About Rationally Speaking

Rationally Speaking is a blog maintained by Prof. Massimo Pigliucci, a philosopher at the City University of New York. The blog reflects the Enlightenment figure Marquis de Condorcet's idea of what a public intellectual (yes, we know, that's such a bad word) ought to be: someone who devotes himself to "the tracking down of prejudices in the hiding places where priests, the schools, the government, and all long-established institutions had gathered and protected them." You're welcome. Please notice that the contents of this blog can be reprinted under the standard Creative Commons license.

Thursday, May 16, 2013

On signalling

by Ian Pollock

[Note: neither character in this dialogue represents any particular person, apart from the “people” who are always arguing about philosophy in my own head. I acknowledge that this conversation is totally unrealistic as a real-life event, for several reasons. (Some of which are signalling reasons!)]

Salviati: Hello, Sagredo!

Sagredo: Hi, Salviati. Good to see you. I’m just finishing up a blog post.

Salviati: On what?

Sagredo: It’s about this creationist, biblical literalist theory to the effect that there was once a water canopy around the earth. The idea is to account for where all the water in Noah’s flood came from.

Salviati: I take it you disapprove of the theory.

Sagredo: Of course! I don’t understand how anyone could possibly believe it.

Salviati: ...Hm, hang on a minute. I want to press you on something.

Sagredo: What is it?

Salviati: Well, when people say they can’t understand so-and-so’s behaviour, they can mean two things. Either they mean that they literally don’t comprehend the behaviour, or they mean they understand it in an intellectual sense, but are trying to morally distance themselves from it.

Sagredo: I think I mean that I literally don’t comprehend people who believe in stuff like this.

Salviati: Then you are at a disadvantage in writing your blog post, aren’t you - I mean, you should understand your opponents’ motivation, if possible. Anyway, what’s your argument?

Sagredo: A scientific takedown of the theory, followed by a critique of its motivations as being basically designed to reduce cognitive dissonance. Creationists know that all the water in the biblical flood needed to come from somewhere, so they invented this wild theory to account for it.

Salviati: Wait, so you do understand people who believe this stuff! They’re just reducing cognitive dissonance.

Sagredo: I guess, in a sense, yes.

Salviati: Why are you writing this post in the first place?

Sagredo: To combat this crazy pseudoscientific drivel.

Salviati: Do you know people who take this water canopy theory seriously?

Sagredo: Well, not personally, but some creationists buy into it.

Salviati: Some? Meaning it is critiqued by the creationists themselves?

Sagredo: Yes, to a certain extent.

Salviati: ...Okay, do you mind hearing some criticism?

Sagredo: I guess not. I’ll try to take it constructively.

Salviati: All right, I am just going to be blunt, then. This blog post you are writing is pure signalling.

Sagredo: Oh, not this “signalling” crap. Since when are you Robin Hanson?

Salviati: So you’re familiar with this line of argument.

Sagredo: I’ve heard it a lot from various “cynics.” But go ahead and make your case, I guess.

Salviati: All right. Here is the situation as I see it. You belong to an intellectual community of skeptics and atheists, in which creationism is the standard example of terrible pseudoscience.

Sagredo: Well, it is, and lots of people believe in it. The cost we are paying for science illiteracy is huge.

Salviati: Sure, I mostly agree, but that is another discussion. The point is, you belong to a group where you can gain status pretty easily by beating up on creationism - do you deny this?

Sagredo: I don’t deny it, but that’s not my motivation in writing this post.

Salviati: To be totally and suicidally frank, I don’t believe you. Here’s why: this blog post is focused on an incredibly tiny problem, and it’s actually pretty much useless in terms of dealing with that problem. I think you know that, in your heart of hearts.

Sagredo: Science illiteracy is not a tiny problem!

Salviati: But this particular crazy theory is. As far as I understand from you, only a small minority of creationists believe in it. Granting that combating creationism is important, combating an unpopular idea among creationists just seems like a really weird goal. It’s like you’re fighting a war and you decide to concentrate on intercepting the enemy’s supply of brussels sprouts - that’ll break their spirit!

Sagredo: It’s part of a general pushback against pseudoscience. You have to do that kind of stuff piecemeal.

Salviati: Sure, sure. The other thing is that even if we assume that it’s really important to debunk the water canopy theory, I’m going to take a wild stab in the dark and say it’s already been done.

Sagredo: It has, but the message needs to be repeated in order to get through.

Salviati: Right, but there are diminishing returns here. That’s what I meant when I said your post is probably close to marginally useless. Being a doctor is uncontroversially important, right? Doctors save lives. But becoming a doctor only saves marginal lives if there’s a shortage of doctors in the area where you’re going to work - otherwise, you’re just replacing or supplementing some other person who was a doctor anyway, and who would have saved those lives anyway. That’s a roundabout way of saying that I think your post is one of many debunkings of the canopy theory, so it’s not making much difference above and beyond what the others are doing. And apart from that, as I said, the canopy theory just seems like a really tiny problem in the first place.

Sagredo: Well, so what? I suppose this isn’t the most important problem in the world, and I may not be combating it in the optimal way. I guess I just find the subject interesting. Didn’t you write some sort of gobbledigook about metaethics last week? Don’t tell me that solves some pressing problem of humanity. I still don’t buy the idea that my post is just “signalling.” I promise you, getting praise from fellow skeptics wasn’t going through my head at all when I thought of the idea.

Salviati: First of all, when I said that your post was pure signalling, I wasn’t accusing you of something unusually terrible. I think that a huge part of the stuff all people say and do is signalling - it’s in our nature as social creatures. That applies to myself as well: to the extent that I can understand my motivations for writing, some of them seem to be aiming at a sort of intellectual showing off. I just respect you and I want you to be a bit more self-aware about your own motivations. Like me, you obsess over logical fallacies and cognitive biases, but you seem to be blind to these other psychological forces that are driving your own behaviour.

Sagredo: I really wasn’t thinking about getting praise when I decided to write the post. Honestly!

Salviati: I both believe you and do not believe you. I believe you in that I don’t think you explicitly thought “let’s beat up on some dumb creationists to get validation from the in-group.” As in, I don’t think those words passed through your stream of consciousness. But I do think it’s very likely that the unconscious expectation of praise and status points was a major factor in what made this idea seem like a good one to you.

Sagredo: That’s totally unfalsifiable!

Salviati: It is invisible, true, even to your own introspection. But we can infer it from your behaviour. Again, I don’t want to beat up on you too much, but when somebody chooses to devote several hours of time and energy to “solving” an incredibly tiny problem in an incredibly ineffective way, and you know that they are smart enough to be able to realize this, you start looking for other explanations for their behaviour. That’s why I look at all the people writing reams of stuff about how Bigfoot doesn’t exist and can’t help but shake my head. Signalling fits nicely as an explanation, though. As an analogy... if a person says that something is 90% probable, and then it doesn’t happen, they could just be unlucky. But if they keep saying “90% probable” and these things keep not happening, you can infer that the person is likely overconfident.

Sagredo: When I say it’s unfalsifiable, I don’t just mean that you’re inferring something about my psychology. Being psychoanalyzed is really annoying, but sometimes it’s possible to do that. I mean that there is literally no way of proving it wrong. As for how that relates to this water canopy thing...

Salviati: Maybe we should stop talking about your blog post, so that the argument’s conclusion doesn’t involve any embarrassment for you personally. No offense, but that would lower the emotional stakes for you.

Sagredo: Yeah, ok, I am kind of feeling like an insect pinned in a display case right now. Well, one of Hanson’s examples I remember is charity. He says “charity is not about charity” and stuff like that. The idea is that people give to such ineffective charities, in such ineffective ways, that the whole point of charity must not be to help people or causes, it must just be to look good.

Salviati: I’m with you so far.

Sagredo: But then you have this whole movement for effective charity centered around Givewell and Giving What We Can and other organizations. What about the people donating to them? And then the Hansonian reply (apart from other criticisms of the ways people donate) is that these people are just signalling to more sophisticated in-groups. Do you see how that looks a lot like epicycles?

Salviati: Sure, superficially.

Sagredo: But that’s not the worst of it! The “everything is signalling” crowd still needs to explain altruistic acts by individuals who are totally anonymous and know that they are.

Salviati: Right. One idea here is that the people who are best at showing off in front of groups are not the people who are most explicit about showing off, but instead the people who are totally oblivious to the fact that they’re showing off, and totally convinced that their behaviour is motivated by other, good reasons. So donating anonymously to charity is a way to, as it were, convince yourself of your own altruism, in order to signal more effectively at others in other circumstances.

Sagredo: Wow, seriously? ...You might as well just be a Freudian at this point.

Salviati: Well, as a human, given that you are going to be deceptive, a really effective way to do it is to actually believe the deception (at least at the conscious level). It’s cognitively cheaper than trying to be some straight up Machiavellian plotter who believes A but says B - you don’t have to watch your tongue or your body language if you “believe” your own lies. And if the “beliefs” involved don’t come in contact with reality too often, it doesn’t hurt your interests very much if they’re technically false. But I concede that that explanation is a bit suspicious-looking and seemingly unfalsifiable. By the way, another theory that occurs to me for explaining anonymous donations is that they often aren’t very anonymous after all. Maybe anonymous donors usually do tell key people in their in-group and trust the word to spread. That might be worth investigating empirically.

Sagredo: What is the big appeal of this theory, anyway? Why are you seeing signals behind every bush like some phallus-obsessed Freudian?

Salviati: I’ve noticed that tendency myself, yes. But I honestly do think that you will just fail to understand people if you don’t take signalling motives into account all the time. Fail miserably. And you will also be selectively blind to your own motivations. We could talk about all sorts of behaviours, from table manners to military strategy... Here’s a throwaway example. Why does everybody talk about crime going up and being out of control, when we both know that in North America and the UK the crime rate has been going down for a while?

Sagredo: They’re misinformed by politicians and members of the media, who are themselves either misinformed or looking for votes.

Salviati: Okay, but why do you think things are always skewed in the direction of high crime as opposed to low crime? Well, look, here’s what we can say from the perspective of signalling. A signal is basically an action (often a speech-act) A that also carries information about some hidden variable B. In the case of politics, the “hidden variable” typically refers to inferred personal characteristics of the speaker, characteristics that look good or bad to their in-group.

Sagredo: How does this apply to the crime issue?

Salviati: So let’s say you’re a member of the public who doesn’t know very much about crime rates. Some high-profile shooting happens in your city’s downtown, and Politician #1 comes on TV and says “Crime is out of control! We must do something about shootings like this! We need more police, more censorship of video games, parents need to teach traditional values...” you know the usual social-conservative spiel. On the left, it would focus more on gun control and other causes celebres, but otherwise it’d be pretty similar. Then Politician #2 comes on TV and says “Actually, crime is pretty much under control, and getting more so every year. This act was terrible, but our laws are necessarily a balance between freedom and security, not to mention between security and scarce resources that can be used elsewhere. I don’t think we should take any additional steps in the wake of this shooting, current laws are adequate.”

Sagredo: Yeah, point taken, I see how the first politician looks much better as a person. The second politician sounds complacent, even callous. And that’s before somebody comes up with a devastating meme about their indifference to human life, referencing the recent shooting as emotional ammunition.

Salviati: It can be even worse in extreme cases! How would you like to argue the case against some new anti-paedophilia law? In the wake of a high-profile case whose details have been luridly reported by the media? It’s an especially horrifying self-perpetuating phenomenon because even people who agree with you will wonder, deep down, what kind of person you are if you’re willing to, ahem, stick your neck out for paedophiles. Free speech advocates run into the same problem, because in practice, they end up defending people like Holocaust deniers, or those tiresome artists who take a crap on a crucifix and put it in a gallery, or anti-Islamic types who are often motivated by xenophobia more than principle. Signalling incentives are horrible, so unless corrected, society tends to drift in whatever direction it’s easiest to moral-grandstand in favour of.

Sagredo: Do you have some solution for this problem?

Salviati: Not a comprehensive one. But at least we can raise people’s consciousness to the problem, so that the pundit who goes on TV and says something like “If this law saves even one child from a horrible fate, it will be worth any cost whatsoever!” gets shouted down for cheap, empty signalling. But on an individual level, it’s just another piece of the puzzle for understanding how the social world works. That’s why I decided to harass you about it, because I think you’d want to know more about how people tick, including yourself. There are a lot of other details to it, and more botanizing that you can do about signalling incentives, harmful signalling arms races, and application to other stuff besides politics, like philosophy or education. This is just the elevator pitch.

Sagredo: I guess I see your point; I can imagine how this idea would apply to lots of stuff. But I still think there is a huge danger of over-application and unfalsifiability.

Salviati: I agree with you there - it’s a powerful idea, but it can definitely eat your brain if you’re not careful.


  1. So if I comment in favor of this article I am signalling that I already knew about this, that I am in the "in-group" that knows about human irrationality. And if I comment against this article I am signalling that my knowledge of human irrationality is such that I don't get affected by signaling any more.
    Therefore my comment about it is...

    I like bacon.

    1. Way to conspicuously signal your irreverent wit & internet meme-savvy. ;)

  2. I'd like to signal that I really enjoyed the dialogue.

    Even if it's difficult to know whether signaling motivations are at play in one's own actions, the idea of rigorously thinking about our motivations is a good one. Even if Sagredo doesn't end up agreeing with the idea that his motivations have to do with signaling, he can at least come to the conclusion that his water canopy post isn't a good use of his time and energy.

    1. I agree. Probably we should choose our battles; there's no point in overanalyzing your motivation for getting vanilla vs chocolate ice cream. But if we're talking about a fairly consequential act (more than a few hours of time or money), a little explicit reasoning that includes considering your motivations is a good idea.

      Also, the key thing in my books is not *whether* you're motivated by signalling (you may even want to USE it as a motivation - that's why people brag about their exercise programs on facebook) - it's whether, once you remove the signalling motivation, your act is in some sense vacuous or very suboptimal.

  3. Signaling seems like an interesting idea to me. My guess, however, is that it's a useful metaphor but limited. Like Psychoanalysis, it's probably correct in the broadest possible sense, for instance, that we are not aware of all of our motives, we are self deceptive for good reasons, etc. But here are some other overlapping motives that could account for the same types of behaviors. 1) laziness, it's easier to make an argument that you've made before against people whose likely response you've already encountered 2) familiarity 3) low risk behavior. Attacking fringe nincompoops has little downside to it, should you be wrong. Compare that to more substantial efforts like fighting Global Warming, ending world hunger. If done wrong, those loftier pursuits could have bad unintended effects.

    I'm not sure it's signaling, but it was interesting at NECSS to see Michael Shermer start his talk with graphic pictures of female genital mutilation. This rabble rousing imagery and attitude may have been more about trying to win an argument than trying to present one.

    1. >Attacking fringe nincompoops has little downside to it, should you be wrong. Compare that to more substantial efforts like fighting Global Warming, ending world hunger. If done wrong, those loftier pursuits could have bad unintended effects.

      That's true, but I think the fact that the [small effect, large effect] and [low risk, high risk] axes are lining up here is an artefact of the examples chosen. There are plenty of slam dunks out there for advocacy that promise a higher return on investment. For example, effective charity.

      >This rabble rousing imagery and attitude may have been more about trying to win an argument than trying to present one.

      Yes, I think what's happening here is that he's positioning himself as the anti-FGM person so that his interlocutor, in order to argue the point, will have to appear to be "pro."

  4. A lot of signals is generated by "members of the media" from many times very little information.

    That seems to be true, but maybe hard to quantify.

  5. Great stuff. I'd note that the "Freud!" reaction is not only somewhat common, it's basically anti-scientific. You don't have to be a Freudian to think that unconscious motives play a large determining role in human behavior: just ask nearly any practicing cognitive scientist. No phallus-theories there, just an inference to the best explanation of observable human phenomena (i.e. major mismatches between people's self-reported motives and their actual behavior).

    Another huge virtue of this post is that you avoid demonizing signalling: such a pervasive feature of our social psychology is unlikely to be some random destabilizing force. It is very a likely an evolved psychological tendency which has been harnessed by culture for all kinds of good.

    1. I'm not sure I completely agree that Ian avoids demonizing signaling. What about when Salviati says "Signalling incentives are horrible"? (I take it that Salviati speaks for Ian, as he did for Galileo.)

      And I'm not certain that signaling qua signaling is horrible. (Perhaps that makes me Simplicio!) It seems to me that in at least some circumstances, signaling could be a good thing, because it could lead individuals to coordinate their efforts towards some good end.

      Let's say I see a group of people engaging in signaling behavior. I am more likely to join them, magnifying their efforts. That could be a good or a bad thing, as I see it. To determine which, we must ask this: why am I more likely to join them? What is my motivation?

      I can think of two possible motivations. The first is simply that I want to be part of the group. In that case, I might wind up joining just for the sake of joining, without really thinking about the effects of my behavior. That would be bad; in this case, signaling has led to group-think. The behavior being signaled could have awful consequences, but I would have no idea because I don't really care about that. I just want to be part of the group.

      The second motivation is that I genuinely believe that the group's behavior is directed toward a good end. I've thought about it carefully, and I think their activities are worthwhile; their goals are my goals, and I want to join them. Furthermore, their goals cannot be achieved by autonomous individuals; they can only be achieved by coordinated activity.

      The difficulty here is that the second motivation implies the first. If I genuinely believe that a group's goals are my goals, and if I can't achieve my goals without joining the group, then I want to join the group. Furthermore, I want others to join the group, and I would be working against my goals if I were to silently join.

      Think of a boycott. If I were to silently stop buying widgets from the Malevolent Widget Company (MWC), I might do a modicum of good. But there might be hundreds of people -- thousands of people -- millions of people -- who have no idea that there's anything bad about MWC widgets. They might not realize that M stands for "malevolent," and if they do, they might just think it was a joke. My personal refusal to buy MWC widgets won't change any of MWC's practices; as far as they're concerned, I could be refusing to buy their widgets because I don't like the color. And anyway, they aren't even paying attention to my purchasing habits; they're paying attention to broad statistical patterns. As long as a large segment of the population thinks that MWC widgets are perfectly fine and dandy, they won't change a bit.

      But if I start signaling -- if I start telling people about how bad MWC's widgets really are, and talking about how I'm refusing to buy them, and encouraging others to do the same -- and if they follow suit -- then I create the possibility that MWC might notice us and do something about it.

      Now perhaps this doesn't really count as signaling. We have two different behaviors: signaling1, based on the first motivation (i.e. joining a group and publicizing your membership just to be part of the group.), and signaling2, based on the second (joining a group and publicizing your membership because you believe the group does good). Perhaps what people mean by "signaling" is really signaling1, and not signaling2 at all.

      But making that distinction seems hard -- because (as I argued above) signaling2 automatically entails signaling1! And furthermore sometimes an instance of signaling may be a genuine instance of signaling2, but is based on erroneous logic, confusion, or ignorance. How do we distinguish that from signaling1? So my feeling is that both these behaviors ought to count as signaling; and so sometimes signaling is good, and sometimes it's bad. We just have to do our best to distinguish between the two.

    2. >I'm not sure I completely agree that Ian avoids demonizing signaling. What about when Salviati says "Signalling incentives are horrible"? (I take it that Salviati speaks for Ian, as he did for Galileo.) And I'm not certain that signaling qua signaling is horrible. (Perhaps that makes me Simplicio!)

      Salviati speaks for one of my mental subagents; Sagredo for the other. Hence my position is ambivalent.

      However, no, I do not want to condemn signalling. On the contrary, I think it's just how we work as social creatures. But I think that it's very easy to fall into signalling spirals that are deeply unhelpful (although some such spirals are useful).

      My comment about "horrible signalling spirals" refers to such a case - the crime example.

      I agree with the rest of your analysis in your comment, just wanted to clarify my position.

  6. My dog has intimate knowledge of my signalling. He gets wildly excited when I put on my running shoes.
    And I have just signalled again.

  7. My dog has intimate knowledge of my signalling behaviour. He gets wildly excited when I put on my running shoes.
    And I have just signalled again.

  8. And then there is turkey bacon, possibly flaunted as a signal to the Less Porcine crowd. Thoughts? Can you tell this is not a fun AM commute?

  9. As Salviati recognizes, signalling applies especially to areas like politics, (some) philosophy and education – I guess because values are centrally involved. I find ideological discussion increasingly depressing as it is so often so patently about posturing and positioning and feeling good about oneself.

    I know one has to talk about values to some extent, but this was one area, I think, in which Wittgenstein was wise (e.g. in rejecting philosophical ethics). And Quine. And Descartes. And a whole bunch of philosophers in fact.

    It is one of the beauties and consolations of serious scientifically (or mathematically) oriented discussion that it usually only gets affected by this sort of thing at the margins.

    In fact, this little dialogue is making a clean, clear, scientific point.

    1. It is indeed depressing, yes.

      One of the great things about philosophy is that the culture is such that one does not typically receive much censure for entertaining unpopular (or popular) ideas.

  10. I'm not sure I totally catch what angle or angles the article has. To kind of riff on Philip, above, another term for signalling, when deliberately done by some politicians or social groups via the media, is "dog whistling." (Don't know if the term is similar in Britain or not, Ian.)

    And, given that a lot of dog whistling here in the US hinges on racial issues, it's not a minor matter there, no more than the issue of willful scientific illiteracy transcends Bigfoot and water canopies to fighting against stem-cell research or mitigating/combating global warming/climate change.

    That said, here in the US, free speech includes the right to dog whistle. It still, on paper, includes the right to report news even when based on leaks by members of the government, even though George Bush and Alberto Gonzales don't think so.

    >>What? They're not in office any more? Oh, sorry.<<

    1. I think dog whistling is a special case, but yes, an illustrative one.

      My angle is something along the lines of "cultivate the ability to distinguish posturing from substance in the words and actions of yourself and others."

      The impetus is my observation that this is not so easy, especially as regards oneself and one's in-group. We are apt to take our posturing for serious philosophizing.

    2. Oh, I'll agree on your last graf. I identify myself as a left-liberal (by American lights), but that I'm skeptical about claims of likelihood of success of left-liberalism, especially on certain economic amelioration.

  11. I think you could make an evolutionary argument for why signaling is a likely motive behind many costly acts. If a costly act doesn't provide any direct benefit to the actor or to his/her ingroup, then the only other way such an act could be selected for is if it signalled some desirable quality. Any motivation to perform effortful acts that did not confer direct benefits or signal some desirable quality would have been selected out. Granted, not every behavior stems from some kind of adaptation, but for any given behavior, it's a reasonable first guess.

  12. Not that signalling isn't a real thing, but most of the time when people say I don't understand how could someone believe in this nonsense, even when they say they literally can't understand why someone could believe that nonsense, they are simply expressing their indignation. For some people, insulting other people is pure fun.


Note: Only a member of this blog may post a comment.