Phonological working memory and finiteness marking in typical development

Abstract

The goal of this work was to characterize the maturational trajectory of two core developmental language functions: phonological working memory (PWM) as indexed on nonword repetition (NWR) tasks, and finiteness-related grammatical processing, as indexed by grammaticality judgment. These fundamental language abilities make particularly alluring candidates for investigation due to their central role in language development and academic achievement as well as their theoretical basis. Moreover, PWM and finiteness-marking appear to have genetic bases (e.g., Bishop, Adams, & Norbury, 2006) are powerful markers of language impairment (Conti-Ramsden, Botting, & Faragher, 2001), highlighting the clinical significance of these abilities. PWM, the capacity to temporarily store and flexibly operate on units of auditory information in the service of a goal, is a developmental ability central to language acquisition. As assessed using NWR tasks, PWM has been shown to be instrumental for the development of both spoken and written language (e.g., Baddeley, Gathercole, & Papagano, 1998), and weaknesses in this system are not only associated with problems of language and literacy but are a common correlate of communication deficits in a number of developmental disorders. Despite their demonstrated clinical and theoretical significance, however, there is a surprising paucity of studies examining NWR in a wide, continuous age range including childhood and young adulthood. In the current work, we administered several assessments of NWR to a wide sample of typically-developing children and adults age 5-35 in order to discover the shape of the developmental trajectory of this skill and the age at which proficient levels of performance are achieved. Across several measures varying in the nature and length of the stimuli, the maturational trajectory of NWR was characterized by rapid growth for younger ages which sharply transitions to relatively stable levels. The transitional age between mature and immature performance was found to lie in the 8-11 year old age range, suggesting that NWR ability develops over the later elementary school years. Consistent with prior work, the effect of stimulus length was greater in younger participants, and the tasks involving stimuli which do not resemble real English words were found to be more challenging for all ages than those involving wordlike stimuli. The ability to appropriately mark tense on verbs is also crucial to language development. Children in the Optional Infinitive (01) stage of language acquisition interchangeably use finite (marked for tense and agreement) and non-finite (infinitival) verb forms in clauses requiring finiteness; likewise, both finite and non-finite forms are accepted as grammatically correct in clauses mandating the finite form. Although appropriate use of finiteness has previously been thought to be in place by the time children enter formal schooling (Rice & Wexler, 1996), our recent work (Kovelman et al., 2014) challenged this notion, as linguistically-proficient adults found sentences containing finiteness errors more difficult to process than sentences containing non-developmental agreement errors or grammatically correct sentences. As yet, no one has examined the continuous progression of 01-related processing from early childhood to adulthood using a receptive (or any) measure. In order to discover the shape of the maturational trajectory of finiteness processing and the age at which proficient performance is achieved, in the current work, we administered a grammaticality judgment task involving sentences with developmental errors of finiteness and control grammatical errors to a large sample of typically developing participants age 5-35. Similar to the case of NWR, the shape of the maturational trajectory of finiteness was marked by initially rapid growth transitioning to stable performance. Not only were sentences with errors of finiteness found to be more difficult than those with non-developmental errors, but this condition was characterized by slower developmental growth and an older transition to mature performance than other conditions. Adult-like levels of performance on sentences with finiteness errors were achieved around age 8, suggesting a more protracted developmental course for this ability than previously believed (c.f. Rice, Wexler, & Hershberger, 1998). Taken together, it is hoped that this work will increase our understanding of the developmental trajectories of finiteness-based grammatical processing and PWM. We hope this work will impact early identification of weaknesses in these systems such that appropriate interventions can be implemented.