What I Tell You Three Times Is True

by Scott Alexander 4 min read2nd May 200932 comments


"The human brain evidently operates on some variation of the famous principle enunciated in 'The Hunting of the Snark': 'What I tell you three times is true.'"

   -- Norbert Weiner, from Cybernetics

Ask for a high-profile rationalist, and you'll hear about Richard Dawkins or James Randi or maybe Peter Thiel. Not a lot of people would immediately name Scott Adams, creator of Dilbert. But as readers of his blog know, he's got a deep interest in rationality, and sometimes it shows up in his comics: for example, this one from last week. How many people can expose several million people to the phrase "Boltzmann brain hypothesis" and have them enjoy it?

So I was very surprised to find Adams was a believer in and evangelist of something that sounded a lot like pseudoscience. "Affirmations" are positive statements made with the belief that saying the statement loud enough and long enough will help it come true. For example, you might say "I will become a syndicated cartoonist" fifteen times before bed every night, thinking that this will in fact make you a syndicated cartoonist. Adams partially credits his success as a cartoonist to doing exactly this.

He admits "it sounds as if I believe in some sort of voodoo or magic", and acknowledges that "skeptics have suggested, and reasonably so, that this is a classic case of selective memory" but still swears that it works. He also has "received thousands of e-mails from people recounting their own experiences with affirmations. Most people seem to be amazed at how well they worked."

None of this should be taken too seriously without a controlled scientific study investigating it, of course. But is it worth the effort of a study, or should it be filed under "so stupid that it's not worth anyone's time to investigate further"?

I think there's a good case to be made from within a rationalist/scientific worldview that affirmations may in fact be effective for certain goals. Not miraculously effective, but not totally useless either.

To build this case, I want to provide evidence for two propositions. First, that whether we subconsciously believe we can succeed affects whether or not we succeed. Second, that repeating a statement verbally can make the subconscious believe it.

The link between belief in success and success has progressed beyond the motivational speaker stage and into the scientific evidence stage. The best-known of these links is the placebo effect. For certain diseases, believing that you will get better does increase your probability of getting better. This works not only subjectively (ie you feel less pain) but objectively (ie ulcers heal more quickly, inflammation decreases).

The placebo effect applies in some stranger cases outside simple curative drugs. A placebo stop-smoking pill does increase your chance of successfully quitting tobacco. Placebo strength pills enable you to run faster and lift more weight. Placebo alcohol makes you more gregarious and less inhibited1. Placebo therapies for phobia can desensitize you to otherwise terrifying stimuli.

There are some great studies about the effect of belief in school settings. Pick a student at random and tell the teacher that she's especially smart, and by the end of the year she will be doing exceptionally well; tell the teacher that she is exceptionally stupid, and by the end of the year she'll be doing exceptionally poorly. The experimenters theorized that the teacher's belief about the student's intelligence was subconsciously detected by the student, and that the student was somehow adjusted her performance to fit that belief. In a similar study, minority students were found to do worse on tests when reminded of stereotypes that minorities are stupid, and better when tested in contexts that downplayed their minority status, suggesting that the students' belief that they would fail was enough to make them fail.

Belief can also translate to success when mediated by signals of dominance and confidence. We've already discussed how hard-to-fake signals of confidence can help someone pick up women2. Although I don't know of any studies proving that confidence/dominance signals help a businessperson get promoted or a politician get elected, common sense suggests they do. For example, height does have a proven effect in this area, suggesting that our ancestral algorithms for assessing dominance play a major role.

MBlume has already discussed how one cannot simply choose to consciously project dominance signals. The expressions and postures involved are too complicated and too far from the normal domain of conscious control. He suggests using imagination and self-deception to trick the subconscious mind into adopting the necessary role.

So I hope it is not too controversial when I say that subconscious beliefs can significantly affect disease, willpower, physical strength, intelligence, and romantic and financial success.

The second part of my case is that repeating something makes the brain believe it on a subconscious level.

Say Anna Salamon and Steve Rayhawk: "Any random thing you say or do in the absence of obvious outside pressure, can hijack your self-concept for the medium- to long-term future." That's from their excellent post Cached Selves, where they explain that once you say something, even if you don't really mean it, it affects all your beliefs and behaviors afterwards. If you haven't read it yet, read it now: it is one of Less Wrong's growing number of classics.

There's also this study which someone linked me to on Overcoming Bias and to which I keep returning. It demonstrates pretty clearly that we don't have a lot of access to our own beliefs, and tend to make them up based on our behavior. So if I am repeating "I will become a syndicated cartoonist", and my subconscious is not subtle enough to realize I am doing it as part of a complex plot, it might very well assume I am doing it because, well, I think I will become a syndicated cartoonist. And the subconscious quite likes to keep beliefs consistent, so once it "discovers" I have that belief, it may edit whatever it needs to edit to become more consistent with it.

There have been a few studies vaguely related to affirmations. One that came out just a few weeks ago found that minorities who wrote 'value affirmation' essays did significantly better in school (although the same effect did not apply to Caucasians) . Another found that some similar sort of 'value affirmation' decreased stress as measured in cortisol and other physiological measures . But AFAIK no one's ever done a study on Adams-variety simple personal affirmations in all of their counter-intuitive weirdness, probably because it sounds so silly, and I think that's a shame. If they works, it's a useful self-help technique and akrasia-buster. If they don't work, that blocks off a few theories about how the mind works and helps us start looking for alternatives.



1: A story I like: in one of the studies that discovered the placebo effect for alcohol, one of the male participants who falsely believed he was drunk tried to cop a feel of a female researcher's breasts. That must have been the most awkward debriefing ever.

2: Here I'm not just making my usual mistake and being accidentally sexist; I really mean "pick up women". There is less research suggesting the same thing works on men. See Chapter 6 of The Adapted Mind, "The Evolution of Sexual Attraction: Evaluative Mechanisms in Women".