New research is suggesting that one of the reasons people stammer is that they're overly concerned about speaking fluently - they try too hard to monitor their speech for minor errors and are too strict about what is acceptable.
By Robin Lickley, Melanie Russell and Martin Corley.

One of the big questions is why do people stammer? Despite decades of research, there's still no clear answer. One current view of speech production suggests that people who stammer do so because they are constantly trying to avoid ... stammering!

Speaking is a fast and creative process, which is naturally error-prone. We constantly monitor our own speech for errors and we stop, correct the errors and start again whenever we need to. This happens all the time. We can detect errors that we have just spoken, then backtrack to correct them; but we can also anticipate errors and stop before they're actually produced. In stopping and starting, we create dysfluencies (hesitations, reformulations and repetitions). In the case of stammerers, this theory suggests these minor dysfluencies have the same effect as errors - detecting an error leads to dysfluency, which leads to a new error being detected, which leads to more dysfluency ... it's a vicious circle, a hypothesis proposed by two researchers from Utrecht, in the Netherlands, Vasic and Wijnen.

Effort - trying too hard

Monitoring demands a certain amount of attention. The Vicious Circle Hypothesis suggests that people who stammer pay too much attention to monitoring their speech. Therefore, speaking while performing another task should reduce the amount of effort available to put into monitoring and thus reduce stammering.

In the first part of their study, Vasic and Wijnen tested this by asking 22 adults who stammer to describe the content of a newspaper story that they had read previously. They compared dysfluency rates in a speaking-only condition to rates in a dual-task condition, where speakers had to speak and play a simple computer game at the same time.

Stammering was significantly reduced in the dual-task condition - supporting the view that distracting attention away from monitoring reduces dysfluency.

Focus - looking for errors

The Vicious Circle Hypothesis also suggests that people who stammer focus too much on disruption, like minor pauses, both in 'overt' and 'covert' speech. If this is so, then monitoring for something else in one's own speech should also reduce the amount of stammering, by shifting the focus away from the stammer itself.

The second part of Vasic and Wijnen's study tested this by asking people to listen out for the word "die" (Dutch for "the") in their own speech while they retold the newspaper story, pressing a button each time they heard it. The effect was less stammering. In fact, the amount of blocking was reduced with respect to the computer game condition, too. However the amount of word repetition rose compared to all other conditions. But in a control group of people who don't stammer, the word repetition rate also rose. If you imagine doing this task yourself, you can see why. If you stop to press a button every time you say the word "the", your natural inclination will be to stop speaking briefly and return to the (*press button*) ... the beginning of the phrase, thus producing a repetition.

Threshold - being too strict

Another idea of the Vicious Circle Hypothesis is that people who stammer are very strict on what is regarded as acceptable speech, in particular worrying too much about minor disruptions to the flow of speech.

We've tested this hypothesis in an experiment in Edinburgh. We asked people who stammer and people who don't to listen to short samples of speech which either contained short part-word repetitions (you go d-down to the left) or were fluent (you go down to the left). All speech samples were taken from recordings of natural spontaneous speech. The listeners' task was to rate how fluent they considered the samples to be.

We found that people who stammer are indeed more sensitive than other people to dysfluencies in other people's speech: they judged fluent speech as more fluent and dysfluent speech as more dysfluent than did non stammerers.

Our findings fit in nicely with those of Vasic and Wijnen, in support of the Vicious Circle Hypothesis. They found that both reducing the effort applied to monitoring, and focussing the person on something other than error detection, reduced the amount of stammering, while we found that adults who stammer are less accepting of dysfluent speech than others.

What implications are there for therapy? We currently have no suggestions for 'monitoring therapy'. But it is certainly interesting to speculate to what extent various types of intervention serve to engage the speaker in a dual task (like breathing in a special way) and divert attention from the task of self-monitoring.

*Both studies described here are to appear in Hartsuiker, R.J., Bastiaanse, R., Postma, A. & Wijnen, F. (Eds), Phonological encoding in normal and pathological speech. Hove (East Sussex): Psychology Press.

Dr Robin Lickley is a lecturer in the Department of Speech & Language Sciences at Queen Margaret University College, Edinburgh, and he is also a member of the Research Committee of the British Stammering Association.

Melanie Russell (MA) is a student in Speech Pathology and Therapy at Queen Margaret University College.

Dr Martin Corley is a lecturer in the Department of Psychology at the University of Edinburgh.

Reprinted with permission from 'Signal'.

From the Autumn 2002 edition of Speaking Out