Thursday, June 05, 2008

Learning to Be Thoughtless

Social norms are unwritten laws. Did you ever go into a grocery store and took somebody else's full cart? Went into a restaurant and ate with your fingers? Went straight to the head of the queue? Went to work wrapped into nothing but a towel? Probably not. And very likely, you haven't spend very much time thinking about it either.

That the effort people make thinking about social norms is less the better established these norms are, is the result of this quite amusing and interesting article

(If you don't have access to the journal, try this draft, which is pretty much similar to the journal version, except that the figures are better.)

It uses a particularly simple model in which `individuals' are represented by simple agents sitting on a circle who chose what to do in a situation (A or B, eat with the fingers or use the fork, wear a towel or a dress). They have a certain search-radius r, that is the number of steps around them in which they observe what the others do. To chose a behavior, in each time-step they do the following: Check out what the neighbors are up to within the search radius, and figure out what the majority does. Then increase search radius to r+1. If the result changes, increase search radius by another step and so on, until it either covers all others or the result does no longer change. Otherwise, if increasing the search radius doesn't result in a change, try decreasing it by one step. If that yields the same result, update search radius to the lower value and figure out whether further decreasing is possible. Adjust behavior to that of the majority within search radius.

It is further assumed that there is some kind of randomness in the system, in that occasionally somebody will just feel like ignoring what the social norm is.

The result of running this model isn't so surprising: If a social norm is already well established, the search radius will decrease into a state where everybody just `knows' what to do. If the system is perturbed by some random shock everybody tries to find out what to do now, and after a while things will settle down in a possibly different, but equally `thoughtless' state.

So, here's a social norm I never understood. Why are letters of recommendation confidential? So confidential, not even the person who the letter is about gets to see them? Having read several letters about myself, I can certify it doesn't kill you. Having read piles of letters about others I still fail to see what the secrecy is good for. The only explanation I can come up with is it's a social norm. As far as I am concerned, it's bad enough there's so much sociology involved in academic hires, somewhat more transparency would certainly be beneficial.

Anyway, what do we learn from this? Question the norms every now and then. It keeps your society mentally flexible and avoids its settling into a thoughtless state.

Related: On the Emergence of Lies


  1. You are describing an extra potential use of "Inverse Square law" as it has never before been seen?:)

    If this is to be the case, then you had to see the relationship of GR in ways that had never been thought possible?:)

    Also, you had to know there is a ethical standard about privacy issues that each and every citizen is guaranteed under the freedom and of information, and held, under the Canadian Constitution of Right and Freedoms.

    Employers have these records and you have a right to them to inspect for inaccuracies introduced without your knowledge.

  2. The reason that they should be confidential is because otherwise most letter-writers would temper their criticism too much, and make the whole exercise a less valuable source of information for prospective employers.

    Thanks for the link to the paper though. I too have been very interested in understanding why the reason behind certain cultural norms (of which I include accounting rules and even legal rules as a subset) are forgotten over long periods of time. For example, the 15th Century Venetians had an excellent understanding of the reasons for having a patent system, which 21st Century Silicon Valley seems to have forgotten.

    You might enjoy my blog --

    I work on patents, but I'm generally interested also introducing simple dynamical models (such as coupled oscillators) into the mental toolkit for economists.

  3. otherwise most letter-writers would temper their criticism too much

    That's a social norm as well. Criticism isn't personal, we all have to live with evaluations of our work. There's ways to do it, and ways not to. What good it is to criticise if the person who is criticised doesn't know about it? How are you supposed to learn about your strengths and weaknesses?

  4. Hi Plato,

    That wasn't the point. It's not that I want so urgently to see my own letters, I just think this letter-collecting business has in some areas grown out of proportion. If so much attention is paid to letters, it signals to me that people in hiring committees don't have the courage to judge on their own (or don't have the time, or maybe are just not able to, who knows). What does it say if a researcher doesn't have a single letter from a really important person in his or her field? Well, it very possibly says he hasn't met many of them. What does this say about his promises? Nothing. I think if this process was more transparent it would taste less like an auction.



  5. Bee,

    I completely agree that it's not the most efficient cultural norm from the point of view of encouraging learning, especially by the prospective employee.

    The problem is that there are other competing values that the cultural norm of letter writing has to meet, especially the avoidance of violent conflicts. Although in an ideal world, everybody would be ready to say negative criticism to the criticised person's face, in the world we live in that is seldom the case. People are afraid of retribution -- or even simply of being proven wrong.

    As to the practical question you've raised -- about how the prospective employee is supposed to learn from job to job -- I would say that the process becomes something like the review of journal results, right? Nobody publishes negative results -- you have to infer them.

    It seems like we share a more idealistic view of how things might be different though, so I'll just share this thought as well.

    Wouldn't it be great to have a journal in each field devoted entirely to negative results? It would de-stigmatize the publication culturally, and avoid a lot of wasted time -- especially in the softer disciplines of economics, sociology, etc.

    I'm quite confident that more has been learned in science from negative results than positive. Why not publish some?

    Too much time and effort is wasted by people worried about others' perceptions of their work. Science is too important to be hindered by those kinds of concerns, no?

  6. I like the idea of having a journal devoted to negative results. It is true that one learns a lot out of failed approaches. I always find it stupid if everybody has to do the same thing to arrive at the conclusion that it doesn't work. It would also erase a lot of pseudoscience if people could look up why their great idea doesn't make sense.

    I think 'violent conflicts' is quite an exaggeration. To begin with, the letters that I've read are already entirely without negative words. The task is mostly about figuring out the fine shades of praise, which again can differ from one writer to the next. One thing e.g. that always strikes me is that letters from Germans are typically way more reserved in praise than those from US Americans unless of course the writer knows about this and 'Americanizes' when appropriate, no kidding. What I am trying to say is that I find this psychological discussion of how what X said about A compares to what Y said about B, given that X is from C, and Y has never been in D and so probably doesn't know F who however has worked with X and knows B is a student of Y and so on not only a waste of time but also a destraction from what the focus should be on, namely the candidate and his work. I'd think if everybody knows what others typically write this would lead very fast to a clearer outcome of the lettering process. There will of course always been that kind of information that is traded privately anyhow. The question is how much weight that should have.

  7. Q and A's are usually drafted for a reason?:)

  8. Way cool article. Explains some of why it's so hard to change a social norm, even when most people agree it's a bad idea.

    In addition to what riemannzeta said, I've always felt the confidentiality of letters also had a class component. A letter of recommendation is a credential furnished by one person to another about the quality of a third person who is generally lower in the hierarchy, and it isn't for them to know what their betters say about them. The lack of transparency reinforces the power associated with ones position in the hierarchy.

    These days it seems emailing references and having a conversation would be easier and make it easier to adjust for norms of praise.

  9. Bee,

    The fact that physical violence seems like an exaggeration may in fact be an example of the way that social norms tend to weaken as the reasons for them start to fade from memory and are not rearticulated.

    Some of the people who sailed to the colonies in the U.S. (and many of their ancestors) had been subject to physical violence for the expression of their negative views about others. The memory of that violence was no doubt part of the reason that these people demanded that freedom of speech and religion were embedded into the U.S. constitution.

    Nowadays many people in the U.S. tend to think of the 1st amendment as a positive right -- i.e., that I get to say whatever I want. The Founding Fathers understood it more as a negative right -- i.e., that I have to listen to you even if you're wrong.

    But I can't dispute what you say about actual letters. I would suggest that the data shows that many writers do not trust the purported anonymity of the process. Perhaps they are all aware of the same thing that plato pointed out in an earlier comment -- namely, that the employees can later see the letters by asking for their file.

    The way to change this kind of cultural norm is to demonstrate with an experiment how including constructive negative criticism in letters is actually valuable for everybody over the long-term. But by definition, that experiment is going to be long-term and large-scale. Perhaps that's why nobody has figured out how to get out of the equilibrium.

    As an aside, I am pleased to see that the Germans are not as prone to the same error. I believe the Germans must have something embedded in their culture that teaches the value of listening to criticism. Negative feedback always carries more useful information than positive -- which, almost by definition is already reflected in the behavior of an individual or institution.

  10. Hi Plato

    Q and A's are usually drafted for a reason?

    If this is supposed to say I did not answer a question you were asking, then the reason is that I did not notice there was a question addressed at me. You have the habit of ending statements with a question mark, which leaves me wondering a) is this a question? b) if so, is it a rhetorical question or does it require an answer? and c) at whom is it addressed? And that leaves asides the problem of figuring out what you are saying to begin with, an interpretational higher order complication for which I usually don't have the time. What I'm trying to say is: if you want me to answer a question, I would greatly appreciate if you could use plain stupid Bee-speech.



  11. Hi Rillian,

    Explains some of why it's so hard to change a social norm, even when most people agree it's a bad idea.

    Yeah, unfortunately they didn't add a recipe how to kick people out of their thoughtlessness.

    The lack of transparency reinforces the power associated with ones position in the hierarchy.

    Right. Another reason to get rid of it. This is science, classes aren't supposed to matter, social hierarchies distort resoning. There is of course a correlation between experience and ranking in the hierarchy, but it's the experience that matter's, not the ranking, and a correlation isn't an identity. Best,


  12. Hi RiemannĪ¶,

    One would think the conclusion should then be that open criticism is worthwhile. Yes, I too guess that part of this is due to cultural norms (see e.g. here). It seems to me that feedback you get from Germans when asked is often more honest, which Americans will easily consider offending. E.g. if somebody asks me how their talk was, it might very well be I'll tell them I found the reasoning very confused or the slides not insightful or whatever. If the speaker is American, I'll go tell them the talk was great no matter what because otherwise I'll be the unfriendly grumpy German and so on. Either way, this probably isn't going to change any time soon, and though it's a topic I always find interesting it wasn't so much what I was aiming at here.

    Consider for example you had a chance to find out what a referee wrote in his letters over the years. Not only would you know how he generally describes people (which is also good to know when you are looking for a referee), but you would also know whether he e.g. generally thinks all his students are totally great, or how well his judgement about somebody's promise turned out to be. In short, it would be much clearer what to pay attention to, without having to rely on a second set of persons who know the referee and will offer their interpretation based on their knowledge of his, upon which you then have to take into account their relation to the referee and so on. See what I mean? It's a multi-layered social game that is inappropriate to the subject.

    Perhaps that's why nobody has figured out how to get out of the equilibrium.

    Courage I'd say. The conclusion of that paper is based on the meta-norm that people pay attention to what their neighbors do to begin with. A bit more wiggle room, a bit more trying around, and possibly we'd have a chance to find a different equilibrium. Best,


  13. Bee, you are not a representative sample of the subject of reference letters. By no means. You are brilliant, skilled, well-read, you know people high up; you are even attractive. No wonder you get good letters!

    I am sure the letters you got to read did not kill you, besides: there is a clear selection bias there. You probably did not manage to put your nosy hands on those very few that might have been less than exstatic ;-)


  14. Ha, the Italian Casanova drops in! Tommaso, I assure you none of the letters I came across contained the words brilliant or well-read, and I know more people deep down than high up. Besides this, what's important would be those people know me not the other way round. In fact, some of what I read gave me something to think about. Anyway, it's my nose that's nosy, my hands are handy ;-) Best,


  15. Bee,

    A question mark always supposes a place from which one may embark?:)

    Q and A's are usually drafted for a reason

    A way, to arrive at a definition, and explanation of such letters and how they are written.

    It was meant to suggest, that while such negatives are meet, they are also meant to keeping open to the full use of any suggestions. Keep on Bee-ing.

    A way I guess, in which to look at negatives while still encouraging one to move ahead.

    So, criteria had already been set, for any judgements to ensue. These may have been thought in regard to such Q and A's which are devised in this case to see if one did indeed fit into such criteria.

    So where are you headed?:)


  16. Hi Bee,

    Another nicely written post on an interesting and yet seldom thought about subject, no pun intended (well perhaps a little). What I particularly liked was the related paper since I have to admit I haven’t explored or contemplated the social graces from the scientific viewpoint often.

    What stood out for me in this paper more then anything else was the effects of the introduction of noise (mixed behavior) as it relates to the pattern formed in social norms. As the noise was introduced such as in figure 4, the pattern formed changed from one that was lineal and regular to one that takes on what appears to have more of a pattern corresponding to chaotic complexity, with even a island being formed, as in a strange attractor. When the noise was further increased the fractal aspect becomes more divergent (figure 5). Then in the final figure the system is maximally randomized and system converts again to random.

    For me I see that in some aspects this could equated with social integration, like experienced when there is immigration from other countries, bringing about the introduction and integration of different social norms and behaviors. I might be all wet in thinking this, yet the islands formed with the lower levels of noise could relate to how ethnic groups tend to form tighter social bonds and neighborhoods among themselves at the exclusion of others corresponding with very little thinking (yet fractal). As the integration level (noise) increased a bit more, this in turn produces even greater clumping (still fractal),with the related thought levels increasing yet becoming more random and divergent. When the level was turned to maximum with no recognized pattern of norm the thought process increased again and went totally random. The first two (4 and 5) might serve to be what’s recognized as segregation and ghettoization.

    I would have liked to see what would occur if both runs 4 and 5 would have been extended with more interations, that is as to represent time. That is I’m thinking as this would serve to simulate world wide integration over extended time. Of course all this most likely indicates I’m trying to hard to read the tea leaves:-) None the less it is interesting to imagine if this could apply.



  17. I disagree that transparency is necessarily a good thing.

    The one reference about me I have ever seen was extremely embarrassing. I had some problems after being read the reference about me which was to be sent to three universities when I was 18 and wanting to read physics (under the UCCAs application scheme at that time you could apply for a choice of three). It was written by Father Gilheney, the Headmaster of Salesian School in Chertsey. He met me and read it out before sending it off. It was the only time I ever met him (I was at that school from 1982-90, eight years, and he was Headmaster for all that time). I think he met every student applying for college and read them the reference, and usually they were very pleased to hear what he wrote.

    However, the reference about me seemed entirely inaccurate, and was terribly embarrassing when afterwards I attended an interview at the one university which seemed impressed enough to call me for an interview. I think he got his information from those teachers who had never really spoken to me at all for more than five seconds. ("Nigel is very quiet and lacks confidence..." for example, when in fact I was only quiet in classes because it was pointless to argue or ask for more evidence, because no teacher had any time to spend on digressions. I've always been confident and assertive - not quiet - whenever there is something that's really seems to be worth saying.)

    The reference praised supposed attributes of mine which were entirely imaginary (politeness/retardedness for instance, when the only reason I wasn't outspoken was because it wouldn't achieve anything positive in that situation), while it ignored my real strengths (which obviously weren't apparent in class, where I was trying to learn, rather than to display impressive characteristics).

    It felt embarrassing to even try to press him to edit it during that five minute meeting. I think he made a mistake in reading it out to me, and automatically expecting me to concur. I later felt that I had to explain to the university admissions tutor why the reference was totally wrong. Doing so was actually a disaster because it then turned out that that particular admissions tutor (the only one who wanted me to attend an interview) had been highly impressed with the (misleading) reference, and was less than happy to hear me reviewing errors in the supposedly confidential reference!

    So in some ways it's probably better if references are kept confidential, just to avoid depressing arguments and corrections. Suppose that you write a reference and the person it concerns disagrees with it. How far should you then edit the reference to appease that person? If you can't do so (either for reasons of ethical principle or time constraints when you have to produce many references), it may cause some embarrassment and unhappiness. Ignorance can be a happier condition, in some cases.

  18. Hi Nige,

    I am really sorry to hear about your unpleasant experience. However, it makes me wonder whether the situation would have occurred this way weren't it that the letter had been thought to be confidential. Possibly, the person writing it meant to do you a favor, or maybe was confused or just generally not particularly good in his assessment of other persons. In the former case, he might have made more effort to provide a clearer profile. In the latter cases, knowing how accurate his previous letters were might have helped the person reading it in knowing how seriously to take it.

    See, at least you had a change to address the incorrectness. It might sound cynical, so my apologies, but wouldn't it have been worse had you been hired based on an inappropriate letter? I don't think a letter should be 'corrected', but one has to understand what it contains. It is an expression of an opinion, not a psychological profile or an absolute statement of qualification. If you know what it contains, you at least have the possibility to address criticism, or explain were the opinion might have come from, and/or whether you intend to improve in the criticised areas etc. Best,



COMMENTS ON THIS BLOG ARE PERMANENTLY CLOSED. You can join the discussion on Patreon.

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