Nordlyd: Tromsø University Working Papers on Language & Linguistics (Apr 2015)

Introduction to double issue 41.1–2 on Features

  • Peter Svenonius,
  • Martin Krämer

DOI
https://doi.org/10.7557/12.3415
Journal volume & issue
Vol. 41, no. 1

Abstract

Read online

This special double issue (41.1 and 41.2) contains 11 articles on the formal properties of linguistic feature systems, all of which were presented at a conference in Tromsø in the fall of 2013. The issue was jointly edited by Martin Krämer, Sandra Ronai, and Peter Svenonius. A version of the original call for papers posted in 2013 follows. All formal models of linguistics assume sets of features in terms of which generalizations can be stated. But the nature of the features themselves is often not explicitly addressed. In this special double issue of Nordlyd we focus on the nature of features across phonology and syntax and related domains of linguistics. One group of questions concerns the ‘grounding’ of features in substance or content. For example, phonological features may be grounded in phonetics, and syntactic features may be grounded in semantics. Innatist traditions have sometimes posited innate universal inventories of grounded features. The ‘substance-free’ movement in phonology argues instead that the formal properties of features can and should be radically dissociated from their grounding in content. Sign language phonology would seem to support this position, as the featural system of sign language phonology operates with a completely different set of articulators from those used in spoken languages. Minimalist syntax also frequently promotes the dissociation of formal properties of features from their content (as in the proposal that tense is simply one of a variety of ways in which Infl may be ‘grounded,’ favored in Indo-European languages but with various other languages opting for other content for Infl). Such proposals raise many questions concerning how feature systems are constrained to be uniform across languages and to what extent they are free to vary. The radically opposing view in phonology denies the existence of categorical features altogether and attempts to model phonological patterns as statistical computation of phonetic data. The formal structure of features raises another set of questions. Complex patterns of feature locality gave rise to feature geometries in phonology, and these have been developed further to account for dependencies among features, not only in phonology but also in syntax. Cartographic work typically assumes linear hierarchies. To what extent are the various geometries and hierarchies motivated, and how might they be grounded in a broader explanatory theory? Interacting with these questions about the “geometric” relations among features is the algebraic structure of the features. For example, it is often assumed that privativity, in which opposition is marked by presence versus absence, is conceptually simplest and therefore the zero hypothesis. While in phonology the pendulum currently swings towards privativity, recently arguments have come from morphosyntax that features have binary values. While apparent ternary patterns in phonology have been taken as arguments in favor of binarity, such patterns have more recently been accounted for by reference to class nodes. Theories such as HPSG or Government Phonology assume much more complex relations among features (with HPSG even allowing feature-value matrices in which the values are feature-value matrices, extending to a kind of feature recursion, and GP positing government and licensing relations between features and positions). In this volume, a selection of researchers address these and other questions about the nature of features in linguistic theory.