We have long been interested in the interactions between fluency and language. For stuttering, there has been continued scrutiny of questions such as how language processing demands affect fluency in children and adults, and whether stuttering onset is somehow related to major developments in children’s language development systems. Some of these issues are surveyed in Bernstein Ratner (1997), a chapter in the textbook The Nature and Treatment of Stuttering: New Directions by Curlee & Siegel. In general, our recent research suggests that at onset, stuttering children have slightly depressed language abilities even as measured by clinical testing devices, which are not meant to detect subtle levels of language processing ability (see Bernstein Ratner & Silverman, Journal of Speech, Language and Hearing Research, in press). Further, when children who have just begun to stutter produce disfluent utterances, such utterances are statistically more likely to contain grammatical errors. That is, all children from about 2-4 years of age have trouble with some aspects of expressive grammar (missing articles, mis-matched helping verbs, tenses, etc.). However, in children who stutter, sentences that are fluent are likely to NOT contain such errors, while sentences with stuttering events are highly likely to contain them, suggesting that it is emerging aspects of grammatical development that stresses fluency in these children. The work we have just completed (Bernstein Ratner & Silverman, in press; Miles & Bernstein Ratner, in review) suggests that parental language interaction styles do not appear to distinguish stuttering children from fluent peers, despite the fact that parents are often asked to adjust the complexity of language addressed to stuttering children.
In our more recent work, we are starting with language disordered children, rather than children who stutter. We reason, from preliminary research conducted by Nancy Hall at the University of Maine, that children who have expressive language problems may not be able to maintain a level of fluency comparable to that of children with normal language development. To date, we have two completed studies that finds interesting patterns of fluency development in these children.
Both sets of studies used a well-known longitudinal database of 32 pairs of children, half of whom were selected at age two because they had normal comprehension but severely delayed expressive language, and half of whom were normally developing children matched to them by age and gender. These children have been followed for almost a decade, and their progress detailed in many publications (Rescorla, 1989, Rescorla & Fechnay, 1996; Rescorla & Bernstein Ratner, 1996; Pharr, Rescorla & Bernstein Ratner, in press).
In the first of the studies, we selected five random pairs of children from the sample at age 3 (Hodge, Rescorla & Bernstein Ratner, ASHA convention 1999). These children were distinct in their expressive language abilities, and demonstrated very different profiles of fluency. In fact, the least disfluent language-impaired child was much more disfluent than the most disfluent language-normal child. More interestingly, the children with language-impairment showed more stutter-like disfluencies (sound, part-word and whole word repetitions, blocks and prolongations), rather than normal-type disfluencies (hesitations, filled pauses, revisions), and the reverse was true of the normally developing children. Finally, there was a very high correlation between the fluency of the child and the fluency of the mother; while we cannot explain this phenomenon right now, the potential for either a genetic language-linked fluency trait or the effects of adult modeling on child speech fluency could exist.
In the second study (Boscolo, Bernstein Ratner & Rescorla, in review), we examined a much larger number of these children five years later, when they were eight years old. In language samples from 22 pairs of children, we duplicated many aspects of Hodge, et al.’s work. While the frequency of disfluencies in both groups by age eight was much lower, the children with histories of language impairment still demonstrated significantly more disfluency, and they were statistically more likely to use stutterlike disfluencies.
Many of us working in the field of fluency are commonly queried about children receiving services for language disorders who "stutter." Typically, their profiles do not include the struggle and associated emotional and cognitive states that accompany true stuttering, and it has been thought that these children were misdiagnosed: they are not stuttering children, merely disfluent language users. Our work confirms this impression, but we note that the higher incidence of stutterlike behaviors in this population could indeed be challenging for many clinicians who have limited experience working with true stuttering.
Continuing work: We are continuing to examine this cohort in its entirety, throughout its followed range (ages 2-11), to ask the following questions:
How does fluency develop in children with normal and atypical language? We note that even in normal children, fluency development has primarily been studied cross-sectionally, rather than longitudinally; this cohort allows us to track fluency development in individual children across a broad span of time.
Is fluency ability linked to language ability?
Do children with poor expressive language skills demonstrate more disfluency, particularly of types more commonly associated with stuttering, than do children with normal language abilities?
Do children with differing levels of language ability experience fluency breakdown at different points in speech production? Our past work with stuttering children (Bernstein, 1981) identified crucial areas of fluency difficulty not experienced by nonstuttering children, such as initiation of verb phrase structures. In the current work we are completing with a larger cohort of the three year old Rescorla children, it appears that children with language production delays experience disfluency at the beginnings of utterances ("trying to get going"), while the language-normal children start utterances fluently, but experience fluency breakdown later on.
Finally, can we bring the continued research on language abilities in stuttering children and fluency abilities in language-impaired children together to help pinpoint better models and testable hypotheses about the underlying nature of the speech production deficiency that produces stuttering?
As usual, I welcome your thoughts and comments, or inquiries about our other research in progress, including the study of children at stuttering onset, parental behaviors and stuttering, and patterns of stuttering in bilingual speakers.