A Theory of Stuttering
A Theory of Stuttering
Whispering was found to be an effective FC in several studies (e.g., Cherry & Sayers, 1956; Commodore, 1980; Commodore & Cooper, 1978; Johnson & Rosen, 1937; Perkins et al., 1976). In whispering, speech is not vocalized; that is, vocal cords do not vibrate, but articulation modulates the sound of the air flowing through the vocal tract. Whispered speech can be heard by the speaker; it is an altered manner of speaking, associated with unaccustomed auditory feedback. Just as amplified auditory feedback (see Section 3.8), the attenuated auditory feedback during whispering may attract the speaker’s attention more than accustomed auditory feedback does.
Whispering is comparable with incomplete auditory masking by noise insofar as both FCs attenuate auditory feedback. A further FC attenuating auditory feedback is speaking with blocked ears, which limits auditory feedback to the bone-conducted component. The German doctor Sandow (1898) already recommended stutterers plugging their ears with cotton to eliminate stuttering. In all these cases, enhanced fluency was commonly attributed to distraction from auditory feedback, but I think the opposite is true. Attenuated auditory feedback, like other kinds of altered auditory feedback, is an unaccustomed stimulus that attracts the speaker’s attention.
Mouthing, also referred to as lipped speech or pantomime speech, is articulation without (or with extremely low) exhalation, such that the air does not generate audible sound when flowing through the vocal tract. Mouthing reduces stuttering by nearly 100% (Commodore, 1980; Commodore & Cooper, 1978; Hudock et al., 2015; Perkins et al., 1976). Perkins assumed that the lack of phonation simplified motor coordination, which make stutterers fluent when mouthing. This account is not convincing, since other FCs that not reduce but rather increase the demands of motor coordination, such as choral speech, shadowing, or DAF, still reduce stuttering. Difficulty in motor coordination seems thus not to be the cause of stuttering.
Why does no stuttering occur during mouthing? Mouthing is similar to complete auditory masking insofar as, in both FCs, articulation takes place, but one’s speech is not heard—not externally. In several studies, it was found that the speaker “hears” his or her voice and words internally in these conditions (Smith, Reisberg, & Wilson, 1992; Smith, Wilson, & Reisberg, 1995). It may thus be that stutterers are fluent when mouthing or with complete auditory masking not because they do not hear themselves speak, but because they do hear themselves speak internally. The same may be true for cases like those reported by Van Riper (1982, pp. 383–384), in which life-long stuttering disappeared after hearing loss.
Of course, the ability to hear ourselves speak internally did not develop to enable mouthing or the self-monitoring of speech in extreme noise. Its main function is to enable inner speech, namely, when thinking (in words) or reading silently—where stutterers usually do not experience disfluency. Verbal thinking and silent reading have commonly not been regarded as FCs, but I think that the reason why stutterers are fluent when mouthing or under complete auditory masking, on the one hand, and silently reading or thinking, on the other hand, is one and the same.
It is astonishing that inner speech and inner hearing have been ignored in stuttering research, and that there was the belief stutterers would not hear themselves speak when mouthing or when their external auditory feedback was totally masked by white noise. In psycholinguistics, researchers naturally assumed that we hear and monitor our speech internally when external auditory feedback is not available, and they used mouthing and auditory masking to investigate features of inner speech and inner hearing (e.g., Postma & Noordamus, 1996; Oppenheim & Dell, 2010).
Admittedly, the fact that stutterers have no difficulty during silent reading and verbal thinking appeaed trivial: they do “not really” (not overtly) speak. But the difference between inner and overt speech is not as great as it may appear at first glance. There are many similarities between inner and overt speech, and motor control is also involved in inner speech (see Section 1.5. about inner speech), Therefore, the fact that stutterers experience no disfluency during silent reading and verbal thinking is not trivial and calls for explanation.
According to the model proposed by Tian and Poeppel (2010, 2012), inner speech depends on a close coupling between motor and auditory system. Smith, Wilson, and Reisberg, (1995) already found an “inner-ear/inner-voice partnership” to be essential for inner speech; that is, a close coupling between speech production (formulation) and speech perception. The inner voice cannot work without the inner ear; inner speech is speaking and listening, both in one.
In overt speech, by contrast, coupling between production and perception is not essential because overt speech is usually produced for others, not for oneself. Listening to oneself seems task-irrelevant. However, sufficient attention to auditory feedback is necessary for auditory-motor integration, and the findings of Lazzari et al. (2024) suggest that normal speakers cannot ignore their auditory feedback when producing something audible through movement. The same study also revealed that stutterers can ignore their auditory feedback, and the deactivation of auditory association areas (see Table 1, among others, suggests that they ignore it when speaking in normal conditions. This impairs proper auditory-motor integration and causes stuttering.
When mouthing, speaking with complete auditory masking, or after hearing loss, stutterers are not fluent because they don’t hear themselves speak. They do hear themselves speak internally, just as when silently reading or thinking, and they need to pay attention to that internal auditory feedback because it doesn’t work without attention. During overt speech in normal conditions, stutterers can ignore their external auditory feedback, but it is impossible to ignore internal auditory feedback when speaking while external auditory feedback is not available. Internal auditory feedback of speech is the same as the auditory imagination of one’s speech, and to imagine something, in any modality, is impossible without attention to the imagination.
Summarizing the sections 3.10–3.12, we can state that Borden (1979) was wrong with her argument against theories that assume a role of auditory feedback in speaking and stuttering. She did not consider that, when external auditory feedback is not available, we hear and monitor our speech internally. This internal auditory feedback is impossible without attention to the ‘inner voice’; which ensures sufficient attention and proper processing of auditory feedback information and unimpaired auditory-motor integration and prevents stuttering (read more).
The theory proposed on this website assumes that stuttering is caused by insufficient attention to and, due to that, poor processing and integration of (external) auditory feedback. However, the existence of an internal auditory feedback loop ,the use of which prevents stuttering, raises the question: Can that internal auditory feedback compensate for a deficit in the processing of external auditory feedback? If so, the assumption that a problem with the processing of external auditory feedback causes stuttering would be implausible.
Such a compensation is possible only if both, external and internal auditory feedback, are available concurrently. This seems not to be the case. Smith, Reisberg, and Wilson (1992) found that external auditory input interfered with the use of internal auditory feedback. The latter was the more disrupted, the more a concurrent external auditory input was phonologically similar to what should be heard internally. White noise did not impair the inner ear, but externally presented speech stimuli blocked the perception of inner speech. This suggests that internal auditory feedback is not available when one’s own speech is externally heard, since both speech signals are phonologically the same.
This view is supported by Vigliocco and Hartsuiker (2002), who noted that if internal and external auditory feedback worked concurrently, we would hear our speech twice, once by the ‘inner ear’ and once externally. There would be a time lag between the two signals because the two feedback loops take different amounts of time. According to Lackner and Tuller (1979), internal auditory feedback is slower than external auditory feedback.
Finally, the Lee effect suggests that external and internal auditory feedback do not work concurrently. As already mentioned, a delay in the auditory feedback of speech of about 200 ms evokes speech disfluencies in healthy individuals. Obviously, internal auditory feedback cannot compensate for the control problems caused by the DAF. This is not surprising because the external and internal signals are phonologically very similar, despite the delay of the external signal.
This also explains the apparently paradoxical fact that normal speakers are disfluent when their external auditory feedback is delayed by 200 ms, but not when it is completely masked by noise. In the latter case, but not in the former, speech control can shift to internal auditory feedback.
Some FCs, among them choral reading, reduce stuttering even at speech onset where auditory feedback is not yet available. This seems to contradict the present theory. However, the anticipation of hearing a co-speaker’s voice, a metronome beat, or altered auditory feedback may already modulate the allocation of attention. Suppose that stuttering at speech onset is triggered by the anticipation of stuttering (see Section 2.5). Anticipation of stuttering is the anticipation of feeling a speech-motor inhibition; it draws attention inward, to proprioception. In contrast, the expectation of an auditory stimulus, for example, a co-speaker’s voice, draws attention outward, to hearing, and distracts from proprioception and stuttering anticipation.
The idea that expecting to hear a co-speaker’s voice, a metronome rhythm, or one’s own speech in an altered manner draws attention to auditory feedback prior to speech onset gains support from the findings of Daliri and Max (2018). The lack of pre-speech auditory modulation (PSAM) in stutterers, compared to controls, before speech onset suggests that stutterers do not expect auditory input when they start speaking; their attention is not directed to hearing. But the expectation of DAF normalized PSAM in the stutterers; that is, expecting to hear their speech with a delay drew their attention outward, to hearing. What goes for DAF may go for other FCs as well; the expectation of a co-speaker’s voice or another auditory stimulus may draw attention to hearing at speech onset, thereby distracting from proprioception and the anticipation of stuttering.
My assumption that, other than during habitual overt speech, (internal) auditory feedback is properly processed during inner speech is consistent with an astonishing finding reported by Ingham (2001). In normally fluent speakers, as a group, auditory association areas (BA21/22) were activated on the right hemisphere during overt reading. In stutterers, by contrast, those areas were considerably deactivated (compared to resting state) bilaterally during overt (and stuttered) reading (Fig. 2, p. 504 in the study). But during inner speech (by Ingham referred to as “speech imagery”), that auditory deactivation disappeared on the left hemisphere and was reduced on the right. That is, auditory association areas were more active during inner speech than during overt speech in the stutterer group.
Unfortunately, there were two limitations in that study. First, the groups were very small (4 stutterers and 4 controls). Second, the study was not done to examine inner speech in stutterers but to demonstrate that the brain activation associated with stuttering does not depend on overt speech, that is, on muscle movements. That’s why the stuttering participants were asked to imagine stuttered reading. However, since it was merely the imagination of pseudo-stuttering, results would probably have been similar with normal (fluent) silent reading.
(return)