EXPRESS: Speaker Variability, But Not Bilingualism, Influences Cross-Situational Word Learning.

When learning new words, listeners must contend with multiple sources of ambiguity and variability. Research has revealed that learners can resolve referential ambiguity by tracking co-occurrence statistics between words and their referents across multiple exposures over time - a process termed cross-situational word learning (XSWL). However, the degree to which variability in the input, such as input from multiple speakers, and variability in learner experience, such as bilingual language experience, modulate XSWL remain unclear. In the present study, we examined the effects of speaker variability in cross-situational word learning performance in monolingual adults and bilingual adults with a range of second language backgrounds and language acquisition histories. Results revealed above-chance word learning in both the single and the multiple speaker conditions across language groups. An advantage for word learning was observed in the single speaker condition but the effects of bilingual language experience were null. This research adds to the limited body of work dedicated to extending theories of statistical learning to account for variations in both input and learner characteristics as well as their interactions.

Constraints on novel word learning in heritage speakers.

Introduction: Recent research on word learning has found that adults can rapidly learn novel words by tracking cross-situational statistics, but learning is greatly influenced by the phonological properties of the words and by the native language of the speakers. Mandarin-native speakers could easily pick up novel words with Mandarin tones after a short exposure, but English-native speakers had specific difficulty with the tonal components. It is, however, unclear how much experience with Mandarin is needed to successfully use the tonal cue in word learning. In this study, we explored this question by focusing on the heritage language population, who typically are exposed to the target language at an early age but then develop and switch to another majority language. Specifically, we investigated whether heritage Mandarin speakers residing in an English-speaking region and speaking English as a dominant language would be able to learn novel Mandarin tonal words from statistical tracking. It helps us understand whether early exposure to the target feature is sufficient to promote the use of that feature in word learning later in life. Methods: We trained 30 heritage Mandarin speakers with Mandarin pseudowords via a cross-situational statistical word learning task (CSWL). Results and discussion: Heritage Mandarin speakers were able to learn the pseudowords across multiple situations, but similar-sounding words (i.e., minimal pairs) were more difficult to identify, and words that contrast only in lexical tones (i.e., Mandarin lexical tone) were distinguished at chance level throughout learning. We also collected information about the participants' heritage language (HL) experience and usage. We did not observe a relationship between HL experience/usage and performance in tonal word learning, suggesting that HL exposure does not necessarily lead to an advantage in learning the target language.

Children's subjective uncertainty-driven sampling behaviour.

Are children and adults sensitive to gaps in their knowledge, and do they actively elicit information to resolve such knowledge gaps? In a cross-situational word learning task, we asked 5-year-olds, 6- to 9-year-olds and adults to estimate their knowledge of newly learned word-object associations. We then examined whether participants preferentially sampled objects they reported not knowing the label in order to hear their labels again. We also examined whether such uncertainty-driven sampling behaviour led to improved learning. We found that all age groups were sensitive to gaps in their knowledge of the word-object associations, i.e. were more likely to say they had correctly indicated the label of an object when they were correct, relative to when they were incorrect. Furthermore, 6- to 9-year-olds and adults-but not 5-year-olds-were more likely to sample objects whose labels they reported not knowing. In other words, older children and adults displayed sampling behaviour directed at reducing knowledge gaps and uncertainty, while younger children did not. However, participants who displayed more uncertainty-driven sampling behaviour were not more accurate at test. Our findings underscore the role of uncertainty in driving 6- to 9-year-olds' and adults' sampling behaviour and speak to the mechanisms underlying previously reported performance boosts in active learning.

The effects of speaker and exemplar variability in children's cross-situational word learning.

Cross-situational word learning (XSWL) - children's ability to learn words by tracking co-occurrence statistics of words and their referents over time - has been identified as a fundamental mechanism underlying lexical learning. However, it is unknown whether children can acquire new words when faced with variable input in XSWL paradigms, such as varying object exemplars and variable speakers. In the present study, we examine the separate and combined effects of exemplar and speaker variability on XSWL in typically developing English-speaking monolingual children. Results revealed that variability in speakers and exemplars did not facilitate or hinder XSWL performance. However, input that varied in both speakers and exemplars simultaneously did hinder children's word learning. Results from this work suggest that XSWL mechanisms may support categorization and generalization beyond word-object associations, but that accommodating multiple forms of variable input may incur costs. Overall, this research provides new theoretical insights into how fundamental mechanisms of word learning scale to more complex and naturalistic forms of input.

The role of bilingualism in paired-associate and cross-situational word learning.

In adulthood, novel words are commonly encountered in the context of sequential language learning, and to a lesser extent, when learning a new word in one's native language. Paired-associate (PAL) and cross-situational word learning (CSWL) paradigms have been studied separately, under distinct theoretical umbrellas, limiting the understanding of the mechanisms underlying the learning process in each. We tested 126 monolinguals and 111 bilinguals on PAL and CSWL, manipulating familiarity and measuring verbal working memory. Results revealed highly similar learning performance across groups, both demonstrating better performance in PAL than in CSWL, similar sensitivity to familiarity, and similar reliance on phonological working memory. We observed a trend such that bilinguals outperformed monolinguals in PAL but not in CSWL, but this trend was weak. Findings indicate limited effects of bilingualism on word learning in adulthood and suggest highly similar word learning mechanisms in learners with different linguistic experiences.

What have we learned from 15 years of research on cross-situational word learning? A focused review.

In 2007 and 2008, Yu and Smith published their seminal studies on cross-situational word learning (CSWL) in adults and infants, showing that word-object-mappings can be acquired from distributed statistics despite in-the-moment uncertainty. Since then, the CSWL paradigm has been used extensively to better understand (statistical) word learning in different language learners and under different learning conditions. The goal of this review is to provide an entry-level overview of findings and themes that have emerged in 15 years of research on CSWL across three topic areas (mechanisms of CSWL, CSWL across different learner and task characteristics) and to highlight the questions that remain to be answered.

Speech Segmentation and Cross-Situational Word Learning in Parallel.

Language learners track conditional probabilities to find words in continuous speech and to map words and objects across ambiguous contexts. It remains unclear, however, whether learners can leverage the structure of the linguistic input to do both tasks at the same time. To explore this question, we combined speech segmentation and cross-situational word learning into a single task. In Experiment 1, when adults (N = 60) simultaneously segmented continuous speech and mapped the newly segmented words to objects, they demonstrated better performance than when either task was performed alone. However, when the speech stream had conflicting statistics, participants were able to correctly map words to objects, but were at chance level on speech segmentation. In Experiment 2, we used a more sensitive speech segmentation measure to find that adults (N = 35), exposed to the same conflicting speech stream, correctly identified non-words as such, but were still unable to discriminate between words and part-words. Again, mapping was above chance. Our study suggests that learners can track multiple sources of statistical information to find and map words to objects in noisy environments. It also prompts questions on how to effectively measure the knowledge arising from these learning experiences.

Paired-associate versus cross-situational: How do verbal working memory and word familiarity affect word learning?

Word learning is one of the first steps into language, and vocabulary knowledge predicts reading, speaking, and writing ability. There are several pathways to word learning and little is known about how they differ. Previous research has investigated paired-associate (PAL) and cross-situational word learning (CSWL) separately, limiting the understanding of how the learning process compares across the two. In PAL, the roles of word familiarity and working memory have been thoroughly examined, but these same factors have received very little attention in CSWL. We randomly assigned 126 monolingual adults to PAL or CSWL. In each task, names of 12 novel objects were learned (six familiar words, six unfamiliar words). Logistic mixed-effects models examined whether word-learning paradigm, word type and working memory (measured with a backward digit-span task) predicted learning. Results suggest better learning performance in PAL and on familiar words. Working memory predicted word learning across paradigms, but no interactions were found between any of the predictors. This suggests that PAL is easier than CSWL, likely because of reduced ambiguity between the word and the referent, but that learning across both paradigms is equally enhanced by word familiarity, and similarly supported by working memory.

The effects of bilingualism on children's cross-situational word learning under different variability conditions.

In the current study, we examined the separate and combined effects of exemplar and speaker variability on monolingual and bilingual children's cross-situational word learning performance. Results revealed that children's word learning performance did not differ when the input varied in a single dimension (i.e., exemplars or speakers) compared with a condition with no variability independent of their linguistic background. However, when performance in conditions that varied in a single dimension (i.e., exemplars or speakers) was compared with a condition that varied in multiple dimensions (i.e., exemplars and speakers), bilingual word learning advantages were observed; bilinguals were more likely to learn word-referent associations than monolinguals. Together, results suggest that children can learn and generalize word-referent associations from input that varies in exemplars and speakers and that bilingualism may bolster learning under conditions of increased input variability.

Explaining L2 Lexical Learning in Multiple Scenarios: Cross-Situational Word Learning in L1 Mandarin L2 English Speakers.

Adults commonly struggle with perceiving and recognizing the sounds and words of a second language (L2), especially when the L2 sounds do not have a counterpart in the learner's first language (L1). We examined how L1 Mandarin L2 English speakers learned pseudo English words within a cross-situational word learning (CSWL) task previously presented to monolingual English and bilingual Mandarin-English speakers. CSWL is ambiguous because participants are not provided with direct mappings of words and object referents. Rather, learners discern word-object correspondences through tracking multiple co-occurrences across learning trials. The monolinguals and bilinguals tested in previous studies showed lower performance for pseudo words that formed vowel minimal pairs (e.g., /dit/-/dɪt/) than pseudo word which formed consonant minimal pairs (e.g., /bɔn/-/pɔn/) or non-minimal pairs which differed in all segments (e.g., /bɔn/-/dit/). In contrast, L1 Mandarin L2 English listeners struggled to learn all word pairs. We explain this seemingly contradicting finding by considering the multiplicity of acoustic cues in the stimuli presented to all participant groups. Stimuli were produced in infant-directed-speech (IDS) in order to compare performance by children and adults and because previous research had shown that IDS enhances L1 and L2 acquisition. We propose that the suprasegmental pitch variation in the vowels typical of IDS stimuli might be perceived as lexical tone distinctions for tonal language speakers who cannot fully inhibit their L1 activation, resulting in high lexical competition and diminished learning during an ambiguous word learning task. Our results are in line with the Second Language Linguistic Perception (L2LP) model which proposes that fine-grained acoustic information from multiple sources and the ability to switch between language modes affects non-native phonetic and lexical development.

Combining statistics: the role of phonotactics on cross-situational word learning.

Language learners can rely on phonological and semantic information to learn novel words. Using a cross-situational word learning paradigm, we explored the role of phonotactic probabilities on word learning in ambiguous contexts. Brazilian-Portuguese speaking adults (N = 30) were exposed to two sets of word-object pairs. Words from one set of labels had slightly higher phonotactic probabilities than words from the other set. By tracking co-occurrences of words and objects, participants were able to learn word-object mappings similarly across both sets. Our findings contrast with studies showing a facilitative effect of phonotactic probability on word learning in non-ambiguous contexts.

Music Perception Abilities and Ambiguous Word Learning: Is There Cross-Domain Transfer in Nonmusicians?

Perception of music and speech is based on similar auditory skills, and it is often suggested that those with enhanced music perception skills may perceive and learn novel words more easily. The current study tested whether music perception abilities are associated with novel word learning in an ambiguous learning scenario. Using a cross-situational word learning (CSWL) task, nonmusician adults were exposed to word-object pairings between eight novel words and visual referents. Novel words were either non-minimal pairs differing in all sounds or minimal pairs differing in their initial consonant or vowel. In order to be successful in this task, learners need to be able to correctly encode the phonological details of the novel words and have sufficient auditory working memory to remember the correct word-object pairings. Using the Mistuning Perception Test (MPT) and the Melodic Discrimination Test (MDT), we measured learners' pitch perception and auditory working memory. We predicted that those with higher MPT and MDT values would perform better in the CSWL task and in particular for novel words with high phonological overlap (i.e., minimal pairs). We found that higher musical perception skills led to higher accuracy for non-minimal pairs and minimal pairs differing in their initial consonant. Interestingly, this was not the case for vowel minimal pairs. We discuss the results in relation to theories of second language word learning such as the Second Language Perception model (L2LP).

Cross-Situational Word Learning With Multimodal Neural Networks.

In order to learn the mappings from words to referents, children must integrate co-occurrence information across individually ambiguous pairs of scenes and utterances, a challenge known as cross-situational word learning. In machine learning, recent multimodal neural networks have been shown to learn meaningful visual-linguistic mappings from cross-situational data, as needed to solve problems such as image captioning and visual question answering. These networks are potentially appealing as cognitive models because they can learn from raw visual and linguistic stimuli, something previous cognitive models have not addressed. In this paper, we examine whether recent machine learning approaches can help explain various behavioral phenomena from the psychological literature on cross-situational word learning. We consider two variants of a multimodal neural network architecture and look at seven different phenomena associated with cross-situational word learning and word learning more generally. Our results show that these networks can learn word-referent mappings from a single epoch of training, mimicking the amount of training commonly found in cross-situational word learning experiments. Additionally, these networks capture some, but not all of the phenomena we studied, with all of the failures related to reasoning via mutual exclusivity. These results provide insight into the kinds of phenomena that arise naturally from relatively generic neural network learning algorithms, and which word learning phenomena require additional inductive biases.

Combining statistics: the role of phonotactics on cross-situational word learning

Abstract Language learners can rely on phonological and semantic information to learn novel words. Using a cross-situational word learning paradigm, we explored the role of phonotactic probabilities on word learning in ambiguous contexts. Brazilian-Portuguese speaking adults (N = 30) were exposed to two sets of word-object pairs. Words from one set of labels had slightly higher phonotactic probabilities than words from the other set. By tracking co-occurrences of words and objects, participants were able to learn word-object mappings similarly across both sets. Our findings contrast with studies showing a facilitative effect of phonotactic probability on word learning in non-ambiguous contexts.

Is 10 Better than 1? The Effect of Speaker Variability on Children's Cross-situational Word Learning.

The current study examined the effect of speaker variability on children's cross-situational word learning (XSWL). The study also examined the role of bilingual experience and sustained attention. Forty English monolingual children and 40 Spanish-English bilingual children ages 4-7 completed a XSWL task in a Single Speaker Condition and a Multiple Speaker Condition. Results indicated that speaker variability neither facilitated nor hindered XSWL. While monolingual children outperformed bilingual children, speaker-variability effects did not fluctuate across the two language groups. Notably, exposure to multiple speakers facilitated XSWL in children with poorer sustained attention skills, suggesting that variability in the input may be especially useful to children with poorer cognitive processing abilities.

Multiple components of statistical word learning are resource dependent: Evidence from a dual-task learning paradigm.

It is increasingly understood that people may learn new word/object mappings in part via a form of statistical learning in which they track co-occurrences between words and objects across situations (cross-situational learning). Multiple learning processes contribute to this, thought to reflect the simultaneous influence of real-time hypothesis testing and graduate learning. It is unclear how these processes interact, and if any require explicit cognitive resources. To manipulate the availability of working memory resources for explicit processing, participants completed a dual-task paradigm in which a cross-situational word-learning task was interleaved with a short-term memory task. We then used trial-by-trial analyses to estimate how different learning processes that play out simultaneously are impacted by resource availability. Critically, we found that the effect of hypothesis testing and gradual learning effects showed a small reduction under limited resources, and that the effect of memory load was not fully mediated by these processes. This suggests that neither is purely explicit, and there may be additional resource-dependent processes at play. Consistent with a hybrid account, these findings suggest that these two aspects of learning may reflect different aspects of a single system gated by attention, rather than competing learning systems.

Sampling to learn words: Adults and children sample words that reduce referential ambiguity.

How do learners gather new information during word learning? One possibility is that learners selectively sample items that help them reduce uncertainty about new word meanings. In a series of cross-situational word learning tasks with adults and children, we manipulated the referential ambiguity of label-object pairs experienced during training and subsequently investigated which words participants chose to sample additional information about. In the first experiment, adult learners chose to receive additional training on object-label associations that reduce referential ambiguity during cross-situational word learning. This ambiguity-reduction strategy was related to improved test performance. In two subsequent experiments, we found that, at least in some contexts, children (3-8 years of age) show a similar preference to seek information about words experienced in ambiguous word learning situations. In Experiment 2, children did not preferentially select object-label associations that remained ambiguous during cross-situational word learning. However, in a third experiment that increased the relative ambiguity of two sets of novel object-label associations, we found evidence that children preferentially make selections that reduce ambiguity about novel word meanings. These results carry implications for understanding how children actively contribute to their own language development by seeking information that supports learning.

A neural network architecture for learning word-referent associations in multiple contexts.

This article proposes a biologically inspired neurocomputational architecture which learns associations between words and referents in different contexts, considering evidence collected from the literature of Psycholinguistics and Neurolinguistics. The multi-layered architecture takes as input raw images of objects (referents) and streams of word's phonemes (labels), builds an adequate representation, recognizes the current context, and associates label with referents incrementally, by employing a Self-Organizing Map which creates new association nodes (prototypes) as required, adjusts the existing prototypes to better represent the input stimuli and removes prototypes that become obsolete/unused. The model takes into account the current context to retrieve the correct meaning of words with multiple meanings. Simulations show that the model can reach up to 78% of word-referent association accuracy in ambiguous situations and approximates well the learning rates of humans as reported by three different authors in five Cross-Situational Word Learning experiments, also displaying similar learning patterns in the different learning conditions.

The company objects keep: Linking referents together during cross-situational word learning.

Learning the meanings of words involves not only linking individual words to referents but also building a network of connections among entities in the world, concepts, and words. Previous studies reveal that infants and adults track the statistical co-occurrence of labels and objects across multiple ambiguous training instances to learn words. However, it is less clear whether, given distributional or attentional cues, learners also encode associations amongst the novel objects. We investigated the consequences of two types of cues that highlighted object-object links in a cross-situational word learning task: distributional structure - how frequently the referents of novel words occurred together - and visual context - whether the referents were seen on matching backgrounds. Across three experiments, we found that in addition to learning novel words, adults formed connections between frequently co-occurring objects. These findings indicate that learners exploit statistical regularities to form multiple types of associations during word learning.

The Pursuit of Word Meanings.

We evaluate here the performance of four models of cross-situational word learning: two global models, which extract and retain multiple referential alternatives from each word occurrence; and two local models, which extract just a single referent from each occurrence. One of these local models, dubbed Pursuit, uses an associative learning mechanism to estimate word-referent probability but pursues and tests the best referent-meaning at any given time. Pursuit is found to perform as well as global models under many conditions extracted from naturalistic corpora of parent-child interactions, even though the model maintains far less information than global models. Moreover, Pursuit is found to best capture human experimental findings from several relevant cross-situational word-learning experiments, including those of Yu and Smith (), the paradigm example of a finding believed to support fully global cross-situational models. Implications and limitations of these results are discussed, most notably that the model characterizes only the earliest stages of word learning, when reliance on the co-occurring referent world is at its greatest.