. The most obvious example of that is the largeness of the visible universe (and evidence from the uniformity of the CMBR that the universe extends far beyond what's visible), but it gets worse when you consider quantum mechanics. You can reduce the size of the state vector describing a system to a finite number with some plausible assumptions about quantum gravity, but it's a rather large finite number. You could imagine a Planck-scale grid, but you get a smaller result (hence better for your argument) if you suppose the number of degrees of freedom of a system is limited to the Bekenstein entropy, which only grows as the square of the radius. This entropy works out to about (r / 10−35 m)2. If you take r = 1 angstrom, you get an entropy of 1050, which suggests that an isolated system about the size of an atom needs e1050 ≈ 101050 complex numbers for its description. That's without any other worlds; it's just what seems to be needed for ours. In fact, supposing there are other actual worlds doesn't make these estimates any larger.

You can't sweep arbitrarily many improbabilities under the rug by supposing existence is large enough, because eventually you run into the Boltzmann brain problem. But you can explain 20,000 unlikely accidents that way.

  • In our universe, the rug could be pretty big.
    – Scott Rowe
    Commented Apr 3 at 1:12

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .