The anthropic principle states that I'm a typical sapient-level observer. What this implies is that if there are two groups of sapient-level observers, then I am most likely to find myself in the larger group. For ex: in a two-party country where voting is mandatory, my favorite candidate is most likely to win. Similarly, if there are two groups of sapient-level observers (us and aliens), then I am most likely to be part of the larger group. Since I am part of mankind and not part of the trillion member alien race that has colonized the galaxy, there's a high chance that such an alien race does not exist. The only other probable scenario is that there are many planet-bound sapient-level species to counteract the existence of a single godly alien race. In such a case, it wouldn't be too rare to find myself as part of a planet-bound sapient-level species.

Is there a way to know which one of the two scenarios is more likely? For ex, wiping my memories and then doing a self-check, noting down which group I'm a part of, and then repeating the process over and over again?

  • 9
    Anthropic principle says that the universe's laws are compatible with your existence, it does not say anything about aliens, their numbers, or colonizing galaxies.
    – Conifold
    Aug 18, 2020 at 12:15
  • 3
    @Conifold The phrase "anthropic principle" is often used for the stronger statement that you should treat yourself as a uniform random sample from the set of "conscious observers" or something of that sort. Look at the literature on the doomsday argument for example.
    – benrg
    Aug 18, 2020 at 17:11
  • 2
    I am born in Switzerland, if there really were billions of people outside of Switzerland I probably would have been born there, therefore, the rest of the world does not exist ? Or Switzerland does not exist ? Or... whaaat ?
    – armand
    Aug 18, 2020 at 23:08
  • 1
    you're lying you're an alien @armand
    – user47711
    Aug 18, 2020 at 23:13
  • 1
    @ask_hole I stand corrected. All your base are belong to us.
    – armand
    Aug 19, 2020 at 3:47

4 Answers 4


Roughly, the anthropic principle identifies the fact that the observations we make about the universe are conditioned on the universe supporting observers—that is, conscious, sentient organisms capable of making observations.

Similarly, if there are two groups of sapient-level observers (us and aliens), then I am most likely to be part of the larger group. Since I am part of mankind and not part of the trillion member alien race that has colonized the galaxy, there's a high chance that such an alien race does not exist.

This is incorrect. Let me put it this way:

Even if aliens outnumber humans 100 to 1, some humans still must exist and thus some humans must be able to make claims and observations (similar to the anthropic principle). While it is true that it's more likely that you would find yourself living in an alien civilization, some life would still have to be human, and thus the prospect of you being one of those humans is not so unbelievable after all—someone had to do it.

So, no, we cannot make a claim about the plausibility of alien life just by you being a human.

  • What about the "all ravens are black" thought experiment then? Finding a black raven increases the chance of "all ravens are black" being true. And how does the conditional "if there are two groups," have anything to do with the conditional "if an observer finds himself in a group," ? Those are different and truths about one does not have any effect on the truth of the other.
    – user39495
    Aug 18, 2020 at 20:28
  • The self-sampling assumption is a widely used version of anthropic reasoning (for example, it's used to argue against cosmologies in which Boltzmann brains would vastly outnumber evolved observers in low-entropy eras of the universe), and it does allow for Bayesian updating on hypotheses about the numbers of various types of observers unlike ourselves--see my answer here
    – Hypnosifl
    Aug 19, 2020 at 20:33
  • @Jurhas I don't see how that's relevant. The claim "aliens do not exist" does not become more likely to be true simply because the question-asker is a human observer. The point is that even if aliens outnumber humans 100 to 1, both groups exist and produce observers. Thus, why shouldn't the person asking the question (i.e. the human observer) be a part of the 1?
    – natojato
    Aug 20, 2020 at 0:58
  • @natojato, the claim isn't "alien don't exist", but "all observers are human", just like "all ravens are black".
    – user39495
    Aug 20, 2020 at 6:02
  • Aliens are not humans, therefore if all observers are human, aliens don't exist? Aug 20, 2020 at 21:52

There are different types of anthropic reasoning, and to make it precise it is necessary to specify the sampling assumption through which you / your body was chosen -- for example the SSA vs SIA question. In principle one might have a model where humans have some property that makes them more likely to be sampled -- for example, if the alien race consisted of intelligent machines without any consciousness, would we still expect them to be part of the "sampled" class? You describe the sampled class as "sapient-level observers" but one might have a different theory. As another example, one might have a model where first we sample an intelligent species (same probability for every intelligent species) and then we sample one member from that species. In that case the fact that there are fewer humans than colonizing aliens does not make it less likely that a human would be sampled.

That aside, back to your original question which supposes that every intelligent entity is equally likely to be sampled regardless of species. Under that assumption, I'm inclined to agree with you that the fact that I'm here is some evidence that the number of planet-bound intelligent creatures is at least of the same order of magnitude as the number of colonizing intelligent creatures, but it seems we can conclude little more than that and that is consistent both with there being many of the former and with there being none of the latter. There may of course eventually be other forms of evidence, e.g., we may eventually find out that colonization is effectively impossible due to some kind of space radiation or something. But I think the main takeaway is that we still don't really know how anthropic reasoning is best done, which is why for example the Doomsday argument and the Sleeping Beauty puzzle remain controversial.

  • 1
    I'll have to read Bostrom's book. I'm currently reading Carter's paper about the implications of anthropic principle. I agree that there's a sort of selection bias within the selection bias, ad infinitum, a sort of meta anthropic reasoning. And it must be taken into account by any observer making claims born out of anthropic reasoning. Because not all humans make these claims, nor are even aware of this to begin with so they have to be excluded. If you are here, it implies the number of colonizing intelligent creatures is limited by an upper bound within an OoM. The lower bound can be 0.
    – user39495
    Aug 18, 2020 at 20:40

So it seems the question is bounded to sentient species in our galaxy (the milky way), and assumes a variant of the Drake equation that there are many life-supporting planets in our galaxy, and sentient life could have arisen on any of them, and in theory a sentient species might colonize the galaxy. It also assumes that humanity has not yet been in contact with any alien species.

And the question is whether by the anthropic bias we have to believe that our galaxy is actually dominated by some space-traveling species or whether we have to believe that no species has successfully colonized large areas of our galaxy.

I believe by the anthropic bias we have to assume that we are not unlikely observers, so the observations we make about our own species are most likely to be true for most species. Since colonizing space does not seem to be easy for us, it would seem most probable that it is neither more easy for other species, so a dominating species colonizing large parts of our galaxy seems statistically unlikely based on the anthropic bias.

Note the "anthropic bias" is called bias for a reason, it favors one hypothesis based on knowledge being limited, and if we have more knowledge, we can make better estimates of the truth.

So every astronomical observation we make will hugely improve our reasoning over the mere anthropic bias. As an example the assumption that we are not an unlikely/exceptional randomly picked observer might be falsified if we can determine earth today is an unusual place compared to other planets or times.

This is related to Fermi Paradox:

  • If there is a dominant alien species colonizing or galaxy, why haven't we seen it?
  • If there is a dominant alien species colonizing or galaxy, why am I as randomly picked observer not part of it?

Personally I'd say that the Anthropic Principle is not the type of thing that can be said to prove anything.

There's a very trivial and true version that says roughly that a statement must be formulated within a context that permits it being formulated (e.g. if I'm making a claim then the universe must be such as to permit this to occur).

Any claims based on the AP beyond this utterly trivial and banal observation are at best tendentious and at worst completely unwarranted. No empirical claim (such that there are/aren't aliens) can be proven by the AP.

  • How so? Can't you prove that "all ravens are black" by the simple fact that you don't find white ravens?
    – user39495
    Aug 22, 2020 at 8:55
  • No way. You certainly can't prove that all ravens are black by that method, unless you have an auxiliary premise that you would definitely have found the non-black raven if there were one - and you can't know this so you can't have that premise. And that isn't an application of the anthropic principle anyway. Aug 22, 2020 at 9:18