A variety of studies have reported that learners of English as a second language find subject relative clauses easier to produce and comprehend than direct object relatives, but it is unclear whether this preference should be attributed to structural factors or to a linear distance effect. This paper seeks to resolve this issue and to extend our understanding of SLA in general by investigating the interpretation of subject and direct object relative clauses by English-speaking learners of Korean, a left-branching language in which subject gaps in relative clauses are more distant from the head than are object gaps. The results of a comprehension task conducted with 53 beginning and intermediate learners point toward a strong preference for subject relative clauses, favoring the structural account.

It is widely recognized that the processor has a key role to play in creating and strengthening the mapping between form and meaning that is integral to language use. Adopting an emergentist approach to heritage language acquisition, the current study considers the extent to which the operation of the processor can contribute to an account of what is acquired, what is subsequently retained or lost, and what is never acquired in the first place. These questions are explored from two perspectives. First, morphosyntactic phenomena for which there is apparently substantial input are considered, with a focus on the relevance of salience, frequency, and transparency to the establishment of form-meaning mappings. Second, a phenomenon for which there appears to be relatively little input (i.e., scope) is examined with a view to understanding its fate in heritage language acquisition. In both cases, the emergentist perspective appears to offer promising insights into why heritage language learners succeed—and fail—in the way that they do.

I propose that the course of development in first and second language acquisition is shaped by two types of processing pressures—internal efficiency‐related factors relevant to easing the burden on working memory and external input‐related factors such as frequency of occurrence. In an attempt to document the role of internal factors, I consider a series of case studies involving contrasts that are rarely instantiated in the input, yet show early mastery. I conclude with some general remarks about the nature of development and the possibility that it unfolds more uniformly than often suggested.

an anonymous referee for this journal, and various attendees at talks that I have given on the ideas summarized in this paper. I also thank Marie O'Grady for her help in conducting one of the experiments reported here. 2 Chan et al.'s study involved a total of 200 monolingual learners (aged 2;6-4;6) of English, German and Cantonese. SVO is the dominant order is all three languages. However, whereas English employs this order almost exclusively, German employs SOV order in embedded clauses and exhibits considerable word order variation related to focus and topicality. Cantonese also permits focus-and topic-related variation in word order, as well as frequent argument drop, both of which reduce the number of SVO sentences.

Syntax constitutes a challenging area for emergentist research, since traditional grammar-based frameworks have reported significant success in their analysis of many important phenomena. This chapter considers a number of those phenomena from an emergentist perspective in order to show how they can be understood in terms of the interaction of lexical properties with a simple efficiency-driven processor, without reference to grammatical principles. It proposes ideas that rest on two key claims: (i) syntactic theory can and should be unified with the theory of sentence processing; and (ii) the mechanisms required to account for the traditional concerns of syntactic theory (e.g., the design of phrase structure, pronoun interpretation, control, agreement, contraction, scope, island constraints, and the like) are identical to the mechanisms which are independently required to account for how sentences are processed from ‘left to right’ in real time.