Universals—everybody wants them! After all, explanation is superior to description in science, or so they say, and something called universals stands a good chance of providing a principled explanation for observable linguistic phenomena, whether cross-linguistically valid generalizations about properties of language, about linguistic performance, or about first or second language acquisition. For some time, generative theoreticians have had a de facto patent on universals as they apply to language and its use and acquisition, with Greenbergian linguistic typologists as well as functionalists and psycholinguists of various persuasions in close competition. A special issue on universals and second language acquisition must pay serious attention to the case for universals as they might apply to second language acquisition (SLA) within each of these traditions. It is for this reason that this issue includes as many different representative perspectives as is possible, given the limitations of space.

In this article, the motivation for Universal Grammar (UG), as assumed in the principles and parameters framework of generative grammar (Chomsky, 1981a, 1981b), is discussed, particular attention being paid to the logical problem of first language acquisition. The potential role of UG in second language (L2) acquisition is then considered. Three different positions are reviewed: (a) the claim that UG is not available to L2 learners; (b) the claim that UG is fully available; and (c) the claim that the L2 learner's access to UG is mediated by the mother tongue. This raises the issue of what kind of evidence can be used to decide between these three positions. Recent experimental research which argues for one or another of these positions by investigating the L2 status of the Subjacency Principle is reviewed, and the implications of this research are discussed.

This article discusses the relationship of first language (L1) and second language (L2) acquisition. First, different approaches to comparing L1 and L2 development are summarized. Then, I argue for a particular version of the fundamental difference hypothesis. Finally, I illustrate the hypothesis with some evidence from the acquisition of German syntax. It is claimed that the Universal Grammar (UG) approach provides a theoretical framework to explain differences between L1 and (adult) L2 development. In particular I argue that the observed L1/L2 differences can be accounted for by assuming that adult L2 learners cannot use principles of UG as a learning device in the same way as L1 learners use them.

This article examines the role played by functional approaches to linguistics in understanding second language acquisition (SLA). Central premises and tenets of functional approaches are described, and several key theoretical problems with functional efforts are detailed. The problem of referential management (the selection of nominal vs. pronominal NPs) in second language discourse production is examined. The general conclusions are drawn that (a) functional approaches to linguistics have a significant role to play in SLA studies, but (b) functional universals are insufficiently grounded theoretically and empirically at this point to contribute more than heuristic guidance to SLA theory.

This article examines the implications of connectionist models of cognition for second language theory. Connectionism offers a challenge to the symbolic models which dominate cognitive science. In connectionist models all knowledge is embodied in a network of simple processing units joined by connections which are strengthened or weakened in response to regularities in input patterns. These models avoid the brittleness of symbolic approaches, and they exhibit rule-like behavior without explicit rules. A connectionist framework is proposed within which hypotheses about second language acquisition can be tested. Inputs and outputs are patterns of activation on units representing both form and meaning. Learning consists of the unsupervised association of pattern elements with one another. A network is first trained on a set of first language patterns and then exposed to a set of second language patterns with the same meanings. Several simulations of constituent-order transfer within this framework are discussed.

On the basis of a critical analysis of the articles in this issue, we argue for a shift in emphasis in the investigation of universals in second language acquisition (SLA). To construct a psycholinguistically plausible theory of acquisition, research must proceed simultaneously on a number of different fronts: the elaboration of (a) a theory of computation consistent with human biology, (b) a theory of structural universals and variation, (c) a theory of meaning, and (d) a theory of pragmatics which must tie in with both a theory of cognitive development and a theory of sociocultural knowledge. None of the theories popular at present address all of these issues. We suggest that there is much room for consensus, but that achieving it will require keeping an eye firmly focused on the long-term objectives.

To the extent that language universals represent internal properties of human beings, one would expect them to manifest themselves in first language (L1) acquisition. Whether they should also manifest themselves in second language (L2) acquisition depends on whether or not language universals remain accessible to adults. The relation between L2 acquisition and language universals research is examined with respect to three phenomena: extraction, where the subset principle makes interesting predictions, although the second language data are far from clear; structure dependence, where there is evidence for the continuation beyond puberty of a general formal property of language; and the distribution of overt reflexives, which has a functional basis and therefore poses the interesting question whether or not this functional basis is accessible to language learners.

The first part of this commentary discusses the minimal requirements that any serious theory of language acquisition must meet. It must take into account the particular properties of the human language processor and the (linguistic and nonlinguistic) input, as well as the specific motivation which causes the learner to apply the former to the latter. Neglecting, or even not keeping constant, some of these factors leads to a very distorted picture of the nature of language acquisition. In particular, claims about the difference between first (L1) and second language (L2) acquisition that ignore major variation in one of these components go astray.