Investigating English Affixes and their Productivity with Princeton WordNet

Such a rich language resource like Princeton WordNet, containing linguistic information of different types (semantic, lexical, syntactic, derivational, dialectal, etc.), is a thesaurus which is worth both being used in various language-enabled applications and being explored in order to study a language. In this paper we show how we used Princeton WordNet version 3.0 to study the English affixes. We extracted pairs of base-derived words and identified the affixes by means of which the derived words were created from their bases. We distinguished among four types of derivation depending on the type of overlapping between the senses of the base word and those of the derived word that are linked by derivational relations in Princeton WordNet. We studied the behaviour of affixes with respect to these derivation types. Drawing on these data, we inferred about their productivity.