Frank R
Department of Cognitive Science, John Hopkins University, Baltimore, MD 21218, USA.
Cognition. 1998 Jun;66(3):249-301. doi: 10.1016/s0010-0277(98)00024-9.
One traditional view of the time course of language acquisition holds that a child's difficulties in learning her language are due to general processing, memory or conceptual limitations. As the child's cognitive capacities expand, so do her abilities to use her already-acquired grammar or to recognize additional properties of her soon-to-be native language. Recent work in the study of child language, however, has discovered the existence of a number of characteristic stages, transitions between which are best described in the terms of linguistic theory proper. These stages are surprising under a view of language acquisition according to which developmental delay derives from general cognitive limitations, which cannot characterize difficulties explicable only in language-specific terms. At the same time, current linguistic theories so severely restrict the variation among possible human grammars that there remains little reason why there should be any learning problem at all or characteristic developmental stages. In this paper, I propose that these two views can be reconciled. I show that children's difficulties with a wide range of syntactic constructions, which are indeed best defined in linguistic terms, should nonetheless be derived from limitations on the child's ability to deal with processing load and formal representational complexity. I suggest however that this can be done only in the context of a particular view of syntactic representation, one which is articulated in the terms of the formal system of tree adjoining grammar (TAG). I demonstrate how precisely those difficulties that children experience in the acquisition of relative clauses, adjectival modification, control constructions, raising, wh-questions and the obligatoriness of finite inflection can be traced to the complexity associated with one of the TAG combinatory operations, adjoining. This proposal relates this apparently disparate set of constructions in a novel way and provides us with a new type of explanation for the time course of syntactic development in terms of the complexity of formal grammatical devices.
关于语言习得时间进程的一种传统观点认为，儿童在学习语言时遇到的困难是由于一般的处理、记忆或概念限制。随着儿童认知能力的发展，她运用已习得语法或识别即将成为其母语的其他属性的能力也会相应提高。然而，儿童语言研究的最新成果发现了一系列特征阶段的存在，这些阶段之间的过渡最好用适当的语言学理论来描述。根据语言习得的一种观点，发育迟缓源于一般认知限制，而这些阶段在这种观点下是令人惊讶的，因为一般认知限制无法解释仅能用特定语言术语说明的困难。同时，当前的语言学理论对可能的人类语法之间的差异限制非常严格，以至于几乎没有理由认为会存在任何学习问题或特征性的发展阶段。在本文中，我提出这两种观点可以调和。我表明，儿童在各种句法结构上遇到的困难，尽管确实最好用语言学术语来定义，但应该源于儿童处理处理负荷和形式表征复杂性能力的限制。然而，我认为这只能在句法表征的特定观点的背景下完成，这种观点是根据树邻接语法（TAG）的形式系统来阐述的。我展示了儿童在习得关系从句、形容词修饰、控制结构、提升、wh-疑问句和限定性屈折的强制性方面遇到的那些困难，如何能够精确地追溯到与TAG组合操作之一即邻接相关的复杂性。这一提议以一种新颖的方式将这组明显不同的结构联系起来，并就形式语法手段的复杂性为句法发展的时间进程提供了一种新型的解释。
