Analysing Afrikaans lexical blends using Levenshtein distances

The utility of language is not limited to its communicative function as can be illustrated by constructions like hangry: Two words (hungry and angry) are combined to generate a new construction that describes a state of being angry due to being hungry. These constructions are known as lexical blends. Language users can create blends for purposes ranging from literary effect to displaying linguistic creativity.In this paper Afrikaans blends (e.g., kapoen as a blend of kak 'shit’ and pampoen ’pumpkin’) are investigated. Context is given with reference to available studies before the analysis of a dataset of Afrikaans blends is undertaken. The collected data is analysed using the Levenshtein distance metric, a type of edit distance that measures the similarity between two strings in terms of the number of single-character edits to illustrate similarity between source words and blends. The following hypothesis is investigated: Whether the shorter source word in a blend contributes more to the blend. From the available data we cannot confirm a positive tendency toward this hypothesis and argue that we require more data before any kind of conclusion can be drawn. Still, this study shows to what degree edit distance measuring can be employed to lay the foundation for the description of Afrikaans blends.