Most of the time, when we speak our native language we do not encounter difficulties in making the subject of our utterance agree with the verb. Native speakers of languages that mark gender are also very accurate in making determiners and adjectives agree with nouns. In other words, the production of number and gender agreement rarely goes wrong for speakers of languages that instantiate these types of agreement. On the other hand, agreement errors tend to be relatively frequent in the speech of second language (L2) learners, especially when the native language of the learner does not instantiate the particular type of agreement in question. In fact, some types of agreement seem to cause lasting difficulties for learners, such that even at high levels of proficiency and after many years of experience with the L2, their accuracy in agreement production is significantly lower than that of a native speaker. The overarching goal of this research is therefore to determine whether these differences between native speakers and L2 learners in agreement production accuracy stem from qualitative differences in how the psycholinguistic mechanisms of agreement function in the two groups, or whether they are simply due to quantitative differences in each group's susceptibility to processes that allow agreement errors to occur.
One factor that seems to play a role in how accurately native speakers produce number and gender agreement is the degree to which the morphology of a language clearly expresses number and gender features. Research with native speakers of various languages suggests that speakers of languages with relatively rich number and gender morphology tend to make fewer agreement errors than speakers of languages with relatively poor morphology. This also applies within a language: When a speaker produces an utterance that contains richer or more transparently-marked number or gender morphology, he or she is less likely to make an agreement error than when he or she produces an utterance that contains less rich or more opaquely-marked number or gender morphology. For example, when a native speaker of Spanish produces the utterance in (1), he or she is less likely to incorrectly mark the adjective escondida (hidden) as feminine in comparison to when he or she produces the utterance in (2). This is because in (1), the noun with which the adjective needs to agree is transparently-marked as masculine (-o), whereas in (2), it is not.
(1)*El libro(masc) en la iglesia esta escondida(fem)
'The book in the church is hidden.'
(2)*El tunel(masc) en la iglesia esta escondida(fem)
'The tunnel in the church is hidden.'
In this talk, Prof. Foote will present the results of two studies that examine the role of morphology in the production of agreement in L2 Spanish learners. She will show that while mechanisms of number agreement seem to function similarly in native speakers and higher-proficiency L2 learners, in that morphological richness leads to decreased susceptibility to number agreement errors in both groups, the pattern of gender agreement errors is more complex. It may be the case that learners pattern like native speakers only when the non-default gender is in play.