Children often pick up a language and learn different elements of grammar in a predictable way. What does it look like in young speakers of Hindi?
By Heraa Hashmi
Course: Morphology and Syntax (Ling 4420)
Advisor: Jared Desjardins
LURA 2019听
Like learning math, from simple counting to solving complex equations, language learning happens in a similar stepwise fashion, with each new step building on previous knowledge. Brown (1973) outlines these steps in stages, illustrating what the acquisition of English morphology looks like in children learning English as their first language. A neurotypical child will usually first learn how to produce the present tense (-ing) form between 27-30 months old, then prepositions like in and out, then how to produce regular plural noun forms (-s), and so on. Stages are identified by Mean Length Utterances (MLUs): the average number of morphemes comprising sentences spoken by a child. Intuitively it is known that as children get older, the sentences they utter become longer and more complex.
Much of the research conducted and available in the field of language acquisition in the West tends to be Eurocentric; in particular, the development of theories is largely based on English language data and studies. I wanted to apply Brown鈥檚 (and other researchers in the field鈥檚) methodology involving stepwise acquisition and MLUs to young, native speakers of听Hindi, and evaluate the results in comparison to existing literature concerning native English speakers. To that end, I conducted a pilot analysis of morphological development in Hindi speakers, primarily looking at mean length utterances and morphological 鈥榚rrors鈥 (like incorrect verb tense/number agreement).
I contacted eight families in the state of Bihar, India, and obtained eight two-minute long recordings of family members talking to children ages 36 to 51 months old. I eliminated data from those who had parents who predominantly spoke a language other than Hindi at home (e.g. Tamil), because that introduced a variable my study would not be looking at (bilingualism), or who were too young and produced mostly phonological errors, as it was difficult to attribute errors (for example, in incorrect agreement) to a pronunciation issue or a morphological one. This left me with a sample size of four recordings, and I subsequently analyzed their speech by following MLU protocol, marking morphemes by means of forward slashes and any errors marked in red.
Example analysis:
The analysis found that the children in the study were either at average or above the MLU range for English speaking children, meaning their sentences were, on average, longer than their English-speaking counterparts in the same age range. I hypothesize that this might be because Hindi verbs carry multiple morphemes (inflectional affixes), whereas English has analytic tendencies making use of individual lexical units. In the sentences in the data, Hindi speaking children had to use more morphemes, especially with verbs, than would be required in English to express the same thing.
Regarding errors in speech, the overwhelming majority involved verbal morphology. When looking at verb forms, over half were produced correctly. Out of the 33% that were not, about 19% were due to incorrect agreement with tense, gender, or number, and 14% due to incorrect constructions with the auxiliary verb 鈥榯o be鈥. The sample size was not large enough to see if age played a role in the pattern of constructions; however, compared to English speaking children, the children in this study produced more errors in verbs, which are more morphologically complex than in English.
This pilot study implicated several different ideas, the most notable being that MLU may depend on the typology of language. However, there is not enough literature on utterance lengths in other typologically diverse languages or patterns of morpheme acquisition. While my study focused on errors compared to what a typical morphological development pattern looks like in English, it did reveal another interesting observation. Due to more complex verbal morphology in Hindi, even children around the age of 4 and 5 do not produce correct structures almost half of the time, whereas their English-speaking counterparts, by that age, have mastered auxiliaries (linking verbs, like 鈥榳as鈥 in 鈥渟he was running鈥) and copulas (words connecting a subject and its complement, like 鈥榠s鈥 in 鈥渉e is tall鈥).
The sample size was not large enough to make any conclusive statements, and there were several potential sources of error. For instance, I did not account for any degree of bilingualism in the participants, and because of the small nature of the study, errors could be attributed to specific language disorders and delays as outliers. A future study is this area should be carried out on a larger-scale, mapping out morphological development across infanthood, toddlerhood, and preschool years. A study of this scope could be helpful for both educators and speech-language pathologists. In addition, it could potentially reveal whether the typology of language, and particularly verbal constructions, significantly impacts language acquisition, and how. Developing a holistic set of guidelines as to what typical Hindi acquisition looks like and establishing a baseline for morphological development has the potential to contribute to other branches of linguistics, such as psycholinguists, sociolinguistics, and cognitive science.
Opening photo credit: