Suppose we go to Mars and come across what appears to be a highly advanced technological device of extraterrestrial origin. Let's call this object X. Would the claim "X was designed" be falsifiable? That is, if we assume that the hypothesis "X was designed" is true, are there any testable predictions that could be made such that if those predictions are shown to be false, the starting hypothesis that "X was designed" would be shown to be false?

I'm proposing this thought experiment in the context of recent discussions about the scientific status of Intelligent Design, whose primary focus is the apparent design observed in biology. Likewise, if we observe apparent design in an object found in extraterrestrial conditions, would it be possible to establish scientifically that such an object came about by design and not by natural processes?

Clarifying what I mean by testable prediction

By testable prediction I mean inferences of the form: If X is true, then Y should be expected to happen or have happened, such that we can arrange some sort of experiment to verify whether Y actually happens (or has happened).

In this particular example, testable predictions would be of the form: if this extraterrestrial object was designed, [fill in the blank] should be expected to happen (or have happened), which naturally raises the question: what would we expect to happen or have happened if an object was indeed designed as opposed to having come about by natural processes?

  • 9
    I doubt relocating Paley's watchmaker analogy to Mars will lead to anything that has not already been said.
    – sdenham
    Jun 5 at 18:44
  • 1
    Sure. After scientists study processes on Mars that shape natural formations and create a model that shows how the "device" emerged from them, see geofacts.
    – Conifold
    Jun 5 at 19:11
  • 9
    I'm wondering if this Question was designed for something.
    – Scott Rowe
    Jun 5 at 20:03
  • I'm wondering what an advanced extraterrestrial device would look like.
    – Ludwig V
    Jun 5 at 20:25
  • 3
    Is "aliens hiding monoliths on Mars for fun" not itself a natural process? Jun 6 at 5:53

8 Answers 8


Scientific theories and falsification are not generally questions about specific instances, but applied to general processes and statements about nature. Theories make predictions, then you perform experiments to see if the results are consistent with the theory's predictions (adding support to the theory) or inconsistent (falsifying it).

But for specific instances, the "scientific method" above doesn't really apply. For some random thing you find, a scientist would start by trying to classify it among known objects that it's similar to. If the processes that create these objects are known, the initial assumption would be that it was created using similar processes -- the "if it walks like a duck" method. So if you find a clock, and you know that clock-making is something that people do, it's most likely that it was created by a clockmaker.

Conversely, if you find a living thing, the obvious assumption would be that it was created by evolution through natural and/or directed selection, since we know that this is the usual process that produces new organisms.

Once you make that initial assumption, you can then examine the object more closely, to see if you can determine more specifically how the creation process applies.

If the object doesn't bear any resemblance to things that we already know about, or you can't figure out the specifics, we have a mystery. If an existing theory comes close to explaining it, we may need to make adjustments to that theory to accomodate it; for instance, once we discovered the genetic mechanisms underlying evolution, there were tweeks to accomodate genes jumping between species.

If not that, we need to do more investigation. Hopefully we can eventually find more instances of this new object (and the same thing applies to phenomena). With more examples, we try to find some commonality (essentially creating a new class of objects), and form hypotheses about this class. Only then can we make predictions about the class that can be confirmed or falsified.

Of course, statements about an individual object can be true or false, but it isn't always possible to prove all such statements. The claim "X was designed" can be shown to be true if you can find the designer. But as the saying goes, you can't prove a negative: if you can't find the designer, you can't be absolutely sure. But if you know natural processes that commonly produce the object, it's not unreasonable to assume that it was not designed until something contradicts that assumption.

As a simpler thought experiment, consider Nelson's "grue", an object that's green before time t and blue after t. If you encounter a green object before t you can wait until t and see if it becomes blue. But if you encounter a blue object after t, you can't determine whether it was green before t. (And both these statements ignore the infinite times surrounding the observations, for simplification purposes.) However, if there are other objects that have been determined to be grue, and this object shares many properties with it, we may naturally assume it's also grue (similar to the way we conclude a creature is a dog or cat based on superficial observations).

We can't always be sure of everything. There are still plenty of holes in our knowledge. Science is a process of progressively filling in those holes, although it also reveals new holes. For instance, as we discovered more details about cosmology, we learned that dark matter and dark energy exist, and now we're investigating what they're made of.

  • 1
    This misrepresents science. Medical diagnosis is absolutely science - it's among the oldest parts of it. Which is to say, making a hypothesis & testing it about a specific instance. There is a perennial problem with 'physicsisation' of science, & lack of knowledge about how much methods vary among different sciences. It seems remarkable you mention finding a clock, but not the Blind Watchmaker version of the argument-by-design. The grue thing implies never gaining a ground-up picture of causal processes, but science aims at that, not just taking colours for granted, before or after time t.
    – CriglCragl
    Jun 7 at 18:35
  • 1
    The universe is an "individual object," and the laws of physics are properties of it, and thus by your reasoning the scientific method wouldn't apply to the universe. The only (relevant) difference between the universe and a small object is scale. Bayesian inference applies equally well to objects of any scale.
    – causative
    Jun 7 at 18:46

Yes; perhaps X was the result of random evolution of self-reproducing machines, like evolution on Earth but for machines. In that case, finding sufficient evidence of the evolution process (fossils/ancestors, other machines that are observed to reproduce with random variation) would falsify the notion that the device was designed.

  • 1
    But what would be the testable prediction in that case? The testable prediction would be "If X was designed, you would never find sufficient evidence of evolution"? Because that's exactly what ID proponents claim.
    – Mark
    Jun 5 at 17:24
  • What I mean is that, you are basically suggesting one possible way to find out that the object was not designed, but in so doing you are making a case for the falsifiability of ID as well. Is your position that ID is falsifiable? And there is also the issue that there is no positive prediction being made in favor of the design hypothesis, only a prediction about a competing hypothesis (evolution).
    – Mark
    Jun 5 at 18:43
  • Maybe humans are randomly designed self-reproducing machines.
    – Scott Rowe
    Jun 5 at 20:07
  • 4
    @Mark Yes, the prediction is, under the assumption X was designed, you would never find sufficient evidence of evolution. There are positive predictions as well - that you would find sufficient evidence of it being designed if you looked in the right places, e.g. looked through the patent library of the alien civilization that you propose made it. It's very hard to directly falsify though, because even with lots of evidence of machine evolution through natural selection it would still be very hard to completely rule out an alien somehow manually intervening at some point in the process.
    – causative
    Jun 5 at 21:04
  • 1
    @CriglCragl Sure, it could have been placed there by humans secretly. Actually that's probably the most likely explanation for an artifact on Mars that appears to be of extraterrestrial origin.
    – causative
    Jun 7 at 19:10

There are different branches of science doing different things, which are all "scientific" in a way, but still different. The requirement of making falsifiable predictions is of utmost importance for natural sciences, the branch of science investigating natural (non-designed) phenomena.

For investigating a single technological device, it is unlikely that anyone would consider the question of whether it evolved, and so the origin of such a device would be investigated with applied science instead. This would be a bit similar to forensics when investigating a crime scene to find out what happened, e.g. based on fingerprints. In applied science, making falsifiable predictions is not required to "solve a case". (Not all of forensics is applied science of course, some researchers in criminology need to find out things like how long a poison takes to kill, but this is done independently of solving individual cases.)

So the question of how such a thing that looks like alien technology came into existence would first be addressed by applied science without making falsifiable predictions. And only in the unlikely case that applied science finds the device emerged naturally would natural sciences become involved in finding out how, trying to find theories and falsifiable predictions.

So the question is flawed because it presupposes that applied sciences work like natural sciences. A better thought experiment than in the question could possibly be created.

  • Could ID be considered "applied science" instead of "natural science"?
    – Mark
    Jun 6 at 6:19
  • 1
    Feel free to ask new questions as questions, not comments. Also feel free to read up on a topic before asking more questions.
    – tkruse
    Jun 6 at 6:23

There are a lot of valid, well-designed experiments which may be expensive, inconvenient, unethical, or demanding an unrealistic amount of patience, but I put it to you that these mere practical considerations are just details of implementation.

As such, one could arrange a series of experiments that reproduce the original natural conditions and natural processes on Mars, and see if one ends up with an alien iPod at the end.

Even this would only falsify the narrower hypothesis that "X was necessarily designed", which may or may not be sufficient, depending on who is asserting what to whom.

  • 2
    We could have, like, an infinity of universes, and try all these experiments and see what happens. Design by not designing. Very post-modern.
    – Scott Rowe
    Jun 6 at 0:01
  • 1
    Like the Miller-Urey experiment?
    – Barmar
    Jun 6 at 14:07

I am a designer and I like to think that I am intelligent enough to shed some light on the subject of design.

All designs aim to serve a purpose. There is no design without any purpose the design is supposed to fulfil. Random unintentional configuration of matter is no design, serves no purpose.

If you claim that something is designed, you must establish the purpose for which the design has been designed.

If you cannot do that, but you know the designer, then you can ask him/her.

  • 5
    Unfortunately, evolution tends to fulfill purpose, too. We have a mouth, with the purpose to eat. We have eyes with the purpose to see, and ears with the purpose to hear. Even for things that are not alive, we are likely to encounter something that looks like it has a purpose, because things naturally stabilize towards an equilibrium. For instance, the bed of a river appears to have a purpose: to let the river flow from its source down to the sea. A volcano appears to have a purpose: to take pressure off of an internal chamber. Etc.
    – Stef
    Jun 5 at 19:49
  • So, when we encounter an object in space: if we cannot see its purpose, maybe it has a purpose but we just don't understand the purpose. If we can see its purpose, then maybe it naturally evolved or stabilised into something that fulfilled that purpose, without being designed by aliens.
    – Stef
    Jun 5 at 19:50
  • As for "If you know the designer, then you can ask", that won't help deciding whether an object found in space was designed or not, if we haven't encounter aliens yet.
    – Stef
    Jun 5 at 19:51
  • I realise my comments above are a bit too all-or-nothing. I agree that finding or not seeing the purpose of an object will be evidence in order to determine whether it was designed by aliens or not. But this evidence won't be definitive.
    – Stef
    Jun 5 at 19:53
  • 4
    @Stef Evolution fulfils no purposes, there is no designer. Evolution is a series of random mutations, some of which are beneficial adaptations to the environment improving the odds of survival and reproduction. We can use evolved natural things for some purposes, but we do not design them, the design itself does not serve any purpose. Selective breeding of living things is kind of a grey zone, there is some design for a purpose involved. Jun 6 at 4:02

William Dembski has made the claim, “evolutionary algorithms cannot generate specified complexity.” Therefore, if the designer was the product of either biological evolution or an evolutionary algorithm, that would refute his assertions about a “design inference.” (Which he has largely abandoned anyway, in response to counterexamples he could not refute.)

So, if we met a Martian who told us, “Oh, yes, the natives of the planet built it. They evolved, you know; we watched for millions of years without interfering, until they achieved sapience,” that would falsify the claims of most ID proponents, such as Dembski and Michael Behe. So would finding a supercomputer that did, and said, “I did it. I am an AI created by a reinforcement-learning algorithm.” Of course, so would any AI being able to create even a single bit of what they call “complex” information in response to a prompt.

However, both Dembski and Behe have a history of moving the goalposts in response to irrefutable examples of evolution and evolutionary algorithms doing things they claimed were impossible. It’s within the realm of possibility that they’d attempt to claim that nothing on Mars is an example of Complex Specified Information, or more likely, that it must really have been made some other way. At various times, they’ve asserted that there is some kind of “actual complex specified information” that could not have been created by an evolutionary algorithm (directly or indirectly), and some other category of “apparent complex specified information” that can. That is unfalsifiable; Dembski has given no infallible Turing Test to tell them apart and infer what he calls “design.”

  • Wikipedia on the term Specified Complexity: en.wikipedia.org/wiki/Specified_complexity
    – Contango
    Jun 6 at 10:18
  • Even with a "Turing test" that could decide a particular question, you might just have to wait forever for the answer. Maybe we could move on to other concerns in the meantime.
    – Scott Rowe
    Jun 6 at 10:26
  • @Contango Is there a particular statement in there that you would like to call my attention to?
    – Davislor
    Jun 6 at 13:00

Suppose you found a working device on Mars that appeared to be a solar powered calculator of some sort, with a keypad and display and so on. I think you could be fairly confident that it was designed. If you made that claim, there would be no way to falsify it in a black-and-white sense. Even if it were possible to show that working solar powered calculators could come into being through a natural process, that would not prove that the individual calculator in question had come to being that way. Unless, of course, you also found CCTV footage on Mars showing the calculator gradually taking shape under the influence of a natural process.

What I have described above is a (very) extreme case in which any sane person would conclude with the minimum of reflection that the object was designed. In principle, however, you could create a spectrum of scenarios in which the technical sophistication of the device was gradually reduced, until at the other end of the spectrum the object might be just a rock of a certain shape, about which you could conjecture that the rock was fashioned into that shape by some intelligent being with a particular purpose in mind, but equally it could have acquired its shape naturally. Between the two extremes of the spectrum would be any number of cases in which the possibility of a deign seems more or less likely. It will always be a judgement, rather than a matter of absolute proof.


Some of these answers are clever, but are overcomplicated. Yes, one can empirically verify a lack of design in things-that-look-designed in principle, but the devil is in the details of which your question is light.

If Elon Musk stumbling around in a cave comes across something that seems to straddle natural process and artificial processes in its origin, then one in theory could attempt to observe natural processes free of agency occur again. Thus, an experiment designed to observe natural processes which documented and explained said natural processes arriving at the artifact-apparent would be strong disconfirmation that an agent has designed the object. The philosophical sticky wicket is in demarcating "natural processes free of agency" since agents often cultivate the conditions of natural processes to benefit. There are in fact natural processes that generate what appear to be artifacts (YT). Ultimately, by validating that these artifacts-apparent are in fact free of agency, it then becomes part of the scientific canon of explanation.

If one comes across an artifact-apparent on Mars, but suspects natural processes free of agency are responsible, one simply goes about collecting observations of the development to ensure agency is not involved. Obviously, the parallel with evolution has certain practical constraints, but speculatively, a long-lived intelligence could bear witness to abiogenesis, evolutionary pressures, and the emergence of human-level intelligence. If in the future, humanity went out among the stars armed with the secrets of sustaining biological organisms indefinitely, it might be able to add to the annals direct scientific evidence. This is a trope in science fiction, in fact.

  • So, in short, you propose experiments to test the chance hypothesis. If those experiment succeed, then the design hypothesis can be disregarded as unnecessary. If those experiments fail, the design hypothesis is still viable. If someone adheres to the design hypothesis, they would predict that all attempts to verify the chance hypothesis will necessarily fail. It's a prediction about the failure of competing hypothesis. This resembles a lot the ID vs. evolution debate. ID proponents predict that all evolutionary explanations are faulty/bound to fail. Hence the need to attack evolution.
    – Mark
    Jun 7 at 17:20
  • @Mark I suggest that you exercise caution by using a caveat. Scientific explanation doesn't work based on the truth conditions of a simple logic claim. What makes a theory robust are the factors you list in your last question. In isolation, a design advocate has a fair criticism in saying that no one has witnessed abiogensis and the evolution to humanity. That's not only a fair claim, it's solid. But to conclude, evolution fails to explain Homo sapiens because no one has is specious. It's one avenue of attack and defense. Aliens evolving people is scientifically speculative, but defensible...
    – J D
    Jun 7 at 17:25
  • so if one argues goes beyond the political intent of ID and comes up with a metaphysically viable version (and make no mistake, it's metaphysically viable to claim that humans have been designed) and discovers evidence of extraterrestrial existence (as this former US intelligence offer alleges to Congress), then creationists and heterodox thinkers will be put out, and Trekkies the world over will rejoice. In fact the other big A-field is artificial life.
    – J D
    Jun 7 at 17:30
  • The struggle in ID is between a theistic flavor of supernaturalists (God pushes button on earth) and naturalists (God might have made the buttons, but that's a matter of faith). Atheism isn't the only metaphysical viable interpretation of materialism, much to the chagrin of New Athiests.
    – J D
    Jun 7 at 17:31
  • So, the sustained attack on natural selection by creationists (in the guise of ID) is an attempt to get a politically viable version of theological interventionism into public schools, and to appeal to the natural intuition to believe in divine agency (who doesn't believe in growing up that there are supernatural forces at work?) to attack the pact between scientists and theologians to observe the non-overlapping domain of magesteria (Gould) which has been stable. Curiously, in Islam, ID is very popular even in secular societies....
    – J D
    Jun 7 at 17:55

You must log in to answer this question.

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