Proceedings

Gang effects: the perspective from variation

---

Abstract

This paper explores the implications of the Strict Domination Generalization (SDG), which derives from Labov(1969). Informally, it says that the phonological factors that condition the rate of application of variable phonological processes can always be ranked in a strict domination hierarchy such that higher ranked factors outweigh the effects of all lower ranked factors in determining rates of application.

We show that the SDG effectively makes a distinction between two kinds of gang effects among factors: it permits superset gang effects, in which addition of factors favoring application of a process increases its rate of application; but it bans disjoint gang effects, in which two individually weaker factors combine to yield a higher rate of application than a single stronger factor.

We then observe that SOT predicts minimal gang effects of either type while ME can be restricted to exclude disjoint gang effects while permitting a wide range of superset gang effects by restricting the possible weight vectors. However, we do not currently have a way to prune ME weights via a general condition but need to tailor pruning to the specific constraints of each case study. 

Keywords: token variation, strict domination, gang effects, stochastic Optimality Theory, Maximum Entropy grammars

How to Cite:

Flemming, E. & Magri, G., (2026) “Gang effects: the perspective from variation”, Proceedings of the Annual Meetings on Phonology 2(1). doi: https://doi.org/10.7275/amphonology.3717