The experiment went like this.
Researchers showed participants a video of Derren Brown performing a mental influence piece. In the video, Brown used a combination of verbal cues — certain words repeated at specific moments — and physical gestures while asking a spectator to think of a playing card. The target card was the Three of Diamonds. The researchers wanted to know: did the primes actually work? Could verbal and gestural cues embedded in a seemingly casual conversation significantly increase the probability that someone would think of a specific card?
The answer was yes. And the magnitude was startling.
In control conditions — where participants simply named a card with no priming — the Three of Diamonds appeared at roughly its base rate, a little under two percent (one out of fifty-two cards). In primed conditions, the rate jumped to over twelve percent. That is more than six times the expected rate. A casual-seeming conversation, loaded with the right cues, made a specific card choice six times more likely.
When I encountered this finding in the research of Gustav Kuhn and Alice Pailhes, I had to read it three times. Not because I did not understand it, but because the implications were so large that I wanted to make sure I was not misinterpreting the data.
Six times more likely. From a conversation.
What Is Priming?
Priming is one of the most well-established phenomena in cognitive psychology. In its simplest form, it means that exposure to one stimulus influences the response to a subsequent stimulus, without conscious awareness.
The classic demonstration is the word-completion task. If you read a passage about food and then are asked to complete the fragment S-O-_-P, you are more likely to say SOUP than SOAP. If you read a passage about cleaning, you are more likely to say SOAP. The prior context activates certain concepts in memory, making them more accessible and more likely to come to mind when needed.
Priming operates entirely through System 1. The person being primed does not know they are being primed. They experience their response as spontaneous and self-generated. If you ask someone who just completed S-O-U-P why they chose that word, they will give you a reason — “I was hungry,” “I like soup,” “it just came to mind” — but they will almost never say “because I just read about food.” The actual cause of their choice is invisible to them.
This invisibility is what makes priming so powerful and so relevant to performance. The spectator does not experience influence. They experience a free, spontaneous choice. The primes are below the threshold of conscious detection.
The Three of Diamonds Study
Let me go deeper into the Derren Brown experiment, because the details matter.
The researchers analyzed the specific primes Brown used in his televised performance. There were verbal primes — words and phrases embedded in the conversation that contained phonetic or semantic connections to “three” and “diamonds.” There were also gestural primes — hand movements that subtly traced diamond shapes or held up three fingers at key moments.
The researchers then isolated these components to test their individual and combined effects. They found that both verbal and gestural primes independently increased the rate of choosing the Three of Diamonds compared to control conditions. The combined effect was the strongest.
But here is what I find most fascinating: the effect was not binary. It was not the case that priming either worked or did not work. Instead, priming shifted the probability distribution. In the control condition, the Three of Diamonds was chosen by about two percent of people. In the fully primed condition, it was chosen by about twelve percent. That means eighty-eight percent of people still chose something else. Priming did not control the choice. It influenced the probability.
This distinction matters enormously for understanding what priming can and cannot do. It is not mind control. It is not guaranteed. It is a probabilistic influence that tilts the odds in a specific direction. The tilt can be substantial — six times the base rate is a dramatic shift — but it is still probabilistic.
How I Started Thinking About This
After reading the research, I started listening to my own conversational patterns during performances with new ears.
I was in a hotel room in Linz — laptop open, a deck of cards on the desk beside me, a recording of my most recent performance playing on my phone. I was not listening for technique. I was listening for language. Specifically, I was listening for the inadvertent primes I was already delivering without knowing it.
And they were everywhere.
In casual conversation with spectators before an effect, I was using words, phrases, and descriptions that were loaded with associations. Some of these were helpful — they happened to prime concepts related to the effect I was about to perform. Others were neutral. And some, I realized with a sinking feeling, were actively counterproductive — they primed concepts that worked against what I was trying to do.
This was an eye-opening moment. I had been treating the conversational lead-up to an effect as filler — warm social interaction that served no specific purpose beyond building rapport. But from a priming perspective, every word in that conversation was doing something. Every concept I introduced was activating related concepts in the spectator’s mind. Every story I told was setting the stage, whether I intended it to or not.
The question was not whether I was priming my spectators. I was always priming them. The question was whether I was priming them intentionally and effectively, or accidentally and randomly.
The Subtlety Problem
One of the things that concerned me when I first started thinking about conversational priming was subtlety. How embedded do the primes need to be? Can they be so subtle that the conscious mind does not notice them, or does the spectator need to actually process the relevant words for priming to work?
The research offers a nuanced answer. Primes work best when they are processed but not noticed as primes. The spectator needs to hear and understand the words — priming does not work if the stimuli are completely subliminal, below the threshold of perception. But the spectator should not realize that the words are doing something specific. The moment a spectator thinks “he keeps saying that word — is he trying to influence me?” the priming effect collapses, because System 2 has been activated to counter it.
This creates a design constraint that I find endlessly challenging and interesting. The primes must be visible enough to be processed but invisible enough to avoid detection. They must be embedded naturally in conversation, so that hearing them feels unremarkable, but they must be specific enough to activate the right associations.
The skill, I have come to believe, is in the naturalness of the delivery. A prime embedded in a genuine, flowing conversation is invisible because the conversation itself provides perfect cover. A prime forced into an awkward pause or an unnatural sentence stands out precisely because it does not fit the conversational context.
I spent weeks in hotel rooms across Austria working on this. Not on technique, but on conversational flow. I would practice telling stories — stories about my day, about something I had read, about an interesting thing that happened at a previous event — and I would thread specific words and concepts through those stories. Then I would listen back and ask myself: does this sound natural? Would I say this if I were not performing? Does the word feel embedded or inserted?
The difference between embedded and inserted is everything. An embedded prime disappears into the sentence like a note disappearing into a chord. An inserted prime sticks out like a wrong note in a melody. The listener may not consciously identify what is off, but something feels unnatural, and that feeling of unnaturalness is enough to activate suspicion and engage System 2.
Semantic vs. Phonetic Priming
The research distinguishes between two types of priming that are relevant here.
Semantic priming works through meaning. If I talk about precious stones, gems, and jewelry, I am semantically priming the concept of “diamond” — including the diamond suit of playing cards. The connection is through meaning and association.
Phonetic priming works through sound. If I use words that contain sounds similar to the target — “free” rhyming with “three,” for example — the phonetic similarity activates the target word even without a semantic connection.
Both types work, and combining them amplifies the effect. The Derren Brown study found that the verbal primes included both semantic and phonetic elements. The researchers’ analysis suggested that the combination was more effective than either type alone.
For me, the practical takeaway was that conversational priming is most effective when it operates on multiple channels simultaneously. A story that is semantically related to the target, told with natural emphasis on phonetically related words, and accompanied by gestures that reinforce the visual associations, creates a web of activation that makes the target concept highly accessible without any single element being detectable as a prime.
The Limitations Matter
I want to be honest about the limitations of priming, because it would be easy to read the “six times more likely” headline and conclude that priming is an infallible tool. It is not.
First, as I mentioned, priming is probabilistic. Six times the base rate still means that the majority of people will choose something other than the target. Priming shifts odds; it does not determine outcomes.
Second, priming effects are sensitive to context. A prime that works in a relaxed social setting may not work in a formal or stressful one. A prime that works in a face-to-face conversation may not work through a screen. The environmental variables matter.
Third, individual differences are real. Some people are more susceptible to priming effects than others. People who tend to be more analytically minded, or who are in an analytical mode at the moment of choice, are less susceptible. People who are relaxed, social, and operating in System 1 are more susceptible.
Fourth, awareness of priming partially inoculates against it. People who know about priming — and this includes many educated, well-read audience members — can consciously resist it, though the research suggests that even aware participants are still influenced to some degree. Knowing about a cognitive bias reduces its effect but rarely eliminates it entirely.
These limitations are not discouraging. They are clarifying. They tell me that priming is one tool among many, not a magic bullet. It is most effective when combined with other psychological principles — population stereotypes, choice architecture, social dynamics — and least effective when relied upon in isolation.
The Consultant’s Parallel
In my consulting work, I encounter priming all the time, though we use different language for it. When a company designs a marketing campaign, the images, words, and cultural references they choose are primes. They activate associations that shape how consumers perceive the product. When a negotiator frames a discussion, the metaphors and examples they choose prime the other party’s thinking about the deal.
Understanding priming in the performance context made me a better consultant, and understanding it in the consulting context made me a better performer. The underlying psychology is identical. The application differs, but the principle — that prior exposure to concepts shapes subsequent decisions — operates everywhere.
At a keynote in Graz last year, I was speaking about innovation strategy and I included a segment on cognitive priming in marketing. I explained the research, showed the data, and watched a room of business executives have the same reaction I had in that hotel room in Linz: a mixture of fascination and discomfort. Fascination because the effect sizes are genuinely impressive. Discomfort because it suggests our decisions are less autonomous than we like to believe.
Then, later in the keynote, I performed a mentalism piece. I did not reference the priming segment. I did not draw any explicit connection. But I had structured the earlier portion of the talk to serve a dual purpose — to educate the audience about priming while simultaneously demonstrating its principles in real time.
The reaction, when the effect landed, was different from what I typically get at corporate events. There was the usual astonishment, but layered on top of it was a deeper recognition. They had just learned about priming. And now they had just experienced something that, despite their new knowledge, they could not explain. The intellectual understanding did not protect them from the experience. It amplified it.
That is the real power of priming research. Not just as a tool for performance, but as a window into the architecture of human choice. The choices we make are shaped by contexts we do not notice, by words we process but do not remember, by associations that operate below the waterline of consciousness.
Six times more likely. From a conversation that felt like nothing at all.