Young children learn words rapidly and amongst substantial environmental variation. How they manage to do so, and with relatively consistent results, is the topic of much debate in the developmental literature. Recent research has turned to the environmental variation surrounding children, and the vital information it may hold to help children to learn words more efficiently by way of statistical learning. The responsiveness of caregivers to this variation however – and the subsequent effects of the cues that they provide in real time – remains under-investigated.

The first part of this thesis investigates two key questions: 1) do caregivers alter the cues they provide in response to variation in the environment during word learning, and 2) does both environmental variation and caregiver response affect how their children learn words? The results of Study 1 demonstrate not only that caregiver cue use is dependent upon the amount of variation present, but also that children can learn more effectively in the face of variation. Study 2 then explores how these cues affect learning in real time, addressing the following questions: 1) how do adult learners make use of visual cues in relation to auditory labels as the learning process unfolds temporally, and 2) does interfering with this process affect word learning? This second study shows that it matters most when such cues occur in relation to the given label as the word learning process unfolds in time. These studies use a multi-disciplinary approach (computational modelling, child and adult experimental studies, and eyetracking) to address the multi-factorial process of language acquisition, and show how investigating the interaction of cues with environmental variation
and within-trial learning processes can help us understand how children manage to learn words so consistently.

The multi-factorial process of word learning is then further explored through the lens of atypical language development, and offers a longitudinal perspective of word learning. Whereas part 1 of the thesis addresses receptive ability in cross-sectional studies, part 2 addresses the additional effects that expressive ability have on word learning processes over time. Late talkers are children who are developing typically with the exception of significant expressive language delay, producing fewer words than approximately 90% of their peers. Their unique deficit offers the chance to elucidate the differences between
receptive and expressive language, and to study how language scaffolds development in other domains, such as symbolic understanding of pictures. However, late talking is also problematic: it is a risk factor for Developmental Language Delay, yet late talking children are notoriously heterogenous as a group, making predicting outcomes difficult. Crucially, determining whether or not late talking children utilise word learning mechanisms differently
to typically developing children can provide an evidence-base for predicting outcomes from a clinical perspective.

The second part of this thesis reports a longitudinal study over 2 years in a cohort of late talking and typically developing children. Two research questions are examined: 1) do late talkers show deficits in word learning mechanisms as compared to typically developing children? 2) do late talking children show an impaired understanding of pictorial symbols as a result of their language delay, and how does expressive language affect symbolic understanding more generally?

This longitudinal study is unique in that it takes into account individual variation within the sample, and it also provides further evidence that a multiple hit hypothesis may best reflect the data, where a deficit in one area of ability does not necessarily lead to poor outcomes unless further deficits in other areas are present (i.e. there are multiple hits to language development ability). Study 3 shows that late talking children are impaired in some, but not all, word learning mechanisms; even when late talking children reach typical expressive vocabulary levels, their phonological abilities still lag behind those of their peers and they may struggle to retain statistical information, although certain key receptive abilities remain intact. Study 4 reports that although late talking children show deficits in symbolic understanding of pictures, their development in this domain follows a delayed trajectory, rather than one that is functionally different to typically developing children. The results also indicate that expressive and receptive language skills differentially support symbolic understanding of pictures, mediated by individual variation in social ability.

By examining language acquisition through typical and atypical development, this thesis aims to not only advance understanding of word learning as a process that inevitably involves, and makes use of, variation that exists in a child’s environment, but also examines how expressive language ability – arguably the most clearly observable outcome of word learning for caregivers and early years professionals – interacts with how children come to understand the world around them.