Background: Research spanning more than two decades has emphasised the lexical deficits of late talkers. However, late talkers have been found to have associated delayed phonological acquisition. Given the close connection between these two linguistic domains, it may be that the late language emergence often observed in these children, arises from deficits in their underlying phonological processing system. This thesis explored the longitudinal relationships between the phonological and lexical development in typically developing toddlers (TD) and those who fit the criteria of late talkers (LT), in light of a psycholinguistic speech processing framework.

Methods: The cohort comprised 168 children aged 2;0 (years; months) at intake who were reassessed when they were about 3;6 and 5;0 years, on measures of phonological accuracy and expressive language. Phonological accuracy (expressed in terms of a percentage of consonants correct) was used as the main behavioural indicator of children‘s phonological development and was measured in two conditions; in a test of nonword repetition (NWR), and a standardised picture naming/articulation test. Children‘s lexical development was assessed using standardised tests of language. Relationships between phonology and expressive language were derived based on correlation and regression analyses of groups‘ scores, as well as in the varied clinical profiles characterised by children‘s abilities in one domain of language relative to the other. With the dataset, analysis of concurrent correlations was conducted in order to identify and compare statistical significance between individual measures of phonological accuracy and the lexicon at each time-point for TD children and LTs. Regression analyses were conducted to identify the proportion of variance in expressive language explained by each measure of phonological accuracy in TD children and LTs. Differences between TD and LT groups in mean scores for phonology and expressive language at each time point were analysed to determine statistical significance.

Results and conclusions: Late talkers‘ performance on a range of measures was significantly different to that of their typically developing peers at all time points. Results indicated that the patterns of individual and combined relationships between phonological accuracy and expressive language also differed between TD and LT children across development. Sufficient phonological representations and motor programs were prerequisites for expressive language development at age 2;0. By age 3;6 and 5;0 continued vocabulary acquisition and expressive language development increasingly relied on their ability to employ phonological units for generating new / nonwords (i.e., the motor programming facility of their speech processing system).

The LTs were found to form a heterogeneous group with varied profiles across development. The emergence of subgroups of LTs and observed shifts in their patterns of phonological relative to expressive language over time, suggested differential underlying deficits in terms of access to different levels of the processing system depending on their phases of development and profiles at different ages. By age 5;0 although the early language difficulties for a majority of LTs resolved, more than half manifested delayed phonological development indicating persistent immaturity in motor programs. The corollary of persisting phonological difficulties in children is that it places them at risk for literacy difficulties at school age. Implications for clinical practice and research were discussed.