The debate on whether creole acquisition proceeds via small steps, involving undergeneralization, or in large leaps, involving overgeneralization, is in this paper explored with particular reference to the treatment of the so-called "predicate marker" i in Tok Pisin by a group of first-generation creole speakers.