Based on my (non-expert, and many years ago) survey of the literature + updates from the replication crisis. I’m a software engineer with a special interest, not a cognitive scientist or a neuroscientist or a behavioral economist or something, so if you are an expert and I got something wrong, let me know.
The popular image of subliminal messages as an irresistible technological super-hypnosis is (obviously) completely false. However, subliminal messages can exert influence in interesting ways.
There are a few reasons why we know the ‘technological super-hypnosis’ thesis is false. One is empirical: the technology to do single frame cuts was pretty widely available, so this is an easy experiment to replicate, and it failed replication immediately. It owes its continued vitality to a lurid mass market paperback and its incurious readerbase, in a kind of Andrew Wakefield-esque series of events. Another reason is genealogical: this idea was modeled on the idea that hypnosis produced suggestibility by being a “direct line to the subconscious”, a super-powerful (often super-natural) domain of perfect memories and perfect naievety, easily corrupted and manipulated by cunning foreign invaders (the high-eyebrowed slavic-accented hypnotist with the ornate gold watch). The model of hypnosis attributable to Svengali (and the play, Trilby, in which he is a character) is a cod-freudianism even less correct than freudianism; there’s no psychic censor to bypass. So we’ve got a protoscience being appropriated by a pseudoscience, misapplied to a different set of…
