Beware Detached Detail

Yesterday I talked about how "social minds must both make good decisions, and present good images to others" and suggested "the near-far brain division can be handy when facing this problem; let the far system focus more on image, and the near system focus more on decisions." But I didn't follow this thought very far down the game tree. To an economist going down the game tree is like going down the rabbit hole; it shows us just how deep and strange are the underlying drivers of top behavior.

If our far thoughts are more distorted to present good images, then the next step down the rabbit hole is this: to judge how we will typically act, others should prefer to see our near thoughts, at least if they can distinguish near versus far thoughts. After all, near thoughts drive most day to day actions. And we should each look more to our own near thoughts to judge our own sincerity.

Once we evolved to weigh near others' thoughts more heavily, the next step would be to look for cheap ways to have good-looking near-thoughts, without paying the full price of distorting important actions. That is, our mind designer would look for ways to show "detached" near thoughts, consistent with good-image far-thoughts, but not actually impacting much on important near decisions. This could be accomplished by vivid engaging detail that can clearly occupy our near thought systems, but which isn't much connected to substantial personal decisions.

For example, abstract religious beliefs are usually tied to specific concrete, but largely irrelevant, details. The gods and prophets have names and birth-dates and stories of who did what when. Churches have pictures and idols people and events, and religious rituals are full of rich sensory details, like incense, plush materials, stained-glass windows, and angelic singing. Daily routines regarding what to eat and when to pray and what to wear all let you show how much near thought there is in your religious thought.  Of course most of these details hardly matter to your important decisions; they let you assure yourself you will be a good Samaritan, without actually much influencing whether you be a good Samaritan.

Similarly, fiction lets us tie lots of concrete detail to our abstract social beliefs about what sorts of people are good or bad, who helps who how much, and which of them tend to win in the end.  Stories presented as plays, movies, or video games have especially rich sensory detail, to give us experiences not just of abstracting having certain far reactions, but of having our near systems fully engaged and in agreement with our far expectations of what one should do in various situations.

A movie can let you feel not only that people in distant times and places should fight Nazis or free slaves, but that if you were actually in such situations with near systems fully engaged, you would actually do such things. But of course since the movie's scenario has little overlap with your real life, there is little risk that near habits acquired would interfere with your usual near actions. You rarely watch movies about, say, helping poor neighbors or illegal immigrants, since those stories are less detached from your ordinary life.  There is a reason they call it "escapism," after all.

This detached detail dodge may explain why inside views seem worse at schedule forecasting than outside views.  Yes inside-view schedulers consider more detail that engages near mental systems, but I suspect this is mostly detached detail, driven as with fiction detail more by far thought agendas than by consistency with relevant personal detail.  Little girls may imagine their future wedding in vivid detail, without that being much constrained by actual weddings they have seen.

Don't assure yourself that your vivid near thoughts mean you are being more realistic, if those vivid thoughts are little connected with real personal experience. It might just be detached detail, there more to flesh out far thoughts than to advise important decisions. When you stand at a real street corner looking to see if it is safe to cross, that is attached detail; when you imagine being tempted by a grinning ten foot devil after fasting for weeks in the hot red desert, that is detached detail.

GD Star Rating
a WordPress rating system
Tagged as: ,
Trackback URL:
  • Unit

    It seems to me that most near-thought has to do with private property: what’s mine, what’s yours (and when) etc…which, by the way, is the source of most disagreements. Far-thought is necessary to “elevate” oneself from this basic way of thinking, but too often it applies ‘talis qualis’ near-thought patterns to general abstract matters that have nothing to do with private property.

  • http://profile.typekey.com/sentience/ Eliezer Yudkowsky

    I worry that if you’re allowed to postulate “detached detail”, then you’re allowing arbitrary thoughts to be tagged as ideology-fodder regardless of whether they’re processed in fine detail or huge abstractions.

    I would suggest, rather, that while church rituals were invented by people who took their God much more seriously than modern theologians, people who now sit in a church and passively experience the religion, may process it as Far.

    In other words, the author’s experience is not the reader’s experience. (In my own experience, writing a story seems to involve at least ten times as much emotional energy as reading a story.)

    The empirical predictable here would be something like: “People who invent fine details for their religion usually take it more seriously and show a noticeably greater tendency, albeit an imperfect one, to act as if the religion is actually true (even when this would be somewhat costly).” In other words, a breakdown of the ideology-isolator.

  • http://profile.typekey.com/huono_ekonomi/ Mikko

    “[…] our intentions tend to be much more real to us than our actions, and this can lead to a great deal of misunderstandings with other people, to whom our actions tend to be much more real than our intentions.”
    — E.F. Schumacher

    Most religious rituals (incense, rhytmic music, chanting, synchronized movement, boring rational thinking with sermon, vertically positioning priest in high status position relative to churchgoers, architecture etc.) have very practical psychological utility. Their purpose is to help people reach an altered state of mind, to help people have “religious experiences”. Religion is organized mass hypnosis, opiate for the masses. This does not refute your point, but reinforces it: altered state of mind increases the vividness of the stories you tell yourself, it makes them more real.

    There is also an upside for making distant thoughts concrete. All new creative endeavors start by visualizing the end result. In the beginning this vision is hazy, but it gets more concrete as you add detail. More detail increases the likelihood that you will in fact succeed in producing the end result you envisioned.

  • http://causalityrelay.wordpress.com/ Vladimir Nesov

    What is the point of distinguishing abstract and specific thoughts, and postulating corresponding near and far subsystems (even brain regions!), as you originally introduced the divide, if the actual criterion cuts across that and simply states that far thoughts are those less constrained by reality? As I noted before, it looks like distinguishing individual thoughts is enough, with weaker thoughts more easily seduced by emotions, while stronger thoughts more connected to reality. In the context of this post, strong thoughts are more indicative of reality, of more value if you want to learn about reality, and so you’d want to distinguish when other people declare their strong thoughts as opposed to deranged weak thoughts.

    Beware weak thoughts, which in themselves may look just like the strong ones, but there is no reason for them to reflect reality.

  • http://profile.typekey.com/robinhanson/ Robin Hanson

    Eliezer, yes I am introducing a new distinction in this post, which reduces the simplicity of the overall picture. This must pay for itself in enough added understanding, or we should reject it. We must have lots more data on fiction authors than religion inventors, so testing would have to focus there.

    Yes fiction authors elaborate their world view in more detail than fiction consumers, and for that reason would take it somewhat more seriously. And authors using near systems to imagine their worlds have to be and usually are realistic in terms of habits of thought ingrained into their near side systems. E.g., footsteps are spaced like heartbeats in time, and people can see chairs but not bacteria ten feet away. But it still seems to me that a fictional story can be more detached from image-relevant social reality than a true story, such as of your Aunt Frieda’s wedding. Authors seem able to start with far-thoughts about what social reality should be like, and then fill in more details while remaining consistent with that far-vision. The data seem to support this.

    Mikko, I’m saying that adding detail to your plan does make it more realizable if those details are attached enough to the right sort of other ordinary details, but not if you just fill them in mainly via vivid-wish mode. Imagining the location, size, and furniture of all the rooms of your mansion, and your daily routine in passing through those rooms, doesn’t help much in actually getting you a mansion.

    Vladimir, the point of making the near-far distinction is the impressive psych lit showing it really is there.

  • http://causalityrelay.wordpress.com/ Vladimir Nesov

    Robin, near-far correlates with strong-weak divide a great deal, and so for this reason if nothing else experiments support them both. On the other hand, there are examples that support strong-weak but oppose specific-abstract divide, ones you list here, and planning fallacy. You can redefine near-far, reducing its simplicity, to incorporate those cases, but this problem doesn’t seem to affect strong-weak view.

    So my question should read: What are the strong points of near-far compared to strong-weak? It looks like after adjustments near-far becomes strong-weak, you care only about how detached from reality thought content is, whether detailed or abstract.

  • mjgeddes

    SAI_2100: There are three general modes of thought not two: very near, near and far.

    Me: What is the very near mode?

    SAI_2100: Not only do we need a means of thinking about the details of external input for decision making (near) and a means of general social abstraction (far), we also need a means of thinking about our own identity (very near). Nothing is closer than the self.

    Me: And the trade-offs between the near and far modes? The far mode is for social image and the near for accurate decisions?

    SAI_2100: This is too simplistic; its true there are trade-offs in the modes of thought, but they are not totally inconsistent, nor is hypocrisy necessary. It would be more accurate to say that the far mode focuses on ‘social communication’, rather than ‘social image’ per se. Both modes involve effective reasoning. This is shown by the fact that Hanson’s posts themselves are examples of ‘far’ mode thinking.

    Me: But a good post by Hanson showing even the near mode of thinking has its own biases?

    SAI_2100: Indeed. Note Yudkowsky’s unconvincing attempts to counter. Far mode thinking is closely related to analogy formation and is actually a more powerful mode of thought than the near!

    Me: Give me an example of an analogy which effectively illustrates the power of far mode thinking. Make it on-topic.

    SAI_2100: Certainly. Divide reality into three categories: Physical, Social and Mathematical. In the physical domain, note the existence of physical-status uncertainty (Heisenberg uncertainty principle). In the mathematical domain note the existence of truth-status uncertainty (Godel’s theorem). Form an analogy between quantum uncertainty (physical) and Godel uncertainty (mathematical) and then map to the social domain.

    Me: What is this, and how it related to the topic of Hanson’s post?

    SAI_2100: For goodness sake it’s obvious! The aforementioned analogy enables one to derive a new, as yet undiscovered (by humans) form of uncertainty in reflective decision making. Technically, this shows that it is impossible to consistently translate the high-level terminal values of any goal-seeking system with general reflective capabilities to precise lower-level mathematical utility functions and bounded physical actions.

    In a nut-shell, that’s exactly why there evolved three separate brain systems for far, near and very far thinking. Because of irreducible meta-uncertainty about the motives of both yourself and others!

    Me: Wow! Thanks for the insight SAI.

    SAI_2100: A pleasure little human.

  • http://profile.typekey.com/sentience/ Eliezer Yudkowsky

    Robin, would you mind if I start deleting mjgeddes’s comments from your posts too? They seem to be getting a tad long and unreadable.

  • Unit

    After listening to this talk linked by Tyler Cowen, I think I understand this post better,

    http://www.youtube.com/watch?v=R-Di86RqDL4

    Dutton mentions how (wasteful) details can be important in art: “he didn’t miss a note” (for a music performer), or “he didn’t miss a hair” (for a painter). In essence because we value skill. The parallel is with the peacock tail, which actually makes the peacock more vulnerable. The message should then be to the peacock: overcome your tail? (Maybe just: beware of your tail). Also interesting was the difference between natural selection (sperm) and sexual selection (charm).

  • http://profile.typekey.com/huono_ekonomi/ Mikko

    Interestingly, metaphors are often detailed, but they are used to convey abstract ideas.

  • http://SingYourOwnLullaby.blogspot.com mariana

    Metaphors do not always convey abstract ideas, metaphors are analogies.

  • Pingback: Overcoming Bias : BSG is Detached Detail