This book explores the goodness of fit, or lack thereof, between linguistic phenomena—more specifically in lexical semantics—and the mathematical and computational models that seek to account for those phenomena. In particular, the largely predominant vector-based paradigm and its construction based on massive scanning of text corpora are examined.Following the structuralists, this book defends a complementary approach grounded in the need to take into account the major paradigmatic and syntagmatic relations that structure language and its acquisition. To address this need, a formal framework is proposed that relies on the notions of paradigmatic and syntagmatic arborescences in order to account for context-based variations in meaning, semantic isotopies, and the effects of negation. This model is parsimonious in terms of both the size of the required corpora and that of the arborescences themselves.The target audience comprises researchers and experts in the field. This book is also beneficial for graduate and post-graduate students.
