A much debated issue in linguistics involves the extent to which the forms of languages(phonemes,morphemes,words,phrases and syntactic structures)have been shaped by the functions that languages perform.This paper beg...A much debated issue in linguistics involves the extent to which the forms of languages(phonemes,morphemes,words,phrases and syntactic structures)have been shaped by the functions that languages perform.This paper begins by defining what these"functions"are:the forms of language must express(i.e.be capable of being paired with)an infinite set of meanings in a consistent way;these form-meaning pairings must be readily processable by language producers and comprehenders in real time;and they must be readily learnable.This paper then focuses on the processability function,and enumerates some problems that complicate discussions of whether processing has actually played a role in shaping grammars,especially in morphology and syntax.Most of these problems,it is argued,are the result of unresolved issues in current theories regarding e.g.the precise relationship between production and comprehension,the measurement of working memory load,and the relationship between prediction and the integration of preceding and following items in on-line processing.I argue here that the linguistic question of explaining why the world’s grammars are the way they are,and understanding the role of processing in explaining grammatical forms,need not,and cannot,wait for these big general issues to be resolved in psycholinguistics.Instead,we can adopt a more empirical approach that compares patterns in grammatical forms and structures across languages with relevant empirical patterns in usage data within languages,from corpora and experiments.This more empirical approach allows us to test whether there is,or is not,a match between grammatical data on the one hand,and usage data on the other,that are the products of the processing mechanisms that psycholinguistic theories are trying to def ine.These patterns can be described in theory-general or-neutral ways and they can shed light on whether grammatical forms have been shaped by performance,without us having to f ill in all the details of the contributing processing theories.This paper il展开更多
文摘A much debated issue in linguistics involves the extent to which the forms of languages(phonemes,morphemes,words,phrases and syntactic structures)have been shaped by the functions that languages perform.This paper begins by defining what these"functions"are:the forms of language must express(i.e.be capable of being paired with)an infinite set of meanings in a consistent way;these form-meaning pairings must be readily processable by language producers and comprehenders in real time;and they must be readily learnable.This paper then focuses on the processability function,and enumerates some problems that complicate discussions of whether processing has actually played a role in shaping grammars,especially in morphology and syntax.Most of these problems,it is argued,are the result of unresolved issues in current theories regarding e.g.the precise relationship between production and comprehension,the measurement of working memory load,and the relationship between prediction and the integration of preceding and following items in on-line processing.I argue here that the linguistic question of explaining why the world’s grammars are the way they are,and understanding the role of processing in explaining grammatical forms,need not,and cannot,wait for these big general issues to be resolved in psycholinguistics.Instead,we can adopt a more empirical approach that compares patterns in grammatical forms and structures across languages with relevant empirical patterns in usage data within languages,from corpora and experiments.This more empirical approach allows us to test whether there is,or is not,a match between grammatical data on the one hand,and usage data on the other,that are the products of the processing mechanisms that psycholinguistic theories are trying to def ine.These patterns can be described in theory-general or-neutral ways and they can shed light on whether grammatical forms have been shaped by performance,without us having to f ill in all the details of the contributing processing theories.This paper il
