On two blog posts of Jerry Coyne

A few months ago, I got to know Jerry Coyne, the recently-retired biologist at the University of Chicago who writes the blog “Why Evolution Is True.” The interaction started when Jerry put up a bemused post about my thoughts on predictability and free will, and I pointed out that if he wanted to engage me on those topics, there was more to go on than an 8-minute YouTube video. I told Coyne that it would be a shame to get off on the wrong foot with him, since perusal of his blog made it obvious that whatever he and I disputed, it was dwarfed by our areas of agreement. He and I exchanged more emails and had lunch in Chicago.

By way of explaining how he hadn’t read “The Ghost in the Quantum Turing Machine,” Coyne emphasized the difference in my and his turnaround times: while these days I update my blog only a couple times per month, Coyne often updates multiple times per day. Indeed the sheer volume of material he posts, on subjects from biology to culture wars to Chicago hot dogs, would take months to absorb.

Today, though, I want to comment on just two posts of Jerry’s.

The first post, from back in May, concerns David Gelernter, the computer science professor at Yale who was infamously injured in a 1993 attack by the Unabomber, and who’s now mainly known as a right-wing commentator. I don’t know Gelernter, though I did once attend a small interdisciplinary workshop in the south of France that Gelernter also attended, wherein I gave a talk about quantum computing and computational complexity in which Gelernter showed no interest. Anyway, Gelernter, in an essay in May for the Claremont Review of Books, argued that recent work has definitively disproved Darwinism as a mechanism for generating new species, and until something better comes along, Intelligent Design is the best available alternative.

Curiously, I think that Gelernter’s argument falls flat not for detailed reasons of biology, but mostly just because it indulges in bad math and computer science—in fact, in precisely the sorts of arguments that I was trying to answer in my segment on Morgan Freeman’s Through the Wormhole (see also Section 3.2 of Why Philosophers Should Care About Computational Complexity). Gelernter says that

  1. a random change to an amino acid sequence will pretty much always make it worse,
  2. the probability of finding a useful new such sequence by picking one at random is at most ~1 in 1077, and
  3. there have only been maybe ~1040 organisms in earth’s history.

Since 1077 >> 1040, Darwinism is thereby refuted—not in principle, but as an explanation for life on earth. QED.

The most glaring hole in the above argument, it seems to me, is that it simply ignores intermediate possible numbers of mutations. How hard would it be to change, not 1 or 100, but 5 amino acids in a given protein to get a usefully different one—as might happen, for example, with local optimization methods like simulated annealing run at nonzero temperature? And how many chances were there for that kind of mutation in the earth’s history?

Gelernter can’t personally see how a path could cut through the exponentially large solution space in a polynomial amount of time, so he asserts that it’s impossible. Many of the would-be P≠NP provers who email me every week do the same. But this particular kind of “argument from incredulity” has an abysmal track record: it would’ve applied equally well, for example, to problems like maximum matching that turned out to have efficient algorithms. This is why, in CS, we demand better evidence of hardness—like completeness results or black-box lower bounds—neither of which seem however to apply to the case at hand. Surely Gelernter understands all this, but had he not, he could’ve learned it from my lecture at the workshop in France!

Alas, online debate, as it’s wont to do, focused less on Gelernter’s actual arguments and the problems with them, than on the tiresome questions of “standing” and “status.” In particular: does Gelernter’s authority, as a noted computer science professor, somehow lend new weight to Intelligent Design? Or conversely: does the very fact that a computer scientist endorsed ID prove that computer science itself isn’t a real science at all, and that its practitioners should never be taken seriously in any statements about the real world?

It’s hard to say which of these two questions makes me want to bury my face deeper into my hands. Serge Lang, the famous mathematician and textbook author, spent much of his later life fervently denying the connection between HIV and AIDS. Lynn Margulis, the discoverer of the origin of mitochondria (and Carl Sagan’s first wife), died a 9/11 truther. What broader lesson should we draw from any of this? And anyway, what percentage of computer scientists actually do doubt evolution, and how does it compare to the percentage in other academic fields and other professions? Isn’t the question of how divorced we computer scientists are from the real world an … ahem … empirical matter, one hard to answer on the basis of armchair certainties and anecdotes?

Speaking of empiricism, if you check Gelernter’s publication list on DBLP and his Google Scholar page, you’ll find that he did influential work in programming languages, parallel computing, and other areas from 1981 through 1997, and then in the past 22 years published a grand total of … two papers in computer science. One with four coauthors, the other a review/perspective piece about his earlier work. So it seems fair to say that, some time after receiving tenure in a CS department, Gelernter pivoted (to put it mildly) away from CS and toward conservative punditry. His recent offerings, in case you’re curious, include the book America-Lite: How Imperial Academia Dismantled Our Culture (and Ushered In the Obamacrats).

Some will claim that this case underscores what’s wrong with the tenure system itself, while others will reply that it’s precisely what tenure was designed for, even if in this instance you happen to disagree with what Gelernter uses his tenured freedom to say. The point I wanted to make is different, though. It’s that the question “what kind of a field is computer science, anyway, that a guy can do high-level CS research on Monday, and then on Tuesday reject Darwinism and unironically use the word ‘Obamacrat’?”—well, even if I accepted the immense weight this question places on one atypical example (which I don’t), and even if I dismissed the power of compartmentalization (which I again don’t), the question still wouldn’t arise in Gelernter’s case, since getting from “Monday” to “Tuesday” seems to have taken him 15+ years.

Anyway, the second post of Coyne’s that I wanted to talk about is from just yesterday, and is about Jeffrey Epstein—the financier, science philanthropist, and confessed sex offender, whose appalling crimes you’ll have read all about this week if you weren’t on a long sea voyage without Internet or something.

For the benefit of my many fair-minded friends on Twitter, I should clarify that I’ve never met Jeffrey Epstein, let alone accepted any private flights to his sex island or whatever. I doubt he has any clue who I am either—even if he did once claim to be “intrigued” by quantum information.

I do know a few of the scientists who Epstein once hung out with, including Seth Lloyd and Steven Pinker. Pinker, in particular, is now facing vociferous attacks on Twitter, similar in magnitude perhaps to what I faced in the comment-171 affair, for having been photographed next to Epstein at a 2014 luncheon that was hosted by Lawrence Krauss (a physicist who later faced sexual harassment allegations of his own). By the evidentiary standards of social media, this photo suffices to convict Pinker as basically a child molester himself, and is also a devastating refutation of any data that Pinker might have adduced in his books about the Enlightenment’s contributions to human flourishing.

From my standpoint, what’s surprising is not that Pinker is up against this, but that it took this long to happen, given that Pinker’s pro-Enlightenment, anti-blank-slate views have had the effect of painting a giant red target on his back. Despite the near-inevitability, though, you can’t blame Pinker for wanting to defend himself, as I did when it was my turn for the struggle session.

Thus, in response to an emailed inquiry by Jerry Coyne, Pinker shared some detailed reflections about Epstein; Pinker then gave Coyne permission to post those reflections on his blog (though they were originally meant for Coyne only). Like everything Pinker writes, they’re worth reading in full. Here’s the opening paragraph:

The annoying irony is that I could never stand the guy [Epstein], never took research funding from him, and always tried to keep my distance. Friends and colleagues described him to me as a quantitative genius and a scientific sophisticate, and they invited me to salons and coffee klatches at which he held court. But I found him to be a kibitzer and a dilettante — he would abruptly change the subject ADD style, dismiss an observation with an adolescent wisecrack, and privilege his own intuitions over systematic data.

Pinker goes on to discuss his record of celebrating, and extensively documenting, the forces of modernity that led to dramatic reductions in violence against women and that have the power to continue doing so. On Twitter, Pinker had already written: “Needless to say I condemn Epstein’s crimes in the strongest terms.”

I probably should’ve predicted that Pinker would then be attacked again—this time, for having prefaced his condemnation with the phrase “needless to say.” The argument, as best I can follow, runs like this: given all the isms of which woke Twitter has already convicted Pinker—scientism, neoliberalism, biological determinism, etc.—how could Pinker’s being against Epstein’s crimes (which we recently learned probably include the rape, and not only statutorily, of a 15-year-old) possibly be assumed as a given?

For the record, just as Epstein’s friends and enablers weren’t confined to one party or ideology, so the public condemnation of Epstein strikes me as a matter that is (or should be) beyond ideology, with all reasonable dispute now confined to the space between “very bad” and “extremely bad,” between “lock away for years” and “lock away for life.”

While I didn’t need Pinker to tell me that, one reason I personally appreciated his comments is that they helped to answer a question that had bugged me, and that none of the mountains of other condemnations of Epstein had given me a clear sense about. Namely: supposing, hypothetically, that I’d met Epstein around 2002 or so—without, of course, knowing about his crimes—would I have been as taken with him as many other academics seem to have been? (Would you have been? How sure are you?)

Over the last decade, I’ve had the opportunity to meet some titans and semi-titans of finance and business, to discuss quantum computing and other nerdy topics. For a few (by no means all) of these titans, my overriding impression was precisely their unwillingness to concentrate on any one point for more than about 20 seconds—as though they wanted the crust of a deep intellectual exchange without the meat filling. My experience with them fit Pinker’s description of Epstein to a T (though I hasten to add that, as far as I know, none of these others ran teenage sex rings).

Anyway, given all the anger at Pinker for having intersected with Epstein, it’s ironic that I could easily imagine Pinker’s comments rattling Epstein the most of anyone’s, if Epstein hears of them from his prison cell. It’s like: Epstein must have developed a skin like a rhinoceros’s by this point about being called a child abuser, a creep, and a thousand similar (and similarly deserved) epithets. But “a kibitzer and a dilettante” who merely lured famous intellectuals into his living room, with wads of cash not entirely unlike the ones used to lure teenage girls to his massage table? Ouch!

OK, but what about Alan Dershowitz—the man who apparently used to be Epstein’s close friend, who still is Pinker’s friend, and who played a crucial role in securing Epstein’s 2008 plea bargain, the one now condemned as a travesty of justice? I’m not sure how I feel about Dershowitz.  It’s like: I understand that our system requires attorneys willing to mount a vociferous defense even for clients who they privately know or believe to be guilty—and even to get those clients off on technicalities or bargaining whenever they can.  I’m also incredibly grateful that I chose CS rather than law school, because I don’t think I could last an hour advocating causes that I knew to be unjust. Just like my fellow CS professor, the intelligent design advocate David Gelernter, I have the privilege and the burden of speaking only for myself.

120 Responses to “On two blog posts of Jerry Coyne”

  1. Enzymes Says:

    Speaking strictly about the 10^(-77) estimate, his essay doesn’t mention how enzymes, aiding or restricting, could affect which mutations go through.

  2. Jacob Says:

    On the topic of complexity and evolution, I’m curious if you’re familiar with Leslie Valiant’s argument that Darwinian evolution should be treated as a learning process which can only produce a subset of PAC-learnable functions (“evolvable algorithms”). He points out that it seems implausible that all of the complicated structures life has produced are evolvable in this sense, and thus that we need new science to explain how we got this way.

    This seems like a much more computationally robust version of what Gelertner is saying.

  3. Paul Topping Says:

    “OK, but what about Alan Dershowitz?” What indeed? This is something I have also wondered about. A few years ago he seemed like a pretty reasonable interviewee for legal matters on CNN, notable perhaps for frequent disagreements with Jeffrey Toobin, CNN’s main legal consultant. Then he seemed to shift his orbit, arguing Trump’s side on more and more issues, finally disappearing completely from CNN. Now I hear, and you mention, he was involved in Epstein’s 2008 legal team. Has Dershowitz sold his soul or does he just want to prove that everyone deserves a legal defence and he’s the guy to do it?

  4. Scott Says:

    Paul #3: Excellent question. I talked with Dershowitz once at a party, but briefly, and certainly not in a way that would give me any insight into these questions. Speaking as someone who often found myself in strong agreement with him, I’m not really a fan of the turn he seems to have taken over the last few years.

  5. Scott Says:

    Jacob #2: Yes, I know Les well and am familiar with his work on evolvability—see for example Section 3.2 of WPSCACC.

    I actually agree with the point that right now, we don’t have any deep principles that would let you predict about how long natural selection should be expected to take to evolve some functionality. A million years? A billion? Right now, one can often do no better than to shrug and say, “well, I guess this MUST’VE been possible to evolve in 5 million years, because 5 million years is how long we see in the fossil record.” I agree that these are profound, fundamental questions, which stand in roughly the same relationship to Darwin and Wallace’s original insight as complexity theory (which ALSO ends up being quite hard…) stands to computability theory.

    Also, of course, even to entertain an idea like the anthropic multiverse, is to entertain the possibility of sheer luck priming the pump of natural selection a little bit. At which point, the main questions remaining are quantitative (just how much luck are you willing to allow in your theory?) and semantic (do you call the luck the “anthropic principle” or do you call it “God”?). It seems to me that, for natural selection to constitute a non-vacuous explanation at all, it needs to have done the overwhelming majority of the work, and indeed many lines of evidence strongly suggest that it did. The question of whether natural selection did 100% of the improbability pumping or “merely” 99.99% of it might never be empirically answerable—you might always need to appeal to Occam’s Razor to bridge the gap between the one and the other.

    The one part of this “pared-down ID thesis” that I really fundamentally disagree with—and it’s a central part—is the idea that you can PROVE that natural selection can’t have done 100% of the work, by looking at some complex adaptation, scratching your head for a while, and concluding that you can’t see how it could’ve been done. Like, OF COURSE you can’t do that! What could possibly make you foolhardy enough even to think about trying it? Do you have ANY IDEA how poorly the appeal to personal incredulity fared the last few hundred times people invoked it for this exact purpose? (If not, read Dawkins or Coyne or almost any other neo-Darwinian; they’ll be happy to give you examples.)

  6. Jacob Says:

    Right, I read WPSCACC and should have remembered you citing him! I think the Anthropic Principle can take you pretty far, if you let it, when it comes to abiogenesis and to the development of intelligence, but not when it comes to things that either seem to have evolved multiple times (like eyes in cephalopods and mammals) or that aren’t in our direct evolutionary path or both. I mean luck can explain anything but not in a useful way that lets you build inferences.

    Ultimately ID proponents would love for us to think that Darwinian natural selection is the only possible natural process that evolution can use, but we already know of others (horizontal gene transfer, epigenetic inheritance, and gene drives for instance). It’s not a binary choice between “natural selection explains everything” and “God did it”. To your point, the question to ask is *how much* do we have to rely on other mechanisms to explain the observed facts? Are these just quirks of biology or do they do a lot of explanatory work?

    My mind boggles when I think about the work that will be necessary to be able to look at an organism and say “here’s the function it computes” in a way that lets us really look at its evolution from a CS perspective. But it beats gesturing and sputtering about watchmakers for sure.

  7. anon Says:

    “the question of how divorced we computer scientists are from the real world”

    There are some oldish survey data showing that mathematicians twice likely to believe in God compared physicists. So it may have some effect if someone is Platonist (in weekdays), in accepting revelations by faith…


  8. Hyman Rosen Says:

    The stupidity of ID is that beyond these appeals to complexity (that always turn out to be wrong anyway), it takes them about two seconds to pivot from that to burning bushes and virgin births and bleeding statues, as if disproving evolution (which they have not done) ratifies all the supernaturalist garbage that they want to believe is true.

  9. Edan Maor Says:

    This is somehow the first time I’ve seen your video on free will vs. determinism, and I have to say I found it excellent. Most of your view mirrors my view, but then you go far beyond it into the practical consequences derived from the base assumptions, specifically that the abstract, to-my-mind pointless word game of what “free will” even means, can be turned into a very practical of question of what-we-can-do-in-practice, which suddenly *has* meaning.

    Anyway, all this to ask – are you planning on replying directly to Jerry Coyne re: free will vs determinism?

    (For the record, I haven’t read The Ghost in the Quantum Turing Machine, sounds like I should get around to it!).

    Unrelatedly, your thoughts on meeting “titans of industry” reminded me of this old blog post from Eliezer Yudkowsky (titled “Competent Elites”): https://www.lesswrong.com/posts/CKpByWmsZ8WmpHtYa/competent-elites. Was wondering what you think of it and how it fits with your experience.

    Key quote (edited):

    > I was invited once to a gathering of the mid-level power elite, where around half the attendees were “CEO of something”—mostly technology companies, but occasionally “something” was a public company or a sizable hedge fund. […]

    > [T]hese people of the Power Elite were visibly much smarter than average mortals. In conversation they spoke quickly, sensibly, and by and large intelligently. When talk turned to deep and difficult topics, they understood faster, made fewer mistakes, were readier to adopt others’ suggestions.

    > No, even worse than that, much worse than that: these CEOs and CTOs and hedge-fund traders, these folk of the mid-level power elite, seemed happier and more alive.

  10. Former Arnold lab postdoc Says:

    “a random change to an amino acid sequence will pretty much always make it worse”

    That’s certainly news to those of use who have worked in directed evolution. Usually screening somewhere between 1000 and 10,000 random mutants is enough to find a variant with an improvement in the property of interest. A fact that’s very fortunate for people who were working back before the lab had high throughput screening robots.

  11. Scott Says:

    anon #7: Right, I knew that mathematicians are more religious than other scientists, although still a lot less religious than the general population. But I think it also matters enormously what kind of religiosity we’re talking about. There are many people who nominally believe in God, go to church or synagogue for the tradition and community, etc., but who in their “revealed preferences” (and “revealed beliefs”), I’d consider much, much closer to atheists or agnostics than to “believers” as the latter would’ve been understood for most of human history.

  12. Scott Says:

    Former Arnold lab postdoc #10: Thanks!! I knew that even if one granted the claim about individual point mutations virtually always being harmful, the IDers’ desired conclusion didn’t follow. I found it harder to judge the claim itself, and find it interesting to hear you say it fails as well.

    (Well, I guess a crucial part of the story is also changing environments that make previously harmful mutations suddenly beneficial—as we already see in different human populations over the last 100,000 years, as natural selection adjusted malaria resistance vs sickle cell risk, lactose tolerance, melanin, etc etc in response to local conditions.)

  13. Scott Says:

    Edan Maor #9: Yeah, I remember that “Competent Elites” essay as one of the posts by Eliezer that the Sneerclubbers ridiculed/attacked the most heavily.

    I mean, I’ve met leaders in the business and tech worlds for whom Eliezer’s description rang close to true: bubbling with energy and intelligence, incredible listeners, lightning fast at absorbing and acting on new information, etc. I’ve also met prominent business and tech people for whom the description was, to put it mildly, not true. But I can easily believe that Eliezer wrote that essay shortly after attending a party populated by the former type of person, and was honestly relating his experience.

  14. Daniel Weissman Says:

    With Pinker, I wish you hadn’t said that he was being attacked for being in a photo with Epstein. As he mentions in his post on Jerry’s blog, he contributed linguistic analysis to one of Dershowitz’s legal documents defending Epstein. (He also mentions that he rode on Epstein’s plane and used to run into him fairly often at Harvard.) I’m not saying that that justifies the attacks on him, but I think it’s important to address or at least acknowledge the strongest argument against him rather than the weakest.

    With Dershowitz, I think his advocacy of collective punishment and legalized torture may have contributed more to harming the world than his defense of Epstein.

  15. Daniel Weissman Says:

    I realized that the last sentence in my comment #14 above was hard to parse. I’m saying that I think Dershowitz harmed the world with his defense of Epstein, and even if you think that we should give him a pass for this to protect the general rule that people accused of crimes should have effective defenses, he’s also done other really bad stuff that can’t be justified in this way.

    (Not saying that people can’t be friends with bad people or whatever.)

  16. Scott Says:

    Daniel Weissman #14: Pinker says that Dershowitz asked him for a linguistic analysis, so he provided one as a favor (as he normally did), without being aware that it had anything to do with the Epstein case. (But that he still regrets his peripheral involvement, knowing what he knows now.) So if this is the “strongest argument” against Pinker, then the case seems damned weak.

    I did do a fair bit of Twitter “research” before posting this, and Pinker’s critics kept displaying the photo of him with Krauss and Epstein over and over and over, so they clearly considered it damning. This despite the facts that (1) Epstein is known to have aggressively sought out photo-ops with everyone in sight, and (2) in my experience, few academics ever say no to photo requests.

    (Despite my own “celebrity” being <1% of Pinker’s, I get asked often enough to pose with students, etc., for instance after giving lectures, and have never once said no. Who knows how many of the students may have been monstrous criminals?)

    Regarding torture, I think the question is slightly more complicated than your description suggests. In practice, virtually no one (including you, presumably) would oppose the use of torture in the ticking-bomb-hidden-in-an-orphanage scenario. So the position opposed to Dershowitz’s is actually a subtle and non-obvious one. It’s that, even though torture would be justified in such a case, because of the potential for abuse that fact should never be codified into law, and instead law itself should simply be set aside if absolutely necessary.

  17. Daniel Weissman Says:

    Scott #16: Hey, could you quickly re-read my comment #14? I don’t think you’re disagreeing with anything I said, including the last bit, where I made sure to include the word “legalized” to cover exactly your point.

  18. Scott Says:

    Daniel #17: OK, reread, and glad to know that we’re in agreement!

  19. anon Says:

    In regards to biology and free will:
    You should check out https://en.wikipedia.org/wiki/Anosognosia – it’s a neurological disorder where due to brain damage, paralysis occurs and the person is unaware of its existence. When asked to move a limb, the person will unconsciously make up excuses like “i don’t want to move it” or “i’m tired”.

  20. ppnl Says:

    Dershowitz may be a dirt bag but not because he defended Epstein. The more unplesant the facts the more honor falls to Dershowitz as long as he does not violate the ethical principles of his position. Anybody remember the central park five? A lawyer is not guilty by association with their client. That way lies madness.

    Pinker has nothing to be ashamed of. We should all be willing to help insure that even guilty people get the best defense possible.

  21. Johnnie Farragut Says:

    Hi Scott. Wanted you to know that I’m a 9/11 truther. Don’t know exactly who carried out the attacks or why but am sure it was not principally Osama and friends. William Ramsey has a nice book about this. Please don’t dismiss the idea simply because you think it is undignified or that people would consider you a credulous fool.

  22. Scott Says:

    Johnnie #21: Out of genuine curiosity, are you also a moon landing truther? Sandy Hook truther? Lincoln assassination truther? Holocaust revisionist? Did WWII happen? Did George Washington live? How do you know?

    What I’m trying to get at is: what (if any) are the cases where you say, “I have no reason to doubt the overwhelming historical consensus about this event”? Where a gigantic conspiracy to falsify the consensus, fabricate records, cover up having done so, etc.—a conspiracy that never stumbles and collapses, that keeps the knowledge of its own existence forever on the fringes—where that strikes you, too, as too implausible to survive Occam’s razor?

    Then, whatever your answer, what features of those other cases made the historical consensus work reliably for them, even though (to your mind) it failed so dramatically with 9/11?

  23. Johnnie Farragut Says:

    Scott, here’s the thing: the people who did 9/11 are laughing right now because they know that they could literally get on television and explain precisely how and why they brought down the towers and people like you would not believe it–simply because they don’t want to appear foolish. TWA 800 is an indisputable instance of this phenomenon. The government says the explosion was an accident but literally hundreds of reliable eye witnesses saw a missile take out the plane. No big conspiracy here: just a screw up by a navy warship. But the NYT and every other news outlet essentially refused to cover the story. My standards for judging the veracity of historical events are EXACTLY the same as yours except that I’m willing to look foolish if the truth so compels me.

    Please look into TWA 800 and if you agree it was a coverup say so here! We got to spread the word!

  24. Scott Says:

    Johnnie #23: Sorry, but I’m banning you from further participation on this thread—not because of your views, but because of your non-responsiveness to my questions.

  25. Jo Says:

    I’m not sure Epstein is devastated by Pinker’s dismissal. Is Epstein a patron of the sciences etc. or is he just a con man? Manipulating “important” people (including academics) to have a social entourage and social credentials as a shield and asset, not unlike he elaborately manipulated teenage girls? In this case, it was just a “show” and he doesn’t care what Pinker or others think.

    Regarding Pinker’s attackers on Twitter, I didn’t “research” down this particular rabbit hole (I have more entertaining ways to waste my time 😉 ) but I would guess it’s the usual stuff: outrage, bad faith arguments etc.
    What you seem to forget (?) is that social media (Twitter, Facebook) works for conveying emotions, not facts and Cartesian reasoning. It was built that way, because emotions are simpler, sell more, and “engage” people more, are more immediately “viral”. Once you accept this, you realize that it’s difficult to have a productive conversation on these media. And that it very often worsens fast, whatever the subject, political opinions of the participants etc. (As an aside is one of the main differences between “alt right twitter”and what you call “woke twitter” the level of organization? It seems more industrialized on the alt right side)
    So! Real conversations seldom happen on Twitter… But is Twitter really necessary to our modern life? There are still blogs like yours to get a message “out”, Twitter can work for “technical” or purely informative/work stuff, and two decades ago, we didn’t have all this and were functioning anyways. The only people interacting with a lot of strangers were people/politicians and they had press offices dealing with the crazies.
    Re: why the attacks focus on the picture of Epstein/Pinker? Because a picture is what’s viral on Twitter/Facebook! Again it’s engineered this way and a picture allows an instant “gut reaction”, allows to propagate an emotion quicker than a five-point thesis.

  26. Craig Says:

    The macro evolution theory is that life came out of mud zillions of years ago via some type of spontaneous generation. Yet nothing like this has ever been observed today. Louis Pasteur even proved that spontaneous generation does not happen. So it is trivial that intelligent design is the only possible explanation for life. I have never heard anyone refute this.

  27. Craig Says:

    As for 9/11 truthers, I know a firefighter who went to the Pentagon that day. According to him, there is no question that a plane hit the Pentagon that day. He saw the wreckage. I have known him for more than 40 years and I trust him.

  28. Michael Says:

    @Scott#16- I don’t think Pinker gets off so easily. He knew that Dershowitz was a defense attorney- why did he think Dershowitz was asking? I realize that getting someone off on a technicality is only morally questionable if the defendant is actually guilty but Pinker arguably should have asked more information about the case before writing the analysis. He was guilty of treating this like an abstract problem and didn’t consider any possible negative consequences in real life.

  29. Scott Says:

    Craig #26: Banned from further participation in this thread, for confusing evolution by natural selection (i.e., the topic we were talking about) with abiogenesis (something that’s indeed never been observed, just like the creation of life by an intelligent designer).

  30. Mikko Kiviranta Says:

    I read some years back a book “Arrival of the fittest” by the evolutionary biologist Andreas Wagner. His message was that the myriad-dimensioned space of amino acid configurations as proteins does not consist of isolated points of “useful” configurations, separated by vast regions of un-livability. Rather, the space is full of filaments and membranes providing gradual pathways from a configuration to another, without un-livable interruptions. Same applies to the configuration space of the possible metabolic pathways.

    These results supposedly come from computational biochemistry. It is in my todo-list to read some of Wagner’s scientific papers, to get an idea how the heck one can obtain such configuration-space maps computationally. I was under impression that already protein folding is pretty much an intractable problem, let alone the question of what metabolic pathways these proteins might form.

  31. Scott Says:

    Michael #28: So, suppose the world’s guiltiest defendant is on trial—I dunno, Jerry Sandusky. And suppose that for some reason, the prosecution has chosen a line of attack that partly depends on the false belief that quantum computers can search N-element unsorted lists in O(log(N)) time. I’m called to the stand as an expert witness.

    What’s the moral course of action here: do I lie, tell the truth, or refuse to testify?

    Note that, as difficult as that conundrum might be, it’s very different from the situation Pinker faced. The latter was more like: some members of the legal profession knock on my door, wanting to know for some weird reason whether quantum computers can search N-element lists in O(log(N)) time. In this case, it seems completely obvious that I just tell them the truth and send them away.

    Yes, the information I provided might be used to free Sandusky. But it might also be used to free some poor black kid who was thrown in jail for marijuana possession. Even without knowing how it will be used, though, it seems clearly better on average for the judge and jury to have true rather than false information on which to base their decisions!

    And if I later found out that my answer was used, even in some peripheral way, to help free Sandusky? Well, I’d probably feel tremendous regret about that, just like I’d feel tremendous regret about buying a flight for a loved one that ended up crashing into the ocean. But it’s not obvious that it would be actionable regret—i.e., the kind that tells me what I ought to do differently the next time.

    You can coherently maintain, it seems to me, that Pinker should never have become friends with Dershowitz in the first place—since he should have foreseen (with superhuman prescience?) that friendship with a world-famous defense attorney might someday put him in a morally compromising situation just like this one.

    But once they were friends—well, if a friend came to me with a technical question in my field of expertise that I knew the answer to, it’s hard for me even to imagine a state of mind where I don’t just immediately tell them. “Maximally nerd-snipeable” is not the only possible way for a human being to be, but it’s certainly the way I am! And even if you put me behind a Rawlsian veil of ignorance, I’m still going to choose to live in a world where some people can freely share their best approximation to the truth without needing to know who benefits from it.

  32. askew2 Says:

    Mikko #30: That’s a great book, and Wagner also has a nice presentation on YouTube somewhere.

    One impression that I got that might be interesting to someone like Scott is that there is a mathematical component to Wagner’s results: the extremely high dimensionality of genome space makes it surprisingly easy to find a feasible (survivable) path from point A to point B, and in fact that is the default in some sense. The space of survivable genomes is, as you say, less separated than our low-dimensional intuition might lead us to believe.

  33. Richard Gaylord Says:

    “.argument from incredulity”. great phrase. did you come up with it?

    As for Gelertner, it’s a pity to watch the creator of Linda descend to his current views

    it makes me wonder if being the last unabomber victim damaged more than his body ( jhttps://yaledailynews.com/blog/2007/01/29/unabombers-act-still-affects-prof-gelernter/).

  34. Sniffnoy Says:

    Scott #31: Agreed on the moral principle, but funny you should use that example… is Jerry Sandusky actually guilty of the crimes he was convicted of? That’s really not so clear at all! (Or, for a shorter version I turned up while trying to find this article again, seems that Jerry Coyne blogged about it!)

  35. Bunsen Burner Says:

    Michael #28

    Even if Pinker knew who Dershowitz was defending he still had a duty as an academic to provide expert technical help when called upon. I would be disappointed in any academic that decided to prejudge a case or defendant before the judge/jury had. It shows the extent to which intellectual life has degraded that many people struggle with the basic concepts that make modern liberal societies possible.

  36. Scott Says:

    Richard Gaylord #33: I believe the phrase “argument from personal incredulity” is due to Dawkins.

    As for Gelernter, let’s try to avoid speculations of a personal nature, and keep the focus more on his work and arguments. Thanks!

  37. fred Says:

    Regarding evolution.

    1) Doesn’t the term “natural selection” suggests that there’s some sort of “choice” going on? Or at least a way for the universe to select between multiple paths of “potentiality”.
    But if the universe is deterministic (i.e. everything is made of subatomic particles that don’t have any choice, with a without a bit of pure randomness), this notion is really just a human brain construct, on the same level as the idea of free will: even if it’s useful and practical for us to view the world as interactions between free independent agents, it’s nonetheless an illusion.
    There’s something else that’s going on, at a more basic level, related to the laws of physics and the inevitable rise of complexity.

    2) Even if the appearance of conscious life on earth is a mathematical impossibility (from a solution space point of view), it’s only impossible if the universe really follows a single path of evolution. But if the Many-Worlds interpretation of QM is true, then the appearance of life is inevitable even if it happens only in the tiniest subset of all possible paths. And a failure to discover any other trace of life in the universe could be a hint that the MW interpretation is true (as opposed to a hint that “intelligent design” is true).

  38. Scott Says:

    Sniffnoy #34: WOW!!!!

    Before reading Crews’s piece, I had had no idea that the case against Sandusky was based so heavily on the discredited “recovered memories” technique, or that the accusers’ original recollections were so often innocuous.

    Frederick Crews is one of my heroes for his Postmodern Pooh book, and his Skeptic article is required reading for anyone who followed the Sandusky case. Honestly, the thought that I, too, might have unwittingly joined a mob rushing to assume guilt made me tense up in my abdomen as I read.

    I could edit my earlier comment to replace Sandusky by a more inarguable “world’s guiltiest defendant”—James Holmes? Adolf Eichmann?—but I think I’ll leave it as is. After all, my possibly too-careless choice was why Crews’s piece was brought to my attention at all.

    I’m still troubled by all sorts of details from the affair that Crews doesn’t address—like the time Bob Costas asked Sandusky point-blank on TV whether he was sexually attracted to young boys, and Sandusky hemmed and hawed for 16 seconds before finally answering “no.” (As opposed to: “Hell no! Are you crazy?!?”) In any case, though, there’s a whole range of intermediate possibilities, wherein Sandusky (e.g.) really did engage in kissing, groping, etc. of boys, but the prosecution felt the need to beef up its case with therapist-implanted memories of rape.

    Thank you for bringing this to my attention.

  39. fred Says:

    Scott #5

    “Also, of course, even to entertain an idea like the anthropic multiverse, is to entertain the possibility of sheer luck priming the pump of natural selection a little bit. At which point, the main questions remaining are quantitative (just how much luck are you willing to allow in your theory?) and semantic (do you call the luck the “anthropic principle” or do you call it “God”?).”

    I was just wondering about that.
    It’s possible that one of the early links in the evolution of life is very unlikely, e.g. the apparition of intelligence.
    But from our own observation, it’s seems that once carbon based intelligence appears, silicon based AI follows shortly, possibly leading to a very different kind of “natural selection” going forward.
    And if we imagine that general AI is possible, and if we assume it would be conscious like we are, its main interest will likely be the study of its own consciousness – unlike us, it will be able to arbitrarily modify its own brain and see what it feels like (unless we prevent it from doing this?). It’s not clear at that point whether this new form of life would be still be interested in spreading physically in the outside material universe vs spreading in simulated universes within itself.

  40. Former Arnold lab postdoc Says:

    Mikko Kiviranta #30: Indeed. We reviewed a lot of the literature on this in group meetings back when I was in Arnold’s lab. Sequence space is both unimaginably huge *and* pretty compact, depending on how you think of it. For a 100 amino acid chain there are 10^130 sequences, which is certainly a big number. But on the other hand, every sequence is at most 100 steps away from every other, and most are much closer than that.

    Throw in the fact that there are neutral networks extending throughout this space and occasionally intersecting with each other, and sequence space starts to look more like a jungle than an empty desert.

  41. don Says:

    A stunning example of local optimization of protein sequences is the adaptive immune system of vertebrates. In their lifetimes, individuals independently generate an enormous repertoire of immunoglobulin sequences that recognize diverse disease antigens, etc. I.e. solutions to a disease challenge are found on a time scale of days or weeks. The molecular mechanism is straight-forward: 1) somatic DNA rearrangements that assemble immunoglobulin genes from a combinatorial library of gene segments coupled with 2) random mutations inserted at the segment junctions by error-prone DNA repair. The system evolved from cut and paste transposons that are a ubiquitous feature of all genomes.

  42. Gabriel Says:

    The anthropic principle is not a good explanation at all for evolution. The evolutionary tree of life is chock-full of huge, splendid branches that are “unnecessary” for our energence. For example, there have been many mass extinctions at different time periods. Each one included an amazing variety of creatures that got extinct. (It’s not just one mass extinction of dinosaurs and that’s it, as I imagined.) I learned this from watching YouTube videos of PBS’s “Eons”.

  43. ppnl Says:


    So would you have refused to furnish aid to the defense of the central park five?. Your moral responsibility is to say the truth. You are only in morally questionable territory when you pick and choose when to speak the truth.

    Getting someone off on a technicality is not morally questionable. If a cop beats a confession out of someone then that confession should not be used. It should not be used even if it lets a guilty person get away. We have things like Miranda exactly because of abusive practices by the police.

  44. Scott Says:

    Gabriel #42: Yeah, that point was already made higher in the thread. The anthropic principle can at most be used to “explain” (I use the word advisedly) the emergence of humans, not the whole rest of the tree of life. From an anthropic perspective, the rest of the tree could be seen as “just a byproduct”—in the sense that any evolutionary process that produced us would necessarily produce lots of other stuff. This is one reason among others why many people dislike anthropic reasoning and try to avoid it when possible.

  45. Scott Says:

    ppnl #43: Alas, it’s now the general educated opinion that the rights of the accused is a profoundly important concept—except with crimes of a sexual nature committed by people high(ish) in the privilege hierarchy, where anyone who breathes a word about the rights of the accused must secretly cheer their crimes. I’d describe this situation as inherently unstable, except for the fact that there are even larger hypocrisies that persisted for hundreds or thousands of years.

  46. ppnl Says:


    Using Sandusky actually makes the point clearer. From Fatty Arbuckle to the central park five we have a habit of letting moral panics distort justice.

    But Sandusky is very probably guilty. At the very least he engaged in some very creepy behavior that should have got him fired. “Very probably” only fails in a court room. I would have no problem helping with his defense but I would never allow him unsupervised interactions with children.

  47. Scott Says:

    ppnl #46:

      I would have no problem helping with his defense but I would never allow him unsupervised interactions with children.

    That sounds reasonable to me (not that there’s any danger of the latter anymore).

    I wish the police and prosecutors had pressed the case without, apparently, contaminating the witnesses so heavily that the original, “unedited” picture of what happened might now be forever lost. There’s probably a lesson there.

  48. william hird Says:

    @Hyman Rosen comment #8
    In a universe ruled by the Second Law, how is it that nature (natural selection) developed the ability to encode all the information to reproduce a complete human being into a cell the size of a period at the bottom of this question mark ? That’s a LOT of information packed into a small space, hard to believe that it is just a result of billions of years of “accidents” 🙂

  49. Scott Says:

    william hird #48: No, a billion years of accidents + saving the accidents that worked.

    Local search is an amazing class of algorithms! Do some reading!

    Incidentally, the computer industry has now achieved comparably impressive information storage densities, in less a hundred years blundering around plus saving the ideas that worked. (Though admittedly, the blundering around wasn’t quite as blind as evolution’s was…)

  50. Jorge Says:

    A few others objections to Gelernter’s #1 come to mind:

    – An amino acid change will generally make things worse… in modern organisms. During the early stages of gene optimization during evolution, random amino acid changes were probably predominantly neutral or slightly positive. v0.01 of DNA polymerase was undoubtedly complete crap compared to what we have now.

    – A lot of organismal evolution happens not at the level of amino acids, but at regulatory nucleic acid sequences. These control where and when a protein or regulatory RNA is expressed, and that can have major consequences for body plan and development, all without altering the target protein’s sequence.

    As for the rest of the post, it just underlines why Twitter is a complete trash fire and I will never post on it. It has a multiplicative effect on one’s intentional (but misguided) opinions or unintentional social faux pas, and serves as a perfect platform to be attacked by both really obnoxious antagonistic people with agendas AND bot amies from god-knows-where intending to sow social discord. The best discourse now happens in on podcasts and the comment sections of carefully curated places in the blogosphere.

  51. fred Says:

    Scott #44
    “Yeah, that point was already made higher in the thread. The anthropic principle can at most be used to “explain” (I use the word advisedly) the emergence of humans, not the whole rest of the tree of life. From an anthropic perspective, the rest of the tree could be seen as “just a byproduct””

    By that same logic the anthropic principle can be used at most to explain your very own emergence, as consciousness experiencing Scott Aaronson reading this sentence in this very moment, and everything else is a byproduct.

    Remember this as you will witness yourself aging on and on, avoiding countless accidents by just a hair, getting “terminal” diseases that will slowly grind you down but never put you out of your misery, while everyone else around you will die at a reasonable age… until finally you’re so old that senility sets in, dissolving all your memories and cognitive abilities, and there’s no longer anything left (in all branches of the MW) in consciousness of what it’s like to be Scott Aaronson.

  52. Lou Scheffer Says:

    Arnold lab postdoc #10:

    More support for evolution mutations being helpful at much more than 1 in 10^77 rate: We can see evolution happening in fruit flies in natural habitats over seasonal scales, as selection pressures change. See, for example “Rapid seasonal evolution in innate immunity of wild Drosophila melanogaster”, at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5784205/ . And if you can get any change you can see at all in 1 year, that change to the billionth power is how much difference you can get over billions of generations. I think this is where the incredulity comes from – an octopus and an ear of corn seem completely different, but when you look closely, a tiny change iterated billions of time can evolve both from a common ancestor.

  53. Michael Says:

    That article by Crews is a complete distortion of what happened. HJ Hornbeck has an article here debunking Crews’s article:
    Crews’s article is a classic case of anomaly hunting. A lot of his article is devoted to accusing Sandusky’s accusers of being greedy, or being bad people.
    Moreover, not all of the accusers allegedly underwent recovered memory therapy. Let’s look at them in the order of his article:
    2- Zachary Konstas- Crews says that he underwent psychotherapy and that “may” account for his changing his story from saying nothing happened. There’s a big difference between psychotherapy and recovered memory therapy. If he spent hours on the couch whining about his mom being a bad mother, I don’t think that would make his testimony less accurate. A more likely explanation of his behavior is that he was scared of Sandusky.
    5- Michael Kajak- there’s no allegation that he underwent recovered memory therapy at all- just a claim that he might be motivated by money.
    6- Jason Simisko- Crews claims that recovered memory was invoked because Simisko said “I tried to block this out of my brain for years”. And if I say that “the horrors of the Holocaust were inconceivable”, would Crews assume that I am unable to imagine gas chambers?
    8- Sabastian Paden- Crews says that he testified “with or without therapeutic prodding”. In other words, there’s no evidence that he saw a therapist.
    9- Ryan Rittmeyer- again there’s no allegation that he underwent recovered memory therapy, just a claim that he might be motivated by money.
    A lot of Crews’s article is based on saying that Sandusky couldn’t have committed these crimes because he’s a good man, that his accusers and their families are bad men and that the accusers got the details wrong years later. IOW, the excuses that criminals usually make.
    Crews complains that the accusers changed their stories. As this website on the Sandusky case explains:
    “I spoke with myriad experts on child molestation—all of whom independently agreed that the vast majority of victims adjust their stories numerous times. That the norm is to change your story, not stick to one”.
    Finally, let’s get to the crux of the matter- Crews’s argument that some of the witnesses were subjected to recovered memory therapy. Here’s what the appeals court found:
    “At the PCRA hearing, D.S., A.F., and B.H., testified, as did Dr. Gillum, A.F.’s therapist, and Dr. Loftus, Appellant’s expert witness. The PCRA court summarized its findings as follows:

    During his direct testimony, [Dr.] Gillum . . . plainly and credibly stated, “I don’t deal with repressed memory [and] I don’t work with anyone who claims to have repressed memories or anything along those lines.” He further articulated his negative assessment of repressed memory therapy and why he did not engage in it. While [D.S.] acknowledged that he and his therapist had discussed methods of unearthing repressed memories, . . . he stated definitively that he had not undergone that type of therapy prior to [Appellant’s] trial.
    Dr. Loftus had a different opinion based on “impressions” from [Dr.] Gillum’s book, statements [D.S.] made two years after the trial, and the fact that the victims whose excerpted trial testimony she reviewed did not give consistent stories to the police, the grand jury, and the trial jury. Having been rendered after an uncritical review of an absurdly incomplete record carefully dissected to include only pieces of information tending to support [Appellant’s] repressed memory theory, however, that opinion was entirely ineffective to rebut Gillum’s and [D.S.]’s definitive denials.”
    “Moreover, despite Appellant’s attempts to establish otherwise, the PCRA court’s conclusion that the victims did not undergo repressed memory therapy prior to trial is supported by the record”.

  54. William Hird Says:

    @ Scott # 49
    Yes Scott , how does nature “know how to save the accidents that work”, what does that mean? You are relating this to some kind of survival function? Nature encoded all the information required to create a whole human being into a volume the size of a dot so that what, it is too small for predators to find and eat for snacks 🙂

  55. Scott Says:

    William #54: It means natural selection.

    From now on, all comments dealing with natural selection will be left in the moderation queue if they don’t demonstrate a basic comprehension of the idea—and in particular, that it’s not just “random accidents,” but rather a local optimization algorithm, optimizing for the ability to make copies.

    As the social-justice types like to say, this is not a 101 space. 😀

  56. Roe Says:

    The thing with Pinker advising Dershowitz is a bit more fishy than Scott writes, I think. The specific piece of ‘linguistic opinion’ which Pinker gave for the case was about the meaning of the law about using the mail to induce/persuade minors to sex. If a close friend came to Scott’s office to ask him about log(N) quantum search, I wouldn’t be surprised if he quickly said, “not possible, and you can quote me”. But if someone looking for a linguistic opinion goes to the office of a close friend in the department of psychology (? — granted, they have some linguistics expertise, but why not ask a Semantics expert from the department of linguistics) and asks about the narrow reading on the meaning of a statute re luring minors to sex… you think the person would say, “Sure, here’s my reading, and I have absolutely no further questions for you about what exactly this is about…” ?

  57. fred Says:

    Scott #55

    “it’s not just “random accidents,” but rather a local optimization algorithm, optimizing for the ability to make copies.”

    At this point, do we know if the apparition of the first viable self-replicating structure (with an ability to mutate, some proto-DNA) was a random accident or is there an even more fundamental algorithm at work (e.g. linked to the rich chemistry of the carbon atom, etc)?

  58. lewikee Says:

    Apologies for incoming digression. I understand if this doesn’t make moderation. Whenever the anthropic principle is brought up (the more useful, weak anthropic principle), it reminds me of the frustration I have with people who think life must clearly be plentiful throughout the universe. How can they know the odds of abiogenesis?

    Its probability is upper-bounded by its empirical (lack of) observation. But isn’t its lower bound clearly arbitrarily small? All we know is that it is greater than zero. But the amount by which it is greater than zero can be as low as the smallest number you could possibly conceive of. So we have it ranging anywhere between some arbitrarily small ε to the empirical upper bound, and with no idea as to the probability density function. A consequence of being presented with a sample size of one.

    Given this, why do so many people (in their Drake equations, mostly) just assume the upper bound and posit a universe teeming with life? The correct approach should just be: we have no idea how many more abiogenesis events could have happened elsewhere in the universe. This kills the Drake equation, of course.

  59. Scott Says:

    fred #57: It’s not known how the first self-replicating molecules arose (to put it mildly). The whole field of origin-of-life research exists to try to figure that out.

    We do know that many of the chemical building blocks of terrestrial life are abundant in the solar system, and that lots of organic compounds (including amino acids) can be formed from inorganic matter via chemical reactions that would’ve been common in the early earth. (That was the conclusion of the celebrated Miller-Urey experiment from the 50s.)

    Many people hypothesize that RNA preceded DNA, and that the current jury-rigged system (wherein DNA gets transcribed into less-stable RNA) is a relic from the path that self-replicating molecules took ~3.5 billion years ago. (Or however long ago it was—that’s also not known with precision.)

    More fundamentally, it’s not even universally agreed what would count as a “solution” to the origin-of-life problem, and how such a solution could ever be conclusively tested. Also, exactly how much does a proposed “solution” get to say: “look, we know this particular step had a tiny probability, but it’s a vast universe and it only had to happen once in billions of years! And if the coin hadn’t come up ‘heads’ a bunch of times here on earth, then we wouldn’t be here on earth talking about it!”

    Of course, everything would get on a vastly more solid footing if it were ever discovered that life arose multiple times (either on earth or elsewhere in the universe), so that it became possible to do comparative analysis. Absent that, though, there’s still lots of fascinating science (theoretical, experimental, observational…) that’s motivated in one way or another by this mystery.

  60. Scott Says:

    lewikee #58: I violently agree with you about our radical ignorance of the probability of abiogenesis. But I don’t think the problem here is with the Drake equation—that equation is just a memorable way to organize one’s ignorance, and to identify the subquestions where one could start making progress. (To take one example, estimates of the number of “earthlike” planets in Goldilocks zones have vastly improved over the last ~20 years, now that we have the ability to survey thousands of solar systems.) The problem, rather, is with people who indulge in garbage-in-garbage-out—e.g., plugging in completely made-up numbers for the latter terms in the Drake equation, and then taking the resulting estimates seriously because they “came from an equation.”

  61. fred Says:

    Scott #59

    Nice, and at least we now know that earth-like planets aren’t rare at all.

  62. fred Says:

    I can imagine that life, as a runaway process of self-replicating/assembling molecules, could be totally unique to earth, a fluke.

    But once I consider that life also happens to “harbor” consciousness, it seems a bit too coincidental that two such extreme “mysteries” would only happen once, and in the same place.

    Of course it’s possible that the two things are somehow dissociated, i.e. consciousness is a “universal” property.

    Then it gets even crazier if you start considering that life on earth can also export itself to any parts of the universe (in theory) while also leading to processes that can simulate reality and life itself.

  63. ppnl Says:

    fred #57

    It isn’t at all clear what you can even mean by “random accident”. From a chemistry point of view I ‘m not sure that is even coherent. Take Buckminsterfullerene for example. It is composed of 60 carbon atoms arranged in the shape of a soccer ball. The chances that 60 carbon atoms will find themselves in this configuration in some mixture of carbon atoms at equilibrium is essentially zero. That’s just not how chemistry works. But put those carbon atoms in a particular out of equilibrium state and the formation of Buckminsterfullerene is not only possible but inevitable.

    Buckminsterfullerene was discoverer by accident while running experiments trying to understand the chemistry of the gas flow out of old carbon rich stars. It turns out that they accidentally recreated the specific out of equilibrium conditions that allow the formation of Buckminsterfullerene. It turns out that carbon rich stars have been making Buckminsterfullerene that way for billions of years.

    Now the thing is, it would have been very hard to work out how to make Buckminsterfullerene on purpose from first principle. The reaction dynamics are just too complex.

    The origin of life has the same problem but far worse. People are looking at many things to find the answer. Electric sparks through a gas mixture, autocatalytic sets, clay minerals with catalytic properties, protocell membranes formed from lipids, self catalyzing RNA molecules…

    Nobody knows the answer. But the one thing I doubt anyone is looking at is unstructured random chance.

    Some form of selection plays a part in most of the attempts. You have to have replication before you have selection but that replication does not have to be very efficient or accurate. Nonliving things can replicate structure. Crystals and polymers for example.

  64. toto Says:

    #5: ah, so refreshing to read that again. Thanks for the reminder!

    #19: anosognosia is not necesseraly associated with paralysis. It can happen, for example, in blind or early demencia persons. As for the paralysis case they usually (pretend)/(actually believe in all good faith) they see/think normally even while we can demonstrate blindness/cognitive limitations. When facing impossibility some will tone down their claim of normality, some will tone down the mistakes, some will get angry with the questionner -at which point you went too far).

  65. Raoul Ohio Says:

    #58, #60.

    I have always thought the Drake equation is a joke for the reason that lewikee does: there is a factor about which we know nothing. Also agree with Scott’s “just a memorable way to organize one’s ignorance”, although “earthlike”, etc., are not fundamental.

    I have been following the story for five or six decades, and my guess for how often life springs up continues to increase, and is likely to exist in several places in our solar system. On the other hand, my guess for the fraction of stars with advanced civilizations continues to drop as we discover more things that can go wrong.

  66. ppnl Says:

    lewikee #58

    Yes, the probability of life can be arbitrarily small making life rare. But on Earth at least life happened in an eye-blink after the earth cooled enough to allow it. That argues for a fairly high probability of abiogenesis. But other things can kill life.

    We also need to precise in what we mean by “plentiful” life. I think a star trek universe with thousands if not millions of advanced species is pretty much ruled out. But what about a thousand intelligent species over the last ten billion years with most of those extinct? Do you call that common or rare?

    In any case I think intelligent life is rare on any human time and space scale. If we could examine all the life in our galaxy I suspect we would see mostly microbes and maybe a few gods. The chances that there is anything at our level of technology is essentially zero.

    Microbial life may also be rare but I doubt abiogenesis is the gateing issue. The universe is violent and unstable on the relevant time scale.

  67. Lou Scheffer Says:

    Among the other ways that biology offers much higher rates of change than 1 in 10^77 is through gene duplication. Caused by copying errors during replication, this is thought to occur at 10^-7 to 10^-3 extra copies per gene per generation. Once duplicated, one copy can evolve to perform new functions without fitness constraints. This is an enormously more efficient path to generating new genes compared to a huge number of individually useless neucleotide insertions (which would indeed be unlikely). https://en.wikipedia.org/wiki/Gene_duplication

  68. GA Says:

    an often underestimated force in natural selection, which is often overlooked by creationists and other evolution deniers, is that of mate selection.
    the fact that in order to have offsprings you need to cooperate with someone else, who might
    select you or select someone else, actually brings kinds of intelligence into the loop. You don’t need the whole feedback loop of a mutation statistically increasing survival chance – all you need is to evolve a good enough intelligence to recognize which mates have better chance of success. instead of giraffes slowly growing neck because each time it increases their survival chance, you get a consistent growth because each time only the tallest get to mate.

    from the moment nature developed intelligence, even small one, the process became guided by this intelligence. it’s a kind of an intelligence design, but not by someone external – but by the species itself. The species itself decided which traits are good and bad, by choosing who to mate with.

  69. mr_squiggle Says:

    Regarding the odds of abiogenesis, lewikee, one thing is that life seems to have arisen on Earth pretty soon (on a geological timescale) after conditions became suitable for it to function.
    This suggests that if conditions are right on a planetary basis, it’s quite likely to show up.
    Of course, in some sense this does just move the anthropic principle goal-posts around a little; although people talk about the goldilocks zone, we don’t know how precise the conditions have to be really. It might be, for example, that having a large moon is important in setting up the right conditions via tides.

    Also, I’d like to add a comment to Craig #26, in case someone refuting his case makes a difference:
    Suppose that spontaneous generation does happen, but only under some specific conditions, and very rarely. Louis Pasteur set up flasks of … what, a few liters of liquid? The Earth has of the order of a billion km^3 of water, in many different environments (mostly it’s salt-water, of course, but even small fractions of that volume are a lot). Plenty of energy input. And it had millions of years for something to happen. The thing is – when it eventually did, it could spread. It only had to happen once.
    Pasteur’s experiments (and any repeats made since) show that spontaneous generation isn’t a common, everyday occurrence (which is what people thought at the time). They don’t preclude it as a ‘rare’ event.

    (I’m sorry Scott if you feel this is a digression. But it does at least respond to an argument raised.)

  70. JimV Says:

    My immediate reaction was : #2 is a crock (which I won’t go into as I see previous commenters have dealt with it; and #3 is a crock because “Intelligent Design” in the real world uses exactly the same mechanisms (albeit more efficiently) as biological evolution.

    That is, it seems to me that when you’re trying to come up with something new (new computer program, new machine, etc.) you try something based on what you’ve seen or done before with a slight variation, and when that doesn’t work, you try something else, and when that doesn’t work you try something else, until either you find a solution or give up. As a mechanical design engineer and computer programmer for 38 years, that’s the only method I’ve used or seen used. That’s how (new) design works, I am very sure, and probably how intelligence works (if we could see what is going on in the background as unmonitored neurons are firing). There is no supernatural magic to it. We don’t create (ex nihilo) solutions, we find them by searching for them. So the so-called Intelligent Design Theory is a theory of magic without any basis developed by people who don’t understand either design or intelligence. (They might at least start by studying the history of design, e.g., Edison’s 1000 attempts to find a practical light-bulb filament.)

    Like biological evolution, technological evolution snowballs. Modern humans have been around a long time (somewhere between 20,000 and 200,000 years depending on what soft tissue changes might have occurred over that span which skeletons can’t tell us), but we only have evidence of the wheel-and-axle going back 6000 years. That’s a long time for “intelligent design” to find a simple mechanism (itself evolved from rollers under heavy loads). But once you have that, you can have gears, capstans, watermills, windmills, pulleys, etc. Then thousands of years later you look around you and think what geniuses humans are, without realizing all the trial and error that was involved.

    There is an article explaining Einstein’s Zurich Notebook online, which shows how he developed the equations of General Relativity. He tried several things that didn’t work before finding one that did.

    A few days ago I had to install another three Windows updates on my 12-year old Windows 7 laptop (which started as a Windows Vista laptop), making the total somewhere around 300 updates. So much for intelligent computer-program design.

    That pocket watch that Paley imagined finding in a forest which told him that trees and watches were both designed should have told him that they both evolved, had he been familiar with the design history of clocks.

  71. Wes Hansen Says:

    I don’t think intelligent design proponents reject evolution, they simply add intelligence. Consider the following from Michael Behe’s site:

    “Behe uses the term Darwinian evolution to distinguish it from evolution meaning simply change over time, which is not controversial and which he accepts. Darwinian evolution, on the other hand, is claimed to be the result of unguided, naturalistic processes of random mutation and natural selection, which he sees as severely limited.”

    I think Gerry Spence’s response to Alan Hirschfield, regarding his decision to defend – pro bono, Randy Weaver, is relevant to this discussion; it can be found here. It’s quite long, but I would reproduce a short excerpt:

    “I also feel your pain, my friend. Yet I know that in the end, if you were the judge at the trial of Adolph Eichmann, you would have insisted that he not have ordinary council, but the best council. In the same way, if you were the judge in Randy’s case, and you had a choice, I have no doubt that despite your own pain you might well have appointed me to defend him. In the end you must know that the Holocaust must never stand for part justice,or average justice but for the most noble of ideals–that even the enemies of the Jews themselves must receive the best justice the system can provide. If it were otherwise the meaning of the Holocaust would be accordingly besmirched.

    Alan, I agree with your arguments. They are proper and they are true. I agree that my defense of Randy Weaver may attach a legitimacy and dignity to his politics and religion. But it may, as well, stand for the proposition that there are those who don’t condone this kind of criminal action by our government. I view the defense of Randy Weaver’s case as an opportunity to address a more vital issue, one that transcends a white separatist movement or notions of the supremacy of one race over another, for the ultimate enemy of any people is not the angry hate groups that fester within, but a government itself that has lost its respect for the individual. The ultimate enemy of democracy is not the drug dealer or the crooked politician or the crazed skinhead. The ultimate enemy is the new king that has become so powerful it can murder its own citizens with impunity. To the same extent that Randy Weaver cannot find justice in this country, we too will be deprived of justice. At last, my defense of Randy Weaver is a defense of every Jew and every Gentile, for every black and every gay who loves freedom and deplores tyranny.”

    I’m a Buddhist, but Amen to that . . .

    And that’s why I’m a Libertarian: Conservatives and Liberals only care about individuals to the extent those individuals can be used to advance their position in the power game. They rob individuals of their humanity in order to advance a freaking agenda! Yeah, go humans. Here’s my question: Are humans the worst that evolution can do, or are there even nastier entities out there in the vast regions of space?

  72. ppnl Says:

    Wes Hansen #71

    Unfortunately I find it hard to see intelligent design as anything but a dumpster fire of bad ideas. Sorry, I’m not trying to offend you. But it really is a mess.

    Behe, of irreducible complexity fame, believes in an old Earth and common descent. The complexity idea didn’t catch on to say the least.

    William Dembski tried to fix irreducible complexity with his “specified complexity” idea and failed utterly. He also believed in an old earth but denies common descent. Until he recanted and became a young Earth creationist. Then he switched back. Apparently he had been pressured into young earth creationism. But anyway how do you look at life on earth and deny common descent? How is that anything but a flat out rejection of evolution?

    Then we have Paul Nelson. He is a flat out young earth creationist. But he is responsible for exposing the dumpster fire. Nelson Quote:

    ” Easily, the biggest challenge facing the I.D. community is to develop a full-fledged theory of biological design. We don’t have such a theory right now, and that’s a real problem. Without a theory, it’s very hard to know where to direct your research focus. Right now, we’ve got a bag of powerful intuitions and a handful of notions, such as irreducible complexity, but as yet, no general theory of biological design. ”

    No coherent theory at all. So what is ID if it is not a theory? It was intended as a big tent under which all forms of creationism could gather for mutual support and political power. Calling it a scientific theory gave it a thin patina of scientific respectability. Or as Leonard Krishtalka said:

    “Intelligent Design Is Creationism in a Cheap Tuxedo.”

    Anyway I always wondered what that “intelligent designer” was doing for the entire 3.5 billion years of life on earth. Did it really spend several billion years tinkering with bacteria before moving on to worms and snails for countless millions of years? What possessed it to play with giant monsters with big teeth for 250 million years before killing them with a giant asteroid? If it turns out that there is an intelligent designer…the worst that you can say about him is that basically he’s an underachiever. With apologies to Woody Allen.

  73. Jacob Says:

    How are there so many science deniers in this comment section! Of all the blogs to come to! We even got a 9/11 truther!

  74. Bunsen Burner Says:

    mr_squiggle #69

    This is a typical mistake in the analysis of unique events. We aren’t looking at single probabilities but compound ones. We can’t talk of the probability of life arising but only of life arising and leading to the development of us. We don’t know how long it takes life to develop an intelligent, technological species, but it only has a finite time to do this, and it took billions of years on Earth. In that case it’s possible for the probability of life arising to be exceedingly small, but on any world with intelligent life for it to arise early.

  75. Bunsen Burner Says:

    Scott #60

    I completely disagree that we have any meaningful estimate of “earthlike” worlds. I mean why the quotes? Because they’re not earthlike? We have no idea how many of these planets have a magnetic field, continental drift, a large moon to stabilise rotation, and man other qualities. Heck, even topological features such as shallow oceans, are believed to have been important for several geochemical processes. Also, Goldilocks zones are not static. They change with the evolution of the star, and the eccentricity of the orbit. How many of these worlds stay in a relatively stable, near circular orbit, in the Goldilock zone for long enough for life to arise?

  76. fred Says:

    Jacob #73

    I’m the worst (or best?) science denier of all: I believe in superdeterminism!
    The initial conditions at the time of the Big Bang are to blame for my writing this post!

  77. James Cross Says:

    I think there is a sort of graphical curve relating age to how likely you are to persuaded by bad ideas and bad people.

    When you are young the probability is greater. You’re naive. As you get older, it drops. You become more experienced and skeptical Then as you get older still the probability begins to grow again. Mental capacities diminishing?

    This explains how people like Lynn Marguilis can become a 9/11 truther. It also accounts for the large number of scientists in fields other than climate science can become climate science deniers.

  78. Raoul Ohio Says:

    The Pinker + Epstine saga is covered in today’s “Higher Education”:


  79. James Cross Says:

    Regarding odds for life and intelligent life.

    We do have a rough guide from the history of the earth. Life appeared within about a billion years or less after the earth formed. So probability for life elsewhere is like good. However, complex life required more than 3 billion years for it to form. Probability for complex life much less. Probability for intelligent life. We are still awaiting the results. 🙂


  80. Scott Says:

    Bunsen Burner #75: I completely agree about all the uncertainties you mentioned. But, like, 25 years ago we didn’t even know whether rocky planets with atmospheres were anomalous to our solar system, and now we know they aren’t, right?

  81. jonas Says:

    Scott: I see you mentioned “a technical question in my field of expertise” and “anthropic principle”. I have a question.

    In your 2005 article “NP-complete Problems and Physical Reality”, section 9, you talk about the computational model where you can do a polynomial time classical calculation and post-selection. It is known that this model is more powerful than NP.

    David Madore just brought up post-selection on his blog, to which a3nm asked (“http://www.madore.org/cgi-bin/comment.pl/showcomments?href=http%3a%2f%2fwww.madore.org%2f~david%2fweblog%2f2019-07.html%23d.2019-07-16.2608#comment-25797”) what class of computational problems you can solve with a computer and post-selection. If I understand correctly, he’s asking about a model that is equivalent to the one that you talk about in that section. David answers that it can solve at least NP problems by randomly guessing a witness, and destroying the world if the witness is not verified successfully. He then argues that you can also solve P^{NP} problems: just use post-selection as an oracle to solve NP problems. I think the latter part of this argument is wrong, mostly because if it were that simple, then your article and the complexity zoo (“https://complexityzoo.uwaterloo.ca/Complexity_Zoo:B#bpppath”) would say that this model can solve P^{NP} problems, but they don’t seem to do so. I tried to explain why the argument is wrong, but my explanation isn’t very clear. Can you explain to us more clearly why this argument fails?

  82. Raoul Ohio Says:

    Issues like those Bunsen Burner #75 lists (+ wild planetary orbits) were mostly recognized recently. This is why my guess for number of advanced civilizations has gone way down.

    In the other direction, the possibility of tidal heating, also recognized recently, greatly expands the number of places where primitive life might be found. This is why my guess for total life instances / star has gone way up.

    Can anyone imagine a life instance not bound to one or more stars?

  83. Scott Says:

    Raoul #78: That’s one of the most depressing articles I’ve ever read.

    “Unlike the responsible mainstream in the humanities, which considers our entire civilization to be monstrous and in need of violent revolution, Pinker has long isolated himself with incendiary data claiming to show that bourgeouis ideals like liberal democracy, equal rights under law, science, reason, and market economies have lifted billions of people out of poverty. Reading between the lines of his books, one can even find occasional hints that he sympathizes with the inflammatory fringe theory that biology and evolution might be relevant to human nature. Now, however, that a photo of Pinker at the same luncheon as Krauss and Epstein is being shared on Twitter, we can rest assured that the tide of history is turning against these refuted beliefs and the reactionary Caucasian males who advocate them.”

    If and when they succeed in unpersonning Pinker, I think that will be a good time to mark the moment when intellectual freedom is over in academia, outside of the hard sciences.

  84. fred Says:

    James #77

    It’s probably also the case that many people who are ambitious and yearning to make a mark in the world may feel (maybe unconsciously) that they’re better off taking a chance on fringe stuff, outside of the mainstream ideas (especially if their time is running out).

    When it comes to global warming/climate change in particular, my very own (conspiracy) theory is that many of its deniers (esp in the rich and powerful) actually do believe in it, they simply do not want to do anything about it, on purpose: they don’t see it as a problem but as its own solution (i.e. curb overpopulation).
    The question really isn’t whether you believe in climate change but whether you believe that the earth can sustain 10+ billion humans, all enjoying a first world life style.

  85. Scott Says:

    jonas #81: That is indeed a technical question in my field of expertise! One whose answer I know and will share! 🙂

    The difficulty in simulating PNP is that you need to make multiple, adaptive queries to the NP oracle (i.e., each query might depend on the outcome of the previous queries). But in PostBPP (that is, using a classical computer with postselection), you only get to do a single, “global” postselection, not many separate postselections interspersed with non-postselected computations. If you try to simulate the latter, you’ll find that imposing a single postselection hopelessly biases the probabilities. And while you can compensate for that to some extent using amplification, if you want more than log(n) adaptive queries to the NP oracle, then it turns out that you need more than a polynomial amount of amplification, so the algorithm is no longer efficient.

    This might sound like a mere technical issue that there’s some clever way around. But from work by Beigel, it follows that there’s an oracle relative to which PNP⊄PostBPP, which means that any proof of that containment would need to be non-relativizing.

    As the above discussion already suggested, you can simulate polynomially many parallel queries to an NP oracle—that is, queries that don’t depend on each other—in PostBPP. In fact it’s known that

    PNP|| ⊆ PostBPP ⊆ BPPNP||,

    where the || denotes that the queries to the NP oracle have to be made in parallel. This means in particular that, under plausible derandomization assumptions, PostBPP would be exactly equal to PNP|| (which, however, might still be smaller than PNP).

  86. ppnl Says:

    Raoul Ohio #82

    If a life bearing planet is ejected from its orbit microbial life could survive for a billion years on geothermal energy. But I doubt it could originate in deep space. A gas giant may provide enough heat to make life possible.

  87. Wes Hansen Says:

    To ppnl:

    I’m not qualified to engage in any kind of debate regarding Michael Behe or Intelligent Design as a semi-formal institution; I don’t follow Behe’s work or that of any other proponent of Intelligent Design for that matter (Brian Josephson being an exception – and William Tiller if you consider him an Intelligent Design proponent). The quote was a simple convenience, found with a standard Google search! I don’t know what people really mean by intelligent design formally, but it is not the case, in general, that intelligent design requires an “intelligent designer!” I think you are erroneously suggesting a false equivalence between intelligent design, in general, and Creationism – which requires an obviously intelligent creator! As JimV stated above, mechanical design here on planet Earth is quite obviously a case of intelligent design – who is the creator there? Suggesting that there need be an integrated, intelligent entity in order for evolution to be intelligent is, I believe, an error of thought grounded in Western religious and philosophical tradition.

    With regards to the rest of your comment, I thought for an extended moment that you were speaking about String Theory, or M Theory, or K Theory, or . . . are we at Z yet!?!

    The formal scientific definition of theory is quite different from the everyday meaning of the word. It refers to a comprehensive explanation of some aspect of nature that is supported by a vast body of evidence. Many scientific theories are so well established that no new evidence is likely to alter them substantially. For example, no new evidence will demonstrate that the Earth does not orbit around the sun (heliocentric theory), or that living things are not made of cells (cell theory), that matter is not composed of atoms, or that the surface of the Earth is not divided into solid plates that have moved over geological timescales (the theory of plate tectonics)…One of the most useful properties of scientific theories is that they can be used to make predictions about natural events or phenomena that have not yet been observed.”

    So you can’t really fault them for not having a coherent theory! Not without finding fault with a significant number of others! There are many open questions and I try to keep an open mind, ideally one free from unreasonable criticism (except when it comes to Platonism); strong faith drove Newton, and Einstein as well, although in a different way I believe.

    I’ve often wondered how that little pissant, Lee Harvey Oswald, managed to redirect JFK’s motorcade at the last minute; or if he didn’t personally redirect the motorcade, how the little pissant, working completely independently, KNEW it was going to be redirected. I’m not a 9/11 Truther, but the kind of people who fly around in Epstein’s plane and invite Presidents to their state tend to play by a different set of rules; I think this is what Michael Harris was referring to in his Mathematics Without Apologies with the term F%$k You Money.

  88. STEM Caveman dissents Says:

    As sinister as Epstein appears, so far his only established (and most of his alleged) sexual predation consists of being a rather milquetoast employer of amateur prostitutes aged 14-17.
    Whether that’s awful or not depends on one’s views on prostitution and age of consent. I suspect there are civilized places where his conduct would have been legal.

    Of course there are also suspicions of nonsexual impropriety by Epstein such as blackmail, financial fraud, and spying; and a brand new accusation of rape that is unusually sketchy and impossible-to-prove even by the standard of televised rape allegations against disgraced billionaires. Dershowitz has just gotten court records unsealed and made dangerously specific falsifiable claims about what in those records will destroy the credibility of the woman who claimed Epstein forced her (as a “sex slave”) into participating in orgies on his private island. It’s not at all a given, on present evidence, that Epstein or his sex ring were involved in any violence or coercion.

    There will be lots more material about Epstein coming out, but the pearl-clutching going on for years about him being a “sex offender”, “child sex abuser”, and “molester” is hysterical moral panic out of proportion to what has been known or credibly alleged.

    I don’t know Epstein, have never met him or anyone associated with him other than a few impersonal interactions with Steve Pinker and Alan Dershowitz when living in Cambridge. I do think I know a moral panic when I see one.

  89. Artem Kaznatcheev Says:

    I am always worried when reckless computer scientists wander into biology and claim to prove some sort of results without understanding the relevant biology (or, I guess even without understanding the relevant computer science in the case of Gelernter). This can discredit the real contributions that theoretical computer science can make to biology.

    For example, I’ve recently used PLS-hardness results (and also some properties of simplex algorithms) to show how to use the constraint of computational complexity to address challenges that are actually relevant to theoretical biology, like understanding open-ended evolution or the maintenance of costly learning and cooperation. And I’ve worked very hard to publish this in a language and venue that evolutionary biologists actively read and respect, while not compromising too much on the cstheory rigour (although I did have to move all the proofs to the appendix). This is an uphill struggle (that I think you once described as digging a tunnel from Europe to North America — but let met not mix too many metaphors) and it is made more difficult when ‘prominent’ computer scientists go out and say nonsense.

  90. mr_squiggle Says:

    Bunsen Burner #74
    I’m actually not sure what you are arguing here.
    I didn’t discuss intelligence at all; I was only talking about the origin of life, not its development. This is also the case upstream in the threads of conversation it was addressing.
    I do not agree with what I interpret as your line of thinking – that you can’t consider individual elements of a series of events. Of course you can.
    For example, in chemists break down complex reactions into a series of steps. One key simplification is to think about some steps being ‘fast’ and some ‘slow’. The slow steps determine the progress of the overall reaction.

    I also think you’re wrong on not knowing how long it takes an “intelligent, technological species” to arise. We have an example – which gives us an upper bound on the minimum time required.
    I think you /might/ be saying that we only have the one known case and it might not be typical. If so, I agree – but I think that the probabalistic sense of it is still useful.
    My understanding is that there is evidence that life was present on the Earth practically as soon as it could possibly been present – almost as soon as the oceans were formed.

    Therefore, _given the right conditions_, this seems likely to occur within a reasonable time; it is a fast step.
    The origin of the Eukaryotes then took a very long time, absolutely, and I certainly wouldn’t be eager to claim that intelligence was probable from that point on. But the origin of life, that does seem pretty sure-fire.

  91. Joshua Zelinsky Says:

    Scott, #85, can we show that P^NP being in PostBPP would cause collapse of the polynomial hierarchy? The zoo entry mentions that if NP^NP is in PostBPP, then the hierarchy collapses, but that’s a weaker a statement.

  92. Scott Says:

    Joshua #91: That’s a superb question. I don’t know—nor do I even know whether PNP=PNP|| is known to cause a collapse of PH or anything similar. Does anyone else here?

  93. James Cross Says:

    Back to ID for a moment.

    Part of the issue I think is that I am not sure we actually understand what intelligence is. In particular, does intelligence require a conscious being?

    I have been operating on a belief that it does not. What got me thinking in this direction was this paper:



    Recent advances in fields ranging from cosmology to computer science have hinted at a possible deep connection between intelligence and entropy maximization, but no formal physical relationship between them has yet been established. Here, we explicitly propose a first step toward such a relationship in the form of a causal generalization of entropic forces that we find can cause two defining behaviors of the human ‘‘cognitive niche’’—tool use and social cooperation—to spontaneously emerge in simple physical systems. Our results suggest a potentially general thermodynamic model of adaptive behavior as a nonequilibrium process in open systems.

    Now there is good chance I am not understanding this paper or maybe the research is bad. But if we begin to think of intelligence as being able to arise in physical systems without conscious beings then we can explain, for example, the seemingly intelligent behavior of slime molds. Life in that case could be a part of intelligent design but it would be a design without a designer.

  94. JimV Says:

    Belated reply to Wes Hansen at 71, re: “Darwinian evolution, on the other hand, is claimed to be the result of unguided, naturalistic processes of random mutation and natural selection, which he [Behe] sees as severely limited.”

    Again, limited compared to what? What other method is there? Magic? Everywhere I look in the history of human design I see trial and error.

    At the GE Corporate Research and Development Lab (back before Welch when GE did research), a cat knocked over two beakers late at night and their contents mixed. That is how Lexan was invented. The flowpath of the GE-90 jet engine was designed by a computer search algorithm, trying different configurations randomly.

    All you need is:

    1) A way to generate variations in successive generations of something (organism, machine, computer program). The variations can be systematic or random or a mixture, but it turns out some randomness is good because systematic methods can systematically ignore some of the solution space.

    2) Some selection criteria to separate failures from neutrals from successes (survival and reproduction in the wild, survival and reproduction in the marketplace, etc.). (That’s what “guides” the process.)

    3) Some form or forms of memory to pass successes on through time (genes, neurons, textbooks, hard drives, etc.).

    You’ve seen cars and washing machines and phones evolve in your lifetime. There is no magic. “Intelligent Design” and evolution are basically the same process. If there is something different about ID, I challenge the IDers to tell us what it is, how it works, and how it is distinguishable from evolution. Take Edison’s design of the lightbulb as an example–please. By the way, have Behe or the others ever designed something complex that worked, in one try? What makes them experts in design or intelligence?

    (It just seems so simple that I should be able to make people see it–but I’ll stop now, in this thread anyway.)

  95. Scott Says:

    JimV #94: They kept cats in GE’s research labs??

  96. James Cross Says:


    Don’t all labs have cats?

  97. Bunsen Burner Says:

    mr_squiggle #90

    I thought the point was quite clear. The early origin of life on Earth tells you nothing about its probability because a lot more has happened than that, intelligent life also evolved within the bounds of our habitability window. So we really have a conditional probability but nothing to tell us what the values of individual parts are. I believe the argument was formalized by Brendon Carter, but here is a great quote from Paul Davies on the topic,

    “In other words, life on Earth had to get going pretty fast, or there wouldn’t have been enough time for intelligent observers like us to hit the scene before the sun became a red giant”

    And we don’t haven bounds on any minimum or maximum time. Just the fact that it happened like so once. The variance of a statistical sample of one is infinity.

  98. Scott Says:

    Bunsen Burner and mr_squiggle: From a Bayesian standpoint, the fact that life started relatively soon after the earth was cool enough to support it is certainly a data point. I.e., it’s something potentially relevant that we learned in the last century, that wasn’t obvious a priori and that could’ve turned out otherwise. It’s just that it isn’t a very good data point, since (as has already been well explained above) it’s badly contaminated by selection bias. Even a single other data point, one not connected to our own existence, would be enormously more useful for estimating the frequency of life in the universe.

  99. Wes Hansen Says:

    Jim V.:

    Again, I can’t speak for Behe and the ID crowd, but I think he is referring to explanatory power. The rest of your comment is just one long false equivalence. If design is equivalent to Darwinian evolution, then why Directed Evolution? Notice the phrase “protein engineering!”

    I am the sole named inventor on a number of patents; I am also a fine artist – follow the link; I have also invented novel mathematics. So I have an experiential understanding of creativity and the design process. Sure, there is trial and error involved, ideally at the drawing table or in the CAD environment, but conscious intelligence brings intuition to bear – the evolution is intuitively directed in the field of consciousness. Why “conscious” intelligence and not simply intelligence? Because, as William Tiller often points out, a key characteristic of consciousness is the manipulation of information in an intelligible manner – puzzles, games, mathematics, engineering. We can design neural nets to make novel associations, but these associations only become meaningful in the field of consciousness. Does the Universe at large require an analogous meaning? I don’t know, but neither does anyone else! And I see no reason to let unwarranted dogma limit scientific inquiry; as Warren Buffett said, “What we learn from history is that people don’t learn from history.” I think this comment thread proves the point:

    “He compared the lack of acceptance of telepathy by scientists such as himself with the way in which the echo-location system had been discovered in bats, followed by its rapid acceptance within the scientific community in the 1940s. In fact, as I later discovered, Lazzaro Spallanzani had shown in 1793 that bats rely on hearing to find their way around, but sceptical opponents dismissed his experiments as flawed, and helped set back research for well over a century.”

    That’s just one example from an interaction between Richard Dawkins and Rupert Sheldrake.

    And I think that this “explanatory power” question is the source of the extreme resistence to the so-called “para”normal data set. Meta-analyses are performed – and conclusive, but then updates are necessitated due to unreasonable criticism; there is clearly a double-standard here. If the question is so obviously closed, then why the extreme resistence? Why the crazy double-standard?

  100. Daniel Says:

    Off topic..have you read Lee Smolin’s latest book ? What do you think ?

  101. ppnl Says:

    James Cross #93

    Part of the issue I think is that I am not sure we actually understand what intelligence is. In particular, does intelligence require a conscious being?

    Well that’s called the hard problem of consciousness. The strong AI position is that there is no difference between general intelligence and consciousness. Many people don’t like that but it is incredably hard to come up with a coherent alternative.

  102. Scott Says:

    Daniel #100: No, I haven’t.

  103. James Cross Says:

    #101 ppnl

    Consciousness is a distinct issue.

    Sometimes intelligence and consciousness are used more or less to mean the same thing, usually when we are referring to human uniqueness, or when we want to make a strong AI argument. But my argument is that intelligence exists independently from and prior to consciousness. Consciousness enhances intelligence, at least in biological entities, but machines which are not conscious can become as intelligent or more intelligent than humans.

  104. fred Says:

    James #103

    As human beings, the one fundamental base fact about reality is that sounds, sights, emotions, pain, pleasure, rational thoughts, irrational thoughts, dreams, sense of volition, sense of self, the ego,… all appear as a modification of a space we call consciousness. Once all conceptualization is stripped down, we’re left with an irreducible sense of “knowing”, as things are arising, which is what it is to “exist” in the present moment.

    Truly understanding the origin of this is outside the bounds of science (even describing it with words feels beyond the bounds of language, at least for me :P) – reductionism (i.e. breaking everything into relations between concepts) can maybe predict when I’m feeling the color red, but not why there’s such a thing as “feeling” in the first place.

    If not for consciousness, why would we even care about whatever happens to us and others? (like the difference between a book describing the holocaust in every possible minute detail vs actual people living it).
    It is only because we think other people experience existence just like we do that we feel compassion for them.
    So consciousness is truly the most precious thing in the universe.

    But what really matters is figuring which systems are conscious, not why they are conscious, and maybe advanced AIs will help with this, since they could be able to self-modify in order to “probe” their own experience of what it’s like to exist.

  105. Adam Lewis Says:

    There have been about of the order of 10^10 humans alive since the discovery of antibiotics, and each one of us hosts around 4*10^13 bacteria, so the cumulative total of bacteria hosted by humans in the antibiotic era is is of the order of 4*10^23. If Galertner’s 10^-77 figure were correct, none of them would ever have developed resistance.
    Unless, of course, there is a remarkably spiteful intelligent designer out there deliberately creating resistant TB to murder kids in the poorest places and torturing hospital patients with MRSA. in which case, I’ve no intention of praying to him/her.

  106. Scott Says:

    Adam Lewis #105: Gelernter would say that taking some strain that happens to be antibiotic-resistant, and making it wildly more common at the expense of other strains, is just “microevolution” of the sort he has no objection to—that the objection is only to “macroevolution” that invents completely new proteins. Of course, the Darwinian response is that “macroevolution” is nothing more than “microevolution” continued for orders of magnitude more time.

  107. fred Says:

    Sorry for the off-topic question regarding NP stuff.
    But is it the case (and/or obvious) that 3SAT instances that map onto a tree (*) are in P?

    (*) i.e. draw each individual var as a node, then draw each clause as a new node with 3 edges to its variables, and then the resulting graph has no cycle.

  108. Scott Says:

    fred #107: Yes, I believe treelike SAT instances are in P, via dynamic programming.

  109. JimV Says:

    2nd and I hope last reply to Wes Hansen, re: “The rest of your comment is just one long false equivalence. If design is equivalent to Darwinian evolution, then why Directed Evolution ”

    I gave a three-part algorithm which I believe applies to biological evolution, human design (technological evolution), and quite well to your example of “Directed Evolution”. (There is no reason the algorithm can’t be nested so that creatures created by evolution use evolution.) To show a false equivalence, you need to show how the algorithm does not apply to human design (whereas your own example shows how it does apply, as do the examples I gave).

    I would hope that anyone who has done complex technical work would have plenty of examples from their own experience. Or recall how you taught yourself to play tic-tac-toe.
    As for art, I liked the quote from Solzhenitsyn’s “The First Circle” when the painter is asked how he knows a painting is finished, something like, “A painting is never finished, but I stop working on it when I notice I am making it worse instead of better.”

    I did not discuss the mechanisms used to perform the algorithm, other than giving a few examples. The mechanisms vary a lot from case to case. I hope that is not what is giving you difficulty. As you probably know, Turing showed that a computer with a one-bit processor (and a lot of memory) can do any calculation that a super-computer can, it will just take a lot longer. A bubble-sort, which I first did in Fortran on a mainframe for my Fortran 101 class, is the same algorithm if done with pencil and paper, or an abacus. An algorithm can be implemented more efficiently with better mechanisms, but it is still the same algorithm.

    As for explanatory power, the god-hypothesis or Intelligent-Designer-hypothesis raises more questions than it answers, as it postulates an entity for which there is no reliable direct evidence (where is it hiding?), who uses powers unknown to physics (how does work?), for reasons unknown to mortals (why does it do things?). There is no who, what, how, where, or when explained by that “explanation”. It really is just an excuse for not having an explanation, which I like to call the “God ate my homework” excuse.

    If there were some magic about human design which could be extrapolated to such an entity … but there isn’t. We use the same algorithm as the one which explains biological evolution, therefore confirming it not refuting it.

    My thoughts on the so-called “problem of consciousness” have bored readers of this blog before, so I will resist pursuing that digression.

    P.S. to Scott: upon further research I found this after googling “the cat who invented Lexan”:

    “The legend, which was apparently invented for advertising purposes, is that a cat knocked over a lab beaker … The more mundane truth appears to be that a fellow researcher was the culprit who knocked over the beaker.”

    So your skepticism was well-founded, and I was guilty of believing a false advertisement, for which I am sorry.

  110. jonas Says:

    So it turns out that in the blog post I linked, David defined two different computational models. The second one is the simple anthropomorphic version, where you can destroy the current universe, and another universe will be chosen with different values for random choices. The first one is more tricky, because it lets you destroy the current universe partially, only starting from a particular time, so that all random choices after that time will be rerolled, but random choices before that time are kept (at least until you choose to redo the universe from an earlier time). I’m not sure, but it’s possible that this second one can do more powerful computations in polynomial time, because David’s argument works for it to prove that it goes all the way up the polynomial hierarchy.

  111. STEM Caveman dissents Says:

    New material coming out on Epstein (much of it hilarious) continues to support the “moral panic” theory.

    Shocking revelations include hiring a Florida high school girl to massage him, who didn’t wanna sex, so he repurposed her by paying her a 200 USD finder’s fee for every new masseuse she brought in (advertising to these new employees in advance that sex was part of the deal). The procurer-girl was pretty happy talking about all this to the FBI, almost boasting, and seems not to have been aware of legal or moral problems in what she and her friends were doing with Epstein.

    Epstein has serious legal troubles, and I don’t expect the media to take his side or rationalize the behavior by pointing to the age of consent in the Netherlands. But it strikes me as an equally serious meta-scandal that news media with readership in the millions can push the narrative for 15 years that Epstein is literally a Serial Child Molester Rapist, when he is merely a rich dweeb who struck reasonable and comfortable deals (in the eyes of both parties) for sex with teenage girls.

  112. Lou Scheffer Says:

    Here’s another way biology proceeds faster than Gelernter’s estimate – neutral (cryptic) mutations accumulate, enlarging the range and speed of adaptive mutations when they are needed.
    See “Tales from the crypt(ic)” in Science

  113. Laurence Says:

    Two points:
    1. When dealing with people like Gelernter, I always remember what Richard Dawkins wrote in “The Selfish Gene”. In his end-notes to Chapter 4 where he was criticising Fred Hoyle and Chandra Wickramasinghe’s theory about influenza viruses coming from space, he wrote “Publishers should correct the misapprehension that a scholar’s distinction in one field implies authority in another. And as long as that misapprehension exists, distinguished scholars should resist the temptation to abuse it.”

    2. While abiogenesis is still an unsolved problem, Nick Lane (professor of biochemistry at University College, London) has come up with an interesting idea in his book “The Vital Question” (2015) where he suggested that life originated as a result of natural proton gradients, such as at vents in the sea-bed (white smokers) that emit alkaline fluids into the acidic early sea. It certainly seemed to me that thinking about life from the viewpoint of it pumping protons across membranes (which is a common characteristic of all cellular life on Earth) is a useful corrective to the concentration on nucleic acids and their evolution.

  114. Wes Hansen Says:

    JimV #109

    I didn’t know until just now that you had replied or I would’ve responded sooner! The difference between neo-Darwinism and Intelligent Design is really in the word “directed;” that’s the point I was trying to make! Find me one neo-Darwinist who admits teleology, I challenge you; I believe, on the other hand, that pretty much every Intelligent Design proponent accepts teleology and certainly teleology plays a huge role in both “directed evolution” and human design. But teleology is not equivalent to Intelligent Design in that there are evolutionists, John Stewart being a good example, who accept teleology but not ID; in other words, ID implies teleology but teleology does not imply ID.

    “As for explanatory power, the god-hypothesis or Intelligent-Designer-hypothesis raises more questions than it answers, . . .”

    Does it really now? Then why do neo-Darwinist struggle so hard to suppress and discredit so much empirical data, while Intelligent Design proponents such as Rupert Sheldrake do the exact opposite? There are many open questions! No one has answered the main question posed by Schrodinger in his book What is Life?; that being, “How does form maintain under the onslaught of molecular chaos?” Can neo-Darwinism address the questions posed by the accelerated learning experiments with rats? And are these experiments not most likely related to the Flynn Effect seen in humans? From the rat link:

    “Consider the following experiment. Animals of an inbred strain are placed under conditions in which they learn to respond to a given stimulus in a characteristic way. They are then made to repeat this pattern of behaviour many times. Ex hypothesi, the new behavioural field will be reinforced by morphic resonance, which will not only cause the behaviour of the trained animals to become increasingly habitual, but will also affect, although less specifically, any similar animal exposed to a similar stimulus: the larger the number of animals in the past that have learned the task, the easier it should be for subsequent similar animals to learn it. Therefore in an experiment of this type it should be possible to observe a progressive increase in the rate of learning not only in animals descended from trained ancestors, but also in genetically similar animals descended from untrained ancestors. This prediction differs from that of the Lamarckian theory, according to which only the descendants of trained animals should learn quicker. And on the conventional theory, there should be no increase in the rate of learning of the descendants of untrained or trained animals.

    To summarize: an increased rate of learning in successive generations of both trained and untrained lines would support the hypothesis of formative causation; an increase only in trained lines, the Lamarckian theory; and an increase in neither, the orthodox theory.

    Tests of this type have in fact already been performed. The results support the hypothesis of formative causation.

    The original experiment was started by William McDougall at Harvard in 1920, in the hope of providing a thorough test of the possibility of Lamarckian inheritance.


    One of these critics, F.A.E. Crew, of Edinburgh University, repeated McDougall’s experiment with rats derived from the same inbred strain, using a tank of similar design.


    Fortunately, this was not the end of the story. W. E. Agar and his colleagues at Melbourne University carried out the experiment again, using methods that did not suffer from the disadvantages of Crew’s. Over a period of 20 years, they measured the rates of learning of trained and untrained lines for 50 successive generations. In agreement with McDougall, they found that there was a marked tendency for rats of the trained line to learn more quickly in subsequent generations. But exactly the same tendency was also found in the untrained line.”

    Agar et. al. lost their controls! And, upon reflection, it turned out that McDougall had too! The original papers are referenced in the link. But most people probably are not even aware of these experiments, experiments conducted from 1920 to 1954 with incredible and inexplicable results; I wonder why that is!?!

    “If there were some magic about human design which could be extrapolated to such an entity … but there isn’t.”

    What makes you so certain?

  115. Nathanael Says:

    Pinker is a poseur whose lingusitics is almost as bad as Chomsky’s. He’s guilty of peddling fake science, repeatedly. I don’t know why linguistics attracts cranks who don’t actually learn multiple languages; both of them have made linguistics assertions which were proven false before they made them, but won’t retract them.

    That said, there is no evidence of him being any sort of sex criminal, so I assume he isn’t!

    Unlike Michael Shermer, where there is clear and convincing evidence that he’s a rapist, which you can look up. He still has all his jobs.

  116. araybold Says:

    Scott #83 (re: Raoul #78):
    The passage you quote is not (at least now) in the linked article or its comments, but if it was made by an academic of some standing, rather than the journalist or some arbitrary comment-poster, then so much the worse. Then again, there are other academics quoted here with statements that are not much better.

    This article, however, was presented in this forum specifically for providing a little more information about the opinion Pinker provided in Epstein’s 2007 case. This matter does raise some issues, both general and specific, and it is important not to invoke hindsight. The article does not say exactly what was asked, but it seems you would likely deduce that there is a defendant accused of luring minors for the purpose of sexual abuse, and you might reasonably be leery of getting involved.

    On the other hand, you do not know if it involves a child abuser trying to get off on a technicality, or someone being pressured by a dubious charge to accept an equally dubious plea deal. I may be wrong in this, but I suspect that if you asked the lawyers for more details in any such situation, they would demur on sound legal grounds, and if the reading of the law seems clear to you, saying so might seem the right thing to do — after all, trials have judges who are supposed to see through tendentious arguments from either side.

    So I do not see this as an automatic negative for Pinker, though he has the opportunity to clarify things further.

    I am in no danger of being called upon as an expert in anything, but those who are may want to carefully consider how they feel about participating if they do not have the full picture. If in doubt, however, leaning towards more truth and facts is probably a sound policy.

  117. Anon Says:

    “(which we recently learned probably include the rape, and not only statutorily, of a 15-year-old)”

    I followed the link you provided, read the text and watched the video and didn’t see anything that would classify as non-statutory rape. Statutory yes, since the law arbitrarily defines consent away. If you’re willing to redefine consent to suit your moral judgment, you can always call any conduct consensual or nonconsensual so it fits your judgment – regardless of whether it was objectively voluntary for the involved individuals.

    I seems we are at a point where rape has been redefined to include perfectly voluntary sexual interactions. Plus, feminists have called for my real-world rape because I defended a fictional video game with rape in it (RapeDay). What reason is there now left for me to care about the real-world rape of girls or women? If all sex is rape, and I don’t have human rights according to the moralists who control the Overton Window, why should I still give a shit about rape?

    I’m happy that Epstein had fun before he died, and in fact I’m not even convinced he was acting irrationally. You can go through all of your life being bullied out of your sexuality, you can agree with society’s judgment of your character even while you strictly respect everybody’s right to say no or exit the relationship, you can accept that you’re either asexual or a villain. OR you can simply acknowledge that morality is arbitrary, everybody is a hypocrite, crucial concepts like consent are redefined on the fly, and use pragmatic methods to get what you want in life. This is what Epstein did and all I feel is envy that he pulled it off so nicely.

  118. Mike Says:

    Anon #117

    Some feminists said mean things to you on the internet when you trolled them, so now rape is permissible. Quite a rational approach to morality.

  119. Anon Says:

    Mike, I didn’t troll them. I defended a fictional rape game against censorship because it was victimless. I went out of my way to make clear that I support the right of every person not to be raped. In return, they called for my real-world rape. This was not the only time my human rights have been attacked and denied by all kinds of self-proclaimed moralists.

    Consensual sex is being redefined as rape. Even if I go out of my way to distinguish between consensual and nonconsensual conduct, people will just redefine it on the fly. The grotesque construct of “statutory rape” is the most striking example. But there are others too. One commentor claimed that all prostitution is rape “by definition”, because there is not love involved.

    I don’t know all the details of the Epstein case, but I do know I can’t trust the media reports. When they talk about “rape”, “abuse”, “molestation”, “trafficking” etc., they may refer to nonconsensual conduct such as kidnappings, threats, forced sex, but most likely they’ll just use the same language for voluntary sexual transactions. One thing is clear though: The government’s violent attack on Epstein himself was actually nonconsensual.

    I’m okay with raping females at this point. Or children for that matter. I could lie about it, but it would be a lie.

  120. More on my Quillette critique of David Gelernter « Why Evolution Is True Says:

    […] hadn’t been trained in evolutionary biology or even biology (neither were Darwin and Mendel). In a post on his own website, Shtetl-Optimized, computer scientist Scott Aaronson does note the falling-away of Gelernter from […]