How do learners gather new information during wordlearning? We present evidence that adult learners will chooseto receive additional training on object-label associations thatreduce ambiguity about reference during cross-situationalword learning. This ambiguity-reduction strategy is related toimproved test performance. We find mixed evidence thatchildren (4-8 years of age) show a similar preference to seekinformation about words experienced in ambiguous wordlearning situations. In an initial experiment, children did notpreferentially select object-label associations that remainedambiguous during cross-situational word learning. However,this may be explained by some children having relatively highcertainty about object-label associations for which they didnot see evidence disconfirming their initial hypothesis. In asecond experiment that increased the relative ambiguity oftwo sets of novel object-label associations, we found evidencethat children preferentially make selections that reduceambiguity about novel word meanings.

Toddlers are sensitive to the reliability of speakers in their environment \cite{koenig2010}. While previous work suggests that children prefer labels from reliable speakers, it remains unclear how these social representations guide lower-level information-seeking processes that affect speaker preferences. The current study introduces a gaze-contingent eye-tracking paradigm to investigate how children engage in sampling during word learning. Toddlers (22-24m) view videos of two speakers labeling familiar objects; one speaker provides reliable labels and the other speaker provides unreliable labels. Toddlers then sample novel labels from the speakers using a gaze-contingent interface: only the speaker they are fixating on provides a novel label. Preliminary data (N = 18) suggests that participants prefer to sample first from the reliable speaker over the unreliable speaker. However, there is little difference in overall sampling preferences. Our findings suggest that toddlers can assess speaker reliability, but remain open to exploring information from unreliable speakers.

Non-adjacent dependencies are ubiquitous in language, butdifficult to learn. Previous research has shown that the presenceof high variability between dependent items facilitateslearning. Yet what allows learning of non-adjacentdependencies even without high variability in interveningelements? One possibility is that learning non-adjacentdependencies highlights similar structures, allowing people tolearn new non-adjacent dependencies that are otherwisedifficult. In two studies, we show how being exposed tolearnable non-adjacent dependencies can change learners’sensitivity to novel non-adjacent regularities that are moredifficult to detect. These findings demonstrate a new way inwhich learning can build on and shape later learning aboutcomplex linguistic structure.

A primary problem for a child learning her first language is that her ungrammatical utterances are rarely explicitly corrected. It has been argued that this dearth of negative evidence regarding the child's grammatical hypotheses makes it impossible for the child to induce the grammar of the language without substantial innate knowledge of some universal principles common to all natural grammars. However, recent connectionist models of language acquisition have employed a learning technique that circumvents the negative evidence problem. Moreover, this learning strategy is not limited to strictly connectionist architectures. What we call Incremental Distributed Prediction Feedback refers to when the learner simply listens to utterances in its environment and makes internal predictions on-line as to what elements of the grammar are more or less likely to immediately follow the current input. Once that subsequent input is received, those prediction contingencies (essentially, transitional probabilities) are slightly adjusted accordingly. Simulations with artificial grammars demonstrate that this learning strategy is faster and more realistic than depending on infrequent negative feedback to ungrammatical output Incremental Distributed Prediction Feedback allows the learner to produce its own negative evidence from positive examples of the language by comparing incrementally predicted input with actual input.

A critical component of language acquisition is the ability to learn from the information present in the language input. In particular, young language learners would benefit from leaming mechanisms capable of utilizing the myriad statistical cues to linguistic structure available in the input. The present study examines eight-month-old infants' use of statistical cues in discovering word boundaries. Computational models suggest that one of the most useful cues in segmenting words out of continuous speech is distributional information: the detection of consistent orderings of sounds. In this paper, we present results suggesting that eight-month-old infants can in fact make use of the order in which sounds occur to discover word-like sequences. The implications of this early ability to detect statistical information in the language input will be discussed with regard to theoretical issues in the field of language acquisition.

Learning the meanings of words involves not only forming connections between individual words and concepts but also building a network of connections across objects and words. Previous studies reveal that infants and adults can learn word-referent links across multiple ambiguous training instances by tracking the statistical co-occurrence of labels and objects (Smith & Yu, 2008; Dautriche & Chemla, 2014). We asked whether adults are sensitive to multiple types of statistical structure in these learning instances by manipulating the frequency with which objects co-occurred with each other during training trials. Across several studies (n=150), we find that adults not only learned to disambiguate label referents, but simultaneously formed connections both between the frequently co-occurring objects themselves and between the labels of frequently co-occurring objects. These findings indicate that learners exploit statistical regularities to form multiple types of associations during word learning.

Infants must develop both flexibility and constraint in their interpretation of acceptable word forms. The current experiments examined the development of infants' lexical interpretation of non-native variations in pitch contour. Fourteen-, 17-, and 19-month-olds (Experiments 1 and 2, N = 72) heard labels for two novel objects; labels contained the same syllable produced with distinct pitch contours (Mandarin lexical tones). The youngest infants learned the label-object mappings, but the older groups did not, despite being able to discriminate pitch differences in an object-free task (Experiment 3, N = 14). Results indicate that 14-month-olds remain flexible regarding what sounds make meaningful distinctions between words. By 17-19 months, experience with a nontonal native language constrains infants' interpretation of lexical tone.

Williams syndrome (WS) is a rare genetic disorder associated with delays in language and cognitive development. The reasons for the language delay are unknown. Statistical learning is a domain-general mechanism recruited for early language acquisition. In the present study, we investigated whether infants with WS were able to detect the statistical structure in continuous speech. Eighteen 8- to 20-month-olds with WS were familiarized with 2min of a continuous stream of synthesized nonsense words; the statistical structure of the speech was the only cue to word boundaries. They were tested on their ability to discriminate statistically-defined "words" and "part-words" (which crossed word boundaries) in the artificial language. Despite significant cognitive and language delays, infants with WS were able to detect the statistical regularities in the speech stream. These findings suggest that an inability to track the statistical properties of speech is unlikely to be the primary basis for the delays in the onset of language observed in infants with WS. These results provide the first evidence of statistical learning by infants with developmental delays.