Pragmatic Treatment for Phonological Disability: Rationale and Procedures

Speech-language pathologists have long been concerned with the factors affecting sound production. The most frequent of these factors studied has been phonetic context (Roe and Milisen, 1942; Sayler 1949; Spriestersbach and Curtis, 1951; Curtis and Hardy, 1959; Siegel, Winitz and Conkey, 1963; Prutting, 1970; Gallagher and Shriner, 1975a and 1975b). Phonetic context has been shown to have a major influence on speech sound production. However, phonetic context is only one of many variables affecting sound production. Sound production often varies within repeated attempts to produce the same word. For example, during speech remediation, specific target sounds within words are sometimes produced correctly and at other times incorrectly (Baer and Winitz, 1968). Furthermore, tests of temporal reliability of articulation testing (Winitz, 1963) have also shown changes in articulatory behavior for the same words in two different test administrations. Method of elicitation has also been investigated as a factor affecting articulation (Faircloth and Faircloth, 1970; Jordon, 1960; Hutchinson, 1972, Dubois and Bernthal, 1978). Investigations dealing with method of elicitation have shown that more structured elicitation procedures, such as imitation, result in fewer misarticulations than spontaneous speech samples. Other elicitation procedures, such as naming and sentence recall, have yielded more errors than imitation but fewer errors than spontaneous speech samples. The general focus of research on method of elicitation has been on the amount of structure in the elicitation procedure and on how much of a model is presented to a child. Research evidence suggests that structure provided during elicitation of responses seems to be the most important factor affecting the number of correct responses in a speech sample obtained. Some investigators feel that speech samples obtained under structured conditions are not valid (Shriberg and Kwiatkowski, 1980). They propose that all testing be done during spontaneous speech because these samples are closest to "real world" situations. It is difficult to argue with their claim. However, it is important to remember that methods of eliciting spontaneous speech samples vary from one investigator to the next, the range of samples deriving from talking about pictures to placing a tape recorder in the home of the child and monitoring verbal output during play. Certainly this wide range of sampling should have varying effects on the nature of the speech sample obtained, depending upon various communication requirements of the speakers. The communicative requirement of the speaker, then, is another factor affecting articulatory behavior. In linguistic