Disambiguating durational cues for speech segmentation

Exploring the "anchor word" effect in infants: Segmentation and categorisation of speech with and without high frequency words

High frequency words play a key role in language acquisition, with recent work suggesting they may serve both speech segmentation and lexical categorisation. However, it is not yet known whether infants can detect novel high frequency words in continuous speech, nor whether they can use them to help learning for segmentation and categorisation at the same time. For instance, when hearing "you eat the biscuit", can children use the high-frequency words "you" and "the" to segment out "eat" and "biscuit", and determine their respective lexical categories? We tested this in two experiments. In Experiment 1, we familiarised 12-month-old infants with continuous artificial speech comprising repetitions of target words, which were preceded by high-frequency marker words that distinguished the targets into two distributional categories. In Experiment 2, we repeated the task using the same language but with additional phonological cues to word and category structure. In both studies, we measured learning with head-turn preference tests of segmentation and categorisation, and compared performance against a control group that heard the artificial speech without the marker words (i.e., just the targets). There was no evidence that high frequency words helped either speech segmentation or grammatical categorisation. However, segmentation was seen to improve when the distributional information was supplemented with phonological cues (Experiment 2). In both experiments, exploratory analysis indicated that infants' looking behaviour was related to their linguistic maturity (indexed by infants' vocabulary scores) with infants with high versus low vocabulary scores displaying novelty and familiarity preferences, respectively. We propose that high-frequency words must reach a critical threshold of familiarity before they can be of significant benefit to learning.

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).

Mathematical Modeling of Learning from an Inconsistent Source: A Nonlinear Approach

Continuing the discussion of how children can modify and regularize linguistic inputs from adults, we present a new interpretation of existing algorithms to model and investigate the process of a learner learning from an inconsistent source. On the basis of this approach is a (possibly nonlinear) function (the update function) that relates the current state of the learner with an increment that it receives upon processing the source's input, in a sequence of updates. The model can be considered a nonlinear generalization of the classic Bush-Mosteller algorithm. Our model allows us to analyze and present a theoretical explanation of a frequency boosting property, whereby the learner surpasses the fluency of the source by increasing the frequency of the most common input. We derive analytical expressions for the frequency of the learner, and also identify a class of update functions that exhibit frequency boosting. Applications to the Feature-Label-Order effect in learning are presented.

Canalization of Language Structure From Environmental Constraints: A Computational Model of Word Learning From Multiple Cues

There is substantial variation in language experience, yet there is surprising similarity in the language structure acquired. Constraints on language structure may be external modulators that result in this canalization of language structure, or else they may derive from the broader, communicative environment in which language is acquired. In this paper, the latter perspective is tested for its adequacy in explaining robustness of language learning to environmental variation. A computational model of word learning from cross‐situational, multimodal information was constructed and tested. Key to the model's robustness was the presence of multiple, individually unreliable information sources to support learning. This “degeneracy” in the language system has a detrimental effect on learning, compared to a noise‐free environment, but has a critically important effect on acquisition of a canalized system that is resistant to environmental noise in communication.

Domain-general mechanisms for speech segmentation: The role of duration information in language learning

Speech segmentation is supported by multiple sources of information that may either inform language processing specifically, or serve learning more broadly. The Iambic/Trochaic Law (ITL), where increased duration indicates the end of a group and increased emphasis indicates the beginning of a group, has been proposed as a domain-general mechanism that also applies to language. However, language background has been suggested to modulate use of the ITL, meaning that these perceptual grouping preferences may instead be a consequence of language exposure. To distinguish between these accounts, we exposed native-English and native-Japanese listeners to sequences of speech (Experiment 1) and nonspeech stimuli (Experiment 2), and examined segmentation using a 2AFC task. Duration was manipulated over 3 conditions: sequences contained either an initial-item duration increase, or a final-item duration increase, or items of uniform duration. In Experiment 1, language background did not affect the use of duration as a cue for segmenting speech in a structured artificial language. In Experiment 2, the same results were found for grouping structured sequences of visual shapes. The results are consistent with proposals that duration information draws upon a domain-general mechanism that can apply to the special case of language acquisition.

This study shows that adults prefer to group both sequences of shapes and sequences of speech with the final item as the one that has the longest duration. This suggests that final-item duration increase is a helpful cue for perceptual grouping in multiple domains—not just language processing. By testing native speakers of languages that use duration differently, we also show that this grouping preference does not arise from language experience.

Regularization of languages by adults and children: A mathematical framework

The fascinating ability of humans to modify the linguistic input and "create" a language has been widely discussed. In the work of Newport and colleagues, it has been demonstrated that both children and adults have some ability to process inconsistent linguistic input and "improve" it by making it more consistent. In Hudson Kam and Newport (2009), artificial miniature language acquisition from an inconsistent source was studied. It was shown that (i) children are better at language regularization than adults and that (ii) adults can also regularize, depending on the structure of the input. In this paper we create a learning algorithm of the reinforcement-learning type, which exhibits patterns reported in Hudson Kam and Newport (2009) and suggests a way to explain them. It turns out that in order to capture the differences between children's and adults' learning patterns, we need to introduce a certain asymmetry in the learning algorithm. Namely, we have to assume that the reaction of the learners differs depending on whether or not the source's input coincides with the learner's internal hypothesis. We interpret this result in the context of a different reaction of children and adults to implicit, expectation-based evidence, positive or negative. We propose that a possible mechanism that contributes to the children's ability to regularize an inconsistent input is related to their heightened sensitivity to positive evidence rather than the (implicit) negative evidence. In our model, regularization comes naturally as a consequence of a stronger reaction of the children to evidence supporting their preferred hypothesis. In adults, their ability to adequately process implicit negative evidence prevents them from regularizing the inconsistent input, resulting in a weaker degree of regularization.

When learners surpass their models: mathematical modeling of learning from an inconsistent source

It has been reported in the literature that both adults and children can, to a different degree, modify and regularize the often-inconsistent linguistic input they receive. We present a new algorithm to model and investigate the learning process of a learner mastering a set of (grammatical or lexical) forms from an inconsistent source. The algorithm is related to reinforcement learning and drift-diffusion models of decision making, and possesses several psychologically relevant properties such as fidelity, robustness, discounting, and computational simplicity. It demonstrates how a learner can successfully learn from or even surpass its imperfect source. We use the data collected by Singleton and Newport (Cognit Psychol 49(4):370-407, 2004) on the performance of a 7-year-boy Simon, who mastered the American Sign Language (ASL) by learning it from his parents, both of whom were imperfect speakers of ASL. We show that the algorithm possesses a frequency boosting property, whereby the frequency of the most common form of the source is increased by the learner. We also explain several key features of Simon's ASL.