Incinerating the Sherlock Holmes Gambit

Is a bad explanation better than no explanation?

Cranks, when faced with issues where science has yet to find a solid answer, often appeal to a classic quote from the literary character Sherlock Holmes that has appeared in different forms in many of his adventures:

When you have eliminated all which is impossible, then whatever remains, however improbable, must be the truth.

This claim, which could be called the Sherlock Holmes gambit has been used in many different contexts:

Sherlock

  • Greg Cochran uses it indirectly to argue for the existence of a virus that turns men gay because he finds explanations based on genes, hormones and selection unpersuasive. This has led race realist JayMan to conclude that the germ hypothesis is “almost certainly” correct despite the fact that no clear evidence exists for that idea. The pathogen has not been identified and no clinical or epidemiological evidence has been presented.
  • John Lennox uses it to argue that the resurrection of Jesus was so improbable that it just had to be true. This line of argument has been forcefully refuted by Richard Carrier.
  • All conspiracy theorists everywhere use it when talking about Malaysia Airlines flight MH370.

However, this argument is flawed in at least four separate ways:

(1) Constitutes a false dilemma

Attacking alternative explanations is not evidence for the proposed explanation. Faulty criticisms against evolution is not evidence for creationism and misguided arguments against quantum mechanics is not arguments for new age mysticism. That would qualify as the fallacy known as false dilemma.

(2) No evidence for the improbable explanation

Nowhere in this gambit is evidence presented for this improbable explanation. It is an evidence-free argument. In fact, it tacitly admits that the prior probability of the claim being made is exceedingly low.

(3) Ignores unknown explanations

Science does not know everything right now. It might never know everything there is to know about the world. Therefore, this gambit ignores the possibility of future explanations. Just because we cannot explain a novel card trick right now does not mean that supernatural powers were involved and not being able to explain engineering anomalies does not mean that a plane was shot down.

(4) Assumes correct understanding of current alternative explanations

The Sherlock Holmes gambit usually include an extremely superficial treatment of alternative explanations. Misunderstandings of evolution, genetics, engineering and other topics are common. Thus, these alternative explanations have not really been excluded as “impossible”. Even if those were currently “impossible”, future discoveries or elucidations might make them more scientifically credible.



Categories: Skepticism

Tags: , , ,

21 replies

  1. The thing about the Sherlock Holmes Gambit is it requires a level of rigour not usually applied by those who try call on it. Most failures are with the condition, “When you have eliminated all which is impossible”. Take, for example, your first example. Not being persuaded by the explanations does not make them impossible. Unless it could be demonstrated conclusively that genetics, epigenetics, and selection can not account for homosexuality then they are all possible explanations. The same holds for Jesus, unless you can prove it impossible that the Jesus story is mythological or solely based on exaggerated real events, you can not progress to more improbable explanations.

    On the conclusion, “whatever remains, however improbable, must be the truth.”, the failure is that what remains must have some evidence for it otherwise you’re left with a situation where a mess of contradictory explanations all “must be the truth”. So in the case homosexuality, there is no evidence whatsoever for a viral cause, there is no evidence at all it is a curse by fairies, there is no evidence at all it is caused by cosmic rays. It’s not possible to claim any of these must be true on the basis another explanation has been deemed impossible.

  2. “Greg Cochran uses it indirectly to argue for the existence of a virus that turns men gay because he finds explanations based on genes, hormones and selection persuasive.” Don’t you mean “unpersuasive”?

  3. The problem is two fold. For one, most don’t understand the meaning of the word “evidence.” Second, few people have a decent grasp of evolutionary theory.

    (Of course, I’d say there’s one problem underlying those, but…)

    If you did understand evolutionary theory, the matter would be obvious, especially once explained to you.

    You would at least gather what the case is here:

    Unless it could be demonstrated conclusively that genetics, epigenetics, and selection can not account for homosexuality then they are all possible explanations.

    Selection: selection eliminates low-fitness phenotypes
    Genetics: low heritability as gathered from twin studies.
    Epigenetics: No evidence of any such effect to begin with, and can’t explain prevalence of such a low-fitness phenotype.

    • Hey JayMan,

      Can you provide any credible scientific evidence for the existence of a gay-causing virus? When I say “evidence” here, I do not mean arguments against other models because arguments against model A is not an argument for model B.

      By “evidence” I mean things like

      (a) large-scale epidemiological associations between a viral pathogen (which one?) and gay male phenotype,

      (b) fulfillment of standard clinical criteria for showing a probable causative relationship between a pathogen and a condition,

      (c) coherence between lab and epidemiological results,

      (d) an empirically plausible mechanism of how this pathogen leads to men being gay,

      (e) a dose-response relationship between the pathogen and the gay male phenotype

      etc.

      Notice that you cannot retreat to uncertainty because you clearly stated that you think that the “gay germ” is “almost certainly” real (source). A core skeptical principle is that the strength of one’s convictions should be in proportion to the evidence. Since the strength of your conviction is very high, then you must think that there exists pretty convincing scientific evidence for the existence of this “gay germ”.

      If you cannot present such evidence, your belief has to be considered irrational and without scientific merit.

    • @Emil Karlsson:

      First of all, if my comments are being moderated, I’m not interested in engaging in this discussion. It’s already not of much substance (however, if the problem is WordPress, please let me know, thanks).

      If you cannot present any such evidence, your belief has to be considered irrational and without scientific merit.

      LOL! You’ve got to be kidding, right?

      Let’s get to the chase. What’s the alternative? The range of broad possible alternatives is rather limited. Can you find one that could potentially work? That’s what eliminate the impossible means. I have to be honest, if you do believe in any of the previously assertive alternatives, you really undercut your stance of being to judge the “scientific merit” of things.

      No, we haven’t found the pathogen itself, but it’s not unexpected that no direct observation of the pathogen has been made (not for the least reason that no one has looked).

      an empirically plausible mechanism of how this pathogen leads to men being gay,

      Not Final! | West Hunter

      You should read that whole post.

      Since the strength of your conviction is very high, then you must think that there exists pretty convincing scientific evidence for the existence of this “gay germ”.

      Since you’re pretty strong in your being unconvinced of the case Cochran and Ewald have made for the gay germ, you should be able to tell me – with your understanding of evolutionary theory and biology – what potential other explanations could exist. Even one will suffice.

    • In my previous comment, I asked JayMan to provide scientific evidence of the existence of a gay-causing virus. Since he has an incredibly strong conviction in the existence of this germ, he must be sitting on pretty strong evidence, right?

      I was disappointed to see that the response JayMan wrote to my challenge did not contain any scientific evidence whatsoever. He did not present any large-scale epidemiological research connecting a viral pathogen to the gay male phenotype. He was not able to point to any research that fulfill standard clinical criteria for showing probable causative relationship between a pathogen and being a gay male. No empirically plausible mechanism was proposed for how a viral infection could cause lasting changes in emotional, romantic and sexual attraction in humans and no dose-response relationship was shown.

      Instead, he continued to promote the false dilemma fallacy: arguments against model A, he argues, is evidence for model B. However, this is not the case. A model stands and falls on its own scientific merit, not by the merits (or lack thereof) of a narrow range of alternatives. As I expected, JayMan retreated to uncertainty by rationalizing the lack of scientific research into this alleged gay-causing pathogen. However, this means that his strong belief in the existence of a gay germ is not in proportion to the evidence available.

      JayMan did link to a blog post written by Gregory Cochran. After reading it carefully, I can say that it does not reference any scientific evidence in favor of the existence of a gay-causing virus.

      In conclusion, I requested that JayMan produce scientific evidence in favor of the existence of a gay-causing virus such as large-scale epidemiological research, fulfillment of standard clinical criteria for a probable causative relationship between a pathogen and being a gay male, an empirically plausible mechanism how a viral infection could cause lasting changes in emotional, romantic and sexual attraction in humans, or a dose-response relationship. He could not provide any of it. Despite this, he persists in his strong conviction. This means that his belief is not in proportion to the evidence and that his position is demonstrably irrational.

      As a last-ditch case, JayMan reinvigorates his false dilemma approach by demanding that I provide an alternative model. However, as pointed out ad nauseam, arguments against model A are not arguments for model B. My skepticism against the “case” made by Cochran is indeed strong, and that is justified by the lack of substantive scientific evidence.

      (My Akismet spam filter has been malfunctioning for several months, retaining real comments and publishing spam ones after I had an influx of very angry anti-psychiatry proponents, so I turned on “Comment author must have a previously approved comment”. However, this does not seem to work that well, since I have to manually approve all comments at the moment, regardless if a poster has a previously approved comment or not. I will try to figure out why…)

      (Edit: I have now discovered that my comment blacklist includes the name “Jay” from when I banned a troll with the email beginning jaytownes82. Since WordPress matches inside words, “Jay” also blocks “JayMan”. I have now removed “Jay”, so this issue should hopefully be fixed. If history is any guide, there are probably more WordPress issues that can get in the way, but I will do my best to fix those as well should they arise. I will also make sure to fish up and post any comments from you should they get stuck in the spam filter again)

  4. Thanks for fixing the comment issue (hopefully that does the trick).

    Perhaps this might help, as I said on Twitter:

    That’s all. Three broad possibilities. (Of course, option D is magic. But if you have to appeal to magic to falsify this…)

    First, obligate male homosexuality has very low evolutionary fitness. Lack of sexual attraction to the opposite sex is going to put a man at a distinct disadvantage evolutionarily (especially when you consider that genetic evidence shows that upwards of 40% of males failed to reproduce, and in modern times, that rate was anywhere from 10-25%; competition was fierce). Even a small fitness hit on average iterated over time will lead to the extinction of the alleles that cause the trait.

    In any case, Option A, genetics, is ruled out by the low heritability. There’s simply no way to reconcile a straight genetic explanation when identical twins are discordant 4 out of 5 times.

    Option B, prenatal environmental is ruled out this way: very low fitness phenotypes are all rare; the rarity is proportional to the fitness hit incurred. Just as people with malformed genitalia, inability to speak, or who are born blind or deaf (sans pathogenic involvement) are very rare, natural selection serves to ensure that any biological failure that will seriously negatively impact fitness doesn’t often occur. Yet, anywhere from 3-5% of all males are gay. Even fairly common genetic ailments (e.g., schizophrenia, autism) occur at most at the ~1% frequency.

    Option C, the postnatal environment, is all that remains. Very few agents fit this bill. It has to be something that doesn’t leave much by way of visible damage to body (save some possible visible brain damage as seen in autopsies). In addition, homosexual men do not seem to suffer from reduced IQ or other disabilities, so whatever damage that occurs must be very specific. The agent in question has to be fairly common, yet somehow able to resist selective pressures preventing this kind of damage from occurring. As well, it’s not something we see in hunter-gatherers, indicating it’s not endemic to humanity. All of these point to one agent: a pathogen.

    I think you’ve been misapplying rational analysis. Direct evidence, while always helpful, is not the be-all end-all when it comes to inferring truth. Logical deduction (well, induction) is important as well. A good skeptic must not only have sound methodology, but must have an excellent grasp of the facts. Hence, as I said in my tweet:

    If you have any questions (as I’d imagine you’d might), please let me know. Just be aware my time is limited, so my response time might be long.

    • In his second response, JayMan seems to tacitly admit that he does not, in fact, have any positive scientific evidence for his position:

      (a) no large-scale epidemiological associations between a viral pathogen (which one?) and gay male phenotype,

      (b) no fulfillment of standard clinical criteria for showing a probable causative relationship between a pathogen and a condition,

      (c) no coherence between lab and epidemiological results,

      (d) no empirically plausible mechanism of how this pathogen leads to men being gay,

      (e) no dose-response relationship between the pathogen and the gay male phenotype.

      Thus, there is little reason to accept his position as scientifically credible. This also demonstrates, beyond a shadow of a doubt, that his conviction in the viral cause of gay male phenotype is not in proportion to the available evidence.

      Instead, JayMan continues his previous line of thought: he starts by claiming that only four explanations exists. His argument is not that those are the currently existing explanations, but that, in principle, those are the only four that can exist. He then goes on to attempt to exclude three of these explanations in isolation. Finally, he concludes that the only remaining explanation (“postnatal environment”) can be true, and the, out of a gigantic range of possible postnatal environmental influences, selects a single one, namely a viral pathogen.

      Unfortunately, this line of argument is flawed in almost every respect:

      – The fact that one can only think of four possible explanations does not mean that these are, in fact, the only possible answers in principle. That would be the fallacy known as argument from incredulity. There could, in theory, be any number of additional explanations. If one wants to argue that these, in principle, are the only four possibilities one shoulders a heavy burden of evidence.

      – It is not enough to exclude alternatives one-by-one because several partial explanations could be operating. That would the be false dilemma fallacy (or in this case a false tetralemma). The classic example here is the creationist claim that regularity, chance or design are the only explanations for, say, the bacterial flagellum. They then go on to exclude regularity alone and chance alone in the naive belief that they have now proven creationism. On the contrary, in this case, regularity and chance would be the relevant explanation.

      – Assuming that the correct explanation is postnatal environment, there needs to be evidence that this postnatal environmental cause is viral. Similarly, the creationist William Lane Craig uses the Kalam cosmological argument to attempt to show that there has to be a cause to the universe. However, there needs to be an argument for why this cause has to be his particular desert deity. He cannot merely assert it to be the case.

      There are also (i) several positive arguments against a viral pathogen being responsible, (ii) multiple errors in the arguments that JayMan makes against the alternative explanations that he discusses and (iii.) it is possible to put forward an alternative sketch of the origins of the gay male phenotype that is supported by at least some empirical evidence that does not involve a viral pathogen.

      However, I would like to postpone those three sub-topics for now until we have come to a consensus on the following two issues:

      – What, if any, scientific evidence (as I have defined earlier) exists for the existence of a viral cause of male gay phenotypes?

      – Is the strength of JayMan’s conviction in proportion to said scientific evidence?

      From my perspective, there seems to be almost no scientific evidence, and JayMan’s conviction is very disproportional to the strength of said evidence. It is much too strong to be justified by the data. Can JayMan comment directly on these two questions?

      (Your comment had two or more links, so it could caught for another reason this time.)

  5. The most honest answer so far is that the scientific consensus is “we don’t know” but it could be a combination or interplay of genetic, hormonal, and environmental factors, rather than one single factor. JayMan is also wrong on the count of it not being endemic to humanity – it’s been around since earliest recorded history. Even if JayMan’s points were all correct however, there would still be zero evidence for it being a virus. You might as well be saying “aliens caused it”.

    • @phytomorph:

      “The most honest answer so far is that the scientific consensus is ‘we don’t know’ but it could be a combination or interplay of genetic, hormonal, and environmental factors”

      I think you should reread my last reply. Slowly.

      “it’s been around since earliest recorded history.”

      A. Because humanity is only as old as recorded history…
      B. That’s not true.
      C. See the Bushmen.

  6. So what “causes” heterosexuality then…?

    Without getting too much on my high horse about the unexamined heteronormativity in this discussion, there are a number of questionable assumptions going on here.

    Firstly, sexual orientation and identity is not an all-or-nothing binary. It is more graduated and fluid than seems to be assumed here: the concept of a fixed homosexual or heterosexual social identity for life is actually a relatively new human idea dating back only to the 19th century, although of course humans and other animals have expressed homosexual as well as heterosexual desire and behaviour for as long as such things have been recorded. Not to mention the various gradations of transgender, which are to be found across most societies, with or without homosexuality. Human sexuality is not a case of either-or – it’s a diverse spectrum whose manifestations appear to be partly inherent in the individual and partly a result of their interactions with the particular society they live in.

    Secondly, what is often presented as “normal” sexuality – lifelong exclusive heterosexual monogamy within a nuclear family – is far from universal in human societies or for that matter among our closest non-human relatives. There are a wide range of mating styles – monogamy, polygyny, polyandry, and/or promiscuity; heterosexual and homosexual bonding; as well as ways of distributing the work of raising offspring and dealing with conflict (including competition for mates) and cooperating in the survival of the group as a whole. It is a fiction as well as a gross oversimplification to assume that heterosexual monogamy is for every individual in every situation the only or even the ideal way of optimising the survival of ones genetic heritage in future generations.

    JayMan’s assumption that homosexuality results in low evolutionary fitness is not necessarily true. If a person contributes to the survival of his or her kin this can just as effectively ensure the long term survival of his genetic heritage as directly trying to reproduce himself. A brother or sister is just as genetically “you” as a son or daughter. And a (blood relative) nephew or niece carries as much of your genetic heritage as your grandson or granddaughter. It is perfectly possible for investment in kin to have as good or better outcomes in long term survival of your genetic heritage than competing for a mate – especially if the outcome of such competition is far from assured. Human conflicts over resources (including mates) can be extremely costly in terms of evolutionary fitness, often lethal, and sometimes catastrophically so for entire societies. Biological mechanisms which potentially reduce such conflicts and promote cohesion within a human group can feasibly result in improved evolutionary fitness for our genes.

    One of the most interesting biological effects on sexual orientation is that of birth order: for each older brother a male has, his chances of having a homosexual rather than heterosexual orientation increase substantially (by around 30% although the studies vary). This effect appears to biological rather than social, and is preserved when younger brothers are adopted into families where they the oldest child. This phenomenon may in fact be consistent with improved evolutionary fitness: after all your older brothers carry much the same genes that you do. In that situation being homosexual could conceivably increase the probability of your genes surviving by decreasing costly competition between brothers for mates while preserving the contribution of all adult members to the wellbeing and ultimate survival of the kinship group as a whole, especially where early mortality is extremely common (as it has been for most of human history).

    The diversity of sexual orientation and behavior is unlikely to have a single cause: there is evidence of moderate effects from biological elements such as genetics and birth order as well as social factors. But I think Jayman’s viral hypothesis unlikely. Aside from the fact that sexual diversity does not appear to distribute in any way like viral infections do, I have trouble conceiving how a virus could have such polymorphous results on sexual orientation and behaviour in so many different species.

    And of course I have a problem with conceiving universal heterosexuality as the incontrovertible norm, and homosexuality as a pathological condition that needs to be explained. But we’re back to the questionable heteronormativity issue…

    • @Snout:

      Do they not teach anyone how to count anymore?

      So what “causes” heterosexuality then…?

      Think really hard…

      Firstly, sexual orientation and identity is not an all-or-nothing binary. It is more graduated and fluid than seems to be assumed here

      Except that the distribution of male sexual orientation, for all we have on the matter, isn’t smooth or even remotely normally distributed. Most men are on one end of the Kinsey scale or (a certain non-trivial minority) on the other end (i.e., the distribution is J-shaped). Few are in between.

      Not to mention the various gradations of transgender, which are to be found across most societies

      First, what does one have to do with the other?

      This goes back to counting. How common are transsexuals? (Hint, less than 0.1%. Indeed, closer to 0.01%.) Such people can be explained by mutational stress. However, gay men? 3-5%. Huge difference.

      JayMan’s assumption that homosexuality results in low evolutionary fitness is not necessarily true. If a person contributes to the survival of his or her kin this can just as effectively ensure the long term survival of his genetic heritage as directly trying to reproduce himself.

      Math: it’s hard.

      One of the most interesting biological effects on sexual orientation is that of birth order: for each older brother a male has, his chances of having a homosexual rather than heterosexual orientation increase substantially (by around 30% although the studies vary).

      Evidence for the birth order effect is questionable. At even if it is real, it would explain only a small fraction of homosexuality.

      I have trouble conceiving how a virus could have such polymorphous results on sexual orientation and behaviour in so many different species.

      Only two: humans and sheep. No other animal has homosexuality like it (i.e., males completely disinterested in females). And these are species that have lived together for thousands of years. Hmmm……

    • There’s a comment from me in spam (feel free to delete this one afterwards).

  7. @JayMan

    A. Because humanity is only as old as recorded history…
    B. That’s not true.
    C. See the Bushmen.

    1. So how can you make any certain comments either way about whether it’s endemic or not in pre-historical humanity?
    2. Yes, sorry, it is true.
    3. Why? A nice cherry pick example?

  8. Funny how JayMan couldn’t answer the questions of Emil. Funnier still, HBD believers have very little evidence for their many speculations (E.G. Genes for ethnic temperament etc.)

  9. “Not to mention the various gradations of transgender, which are to be found across most societies First, what does one have to do with the other?”

    In the context of this discussion, it’s whether genetic predispositions that sometimes result in a non-reproducing phenotype can nonetheless survive and thrive in the face of selection.

    “Evidence for the birth order effect is questionable. At even if it is real, it would explain only a small fraction of homosexuality.”

    As opposed to your viral hypothesis which lacks even questionable positive evidence. And whether birth order might account for 1% or 100% of homosexuality is irrelevant. If it exists at all it refutes your claim that such effects are “impossible”.

    “Only two: humans and sheep. No other animal has homosexuality like it (i.e., males completely disinterested in females).”

    Err, no. Try reading more widely.

    “Math is hard (link)”

    Your math is easy. Too easy. Unfortunately you have applied it to a model of selection that is overly simplistic. In dismissing the birth order and genetic hypotheses as “impossible” you have created a strawman model of selection of behavioural traits.

    1. You seem to be assuming that genotypes map to phenotypes rigidly one to one. This rarely appears to be the case when it comes to the genetic basis of complex behaviour such as sexual preferences – a given genotype can map to multiple very different phenotypes, depending on what other biological and other factors are also operating. And birth order is an example of one such potential biological factor that could act on a particular genotype (or genotypes) to result in a variety of phenotypes. That’s inherent in the hypothesis.

    2. You have not taken into account the effect of scarcity of reproductive opportunities.

    3. You have not taken into account fitness costs involved in males competing for mates.

    4. You also seem to be assuming that the unit of basic unit of evolutionary fitness is the individual. It is not, it is the gene. Individuals are simply the vehicles by which genes transmit over time. It is not the individual phenotype that is passed on – it is the genotype that gives rise to that phenotype – and in the context of behaviour very likely other phenotypes associated with that genotype. The evolutionary fitness of a genotype is the result of the fitness of all its corresponding phenotypes in the relevant setting.

    “Some people have suggested that human homosexuality is an adaptation produced by group selection. I can’t see how this could possibly work. They would have to do stuff for close relatives – lots of stuff. This is a quantitative question: if they concentrated on the closest possible relatives, nephews and nieces, they’d have to cause four more to survive than would otherwise. We’re talking a behavior stronger and more effective than mother love. It doesn’t exist.”

    Your first questionable assumption here is that reproductive opportunities will expand in proportion to the number of males available to take advantage of them. In reality you can have all the available reproductive duties done by a few or even one of a group of closely related males rather than shared among them (as is the case in polygamous human societies). And if they are all likely to be carrying the same genotype (manifesting as different phenotypes according to birth order), then there is no loss of reproductive fitness for the genotype if some of the younger ones carrying it bow out of competing with their older brothers for those finite reproductive resources. Which they may well choose to do in any case, given that older brothers are likely to be bigger and stronger than younger ones at the ages of peak fertility.

    In short, there is no point (in terms of fitness of that gene) with competing with brothers or other kin who carry the same gene – it will be passed on, whoever might win such a competition. However, there is a fitness cost for the gene from such pointless competition in terms of risk of death, serious injury and loss of social cohesion among related males. In this situation a genotype which has a variable phenotype mediated by birth order and which reduces competition without compromising how often the gene makes it to the next generation may actually have a fitness advantage over a different genotype which results in increased fraternal conflict without resulting in increased total reproductive opportunities for the fraternal group.

    Now I am not claiming this is definitely what happens, only that it is just one possible explanation for how a genotype which under the right circumstances results in a gay (or other non-reproducing) phenotype can not only survive selection, but actually prosper. It cannot, at this stage of knowledge, be eliminated as impossible. Which as others have pointed out is the problem with applying the Holmes gambit to science, especially behavioural science. Of course the maxim is logically true (no shit, Sherlock) but practically pretty useless, because science is very good at sorting the probable from the improbable, but poor at definitively identifying the impossible.

    “So what “causes” heterosexuality then…?”
    Think really hard…

    Your non-answer indicates to me that you haven’t understood the question.

  10. I guess that’s about as close to an admission of defeat as is likely with cranks and their pet theories. Bonus irony points for invoking reference to Dunning-Kruger though…well done! By the way JayMan, you can totally use my aliens theory if you want. You could use the alien angle to help explain the lack of any supporting evidence for your virus theory! How about…Alien Engineered Homosexuality Virus from Space! You’ve gotta admit it’s catchy right? I mean, if you’re going to be a crank, I say go for it and c-c-crank it up to 11! Those are always the most entertaining ones…

  11. Oh, so this is why he said (somewhere) that a virus causing homosexuality is the only thing that “makes sense…” I have to admit, this is worse than I thought…

    Amazing post Emil. I’ll have to refer to this in the future as a sort of myth busters toolkit.

%d bloggers like this: