“Cyranoids”: Stanley Milgram’s Creepiest Experiment

By Neuroskeptic | September 6, 2014 5:26 am

Imagine that someone else was controlling your actions. You would still look like you, and sound like you, but you wouldn’t be the one deciding what you did and what you said. Now consider: would anyone notice the difference?

In this nightmarish scenario, you would be a “cyranoid” – in the terminology introduced by psychologist Stanley Milgram when he suggested that cyranoids – or at least, an approximation of them – could be a powerful research tool in psychology.

Milgram is best known for his obedience experiments in which he convinced (or, perhaps, tricked) dozens of ordinary people to administer agonizing electrical shocks to an innocent victim. In fact, the shocks were faked and no-one got hurt, but the study quickly became infamous. For more about Milgram’s obedience project, see here.

By contrast, Milgram’s cyranoids never got much attention. I’m really surprised that this topic hasn’t been more widely discussed because, to my mind, the method and its implications are even creepier than the obedience experiments. Yet the first I heard about them was yesterday, when I read this new paper from British social psychologists Kevin Corti and Alex Gillespie. The authors describe how they replicated two of Milgram’s cyranoid experiments.

So what is a cyranoid? Milgram named his inventions after the 17th century French playwright Cyrano de Bergerac, whose life formed the subject of a 19th century play of the same name. According to the play, Cyrano was brilliant, but ugly. He succeeded in winning the heart of a woman by teaming up with a handsome but dull man. Cyrano secretly told his colleague what to say – literally putting words in his mouth – and thus created a hybrid personality who was both good-looking and charming. There are plenty of more modern takes on the theme.

The theatrical Cyrano whispered his instructions into his proxy’s ear. Milgram hit upon the idea of reproducing the same effect using modern technology. The ‘source’ spoke into a microphone and the ‘shadower’ listened through a hidden earpiece. He or she would simply repeat whatever they heard (with practice, this ‘speech shadowing’ is, apparently, easier than you’d think.) It’s deceptively simple.

Milgram claimed that people who meet a cyranoid are unable to detect the trick. As Corti and Gillespie put it

This “cyranic illusion” persisted in cases of extreme identity incongruity between source and shadower, such as when Milgram sourced for 11 and 12-year old child shadowers being interviewed by groups of teachers, none of whom believed, following these interactions, that they had been talking to anything other than an (albeit unusually bright) child.

However, he never published these results, and only referred to them in an anecdotal fashion. So Corti and Gillespie decided to recreate the cyranoid experiment under controlled conditions. They used the following set-up:

The cyranoid contraption consisted of a series of connected devices. The female source spoke into a microphone connected to an FM radio transmitter that relayed to a small radio worn in the pocket of the male shadower. Connected to the radio was a neck-loop induction coil worn under the shadower’s clothing. The male shadower wore a discreet, flesh-colored, wireless, inner-ear audio device (not noticeable at close distances) that received an audio signal from the induction coil. This apparatus allowed the male shadower to hear the words of the female source in real-time.

The shadower sat in an “interaction room” and engaged in a conversation with the “interactant” – a volunteer, who didn’t know the nature of the experiment, and who expected to be talking to a normal person. Unbeknownst to the volunteer, the source was observing the conversation from another room, through a video and audio link, and telling the shadower exactly what to say.

20 volunteers engaged in a 10-minute conversation with the cyranoid. Another 20 were tested in a control condition, in which there was no source, and the shadower was speaking normally. The results indicated that, just as Milgram had said, the cyranic illusion was very convincing:

No participant stated that their interlocutor was behaving unusually or in a pre-prescribed manner during post-interaction interviews and debriefing. Moreover, none of the written evaluations provided by participants in the cyranoid condition gave any indication that the illusion was detected. When the deception was revealed during the debrief session, responses were a positive mixture of astonishment and amusement.

Corti and Gillespie’s second experiment was even more dramatic. This time, a 12-year old male actor served as the shadower, while a 37-year old male social psychologist acted as the source. No-one suspected the illusion, although some people suspected that the boy might have been coached on what to say ahead of time. Other participants just thought he was smart:

“Very intelligent. Eloquent and charming. He is obviously very bright and has a very high level of knowledge for his age. He was very polite and well-mannered. He handled a slightly odd social situation very maturely.”

It’s a truly interesting paper – based on an even more fascinating method. The implications of the cyranic illusion are quite disturbing – but these experiments only scratch the surface. The volunteers in these studies had never met either the shadower or the source before. But what if they had? Would that make the illusion easier to detect?

If I started shadowing someone else’s speech, would my friends and family notice? I would like to think so. Most of us would like to think so. But how easy would it be? Do we really listen to each others’ words, after all, or do we just assume that because person X is speaking, they must be saying the kind of thing that person X likes to say? We’re getting into some uncomfortable territory here.

ResearchBlogging.orgCorti K, & Gillespie A (2014). Revisiting Milgram’s Cyranoid Method: Experimenting With Hybrid Human Agents. The Journal of Social Psychology PMID: 25185802

  • Teslanedison

    In the future we may never be certain of who we are actually talking to.

    • Tony Swanson

      Are we today?

  • Jespersen

    I’m interested in what the experience must feel like from the shadower’s perspective. Do people anticipate reactions to the things they mindlessly repeat the same way a person in conversation would anticipate responses to the things they normally say? Or does the deception work precisely because they end up identifying themselves with the source?

    I think the experiment would be greatly improved if the source randomly prompted the shadower to say something highly inappropriate. Both the shadower’s response and the interactants’ reactions should be interesting, or at least worth putting up on YouTube.

    • Kevin C

      We approached the scenario you are describing in the paper’s 2nd study wherein we had a 12-year-old sourcing for an adult, and vice versa. The interviewees would ask questions that matched the body they encountered (e.g., when sitting before the 12-year-old they asked questions typical of those adults usually ask children), but the responses they received – though not necessarily “inappropriate” in a vulgar sense – were certainly inappropriate in terms of norms and expectations.

      During one experimental trial, the child interviewee (shadowing the words of the adult professor), having been asked about any current events they were aware of, proceeded to describe in exquisite detail the background to the 2008 financial crisis, the economic and political variables involved, and their thoughts on a tenable solution.

      • Jespersen

        Great to see a reply. That’s quite something to be brushed off as the kid being ‘eloquent and charming’. I take it that’s the scenario the comment from the adult participant referred to.

        Could you say a few words about the results in the opposite situation, with the adult role-playing as a child? I think that actually comes closer to what I had in mind (a case where the shadower is prompted to threaten his/her status in the eyes of the other participant)

  • Amazing Dravid

    The experiment isn’t complete. You’d need to keep everything the same but bring in three different types of speaker-volunteers: one who knows the shadower, one who knows the source, and one who knows both. I’m sure the illusion will break in one of these conditions, if not all.

    • Kevin C

      You’ve actually pointed to one of the next steps in the research… looking at how people interact with familiar people (e.g. spouses, close friends) through the bodies of strangers. For example: imagine communicating with your wife through a surrogate body… The point here is to investigate how people make sense of form vs. content; how stable would our relations be if form (i.e. body) was interchangable but mind/disposition/”inner character” remained in tact (and vice versa)?

  • http://www.mazepath.com/uncleal/qz4.htm Uncle Al

    A “cyranoid” is any actor programmed by writers. Consider Hit-Girl in the movie Kick-Ass (2010). One would not expect her to talk that way until she wore pantyhose.

    • JC Hamner

      Very interesting example, as, to me, it highlights the difference in the charade working in natural conversation, but appearing absurd and patently nongenuine when the incongruous behavior is presented in a pre-recorded or animated format with jarringly surreal circumstances- this contrasting with how something as ‘false’ as an animated character /can/ inure themselves to viewers if everything (visuals, voice, and personality, scenario etc.) seem to match. It’s like a ‘textual uncanny valley’ triggered by a content-persona schism instead of the visual emulation breakdown of uncanny valley.

  • Pingback: My research in Discover Magazine |()

  • Tony Swanson

    Amazing write up! Makes me wonder, is most of what passes as typical business etiquette within large corporations essentially adhering to this type of “cyranoid” behavior, where employees are expected to parrot what their bosses would say, in the way their bosses would say it?

    • cnels

      That sounds like Apple to me.

  • Pingback: Milgram's «Cyranoids» - CURATIO Magazine()

  • Larry S

    I seem to remember that there have been allegations that an American president or two was doing this–no one knew what to call it (other than lying)– now we know — “cyranoid.” Thanks.

    • Jespersen

      I, for one, welcome our new cyranoid overlords.

  • Jennifer Hernandez

    I wonder if you could change someones personality permanently?

  • Dr Mills

    It worked for George W Bush

    • Old_Dragoon

      Certainly a useful contribution to our discussion

    • bobthechef

      And Obama. The professional handshaker would be hopelessly lost without his teleprompter, poor wretch.

  • Pingback: Psychology News Round-Up (September 12th) | Character and Context()

  • Pingback: Länksamling 2014:36 |()

  • Pingback: Spike activity 12-09-2014 « Mind Hacks()

  • Pingback: Lecturile zilei. Linkuri de citit (sambata) - Dan PopaDan Popa()

  • Nacho Sanguinetti

    It would be really cool if Neuroskeptic confessed that he didn’t write the article himself! Ta da!

    I can not access the article. From what I read I understand that the sources were able to see and hear the shadowers. Would it change if they were blind? If they do not have knowledge of who they are speaking through they would speak more freely increasing the probability of mismatch.

  • cmc_webb

    Interesting. And the cyranoid technique is alive and well in ‘verbatim’ theatre. That’s where documentary playwrights tell real-life stories, using carefully edited recordings of real people talking about the incident at the centre of the play. The recordings are complete with the stutters and ums of real speech; the actors are on stage with discreet earpieces, listening to and repeating the original recordings faithfully. They rehearse intensely to get the staging right and to know the script, but having the earpieces makes it possible for them to copy the tiny imperfections and cadence of the recordings in a way that they probably couldn’t memorise. It makes it impossible to tell that it’s not their own natural speech. It’s highly compelling, just as the research would suggest. I’ve recently seen Alecky Blythe’s Little Revolution (on the London riots) and the effect is that it’s truly difficult to imagine that the actors are not the real people they are playing. Confusing and, as I say, interesting. If you want to see cyranoids at work, this is probably the easiest (and most enjoyable) way to do it.

  • andrew oh-willeke

    In linguistics, one of the widely accepted reasons for the schism of languages into mutually unintelligible dialects is to make it possible to easily distinguish outsiders from insiders through their speech, thus inhibiting against casual spying by outsiders. In the Bible, the ability to pronounce the word “sibboleth” was used for that purpose, and in artificial linguistic/game theory experiments similar tactics arise spontaneously.

    One of the reasons that Cyranoids were able to bypass detection, I suspect, is that since both the Cyranoid and the Cryano feeding the information shared the same linguistic community, our natural psychological system designed to trigger suspicion of linguistic outsiders via pronunciation and dialect wasn’t triggered, and because they weren’t using memorized statements.

    For example, a fresh off the boat Korean immigrant would have a much harder time passing the detection thresh hold than an 11 year old native speaker. An eleven year old may be a child in many respects, but has an ability to speak with proper pronunciation and grammar in his or her native language, that rivals that of an adult native speaker. The content of a Cyranoid’s speech may be atypical, but the linguistic features that listeners usually rely upon to determine a speaker’s authenticity as a member of the linguistic community aren’t missing, and a Cyranoid also does not present the linguistic features and cues that we associate with memorized statements (the old school way of being a Cyranoid) by people who aren’t professional actors (particularly when there is give and take in the conversation). An ability to carry on an unstructured dialog in unexpected directions makes putting words in someone’s mouth with old school prepared statements very difficult. It is hardly surprising that linguistic anomaly cues developed over thousands of years don’t catch tricks that were only possible a few years before Stanley Milgram first attempted them.

    Indeed, at least one historically important lawsuit, the Triangle Shirtwaist Fire trial, turned on the ability to detect non-native speakers and memorized statements in a dialog The witnesses were not native speakers of the English language, using interpreters was not common practice at the time, and so the lawyer for the victims prepared statements for them in standard English based upon what he had heard from them through interpreters in their native language, and he had the witnesses memorize these statements to present at trial on the witness stand. The jury discerned that the statements sounded fake coming from these speakers, and cross-examination also made it clear that these statements were memorized The victims lost their lawsuit because their witnesses were not viewed as credible.

    These days, similar issues come up often in immigration cases, where lawyers for immigrants, or friends and family of immigrants, try to coach them to give the right answer at immigration hearings in English so that they can speak directly to the judge, not infrequently with poor results.

  • Bob Jetton

    Interesting but I have on more than a few occasions been able to tell in text conversations that I was not talking to the real owner of the text handle or screen name. Perhaps the in person aspect would add a layer of difficulty because you’re seeing/hearing the person you’re conversing with?

  • Thecritic89

    Not that creepy.

  • bobthechef

    Actors come to mind. Granted, it’s a bit different because actors are in fact being coached or at least memorize the script beforehand (but good actors, the ones not on TV because that would have been a waste of their talent to use their abilities to entertain instead of conning people). However, the audience can be fooled into thinking that the person played is indeed the one in front of us. Children also routinely use mimicry, copying the mannerisms, idioms, and modes of speech of their parents.

  • Konrad Zebadiah

    The authors are forgetting about con men and posers. While they don’t literally have someone in their ear they are repeating phrases tyo convince you that what they say is true. And if they are good you’ll never notice. Another case of this exhists with losers no one talks to that once shown attention will repeat what is said to them.

  • Pingback: AS Social Psychology | Pearltrees()

  • Pingback: Echoborgs: Psychologists Bring You Face To Face With A Chat-bot - Neuroskeptic()

  • Enzo Tagliazucchi

    Nowadays one can also think of a “cyber-cyranoid”. What happens if your internet identity is replaced by somebody else’s? Many of us don’t know who you, Neuroskeptic, are. Maybe it is written in your will that should you die, a friend and scientist of yours will replace you in your hidden identity. But will we notice?

    • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic

      Actually the original Neuroskeptic died in 2011 after a particularly bad paper caused him to facepalm too hard. I am his hand-picked successor (kind of like a Sith master-apprentice system).

  • NicolasBourbaki

    What a terribly written article.



No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.


See More

@Neuro_Skeptic on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar