ChapterPDF Available

Vocabulary Acquisition

Authors:

Abstract

Acquiring the words of one's language is, in principle, a challenging problem, but children solve it surprisingly easily. This article covers (1) the skills (speech perception, speech segmentation, and speech production) that children must acquire before producing their first words; (2) the main theories attempting to explain how the meaning of words is acquired; (3) the role of working memory in learning words; and (4) the acquisition of vocabulary beyond the early years. The tension between nativist and constructivist theories is discussed, as well as the challenges faced by future research.
1
Gobet, F. (2015). Vocabulary acquisition. In James D. Wright (Ed.), International encyclopedia of the
social & behavioral sciences, 2nd edition, Vol 25 (pp. 226–231). Oxford: Elsevier.
Vocabulary Acquisition, Psychology of
Author Contact Information
Fernand Gobet, PhD
Professor of Cognitive Psychology
Department of Psychological Sciences
University of Liverpool
Liverpool L69 7ZA
United Kingdom
fernand.gobet@liv.ac.uk
Abstract
Acquiring the words of one’s language is in principle a challenging problem, but children
solve it surprisingly easily. This article covers (a) the skills (speech perception, speech
segmentation and speech production) that children must acquire before producing their first
words; (b) the main theories attempting to explain how the meaning of words is acquired; (c)
the role of working memory in learning words; and (d) the acquisition of vocabulary beyond
the early years. The tension between nativist and constructivist theories is discussed, as well
as the challenges faced by future research.
Keywords
associative learning; concepts; communication; development; EPAM-VOC; first language
acquisition; latent semantic analysis; nativist theories of language acquisition; semantics;
speech perception; speech production; syntax; working memory
2
Introduction
First words are typically produced between 10 and 18 months of age. Before reaching this
stage, infants must acquire several competencies in phonemic perception, segmentation and
speech production. In spite of considerable empirical research and theoretical debate
(mostly between nativist and constructivist researchers), we still do not know how children
acquire vocabulary and language more generally. This chapter will consider, in order: (a) the
skills that children must acquire before producing their first words; (b) how the meaning of
words is acquired; (c) the role of working memory in learning words; and (d) the acquisition
of vocabulary beyond the early years.
Acquiring pre-vocabulary competencies
There is now good evidence that language acquisition starts before birth and that infants’
perception of speech is affected by prenatal maternal speech. Three days after birth, infants
already display a preference for familiar speech (Decasper and Spence, 1986). But there is
still a long way to go before speech perception and production skills are sufficiently
developed to enable the production of words. At the very least, before acquiring words,
infants must develop three types of competencies: they must be able to perceive, segment
and produce speech.
Speech perception
A phoneme can be defined as the smallest class of sounds that leads, in a specific
language, to differences in meaning. For example, in English, /l/ and /r/ are two phonemes
because replacing one with the other produces words with different meanings, as is
apparent with “royal” and “loyal”. (Note that phonemes are usually indicated by two slashes.)
By contrast, /l/ and /r/ do not affect meaning in languages such as Japanese and Thai and
thus are two instances of a single phoneme.
3
Infants face two sources of difficulty in finding out which phonemes to use (Rowland, 2014).
First, the sound pattern of phonemes lacks acoustic invariance. Phonemes vary
considerably in the way they are pronounced (the variations are called “phones”): people
speak with different accents, different speeds and different pitches of voice (e.g. male vs.
female voices). In fact, it has been found that even a word as short and common as “the” is
pronounced in more than fifty distinct ways (Waldrop, 1988)! Another source of difficulty is
that infants do not receive phonemes as an orderly sequence of sounds that are
independent. On the contrary, these sounds overlap and interact – a phenomenon named
co-articulation. For example, the phoneme /p/ is pronounced differently in “my pet” and in
“your pet”.
Second, different sets of phonemes are used in different languages, as we have seen with
the example of /l/ and /r/ above. It has been estimated that, in the world’s languages, there
are about 600 different consonants and about 200 different vowels. While most languages
use from twenty to thirty-seven phonemes (Maddieson, 1984), the problem faced by infants
is to identify which phonemes are used in their language. Although the input obviously
provides strong cues, this does not fully solve the problem because some groups of phones
constitute a phoneme in some languages but not in others.
Segmentation
The speech signal is continuous and does not come with indications of where the words start
and end. In this respect, it is different from many written languages that use spaces between
words. Thus, one important and challenging task faced by infants is to segment speech into
words – a task which is known as word segmentation. It is important to note that pauses in
speech do not provide reliable cues to word boundaries, because pauses occur both
between and within words.
Research has investigated the kind of cues that children might use to segment speech, and
we briefly discuss the main candidates (see Ambridge and Lieven, 2011, Rowland, 2014, for
4
a more detailed discussion). A first cue is provided by prosody, which concerns the rhythm,
stress and intonation of speech. Several languages have a characteristic stress pattern. For
example, English has a trochaic pattern, where most words consist of a heavy stress
followed by a light stress (e.g. language, infant). However, several languages do not offer a
regular stress pattern (e.g. Inuktitut, an Inuit language of Canada, and French; Rowland,
2014), so this cue is not sufficient by itself.
We have mentioned earlier that the same phoneme can be pronounced with different
variations (phones) depending on the surrounding phonemes and on the position within a
word. Allophonic variation thus provides cues that can be used by infants for word
segmentation. Another cue is offered by phonotactic regularities (i. e., the acceptable
combinations of phonemes in a given language). Some combinations of sounds are not
allowed within a word but could occur between words (or vice versa), and some
combinations of sounds are more likely to occur at the beginning rather than the end of
words (or vice-versa). Thus, phonotactic cues can help segmentation.
Another source of information is provided by words that occur in isolation. Caregivers
occasionally pronounce words that are clearly segmented. Using as criterion a gap of at
least 300 ms in the speech stream after the preceding word and before the following word,
Brent and Siskind (2001) found that 9 per cent of the words met this criterion and were thus
pronounced in isolation.
A final cue is provided by the transitional probabilities between words. Different syllables
have a different probability of following a given word, and if infants learn even
approximations to these probabilities, they can then use them as information to select the
speech stream. That infants can do this, at least in an experimental situation, was shown by
Saffran, Aslin and Newport (1996), who used a continuous sequence of nonsense syllables
(e.g. bidakupadotigolabubidaku …).
5
Experiments have shown that infants can use the cues that we have reviewed; however, by
themselves, these cues are not sufficient for successfully segmenting the speech stream.
Ambridge and Lieven (2011) have therefore proposed that infants are likely to use multiple
cues. In addition, infants gradually understand that speech conveys meaning and use
meaning to isolate words.
Speech production
While infants are precocious in perceiving speech sounds, production of the sounds of their
language progresses only slowly. This is in part due to the immaturity of their vocal organs
(e.g. the infant’s larynx is located higher than the adult’s, making it impossible to pronounce
some sounds). It is generally admitted that speech production improves through a series of
stages. Here we briefly describe the stages proposed by Vihman (1996) (note that the ages
given are approximate).
In the first stage (reflexive vocalisation, from 0 to 2 months), the sounds produced by infants
include sounds expressing discomfort (e.g. crying), vegetative sounds (e.g. burping) and
sounds related to physical activity (e.g. grunts). The second stage (cooing and laughter, from
2 to 4 months) is characterised by the appearance of comfort sounds, which normally occur
when other people smile and talk. Stage 3 (vocal play, from 4 to 7 months) is characterised
by vocal play sounds (e.g. squeals), which are made possible by the maturation and better
control of articulation. This stage also sees the beginning of primitive babbling, with sounds
starting to resemble consonants and vowels. The final stage (from 7 months on) comprises
two parallel sub-stages. In the reduplicated/canonical babbling sub-stage, starting from
seven months of age, infants repeat speech syllables (e.g. da-da-da). In the variegated
babbling sub-stage, infants produce combinations of syllables (e.g. ka-da-bu-da). This
second sub-stage increasingly dominates the first sub-stage over time.
It is interesting that the first words, which tend to occur at this stage, contain the syllables
ma, pa and na, such as in papa and mama. Jakobson (1962) has argued that these speech
6
sounds, which contain bilabials and the open vowel /a/, are the simplest to produce.
Whether they are really intended to refer to parents, as parents commonly think, is
debatable.
Even after the first words have been produced, it takes a long time for infants to learn to
pronounce words correctly. Lust (2006) has identified four types of phonological processes
that produce errors in early child language: omission of sounds (e.g. “broke” is pronounced
as “bok”); substitution of sounds (e.g. “rabbit” is pronounced as “wabbit”); assimilation of
sounds (e.g. “doggy” is pronounced as “goggy”); and repetition of sounds (e.g. “daddy” is
pronounced as “dada”). Several explanations have been proposed to account for these
errors. First, infants mispronounce some words because they misperceive them (Lust,
2006). Second, infants need time to learn and optimise the way they use their speech
organs to articulate words correctly. Third, the production of correct speech sounds depends
on a maturational program, which is universal (Jakobson, 1941/1968). Finally, pronunciation
depends on learning and correctly using word patterns or word “templates”. When these
templates are used incorrectly, errors occur (Macken, 1995).
Learning the meaning of words
Children’s vocabulary learning begins slowly, but rapidly increases – at the age of 16 months
children know around 40 words (Bates et al., 1994), yet by school age children learn up to
3000 words each year (or just above 8 words every day) and a 17-year-old learns up to
10,000 words each year (or more than 27 words every day), mainly through reading (Nagy
and Herman, 1987).
A classic phenomenon reported in most textbooks on child language is the “vocabulary
spurt”: after having learnt around 50 words, children suddenly and dramatically increase the
rate at which they learn new words. However, it seems that this spurt is a statistical artefact.
When the appropriate statistical analysis is carried out, it appears that, with most children,
there is no sudden increase in learning rate (Ganger and Brent, 2004).
7
At face value linking words to their meaning involves overcoming a number of
insurmountable difficulties, of which two have been classically considered as particularly
critical. The first is the reference problem, discussed at length by the philosopher Quine
(1960). When hearing a word (e.g. “gavagai”) in a given situation (e.g. a rabbit scurries by),
how does a child (or a linguist) know what the referent of gavagai is? Is it the rabbit, its
colour, the action of running, some part of the rabbit (e.g. its ear), the time of the day, the
colour of the grass the rabbit is running on, or an indefinite number of other possible
referents? The second problem is the extension problem, which is actually common to the
acquisition of all concepts, not only those labelled by words. Let us assume that a child
learns that “dog” refers to the family dog (”Snowy”). How does she correctly extend the class
of animals and/or objects to which the label “dog” can be correctly applied? How does she
learn that “dog” not only applies to “Snowy”, but also to “Fido” or “Rintintin”? That it also
applies to labradors, greyhounds and bulldogs, but not to cats and coyotes? That it applies
not only to living dogs but also to dogs made of plastic, wood or metal? And that it does not
apply to the parts of a dog, such as its ears or legs?
While these philosophical questions are interesting, they do not seem to bother infants and
children – as we have just seen, they learn new words and their meaning fairly rapidly. Thus,
an important scientific task is to uncover the mechanisms by which children can avoid the
twin problems of reference and extension. A number of theories have been proposed to
answer this question, and we briefly review six of them.
Theories based on innate constraints
A first class of theories (Markman, 1990, Landau et al., 1988, Golinkoff et al., 1994) argue
that the presence of innate constraints helps infants make the correct inferences when a
new word is met. The details differ between the specific theories, and a large number of
innate constraints have been proposed, but the following constraints provide a good sample
of the kind of innate biases that have been proposed. A first constraint is the whole object
8
assumption: when hearing a word, the child assumes that it refers to an entire object (or
organism) rather than to its parts. A second constraint, the mutual exclusivity assumption,
states that objects should have only one name. Thus, for example, if the child knows that the
family cat’s name is Pepsi, she would assume that words spoken in relation to this cat refer
to something other than its name (e.g. they refer to an action or to a part of the cat). A third
example is the noun bias: infants find nouns easier to learn than other kinds of words. A
fourth constraint is the shape bias: children tend to label objects that have a similar shape
with the same word. As a final example, consider the principle of extendibility: children take it
for granted that words do not refer to a single exemplar, but rather to a category containing
exemplars that share similarity.
Assuming such innate constraints meets with a number of theoretical and empirical
difficulties (Rowland, 2014, Ambridge and Lieven, 2011). Theoretically, some constraints
would actually make it difficult to learn many words. For instance, in our family cat example,
the mutual exclusivity assumption makes it impossible for a child who knows that her cat’s
name is Pepsi to learn that it is a “cat”. Another problem is that some of the biases are not
empirically valid. Consider the noun bias assumption, which implies that nouns are learnt
more quickly than verbs. One argument for this (putative) bias is that nouns tend to refer to
individuals and concrete objects, while verbs are less directly related to perceptual input
(e.g. Gentner, 1982). As it is plausible on biological grounds that more concrete stimuli are
learnt more quickly than more abstract ones, if follows that nouns should be learnt more
quickly. However, Ambridge and Lieven (2011) convincingly argue that, while this is true for
some noun-verb pairs (e.g., ball vs. think), it is not true for others. For example, in the pair
situation vs. eat, the noun is less concrete and thus harder to learn. They further argue that
the empirical evidence for this phenomenon is weak, as most checklists used by parents
contain more nouns than verbs. This bias reflects the statistics of the child’s environment: at
least in English, the number of nouns is about five times larger than the number of verbs.
Nevertheless, the differential number of nouns and verbs creates a methodological problem.
9
Ambridge and Lieven also note that experiments that have found that nouns are learnt more
easily than verbs (e.g. Childers and Tomasello, 2002) suffer from the problem that these
studies typically use nouns as objects, which means that nouns are selected for their
concreteness.
Rather than following innate and universal constraints, it seems that children are sensitive to
the semantic characteristics of the language they are learning. This is nicely illustrated in a
study by Choi and colleagues (1999). While English distinguishes between containment (put
in) and support (put on), Korean distinguishes between tight-fit relations (kkita) and loose-fit
relations (nehta). In a preferential looking study with children aged between 18 and 23
months, Choi et al. found that children’s eye movement behaviour was consistent with the
spatial categories of their own language. Thus, the children followed constraints specific to
their own language rather than following innate and universal constraints.
Social pragmatic theories
As with the previous approach, social pragmatic theories include a variety of specific
accounts (e.g., Bruner, 1978, Nelson, 1985, Tomasello, 2003). However, they agree on the
key assumption: innate constraints are not necessary for acquiring new words as children
can use a number of social cues (e.g. gaze, gestures, movements of the entire body and
facial expressions) to guide their inferences. The concepts of joint attention and
understanding of communicative intention are central in this approach. While these theories
are consistent with a fair amount of empirical data, they also suffer from a few weaknesses.
For example, as noted by Rowland (2014), children can learn words before they have
developed the abilities of joint attention and intention reading. In addition, autistic children
learn words to a level similar to that of normally developing children, although they lack
precisely these abilities.
10
Attentional and associative learning
A third account argues that world learning simply requires the use of general cognitive
mechanisms dealing with attention, memory and learning (Samuelson and Smith, 1998).
Two learning mechanisms are particularly important: associative learning, where an
association between two objects is made, and cross-situational learning, where regularities
are learnt across situations. While this approach can account for several phenomena, it does
not tell the whole story, and it is likely that an account based both on social pragmatics and
attentional/associative learning is necessary (Ambridge and Lieven, 2011).
Syntactic bootstrapping
This approach argues that the meaning of some words, especially verbs, cannot be acquired
through a general mechanism because of their complexity. Therefore, children must use
constraints provided by language itself, and they do so by exploiting the information provided
by syntax (Landau and Gleitman, 1985). Critics of the approach point out that syntax
develops much more slowly than proposed, and thus might be of little use in learning new
words (Tomasello, 2003), and that in some languages syntax actually hinders the acquisition
of word meaning. For example, Margetts (2008) shows that in Saliba, an Oceanic language
spoken in Papua New Guinea, two different verbs express the same meaning (e.g., to give)
but are embedded in completely different syntactic structures.
Emergentist coalition model
The approaches we have just described focus on one kind of mechanism at a time. A natural
extension is to propose an account combining all the mechanisms we have discussed so far.
This is what Hirsh-Pasek and colleagues (2000) did in their emergentist coalition model. The
key assumptions are that (a) children are sensitive to a number of cues, including those we
have reviewed above; (b) the weight of these cues changes with age; and (c) constraints
emerge as language develops. As the model incorporates a large number of mechanisms, it
11
can account for a large amount of empirical data. However, champions of the approaches
we have reviewed above retort that the model lacks parsimony, as word learning can be
explained by one single mechanism rather than by a large number of them.
Latent semantic analysis
Compared to the previous approaches, computational approaches offer a different kind of
explanation of word learning. Perhaps the best example is offered by latent semantic
analysis (LSA) (Landauer and Dumais, 1997), so we focus on this approach. Latent
semantic analysis is a mathematical technique suggesting that word meanings are acquired
by exploiting mutual constraints from the language input. With children, these constraints
come from interacting with the environment and hearing speech. Latent semantic analysis
uses the shortcut of using the paragraphs of a text as a proxy for the episodes we
experience in our lives. The central rationale is that the contexts in which a given word does
or does not appear powerfully constrain and determine word meanings. Essentially, what
LSA is doing is computing correlations between words within different contexts. However, the
similarity values estimated by LSA are not simply based upon co-occurrence frequencies,
but depend on a deeper statistical analysis (they are “latent”).
Concretely, in its computer implementation, latent semantic analysis uses as input a corpus
of text, where paragraphs define the relevant contextual units. The analysis of word
occurrences is carried out across all paragraphs, which makes it possible to compute
vectors that represent the meaning and relationships between words. Vectors are
mathematically well suited to drawing semantic comparisons between words. To reduce
dimensionality, latent semantic analysis uses singular value decomposition, which is a
generalisation of factor analysis. Thus, large matrices of word-by-context data (e.g.
encyclopaedias) can be reduced into 100-500 dimensions. Higher-order co-occurrences are
implicitly taken into account by the mechanisms used by latent semantic analysis. Important
contributions of this approach include the description of the acquisition of vocabulary as the
12
construction of a semantic space and the insight that a considerable amount of “learning”
results from indirect induction. However, LSA suffers from two main weaknesses: it does not
use syntax or morphology and it cannot handle negation.
Role of working memory in learning words
Children differ in the speed with which they acquire vocabulary and the number of words
they know. Baddeley et al. (1998) have proposed that the phonological loop component of
working memory (below, phonological working memory) is the main source of these
individual differences. According to these authors, phonological working memory constitutes
a bottleneck in the learning of new words. This is due to the fact that, compared to children
with low phonological working memory capacity, children with high phonological working
memory capacity can maintain more sound patterns in memory. This allows them to learn
words more rapidly. Baddeley and colleagues largely support their view using empirical data
obtained with the non-word repetition task (Gathercole et al., 1994), which they consider to
be a reliable measure of the capacity of phonological working memory. In this task, the
experimenter presents non-words of varying lengths and asks children to repeat them back
as accurately as they can. One advantage of the non-word repetition task, which uses
strings that respect the phonotactic constraints of the language under study, is that it makes
it possible to gather a large amount of data whilst at the same time making sure that children
do not possess any knowledge of the specific strings that are being used as stimuli.
While influential, this theory suffers from a lack of precision, as it is expressed only verbally.
Several computational models have been developed which provide mechanisms that are
better specified. We consider two of these models, one based on connectionism, the other
based on the mechanism of chunking.
Gupta and Tisdale (2009) used a “simple recurrent network” model, consisting of an input
layer, hidden layer, context layer and output layer. In this model, activation patterns
represent well-formed phonemic sequences as part of syllabic slots. The model receives
13
input one syllable at a time. Simulations of the non-word repetition test showed that the
model captures the overall accuracy patterns, length effects and error types (i.e.,
substitution, insertion and deletion).
Chunking, a mechanism by which units are grouped to build increasingly larger units, is
considered a central mechanism of human learning and cognition, including language
(Gobet et al., 2001). With respect to the acquisition of vocabulary, Jones, Pine and Gobet
(2007) have shown that chunking provides a plausible mechanism for how phonemes are
grouped together to form new words. The specific model they use, EPAM-VOC, provides a
detailed specification of the way long-term memory (where knowledge of vocabulary is
stored) and phonological working memory interact. The model learns by receiving as input
child-directed speech and text from story books written for children. It captures key features
of children’s performance in the non-word repetition task: performance is better for shorter
non-words and for wordlike non-words; performance improves with age; and performance is
superior for single consonant non-words compared to clustered consonant non-words
(Jones et al., 2007). Simulations also suggest that long-term knowledge plays a more
important role in explaining individual differences in the acquisition of vocabulary than the
capacity of working memory (Jones et al., 2008). Finally, Tamburelli and colleagues (2012)
examine the pattern of children’s phonological errors within the syllabic domain. They show
that the model captures the fact that children are more likely to make errors at the syllable
onset than in the coda position. Thus, the model is able to acquire at least some of the
phonotactic rules of English.
Together, the simulations with EPAM-VOC show that the input received by children contains
a substantial amount of information, which can be used by relatively simple learning
mechanisms to extract the kind of statistical information likely to be useful for acquiring
language. They also suggest that developmental changes in the way children perform in the
non-word repetition task may reflect differences in the amount of knowledge in long-term
memory rather than maturational changes in the capacity of working memory.
14
Vocabulary acquisition in later language development
Most developmental psycholinguistics has focused on what happens in the first five years of
life, perhaps because it is during this period that the greatest challenges are met. One
consequence is that there has been little research on vocabulary acquisition in late
childhood, adolescence and adulthood. It is however obvious that there are important
changes after early childhood, and that becoming a proficient speaker involves much more
time than becoming a native speaker (Berman, 2007). With respect to vocabulary, there is
both a quantitative and a qualitative change. Quantitatively, as seen earlier, individuals learn
a considerable number of words during late primary and high school years. Many different
sources make it possible to acquire so many words either incidentally or through instruction:
school activities, reading books, the Internet, cinema, television, radio and interaction with
peers. Qualitatively, the vocabulary becomes more marked (Berman, 2007): it uses longer,
more formal and less frequent words, which often have a highly specialized meaning. In
addition, increased use is made of derivational morphology: more words are formed with the
help of affixes and compounding (e.g. un-grateful-ness). Lexical knowledge becomes more
sophisticated, with an understanding and use of synonymy and polysemy. Around the
beginning of high school, teenagers become more sensitive to linguistic registers and are
able to gracefully and appropriately alternate between casual conversation with all its slang
and colloquialisms, and more formal and academic language. In parallel with these changes
in vocabulary use, other aspects of language also develop considerably: morphosyntax and
metacognition (e.g. critical thinking and use of analogies). This evolution continues well into
adulthood, at least with educated individuals.
Conclusion
Like the majority of research on language development, the field of vocabulary acquisition is
dominated by the debate about whether language-specific innate constraints are necessary
or whether it is sufficient to postulate a combination of general cognitive mechanisms and
15
statistical information in the environment. While the latter theories have gained ground in
recent decades, theories based on innate constraints are still championed vigorously.
However, as witnessed by the many criticisms we have levelled against most of the theories
discussed in this chapter, our understanding of the way children acquire vocabulary is still
poor. As argued by Shatz (2007) for language acquisition in general, what is needed for
scientific breakthrough in the field is a better understanding of the information processing
abilities underpinning vocabulary acquisition as well as an understanding of the multiple
interactions involved, including those between cognitive and social development.
References
AMBRIDGE, B. & LIEVEN, E. V. M. 2011. Child language acquisition: Contrasting theoretical
approaches, Cambridge, Cambridge University Press.
BADDELEY, A. D., GATHERCOLE, S. E. & PAPAGNO, C. 1998. The phonological loop as a
language learning device. Psychological Review, 105, 158-173.
BATES, E., MARCHMAN, V., THAL, D., FENSON, L., DALE, P., REZNICK, J. S., REILLY, J.
& HARTUNG, J. 1994. Developmental and stylistic variation in the composition of
early vocabulary. Journal of Child Language, 21, 85-123.
BERMAN, R. A. 2007. Developing language knowledge and language use across
adolescence. In: HOFF, E. & SHATZ, M. (eds.) Handbook of language development.
London: Blackwell Publishing.
BRENT, M. R. & SISKIND, J. M. 2001. The role of exposure to isolated words in early
vocabulary development. Cognition, 81, B33-B44.
BRUNER, J. S. 1978. The role of dialogue in language acquisition. In: SINCLAIR, A.,
JARVELLA, R. J. & LEVELT, W. J. M. (eds.) The child’s conception of language New
York: Springer-Verlag.
CHILDERS, J. B. & TOMASELLO, M. 2002. Two-year-olds learn novel nouns, verbs, and
conventional actions from massed or spaced exposures. Developmental Psychology,
38, 967-978.
CHOI, S., MCDONOUGH, L., BOWERMAN, M. & MANDLER, J. M. 1999. Early sensitivity to
language-specific spatial categories in English and Korean. Cognitive Development,
14, 241-268.
DECASPER, A. J. & SPENCE, M. J. 1986. Prenatal maternal speech influences newborns
perception of speech sounds. Infant Behavior & Development, 9, 133-150.
GANGER, J. & BRENT, M. R. 2004. Re-examining the vocabulary spurt. Developmental
Psychology, 40, 621-632.
GATHERCOLE, S. E., WILLIS, C. S., BADDELEY, A. D. & EMSLIE, H. 1994. The children’s
test of nonword repetition: A test of phonological working memory. Memory, 2, 103-
127.
GENTNER, D. 1982. Why nouns are learned before verbs: Linguistic relativity versus natural
partitioning. In: KUCZAJ, S. (ed.) Language development: Language, cognition, and
culture. Hillsdale, NJ: Erlbaum.
16
GOBET, F., LANE, P. C. R., CROKER, S., CHENG, P. C. H., JONES, G., OLIVER, I. & PINE,
J. M. 2001. Chunking mechanisms in human learning. Trends in Cognitive Sciences,
5, 236-243.
GOLINKOFF, R. M., MERVIS, C. B. & HIRSH-PASEK, K. 1994. Early object labels: the case
for a developmental lexical principles framework. Journal of Child Language, 21, 125-
155.
GUPTA, P. & TISDALE, J. 2009. Does phonological short-term memory causally determine
vocabulary learning? Toward a computational resolution of the debate. Journal of
Memory and Language, 61, 481-502.
HIRSH-PASEK, K., GOLINKOFF, R. M. & HOLLICH, G. 2000. An emergentist coalition
model for word learning: Mapping words to objects is a product of the interaction of
multiple cues. In: GOLINKOFF, R. M., HIRSH-PASEK, K., BLOOM, L., SMITH, L.,
WOODWARD, A., AKHTAR, N., TOMASELLO, M. & HOLLICH, G. (eds.) Becoming a
word learner: A debate on lexical acquisition. New York: Oxford University Press.
JAKOBSON, R. 1941/1968. Child language, aphasia and phonological universals., The
Hague, NL Mouton.
JAKOBSON, R. 1962. Why 'mama' and 'papa'?". In: JAKOBSON, R. (ed.) Selected writings,
Vol. I: Phonological studies. The Hague, NL: Mouton.
JONES, G., GOBET, F. & PINE, J. M. 2007. Linking working memory and long-term memory:
A computational model of the learning of new words. Developmental Science, 10,
853-873.
JONES, G., GOBET, F. & PINE, J. M. 2008. Computer simulations of developmental change:
The contributions of working memory capacity and long-term knowledge. Cognitive
Science, 32, 1148-1176.
LANDAU, B. & GLEITMAN, L. 1985. Language and experience: Evidence from the blind
child, Cambridge, MA, Harvard University Press.
LANDAU, B., SMITH, L. B. & JONES, S. 1988. The importance of shape in early lexical
learning. Cognitive Development, 3, 299-321.
LANDAUER, T. K. & DUMAIS, S. T. 1997. A solution to Plato’s problem: The latent semantic
analysis theory of acquisition, induction, and representation of knowledge.
Psychological Review, 104, 211-240.
LUST, B. C. 2006. Child language, acquisition and growth, Cambridge, Cambridge University
Press.
MACKEN, M. 1995. Phonological acquisition. In: GOLDSMITH, J. (ed.) The handbook of
phonological theory. Cambridge, MA: Blackwell.
MADDIESON, I. 1984. Patterns of sounds, New York, Cambridge University Press.
MARGETTS, A. 2008. Learning verbs without boots and straps? The problem of give' in
Saliba. In: BOWERMAN, M. & BROWN, P. (eds.) Cross-linguistics perspectives on
argument structure: Implications for learnability. New York: Lawrence Erlbaum.
MARKMAN, E. M. 1990. Constraints children place on word meanings. Cognitive Science,
14, 57-77.
NAGY, W. E. & HERMAN, P. A. 1987. Breadth and depth of vocabulary knowledge:
Implications for acquisition and instruction. In: MCKEOWN, M. G. & CURTIS, M. E.
(eds.) The nature of vocabulary acquisition Hillsdale, NJ: Lawrence Erlbaum
Associates.
NELSON, K. 1985. Making sense: The acquisition of shared meaning, New York, Academic
Press.
QUINE, W. V. 1960. Word and object, Cambridge, MA, Technology Press of MIT.
ROWLAND, C. F. 2014. Understanding child language acquisition, London, Routedge.
SAFFRAN, J. R., ASLIN, R. N. & NEWPORT, E. L. 1996. Statistical learning by 8-month-old
infants. Science, 274, 1926-1928.
SAMUELSON, L. K. & SMITH, L. B. 1998. Memory and Attention Make Smart Word
Learning: An Alternative Account of Akhtar, Carpenter, and Tomasello. Child
Development, 69, 94-104.
17
SHATZ, M. 2007. On the development of the field of language development. In: HOFF, E. &
SHATZ, M. (eds.) Blackwell handbook of language development Oxford: Blackwell.
TAMBURELLI, M., JONES, G., GOBET, F. & PINE, J. M. 2012. Computational modelling of
phonological acquisition: Simulating error patterns in nonword repetition tasks.
Language and Cognitive Processes, 27, 901-946.
TOMASELLO, M. 2003. Constructing a language: A usage-based theory of language
acquisition, Cambridge, MA, Harvard University Press.
VIHMAN, M. M. 1996. Phonological development, Oxford, Blackwell.
WALDROP, M. M. 1988. A landmark in speech recognition. Science, 240, 1615.
... For example, short silences in speech provide unreliable cues since they sometimes occur not only between words, but also within words. It is thought that children use a combination of several cue types to segment speech, including prosody, allophonic variation, phonotactic regularities, transitional probabilities, semantics, and words occasionally occurring in isolation (Ambridge & Lieven, 2011;Rowland, 2014;Gobet, 2015). ...
Article
Full-text available
Vocabulary instruction has received serious attention among English language teaching (ELT) researchers for decades. The objective of this research was to find out whether peer assessment can facilitate vocabulary retention. The peer assessment in the current research was delivered through a workshop activity module as an extension of Moodle platform, a most used open-source LMS (Learning Management System). This study was a quantitative study with an experimental design. The data were collected from 59 adult EFL learners participating in the study. The study used a repeated measure design, and tests were administered after each type of instruction, where traditional vocabulary instruction preceded peer assessment instruction. The scores were analysed using the Independent Samples T-test. The analysis results showed that the scores were significantly different. The scores obtained for vocabulary use after peer assessment instruction with peer review were higher compared to those with traditional vocabulary instruction. Therefore, it can be concluded that peer assessment in Moodle workshop activity module can facilitate sufficient vocabulary exposure for better retention. The pedagogical implications of the research are discussed in the article.
Article
Full-text available
The purpose of this study is to investigate the vocabulary acquisition among English as a second language (ESL) Learners. The study is based on grounded theory research designs. Grounded theory allows the study to generate a far-reaching comprehensible theory founded on behaviourism, cognitivism, and constructivism on the central phenomenon (using online games in vocabulary learning); from the results and findings (Creswell, 2012). A systematic design is applied which highlights the use of data analysis procedures of open, axial, and selective coding, as well as the development of logic paradigm or visual picture(s) from the generated theory (Creswell, 2012). The interview protocol consists of three parts: personal information, advantages and disadvantages factors, as well as possible factors that support or inhibit vocabulary acquisition. Based on the findings, there are basically eight (8) descriptor codes, and thirty (30) in-vivo codes. All of these codes are emerged from the inductive coding method. The Emerging possible categories from the first sample: The possible categories that could be derived from the pattern of codes are: a. Personal preferences & knowledge on computer or online games, b. Habitual playtime on playing computer or online games, c. Online games & learning English vocabulary, and d. Factors that could hinder learning English vocabulary.
Article
Full-text available
There are three objectives of this study. First, to investigate the effectiveness of TTS in Mandarin language reading, find out the shortage of TTS, and understand the difference between machine-reading and human reading. Second, the study concludes that TTS technology is mature enough now but still needs to solve some problems in specific criteria such as conveying humanity elements, putting in more human nature and expressing human emotions in the reading process in the future. Third, teaching mandarin reading is still necessary as the TTS system could not, instead of the importance of a teacher in reading and learning thoroughly.
Article
Full-text available
Behaviour problems in early childhood have a lasting impact on cognitive development and education attainment in later adolescence and into adulthood. Here we address the relationship conduct and hyperactivity problems at school entrance, and vocabulary acquisition in adolescence. We compare performance in identical assessments across two generations of British children born 30 years apart in 1970 (n = 15,676) and 2000/2 (n = 16,628) and find that both conduct and hyperactivity problems have a negative association with later vocabulary in both generations. We take advantage of rich longitudinal birth cohort data and establish that these relationships hold once family socioeconomic status and a child’s personal characteristics and earlier vocabulary acquisition are taken into account. We also find that teenagers today achieved substantively lower scores in the vocabulary assessment compared to their counterparts born 30 years earlier, and that this holds across all categories within each of the family and individual characteristics considered in this article. As vocabulary and language skills are key prerequisites for wider learning, we discuss implications the findings have for education policies.
Article
Teaching English to young learners, especially for kindergarten students, is focused on teaching vocabulary because they are still introduced to new language in which it has not ever been learned before. The use of songs in teaching vocabulary for young learners is an effective way to teach English. This study is descriptive qualitative research because (1) It is concerned with context and meaning, (2) the researcher works in natural setting, (3) the researcher is the key instrument in collecting data, (4) the data are presented descriptively, and (5) the data are analyzed inductively. The subjects are kindergarten students and the English teacher. In this study, the instruments used are observation checklist, field note, and interview guide for the four teachers. From the findings, it can be concluded the characteristics of the songs used in teaching vocabulary were the lyrics of the songs were simple and not too long, there was the repetition of the words while singing the songs, the vocabulary used in the lyrics was presented in the meaningful context, the lyrics were based on the theme in the school curriculum, and the rhythm of the songs was fun, The teachers also considered the criteria of the appropriate songs for their students while they were selecting the songs.
Article
Full-text available
Increasing working memory (WM) capacity is often cited as a major influence on children's development and yet WM capacity is difficult to examine independently of long-term knowledge. A computational model of children's nonword repetition (NWR) performance is presented that independently manipulates long-term knowledge and WM capacity to determine the relative contributions of each in explaining the developmental data. The simulations show that (a) both mechanisms independently cause the same overall developmental changes in NWR performance, (b) increase in long-term knowledge provides the better fit to the child data, and (c) varying both long-term knowledge and WM capacity adds no significant gains over varying long-term knowledge alone. Given that increases in long-term knowledge must occur during development, the results indicate that increases in WM capacity may not be required to explain developmental differences. An increase in WM capacity should only be cited as a mechanism of developmental change when there are clear empirical reasons for doing so.
Article
Full-text available
The nonword repetition (NWR) test has been shown to be a good predictor of children's vocabulary size. NWR performance has been explained using phonological working memory, which is seen as a critical component in the learning of new words. However, no detailed specification of the link between phonological working memory and long-term memory (LTM) has been proposed. In this paper, we present a computational model of children's vocabulary acquisition (EPAM-VOC) that specifies how phonological working memory and LTM interact. The model learns phoneme sequences, which are stored in LTM and mediate how much information can be held in working memory. The model's behaviour is compared with that of children in a new study of NWR, conducted in order to ensure the same nonword stimuli and methodology across ages. EPAM-VOC shows a pattern of results similar to that of children: performance is better for shorter nonwords and for wordlike nonwords, and performance improves with age. EPAM-VOC also simulates the superior performance for single consonant nonwords over clustered consonant nonwords found in previous NWR studies. EPAM-VOC provides a simple and elegant computational account of some of the key processes involved in the learning of new words: it specifies how phonological working memory and LTM interact; makes testable predictions; and suggests that developmental changes in NWR performance may reflect differences in the amount of information that has been encoded in LTM rather than developmental changes in working memory capacity.
Article
This paper views lexical acquisition as a problem of induction: Children must figure out the meaning of a given term, given the large number of possible meanings any term could have. If children had to consider, evaluate, and rule out an unlimited number of hypotheses about each word in order to figure out its meaning, learning word meanings would be hopeless. Children must, therefore, be limited in the kinds of hypotheses they consider as possible word meanings. This paper considers three possible constraints on word meanings: (1) The whole object assumption which leads children to interpret novel terms as labels for objects—not parts, substances, or other properties of objects; (2) The taxonomic assumption which leads children to consider labels as referring to objects of like kind, rather than to objects that are thematically related; and (3) The mutual exclusivity assumption which leads children to expect each object to have only one label. Some of the evidence for these constraints is reviewed.
Article
Two general types of accounts have been offered to explain the smartness of young children's word learning. One account postulates that children enter the word-learning task with specific knowledge about how words link to categories. The second account puts the source of children's smart word learning in knowledge about the pragmatics of communication and social interactions. The present experiment tested a third idea: that children's seemingly smart word learning derives from general, indeed mundane, cognitive processes. Forty-eight children from 18 to 28 months of age participated in a task designed to test our alternative explanation as applied to Akhtar, Carpenter, and Tomasello's (1996) finding that children use knowledge of the communicative intents of others to interpret a novel noun. Specifically, we suggest that children's attention to the proper referent was guided by the general effects of a contextual shift on memory and attention. The procedure in the present study was identical to that of Akhtar et al. except that we differentiated the target through a nonsocial context shift. Findings similar to that of Akhtar et al. emerged under the present procedures. These results strongly suggest that general attentional and memorial processes, and not knowledge about the communicative intents of others, may guide young children's word learning. These findings provide one demonstration of how smart word learning may emerge from more ordinary (and dumb) cognitive processes.