Utility-based evaluation metrics for models of language acquisition: A look at speech segmentation

2015 ◽  
Author(s):  
Lawrence Phillips ◽  
Lisa Pearl
Keyword(s):  
Language Acquisition ◽  
Speech Segmentation ◽  
Evaluation Metrics

Which came first: Infants learning language or motherese?

2004 ◽  
Vol 27 (4) ◽  
pp. 505-506 ◽  
Author(s):  
Heather Bortfeld
Keyword(s):  
Language Acquisition ◽  
Word Recognition ◽  
Language Learning ◽  
Learning Process ◽  
Building Blocks ◽  
Speech Segmentation ◽  
Learning Language

Although motherese may facilitate language acquisition, recent findings indicate that not all aspects of motherese are necessary for word recognition and speech segmentation, the building blocks of language learning. Rather, exposure to input that has prosodic, phonological, and statistical consistencies is sufficient to jump-start the learning process. In light of this, the infant-directedness of the input might be considered superfluous, at least insofar as language acquisition is concerned.

scholarly journals Input and uptake at 7 months predicts toddler vocabulary: the role of child-directed speech and infant processing skills in language development

2015 ◽  
Vol 43 (5) ◽  
pp. 1158-1173 ◽  
Author(s):  
ROCHELLE S. NEWMAN ◽  
MEREDITH L. ROWE ◽  
NAN BERNSTEIN RATNER
Keyword(s):  
Language Acquisition ◽  
Language Development ◽  
Speech Segmentation ◽  
Experimental Paradigm ◽  
Language Outcomes ◽  
Maternal Input ◽  
Independent Effects ◽  
Maternal Child

ABSTRACTBoth the input directed to the child, and the child's ability to process that input, are likely to impact the child's language acquisition. We explore how these factors inter-relate by tracking the relationships among: (a) lexical properties of maternal child-directed speech to prelinguistic (7-month-old) infants (N = 121); (b) these infants' abilities to segment lexical targets from conversational child-directed utterances in an experimental paradigm; and (c) the children's vocabulary outcomes at age 2;0. Both repetitiveness in maternal input and the child's speech segmentation skills at age 0;7 predicted language outcomes at 2;0; moreover, while these factors were somewhat inter-related, they each had independent effects on toddler vocabulary skill, and there was no interaction between the two.

Modeling the contribution of phonotactic cues to the problem of word segmentation

2010 ◽  
Vol 37 (3) ◽  
pp. 487-511 ◽  
Author(s):  
DANIEL BLANCHARD ◽  
JEFFREY HEINZ ◽  
ROBERTA GOLINKOFF
Keyword(s):  
Language Acquisition ◽  
Computational Model ◽  
Experimental Research ◽  
Computational Models ◽  
Word Segmentation ◽  
Experimental Results ◽  
Speech Segmentation ◽  
Advantages And Disadvantages ◽  
Language Universal ◽  
Segmentation Models

ABSTRACTHow do infants find the words in the speech stream? Computational models help us understand this feat by revealing the advantages and disadvantages of different strategies that infants might use. Here, we outline a computational model of word segmentation that aims both to incorporate cues proposed by language acquisition researchers and to establish the contributions different cues can make to word segmentation. We present experimental results from modified versions of Venkataraman's (2001) segmentation model that examine the utility of: (1) language-universal phonotactic cues; (2) language-specific phonotactic cues which must be learned while segmenting utterances; and (3) their combination. We show that the language-specific cue improves segmentation performance overall, but the language-universal phonotactic cue does not, and that their combination results in the most improvement. Not only does this suggest that language-specific constraints can be learned simultaneously with speech segmentation, but it is also consistent with experimental research that shows that there are multiple phonotactic cues helpful to segmentation (e.g. Mattys, Jusczyk, Luce & Morgan, 1999; Mattys & Jusczyk, 2001). This result also compares favorably to other segmentation models (e.g. Brent, 1999; Fleck, 2008; Goldwater, 2007; Johnson & Goldwater, 2009; Venkataraman, 2001) and has implications for how infants learn to segment.

scholarly journals Rapid Statistical Learning Supporting Word Extraction From Continuous Speech

2017 ◽  
Vol 28 (7) ◽  
pp. 921-928 ◽  
Author(s):  
Laura J. Batterink
Keyword(s):  
Language Acquisition ◽  
Statistical Learning ◽  
Essential Role ◽  
Reaction Time Task ◽  
Speech Segmentation ◽  
Learning Effects ◽  
Continuous Speech ◽  
Statistical Structure ◽  
Early Step ◽  
On Line

The identification of words in continuous speech, known as speech segmentation, is a critical early step in language acquisition. This process is partially supported by statistical learning, the ability to extract patterns from the environment. Given that speech segmentation represents a potential bottleneck for language acquisition, patterns in speech may be extracted very rapidly, without extensive exposure. This hypothesis was examined by exposing participants to continuous speech streams composed of novel repeating nonsense words. Learning was measured on-line using a reaction time task. After merely one exposure to an embedded novel word, learners demonstrated significant learning effects, as revealed by faster responses to predictable than to unpredictable syllables. These results demonstrate that learners gained sensitivity to the statistical structure of unfamiliar speech on a very rapid timescale. This ability may play an essential role in early stages of language acquisition, allowing learners to rapidly identify word candidates and “break in” to an unfamiliar language.

scholarly journals Exploring the “anchor word” effect in infants: Segmentation and categorisation of speech with and without high frequency words

PLoS ONE ◽  
2020 ◽  
Vol 15 (12) ◽  
pp. e0243436
Author(s):  
Rebecca L. A. Frost ◽  
Kirsty Dunn ◽  
Morten H. Christiansen ◽  
Rebecca L. Gómez ◽  
Padraic Monaghan
Keyword(s):  
Language Acquisition ◽  
High Frequency ◽  
Speech Segmentation ◽  
Critical Threshold ◽  
Control Group ◽  
Head Turn ◽  
Lexical Categories ◽  
Distributional Information ◽  
Preference Tests ◽  
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.

The bottleneck may be the solution, not the problem

2016 ◽  
Vol 39 ◽  
Author(s):  
Arnon Lotem ◽  
Oren Kolodny ◽  
Joseph Y. Halpern ◽  
Luca Onnis ◽  
Shimon Edelman
Keyword(s):  
Working Memory ◽  
Language Acquisition ◽  
Recent Theory ◽  
Cognitive Mechanisms ◽  
Selection Pressures ◽  
Biological Trait ◽  
Memory Bottleneck

AbstractAs a highly consequential biological trait, a memory “bottleneck” cannot escape selection pressures. It must therefore co-evolve with other cognitive mechanisms rather than act as an independent constraint. Recent theory and an implemented model of language acquisition suggest that a limit on working memory may evolve to help learning. Furthermore, it need not hamper the use of language for communication.

Assessing the Link Between Perception and Production in Cantonese Tone Acquisition

2019 ◽  
Vol 62 (5) ◽  
pp. 1243-1257 ◽  
Author(s):  
Peggy Pik Ki Mok ◽  
Holly Sze Ho Fung ◽  
Vivian Guo Li
Keyword(s):  
Language Acquisition ◽  
General Principle ◽  
Choice Task ◽  
Factors Affecting ◽  
Tone Pair ◽  
Tone Production ◽  
Monosyllabic Words ◽  
Early Production ◽  
Forced Choice Task ◽  
Production Accuracy

Purpose Previous studies showed early production precedes late perception in Cantonese tone acquisition, contrary to the general principle that perception precedes production in child language. How tone production and perception are linked in 1st language acquisition remains largely unknown. Our study revisited the acquisition of tone in Cantonese-speaking children, exploring the possible link between production and perception in 1st language acquisition. Method One hundred eleven Cantonese-speaking children aged between 2;0 and 6;0 (years;months) and 10 adolescent reference speakers participated in tone production and perception experiments. Production materials with 30 monosyllabic words were transcribed in filtered and unfiltered conditions by 2 native judges. Perception accuracy was based on a 2-alternative forced-choice task with pictures covering all possible tone pair contrasts. Results Children's accuracy of production and perception of all the 6 Cantonese tones was still not adultlike by age 6;0. Both production and perception accuracies matured with age. A weak positive link was found between the 2 accuracies. Mother's native language contributed to children's production accuracy. Conclusions Our findings show that production and perception abilities are associated in tone acquisition. Further study is needed to explore factors affecting production accuracy in children. Supplemental Material https://doi.org/10.23641/asha.7960826

Rule Invention in the Acquisition of Morphology Revisited

1988 ◽  
Vol 31 (3) ◽  
pp. 425-431 ◽  
Author(s):  
Stephen M. Camarata ◽  
Lisa Erwin
Keyword(s):  
Language Acquisition ◽  
Active Role ◽  
Acoustic Parameters ◽  
Impaired Child ◽  
Language Impaired ◽  
Intensity Parameters ◽  
Suprasegmental Features ◽  
Acoustic Analyses ◽  
Suprasegmental Cues

This paper presents a case study of a language-impaired child who signaled the distinction between English singular and plural using suprasegmental cues rather than the usual segmental form used within the parent language. Acoustic analyses performed within the first study in the paper revealed that the suprasegmental features used to maintain this distinction included various duration, fundamental frequency, and intensity parameters. Acoustic analyses Were also performed on a set of matched two- and four-item plural forms within a second study. The results of these analyses indicated that the same acoustic parameters were used to distinguish two-item plural forms from four-item plural forms. This case of linguistic creativity is offered as further evidence in support of the model of language acquisition that emphasizes the active role children take in the acquisition process. Additionally, the phonological, morphological, and psycholinguistic factors that may contribute to such rule invention are discussed.

Sign in / Sign up

Export Citation Format

Share Document