The Emperor Experiment

Years ago I became intrigued by the idea that some forms of perception involve physical mimicry. One of my college linguistics classes (back in the mid-90s) touched on the motor theory of speech perception, which asserts that when we listen to speech, we reverse-engineer the sounds we hear, identifying each sound with the motion or “gesture” of the vocal tract that would be needed to produce it. So, when I listen to you say a word like “dog,” I recognize the word by preparing to say it myself, or at least by figuring out what I’d need to do to with my mouth in order to make those same sounds.  In fact, some studies show that seeing and hearing speech excites the motor system involved in speech production (to quote the title of one paper by Watkins, Strafella, and Paus) — when I listen to you say “dog,” I exhibit neural activity similar to the act of speaking.

When I first heard about these ideas, I wondered if there was a way I could test them myself, or at least explore the territory through an introspective experiment. If listening to speech, or even imagining speech, involves some degree of physical mimicry, what would it happen if that mimicry were obstructed? Would it be harder to perceive speech, or to imagine speech (two separate but related processes) if I adopted an awkward physical position — a configuration of the vocal tract that would make it difficult to physically mimic what I was hearing?  I came up with the following simple experiment that focuses on imagined or “inner” speech:

The Emperor Experiment

1) Stick your tongue out as far as you can (as when a doctor asks you to say “ahhh”) and keep it in that position.

2) Now try to imagine yourself saying the word “emperor.” Try to hear the word in your mind’s ear (as when you talk silently to yourself) but don’t actually say it out loud or do anything with your mouth besides holding your tongue out.  Although your tongue should remain extended as you do this, you should try to imagine the word pronounced clearly, as you would normally say it (i.e. without an extended tongue).

Stop reading and try it!

Were you able to hear the word “emperor” enunciated without a lisp in your mind’s ear? Or did you find that were only able to hear something like em-pah-wah — the way “emperor” would sound if actually spoken with an extended tongue?

I’ve administered this experiment to countless “subjects” informally over the years, including dozens of unsuspecting conversation partners at coffee shops throughout New England, and approximately 30 classmates in an Artificial Intelligence class I took as a grad student. I find that roughly half of participants react to me as though I’m crazy, not because I’m asking them to do something silly like sticking out their tongue, but because they have no problem imagining a clearly pronounced “emperor” in step 2 and therefore don’t see any point to the experiment. The other half quickly bursts out laughing in step 2, because they discover (often to their great surprise) that they simply cannot imagine a clearly pronounced “emperor” while they keep their tongues extended — they can only imagine a mangled, lisping em-pah-wah. Their imagination is a slave to their tongue!

My own experience was that I could only hear em-pah-wah the first couple of times I tried the experiment, but with some practice, I gained the ability to imagine a clear “emperor” no matter the position of my tongue. So it seems that a physical obstruction interferes with imagined speech for some subjects initially, but it’s possible to learn to separate the imagined sound from what it suggested by the obstruction.

I still wonder why people have such different experiences the first time they try this expirement: why is it initially easy for some but hard for others? And how might other kinds of physical or motor interference affect what we can perceive or imagine? I haven’t had occasion to study “extended tongue effect” formally, and I don’t know whether something similar to the emperor experiment occurs in the research literature: I’d appreciate any references you might send my way.

The emperor experiment came back to mind today after several years of dormancy: I was having coffee this afternoon with a neuroscience researcher (and founder of the startup momedx) who has done some fascinating work on visual perception in people who gain or regain sight after years of blindness. On his first try, my coffee companion reported that he heard “em-pah-wah.”

Diversions, Language


Driving on I-90 from Boston towards the Berkshires yesterday, I spied a curious road sign gleaming in the summer haze:


My mind skipped over three possible interpretations of this grammatically ambiguous statement and landed on a fourth. I didn’t read it as an assertion of the tenet that plows are cautious; I didn’t read it as an admonition to plows to be cautious; and I didn’t take it as a directive to non-plows to avoid plows. Channeling my inner copy editor (always the pessimist), I assumed it must be a simple misspelling of:


It’s not easy making signs (as evidenced by the preponderance of wretched signage in our world), and so I can easily imagine an incautious sign-maker trying to write PLEASE and having it come out as PLOWS.

Unfortunately, the very thought of this has put me in a state where I can no longer see the word please without mentally substituting it with plows (and the result turns out to be grammatical with surprising frequency). And so I invite you, dear reader, to join me in this affliction by listing some of your favorite please/plows substitutions. I will start off with one of my favorites:

Always say plows.