How is Language Acquisition studied?
As we have mentioned, language acquisition is an interesting topic; however, it is only recently that real linguistic research has been generated.
Early studies were diaries of spontaneous speech where the researcher would observe the child speaking in context. Conclusions about language acquisition would then be drawn from this data. This method has advantages and disadvantages.
In the early 20th Century, a huge emphasis was placed on Behaviourist approaches to explaining language acquisition which led to large group studies involving systematic/controlled tests 鈥 The most influential behaviourist theory is that of Operant Conditioning proposed by Skinner (1950s) who generalised his thorough research on rats to explain how children acquire language. (Click here to learn more about Skinner鈥檚 theory)
Towards the late 1970s, linguists began to use more experimental/strict observational techniques. Researchers became more interested, not in the child鈥檚 actual production of language but of their comprehension, for we have to assess what children know about language by the way they use it, i.e., language competence is usually reflected in language performance. When researching child language, researchers are interested in looking at three things: grammaticality judgement, production and comprehension.
The table below summarises the methods used by linguists in order to test these aspects.
Type of Language Performance |
Method |
| Grammaticality Judgement | Sentence Possibilities
The experimenter would produce a number of grammatical and ungrammatical sentences to the child, asking them which ones are 鈥榦kay鈥 e.g., 鈥滼essica is a girl.鈥 Is that sentence okay?鈥 鈥滼essica are a girl.鈥 Is that sentence okay?鈥 |
| Production | Spontaneous Speech (Natural Speech Sampling)
This method is a naturalistic observation as opposed to an experiment. The researcher would record and transcribe the child鈥檚 speech, focusing on a particular aspect of production. Researchers might then want to share their data or compare it to other data in order to test the generalizability of their work. This is made possible by CHILDES (Child Language Data Exchange System: MacWhinney and Snow 1985 and 1990) which maintains copies of various researcher鈥檚 transcriptions of child鈥檚 speech.[1] Elicited Production (EP) Experimental methods can extract language production in a controlled fashion which allows the researcher to target particular aspects of linguistic knowledge. The experimenter might provide the children with a fictional context, provoking them to give responses with an inverted question (Thorton 1996): Experimenter: In this story, the crane is tickling one of the zebras. Ask the puppet which one. Child: Which zebra is getting tickles by the crane? This method, like natural speech sampling provide detailed examples of a child鈥檚 ability to create and organise language. Another EP method focuses on the extent to which children can be 鈥榣ed鈥 to a particular language form. Potts et al. (1979) studied relative clauses, where the child was asked to complete the experimenter鈥檚 sentence: Experimenter: Some children walk to school and some children ride the bus. These children鈥 Child: 鈥 who ride the bus, that ride. Elicited Imitation (EI) In EI tasks, children are required to imitate sentences which have been specifically designed to test particular principles. In EI, the analysis of children鈥檚 鈥渆rrors鈥 may reveal their linguistic understanding. Slobin and Welsh (1973) provide an example of this: Experimenter: [Mozart [who cried]] came to my party Child: [Mozart cried] and [he came to my [party] This example shows that the child is maintaining the meaning whilst abandoning the form (relative clause embedding is converted to a coordinate structure) implying the difficulty children have with embedded clauses. Experimenter: [[The blue shoes] and [blue pencils]] are here Child: [blue pencil are here] and [a blue pencil are here] This example shows that the child is maintaining the form whilst abandoning the meaning. |
| Comprehension | Act Out/ Toy-Moving Task
This method is useful for testing language comprehension because of its simplicity and game like nature. The children may be asked to 鈥榓ct out鈥 a story presented to them by using dolls. The child鈥檚 performance of such tasks provides rich information about whether they comprehend the language they are being asked to interpret. Truth Value Judgement Task This method might include an adult acting out a situation or providing a picture showing a particular action. The child will simply be asked to judge whether this is a possible interpretation for a particular sentence by answering either 鈥榶es鈥 or 鈥榥o鈥. Like the act out tasks, the Truth Value Judgement Task can test for possible limitations on interpretation, but in a more direct way. The task thus provides data which can be quickly and easily summarised. |
Example Experiment
The Wug Test was an experiment carried out by Jean Berko in 1958 which aimed to investigate the acquisition of the plural and other inflectional morphemes in English-speaking children. Berko presented the child with a made-up creature which she named a 鈥榃ug鈥. When presenting another she asked them 鈥楴ow there are two鈥?鈥 on this cue, most of the children tested replied 鈥榃ugs鈥, applying the correct plural suffix 鈥榮鈥 with the allomorph /z/ to the noun despite having never hearing it before. Berko鈥檚 test showed that children have productive rules: they don鈥檛 learn by hearing every possible form (they hadn鈥檛 heard 鈥榃ugs鈥 due to it being an invented word) but by applying linguistic rules that they subconsciously know. Although the derivation of these rules is not addressed by Berko, 鈥楾he Wug Test鈥 could provide possible evidence for Chomsky鈥檚 theory of 鈥楿niversal Grammar鈥
鈥n correspondence to this replicated study, Berko found that children over the age of 4 were able to produce the correct plural suffix, suggesting that at this age, children have internalised a working system of the plural allomorphs in English, and is able to generalise such rules to new cases.[2]
References
[1] Lust, B. (2006). 鈥楬ow Can We Tell What Children Know? Methods for The Study of Language Acquisition鈥 In: Lust, B. Child Language. Cambridge: Cambridge University Press. pp. 123-124.
[2] Berko, J. (1958). 鈥楾he Child鈥檚 Learning of English Morphology鈥 In: Natalicio, D.S. and Natalicio L.F.S., (2006). Learning: A Journal of Research in Language Studies.
What Next?
For more information on any of the terminology listed about studying language acquisition, visit the glossary page here to brush up on your knowledge