A Theory of Stuttering
A Theory of Stuttering
Any theory that claims stuttering is caused by a mismatch between auditory prediction and auditory feedback, irrespective of whether prediction or feedback is the problem, has to explain how stuttering can occur at the very beginning of an utterance where auditory feedback is not yet available. Below, I will propose two explanations: one for symptoms at speech onset in early childhood stuttering, and one for blocks at the first syllable of the first word in later stuttering.
Disfluencies at speech onset are frequent in preschoolers who stutter (Buhr & Zebrowski, 2009; Richels et al., 2010). Bloodstein (2006) writes thst “early stuttering occurs regularly on the first word of a syntactic structure, most often a sentence, clause, or verb phrase” (p. 186). He further (p 187) writes: “A distinctive aspect of early stuttering is that much of it consists of whole-word repetitions (Ambrose & Yairi, 1999; Bloodstein 1960, 1974; Johnson & Associates, 1959).”
Other than a silent block occurring before any sound was produced, the repetition of the first word of an utterance can be caused by a mismatch between auditory prediction and auditory feedback. If the auditory feedback of the first word is incomplete or noisy because of poor feedback processing, the monitoring system may misinterpret this as a speech error and trigger a ‘correct’ repetition of that word, until prediction and feedback match, that is, until the allocation of attention is adapted so that auditory feedback is sufficiently processed (read more).
Hearing the word repetitions may support this adaptation of the attention system by drawing the child’s attention to the auditory channel. This may explain the fact that many children overcome that period of disfluent speech after a short time: stuttering itself—the repetition of whole words—may help their attention system adapt to the new requirements of connected speech.
A further factor supporting spontaneous recovery is that the affected children learn to master syntax and grammar increasingly well. Bloodstein (2006) was right when he attributed the onset of childhood stuttering at the age of about three years to uncertainty in sentence formation. But unlike him, I don’t think that word repetitions at the beginning of sentences or clauses are “hesitations” due to uncertainty. Hesitation is a conscious behavior, something a person does. By contrast, stuttering is something that happens to a person, entirely against his or her will.
Therefore, I propose that uncertainty in sentence planning causes the child to pay much attention to it at the beginning of an utterance; this reduces the capacity for the processing of auditory feedback, as the allocation of attention is the allocation of perceptual and processing capacity. It is thus not uncertainty in sentence planning itself that causes early stuttering, but the excess of attention to sentence planning due to that uncertainty.
In my observation, blocks, prolongations, and sound repetitions on the first syllable of the first word are rare in adults, but if they occur, a possible explanation could be as follows. Originally caused by poor processing of auditory feedback, a mechanism of persistent stuttering may have developed that can secondarily also be triggered by the anticipation of stuttering. This could also happen at speech onset, where, according to my experience, anticipation of stuttering is particularly likely.
The idea that stuttering may be evoked by anticipation is not new; it was first proposed by Bloodstein (1958). It must, however, be emphasized that the link between the anticipation and the occurrence of stuttering can only have arisen as a secondary symptom, after a child has repeatedly experienced stuttering.
Seen from a theoretical viewpoint, the idea that stuttering at speech onset is caused by anticipation seems to be incoherent with the theory that it is primarily caused by invalid error signals. However, the start of speaking despite the anticipation of stuttering— starting an action despite the expectation that it will fail—may also evoke an error signal in the cerebellum. And recall that error signals due to poor feedback processing are ‘invalid’ only from an observer’s perspective (the speaker made no mistake). There is no difference between valid and invalid error signals in the brain; both have the same inhibitory effect.
If the above explanation is correct, then anticipation is a secondary way to trigger stuttering, not only at the onset of an utterance. This corresponds to not a few stutterers’ experiences. However, many stuttering events occur without anticipation; that is, anticipation cannot be the primary cause of stuttering.
Previously, I had proposed a different explanation on this page, namely, that poor processing of the sensory feedback of breathing may cause stuttering at speech onset. The idea was that inhalation precedes the start of speaking; thus, disrupted feedback of this inhalation phase could cause an error signal, which then inhibits speaking to start. I still think that a misallocation of attention can also impair the integration of the sensory feedback of breathing; this would explain the breathing irregularities observed in stutterers (Zocchi et al., 1990). However, I do no longer believe that impaired feedback of breathing can evoke error signals.
A further idea was that an error signal may occur when the person starts speaking, although the speech network is not yet ready (not sufficiently prepared, particularly regarding the attention system). But this would require the existence of a meta-level, from which the brain could assess its own readiness and that of its networks and other components. I think that there is no such a meta-level; therefore, the brain has no knowledge of network readiness and cannot evaluate the premature start of speaking as an error.
I still think that the speech network is not well-prepared, particularly not adjusted to the processing of auditory feedback in stutterers at speech onset. This is suggested by the lack of pre-speech auditory modulation in adults who stutter (Max & Daliri, 2019). And this may cause stuttering on the second or third word of an utterance, or even on the second syllable of the first word, if that syllable is stressed, but not on the first syllable of the first word.
I think that part-word repetitions and whole-word repetitions slightly differ in terms of their causation. In Section 2.1, I have written that the word following that on which the error signal has occurred is blocked after a reaction time. This can either cause the repetition of the initial sounds of that blocked word or a tonic symptom.
A whole-word repetition may occur if the reaction to an invalid error signal does not yet affect the next following word. That may depend on the speech rate. Young children’s articulation is not fully automated; they don’t speak very fast, and this could be the reason that, responding to an invalid error signal, the “wrong” word is repeated.
According to my theory, the stuttering mechanism is a malfunction, a “perversion” of the normal error repair mechanism, and it is normal to repeat a mispronounced word correctly. Just as we automatically repeat, e.g., a grasping movement that missed its target, we automatically repeat a word if the brain, because of impaired auditory feedback, “believes” that the speech movement missed its target and the word has not been produced correctly.
Whole-word repetitions seem thus to be a mediate stage between normal error response and “fully developed” stuttering. This corresponds to their typical occurrence in early childhood stuttering.
(return)