Type-token models: a comparative study

Abstract

The type (V) – token (N) relationship has been studied for almost a century. Although a number of models have been developed to examine this relationship, comparative studies have been rare. Thirty-six published and 14 new models were examined using 28 Latin texts. The results were similar to those found in other languages. The best fitting models were 1/V = a + b/N^c where a, b and c are constants. The ratio V/N was also found to be well fitted by the relation a + b log N.

Acknowledgements

I would like to thank an anonymous referee for his/her comments which have significantly improved this communication and for suggesting the new model.