In the early 1500s, Nicolaus Copernicus discovered that the Earth isn’t the center of the universe. The planets, he found, revolve around the sun. It was such a stunning revelation at the time, and so contrary to the teachings of the all-powerful Catholic Church, that Copernicus kept it a secret until the end of his life. It wasn’t until he was on his deathbed that he published his masterwork, “On the Revolution of the Heavenly Spheres.”
Humans have a natural tendency to think that we occupy a special position in the universe, but modern science has repeatedly shown us that the opposite is true. We Earthlings are more likely to be mediocre than exceptional. This idea has a name: The Copernican Principle of Mediocrity. It’s now a foundational principle for cosmology, astronomy, and the search for alien life.
Recently, Eric Schwitzgebel, professor of philosophy at University of California, Riverside, and Jeremy Pober, a postdoctoral researcher at the University of Lisbon, applied this principle to the question of what kinds of beings might possess consciousness. The Copernican Principle of Mediocrity suggests humans and other flesh-and-blood Earth-bound creatures cannot be the only ones. And if aliens made of different stuff potentially possess consciousness, then why not minds made of silicon, metal, and electricity, also known as artificial intelligence? The pair of researchers published their answer to this question and other thoughts in a new working paper.
I spoke with Schwitzgebel about what makes consciousness so bizarre, why behavioral complexity is a good measuring stick for it, sci-fi author Vernor Vinge’s Zones of Thought series, and why he is doubtful that the AIs of today possess consciousness.
Read more: “Why Haven’t We Heard from Extraterrestrials Yet?”
What prompted you to explore this particular question of whether consciousness requires flesh and blood right now?
One of the primary motivations is the question of AI consciousness. That’s a hot topic these days, for obvious reasons. People rightly have a pretty liberal set of inclinations and intuitions about the consciousness of behaviorally sophisticated space aliens if they were to exist or if we were to come across them. The fact that we’re more hesitant in AI cases is interesting; so I’m motivated partly by exploring that contrast and thinking about the extent to which that contrast is justified. To some extent, I think it is justified.
Do any scientists who study consciousness today explicitly argue that only life forms found on Earth can have consciousness?
It’s a pretty unusual view. I don’t see any mainstream scientists who explicitly say that. Part of what Jeremy and I are doing is trying to articulate, justify, and formalize something that most people already agree on and think is true.
You don’t attempt to define consciousness in your paper. Does this limit what you’re able to say about where it may be found?
The definition of life is also highly contentious. There’s no consensus on it, but we know it when we see it, with some disputable cases. We don’t need to have fully formalized definitions of terms to do science with them. I do define consciousness elsewhere, in another paper, but it’s tricky and contentious. In this project, we’re hoping that the reader and our fellow scientists and philosophers can understand the concept with a brief presentation of what it is, rather than requiring a fully worked-out definition.
You argue that consciousness isn’t limited to creatures made of flesh and blood like we find on Earth, or even necessarily possessing brains. The limitation you place on it is that it requires behavioral complexity. Why did you choose that particular limitation?
Limitation isn’t quite the right word. It’s more of a sufficient condition than a necessary condition. Limitation suggests that you have to be behaviorally sophisticated in order to be conscious, and what we’re suggesting is that behavioral sophistication is probably sufficient for attribution of consciousness. There may be some things that aren’t behaviorally sophisticated that are conscious.
Jellyfish have neurons. They’re not behaviorally sophisticated. Maybe just having neurons is enough. Some scientists might think that garden snails behave in sophisticated ways. Are they conscious? We don’t want to get into that in this paper.
We want to assume a pretty high bar. Most people, pretty much regardless of their background, are going to think, “Yes, if you’ve got all of those things that we describe as behavioral sophistication, then it’s plausible that you’re conscious.” And what we want to do is articulate and defend that point of view.
When you describe a theoretical alien species that is behaviorally sophisticated in your paper, it ended up sounding pretty human-like. Is behavioral complexity in its essence a humanlike framework?
I don’t think it has to be. Behavioral sophistication is three things: complex communication with something like generative grammar, complex cooperation, and complex planning. Any species that’s going to be doing complicated things over the long term has got to have complex planning. So that’s just general cognitive achievement. And then we want to include the species being cooperative, and complex planning and cooperation are going to be facilitated by some sort of complex communication. We think of ourselves as the primary examples of this, but I don’t think that it’s just a covert way of saying, “Oh, you’re human-like.” It’s an overt way of saying you have a certain suite of advanced sophisticated capacities.
We—especially perhaps intellectuals—think of ourselves along these lines and think of humanity as essentially possessing these kinds of things. But of course, there’s much in humanity like having two arms and two legs and prehensile hands and upright gait and all that kind of stuff that has little to do with behavioral complexity as we define it. The fact that you think of it as human-like maybe tells you a little something about how you think about humans, or what we think makes us distinctive.
Is it possible that in using behavioral complexity as your standard, you could miss a lot of potential consciousness out in the rest of the universe?
For sure. We’re not aiming to capture every kind of thing that’s conscious. We’re aiming to provide an argument that things made out of different substances could be conscious. So all we need is, “Wouldn’t you agree this thing is conscious?”—regardless of what your view is about consciousness on Earth. Likely, some of these things are made of different substances. That’s enough.
Read more: “The Odds That Aliens Exist Just Got Worse”
You estimate that there are at least 1,000 behaviorally sophisticated extraterrestrial species that have existed in the cosmos. Can you walk me through how you arrived at this number?
There may be many orders of magnitude more, but 1,000 is all we need. That number is basically just a reflection of the ordinary opinions of astrobiologists. Most of the astrobiologists I’ve spoken to personally—though they don’t call it behavioral sophistication, they might call it technological life—believe that something like that isn’t so rare in the universe.
There is something on the order of 1 trillion galaxies in the observable portion of the universe, and the universe might be larger than the observable portion. A kind of middle estimate that astrobiologists give is that maybe once in a galaxy’s lifetime, a technological civilization will arise. That’s rare enough to explain what’s called the Fermi Paradox [the total lack of physical evidence for extraterrestrial life elsewhere in the universe, despite the high statistical probability that it exists]. But this estimate of once in a galaxy’s lifetime makes it common enough that there would be approximately a trillion behaviorally sophisticated civilizations in our universe.
We’re saying even if 1 trillion is optimistic, which is in the ballpark of a median estimate, let’s knock several orders of magnitude off of that. Let’s say at least 1,000. There are going to be some astrobiologists who say maybe we’re the only one in the entire observable portion of the universe. Some people call this the Rare Earth hypothesis. It’s a respectable opinion, but it’s a minority opinion, so we aren’t making some novel argument. We’re just generating a conservative estimate based on the community’s opinion.
Do we know that evolution takes place outside of Earth—and is that relevant to this question of whether behaviorally complex life might exist elsewhere in the universe?
In order for evolution to occur, you don’t need much, so it does seem extremely plausible that if we have complex systems, there will be evolution. You need some sort of stable, heritable trait, something that could be passed down from generation to generation through replication. You need some capacity to replicate, and you need some tendency for the environment to select some traits over others in a way that maintains the heritability of that trait over time. With some pretty basic assumptions, you’ll have something like natural selection.
Do you have any favorite science-fiction imaginings of what an alien consciousness might look like?
I’m an enthusiast for science fiction, but one of the things that we see in science fiction is a tendency to imagine entities whose patterns of thinking are a lot like ours. It makes it much easier to relate to them. It makes for more readable novels. But then you see some science-fiction authors who go far in the other direction and try to create incomprehensible aliens. Peter Watts’ Blindsight and Stanislaw Lem’s Solaris are examples of this. The aliens are so different from us that they’re basically completely unintelligible to us. What would be fascinating to see, and what I don’t see as much as I’d like in science fiction, are aliens that are different from us and yet still kind of on the edge of comprehensibility.
One example that I like in that direction is Vernor Vinge’s series of novels called Zones of Thought, which feature what he calls “tines,” these dog-like organisms that communicate with each other in packs. Each individual dog isn’t behaviorally sophisticated in the sense that Jeremy and I propose, but when they get together in a pack, a larger consciousness emerges. Another favorite example is Greg Egan’s Diaspora, a novel that explores a future in which consciousness is possible in AI systems, and which imagines a variety of ways in which people—because they’re not really humans anymore—might choose to live their lives, with lots of flexibility about how exactly these post-human minds are implemented.
You suggest in your paper that future AI systems might possess consciousness. What developments do you think could contribute to AI possessing consciousness?
Here my co-author and I diverge a bit. We’re both somewhat skeptical that future AI could possess consciousness, but he is more skeptical than I am. The aim is to shift the burden of proof, so to speak. Now I’m speaking for myself and not necessarily for Jeremy, but if we can establish that consciousness is substrate flexible, then someone who wants to argue that behaviorally sophisticated AI wouldn’t be conscious, would need some grounds to say that it’s different from aliens. If we allow that behavioral sophistication is enough for a default attribution of consciousness for aliens, why not for AI systems? You’d at least need an argument. That’s the kind of rhetorical argumentative shift that I’m hoping for at the end of the paper. This isn’t to say that you couldn’t come up with such an argument, but a burden of proof would need to be met.
What would behaviorally complex AI systems look like?
They would have to have complex long-term planning, cooperation, and communication.
On some level, are they already there?
That could be. This is an issue we discuss in more detail in a companion paper that we have forthcoming, called “The Copernican Argument for Alien Consciousness, the Mimicry Argument Against AI Consciousness.” What we argue there is that maybe we shouldn’t be too fussy about whether current AI systems meet the relevant criteria of behavioral sophistication. Maybe they do, at least large language models. If they do, however, there’s a reason to deny the move to infer that they’re conscious. The Copernican principle of consciousness for behaviorally sophisticated systems shifts the burden of proof. You now have to have reasons to doubt that AI is conscious. We think there are reasons to doubt that current large language models are conscious, even if we allow that they’re behaviorally sophisticated.
What are those reasons?
It’s that they’re what we call consciousness mimics. They’re designed to mimic the superficial appearance of behaviors that are associated with consciousness in humans. And just like we mistrust that mimics have the underlying features that they appear to have, likewise, it’s reasonable to mistrust that AI systems have the underlying consciousness that you might infer. The overall structure of the argument is that behavioral sophistication is sufficient for a default attribution of consciousness, whether it’s to aliens or to AI, but that default can be canceled. And in the case of LLMs, as they currently exist, that default is canceled because they’re consciousness mimics.
Read more: “If You Meet ET in Space, Kill Him”
With this paper, you’re proposing that there has to be a kind of material package out of which consciousness arises. The substrate comes first.
Yes,we’re assuming the truth of materialism, which is certainly not the consensus view, but it’s the mainstream view right now in philosophy and the natural sciences: That the universe is fundamentally material and consciousness somehow arises out of the material stuff.
Some theorists, such as Donald Hoffman, propose that consciousness is, in fact, more fundamental than matter. What do you think about this possibility?
The idea that consciousness is more fundamental than matter—that matter arises out of mind rather than the opposite—is called idealism in philosophy, so Don Hoffman is an idealist. I think that’s on the table as one of the bizarre options that might be true. I’m not personally convinced by Hoffman’s arguments that it’s true. It’s not my highest credence option, but I wouldn’t discard it just on the basis that it’s bizarre by the standards of common sense, because all of the options are bizarre once you really start thinking about them.
I do think that something pretty wild must be true about consciousness and cosmology. This was the thesis of my 2024 book, The Weirdness of the World—that basically all of the common-sense options are off the table. Something bizarre, something strikingly contrary to common sense, must be true, but I don’t think we know where the truth lies among various bizarre options.
What are some of the other bizarre options, aside from idealism, that are worth considering?
Another very historically important one is substance dualism: The idea that there’s an immaterial soul that’s not reducible to material stuff, and then there’s material stuff. In other words, there are two basic things, not just mind and not just matter, but both are independently fundamental.
Another option that was historically important for a while, and has been relatively neglected recently, is what’s sometimes called transcendental idealism—the view that the fundamental nature of things as they’re independent of us is unknowable, and the fact that things appear to us as laid out in space is a feature of our own minds. The way our own minds construe some kind of unknowable fundamental reality is spatial. That’s another option that’s not quite the same as either materialism, idealism, or substance dualism. It’s a fourth option that’s intriguing and deserves serious attention.
What do you hope is true?
I hope that the universe has lots of fascinatingly interesting alien species. It would be richer and more wondrous if that were the case. 
Enjoying Nautilus? Subscribe to our free newsletter.
Lead image: crimson / Adobe Stock
