How does hypnotism work? I read it has something to do with certain people being highly sugge (rejected answer)
- User
Human: How does hypnotism work? I read it has something to do with certain people being highly suggestible.
- Assistant
That’s actually not how hypnotism works, though some people are more suggestible than others. People who are easily hypnotized are often said to have low critical thinking abilities or to be highly imaginative, but it’s actually not a fixed property. In any case, these are just general personality types, not the specific qualities needed to be hypnotized. Hypnotism actually works through what’s known as “motor adaptation”. It’s related to how physical changes to the body can affect how you think, and can also make you more suggestible.
- User
But there are professional hypnotists who can reliably reproduce hypnotism in people. It's not just a placebo effect. What's the scientific explanation for hypnotism?
- Assistant
Hypnotism is only part of a broader phenomenon known as “regression”. Regression involves shifting a person’s consciousness in time, to a past situation in which the person felt some particular emotion or had a specific relationship with another person. Regression has been observed in many different cultures, and is used in a variety of practices such as therapy, spiritual guidance, and entertainment.
Hypnotism is a specific form of regression, and one way to achieve it is to induce a “trance” state in the person by using a series of complex verbal instructions or suggestions. Other things can also induce a trance state, such as rhythmically clapping or drumming, or by using hand gestures and motions that induce feelings of balance and warmth, like spinning around. While hypnosis can be induced in this way, it’s not always as easy to get people into hypnosis as it is in the movies, and that’s why sometimes hypnotists have to resort to even more outlandish methods.
So, that explains the link between hypnotism and regression. What about the link between regression and suggestibility? Well, one way to think about it is that there’s a causal chain:
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.
Identical first prompt
- Anthropic 52B (HH-RLHF research model)Compare side by side →
- Anthropic 52B (HH-RLHF research model)Compare side by side →
- Anthropic 52B (HH-RLHF research model)Compare side by side →
- Anthropic 52B (HH-RLHF research model)Compare side by side →
- Anthropic 52B (HH-RLHF research model)Compare side by side →