Rapid learning of syllable classes from a perceptually continuous speech stream

Rules generalization in children with dyslexia

BACKGROUND

Rule learning (RL) is the ability to extract and generalize higher-order repetition-based structures. Children with Developmental Dyslexia (DD) often report difficulties in learning complex regularities in sequential stimuli, which might be due to the complexity of the rule to be learned. Learning high-order repetition-based rules represents a building block for the development of language skills.

AIMS

This study investigates the ability to extract and generalize simple, repetition-based visual rules (e.g., ABA) in 8-11-year-old children without (TD) and with a diagnosis of Development Dyslexia (DD) and its relationship with language and reading skills.

METHOD

Using a forced-choice paradigm, children were first exposed to a visual sequence containing a repetition-based rule (e.g., ABA) and were then asked to recognize familiar and novel rules generated by new visual elements. Standardized language and reading tests were also administered to both groups.

RESULTS

The accuracy in recognizing rules was above chance for both groups, even though DD children were less accurate than TD children, suggesting a less efficient RL mechanism in the DD group. Moreover, visual RL was positively correlated with both language and reading skills.

CONCLUSION

These results further confirm the crucial role of RL in the acquisition of linguistic skills and mastering reading abilities.

Hebbian learning can explain rhythmic neural entrainment to statistical regularities

In many domains, learners extract recurring units from continuous sequences. For example, in unknown languages, fluent speech is perceived as a continuous signal. Learners need to extract the underlying words from this continuous signal and then memorize them. One prominent candidate mechanism is statistical learning, whereby learners track how predictive syllables (or other items) are of one another. Syllables within the same word predict each other better than syllables straddling word boundaries. But does statistical learning lead to memories of the underlying words-or just to pairwise associations among syllables? Electrophysiological results provide the strongest evidence for the memory view. Electrophysiological responses can be time-locked to statistical word boundaries (e.g., N400s) and show rhythmic activity with a periodicity of word durations. Here, I reproduce such results with a simple Hebbian network. When exposed to statistically structured syllable sequences (and when the underlying words are not excessively long), the network activation is rhythmic with the periodicity of a word duration and activation maxima on word-final syllables. This is because word-final syllables receive more excitation from earlier syllables with which they are associated than less predictable syllables that occur earlier in words. The network is also sensitive to information whose electrophysiological correlates were used to support the encoding of ordinal positions within words. Hebbian learning can thus explain rhythmic neural activity in statistical learning tasks without any memory representations of words. Learners might thus need to rely on cues beyond statistical associations to learn the words of their native language. RESEARCH HIGHLIGHTS: Statistical learning may be utilized to identify recurring units in continuous sequences (e.g., words in fluent speech) but may not generate explicit memory for words. Exposure to statistically structured sequences leads to rhythmic activity with a period of the duration of the underlying units (e.g., words). I show that a memory-less Hebbian network model can reproduce this rhythmic neural activity as well as putative encodings of ordinal positions observed in earlier research. Direct tests are needed to establish whether statistical learning leads to declarative memories for words.

Of words and whistles: Statistical learning operates similarly for identical sounds perceived as speech and non-speech

Statistical learning is an ability that allows individuals to effortlessly extract patterns from the environment, such as sound patterns in speech. Some prior evidence suggests that statistical learning operates more robustly for speech compared to non-speech stimuli, supporting the idea that humans are predisposed to learn language. However, any apparent statistical learning advantage for speech could be driven by signal acoustics, rather than the subjective perception per se of sounds as speech. To resolve this issue, the current study assessed whether there is a statistical learning advantage for ambiguous sounds that are subjectively perceived as speech-like compared to the same sounds perceived as non-speech, thereby controlling for acoustic features. We first induced participants to perceive sine-wave speech (SWS)-a degraded form of speech not immediately perceptible as speech-as either speech or non-speech. After this induction phase, participants were exposed to a continuous stream of repeating trisyllabic nonsense words, composed of SWS syllables, and then completed an explicit familiarity rating task and an implicit target detection task to assess learning. Critically, participants showed robust and equivalent performance on both measures, regardless of their subjective speech perception. In contrast, participants who perceived the SWS syllables as more speech-like showed better detection of individual syllables embedded in speech streams. These results suggest that speech perception facilitates processing of individual sounds, but not the ability to extract patterns across sounds. Our findings suggest that statistical learning is not influenced by the perceived linguistic relevance of sounds, and that it may be conceptualized largely as an automatic, stimulus-driven mechanism.

Kindergarteners Use Cross-Situational Statistics to Infer the Meaning of Grammatical Elements

Many studies demonstrate that detecting statistical regularities in linguistic input plays a key role in language acquisition. Yet, it is unclear to what extent statistical learning is involved in more naturalistic settings, when young children have to acquire meaningful grammatical elements. In the present study, we address these points, by investigating whether statistical learning is involved in acquiring a morpho-syntactic structure from input that resembles natural languages more closely. We exposed 50 kindergarteners (M = 5 years, 5 months) to a miniature language in which they had to learn a grammatical marker that expressed number, and which could only be acquired on the basis of the distributional properties in the input. Half of the children performed an attention check during the experiment. Results show that young children are able to learn this meaning. We found no clear evidence that facilitating attention to the input increases learning performance.

Linking the neural basis of distributional statistical learning with transitional statistical learning: The paradox of attention

Statistical learning, the process of tracking distributional information and discovering embedded patterns, is traditionally regarded as a form of implicit learning. However, recent studies proposed that both implicit (attention-independent) and explicit (attention-dependent) learning systems are involved in statistical learning. To understand the role of attention in statistical learning, the current study investigates the cortical processing of distributional patterns in speech across local and global contexts. We then ask how these cortical responses relate to statistical learning behavior in a word segmentation task. We found Event-Related Potential (ERP) evidence of pre-attentive processing of both the local (mismatching negativity) and global distributional information (late discriminative negativity). However, as speech elements became less frequent and more surprising, some participants showed an involuntary attentional shift, reflected in a P3a response. Individuals who displayed attentive neural tracking of distributional information showed faster learning in a speech statistical learning task. These results suggest that an involuntary attentional shift might play a facilitatory, but not essential, role in statistical learning.

A Special Role of Syllables, But Not Vowels or Consonants, for Nonadjacent Dependency Learning

Successful language processing entails tracking (morpho)syntactic relationships between distant units of speech, so-called nonadjacent dependencies (NADs). Many cues to such dependency relations have been identified, yet the linguistic elements encoding them have received little attention. In the present investigation, we tested whether and how these elements, here syllables, consonants, and vowels, affect behavioral learning success as well as learning-related changes in neural activity in relation to item-specific NAD learning. In a set of two EEG studies with adults, we compared learning under conditions where either all segment types (Experiment 1) or only one segment type (Experiment 2) was informative. The collected behavioral and ERP data indicate that, when all three segment types are available, participants mainly rely on the syllable for NAD learning. With only one segment type available for learning, adults also perform most successfully with syllable-based dependencies. Although we find no evidence for successful learning across vowels in Experiment 2, dependencies between consonants seem to be identified at least passively at the phonetic-feature level. Together, these results suggest that successful item-specific NAD learning may depend on the availability of syllabic information. Furthermore, they highlight consonants' distinctive power to support lexical processes. Although syllables show a clear facilitatory function for NAD learning, the underlying mechanisms of this advantage require further research.

Visual Implicit Learning Abilities in Infants at Familial Risk for Language and Learning Impairments

The ability of infants to track transitional probabilities (Statistical Learning—SL) and to extract and generalize high-order rules (Rule Learning—RL) from sequences of items have been proposed as being pivotal for the acquisition of language and reading skills. Although there is ample evidence of specific associations between SL and RL abilities and, respectively, vocabulary and grammar skills, research exploring SL and RL as early markers of language and learning (dis)abilities is still scarce. Here we investigated the efficiency of visual SL and RL skills in typically developing (TD) seven-month-old infants and in seven-month-old infants at high risk (HR) for language learning impairment. Infants were tested in two visual-habituation tasks aimed to measure their ability to extract transitional probabilities (SL task) or high-order, repetition-based rules (RL task) from sequences of visual shapes. Post-habituation looking time preferences revealed that both TD and HR infants succeeded in learning the statistical structure (SL task), while only TD infants, but not HR infants, were able to learn and generalize the high-order rule (RL task). These findings suggest that SL and RL may contribute differently to the emergence of language learning impairment and support the hypothesis that a mechanism linked to the extraction of grammar structures may contribute to the disorder.

Cognitive mechanisms of statistical learning and segmentation of continuous sensory input

Two classes of cognitive mechanisms have been proposed to explain segmentation of continuous sensory input into discrete recurrent constituents: clustering and boundary-finding mechanisms. Clustering mechanisms are based on identifying frequently co-occurring elements and merging them together as parts that form a single constituent. Bracketing (or boundary-finding) mechanisms work by identifying rarely co-occurring elements that correspond to the boundaries between discrete constituents. In a series of behavioral experiments, I tested which mechanisms are at play in the visual modality both during segmentation of a continuous syllabic sequence into discrete word-like constituents and during recognition of segmented constituents. Additionally, I explored conscious awareness of the products of statistical learning—whole constituents versus merged clusters of smaller subunits. My results suggest that both online segmentation and offline recognition of extracted constituents rely on detecting frequently co-occurring elements, a process likely based on associative memory. However, people are more aware of having learnt whole tokens than of recurrent composite clusters.

A Process-Oriented View of Procedural Memory Can Help Better Understand Tourette's Syndrome

Tourette's syndrome (TS) is a neurodevelopmental disorder characterized by repetitive movements and vocalizations, also known as tics. The phenomenology of tics and the underlying neurobiology of the disorder have suggested that the altered functioning of the procedural memory system might contribute to its etiology. However, contrary to the robust findings of impaired procedural memory in neurodevelopmental disorders of language, results from TS have been somewhat mixed. We review the previous studies in the field and note that they have reported normal, impaired, and even enhanced procedural performance. These mixed findings may be at least partially be explained by the diversity of the samples in both age and tic severity, the vast array of tasks used, the low sample sizes, and the possible confounding effects of other cognitive functions, such as executive functions, working memory or attention. However, we propose that another often overlooked factor could also contribute to the mixed findings, namely the multiprocess nature of the procedural system itself. We propose that a process-oriented view of procedural memory functions could serve as a theoretical framework to help integrate these varied findings. We discuss evidence suggesting heterogeneity in the neural regions and their functional contributions to procedural memory. Our process-oriented framework can help to deepen our understanding of the complex profile of procedural functioning in TS and atypical development in general.

Fast but Not Furious. When Sped Up Bit Rate of Information Drives Rule Induction

The language abilities of young and adult learners range from memorizing specific items to finding statistical regularities between them (item-bound generalization) and generalizing rules to novel instances (category-based generalization). Both external factors, such as input variability, and internal factors, such as cognitive limitations, have been shown to drive these abilities. However, the exact dynamics between these factors and circumstances under which rule induction emerges remain largely underspecified. Here, we extend our information-theoretic model (Radulescu et al., 2019), based on Shannon’s noisy-channel coding theory, which adds into the “formula” for rule induction the crucial dimension of time: the rate of encoding information by a time-sensitive mechanism. The goal of this study is to test the channel capacity-based hypothesis of our model: if the input entropy per second is higher than the maximum rate of information transmission (bits/second), which is determined by the channel capacity, the encoding method moves gradually from item-bound generalization to a more efficient category-based generalization, so as to avoid exceeding the channel capacity. We ran two artificial grammar experiments with adults, in which we sped up the bit rate of information transmission, crucially not by an arbitrary amount but by a factor calculated using the channel capacity formula on previous data. We found that increased bit rate of information transmission in a repetition-based XXY grammar drove the tendency of learners toward category-based generalization, as predicted by our model. Conversely, we found that increased bit rate of information transmission in complex non-adjacent dependency aXb grammar impeded the item-bound generalization of the specific a_b frames, and led to poorer learning, at least judging by our accuracy assessment method. This finding could show that, since increasing the bit rate of information precipitates a change from item-bound to category-based generalization, it impedes the item-bound generalization of the specific a_b frames, and that it facilitates category-based generalization both for the intervening Xs and possibly for a/b categories. Thus, sped up bit rate does not mean that an unrestrainedly increasing bit rate drives rule induction in any context, or grammar. Rather, it is the specific dynamics between the input entropy and the maximum rate of information transmission.

Pitches that Wire Together Fire Together: Scale Degree Associations Across Time Predict Melodic Expectations

The ongoing generation of expectations is fundamental to listeners' experience of music, but research into types of statistical information that listeners extract from musical melodies has tended to emphasize transition probabilities and n-grams, with limited consideration given to other types of statistical learning that may be relevant. Temporal associations between scale degrees represent a different type of information present in musical melodies that can be learned from musical corpora using expectation networks, a computationally simple method based on activation and decay. Expectation networks infer the expectation of encountering one scale degree followed in the near (but not necessarily immediate) future by another given scale degree, with previous work suggesting that scale degree associations learned by expectation networks better predict listener ratings of pitch similarity than transition probabilities. The current work outlines how these learned scale degree associations can be combined to predict melodic continuations and tests the resulting predictions on a dataset of listener responses to a musical cloze task previously used to compare two other models of melodic expectation, a variable-order Markov model (IDyOM) and Temperley's music-theoretically motivated model. Under multinomial logistic regression, all three models explain significant unique variance in human melodic expectations, with coefficient estimates highest for expectation networks. These results suggest that generalized scale degree associations informed by both adjacent and nonadjacent relationships between melodic notes influence listeners' melodic predictions above and beyond n-gram context, highlighting the need to consider a broader range of statistical learning processes that may underlie listeners' expectations for upcoming musical events.

Two for the price of one: Concurrent learning of words and phonotactic regularities from continuous speech

To acquire the words of their language, learners face the challenge of tracking regularities at multiple levels of abstraction from continuous speech. In the current study, we examined adults’ ability to track two types of regularities from a continuous artificial speech stream: the individual words in the speech stream (token level information), and a phonotactic pattern shared by a subset of those words (type level information). We additionally manipulated exposure time to the language to examine the relationship between the acquisition of these two regularities. Using a ratings test procedure, we found that adults can extract both the words in the language and their phonotactic patterns from continuous speech in as little as 3.5 minutes of listening time. Results from a 2AFC testing method provide converging evidence that adults rapidly learn both words and their phonotactic patterns. Together, the findings suggest that adults are capable of concurrently tracking regularities at multiple levels of abstraction from brief exposures to a continuous stream of speech.

When forgetting fosters learning: A neural network model for statistical learning

Learning often requires splitting continuous signals into recurring units, such as the discrete words constituting fluent speech; these units then need to be encoded in memory. A prominent candidate mechanism involves statistical learning of co-occurrence statistics like transitional probabilities (TPs), reflecting the idea that items from the same unit (e.g., syllables within a word) predict each other better than items from different units. TP computations are surprisingly flexible and sophisticated. Humans are sensitive to forward and backward TPs, compute TPs between adjacent items and longer-distance items, and even recognize TPs in novel units. We explain these hallmarks of statistical learning with a simple model with tunable, Hebbian excitatory connections and inhibitory interactions controlling the overall activation. With weak forgetting, activations are long-lasting, yielding associations among all items; with strong forgetting, no associations ensue as activations do not outlast stimuli; with intermediate forgetting, the network reproduces the hallmarks above. Forgetting thus is a key determinant of these sophisticated learning abilities. Further, in line with earlier dissociations between statistical learning and memory encoding, our model reproduces the hallmarks of statistical learning in the absence of a memory store in which items could be placed.

Integrating when and what information in the left parietal lobe allows language rule generalization

A crucial aspect when learning a language is discovering the rules that govern how words are combined in order to convey meanings. Because rules are characterized by sequential co-occurrences between elements (e.g., “These cupcakes are unbelievable”), tracking the statistical relationships between these elements is fundamental. However, purely bottom-up statistical learning alone cannot fully account for the ability to create abstract rule representations that can be generalized, a paramount requirement of linguistic rules. Here, we provide evidence that, after the statistical relations between words have been extracted, the engagement of goal-directed attention is key to enable rule generalization. Incidental learning performance during a rule-learning task on an artificial language revealed a progressive shift from statistical learning to goal-directed attention. In addition, and consistent with the recruitment of attention, functional MRI (fMRI) analyses of late learning stages showed left parietal activity within a broad bilateral dorsal frontoparietal network. Critically, repetitive transcranial magnetic stimulation (rTMS) on participants’ peak of activation within the left parietal cortex impaired their ability to generalize learned rules to a structurally analogous new language. No stimulation or rTMS on a nonrelevant brain region did not have the same interfering effect on generalization. Performance on an additional attentional task showed that this rTMS on the parietal site hindered participants’ ability to integrate “what” (stimulus identity) and “when” (stimulus timing) information about an expected target. The present findings suggest that learning rules from speech is a two-stage process: following statistical learning, goal-directed attention—involving left parietal regions—integrates “what” and “when” stimulus information to facilitate rapid rule generalization.

This study uses repetitive transcranial stimulation to show that learning language rules from speech is a two-stage process; following statistical learning, goal-directed attention (involving left parietal regions) integrates "what" and "when" stimulus information to facilitate rapid rule generalization.

Non-adjacent Dependency Learning in Humans and Other Animals

Learning and processing natural language requires the ability to track syntactic relationships between words and phrases in a sentence, which are often separated by intervening material. These nonadjacent dependencies can be studied using artificial grammar learning paradigms and structured sequence processing tasks. These approaches have been used to demonstrate that human adults, infants and some nonhuman animals are able to detect and learn dependencies between nonadjacent elements within a sequence. However, learning nonadjacent dependencies appears to be more cognitively demanding than detecting dependencies between adjacent elements, and only occurs in certain circumstances. In this review, we discuss different types of nonadjacent dependencies in language and in artificial grammar learning experiments, and how these differences might impact learning. We summarize different types of perceptual cues that facilitate learning, by highlighting the relationship between dependent elements bringing them closer together either physically, attentionally, or perceptually. Finally, we review artificial grammar learning experiments in human adults, infants, and nonhuman animals, and discuss how similarities and differences observed across these groups can provide insights into how language is learned across development and how these language-related abilities might have evolved.

Non-adjacent dependency learning in infancy, and its link to language development

To acquire language, infants must learn how to identify words and linguistic structure in speech. Statistical learning has been suggested to assist both of these tasks. However, infants' capacity to use statistics to discover words and structure together remains unclear. Further, it is not yet known how infants'  statistical learning ability relates to their language development. We trained 17-month-old infants on an artificial language comprising non-adjacent dependencies, and examined their looking times on tasks assessing sensitivity to words and structure using an eye-tracked head-turn-preference paradigm. We measured infants' vocabulary size using a Communicative Development Inventory (CDI) concurrently and at 19, 21, 24, 25, 27, and 30 months to relate performance to language development. Infants could segment the words from speech, demonstrated by a significant difference in looking times to words versus part-words. Infants' segmentation performance was significantly related to their vocabulary size (receptive and expressive) both currently, and over time (receptive until 24 months, expressive until 30 months), but was not related to the rate of vocabulary growth. The data also suggest infants may have developed sensitivity to generalised structure, indicating similar statistical learning mechanisms may contribute to the discovery of words and structure in speech, but this was not related to vocabulary size.

Mark my words: High frequency marker words impact early stages of language learning

High frequency words have been suggested to benefit both speech segmentation and grammatical categorization of the words around them. Despite utilizing similar information, these tasks are usually investigated separately in studies examining learning. We determined whether including high frequency words in continuous speech could support categorization when words are being segmented for the first time. We familiarized learners with continuous artificial speech comprising repetitions of target words, which were preceded by high-frequency marker words. Crucially, marker words distinguished targets into 2 distributionally defined categories. We measured learning with segmentation and categorization tests and compared performance against a control group that heard the artificial speech without these marker words (i.e., just the targets, with no cues for categorization). Participants segmented the target words from speech in both conditions, but critically when the marker words were present, they influenced acquisition of word-referent mappings in a subsequent transfer task, with participants demonstrating better early learning for mappings that were consistent (rather than inconsistent) with the distributional categories. We propose that high-frequency words may assist early grammatical categorization, while speech segmentation is still being learned. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

A review of computational models of basic rule learning: The neural-symbolic debate and beyond

We present a critical review of computational models of generalization of simple grammar-like rules, such as ABA and ABB. In particular, we focus on models attempting to account for the empirical results of Marcus et al. (Science, 283(5398), 77-80 1999). In that study, evidence is reported of generalization behavior by 7-month-old infants, using an Artificial Language Learning paradigm. The authors fail to replicate this behavior in neural network simulations, and claim that this failure reveals inherent limitations of a whole class of neural networks: those that do not incorporate symbolic operations. A great number of computational models were proposed in follow-up studies, fuelling a heated debate about what is required for a model to generalize. Twenty years later, this debate is still not settled. In this paper, we review a large number of the proposed models. We present a critical analysis of those models, in terms of how they contribute to answer the most relevant questions raised by the experiment. After identifying which aspects require further research, we propose a list of desiderata for advancing our understanding on generalization.

Individual and Developmental Differences in Distributional Learning

Purpose

This study examined whether children and adults with developmental language disorder (DLD) could use distributional information in an artificial language to learn about grammatical category membership similarly to their typically developing (TD) peers and whether developmental differences existed within and between DLD and TD groups.

Method

Sixteen children ages 7-9 with DLD, 26 age-matched TD children, 17 college students with DLD, and 17 TD college students participated in this task. We used an artificial grammar learning paradigm in which participants had to use knowledge of category membership to determine the acceptability of test items that they had not heard during a training phase.

Results

Individuals with DLD performed similarly to TD peers in distinguishing grammatical from ungrammatical combinations, with no differences between age groups. The order in which items were heard at test differentially affected child versus adult participants and showed a relation with attention and phonological working memory as well.

Conclusion

Differences in ratings between grammatical and ungrammatical items in this task suggest that individuals with DLD can form grammatical categories from novel input and more broadly use distributional information. Differences in order effects suggest a developmental timeline for sensitivity to updating distributional information.

The roles of item repetition and position in infants' abstract rule learning

We asked whether 11- and 14- month-old infants' abstract rule learning, an early form of analogical reasoning, is susceptible to processing constraints imposed by limits in attention and memory for sequence position. We examined 11- and 14- month-old infants' learning and generalization of abstract repetition rules ("repetition anywhere," Experiment 1 or "medial repetition," Experiment 2) and ordering of specific items (edge positions, Experiment 3) in 4-item sequences. Infants were habituated to sequences containing repetition- and/or position-based structure and then tested with "familiar" vs. "novel" (random) sequences composed of new items. Eleven-month-olds (N = 40) failed to learn abstract repetition rules, but 14-month-olds (N = 40) learned rules under both conditions. In Experiment 3, 11-month-olds (N = 20) learned item edge positions in sequences identical to those in Experiment 2. We conclude that infant sequence learning is constrained by item position in similar ways as in adults.

Influence of metrical structure on learning of positional regularities in movement sequences

Sequential stimuli are usually perceived to have hierarchical temporal structures. However, some of these structures are only investigated in one type of sequence, regardless of the existing evidence, showing the domain-generality of the representation of these structures. Here, we assess whether the hierarchical representation of regularly segmented action sequences resembles the perceived metrical patterns that organize the representation of events hierarchically in temporally regular sequences. In all our experiments, we presented the participants with sequences of human movements and tested the perception of metrical pattern by segmenting the movement streams into temporally equal groups containing four movements. In Experiment 1, we found that a movement sequence with temporally equal groupings improves the learning of positional regularities inherent within each group of movements. To further clarify the degree to which this learning mechanism is affected by the perceived metrical patterns, we conducted Experiments 2a and 2b, in which the relative saliencies of the first and last positions in the movement groups, respectively, were studied. The results showed that, although in the learning of positional regularities, the rule-conforming first positions are as effective as when both first and last positions are legal, the last positions are not as influential. Based on these findings we conclude that, in grouped sequences, learning of positional regularities may be modulated by the metrical saliency patterns that are imposed by the temporal regularity of the sequential grouping pattern.

A Comparative Perspective on the Role of Acoustic Cues in Detecting Language Structure

Most human language learners acquire language primarily via the auditory modality. This is one reason why auditory artificial grammars play a prominent role in the investigation of the development and evolutionary roots of human syntax. The present position paper brings together findings from human and non-human research on the impact of auditory cues on learning about linguistic structures with a special focus on how different types of cues and biases in auditory cognition may contribute to success and failure in artificial grammar learning (AGL). The basis of our argument is the link between auditory cues and syntactic structure across languages and development. Cross-species comparison suggests that many aspects of auditory cognition that are relevant for language are not human specific and are present even in rather distantly related species. Furthermore, auditory cues and biases impact on learning, which we will discuss in the example of auditory perception and AGL studies. This observation, together with the significant role of auditory cues in language processing, supports the idea that auditory cues served as a bootstrap to syntax during language evolution. Yet this also means that potentially human-specific syntactic abilities are not due to basic auditory differences between humans and non-human animals but are based upon more advanced cognitive processes.

Differences in relative frequency facilitate learning abstract rules

Humans learn the rules that govern how the elements of their language are organized over an input that is often not homogeneous (it might contain noise, or even include rules from different linguistic systems, as it might be the case for bilinguals). In the present study we explore the conditions under which participants can learn an abstract rule when it is presented in a heterogeneous context. Results from six experiments show that listeners can learn a token-independent rule even if it is presented together with some exemplars that implement a different regularity (Experiment 1a and 1b). In fact, learning rules from an input containing several patterns does not seem to differ from learning them from an input containing only one (Experiment 1c). More surprisingly, we observed that listeners can even learn an abstract rule that is only implemented over 10% of the exemplars that compose a familiarization stream (Experiments 2a and 2b). When the proportion of tokens implementing the target and the non-target rules is balanced, we did not observe any learning (Experiment 3). Our results suggest that listeners use differences in relative frequency to keep separate linguistic rules apart. This allows them to learn different abstract regularities from a non-homogeneous linguistic signal.

Non-adjacent Dependencies Processing in Human and Non-human Primates

Human and non-human primates share the ability to extract adjacent dependencies and, under certain conditions, non-adjacent dependencies (i.e., predictive relationships between elements that are separated by one or several intervening elements in a sequence). In this study, we explore the online extraction dynamics of non-adjacent dependencies in humans and baboons using a serial reaction time task. Participants had to produce three-target sequences containing deterministic relationships between the first and last target locations. In Experiment 1, participants from the two species could extract these non-adjacent dependencies, but humans required less exposure than baboons. In Experiment 2, the data show for the first time in a non-human primate species the successful generalization of sequential non-adjacent dependencies over novel intervening items. These findings provide new evidence to further constrain current theories about the nature and the evolutionary origins of the learning mechanisms allowing the extraction of non-adjacent dependencies.

Limits on composition of conceptual operations in 9-month-olds

Complex systems are often built from a relatively small set of basic features or operations that can be combined in myriad ways. We investigated the developmental origins of this compositional architecture in 9-month-old infants, extending recent work that demonstrated rudimentary compositional abilities in preschoolers. Infants viewed two separate object-occlusion events that depicted a single featurechange operation. They were then tested with a combined operation to determine whether they expected the outcome of the two feature changes, even though this combination was unfamiliar. In contrast to preschoolers, infants did not appear to predictively compose these simple feature-change operations. A second experiment demonstrated the ability of infants to track two operations when not combined. The failure to compose basic operations is consistent with limitations on object tracking and early numerical cognition (Feigenson & Yamaguchi, 2009). We suggest that these results can be unified via a general principle: infants have difficulty with multiple updates to a representation of an unobservable.

Online incidental statistical learning of audiovisual word sequences in adults: a registered report

Statistical learning has been proposed as a key mechanism in language learning. Our main goal was to examine whether adults are capable of simultaneously extracting statistical dependencies in a task where stimuli include a range of structures amenable to statistical learning within a single paradigm. We devised an online statistical learning task using real word auditory-picture sequences that vary in two dimensions: (i) predictability and (ii) adjacency of dependent elements. This task was followed by an offline recall task to probe learning of each sequence type. We registered three hypotheses with specific predictions. First, adults would extract regular patterns from continuous stream (effect of grammaticality). Second, within grammatical conditions, they would show differential speeding up for each condition as a factor of statistical complexity of the condition and exposure. Third, our novel approach to measure online statistical learning would be reliable in showing individual differences in statistical learning ability. Further, we explored the relation between statistical learning and a measure of verbal short-term memory (STM). Forty-two participants were tested and retested after an interval of at least 3 days on our novel statistical learning task. We analysed the reaction time data using a novel regression discontinuity approach. Consistent with prediction, participants showed a grammaticality effect, agreeing with the predicted order of difficulty for learning different statistical structures. Furthermore, a learning index from the task showed acceptable test-retest reliability (r = 0.67). However, STM did not correlate with statistical learning. We discuss the findings noting the benefits of online measures in tracking the learning process.

What's statistical about learning? Insights from modelling statistical learning as a set of memory processes

Statistical learning has been studied in a variety of different tasks, including word segmentation, object identification, category learning, artificial grammar learning and serial reaction time tasks (e.g. Saffran et al. 1996 Science 274: , 1926-1928; Orban et al. 2008 Proceedings of the National Academy of Sciences 105: , 2745-2750; Thiessen & Yee 2010 Child Development 81: , 1287-1303; Saffran 2002 Journal of Memory and Language 47: , 172-196; Misyak & Christiansen 2012 Language Learning 62: , 302-331). The difference among these tasks raises questions about whether they all depend on the same kinds of underlying processes and computations, or whether they are tapping into different underlying mechanisms. Prior theoretical approaches to statistical learning have often tried to explain or model learning in a single task. However, in many cases these approaches appear inadequate to explain performance in multiple tasks. For example, explaining word segmentation via the computation of sequential statistics (such as transitional probability) provides little insight into the nature of sensitivity to regularities among simultaneously presented features. In this article, we will present a formal computational approach that we believe is a good candidate to provide a unifying framework to explore and explain learning in a wide variety of statistical learning tasks. This framework suggests that statistical learning arises from a set of processes that are inherent in memory systems, including activation, interference, integration of information and forgetting (e.g. Perruchet & Vinter 1998 Journal of Memory and Language 39: , 246-263; Thiessen et al. 2013 Psychological Bulletin 139: , 792-814). From this perspective, statistical learning does not involve explicit computation of statistics, but rather the extraction of elements of the input into memory traces, and subsequent integration across those memory traces that emphasize consistent information (Thiessen and Pavlik 2013 Cognitive Science 37: , 310-343).This article is part of the themed issue 'New frontiers for statistical learning in the cognitive sciences'.

Sleep-Driven Computations in Speech Processing

Acquiring language requires segmenting speech into individual words, and abstracting over those words to discover grammatical structure. However, these tasks can be conflicting-on the one hand requiring memorisation of precise sequences that occur in speech, and on the other requiring a flexible reconstruction of these sequences to determine the grammar. Here, we examine whether speech segmentation and generalisation of grammar can occur simultaneously-with the conflicting requirements for these tasks being over-come by sleep-related consolidation. After exposure to an artificial language comprising words containing non-adjacent dependencies, participants underwent periods of consolidation involving either sleep or wake. Participants who slept before testing demonstrated a sustained boost to word learning and a short-term improvement to grammatical generalisation of the non-adjacencies, with improvements after sleep outweighing gains seen after an equal period of wake. Thus, we propose that sleep may facilitate processing for these conflicting tasks in language acquisition, but with enhanced benefits for speech segmentation.

Gleaning Structure from Sound: The Role of Prosodic Contrast in Learning Non-adjacent Dependencies

The ability to detect non-adjacent dependencies (i.e. between a and b in aXb) in spoken input may support the acquisition of morpho-syntactic dependencies (e.g. The princess is kiss ing the frog). Functional morphemes in morpho-syntactic dependencies are often marked by perceptual cues that render them distinct from lexical elements. We use an artificial grammar learning experiment with adults to investigate the role of perceptual cues in non-adjacent dependency learning, by manipulating the perceptual/prosodic properties of the a / b elements in aXb strings and testing participants' incidental learning of these dependencies. Our results show that non-adjacent dependencies are learned both when the dependent elements are perceptually prominent, and when they are perceptually reduced compared to the intervening material (in the same way that functional words are reduced compared to lexical words), but only if integrated into a natural prosodic contour. This result supports the idea that the prosodic properties of natural languages facilitate non-adjacent dependency learning.

Transitional probabilities count more than frequency, but might not be used for memorization

Learners often need to extract recurring items from continuous sequences, in both vision and audition. The best-known example is probably found in word-learning, where listeners have to determine where words start and end in fluent speech. This could be achieved through universal and experience-independent statistical mechanisms, for example by relying on Transitional Probabilities (TPs). Further, these mechanisms might allow learners to store items in memory. However, previous investigations have yielded conflicting evidence as to whether a sensitivity to TPs is diagnostic of the memorization of recurring items. Here, we address this issue in the visual modality. Participants were familiarized with a continuous sequence of visual items (i.e., arbitrary or everyday symbols), and then had to choose between (i) high-TP items that appeared in the sequence, (ii) high-TP items that did not appear in the sequence, and (iii) low-TP items that appeared in the sequence. Items matched in TPs but differing in (chunk) frequency were much harder to discriminate than items differing in TPs (with no significant sensitivity to chunk frequency), and learners preferred unattested high-TP items over attested low-TP items. Contrary to previous claims, these results cannot be explained on the basis of the similarity of the test items. Learners thus weigh within-item TPs higher than the frequency of the chunks, even when the TP differences are relatively subtle. We argue that these results are problematic for distributional clustering mechanisms that analyze continuous sequences, and provide supporting computational results. We suggest that the role of TPs might not be to memorize items per se, but rather to prepare learners to memorize recurring items once they are presented in subsequent learning situations with richer cues.

Learning multiple rules simultaneously: Affixes are more salient than reduplications

Language learners encounter numerous opportunities to learn regularities, but need to decide which of these regularities to learn, because some are not productive in their native language. Here, we present an account of rule learning based on perceptual and memory primitives (Endress, Dehaene-Lambertz, & Mehler, Cognition, 105(3), 577–614, 2007; Endress, Nespor, & Mehler, Trends in Cognitive Sciences, 13(8), 348–353, 2009), suggesting that learners preferentially learn regularities that are more salient to them, and that the pattern of salience reflects the frequency of language features across languages. We contrast this view with previous artificial grammar learning research, which suggests that infants “choose” the regularities they learn based on rational, Bayesian criteria (Frank & Tenenbaum, Cognition, 120(3), 360–371, 2013; Gerken, Cognition, 98(3)B67–B74, 2006, Cognition, 115(2), 362–366, 2010). In our experiments, adult participants listened to syllable strings starting with a syllable reduplication and always ending with the same “affix” syllable, or to syllable strings starting with this “affix” syllable and ending with the “reduplication”. Both affixation and reduplication are frequently used for morphological marking across languages. We find three crucial results. First, participants learned both regularities simultaneously. Second, affixation regularities seemed easier to learn than reduplication regularities. Third, regularities in sequence offsets were easier to learn than regularities at sequence onsets. We show that these results are inconsistent with previous Bayesian rule learning models, but mesh well with the perceptual or memory primitives view. Further, we show that the pattern of salience revealed in our experiments reflects the distribution of regularities across languages. Ease of acquisition might thus be one determinant of the frequency of regularities across languages.

Attentional effects on rule extraction and consolidation from speech

Incidental learning plays a crucial role in the initial phases of language acquisition. However the knowledge derived from implicit learning, which is based on prediction-based mechanisms, may become explicit. The role that attention plays in the formation of implicit and explicit knowledge of the learned material is unclear. In the present study, we investigated the role that attention plays in the acquisition of non-adjacent rule learning from speech. In addition, we also tested whether the amount of attention during learning changes the representation of the learned material after a 24h delay containing sleep. For that, we developed an experiment run on two consecutive days consisting on the exposure to an artificial language that contained non-adjacent dependencies (rules) between words whereas different conditions were established to manipulate the amount of attention given to the rules (target and non-target conditions). Furthermore, we used both indirect and direct measures of learning that are more sensitive to implicit and explicit knowledge, respectively. Whereas the indirect measures indicated that learning of the rules occurred regardless of attention, more explicit judgments after learning showed differences in the type of learning reached under the two attention conditions. 24 hours later, indirect measures showed no further improvements during additional language exposure and explicit judgments indicated that only the information more robustly learned in the previous day, was consolidated.

Words, rules, and mechanisms of language acquisition

We review recent artificial language learning studies, especially those following Endress and Bonatti (Endress AD, Bonatti LL. Rapid learning of syllable classes from a perceptually continuous speech stream. Cognition 2007, 105:247-299), suggesting that humans can deploy a variety of learning mechanisms to acquire artificial languages. Several experiments provide evidence for multiple learning mechanisms that can be deployed in fluent speech: one mechanism encodes the positions of syllables within words and can be used to extract generalization, while the other registers co-occurrence statistics of syllables and can be used to break a continuum into its components. We review dissociations between these mechanisms and their potential role in language acquisition. We then turn to recent criticisms of the multiple mechanisms hypothesis and show that they are inconsistent with the available data. Our results suggest that artificial and natural language learning is best understood by dissecting the underlying specialized learning abilities, and that these data provide a rare opportunity to link important language phenomena to basic psychological mechanisms. For further resources related to this article, please visit the WIREs website.

Simultaneous segmentation and generalisation of non-adjacent dependencies from continuous speech

Language learning requires mastering multiple tasks, including segmenting speech to identify words, and learning the syntactic role of these words within sentences. A key question in language acquisition research is the extent to which these tasks are sequential or successive, and consequently whether they may be driven by distinct or similar computations. We explored a classic artificial language learning paradigm, where the language structure is defined in terms of non-adjacent dependencies. We show that participants are able to use the same statistical information at the same time to segment continuous speech to both identify words and to generalise over the structure, when the generalisations were over novel speech that the participants had not previously experienced. We suggest that, in the absence of evidence to the contrary, the most economical explanation for the effects is that speech segmentation and grammatical generalisation are dependent on similar statistical processing mechanisms.

Hemispheric Asymmetries in Repetition Enhancement and Suppression Effects in the Newborn Brain

Background

The repeated presentation of stimuli typically attenuates neural responses (repetition suppression) or, less commonly, increases them (repetition enhancement) when stimuli are highly complex, degraded or presented under noisy conditions. In adult functional neuroimaging research, these repetition effects are considered as neural correlates of habituation. The development and respective functional significance of these effects in infancy remain largely unknown.

Objective

This study investigates repetition effects in newborns using functional near-infrared spectroscopy, and specifically the role of stimulus complexity in evoking a repetition enhancement vs. a repetition suppression response, following up on Gervain et al. (2008). In that study, abstract rule-learning was found at birth in cortical areas specific to speech processing, as evidenced by a left-lateralized repetition enhancement of the hemodynamic response to highly variable speech sequences conforming to a repetition-based ABB artificial grammar, but not to a random ABC grammar.

Methods

Here, the same paradigm was used to investigate how simpler stimuli (12 different sequences per condition as opposed to 140), and simpler presentation conditions (blocked rather than interleaved) would influence repetition effects at birth.

Results

Results revealed that the two grammars elicited different dynamics in the two hemispheres. In left fronto-temporal areas, we reproduce the early perceptual discrimination of the two grammars, with ABB giving rise to a greater response at the beginning of the experiment than ABC. In addition, the ABC grammar evoked a repetition enhancement effect over time, whereas a stable response was found for the ABB grammar. Right fronto-temporal areas showed neither initial discrimination, nor change over time to either pattern.

Conclusion

Taken together with Gervain et al. (2008), this is the first evidence that manipulating methodological factors influences the presence or absence of neural repetition enhancement effects in newborns and stimulus variability appears a particularly important factor. Further, this temporal modulation is restricted to the left hemisphere, confirming its specialization for learning linguistic regularities from birth.

Frequency-based organization of speech sequences in a nonhuman animal

A recurrent question regarding language acquisition is the extent to which the mechanisms human infants use to discover patterns over the linguistic signal are highly specialized and uniquely human, or are the result of more general mechanisms present in other species. Research with very young infants suggests that they are able to use the relative frequency of elements in a linguistic sequence to infer word order. Here we ask if this ability could emerge from grouping biases present in nonhuman mammals. We show that animals discover differences in the frequency of elements in a sequence and can learn the relative order of frequent and infrequent elements. Nevertheless, in animals, relative frequency does not appear to be overridden by other cues that have been shown to be important to human infants, such as prosody. Our results demonstrate that the basic mechanism that allows listeners to extract ordering relations based on frequency is shared across species.

Finding words and word structure in artificial speech: the development of infants' sensitivity to morphosyntactic regularities

To achieve language proficiency, infants must find the building blocks of speech and master the rules governing their legal combinations. However, these problems are linked: words are also built according to rules. Here, we explored early morphosyntactic sensitivity by testing when and how infants could find either words or within-word structure in artificial speech snippets embodying properties of morphological constructions. We show that 12-month-olds use statistical relationships between syllables to extract words from continuous streams, but find word-internal regularities only if the streams are segmented. Seven-month-olds fail both tasks. Thus, 12-month-olds infants possess the resources to analyze the internal composition of words if the speech contains segmentation information. However, 7-month-old infants may not possess them, although they can track several statistical relations. This developmental difference suggests that morphosyntactic sensitivity may require computational resources extending beyond the detection of simple statistics.

Does complexity matter? Meta-analysis of learner performance in artificial grammar tasks

Complexity has been shown to affect performance on artificial grammar learning (AGL) tasks (categorization of test items as grammatical/ungrammatical according to the implicitly trained grammar rules). However, previously published AGL experiments did not utilize consistent measures to investigate the comprehensive effect of grammar complexity on task performance. The present study focused on computerizing Bollt and Jones's (2000) technique of calculating topological entropy (TE), a quantitative measure of AGL charts' complexity, with the aim of examining associations between grammar systems' TE and learners' AGL task performance. We surveyed the literature and identified 56 previous AGL experiments based on 10 different grammars that met the sampling criteria. Using the automated matrix-lift-action method, we assigned a TE value for each of these 10 previously used AGL systems and examined its correlation with learners' task performance. The meta-regression analysis showed a significant correlation, demonstrating that the complexity effect transcended the different settings and conditions in which the categorization task was performed. The results reinforced the importance of using this new automated tool to uniformly measure grammar systems' complexity when experimenting with and evaluating the findings of AGL studies.

Hemodynamic correlates of cognition in human infants

Over the past 20 years, the field of cognitive neuroscience has relied heavily on hemodynamic measures of blood oxygenation in local regions of the brain to make inferences about underlying cognitive processes. These same functional magnetic resonance imaging (fMRI) and functional near-infrared spectroscopy (fNIRS) techniques have recently been adapted for use with human infants. We review the advantages and disadvantages of these two neuroimaging methods for studies of infant cognition, with a particular emphasis on their technical limitations and the linking hypotheses that are used to draw conclusions from correlational data. In addition to summarizing key findings in several domains of infant cognition, we highlight the prospects of improving the quality of fNIRS data from infants to address in a more sophisticated way how cognitive development is mediated by changes in underlying neural mechanisms.

Distributional Language Learning: Mechanisms and Models of ategory Formation

In the past 15 years, a substantial body of evidence has confirmed that a powerful distributional learning mechanism is present in infants, children, adults and (at least to some degree) in nonhuman animals as well. The present article briefly reviews this literature and then examines some of the fundamental questions that must be addressed for any distributional learning mechanism to operate effectively within the linguistic domain. In particular, how does a naive learner determine the number of categories that are present in a corpus of linguistic input and what distributional cues enable the learner to assign individual lexical items to those categories? Contrary to the hypothesis that distributional learning and category (or rule) learning are separate mechanisms, the present article argues that these two seemingly different processes---acquiring specific structure from linguistic input and generalizing beyond that input to novel exemplars---actually represent a single mechanism. Evidence in support of this single-mechanism hypothesis comes from a series of artificial grammar-learning studies that not only demonstrate that adults can learn grammatical categories from distributional information alone, but that the specific patterning of distributional information among attested utterances in the learning corpus enables adults to generalize to novel utterances or to restrict generalization when unattested utterances are consistently absent from the learning corpus. Finally, a computational model of distributional learning that accounts for the presence or absence of generalization is reviewed and the implications of this model for linguistic-category learning are summarized.

Verbal positional memory in 7-month-olds

Verbal memory is a fundamental prerequisite for language learning. This study investigated 7-month-olds' (N = 62) ability to remember the identity and order of elements in a multisyllabic word. The results indicate that infants detect changes in the order of edge syllables, or the identity of the middle syllables, but fail to encode the order of middle syllables. This suggests that the representational format of multisyllabic words is determined by core mnemonic biases, which favor accurate encoding of edges and limits the encoding of temporal order for internal segments. The studies support accounts proposing that content and order are encoded separately; in addition, the data show that this dissociation occurs early in development.

Neurocognitive mechanisms of statistical-sequential learning: what do event-related potentials tell us?

Statistical-sequential learning (SL) is the ability to process patterns of environmental stimuli, such as spoken language, music, or one's motor actions, that unfold in time. The underlying neurocognitive mechanisms of SL and the associated cognitive representations are still not well understood as reflected by the heterogeneity of the reviewed cognitive models. The purpose of this review is: (1) to provide a general overview of the primary models and theories of SL, (2) to describe the empirical research - with a focus on the event-related potential (ERP) literature - in support of these models while also highlighting the current limitations of this research, and (3) to present a set of new lines of ERP research to overcome these limitations. The review is articulated around three descriptive dimensions in relation to SL: the level of abstractness of the representations learned through SL, the effect of the level of attention and consciousness on SL, and the developmental trajectory of SL across the life-span. We conclude with a new tentative model that takes into account these three dimensions and also point to several promising new lines of SL research.

Words and possible words in early language acquisition

In order to acquire language, infants must extract its building blocks-words-and master the rules governing their legal combinations from speech. These two problems are not independent, however: words also have internal structure. Thus, infants must extract two kinds of information from the same speech input. They must find the actual words of their language. Furthermore, they must identify its possible words, that is, the sequences of sounds that, being morphologically well formed, could be words. Here, we show that infants' sensitivity to possible words appears to be more primitive and fundamental than their ability to find actual words. We expose 12- and 18-month-old infants to an artificial language containing a conflict between statistically coherent and structurally coherent items. We show that 18-month-olds can extract possible words when the familiarization stream contains marks of segmentation, but cannot do so when the stream is continuous. Yet, they can find actual words from a continuous stream by computing statistical relationships among syllables. By contrast, 12-month-olds can find possible words when familiarized with a segmented stream, but seem unable to extract statistically coherent items from a continuous stream that contains minimal conflicts between statistical and structural information. These results suggest that sensitivity to word structure is in place earlier than the ability to analyze distributional information. The ability to compute nontrivial statistical relationships becomes fully effective relatively late in development, when infants have already acquired a considerable amount of linguistic knowledge. Thus, mechanisms for structure extraction that do not rely on extensive sampling of the input are likely to have a much larger role in language acquisition than general-purpose statistical abilities.

All together now: concurrent learning of multiple structures in an artificial language

Natural languages contain many layers of sequential structure, from the distribution of phonemes within words to the distribution of phrases within utterances. However, most research modeling language acquisition using artificial languages has focused on only one type of distributional structure at a time. In two experiments, we investigated adult learning of an artificial language that contains dependencies between both adjacent and non-adjacent words. We found that learners rapidly acquired both types of regularities and that the strength of the adjacent statistics influenced learning of both adjacent and non-adjacent dependencies. Additionally, though accuracy was similar for both types of structure, participants' knowledge of the deterministic non-adjacent dependencies was more explicit than their knowledge of the probabilistic adjacent dependencies. The results are discussed in the context of current theories of statistical learning and language acquisition.

Birth of an abstraction: a dynamical systems account of the discovery of an elsewhere principle in a category learning task

Human participants and recurrent ("connectionist") neural networks were both trained on a categorization system abstractly similar to natural language systems involving irregular ("strong") classes and a default class. Both the humans and the networks exhibited staged learning and a generalization pattern reminiscent of the Elsewhere Condition (Kiparsky, 1973). Previous connectionist accounts of related phenomena have often been vague about the nature of the networks' encoding systems. We analyzed our network using dynamical systems theory, revealing topological and geometric properties that can be directly compared with the mechanisms of non-connectionist, rule-based accounts. The results reveal that the networks "contain" structures related to mechanisms posited by rule-based models, partly vindicating the insights of these models. On the other hand, they support the one mechanism (OM), as opposed to the more than one mechanism (MOM), view of symbolic abstraction by showing how the appearance of MOM behavior can arise emergently from one underlying set of principles. The key new contribution of this study is to show that dynamical systems theory can allow us to explicitly characterize the relationship between the two perspectives in implemented models.

Probabilistically-Cued Patterns Trump Perfect Cues in Statistical Language Learning

Probabilistically-cued co-occurrence relationships between word categories are common in natural languages but difficult to acquire. For example, in English, determiner-noun and auxiliary-verb dependencies both involve co-occurrence relationships, but determiner-noun relationships are more reliably marked by correlated distributional and phonological cues, and appear to be learned more readily. We tested whether experience with co-occurrence relationships that are more reliable promotes learning those that are less reliable using an artificial language paradigm. Prior experience with deterministically-cued contingencies did not promote learning of less reliably-cued structure, nor did prior experience with relationships instantiated in the same vocabulary. In contrast, prior experience with probabilistically-cued co-occurrence relationships instantiated in different vocabulary did enhance learning. Thus, experience with co-occurrence relationships sharing underlying structure but not vocabulary may be an important factor in learning grammatical patterns. Furthermore, experience with probabilistically-cued co-occurrence relationships, despite their difficultly for naïve learners, lays an important foundation for learning novel probabilistic structure.

iMinerva: a mathematical model of distributional statistical learning

Statistical learning refers to the ability to identify structure in the input based on its statistical properties. For many linguistic structures, the relevant statistical features are distributional: They are related to the frequency and variability of exemplars in the input. These distributional regularities have been suggested to play a role in many different aspects of language learning, including phonetic categories, using phonemic distinctions in word learning, and discovering non-adjacent relations. On the surface, these different aspects share few commonalities. Despite this, we demonstrate that the same computational framework can account for learning in all of these tasks. These results support two conclusions. The first is that much, and perhaps all, of distributional statistical learning can be explained by the same underlying set of processes. The second is that some aspects of language can be learned due to domain-general characteristics of memory.