SEP 25, 2018 | 6:30 PM TO 8:00 PM



The Graduate Center
365 Fifth Avenue




September 25, 2018: 6:30 PM-8:00 PM




Amy Ma presents on a topic in First Language Acquisition:

Pronoun Case Errors in Children

Pronoun case errors have long been treated as a common and prominent developmental error in child language acquisition (Huxley, 1970; Bloom et al, 1975; Budwig, 1989; Radford, 1990). The most common errors are misuse of accusative case for subject position (me* want cookies) and nominative case for object position (She hit *I). Generative accounts posited that case features are innate and always available to children and errors are the result of incomplete syntactic knowledge (Sch├╝tze & Wexler, 1996). Other input-based accounts proposed that case features are acquired through input and statistical learning (Pine et al, 2005; Rispoli, 2005). Previous literature doesn’t agree with each other since each study usually focused on one perspective of pronoun case errors and confined in limited size of data, thus lack the comprehensive analysis of pronoun use and case errors in children’ utterances.

In this paper, we tackled the lack of quantitative data on pronoun case errors by analyzing 110763 pronoun utterances of 389 children from CHILDES (MacWhinney, 2000) who are in between 20-month-age to 48-month-age. Our findings are as follow:

  1. Pronoun case errors are relatively rare.
    Only 525 errors were found in 110763 pronouns used by children, resulting a total error rate of 0.07%. Among all 363 children with single samples, pronoun errors were only found in 42 children; and 40 of them made less than 5 errors. The error rate is slightly higher among the children who made pronoun errors. The children who made pronoun errors produced 100797 pronouns, resulting an error rate of 0.08%.
  2. Pronoun case errors correlate with children’s speech and parent’s input.
    There is a strong negative correlation between children’s pronoun case error rate and total pronoun use and total word use. Similarly, a mild negative correlation was found between parent’s speech and children’s pronoun error rate. However, for each pronoun, the total use and error rate are not correlated. Also, there is no correlation between pronoun case errors and children’s age and MLU.
  3. No pronoun is protected from case errors; however, errors are distributed unevenly on each pronoun.
    The most frequent errors were made on me. There are total 235 me errors were found, making up 44% of all 525 errors. The second vulnerable pronoun is her (136, 26%) and him (63, 12%). Also, 82.7% of the errors were accusative missed as a nominative.
  4. The same children who make accusative errors also make nominative errors; who make errors on first person pronouns also on third person pronouns.
    Although accusative case errors and first person pronouns are more common, 19 children with longitudinal data showed that they made errors on both cases and persons.
  5. Pronoun case errors purported a U-shape development pattern.
    Although pronoun errors were never predominate, the longitudinal data of 25 children confirmed a U-shape development pattern of pronoun case errors that the peak of error rate usually clusters around 30 month of age.

All are welcome!