Most speech pathologists nebulously conceptualized stuttering as an "incoordination" among respiratory, vocal cord, and articulatory mechanisms. However, the precise nature of this incoordination was never spelled out and there didn't seem to be research to support it.

Why, if there was an "incoordination," did it disappear when the patient talked out loud to himself? And why was it not present continuously but only on certain words?

After reviewing the literature, my feeling remained that the physical cause of stuttering lay at the vocal cords and that all the other behaviors seen were a reaction to the constriction of the vocal cords. But the precise nature of this response eluded me, until one day, quite by accident, I found an answer.

There is a door in my office through which I pass each day. The procedure is always the same. I go to the door, put my hand on the doorknob, turn it, pull the door open, walk through, and the door closes automatically behind me. It always works, I always expect it to work, and I'm never disappointed.

But one particular day a water pipe broke in the office above me, and the water seeped slowly through the ceiling and into the wooden door. The door was swollen and stuck in the door frame but I didn't know it because it had been a slow ooze I couldn't see it.

So I went to the door, the way I have every day for years, put my hand on the doorknob, turned it, and pulled. But it didn't move; it was stuck. My initial reaction was to pull harder, and that didn't work. So I pulled ever harder, until finally, I wrenched the door open.

Twenty minutes later, I returned to the door and again found it stuck. I immediately yanked forcefully and it opened.

At that instant I realized I had undergone a conditioning. I had learned, after just a single trial, some twenty minutes earlier, to forcefully yank the door. I had learned to struggle.

The swollen door was equivalent to the locked vocal cords, and my tugging was equivalent to the struggle behaviors of stuttering. The fact that my initial struggle had resulted in success, that is, the door had opened, meant that I had been rewarded for my efforts and thus, in the same situation, would likely struggle again. Which I did.

To relate this to stuttering - John, struggling to say his name (pulling on the door knob), after a few moments, says it (gets the door open), and the act of saying it (getting the door open), becomes the reward for the struggle (the stutter) which enables him to say it.

I now began to suspect that all of the stuttering I observed was learned. And all of the variety I had noted was nothing less than an eloquent testimonial to the heterogeneity of human beings learned struggle behaviors against a common core problem, a spasm of the vocal cords. Here, at last, was the stutter reflex. All the struggle events were now viewed as extricatory and learned.

The pieces of the puzzle were beginning to come together. From my knowledge of the anatomy and physiology of the speaking apparatus I knew that there are small nerve endings within the vocal cords that detect tension and send this information to the brain. When the tension in the vocal cords builds to a critical locking threshold, these nerve endings issue a particular pattern of impulses. I now began to understand that it is this particular pattern of nerve impulses reaching the brain that triggers the stutter reflex.

In response to some form of apparent stress (as yet, unspecified), the stutterer locks his vocal cords; it is an inborn reflex. The locking of the vocal cords, in turn, triggers stuttering, a learned reflex. So there are two reflexes - one inborn, the other learned.

I asked a colleague whose specialty was learning psychology if it was possible that so violent and varied a behavior as stuttering could really be learned. His response was to show me a film of learned self mutilations and to point out that most physically aggressive behaviors are culturally acquired. The struggles I was seeing were mild by comparison. Any behavior that was rewarded could be learned, and the act of speaking which followed the struggle was clearly a sufficient reward to enable the learning of stuttering.

He further pointed out that there was yet another reward commonly present in most human learned behavior and he suspected it to be present for stutterers - anxiety reduction. Anything that would reduce a stutterer's speech anxieties would be rewarded and learned.

I recalled that many stutterers reported that they would often word substitute rather than stutter. As they spoke they mentally looked ahead so that they could shift the conversation away from feared words or change instantly to a word they could pronounce. The successful avoidance of feared words reduced anxiety, was therefore rewarded and learned. Word avoidance was as much of a reflex as the stutter itself. Further, in order to avoid one had to develop the habit of looking ahead and so that, too, became reflexive.

I learned that the stutter reflex was, in fact four learned reflexes. First, the overt stutter; second, word avoidance, third, situation avoidance; and fourth, the habit of looking ahead.