What this covers

Rachell Powell joins Sean Carroll to examine how evolutionary contingency and convergent evolution reshape our understanding of life's diversity, the origins of morality, and the fragility of human culture. The conversation centers on a methodological question: how can we tell which biological traits are law-like, repeating across distant lineages, and which are quirky one-offs? Powell argues that the answer lies in treating convergence as natural experiment—when similar structures or behaviors evolve independently in unrelated organisms, they reveal something robust about how life works. This approach cuts against the grain of 20th-century evolutionary thinking, which tended to emphasize historical accident and contingency so heavily that figures like Ernst Mayr found little paradox in the Fermi paradox.

The conversation ranges across brains, social organization, human morality, and what we can infer about minds we have not yet encountered. Powell traces how convergent evolution appears in unexpected places: not just eyes and fins, but the independent emergence of sophisticated cognition in arthropods, cephalopods, and vertebrates, and strikingly similar norm-enforcement systems in humans and social insects. She argues, however, that human cumulative culture—the ability to retain and refine innovations across generations—is highly contingent, emerging only in the last eye-blink of human history after anatomically modern humans existed for hundreds of thousands of years without it. This distinction matters profoundly for ethics: Powell shows how human morality, shaped by group selection during ancestral intergroup conflict, carries a tribalistic bias that makes our moral progress fragile and easily reversed. She also challenges the assumption that the Copernican principle applies neatly to biology, arguing that it breaks down precisely where contingency dominates and that no globally optimal traits exist—fitness always depends on local context.

Sharpest takeaway

Powell argues that whether evolution is contingent or convergent must be settled by carefully parsing which traits are replicable versus quirky, and that a convergentist, anti-anthropocentric lens reveals law-like patterns in minds and social organization while showing human cumulative culture to be a highly contingent, fragile, and non-inevitable outcome.

  • Physics-style Copernican/mediocrity reasoning breaks down in biology because there are few contentful universal laws and no globally optimal traits.
  • Convergent evolution functions as natural experimental replication, revealing law-like patterns in brains, perception, and even social norms across distant lineages.
  • Human cumulative culture arose only recently after ~99% of human history without it, marking it as contingent and our moral gains as fragile.

The argument · threads10 threads · 40 claims
0.78

Contingency, not law-like necessity, dominates the history of life because fitness depends on local environments and early developmental nodes are highly determinative.

6 pointscentrality 5/5
  • There are no globally optimal traits—no traits that are always fitter—because fitness depends entirely on local environments (intelligence is fitness-enhancing in some contexts, while simply grazing on an easy food source is fitness-enhancing in others), which is a fundamental reason the history of life is not predictable.

    “the big problem is there are no globally optimal traits, meaning traits that are always going to be fit, always going to be optimal.”

  • Stephen Jay Gould articulated contingency via the thought experiment of rewinding and replaying the tape of life from slightly different initial conditions; inspired by the Burgess Shale fauna, he argued that if things had gone slightly differently at the base of the Cambrian, the dominant animal phyla (vertebrates, arthropods, mollusks) would have been relegated to science-fiction possibilia and some other shape of life would have arisen.

    “imagine if we rewind the tape of life and let it play again.”

  • The fundamental obstacle to settling the contingency question is the N=1 problem: with only one history of life and no extraterrestrial data sets, we cannot actually rerun the tape, so distinguishing law-like from contingent outcomes remains largely speculative.

    “The N equals one problem is presumably the biggest problem that we're dealing with when it comes to the contingency question”

  • Cumulative culture was not enabled by a single genetic mutation but by the unpredictable coming-together of perhaps 30 or 40 orthogonal major adaptations—including prerequisites like freed-up hands for fine motor manipulation, which themselves arose for unrelated reasons such as the locomotor efficiency of bipedalism over knuckle-walking—which is the signature of contingency in macroevolution.

    “We could probably come up with like 30 or 40 orthogonal adaptation, major adaptations that are all relevant to creating cumulative culture.”

  • For about 99% of their existence as fully modern persons, humans had the ecological footprint of other social carnivores with tiny numbers; events like the Toba super-eruption ~70,000 years ago may have reduced the effective human population to around 2,000, with global population hovering around 20,000 for tens of thousands of years, showing there was nothing inevitable about the later population and technological explosion.

    “the Toba super eruption, whatever, 75, 70,000 years ago, you have an effective human population down to what, 2000. And then like you have a human population hovering at 20,000 global for like tens of thousands of years afterwards.”

  • Early nodes in a complex developmental cascade during embryogenesis are highly determinative of subsequent outcomes, and once the downstream machinery is laid down those early nodes cannot be changed without disrupting the whole system, which explains why new phyla do not repeatedly originate at large scales of organization.

    “once those early nodes are laid down, and then you lay down all the the the downstream machinery, you can't change the early nodes without disrupting the whole system.”

0.78

Convergent evolution across distant lineages suggests law-like constraints on how brains, cognition, and social norms repeatedly evolve independently.

6 pointscentrality 5/5
  • Convergent evolution—the independent replication of similar biological forms and functions—can be treated as tantamount to natural experimental replications, providing a workable data set in the absence of extraterrestrial information and indicating a law-like necessity inconsistent with strong contingency.

    “these replications, these convergences, these repetitions, can essentially be treated as tantamount to natural experimental replications in the history of life.”

  • There is strong evidence that brains, active bodies, and minds evolved multiple times independently early in animal evolution—first in arthropods (predatory 'seek and destroy' animals in the Cambrian before vertebrates even had heads), triggering arms races, and then independently in cephalopod mollusks and vertebrates—indicating something law-like about the repeated evolution of brains.

    “there's very, there's strong evidence that brains and active bodies and minds evolved multiple times independently early in animal evolution. First in the arthropods”

  • Bees provide astounding evidence of sophisticated cognition—including abstract concepts such as sameness and difference and the ability to transfer those concepts across sensory modalities (e.g., same/different shape to same/different smell)—and evidence for holistic, unified perceptual worldviews with objects bound to properties in space and time, much like our own umwelt, suggesting these features are law-like and were recreated independently in other lineages.

    “There's I think, compelling evidence for abstract concepts in bees, like sameness, difference and then actually transmitting that across sensory modalities”

  • Because we could only ask these questions from a world where terrestrial intelligence evolved, observer selection effects normally prevent projecting our own existence as evidence; but convergence is powerful precisely because we do not have to hail from a world where consciousness and complex cognition evolved multiple times independently, so its repeated independent evolution looks like a law-like phenomenon that gets around observer selection bias.

    “we don't have to hail from a world where consciousness, intelligence, complex problem solving, whatever you want to talk about, evolved multiple times. That starts to really look like you're getting around observer selection effects”

  • Convergence should not be treated as a monolithic, homogeneous phenomenon that simply refutes Gould; instead one must examine the underlying causes of each repetition, because some convergences genuinely cut against contingency while others do not, so the data set must be carefully parsed before drawing grand conclusions about a cosmic biology.

    “it's critical that you sort of look at what the underlying causes of the repetitions are.”

  • Asking why humans diverged from close relatives like chimps yields a 'divergentist' explanation that is powerful but narrow, applying only to a possibly highly contingent one-off case with no law-like lesson; asking instead why two distant lineages arrived at similar endpoints yields a 'convergentist' explanation revealing deeper, more law-like common causes, and both questions are legitimate but the convergentist one is needed for broad evolutionary patterns.

    “that's what I would call like a divergentist explanation. Why did humans diverge from these other close cases?... if you go broader and you think about convergentist explanations... what you're asking is why did these two lineages arrive at similar endpoints?”

0.53

Humans are unique in cumulative culture, not raw intelligence; this contingent ability explains our ecological dominance and relates to the Fermi paradox.

3 pointscentrality 4/5
  • Humans are not, on any reasonable operationalization, smarter than animals like dolphins; the tremendous difference is that humans benefit from cumulative culture rather than superior individual intelligence, and this cultural accumulation is the focal point of contingency and the closest thing to a biological resolution of the Fermi paradox.

    “I don't think humans are smarter than, say, dolphins. I don't. I don't believe that. I do believe that we benefit from cumulative culture and that is a tremendous, tremendous difference.”

  • Cumulative culture—the retention and incremental improvement of innovations across generations so a species need not reinvent the wheel—is the respect in which humans are genuinely unique, but it is highly contingent: anatomically modern humans with all the relevant capacities (language, morality, mind reading, imitation) lacked robust cumulative culture for possibly hundreds of thousands of years, having the ecological footprint of other social carnivores until a very recent eye-blink of history.

    “getting a species with cumulative culture. And that is a respect in which we are unique. But it's worse than that.”

  • Extraordinary levels of cognitive flexibility and intelligence are probably a liability in the long run; far stupider animals are likely to persist in the stratigraphy of the Earth long after highly intelligent humans are gone.

    “I think ultimately extraordinary levels of cognitive flexibility and intelligence is probably a liability in the long run.”

0.53

Social norms are functionally defined rules enforced by subordinates; they exist in social insects and humans but not hierarchical primates like chimpanzees.

3 pointscentrality 4/5
  • Ants share with humans, but differ from chimpanzees, in having an institutionalized form of norm enforcement—effectively police forces—where rules are enforced by subordinates rather than by sheer dominance, which is why ant and human societies are far more cooperative than chimpanzee societies.

    “they do have police forces and they do have what I would call an institutionalized form of norm enforcement.”

  • The best case for genuine social norms outside humans is found not in cetaceans or chimpanzees but in social insects, because a normative society requires rules enforced by subordinates against everyone (rule of all by all), not mere dominant coercion; chimps remain hierarchical with dominance turnover but lack robust subordinate enforcement, which explains why chimpanzee societies are not very cooperative.

    “the best case, I think, for social norms outside of human beings is actually not in the animals you might expect... it's actually in social insects where you get a robust institutionalization of rules of conduct that are enforced by subordinates against everyone in the colony.”

  • Social norms should be defined functionally as rules of conduct that regulate behavior in social groups to stabilize ultra-cooperation, rather than tied to uniquely human cognitive mechanisms; defined this way they are multiply cognitively realizable, so building human-specific cognition into the definition is a fait accompli that wrongly guarantees no other species can have them—analogous to defining vision so narrowly that cephalopods would absurdly count as blind.

    “Social normative structures are social structures that are multiply cognitively realizable, meaning many different cognitive forms could give rise to the same functional structure.”

0.53

Moral progress achieved through reflective normativity is fragile because humans evolved parochial in-group altruism through intergroup conflict.

3 pointscentrality 4/5
  • On the standard evolutionary picture, human in-group altruism evolved through intergroup conflict and a group-level selection process that favored moral groups only because that strategy was adaptive in competition with other groups, producing universal in-group favoritism and out-group antagonism—so the inclusive moralities we identify as progress are difficult to achieve and may be unsustainable given this parochial, tribalistic legacy.

    “in order to get the kind of altruism within the groups, you need these intergroup conflicts... you have essentially what amounts to like in-group favoritism and out-group antagonism, which is extremely universal in humans.”

  • For humans to exercise reflective morality, social conditions must avoid replicating the ancestral cues that trigger out-group antagonism, requiring surpluses, education, and security; because perceived (not necessarily real) scarcity, intergroup competition, or predation can be exploited polemically to trigger xenophobic attitudes, moral gains are fragile and can regress rapidly.

    “it's very easy to turn back the tide and regress rapidly when there's actual or perceptions of scarcity, intergroup competition, predation of one group by another.”

  • Despite a parochial evolutionary legacy, humans have an open-ended capacity for normativity that lets them, under the right circumstances, step back and critique the norms they follow, make consistency judgments, and thereby interrogate and improve their moral systems—which explains post-Enlightenment moral progress such as the rule of law, abolition of slavery, and inclusion of marginalized groups.

    “humans are able to step back, critique the kind of norms that they're following and make consistency judgments and other things that allow them to interrogate and improve our moral systems.”

0.44

The natural world is horrific and competitive; human moral reflection occurs within fundamentally constrained biological and technological bounds.

8 pointscentrality 2/5
  • The natural living world is a horrific place rather than a beautiful harmonious state that human action merely disrupts; the colonialist-style narrative of bad actors mucking up a natural harmony is false, as illustrated by the fact that maintaining a stable fish population requires that the vast majority of offspring die in often horrific ways.

    “there wasn't some kind of harmony that was being disrupted. And is of course the same thing in the natural living world. It is a horrific, horrific place.”

  • Once ethics moves to aggregates—weighing the value of rational persons against sentient non-persons and summing across them—no one knows what to do, because parsing value in nature rests on insanely controversial and unresolved questions, so whether human extinction would be a boon or loss for sentient life overall is genuinely indeterminate.

    “Once you get to aggregates and ethics, no one knows what they're no one knows what to do.”

  • From a macroevolutionary perspective, the one robust inductive inference is that every species' time is limited; there may be no ethical imperative for existence at any cost or in any form, and human extinction is not obviously to be avoided regardless of consequences.

    “if there's any inference you're going to make from induction, it's that all of every species' time is limited.”

  • Comparative cognition was long hampered by an excessive focus on homology—looking for the specific human brain structure associated with a capacity in other lineages and concluding its absence meant the capacity was absent—when in fact lineages can independently evolve different structures performing the same function, as eventually recognized for birds and other animals.

    “historically, there was this like focus in comparative cognition, a very heavy focus on homology.”

  • Using Groundhog Day as a thought experiment, the protagonist progresses from egoism through futile utilitarian maximization to virtue ethics, where he stops trying to maximize good and instead develops his talents and flourishes—suggesting that since we cannot actually aggregate and maximize value, the deepest stance is to understand and empathize with the world while giving oneself space to flourish, which looks like virtue ethics even if it is closer to aesthetics than to ethics.

    “I think virtue ethics is stupid, but I think it's right.”

  • Just as social insects and humans are the only lineages to have repeatedly domesticated fungi and animals, we should ask whether artificial intelligence and computational technologies are our domesticates or whether—through a coevolutionary process in which the domesticator/domesticate distinction is hard to draw—there may come a point where humans become the domesticate without even realizing it.

    “Should we think about them as our domesticate? Or is there going to be a point where we're not even going to realize it and we're going to become the domesticate?”

  • The most profound worldview-shattering insights in modern human history have come from science rather than from philosophy or religion, operating along two arcs: an intuition-shattering demolition of common assumptions about the causal structure of the world (relativity, microphysics, mechanistic adaptation) and a decentering project that displaces humanity from a privileged position to an unremarkable periphery.

    “the most profound worldview shattering insights in modern human history haven't come from philosophy. They haven't come from religion or anything. They've come from science.”

  • The burial of the dead appears across vastly different lineages—Neanderthals (often interpreted via symbolism and supernatural belief) and social insects (which lack such representations)—suggesting a common underlying evolutionary function tied to epidemiology, with the human supernatural justifications being merely the proximate, confabulatory mechanism evolution uses to achieve that function in a flexible species.

    “you see all sorts of social insects burying their dead. And you're like, whoa, hold on a second, what's going on? Obviously, they're not... They're don't they don't have rituals around it.”

0.43

Convergence must be examined trait-by-trait; some convergences refute contingency while others do not, requiring careful parsing before grand conclusions.

4 pointscentrality 3/5
  • Even if we fully understood all the causal levers of human moral psychology, development, and evolution, we might still be feckless against moral regression because of the way emergent patterns work—knowing a demagogue's playbook does not guarantee we can successfully battle it, especially amid new arms-race dynamics created by social media.

    “even if we knew all of the causal levers that go into human moral psychology, development and evolution... we might still be feckless to do anything about it because of just the way these emergent patterns”

  • The most prominent evolutionists of the 20th century synthesis—Ernst Mayr, Dobzhansky, and G.G. Simpson—were very negative about the SETI project because, seeing the big-picture history of life as a sequence of quirky, contingent outcomes, they found little paradox in Fermi's question and were unmoved by Copernican large-numbers arguments.

    “the most prominent evolutionist of the 20th century, like that were involved in the major evolutionary synthesis, it's called. Like Ernest Meyer, Dobzhansky, G.G. Simpson, they were all very negative about the SETI project.”

  • The common-culture image of evolution as a single linear trajectory of progress—from quadruped to knuckle-walker to upright human—is false, and assumptions of progress have infiltrated thinking about biology far more profoundly than thinking about physics, including 20th-century thinking about SETI.

    “you see the what the quadrupedal creature, then the knuckle walking creature, then the creature walking upright, and it looks like a single linear trajectory of progress. We know that all of that is false.”

  • Thought experiments have played important roles in physics but, unlike in physics, have rarely featured in biology—which is itself interesting—yet Gould's evolutionary counterfactual thought experiments were profoundly important in thinking about contingency and law-like necessity in the history of life.

    “In physics, thought experiments have played interesting roles. In biology, no, not that I know of, right? Which is interesting in itself.”

0.35

Biology lacks universal laws with specific predictive content; natural selection is generic and schema-like, unlike physics with its contentful laws.

2 pointscentrality 3/5
  • Unlike physics, which has contentful laws enabling specific predictions, biology has arguably no or very few universal laws; even the principle of natural selection—the closest thing to a law—is so generic and schema-like that it skirts being an a priori truth, contains no specific content, and cannot predict which traits will evolve.

    “In biology, there are arguably no or very few universal laws.”

  • Copernican/mediocrity reasoning holds up well in physics because physical laws are universal (the same laws govern terrestrial and celestial phenomena), but when those modes of inference are extended to biology they break down quickly, with contingency being one of the principal reasons.

    “when you extend those principles, those ways of thinking, those modes of inference that are very familiar in the physical sciences to biology, things start to break down pretty quickly.”

0.21

How social functions emerge developmentally—genetically or culturally—does not determine whether they count as the same trait across lineages.

2 pointscentrality 2/5
  • How a social function emerges in development—whether via relatively fixed genetic mutations (as in social insects) or flexible learning and culture (as in humans)—does not matter for whether the function counts as the same trait, because genes, environment, and culture are all just resources that bias developmental outcomes, with nothing metaphysically special about genes.

    “Genetics and environment and culture are just all resources that bend, shape, bias certain outcomes in ontogeny. There's nothing special about... I mean, well, there's debates about this, but I'm not sure that there's anything special about genes in that regard.”

  • Scientific explanations are question-relative: a complete account of a behavior like Neanderthal burial requires proximate mechanisms (symbolism, beliefs) when explaining that specific behavior, but understanding what the trait fundamentally is and why it arises in evolution requires a broader convergentist perspective—and much cross-talk between disciplines stems from conflating these different questions.

    “it's all question relative. The explanations are question relative. And so that's why you end up getting so much cross talk”

0.18

Evolved traits are separate and replicable, not inseparable bundles; science fiction wrongly treats distinct features as necessarily co-evolved packages.

1 pointcentrality 2/5
  • People commit a 'bundling fallacy' by treating evolved traits as inseparable bundles and projecting the whole bundle (e.g., a cephalopod body plan plus human-like intelligence) onto alien life, rather than parsing which traits are contingent and which are replicable; this conflation underlies science fiction's reliance on humanoid or near-humanoid aliens.

    “this is something that I call like the bundling fallacy, because I couldn't think of a better word for it, but we tend to... We treat evolution as these like bundles of traits.”