<p>Based on the interlanguage hypothesis and with reference to skill learning and central concepts of connectionist language-acquisition theory, this study develops an explanatory model, with the help of which untutored acquisition of grammatical gender in German is shown to be a sequence of meta-individual developmental phases. The empirical evidence consists of linguistic data compiled from interviews with Italian, Portuguese, and Spanish foreign workers carried out within the framework of the ZISA- Cross-Sectional Study. </p><p>The acquisition process is promoted by a combination of two cognitive procedures, one analytic (restructuring) and the other reproductive (automation). Both contribute to the construction of a network of associative <i>form-function mappings</i> which, in the course of the process, assume the character of prefabricated linguistic elements or chunks, thus enabling automatic processing. Accordingly, the gender of a noun is not acquired separately as such, but rather as an integral component of an automatic form-function mapping. </p><p><i>Formula</i> and <i>routine</i> are central concepts. They denote the norm-language (formula) and interlanguage (routine) chunks used by the learner and thereby the subsumed forms of the respective determiners, that is, gender markers. Their creation and substitution, as well as the sequence of their acquisition, are described and elucidated by means of cognitive mechanisms and psycholinguistic principles. </p><p>For the interlanguage routine in particular, but to some extent for the norm-language formula as well, it was possible to confirm the fundamental concept of the interlanguage hypotheses, according to which the learner sets up provisional hypotheses about the perceived elements and gradually approaches the form of the target language, albeit with the assistance of a concept of language acquisition that does not primarily center on any rule-defined morphology of the target language but rather on the phonetic surface-level form of concrete linguistic communication. </p><p>Against this background some key concepts of previous language-acquisition research, such as <i>explicit/implicit, rules, chunks, simplification, omission, </i>and<i> over-generalization</i>, are discussed and partially reassessed.</p>

Based on the interlanguage hypothesis and with reference to skill learning and central concepts of connectionist language-acquisition theory, this study develops an explanatory model, with the help of which untutored acquisition of grammatical gender in German is shown to be a sequence of meta-individual developmental phases. The empirical evidence consists of linguistic data compiled from interviews with Italian, Portuguese, and Spanish foreign workers carried out within the framework of the ZISA- Cross-Sectional Study. The acquisition process is promoted by a combination of two cognitive procedures, one analytic (restructuring) and the other reproductive (automation). Both contribute to the construction of a network of associative form-function mappings which, in the course of the process, assume the character of prefabricated linguistic elements or chunks, thus enabling automatic processing. Accordingly, the gender of a noun is not acquired separately as such, but rather as an integral component of an automatic form-function mapping. Formula and routine are central concepts. They denote the norm-language (formula) and interlanguage (routine) chunks used by the learner and thereby the subsumed forms of the respective determiners, that is, gender markers. Their creation and substitution, as well as the sequence of their acquisition, are described and elucidated by means of cognitive mechanisms and psycholinguistic principles. For the interlanguage routine in particular, but to some extent for the norm-language formula as well, it was possible to confirm the fundamental concept of the interlanguage hypotheses, according to which the learner sets up provisional hypotheses about the perceived elements and gradually approaches the form of the target language, albeit with the assistance of a concept of language acquisition that does not primarily center on any rule-defined morphology of the target language but rather on the phonetic surface-level form of concrete linguistic communication. Against this background some key concepts of previous language-acquisition research, such as explicit/implicit, rules, chunks, simplification, omission, and over-generalization, are discussed and partially reassessed.