Grammatical representations versus productive patterns in change theories

Abstract In this paper, I discuss differences between representational change (i. e. in formal features and structures involved in grammatical competence) and change in quantitative patterns (i. e. in the quantitative properties of the language system in use), as relevant to my approach to incrementation. My approach differs from the standard variationist sociolinguistic approach because I argue that representational Language processing differences between children and adults could also contribute, but I set these aside here. Note that Biberaurer (this volume) also considers these relevant factors to the role of children in change. input-divergence Input-divergence (Cournane 2017) is used very broadly, as a way to capture any child language properties that deviate from the input model the child learns from. This includes what we standardly call child “errors”, without using that term, which assumes that there is a fixed target when learning a language and interim analyses are wrong. Rather “errors” are only such in comparison to the input/intake grammars, so I opt to call these “input-divergent” properties. along the child learning path contributes to quantitative differences between children and older speakers, most importantly the input speakers. In this way, the Inverted U Model (IUM) for incrementation offers an initial sketch of a linking theory between (a) child developmental findings for competence-related changes over acquisitional time in the individual, and (b) the change-in-progress phenomenon of incrementation which describes how usage rates for innovative variants advance relative to conservative variants in speakers in the community over generational time. Maximize Minimal Means (MMM), this volume similarly attributes a principled, creative role in change to the child-learner, offering a linking theory between (a), and (c), discrete changes in representations between grammars in historical time, grounded in Minimalism. I’ll also respond to Westergaard’s (this volume) argument that the IUM’s reliance on child overgeneralization conflicts with a set of linguistic phenomena for which directional, child-driven changes have been proposed, namely syntactic changes characterized by economy or simplification. In syntax, relative to common language change pathways (e. g. biclausal>monoclausal reanalyses), children typically acquire the (potentially) innovative grammatical structure earlier than the conservative one as they develop complexity (e. g. they develop from monoclausal>biclausal). It is indeed not clear how these child interim syntactic structures relate to overgeneralization, if at all. Rather, syntactic innovations are typically attributed to economy principles, and syntactic learning is sometimes characterized as conservative, also not obviously related to overgeneralization. I’ll show that neither economy in change nor child conservativity in syntactic development directly undermine the proposed model, as both are concerned with representational changes in grammars, not differences in quantitative patterns and changes-in-progress (the purview of incrementation and the IUM). Finally I will say a few words on the case study on Norwegian gender-system changes laid-out in Westergaard (this volume). These elicited production data are a valuable contribution to the roles of children in changes-in-progress, and while the data patterns conflict with some aspects of the IUM as proposed, the overall approach of Rodina and Westergaard is in line with a child-learning-centered contribution to the directionality and shape of changes-in-progress.