Authors´ Reply2

The authors would like to thank B. R. Gaines for his careful reading of both the paper¹and the thesis [1]. Dr. Gaines is correct in pointing out that for application to the grammatical inference problem We have found, as Gaines has stated, that making this correction to the program does in fact yield more accurate models. We also concur with Dr. Gaines as to the two typographical errors pointed out in test grammar 1.