Comparing Different Models of the Development of Verb Inflection in Early Child Spanish

Abstract

How children acquire knowledge of verb inflection is a long-standing question in language acquisition research. In the present study, we test the predictions of some current constructivist and generativist accounts of the development of verb inflection by focusing on data from two Spanish-speaking children between the ages of 2;0 and 2;6. The constructivist claim that children’s early knowledge of verb inflection is only partially productive is tested by comparing the average number of different inflections per verb in matched samples of child and adult speech. The generativist claim that children’s early use of verb inflection is essentially error-free is tested by investigating the rate at which the children made subject-verb agreement errors in different parts of the present tense paradigm. Our results show: 1) that, although even adults’ use of verb inflection in Spanish tends to look somewhat lexically restricted, both children’s use of verb inflection was significantly less flexible than that of their caregivers, and 2) that, although the rate at which the two children produced subject-verb agreement errors in their speech was very low, this overall error rate hid a consistent pattern of error in which error rates were substantially higher in low frequency than in high frequency contexts, and substantially higher for low frequency than for high frequency verbs. These results undermine the claim that children’s use of verb inflection is fully productive from the earliest observable stages, and are consistent with the constructivist claim that knowledge of verb inflection develops only gradually.

An Elicited‐Production Study of Inflectional Verb Morphology in Child Finnish

“…First, overall error rates are misleading because they collapse across data from both high-and low-frequency person/number contexts (or, from a constructivist viewpoint, morphological schemas). Similar findings were reported by Aguado-Orea and Pine (2015) in a naturalistic corpus study of two Spanish-speaking children. However, a closer look at the data revealed that this low error rate was composed of an error rate of 0.3% in high-frequency 3sg contexts and of 43.5% error rate in lowfrequency 3pl contexts.…”

“…Furthermore, the pattern of results is very similar to that observed in naturalistic studies of Spanish (Aguado-Orea & Pine, 2015) and Brazilian Portuguese (Rubino & Pine, 1998), suggesting that any occasional misclassifications of errors as well-formed attempts at nontarget person + number forms did not substantially affect the overall pattern observed. Note, however, that because we excluded children who did not produce at least one correct instance of the relevant person/number morpheme, the high error rates observed for certain inflectional contexts cannot solely be a reflection of a tendency to avoid these contexts for pragmatic reasons (e.g., using a 2sg form rather than a 2pl form because the child prefers to describe the actions of the dog alone, rather than the dog and the experimenter).…”

“…While this error rate is already somewhat higher than rates typically taken as evidence for virtually error-free performance (around 5%; Hoekstra & Hyams, 1998;Wexler, 1998), it hides considerably higher rates in certain parts of the system. The pattern is very similar to that observed by Aguado-Orea and Pine (2015) and Rubino and Pine (1998), with a very low error rate for 3sg forms (<1%) hiding rates as high as 32% in other contexts. The pattern is very similar to that observed by Aguado-Orea and Pine (2015) and Rubino and Pine (1998), with a very low error rate for 3sg forms (<1%) hiding rates as high as 32% in other contexts.…”

“…This is simply because, in spontaneous speech, children (and, indeed, adults) tend to use only a small number of verbs, and-in most cases-only one or two inflectional forms of each (Aguado-Orea & Pine, 2015). This is simply because, in spontaneous speech, children (and, indeed, adults) tend to use only a small number of verbs, and-in most cases-only one or two inflectional forms of each (Aguado-Orea & Pine, 2015).…”

“…However, studies of Spanish (Aguado-Orea & Pine, 2015) and Brazilian Portuguese (Rubino & Pine, 1998) have revealed that these low overall error rates actually hide important differences across the verb paradigm. However, studies of Spanish (Aguado-Orea & Pine, 2015) and Brazilian Portuguese (Rubino & Pine, 1998) have revealed that these low overall error rates actually hide important differences across the verb paradigm.…”