Allomorphy: Old Concept, Big Data, New Model

Abstract

Many key concepts that modern linguistics operates with were first introduced in the Structuralist Era, the time when linguists believed in clear-cut oppositions and privative features and did not have access to large electronic corpora. Today, when we apply these Structuralist concepts to linguistic data, which is gradient in nature, we encounter a gap between the data and our theory. I argue that it is now high time to update the old classical linguistic concepts with regard to 1) big authentic data and 2) new robust methodology of qualitative and quantitative analyses. I demonstrate this idea by looking at the concept of allomorphy, crucial for understanding the morpheme and the asymmetry between form and meaning. I will draw much attention to the role of semantics in the allomorphic relationship which has been largely underestimated.