Is Repetitive Speech a Marker of Development or Impairment?: Correlates of Unconventional Language Use

Background: Children with ASD often use unconventional language, including echolalia, neologisms, idiosyncratic phrasing, and pedantic/formal speech. However, there is a tension in clinical realms: is unconventional language best characterized as functional language or as problematic, repetitive behavior? Speech-language therapy generally emphasizes the communicative value of unconventional language, treating it as a developmental achievement (Charlop, 1983; Tarplee and Barrow, 1999). In contrast, behavioral approaches tend to focus on reducing unconventional language, interpreting it as a problem behavior (Lanovaz & Sladeczek, 2012; Neely, Gerow, Rispoli, Lang, & Pullen, 2016). Despite this marked divide, there is a relative dearth of empirical evidence associating unconventional language with language, cognitive, adaptive, or behavioral characteristics.

Objectives: The objective was to address whether unconventional language is associated with higher or lower scores across measures of language, cognition, adaptive functioning and repetitive behaviors.

Methods: Children were selected from a larger sample (Anderson, Liang, & Lord, 2014; Lord et al., 2006). Sixteen children with high levels of unconventional language at ages 2 and/or 3 (based on the Autism Diagnostic Observation Schedule, ADOS) were identified. Each child was individually matched on gender/nonverbal ratio IQ to a child with no evidence of unconventional language on the ADOS. This resulted in an overall sample of 32 children: 16 with high levels of unconventional language (HiUL), and 16 with low levels of unconventional language (LoUL).

Results: The HiUL group had higher ratio verbal IQ than the LoUL group, t(30)=4.13, p<.001. Similarly, HiUL had higher standard scores for the Vineland Communication [t(30)=4.05, p<.001)] and Socialization domains [t(30)=2.49, p<.05], though no group differences were noted in Vineland Daily Living Skills or Adaptive Behavior Composite. On the ADOS, HiUL had lower social affect algorithm scores – indicating lesser impairment – than LoUL [t(30)=2.42, p<.05]. There was no significant difference in the restricted and repetitive behavior algorithm scores. On the Repetitive Behavior Scale-Revised (RBS-R), the HiUL group had a lower mean score (M=21.80, SD=15.89) than the LoUL group (M=11.67, SD=8.74), though the difference was not significant.

However, we note a potential bias in the ADOS, whereby children could be scored as having low levels of unconventional language because of minimal expressive language overall. Therefore, a second round of matches was generated (HiUL: n=6; LoUL: n=6), including only children who showed established single word speech on the ADOS. Table 1 indicates a consistent pattern of results: HiUL showed higher average skills. Mann Whitney tests were significant (p<.05) for the Vineland Communication and Socialization domain scores and Adaptive Behavior Composite.

Conclusions: These findings indicates that -- compared to LoUL peers -- children with high levels of unconventional language in their toddler/preschool years have higher communication skills and lower levels of social impairment. This has important clinical implications: these preliminary findings support the position that unconventional language is associated with developmental progress and should not be discouraged in therapeutic settings. Additional analyses will address associations between early unconventional language and adolescent outcomes.