Language is an extremely resilient human capacity that expresses itself fully even under adverse conditions such as limited cognitive capacities, non-native language models, or lack of access to auditory and/or visual modalities. Numerous studies have addressed the independence of the first language acquisition process from the tangible linguistic input to that process. However, there is a lower bound on what constitutes sufficient input to the first language acquisition process and there are humans with language-ready brains that nonetheless can fail to acquire language. This presentation focuses upon such individuals after they finally come in contact with language input. This paper assumes the nativist premise that all human children are born with language-ready brains capable of creating language and recognizing language-relevant evidence in the environment. In the absence of language-relevant evidence, however, the language-ready brain fails to engage in the first language acquisition process. Our research in Nicaragua shows that language-relevant evidence need not be language. Those findings will be reviewed here, but our focus will be on those cases where language as we know it has failed to emerge. Several essential components of the language acquisition process are proposed: sensitivity to prosody and sequencing that leads acquirers to attend to language-relevant input; 2. awareness of one's ability to copy certain language-relevant stimuli and a tendency to attempt to copy such stimuli; 3. an innate set of language expectations that drive, direct and supplement the first language acquisition process (Chomsky's I-language); and finally, 4. a drive to match the output of one's first language acquisition process to already existing target languages available in the environment (Chomsky's E-language). Evidence from the emergence and non-emergence of language in Nicaragua supports the hypothesis that only the innate language expectations in (3) are critical-period bound. The behavior of adult language isolates who subsequently come in contact with language reveals that components (1) and (2) remain in force, and, while the press to match the language around them central to component (4) remains, in the absence of (3), the capacity to do so is typically lost. The rare cases that succeed offer us yet another insight into the nature of language. Approximating a target language (E-language) after the innate capacity to do so has waned, requires an algorithmic-like, explicitly tutored, learning of the rules of grammar. First language acquisition within the critical periods is blind to general intelligence and cognitive prowess, and native adult language production is virtually effortless. In contrast, late-learning draws heavily upon general intellectual resources. As a result, the language production of late-learners is effortful and highly vulnerable to any increases in cognitive load. The infrastructure that supports late approximation to E-language is argued not to be the same language organ that creates I-language anew in the young child. Faculty

Host: Dr. Geraldine Legendre (legendre@cogsci.jhu.edu)