Summary: In this chapter we present a network-theoretic approach to linguistics. In particular, we introduce a network model of derivational morphology in languages. We focus on suffixation as a mechanism to derive new words from existing ones. We induce networks of natural language data consisting of words, derivation suffixes and parts of speech (PoS) as well as the relations between them. Measuring the entropy of these networks by means of so-called information functionals we aim at capturing the variation between typologically different languages. In this way, we rely on the work of M. Dehmer [Appl. Math. Comput. 201, No. 1–2, 82–94 (2008; Zbl 1152.05361)], who has introduced a framework for measuring the entropy of graphs. In addition, we compare several entropy measures recently presented for graphs. We check whether these measures allow us to distinguish between language networks on the one hand, and random networks on the other.
We found out, that linguistic variation among languages can be captured by investigating the topology of the underlying networks. Further, information functionals based on distributions of topological properties turned out to be better discriminators than those that are based on properties of single vertices.
For the entire collection see [Zbl 1227.94001].