The Acquisition of Brown's 14 Grammatical Morphemes In Children with Autism: A New Look

Background: Brown (1973) found a consistent order in the progress of acquisition of fourteen grammatical morphemes in 3 typically developing (TD) children, subsequently replicated by deVilliers and deVilliers (1973). In contrast, Bartolucci, et al. (1980) and Howlin (1984) found that children with autism (ASD) show a different pattern of acquisition, based on verbal samples taken from 8-year-olds and 10-year-olds.

Objectives: Our study investigates the progress of acquisition of Brown’s grammatical morphemes in children with ASD, as well as the onset of their acquisition in very young children with ASD. 

Methods: We included 18 TD toddlers (MA = 20.6 months at Visit 1, 16 males), and 17 children with ASD (MA = 32.85 months at Visit 1, 16 males); the groups were matched on the CDI at Visit 1 (ASD = 94.1; TD = 118.77). At 6 visits each 4 months apart, children engaged in 30-minute, semi-structured play sessions with their parents, which were transcribed and coded for Brown’s 14 morphemes using CLAN (e.g. progressive   “-ing,” the prepositions in and on, the articles a and the). Because Howlin found no differences between coding for “90% use in obligatory contexts” vs. “percent of correct usage,” we used the latter method.

Results: Thus far, Visits 1, 3 and 5 have been analyzed.  The ASD children were divided into 2 groups based on their joint attention behavior, which was coded separately (Tek et al., 2010).  ASD1 children (n = 8) performed close to TD levels on joint attention whereas ASD2 children (n = 9) engaged in very little joint attention. At Visit 1, both the ASD1 and TD groups produced many morphemes at significantly higher rates than the ASD2 group. Because the ASD1 group was chronologically older, they produced more verb types with the progressive and more article types than the TD group (ps < .05); however, the pattern of morpheme production (i.e., which morphemes were frequent vs. not) was strikingly similar across the TD and ASD1 groups. Interestingly, plurals were more frequent than predicted, probably because object labels were frequently elicited. At Visits 3 and 5, no statistically significant differences were found between ASD1 and TD; however, both of these groups still produced more morphemes than the ASD2 group.  From Visit 3 to Visit 5 the TD group increased in overall morpheme use by an average of 7.69 tokens while the ASD 1 group only increased by 2.81 tokens, which illustrates that although they have a similar pattern at Visit 5, the TD children might surpass the ASD1 group by Visit 6.

Conclusions: Children with ASD who have good joint attention skills do seem to acquire Brown’s 14 grammatical morphemes similarly to TD children, in contrast to ASD children with poor joint attention.  The non-canonical pattern observed by previous researchers may be traced to their inclusion of ASD children who were (a) older and (b) of a wider range of abilities.  Moreover, we conjecture that some differences in Brown’s order of acquisition may be traced to contexts of use.