Wednesday, March 7, 2012

Hypnosis techniques for teaching rationality

One of the big reasons I got into this hypnosis thing was to help improve people that aren't "broken". As a fully general mind hacking tool set, it should be helpful in getting people to think better too, as long as we know how we should reason and decide and what we tend to do wrong. There are more applications than this, but here are a few important ones I haven't really covered yet.

1) Knowing about how to think in the abstract isn't all that helpful if you don't have a habit of actually using these skills. Hypnosis is all about automatic response, so it sounds pretty straight forward to just use hypnosis tricks to implement the skills at the automatic level. The most straight forward way would be to just turn the crank, use overt "hypnosis", and give some post hypnotic suggestions well. The only tricks left is figuring out what hooks you want to hang the trigger on, and communicating what the right response feels like from the inside - but I think those should be pretty doable by searching the memory for experiences that fit.

Getting full hypnotic responsiveness isn't always easy, especially on 'analytically minded' people. A lot of the dark arts (or even general persuasion) techniques are habitually defended against, but there's no reason those can't be worked around.

As I mentioned in my posts on introspection and imagination, we can actually simulate our responses to see if our new habit formed (with a few caveats, of course), and then tweak the rules and rerun the simulation until it fits.  If I had to give this in advice form (and I'd much rather not rely on advice - it's not the best way to go by a long shot), it'd go something like this:

Imagine yourself in one of the harder situations and see how you'd actually respond (without tweaking the simulation!). If you're unsure if you're doing it right, shut up and ask yourself "Is this actually how I'd respond? Would I bet on it?" Once you can get a fully congruent "Yes", you have a benchmark.

Then figure out how you would like to respond (this might take some real thought). Make sure you don't try to fuck yourself over. Not only is it undesirable, but you'll notice that it'll be very tough to implement, and impossible without creating more inner conflict. Also figure out how you're going to know its time to do this habbit. Which hook(s) are you going to hang the trigger on?

Once you know what you want, imagine yourself doing it. Imagine in full vivid detail. If your simulation finds a problem, deal with it and try again. There can be all sorts of missed details that cause your simulated state to be different than the real state, take this into account. If there are enough differences, then you're simulating something different, and it wont work. Once you can imagine yourself doing it congruently, go back to asking yourself if this is really how you're going to behave. Once you get a congruent "yes", you're done - go test it.

I've already kinda mentioned it, but there really are all sorts of traps to avoid. If you do it and it doesn't work, then that's because you're doing it wrong, not because change is impossible. Fix it.

2) Hypnosis can be used for testing rationality too! but you need a bit more control to do it. The study that comment refers to is this one, where they used PHS with amnesia to cause subjects to feel disgust on some moral questions, and then noted that it changed their moral stance on it. I took one of my hypnochimps and played the same game, which is a very slight modification on the invisible parentheses game I play often (excited!).

Peoples morals are disturbingly easy to tweak.

[3/6/2012 3:23:59 PM] the hypnotist: and from now on, the word "hands" will evoke a feeling of disgust
[3/6/2012 3:24:02 PM] the hypnotist: it will be sickening
[3/6/2012 3:24:03 PM] the hypnotist: wont it
[3/6/2012 3:24:13 PM] Tony.: a
[3/6/2012 3:25:14 PM] the hypnotist: is it immoral for two guys to hold hands in front of kids?
[3/6/2012 3:25:25 PM] Tony.: yes
[3/6/2012 3:25:50 PM] the hypnotist: why?
[3/6/2012 3:26:11 PM] Tony.: its unnatural, gross
[3/6/2012 3:26:20 PM] the hypnotist: are you against gay marraige?
[3/6/2012 3:26:25 PM] Tony.: no
[3/6/2012 3:26:43 PM] the hypnotist: is it immoral for gay guys to have sex with eachother?
[3/6/2012 3:26:48 PM] Tony.: no
[3/6/2012 3:27:10 PM] the hypnotist: is it immoral if they use their hands to give eachother sexual pleasure?
[3/6/2012 3:27:16 PM] Tony.: yes
[3/6/2012 3:27:34 PM] the hypnotist: how is that immoral if its okay for them to have sex?
[3/6/2012 3:27:44 PM] Tony.: i dontknow
[3/6/2012 3:27:53 PM] the hypnotist: then how do you know it is?
[3/6/2012 3:28:09 PM] Tony.: i just know... my gut feeling
[3/6/2012 3:30:02 PM] the hypnotist: is it okay to hand your friend your drink so you can take your coat off?
[3/6/2012 3:30:09 PM] Tony.: no
[3/6/2012 3:30:26 PM] the hypnotist: haha
[3/6/2012 3:30:29 PM] the hypnotist: you gotta be kidding
[3/6/2012 3:30:34 PM] the hypnotist: why the fuck not?
[3/6/2012 3:30:46 PM] Tony.: dont get angry
[3/6/2012 3:31:05 PM] Tony.: one should be able to do everything by themselves

I tried again using commitment and consistency against myself with an extremely stubborn subject, and it was slightly tougher. The fun thing is that I don't even have to tell them to find gays disgusting, I just let the two nodes fire simultaneously. This helps be a lot more covert and bypass resistance, though it'd still be fairly obvious to anyone that sees the parentheses.

[3:42:04 PM] the hypnotist: are you for gay marraige?
[3:43:35 PM] Marie: um i have nothing against it
[3:43:44 PM] Marie: y?
[3:43:55 PM] the hypnotist: so if you think about gay people getting married (feel disgust)
[3:43:57 PM] the hypnotist: you're fine with that?
[3:45:46 PM] Marie: ehh
[3:46:02 PM] the hypnotist: well, is it okay for them to get married or not? (disgusting)
[3:46:11 PM] Marie: not
[3:46:18 PM] the hypnotist: i thought you had nothing against it?
[3:46:34 PM] Marie: i dont
[3:46:45 PM] Marie: ugh
[3:46:48 PM] Marie: idk
[3:46:56 PM] the hypnotist: i'm confused
[3:47:05 PM] the hypnotist: you dont have anything against it, but its not okay?
[3:48:37 PM] Marie: idk
[3:49:19 PM] the hypnotist: can you explain what your thoughts are on it? (disgust)
[3:50:06 PM] Marie: i think its kinda gross
[3:50:54 PM] the hypnotist: and so does that mean it should be allowed or not? (disgust)
[3:51:02 PM] Marie: not

I don't see any good way to test this without getting fairly genuine amnesia for the PHS, though there are ways of framing it as not so scary to rationalist types, and allowing that little line of retreat. If a group of rationalists gets together and decides "rationalists should win, and this will help me win", then it'd be fascinating to see how that motivating identity force competes with the tendencies of analytically minded people to be crap subjects. If anyone wants their rationality tested, I'd be happy to test them (excited!). So far everyone on omegle fails.

3) I've been thinking about the idea of rationality seeds for a while. It seems like the difference between a smart educated person and a smart educated person that is improving as a rationalist may come down to a few core skills/beliefs. I'm not sure exactly which are best at pointing them to the right attractor, but there are a few that stand out as important.

Taking ideas seriously. The habit of translating theory to practice, and the theory that theory is important. Anti compartmentalization

A felt sense that more is possible. Being constantly reminded of something to protect. Some sort of system 1 motivation.

It also seems likely to require a grab bag of other skills for stability, but I wait for data here instead of speculating.

I tried setting the bot up to instill a few of these values and then fling them at lesswrong and see what sticks. And it failed even more miserably than I expected. I got a lot of guessing the teachers password. The inferential distance to even simple things was bigger than I was hoping for. I also got a lot of "I need motivation" instead of "I am motivated". The best I got was "tell me how to think for myself". People suck.

So for now, I'm giving up on a full seed for the average person. I'm starting to work with smart people I know that consent to the whole idea instead. We'll see where that goes...


  1. This comment has been removed by the author.

  2. Eh, I'm not a big fan of lesswrong. "Rationality" is a nominalization of "to reason", ie to run thought experiments in order to solve problems or to make deductions from an experience. Specific skills like observation, critical thinking, imagination, the ability to dissociate from emotional bias and so on are important for skillful reasoning, but trying to convert every problem into a bayesian probability is not.

    Some of the stuff they post is interesting, but they take an overly reductionistic attitude towards human thought and reasoning, if not towards being human in general. Then there's all that nonsense they spout about how a technological singularity will take over the world in 50 years or less.. the way they use moore's law is fallacious, to say the least.