Hypothesis: In this study, Berko aimed to see whether children possess morphological rules that can be applied to newly acquired words, and specifically nonsense words that carried no meaning but did share phonetic similarities with actual words. If a child knew the correct allophone used for the plural form of a specific word, would the child be able to use this existing knowledge and apply it to the nonsense word that used the same allophone? This is what Berko set out to find out, and his research wasn’t confined to looking at plural forms; he also included: the two possessive forms of the noun, the third person singular, the progressive tense, the past tense, and the comparative and superlative forms of adjectives.
Subjects: A total of 56 children aged from 4 to 7 were chosen from two schools. There were an equal number of boys and girls chosen, to see if sex had any effect on the test results. 12 adults, all college graduates, were also tested. All the subjects were native English speakers.
Test: To test the child’s use of morphological rules, a number of nonsense words were created. The nonsense word was then accompanied by a picture of what the word was supposed to represent, and a text with a blank space for the subject to verbally fill-in. The experimenter would read the test and ask the subject to supply the missing word. For example;
“This is a wug /wΛg/. Now there is another one. There are two of them. There are two ______.”
The Experimenter would note down the subjects answer phonemically and this would be compared to the phonological rules. In the event of ambiguity (such as the plural of *heaf) the answer was compared to those given by the adults. To insure that the subjects were familiar with the phonological rules, words that convey meaning, such as glasses and melted, were included in the test. This procedure was repeated with the aforementioned language features.
Results: For the plural inflexion of *wug, 97% of the first graders answered correctly. For *nizz, 33% answered correctly. These results are typical for the rest of the language features.
Conclusion: The results showed that children of this age range do possess phonological rules. However, these rules are delimited. The results also show that the older children’s answers were slightly better than the younger ones. When comparing the girls’ results with the boys’, there was no significant difference, leading to the conclusion that at this age there is no difference between the sexes when it comes to language acquisition.
Flaws: There was sampling bias in the choosing of the subjects as they were from two schools and this is not representative of the population. There was also a potential for experimenter bias because he was the only one noting the results. To improve this experiment, a true representation of the native English speaking population in the required age range could be used. Furthermore, an increase in experimenters recording the results would reduce the level of experimenter bias as the experimenters could compare their findings.