Chapter 1 Overview and aims

Individual differences in language ability are apparent as soon as children start talking, but it is difficult to identify children at risk for language delay or disorder. Recent work suggests word recognition efficiency—that is, how well children map incoming speech to words—may help identify early differences in children’s language trajectories. Children learn spoken language by listening to caregivers, so children who are faster at recognizing words have an advantage for word learning. This view is borne out by some studies suggesting that children who are faster at processing words show greater vocabulary gains months later (e.g., Weisleder & Fernald, 2013).

We do not know, however, how word recognition itself develops over time within a child. This is an important open question because word recognition may provide a key mechanism for understanding how individual differences emerge in word learning and persist into early language development. Without a developmental account of word recognition, we lack the context for understanding individual differences in lexical processing. Thus, even the big-picture questions are unclear: Do early differences persist over time so that faster processors remain relatively fast later in childhood? Or, is such a question ill-posed because the magnitude of the differences among children shrink with age? In this dissertation, I address this gap in knowledge by analyzing three years of word recognition data collected in a recently completed longitudinal study of 160 children.

In particular, I examine the development of familiar word recognition, lexical competition, and fast referent selection (the ability to map novel words to novel objects in the moment). Through these analyses, I develop a fine-grained description of how the dynamics of word recognition change year over year, and I document how differences in word recognition performance relate to other child-level measures (such as vocabulary and speech perception).

Study 1: Familiar word recognition and lexical competition

Specific Aim: To characterize the development of familiar word recognition and lexical competition, I analyze data from a Visual World Paradigm experiment, conducted at age 3, age 4, and age 5.

In these eyetracking experiments, children were presented with four images of familiar objects and heard a prompt to view one of the images. The four images included a target word (e.g., bell), a semantically related word (drum), a phonologically similar word (bee), and an unrelated word (swing). In Chapter 5, I use a series of growth curve analyses to describe how children’s familiar word recognition develops year over year. Children in this cohort cover a range of vocabulary scores at age 3, and this variability allows me to investigate individual differences in vocabulary and word recognition over time and assess the predictive value of these measures. Of interest was how individual differences at age 3 persisted into age 5 and how these differences related to vocabulary measures at later ages. In Chapter 6, I examine the children’s looks to the distractors to study the developmental course of lexical competition from similar sounding and similar meaning words. Increases in sensitivity to competing words reveal how lexical competition effects emerge as a byproduct of learning new words and developing more efficient phonological and lexical representations. As I argue in Chapter 7, increased sensitivity to lexical competitors supports familiar word recognition because children become more efficient at activating a named word and related words. When children err, they become more likely to err on a lexically relevant alternative.

Study 2: Referent selection and mispronunciations

Specific Aim: To characterize how fast referent selection develops longitudinally, I analyze data from a looking-while-listening mispronunciation experiment, conducted at age 3, age 4, and age 5.

Not every word children hear are familiar to them. They may hear entirely new words, or they may hear variations and corruptions of familiar words. How children respond to both kinds of words is informative, as I review in Chapter 9. I describe this eyetracking experiment in detail—based on White and Morgan (2008) and Law and Edwards (2015)—in Chapter 10. Children saw an image of a familiar object and an unfamiliar object, and they heard either a correct production of the familiar object (e.g., soup), a one-feature mispronunciation of the familiar object (shoup), or a novel word unrelated to either image (cheem). The correct productions tested familiar word recognition and the nonwords tested fast referent selection. The mispronunciations tested the child’s phonological categories by showing whether the child permitted, rejected, or equivocated about mispronunciations.

I use growth curve analyses to study how children’s responses to the three word types changed over time. In Chapter 11, I examine familiar word recognition and fast referent selection for novel words to determine which feature of lexical processing better predicts vocabulary growth. I compare these two conditions directly to look for dissociations or asymmetries in these forms of processing within children as a way to empirically assess the claim that “novel word processing (referent selection) is not distinct from familiar word recognition” (McMurray, Horst, & Samuelson, 2012). In Chapter 12, I examine how children interpreted mispronunciations of the familiar words at each age and study how individual differences in vocabulary and speech perception related to children’s responses to the mispronunciations. I also report how children at age 5 are better able to retain nonwords than mispronunciations of familiar words.

Summary

This project investigates how word recognition develops during the preschool years. There has been no published research studying word recognition longitudinally after age two. Furthermore, this project also examines word recognition in two experimental tasks that tap into different aspects of word recognition. Specifically, a four-image experiment with semantic and phonological foils allows me to study how lexical competition develops, and a two-image experiment with nonwords and mispronunciations enables me to study how children’s responses to unfamiliar words develop over time as well. Chapter 15 reviews the results from both studies in terms of lexical processing as well as the main contributions of this project. Findings show how individual differences in lexical processing change over time and reveal how low-level mechanisms underlying word recognition mature longitudinally in children. These findings have translational value by studying processing abilities that subserve word learning and by assessing the predictive relationships between early word recognition ability and later language outcomes.

References

Weisleder, A., & Fernald, A. (2013). Talking to children matters: Early language experience strengthens processing and builds vocabulary. Psychological Science, 24(11), 2143–52. doi:10.1177/0956797613488145

White, K. S., & Morgan, J. L. (2008). Sub-segmental detail in early lexical representations. Journal of Memory and Language, 59(1), 114–132. doi:10.1016/j.jml.2008.03.001

Law, F., II, & Edwards, J. R. (2015). Effects of vocabulary size on online lexical processing by preschoolers. Language Learning and Development, 11(4), 331–355. doi:10.1080/15475441.2014.961066

McMurray, B., Horst, J. S., & Samuelson, L. K. (2012). Word learning emerges from the interaction of online referent selection and slow associative learning. Psychological Review, 119(4), 831–877. doi:10.1037/a0029872