"The trigger for stuttering", by Dr Martin F Schwartz

Dr Schwartz has kindly provided an updated version of his presentation "The trigger for stuttering", and I have great pleasure in placing it:

­­THE TRIGGER FOR STUTTERING

A presentation by

Martin F. Schwartz, Ph.D.

 Stuttering is often defined in terms of its overt symptoms.   Typically these are described as repetitions, prolongations and blocks - although a great many other behaviors have been described as well.  In concert with this definition, most people tend to think of stuttering as a disorder of speech.  As a result, cognitive issues, more specifically, the thoughts of people who stutter, tend to receive much less attention. 

One important cognitive issue for people who stutter is their ability to look ahead and ‘see’ feared sounds, words, and speaking situations approaching.   This anticipatory mental activity, which I here call scanning, leads to a variety of behaviors, one of which is the avoidance of feared words and feared speaking situations.  Clinical experience indicates that scanning and avoiding are not present at the onset of stuttering but develop over time to become habits.  

Scanning and avoiding, however, are not the only habits.  A number of studies have shown that most, if not all, of the overt struggles that typically characterize stuttering are also learned.  Part of this understanding comes from learning-theory models of stuttering and part as a result of the successful use of behavior modification techniques in treatment.

If stuttering is composed of learned struggle, scanning and avoidance behaviors, an adequate definition must include all three.  For purposes of this presentation, let us operationally define stuttering as:

All the things people who stutter do when they get ‘stuck’ in their speech and/or all the things people who stutter do to avoid getting ‘stuck’ in their speech.

Now this definition might appear simplistic, but notice what it accomplishes.  First, it tends to acknowledge and embrace the point of view of the person who stutters, not merely the point of view of an observer.  Second, it takes an exclusive orientation away from a catalog of overt struggle behaviors and provides for the very real possibility of important cognitive components. Finally, it strongly emphasizes that overt stuttering, scanning and avoidance behaviors are responses to something else - the ‘getting stuck’ or, what will be hypothesized here to be, an event, that both initiates and perpetuates what is called:  the trigger for stuttering.    

To begin our quest for the trigger, let us start with a rewording of a number of questions that have been asked repeatedly about overt stuttering. In these questions, the concept of a ‘trigger’ is used in place of the traditional focus upon overt behavioral signs.   So, for example, why does the trigger for stuttering occur more often among males than females, why does it tend to run in families, why is it absent or markedly reduced when a person who stutters talks to themselves out loud alone, or when they sing, or whisper, or speak to the rhythm of a metronome, or speak chorally, or against a masking noise, or when using a foreign accent or when talking to pets?  And why does the trigger typically not occur when, in the course of speech development, a child is using single words to express himself - only to appear suddenly when the child begins using sentences?  And when it does appear, why does it typically manifest itself only at the beginning of sentences or phrases rather than being evenly distributed throughout?   

And then there are interesting questions relating to the trigger raised by the cognitive component.  For example, a child who recently started using sentences suddenly begins to stutter.  In response, he reverts to using single words to convey his ideas.  This child is, by the definition given earlier, still stuttering because he is responding to the awareness of his stuttering by electing to avoid using sentences.  How are we to react to this child? 

And, later in life, if this child, as an adult, masters word substitution and situation avoidance to the point where he never stutters overtly but still lives in constant fear of inadvertently stuttering - are we not to call him a person who stutters? 

These last two questions illustrate some of the issues that arise when a traditional overt-symptom-response-only approach is taken with respect to the understanding and treatment of stuttering.

                                                                                                                                                                                     

Given the above, one is lead to an inevitable conclusion, herein offered as a question:

If overt stuttering, scanning and avoiding are learned responses to the occurrence or anticipation of a trigger, wouldn’t it be better to ignore these learned responses and deal with the trigger or the event causing the trigger instead?

The following example illustrates this alternative way of thinking.  Let us relate overt stuttering to the knee-jerk reflex.  In the knee-jerk reflex, striking the patellar tendon affects the quadriceps muscle.  Muscle spindles in the quadriceps muscle are stretched and send an afferent signal to the brain.  The afferent signal is the trigger.  The striking of the patellar tendon is the event that produces the trigger.  The spinal cord responds to the trigger reflexively by sending an efferent signal to contract the quadriceps muscle and to relax its antagonists, the hamstring muscles.  The foot flies up.  The foot flying movement is equivalent to the stutter.

Imagine, however, that the patellar tendon cannot be seen.  All that can be seen is the foot, and its flying movements are inappropriate, unwanted and cause considerable stress whenever they occur.  Attempts to hold the foot down or to modify its movements (so-called traditional speech therapy) often fail.  One day someone discovers the patellar tendon and observes the temporal relationship between it being struck and the foot-flying event. Knowledge of this relationship quickly leads to a solution.

Now one might argue about the validity of this example, since I am attempting to equate a reflex with a learned behavior, but the example still retains its power to explain what is here contended to be a common mistake in therapy (attempting to treat stuttering instead of its trigger and/or the event initiating the trigger). It also explains a similar and common mistake made by researchers who try to uncover the core brain mechanisms underlying stuttering by exhaustively mapping neuro-physiological correlates of learned behavior, rather than the event that is triggering that behavior.  

There is another common mistake that is often made.  People who stutter often believe ‘stress‘ to be the trigger for their stuttering.  To them, it is a simple stimulus-response event.  For example, a feared speaking situation, such as talking on the telephone, elicits a high degree of anticipatory stress (a cognitive response conditioned by prior experience) that, in turn, elicits a stutter.  The logical assumption, derived from this experience, is that anticipatory stress causes stuttering.  A closed loop cycle is proposed, with anticipatory stress triggering stuttering which, in turn, reinforces the anticipatory stress.  In other words, a self-fulfilling prophecy.

But both psychotherapy and traditional speech therapy have long histories of not being very successful in the treatment of chronic stuttering.  Both disciplines often rationalize their relative lack of success by saying that since stuttering is a complex, multi-factorial problem, starting early in life and of long-standing duration, it is highly resistant to permanent extinction.  Rarely, if ever, is there mention made of the possibility of an intervening step, such as a triggering event, that might be going unaddressed.  

It is the contention of this presentation that there is a trigger for stuttering and that it always occurs during the time between the occurrence of stress and an overt stutter or, in the case of an avoidance response, it is a cognitive event, the perceived threat of a trigger, which occurs in response to stress, that evokes the avoidance behavior.

The origin of the trigger

In our quest for the origin of the trigger, we examine the beginnings of stuttering.  We do so for one reason: unencumbered by the overlay of years of conditioning, important clues to the trigger reveal themselves more readily at this early stage.

We begin with general observations about the onset of stuttering.  The first is that most stuttering begins between the ages of two and five, with a major peak of onset of occurrence between two and a half and three and a half years of age.  The second is that most children who stutter have been speaking fluently for a period of time before their stuttering begins.  The average length of time the child has been speaking fluently before the onset of stuttering is seventeen months, but it can vary widely.  Third, many children who stutter can have periods when they are totally fluent.  These periods, sometimes called remissions, can last for days, weeks, months, years or, in the case of outgrowth, forever.  Fourth, at the onset, most stuttering appears at the beginnings of phrases or sentences.  And fifth, by far, the most common initial expressions of stuttering are repetitions.  

Given these observations, it is obvious that the trigger, whatever it is, is not occurring all or even most of the time.  And if the trigger is a response to stress, then the stress apparently comes and goes quickly, and often without apparent rhyme or reason.  It is strange to think, for example, that a stress and its associated trigger could be present on the first word of a sentence and not on the rest of the sentence, or that Johnny could be overheard talking fluently with his action figures when playing alone in the family room, only to stutter instantly when someone comes into the room and asks him a question.  

If we think for a moment about the concept of a trigger we see that for one to occur, a threshold of some sort must be crossed. Whether it be the response of a thermostat to a certain temperature, triggering the turning on of a furnace or an air conditioner, or an ion channel in a cell's membrane, shut when the potential is near its resting level, only to open suddenly when the potential increases to a precisely defined level, the concept of a trigger always implies the concept of a threshold.  

For the purposes of the presentation today, the threshold for the trigger for stuttering is here defined as:

A level of physiological activity capable of eliciting an overt stutter or of provoking an avoidance response.  

When the magnitude of the physiological activity crosses the threshold, an overt stutter occurs; when it approaches the threshold, its threat potential, as a result of prior conditioning, can initiate an avoidance response - but only in children and adults who have learned to scan.  

At the onset of stuttering, as indicated earlier, there is no scanning.  Anticipatory stress, as a learned response, takes time to develop.   This means that at the onset of stuttering there is no attempt to avoid supra-threshold trigger events.  And since, without scanning, the trigger events are essentially all unanticipated, they evoke what are probably the most basic of responses.

It is a contention of this presentation that the most basic of responses are ones that occur whenever there is a sudden, unanticipated stoppage of a learned and generally smoothly functioning forward progression of preprogrammed neuromuscular events.

So, for example, both speech and walking are preprogrammed forward-movement processes.  Both have beginnings, midpoints, and endings.  Walking and speaking are both learned activities and both have central nervous system mechanisms that plan, execute and monitor the respective behaviors.  Once learned, both behaviors appear to be automatic and the users of such behaviors fully expect the ongoing nature of the behaviors to proceed smoothly and without interruption.

Interesting things happen when an unanticipated interruption occurs to a normally smoothly functioning ongoing neuromuscular activity.  Let me give you a personal example:

A number of years ago I was traveling by train from New York to Washington, DC.  I was sitting in a coach next to the dining car.  As people passed through the coach on their way to the dining car, they were required to press a bar on the door at the end of the coach that would cause the door to slide open.  Sitting near that door, I noticed that something was wrong with the bar’s actuating mechanism, for when the bar was pressed, the door did not open immediately, as expected, but only after a delay of several seconds.  I watched how people reacted to this sudden, unexpected stoppage of a normal-ongoing-forward-movement process (the progression of walking to the door, pressing the bar, expecting the door to open immediately and then continuing to walk).

There were generally two types of responses.  First, and by far the most frequent, were those individuals who repeatedly pressed the bar in an effort to get the door to open.  Call these actions repetitions.  The second, much less numerous than the first, were those who applied pressure continuously to the bar in the effort to open the door.  Call these actions prolongations.  Both actions were in response to the sudden stoppage of movement cause by the temporarily stuck door. Call this sudden stoppage a block. 

    

Here, I thought, was an obvious and efficient explanation for the beginning of developmental stuttering.  As indicated earlier, children had been speaking for a while prior to the onset of stuttering.  The learning process had progressed without undue incident.  Speech was becoming smoother and more automatic each day.  Suddenly, in the midst of this ongoing movement process, the progress was abruptly and unexpectedly halted.  (The equivalent to the ‘getting stuck’ part of the definition – in other words, the block.)  The block, for purposes of this presentation is here defined as the core behavior of stuttering.

The reaction to this event was as non-thinking, automatic and similar as the behaviors exhibited by my fellow travelers on the train that day: mostly repetitions with occasional prolongations or simply the core behavior itself, the blocks.  Interestingly, research on the earliest stuttering signs show this to be precisely the case, with the onset typically being mostly repetitions, with occasional prolongations or, in the absence of those, the block itself.

One final point to the story.  One of my fellow travelers that day, a young man in his late teens, dressed in what I would call ‘Elvis’ attire, had occasion to go to the dining car a second time. He was hard to miss.  I noticed that as he approached the bar, almost without breaking stride, he lifted his foot and gave the bar a forceful kick.  One may conclude from this that on the basis of apparently just a single prior experience (one-trial learning), this young man had devised a strategy for dealing with his anticipation of a block. Here, I thought, was the analog for quickly-learned severe, preemptive (anticipatory) stuttering.

Early responses to repetitions

We shall now examine some of the early behaviors developed by children to deal with repetitions since repetitions are, by far, the most common initial expressions of stuttering.  Note that repetitions can persist for variable periods of time (from a day to years) in unmodified form.  This is due to the fact that since they often work to release the ‘stuck’ condition, the block, they are rewarded and thus learned. 

Eventually, however, a child might react to the repetitions, find them unacceptable, and seek to eliminate them.  A number of options are available.  He can, if he scans, choose to eliminate the repetitions by looking ahead and changing feared words and avoiding feared speaking situations.  Or he might confront the ‘getting stuck’ head on with forceful struggles - as ‘Elvis’ did on the train.  Or he might replace the repetitions with other, seemingly less obvious struggle behaviors. It is this last group that interests us as we again look to certain behaviors that appear early on in the development of stuttering but may not persist into the adult form of the disorder.

There are three of note.  The first is swallowing.  Children who stutter will sometimes swallow just before speaking. It is as if the process of swallowing somehow facilitates the onset of speech.  If one asks a bright 7 year-old who has been exhibiting this behavior for a while the following question: “Does swallowing always work?”  They will sometimes smile, revealing that they fully understand the question and usually answer: “not always”.  To which, if one asks, “Does it sometimes take 2 or 3 swallows to get things going?”  They will often say “yes”.  So swallowing, when it occurs, is usually a conscious anticipatory behavior designed to either disable or prevent a block from occurring.

A second behavior is the use of inspiratory gestures.  Brief, deep, rapid inhalations just before the onset of speech.  Sometimes these behaviors occur as a series of sniffs, sometimes as deep oral inhalations.  Again, they appear just before the beginnings of sentences and appear to facilitate speech onsets.  They are less prominent than swallowing and are often seemingly employed without conscious awareness. 

The third behavior is speaking on expiratory reserve volume or what is commonly called: supplemental air.  The child will appear to exhale almost completely and then, on the minimum amount of air left in the lungs, produce as many words as he can.  Again, this is apparently used to initiate speech.  Owing to the forceful nature of this respiratory maneuver, the child is often aware of using it.

Notice that since swallowing, rapid, deep inspirations, and speaking on supplemental air are techniques employed by these children in the silence before speech begins, it would appear they are being used to prevent or disable a block.  This interpretation gains credence by the fact that when these behaviors are employed stuttering symptoms do not generally appear. 

The silence before speech

Since the trigger occurs in response to stress and both apparently occur in the silence before speech, it would seem reasonable to consider what events might be taking place during this time period that could be interpreted as stressful.  There are several obvious candidates.  To begin, if the child is speaking in sentences, as they typically are when stuttering begins, there are a number of choices that have to be made before speech can begin.  The vocabulary, grammar, syntax, intonation and linguistic stress of the entire sentence have to be selected, integrated and sequenced.   And these activities, and the neurological processes that sub-serve them, have to be made in the fractions of a second and in a brain that, on average, is two and a half to three and a half years old.  

It is a contention of this presentation that these complex and diverse activities, occurring quickly and simultaneously in a relatively immature central nervous system, constitute multiple stresses that express themselves as heightened levels of muscle tension in the vocal tract during the silence before speech begins. 

In predisposed individuals, these heightened levels of muscle tension can cross a threshold yielding a potentially speech-arresting condition. The repetitions that then typically ensue are not ‘problems’ saying the first word or syllable; they are, instead, the normal responses to a sudden, unanticipated stoppage of a learned and relatively smoothly functioning forward progression of preprogrammed neuromuscular activities.  They are the repeated pressing of the bar required to open the ‘stuck door’ at the end of the railway car.

Other sources of ‘stress’ in the silence before speech

There are three other less obvious sources of ‘stress’ that occur during the silence before speech.  The first is called: coarticulation.   Co-articulation, as you probably well know, is the influence one sound has upon another during ongoing speech.  For example, in the word ‘queen’ [kwin], notice that before the [k] is spoken, the lips are already rounded for the [w].  This occurs because the [k] requires no lip involvement and so the lip rounding for the [w] is free to move forward in time to facilitate the smooth transition from the [k] to the [w].  This happens continuously and is a normal feature of speech production.

But co-articulation can be more extensive.  It has been shown, for example, that given the correct phonetic environment, the effect of the articulatory requirements for the last word of a sentence can sometimes be detected on or before the first word of that sentence.  This means that continuous speech is not merely a concatenation of one word independently following another in time, but instead, prior to the onset of the sentence, that is, during the silence that precedes it, portions of the articulatory organization for the entire sentence can take place.

The second not-so-obvious event occurring in the silence before speech is the ‘stress’ associated with speaking too rapidly.  More specifically, it is the speed of the first word.  It is called speed stress and it manifests itself in the following manner:  As the length of a sentence is increased, the duration of each of the syllables in the sentence decreases.   This means, as the sentence gets longer, the speaker talks more rapidly.  And although this is true throughout the sentence, nowhere is it apparently more prominent than on the first word of the sentence.   So the speed of the first word is particularly sensitive to the number of words in the sentence. 

Let us consider what happens as the sentence is lengthened and the first word is spoken more rapidly.  To understand this, imagine we are in a physiology lab and are going to study the behavior of the triceps muscle as it functions in an arm-extension maneuver.  Surface electrodes have been attached to the muscle to record its electrical activity.  The three conditions to be studied are a slow-speed extension of the arm, a moderate-speed extension, and a rapid one.  The activity to be examined in detail is the behavior of the triceps muscle a fifth of a second before the movement begins. 

The test is completed and the data collected and analyzed.  Not surprisingly, one finds the greatest degree of electrical activity, that is, the most tension on the triceps, occurs before the rapid extension, a lesser degree of tension before the moderate-speed extension, and the least before the slowest-speed extension.

The basic principle of physiology demonstrated is that in order to have progressively greater initial velocities, one must recruit progressively more muscle fibers before the movement begins and that this recruitment is reflected in the greater average tensions recorded prior to the onset of the faster extensions.

The same is true during the silence before a sentence versus the silence before a word spoken in isolation.  The tension is always greater before the sentence and, within limits, the longer the sentence, the greater the pre-speech tension.  Research shows that the longer the sentence the greater the amount of stuttering  and that most early onset stuttering occurs on the first word of phrases or sentences.   

Stocker developed a treatment for young children based on this understanding.   Her approach was to have children return to using single word utterances to express what formerly had been expressed as stuttered sentences.  This elemental form of speech, relieved of a great deal of its pre-speech linguistic programming and most of its speed stress, was fluent.  Then, gradually, as fluency became more firmly established, she proceeded to two-word utterances, three-word utterances and so on - guiding the child through to progressively greater levels of linguistic complexity and speed stress.

The third event of importance is called the pre-phonatory tuning of the larynx.  There are two types.  In one type, if speech is to start with the correct pitch and loudness, the vocal folds must assume specific postures and tensions before speech.  Were pre-phonatory tuning of the larynx not present, speech onsets would be marked by rapid shifts of vocal fold activity, as cortical control of the larynx, making use of auditory feedback, hunted to find the right initial pitch and loudness.  A moment’s reflection indicates this is not the case and a simple example convincingly makes the point:  If a person is asked to speak in falsetto, they know, a priori, using various proprioceptors, they can do it.  There is no need for the person to hunt for falsetto voice after phonation has begun. 

 

_­

The second type is the position required of the vocal cords for the first sound in an utterance.  As is well known, it is obviously not always the case that the vocal cords come together and touch one another just before speech.  Quite often the opposite is true.  A little less than half of the consonants in English are voiceless, that is, they require the vocal cords to be apart.  So, for example, as we all know, the voiceless ‘p’, ‘t’, and ‘k’ sounds are produced with the vocal cords apart, while their voiced articulatory counterparts, ‘b’, ‘d’ and ‘g’, are produced with the vocal cords together.

The base level tension

All of the above tensions are superimposed upon a fluctuating ‘base level tension.’  The ‘base level tension’ is defined as the tension on the vocal cords when a person is not speaking and not intending to speak.  The sources of this tension are both systemic and psychological.  Some of the systemic ones are: fatigue, nutritional deficiencies, allergies, illnesses, drugs, alcohol, hormone fluctuations (body and brain) and individual reactions to environmental conditions (temperature, humidity, positive ions, pollutants, etc.).  The psychological ones can be a ‘mean’ teacher, a bully, a divorce, loss of a grandparent, a job or a pet – a whole host of events. 

The combined strength of the various systemic and psychological sources, at each moment in time, contributes to the ‘base level tension’.  The sources can be in phase with each other or mostly negative and thus add to one another (producing what are called ‘bad’ speaking days, weeks, months, etc) or out of phase with each other and cancelling one another or mostly positive (producing what are called ‘good’ speaking days, weeks, months, etc).  Thus the variability in the severity of disorder, which is well documented, has to do with shifting base level tensions.

Marshaling the evidence in support of the site of the trigger

There are thus at least ten separate events taking place simultaneously in the silence before phrases or sentences.  They are: the selection of vocabulary, grammar, syntax, intonation and linguistic stress, the effects: of co-articulation, speed of the first word, and the two forms of pre-phonatory tuning of the larynx: correct pitch and loudness at onset and correct voiced/voiceless distinction at onset - and all superimposed upon a fluctuating base level tension.

  

Managing these events correctly represents a substantial undertaking when viewed within the constraints of the time available and the immaturity of the nervous system at the typical age of the onset of stuttering.  Clearly this creates a multiple-stress situation and, as indicated earlier, the most common expression of this stress is an increase in muscle tension.

The question remains, where is the most likely site of the tension?  The purpose of this presentation is to suggest that the answer lies at the narrowest aperture along the vocal tract from lungs to lips: the space between the two vocal cords.  A small increase in muscle tension at this location can, given its crucial role in so many of these activities, create more difficulty in speech than an increase in muscle tension at any other location.  If the critical source of tension were elsewhere, let us say, the lips, the tongue, or the soft palate, the primary initial signs of stuttering, repetitions, would show articulation and/or resonance defects indicative of struggle at these areas, but they do not.  On the contrary, most clinicians often describe the very early onset repetitions as seemingly ‘effortless.’

It is thus the contention of this talk that the pre-speech focus of stress-induced muscle tension is likely at the larynx. When the tension level at the larynx reaches a threshold, the vocal cords become essentially immobilized in either an open or closed position, depending upon whether the intended first sound is voiced or voiceless.  To repeat, this state of vocal-cord immobilization, commonly called a block, is felt to be the core of stuttering.

Swallowing, rapid, deep inspirations, and speaking on supplemental air - all prevent repetitions and prolongations and other learned manifestations of stuttering because they interfere with the buildup of tension leading to the threshold of laryngeal blocks - but they do so in different ways.  For example, in the case of an abrupt, deep, inspiratory gesture, a rapid and large laryngeal opening is produced by a reflexive response of two muscles: the posterior cricoarytenoid and cricothyroid which contract forcefully to accommodate the greater inflow of air.   This command from the brain to these muscles supersedes the intent to speak, with the result that most of the sources of pre-speech tension are temporarily suspended.

  

Speaking on supplemental air (just the opposite of the previous approach) also accomplishes the same thing, but in a more involved way.  On the one hand, sensory receptors in the lungs, diaphragm, muscles between the ribs and the muscles that line the abdominal wall detect the amount of air left in the lungs, sense that it is dropping dramatically below the levels required for normal breathing and begin to prepare the person for an imminent corrective inhalation by initiating a major opening of the vocal folds.  At the same time, muscles that close the vocal cords are at work trying to get sound going.   This conflicting activity interferes with the intent to speak by temporarily suspending most of the normally present pre-speech laryngeal tensions.

Finally, at the end of the swallow reflex, after transport of the bolus of food to the esophagus has taken place, with its concomitant transient deglutition apnea, the voice box drops abruptly and the vocal folds mechanically open widely to enable resumption of respiration.  This behavior, as part of the swallow reflex, like the others, interferes with the preparations that must be made to get ready to speak. 

Swallowing, rapid inspirations, and speaking on supplemental air all engender alterations of the tensions within and around the larynx just prior to the onset of speech that are both independent of and, to a certain extent, incompatible with the speech that follows. The three behaviors are not struggles in speech; they are, instead, activities before speech that temporarily substitute basic, reflexive mechanisms for neurologically higher-order speech commands to the larynx.  Said somewhat differently, they all temporarily suspend the intent to speak and, in so doing, alter most of the pre-speech tensions on the vocal cords.

Further evidence in support of the larynx as the site of the trigger

There are a whole host of other observations that further implicate the larynx as the site of the trigger.  Here are just a few: Occasionally, in the effort to speak without stuttering, a child will resort to whispering.  The altered physiology of the vocal cords during whispering to a more open state would support the role of the larynx as the site of the trigger.   The use of the electro larynx (a device used by laryngectomees, people who have lost their voice boxes to cancer, to produce voice when they cannot produce esophageal speech) has elicited fluent speech when used by non-laryngectomized individuals who stutter.    Changing the pitch of the voice, and thus the internal and external laryngeal tensions and postures, has produced similar effects.   Mouthing words fluently with no attempt at phonation further supports the contention.   Fluent esophageal speech following total laryngectomy with individuals who had stuttered preoperatively has been reported.    The therapeutic use of gentle onsets of phonation to ease into vocal fold vibration also is corroborative.  And the research that shows that individuals who stutter have longer Voice Onset Times (VOT’s) than people who don’t suggests an inherent state of neural-based inertia at the vocal folds.   All of this, admittedly circumstantial, but all, nevertheless, suggestive of a laryngeal involvement.

There is yet a further intriguing piece of evidence that points to the larynx as the site for the trigger.  One of the most common activities for lessening or interfering with the pre-phonatory tensions in and around the larynx is the use of verbal starters.  A starter may be a sound, a syllable, a word or words (real or nonsense) or a phrase that has no contextual linguistic function and is used before the onset of speech to get the vocal cords vibrating in the expectation that they will continue to vibrate directly into and through the feared utterance.  Verbal starters have a reputation for working initially and then failing.  The reason this happens is that after a while, the starter becomes so automatic it becomes incorporated into the overall plan for the sentence.  Said somewhat differently, verbal starters stop working when they begin to assume the role of the first word and start to co-articulate with the rest of the sentence.  It is at this point that the starter becomes part of the intent to speak and thus acquires many of the sources of pre-speech tension that lead to a block.

The trigger for stuttering

It is the contention of this presentation that the trigger for stuttering is a unique pattern of afferent impulses that arise from receptors within and around the larynx that signal the central nervous system that a threshold level for a specific concordance of speech-arresting laryngeal tensions has been reached or soon will be reached.  It is this unique afferent array, or its threatened occurrence, that constitutes the trigger for the learned anticipatory or extricatory behaviors called stuttering. 

Change the afferent array and the conditioned stutter responses cease.  The most efficient way for making changes to the array is to subtract laryngeal tensions or to create new ones, ones that have no cue value.  So, for example, singing, whispering, speaking to the rhythm of a metronome, shouting, swallowing, rapidly inspiring, speaking on residual air, head jerks, coughing (throat clearing), stress-reduction techniques, starters, changing words, etc. all have the capability of changing the afferent array, as do all activities that distract attention away from speech, have no conditioned anticipatory stresses associated with them (like speaking alone or with a foreign accent), and any or all of the activities that reduce linguistic and/or speed demands. 

The Basal Ganglia

Recent research suggests that children who do not stutter are subject to the same linguistic and speech stresses as children who do, but that what separates the two groups may be a genetically-based predisposition for excessive tension to be developed at the larynx in stuttering children.   Current findings implicate the basal ganglia, and its tendency to respond inappropriately when under conditions of stress - as the source of this predisposition.  The reasons for this are not entirely known, and it may, for many, be nothing more than a temporal delay that eventually catches up for most and is not found in adults (this may account for why a substantial majority of children who stutter spontaneously outgrow it), but attempts to understand the mechanism are currently underway.   We do know, however, that since the basal ganglia are critically involved in memory, cognition, learning, sequencing of complex motor commands and have direct connections with the limbic system and projections to and from a laryngeal area on the motor cortex, the basal ganglia, more than any other central nervous system structure, has the potential of being the source of the tensions that ultimately produce the trigger for stuttering.

A final overview

We can summarize this talk by positing a sequence of events that constitutes a model for stuttering based on the elucidation of a trigger for the disorder.  The events are as follows:

In a neurologically-predisposed individual, diverse endogenous and exogenous stresses occur simultaneously to result in a buildup of tension at the larynx that reaches speech-arresting levels in the silence before speech.   A state of vocal-fold immobilization, commonly called a block, then exists. This block constitutes the core of stuttering. 

The most typical immediate response to a block is a repetition.  Repetitions can occur with varying degrees of tension and frequency and often become habits.  After a period of time, which can vary greatly, the child can become aware of and react to the repetitions by engaging in one or more of a number of activities.  These can be avoidance behaviors, direct, forceful confrontational attacks, or more subtle maneuvers such as swallowing, inhaling rapidly and deeply, speaking on expiratory reserve volume or using starters. 

All of these behaviors, with the exception of the core behavior, the laryngeal block, are learned.  The core behavior is represented by a unique afferent array that, through conditioning, becomes the trigger for all of the subsequent learned behaviors associated with it.  Any activity that alters this unique afferent array can prevent the conditioned stutter responses from occurring.    

Concluding remarks

Textbooks on stuttering provide the student with collections of published data. Examining these texts, and studying how they are organized, one discovers a thread common to them all:  Stuttering, in spite of a great deal of research, remains for many a complex, often bewildering, frequently enigmatic and frustratingly confusing disorder. 

A large part of this difficulty stems from a long-standing tradition by both researchers and clinicians to focus on the overt signs of school age and adult stuttering populations.  As indicated earlier, the myriad and heterogeneous accumulations of overlaid struggle behaviors associated with the chronic version of this disorder overwhelm and obscure the underlying core mechanisms.  Admittedly, these struggle behaviors are seductive - and therein lies the trap, for much of the published research on stuttering can be said to be essentially well-designed studies of whether ‘the foot flies to the right, left or straight ahead when the tendon is struck.’ 

It is hoped that awareness of the trigger will prove useful for those seeking to find more productive ways of ordering the knowledge database of stuttering and, by extension, to provide guidance for those seeking more effective methods of treatment.