Journal of Speech and Hearing Disorders, Volume 52, 393-409, November 1987 I CLINICAL APPLICATION OF TWO PHONOLOGICALLY BASED TREATMENT PROCEDURES ANN A. TYLER MARY LOUISE EDWARDS JOHN H. SAXMAN Syracuse University, Syracuse, NY Two phonological process-based treatment procedures were applied in an ongoing clinical program. Subjects were 4 children aged 3: i, 3:8, 4:1, and 5:1. Two subjects were assigned to a minimal pairs contrasting procedure, and 2 were assigned to a modified cycles procedure based on results of a detailed phonological analysis. All children demonstrated marked changes in their phonological systems as shown by the results of pretreatment and follow-up generalization probes. Correct production generalized to sounds affected by the treatment process that were not a focus of training. Correct production of untrained sounds lagged behind that of trained sounds for all subjects. Results support the hypothesis that articulation remediation is enhanced by treating phonological processes as well as the notion that the acquisition of phonology is a gradual process. Both treatment procedures used in this study were found to be effective and efficient, as evidenced by the elimination of up to three phonological processes within 21/2 months for each subject. Knowledge of normal phonological acquisition has aided in developing an understanding of disordered pho- nology. It is now well accepted that children who exhibit multiple misarticulations generally have phonological systems that are as regular as those of children with normally developing speech (Edwards & Shriberg, 1983; Ingram, 1976). Investigations of normal and disordered phonological systems are frequently accomplished by discovering the child's phonological processes (system- atic sound changes affecting entire classes of sounds or sound sequences). The term phonological process has been adopted from Stampe's (1969) theory of natural phonology, which has been loosely interpreted in the description of functional misarticulations. A phonological process accounts for the child's systematic sound changes that result from the common difficulty a class of sounds or sound sequences poses for the speech capacity (Stampe, 1969). Phonolog- ical process analysis provides a method for identifying children's systematic speech sound error patterns. Using any one of several published or informal procedures (Edwards, 1983; Hodson, 1980; Ingram, 198!; Shriberg & Kwiatkowski, 1980; Weiner, 1979), the phonological pro- eesses that account for the child's misarticulations can be identified. In recent years, several procedures have been described that use the results of phonological process analysis in remediating misartieulations of children with functional artieulation disorders (Crary & Hunt, 1983; Hodson & Paden, 1983; Weiner, 1981). A treatment approach based on phonological process analysis involves facilitating the emergence of new sound patterns through treating phonological processes, as op- posed to treating separate sound errors. The major as- sumption of a phonological process-based treatment ap- proach is that remediation is maximized through gener- alization that oecnrs across the sounds affected by a particular process when only a few sounds undergoing that process are taught. It is assumed that the elimination of a few specific sound errors produces change in the underlying process accounting for those errors. There- fore, other errors arising from the process may also be eliminated without direct training (Compton, 1976). Al- though proponents of a phonological process-based treat- ment approach argue that it is efficient and effective, this has yet to be established in empirical clinical studies. Most phonological process-based approaches: to reme- diation have been described in case studies where signif- icant changes in individual children's phonological sys- tems have been reported (Compton, 1975, 1976; Crary & Hunt, 1983; Elbert, 1983; Grunwell, 1983; Hodson, 1983). Several training studies involving single-subject designs (Elbert & McReynolds, 1985; McReynolds & Elbert, 1981; Weiner, 1981) have also investigated the effects of training and/or generalization from a phonolog- ical process perspective. Elbert and McReynolds (1985) investigated generalization across two sound classes (stops and fricatives ) affected by the process of final consonant deletion. They trained sounds in one class and probed to assess generalization to the untrained class for each of 4 subjects. Their results indicated that correct production of the target pattern generalized only to items of the trained class and not to the untrained class. Weiner (1981) used a multiple baseline design in pro- viding treatment of three target processes for 2 subjects. One of the target processes, final consonant deletion, affected sounds from two different classes; therefore, training sounds for this process were selected from each class. Generalization probes were administered to assess the elimination of the target processes in untrained words. Weiner's results indicated that correct production generalized to untrained words that contained both trained and untrained sounds, resulting in a reduction in the frequency of process application for the probes. McReynolds and Elbert (1981) investigated generaliza- tion for 6 subjects who exhibited the process of cluster reduction, although/s/and liquid clusters were not sep- arated in their definRion of cluster reduction. Untrained clusters were probed to assess generalization. Results © 1987, American Speech-Language-Hearing Association 393 0022-4677/87/5204-0393501.00/0