See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/319051251

Theories of Language Development

Chapter · January 2017

DOI: 10.1007/978-1-4614-6435-8_543-3

CITATIONS READS
2 43,446

1 author:

Inge-Marie Eigsti
University of Connecticut
182 PUBLICATIONS   5,937 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Optimal Outcome in Children with Autism Spectrum Disorder View project

Perception-Action Links in ASD View project

All content following this page was uploaded by Inge-Marie Eigsti on 23 October 2018.

The user has requested enhancement of the downloaded file.

Dr. Inge-Marie Eigsti
Theories of Language Development SpringerReference

Theories of Language Development

Definition
Language acquisition presents children with what seems like an important but near-impossible challenge: how to
determine the mappings between meanings in their world and the complex acoustic signal of speech. Children must
segment the running stream of sounds that they hear into words and phrases and determine the regularities, or
systematicities, of the language. This is particularly challenging because of the "many-to-one" mapping problem. That is,
a particular word can be conveyed with a large variety of different acoustic signals; this allows us to understand people
with accents, or who have peanut butter in their mouths, but has made machine language comprehension a strikingly
difficult task. Additionally, speakers do not pause between words; thus, the learner simply hears a continuous acoustic
stream. Children accomplish this feat of learning in the absence of direct instruction and even (e.g., in the case of Quiche
Mayan) while hearing little language directed at themselves (Pye, 1986). As such, the acquisition of human language may
be the single most impressive human intellectual accomplishment (Pinker, 1984).
There are five systematic features of language that a learner must master in order to understand and produce a spoken
or signed language. Note that while this review focuses on spoken languages, there is such significant overlap with
signed languages; see the American Sign Language entry for more information. First, they must learn the phonology -
that is, the inventory of sounds that make up the language. Sounds are considered to be distinct phonemes when they
are "minimally contrastive," such that two words which differ only by that sound and which are recognized as distinct
words ("pin" vs. "pen" is a minimal pair in English, whereas "pint" vs. "pint" with a final "t" accompanied by a puff of air is
non-contrastive). Children learning English must master an inventory of 12 distinct vowels and 24 consonants (depending
on what sounds are included); classical Arabic, 6 vowels and 30 consonants; and Swedish, 17 vowels and 18
consonants. The phonology of a language corresponds to a degree to its written alphabet but is not identical (and indeed,
in English, the alphabet/phonology correspondence is particularly weak, such that the acquisition of reading in English is
a particularly lengthy and error-prone process).
Second, the learner must learn an inventory of words; that is, she must learn what concepts are encoded as words. The
acquisition of a language's vocabulary, or lexicon, is a challenging proposition; some languages convey concepts that, in
other languages, are unspecified (e.g., the concepts melancholy and gloomy are conveyed by a single term in many other
languages). Typically developing children produce first words by around 10-11 months, and typically, these early words
consist of names for specific individuals (mama), objects (dog), or substances (juice); other early-emerging word types
include action verbs (give), adjectives (pretty), and some social expressions (hi, no). Abstract terms, such as mental state
verbs, and those expressing non-present objects or events, are produced later. Children bring a set of useful constraints
or biases to this learning problem; for example, they are guided by an assumption of "mutual exclusivity," which suggests
that novel concepts should be expressed using novel words (Markman, 1990). Recent work indicates that children with
autism spectrum disorders share at least some of these important biases (de Marchena, Eigsti, Worek, Ono, & Snedeker,
2011).
Third, the learner must determine how words are to be meaningfully combined in the language, a system known as
syntax (or sometimes, grammar). That is, in English, a descriptive noun phrase might consist of a determiner (e.g., the),
an adjective (fuzzy), and a noun (rabbit); the combinatorial rules of language are known as syntax. During the second and
third years of life, children's utterances increase in length and complexity and increasingly incorporate function words (if,
the, over); they also incorporate advanced syntactic devices such as relative clauses ( the cookie that I ate yesterday),
yes/no questions (Can I have that cookie?), and if/then statements (If I eat my dinner, can I have a cookie?).
Fourth, children must learn how to mark tense, gender, number, aspect, register (formality or politeness), and so on, via
either phrasal combinations or via morphological markers. Morphemes are the smallest meaningful units of language;
morphological development refers to the development and understanding of these units and how such units are combined
into words. For example, jumper, jumped, jumps, and jumpy, all use the morpheme "jump" in combination with
morphemes -er, -ed, -s, and -y. In many languages (typically those where word order is free to vary), syntactic information
is conveyed primarily through morphological markers; thus, many theories of language structure collapse morphology and
syntax into a single process.
Finally, in addition to phonology, lexical, syntactic, and morphological development, children must learn to use language
as a social tool. Pragmatic language refers to a speaker's ability to select words that are appropriate to the social context

http://www.springerreference.com/index/chapterdbid/334364 12 Dec 2012 15:37

© Springer-Verlag Berlin Heidelberg 2012 1
Dr. Inge-Marie Eigsti
Theories of Language Development SpringerReference

("register," i.e., speaking more formally to a professor than to a peer, using taboo words in appropriate situations);
negotiating when it is one's turn in the conversation; the choice of referential expressions (a vs the); use of indirect
speech (it's very cold in here), sarcasm (nice shirt, Mom), or metaphor (you had that one hidden up your sleeve); repair of
miscommunications; the use of prosody (the melody or rhythm of the voice); the use of perspective in language (he
thought that she was going home); and nonverbal linguistic functions (e.g., eye contact for conversation management,
body language, and gestures). Discourse is a closely related concept, which refers to the production of longer connected
streams of speech, as in the telling of a story or joke. Narrative ability is important for communication as well as for the
structuring of one's own thoughts.

Historical Background
According to Herodotus (Fromkin, Krashen, Curtiss, Rigler, & Rigler, 1974), an Egyptian pharaoh, Psammetichus the
First, conducted one of the first known experiments on language acquisition. Hoping to uncover the "original" language,
he caused two newborn babies to be left with an isolated shepherd, who was ordered not to speak with them. The
shepherd monitored their speech sounds; as the first word uttered sounded like the Phrygian word for "bread,"
Psammetichus concluded that Phrygian was the original or primary language. Later work has also drawn the experiences
of feral or, wild-reared, children in order to explore the process of language acquisition in the absence of or given
inadequate language input (Candland, 1993); one challenge is to disentangle language input from atypical social
experiences.
Another classic formulation described in theories of language acquisition is a problem posed by the philosopher Quine,
commonly known as the "gavagai" thought experiment (Quine, 1960). Quine asks us to imagine standing in an alien
world, next to a native from that world; when a rabbit runs past, the native points and says, " gavagai!" While gavagai
might refer to the concept rabbit, there is a nearly infinite number of alternative interpretations, such as brown, furry,
four-legged, speedy, hopping, delicious, watch out! or, even, an assembly of rabbit parts. How can a learner determine
what semantic concepts are conveyed, when they happen to hear an isolated word spoken?
Faced with the challenges of learning what sounds should cohere to form words, the challenge of determining word
meanings, and the challenge of learning how words should be further combined, one might assume that an important
component of language acquisition is the instruction of adults. Indeed, a typical educated parent in the USA may act
according to the assumption that he or she is teaching an infant to speak. Somewhat surprisingly, however, children
receive very little explicit instruction in their language; what is more, they appear to receive little negative (error-correcting)
feedback about their own speech. That is, when children produce a speech error, it is rarely explicitly corrected, and
children are famously resistant to such corrections. McNeill reports a now-famous anecdote about this resistance to error
correction (McNeill, 1970, p. 106): Child: Nobody don't like me. Mother: No, say "nobody likes me." Child: Nobody don't
like me. (eight repetitions). Mother: Listen carefully: say, "nobody likes me." Child: Oh! Nobody don't likes me.
Among the many methodological innovations in child language research, one bears particular mention: the language
database approach. Research in this field has often made use of "diary studies" in which investigators track the
spontaneous utterances produced by an individual child over time. This is a valuable resource for that individual, but it
becomes a highly useful resource if it is available to others. The CHILDES website project has assembled written
transcripts of children's speech from 34 languages, several clinical populations, and children in bilingual acquisition
contexts. This publically available, searchable, and downloadable database has been used as a resource in more than
3,000 publications from its inception through 2008. It can be accessed at http://childes.psy.cmu.edu/. MacWhinney, who
is the principal investigator on the CHILDES database, spearheaded the development of a language transcription system
(CLAN), also available on that website.

Current Knowledge
As reviewed above, the challenge of learning a language is a formidable one. Given the difficulties, it is remarkable how
uniform the process of acquisition is. It follows a similar developmental progression across children, despite quite striking
differences in the structure of the language being learned, culture, individual differences in intelligence or sociability,
parent factors, and so on. That said, however, not all children develop functional language skills; in addition to the clinical

http://www.springerreference.com/index/chapterdbid/334364 12 Dec 2012 15:37

© Springer-Verlag Berlin Heidelberg 2012 2
Dr. Inge-Marie Eigsti
Theories of Language Development SpringerReference

implications of this fact, the study of delays and deficits in acquisition can help to elucidate the nature of the language
acquisition process, by throwing into sharper relief the developmental course of language acquisition (Cicchetti &
Rogosch, 1996; Curtiss, Katz, & Tallal, 1992).
One of the primary theories of language acquisition is associated with "nativist" approaches to human development. The
enormous complexity of language, and the relatively limited direct "teaching" of language (along with the uniformity of
acquisition of language milestones), has led scholars in the nativist or innate school to propose that language
development proceeds according to the growth or unfolding of an innate language faculty, just as physical development
proceeds in the absence of specific instruction or special input (e.g., Chomsky, 1986; Fodor, 1983). Chomsky formulated
one of the most influential theories of language that applies to the nativist approach, the principles and parameters
framework (and its subsequent iterations and revisions, government and binding and the minimalist approach). In this
model, humans have innately specified knowledge about language: a set of principles shared by all human languages
(e.g., that all grammatical sentences must have a (possibly covert) subject) and a set of parameters that specify how a
given language instantiates specific syntactic qualities (e.g., that the subject should be overtly expressed, or not);
parameters are essentially binary decision processes. Under this account, children require little language input; they only
require enough input to direct them in how to set those particular parameters.
The "constructionist" perspective provides an alternative perspective (see Elman, et al., 1996); in this account, general
and nonlinguistic learning capacities may account for much or all of language knowledge. Although support for related
theoretical positions dwindled after Chomsky's (1959) influential and devastating critique of behaviorism, more recent
research suggests an important role for nonlinguistic learning capacities such as tracking the statistics of words or sounds
in language input (Saffran, Aslin, & Newport, 1996). The history of research in language development has been marked
by disagreement between the nativist and constructionist positions; currently, many researchers in the field accept some
middle ground between the strict nativist and the purely empiricist approaches, suggesting that children may draw on
some innately specified biases as well as on domain-general learning processes.
There are several important models that fall in the middle ground of, or otherwise negotiate between, the nativist and
empiricist perspectives. An example of one usage-based account, the "verb island" hypothesis (Tomasello, 2000), for
example, suggests that children acquire syntactic knowledge gradually, item by item, and that much of their early
knowledge consists of memorized two- or three-word structures (with relatively little transfer of morphological structure
across verbs). Experiments presenting children with novel (nonsense) verbs indicate a sharp change between ages 3 and
4. Specifically, children ages 3-4 are able to extend novel verbs to abstract syntactic categories; for example, they are
able to use a verb heard only in a passive sentence frame and utilize it in an active sentence frame. However, children
ages 2-3 years had difficulty extending verbs between transitive and intransitive frames.
Connectionist approaches to acquisition have also provided a set of tools for addressing the trade-off between the
apparent poverty of syntactic input to the learner (i.e., the "learnability" problem) and the role of innate or structures and
processes in development; thus, while often described as simply a version of empiricism, connectionism and empiricism
are not necessarily identical (Plunkett, 1995). "Connectionism" refers to computational models (either theoretical or
actually implemented using computer learning algorithms) that take input (typically, somewhat abstracted in nature from
actual language input) and feeds that input forward (e.g., through a single layer of units or nodes) to a set of output units.
For example, a network model of past tense formation in English might receive as input an abstracted phonological
representation of the verb stem (jump, swim); activation (i.e., which units are ready to "fire" at a given time) passes from
input units to a connected set of output units; those units receiving the strongest input activations are most likely to be
active in the output (e.g., the corresponding past tense form: jumped, swam). The output activations are compared to the
target activation pattern (e.g., the correct past tense form), and any discrepancies form an error signal which influences
the network connections. Connectionist models have helped demonstrate "proof of principle" that a single learning system
is capable of generating such apparently distinct outputs as the regular past tense form (− ed) along with the irregular
forms. Connectionist models have also helped to generate and test models of concept formation and vocabulary
acquisition (see Plunkett, 1995). Work of the past decade has examined children's ability to make use of distributional
aspects of language input (e.g., Saffran, Pollak, Seibel, & Shkolnik, 2007) and demonstrated both the availability of
potentially informative information in the statistics of language input, as well as learners' ability to access that information.

Future Directions
Researchers in the field of language acquisition have often turned to ASD as a sort of "natural laboratory" in which to test

http://www.springerreference.com/index/chapterdbid/334364 12 Dec 2012 15:37

© Springer-Verlag Berlin Heidelberg 2012 3
Dr. Inge-Marie Eigsti
Theories of Language Development SpringerReference

theories of acquisition because studies of this disorder offer the possibility of examining meaningful differences in ability
across a wide range of language, social, and cognitive domains. At the same time, research in this area must grapple with
the subtleties in performance and ability that are part and parcel of a developmental disorder; development is rarely as
neatly "packaged" as we might hope. For example, in typical development, comprehension seems to outpace production
in early language acquisition; thus, children who do not produce function words (e.g., determiners) show effects in
comprehension and memory when those elements are left out, suggesting some sensitivity and awareness of these
elements. Children with ASD appear to provide a distinct picture, with some data suggesting that their production may
outpace their comprehension. Research in the field must reconcile this somewhat bewildering inconsistency with a finding
in typical development that is, as yet, unexplained.

See Also
American Sign Language (ASL)
Grammar
Language
Language Acquisition
Mean Length of Utterance (MLU)
Phonology
Pragmatics
Prosody
Psycholinguistics
Syntax

References and Readings

Candland, D. (1993). Feral children and clever animals: Reflections on human nature. New York: Oxford.
Chomsky, N. (1959). Review of Skinner's "verbal behavior". Language, 35, 26-58.
Chomsky, N. (1986). Knowledge of language: Its nature, origin, and use. New York: Praeger.
Cicchetti, D., & Rogosch, F. (1996). Developmental pathways: Diversity in process and outcome. Development
and Psychopathology, 8(4), 597-896.
Curtiss, S., Katz, W., & Tallal, P. (1992). Delay versus deviance in the language acquisition of language-impaired
children. Journal of Speech and Hearing Research, 35(2), 373-383.
de Marchena, A., Eigsti, I. M., Worek, A., Ono, K. E., & Snedeker, J. (2011). Mutual exclusivity in autism spectrum
disorders: Testing the pragmatic hypothesis. Cognition, 119, 96-113.
Elman, J. L., Bates, E. A., Johnson, M. H., Karmiloff-Smith, A., Parisi, D., & Plunkett, K. (1996). Rethinking
innateness: A connectionist perspective on development. Cambridge, MA: MIT Press.
Fodor, J. A. (1983). The modularity of mind. Cambridge, MA: MIT Press.
Fromkin, V., Krashen, S., Curtiss, S., Rigler, D., & Rigler, M. (1974). The development of language in Genie: A
case of language acquisition beyond the "critical period". Brain and Language, 1(1), 81-107.
McNeill, D. (1970). The acquisition of language: The study of developmental psycholinguistics. New York: Harper
& Row.
Markman, E. M. (1990). Constraints children place on word meanings. Cognitive Science, 14, 57-77.
Pinker, S. (1984). Language learnability and language development. Cambridge, MA: Harvard University Press.
Plunkett, K. (1995). Connectionism and language acquisition. In P. Fletcher & B. MacWhinney (Eds.), The
handbook of child language (pp. 36-72). Cambridge, MA: Blackwell.
Pye, C. (1986). Quiche Mayan speech to children. Journal of Child Language, 13, 85-100.
Quine, W. V. (1960). Word and object. Cambridge, MA: MIT Press.
Saffran, J. R., Aslin, R. N., & Newport, E. L. (1996). Statistical learning by 8-month-old infants. Science, 274,
1926-1928.
Saffran, J. R., Pollak, S. D., Seibel, R. L., & Shkolnik, A. (2007). Dog is a dog is a dog: Infant rule learning is not
specific to language. Cognition, 105, 669-680.
Tomasello, M. (2000). The item-based nature of children's early syntactic development. Trends in Cognitive

http://www.springerreference.com/index/chapterdbid/334364 12 Dec 2012 15:37

© Springer-Verlag Berlin Heidelberg 2012 4
Dr. Inge-Marie Eigsti
Theories of Language Development SpringerReference

Science, 4, 156-163.

Theories of Language Development

Dr. Inge-Marie Department of Psychology, University of Connecticut Haskins Laboratories,

Eigsti New Haven, Storrs, USA

DOI: 10.1007/SpringerReference_334364
URL: http://www.springerreference.com/index/chapterdbid/334364
Part of: Encyclopedia of Autism Spectrum Disorders

Editor: Dr. Fred R. Volkmar

PDF created on: December, 12, 2012 15:37

© Springer-Verlag Berlin Heidelberg 2012

http://www.springerreference.com/index/chapterdbid/334364 12 Dec 2012 15:37

© Springer-Verlag Berlin Heidelberg 2012 5
View publication stats