Sequence index: Living Luminously
Previously in sequence: City of Lights

You can use luminosity to help you effectively change yourself into someone you'd more like to be.  Accomplish this by fixing your self-tests so they get good results.

You may find your understanding of this post significantly improved if you read the seventh story from Seven Shiny Stories.

When you have coherent models of yourself, it only makes good empirical sense to put them to the test.

Thing is, when you run a test on yourself, you know what test you're running, and what data would support which hypothesis.  All that and you're the subject generating the data, too.  It's kind of hard to have good scientific controls around this sort of experiment.

Luckily, it turns out that for this purpose they're unnecessary!  Remember, you're not just trying to determine what's going on in a static part of yourself.  You're also evaluating and changing the things you repudiate when you can.  You don't just have the chance to let knowledge of your self-observation nudge your behavior - you can outright rig your tests.

Suppose that your model of yourself predicts that you will do something you don't think you should do - for instance, suppose it predicts that you will yell at your cousin the next time she drops by and tracks mud on your carpet, or something, and you think you ought not to yell.  Well, you can falsify that model which says you'll yell by not yelling: clearly, if you do not yell at her, then you cannot be accurately described by any model that predicts that you'll yell.  By refraining from yelling you push the nearest accurate model towards something like "may yell if not careful to think before speaking" or "used to yell, but has since grown past that".  And if you'd rather be accurately described by one of those models than by the "yells" model... you can not yell.

(Note, of course, that falsifying the model "yells" by silently picking up your cousin and defenestrating her is not an improvement.  You want to replace the disliked model with a more likable one.  If it turns out that you cannot do that - if controlling your scream means that you itch so badly to fling your cousin out a window that you're likely to actually do it - then you should postpone your model falsification until a later time.)

Now, of course figuring out how to not yell (let us not forget akrasia, after all) will be easier once you have an understanding of what would make you do it in the first place.  Armed with that, you can determine how to control your circumstances to prevent yelling-triggers from manifesting themselves.  Or, you can attempt the more difficult but more stable psychic surgery that interrupts the process from circumstance to behavior.

Sadly, I can't be as specific as would be ideal here because so much depends on the exact habits of your brain as opposed to any other brains, including mine.  You may need to go through various strategies before you hit on one that works for you to change what you need to change.  You could find that successful strategies eventually "wear off" and need replacing and their edifices rebuilding.  You might find listening to what other people do helpful (post techniques below!) - or you might not.


New Comment
7 comments, sorted by Click to highlight new comments since: Today at 3:03 AM

Cross-posted from Seven Shiny Stories

7. Experiment

Eva bursts into tears whenever she has a hard problem to deal with, like a stressful project at work or above-average levels of social drama amongst her friends. This is, of course, completely unproductive - in fact, in the case of drama, it worsens things - and Eva wants to stop it. First, she has to figure out why it happens. Are the tears caused by sadness? It turns out not - she can be brought to tears even by things that don't make her sad. The latest project from work was exciting and a great opportunity and it still made her cry. After a little work sorting through lists of things that make her cry, Eva concludes that it's linked to how much pressure she feels to solve the problem: for instance, if she's part of a team that's assigned a project, she's less likely to react this way than if she's operating solo, and if her friends embroiled in drama turn to her for help, she'll wind up tearful more often than if she's just a spectator with no special responsibility. Now she needs to set herself up not to cry. She decides to do this by making sure she has social support in her endeavors: if the boss gives her an assignment, she says to the next employee over, "I should be able to handle this, but if I need help, can I count on you?" That way, she can think of the task as something that isn't entirely on her. When next social drama rears its head, Eva reconceptualizes her part in the solution as finding and voicing the group's existing consensus, rather than personally creating a novel way to make everything better. While this new approach reduces the incidence of stress tears, it doesn't disassemble the underlying architecture that causes the tendency in the first place. That's more complicated to address: Eva spends some time thinking about why responsibility is such an emotional thing for her, and looks for ways to duplicate the sense of support she feels when she has help in situations where she doesn't. Eventually, it is not much of a risk that Eva will cry if presented with a problem to solve.

Helpful link to a completely harmless source of information for those who don't get the reference: lampshade hanging.

"Completely harmless" is not an accurate description of TV Tropes Wiki - many people have noted that it can become an immense time sink. In my opinion, this is due to the large amount of material being available and heavily crosslinked - and given high average and moderate variation in data attractiveness, this produces a variable-ratio schedule of positive reinforcement, leading to addictive behavior much like that observed in gambling.

That last link is the only safe one, by the way - it goes to, as opposed to TV Tropes or xkcd, and is very brief.

It's 2 PM and I still haven't done any work today. Thanks. :(

Was that supposed to be funny? is pretty bad - it's full of hard-to-spot sponsored ad links and is a mess when it comes to readability.

All true, but is pretty anti-addictive, unlike Wikipedia.

When I was young, I went to psychologist because I was afraid of the dark. She made me find the origin of my fear, so i enlisted something like: recent burglary in our apartment, war, my imagination, some TV series, and said that all those mixed together can make you be afraid of the dark at night (10 years old kid can do that, yes, though I am not sure if I repeated what i heard from others). So, not after seeing that there is nothing in my backyard, but after seeing the origin of my fear, I stopped.

That is why I think that finding the origin of some behavior is important when trying to stop it. For example, you can find out that your self-respect is causing your immunity to being insulted. So, if you want to be offended by something, you might want to undermine your self-respect first. You can as well imitate, and try to get really angry when someone offends you, but you will possibly feel fake, and eventually, you will go back to your pre-change-attempt behavior.

Your behavior can be learned, as well. If your father was a person-type that always steps aside, you can be like him. In this case, you will have to learn the other behavior, by observing other people doing it, and trying to imitate them (just like you learned from your parents?). These are my personal experiences, though, and I would like to hear some expert's opinion on this topic.

The last thing I want to point out is, be aware that you can make mistakes. You want to fix something in yourself, but you can make a mistake and try to fix something that you don't really want to fix. Like, you see that people who are devious have more success than you do, but this does not mean that you should become devious, for example, it can make you unhappy. (I am not stating anything about being devious, it's example).

New to LessWrong?