ABSTRACT
Polysyllabic words (PSWs), that is, words of three or more syllables, make up a significant proportion of young children's lexicons. Vihman (1996) reported a range from 2% to 27% in children aged 1;0 to 2;0 years from different linguistic backgrounds. Calculations based on Klein's (1981) study indicated an average of 24% of PSWs (predominantly monomorphemic) in the lexicons of five children speaking American English in an age range of 20 to 24 months. However, few tests of speech in citation form include this proportion of PSWs. For example, the proportion of PSWs in The Fisher-Logemann Test of Articulation Competence (Fisher & Logemann, 1971) is 4.5% (5 PSWs). The proportion within The Goldman-Fristoe Test of Articulation (Goldman & Fristoe, 1969) is 9% (4 PSWs). The proportion within the Smit-Hand Articulation and Phonology Evaluation (Smit & Hand, 1997) is 3.8% (3 PSWs). The proportion in the Articulation Survey (Atkin & Fisher, 1996) is 7.6% (5 PSWs). Even the word corpus developed by Grunwell (1987) to account for medial sounds occurring in either syllable-initial or syllable-final position only contains 7% PSWs (14 PSWs). Furthermore, most of the normative data about speech in citation form have been developed from word corpora dominated by monosyllabic words (MSWs) and disyllabic words (DSWs) not PSWs (Ingram, 1976). For example, the American normative data about consonants developed by Smit, Hand, Freilinger, Bernthal and Bird (1990) were developed from a protocol of 81 words, three of which were PSWs. Australian data developed by Kilminster and Laird (1978) and Chirlian and Sharpley (1982) were based on a protocol of 59 words, four of which were PSWs. The normative data developed by Arlt and Goodban (1976) were developed from a protocol of 48 words, one of which was a PSW. The normative data for the Edinburgh Articulation Test (Anthony, Bogle, Ingram & McIsaac, 1971) were developed from 41 words, four of which were PSWs. If the inclusion of significant numbers of PSWs in the normative data results in different, possibly lower norms, then children's speech skills may be overestimated when using tests containing few PSWs.
