LETTER SEQUENCES VARYING IN ORDER OF APPROXIMATION TO JAPANESE
Kenichi Hirata · The Japanese journal of psychology · 1974
Letter sequences of zero- through second-order approximation to Japanese were computer-generated using tables of single-letter and digram frequencies. The single-letter and digram structures of the sequences were tested by calculating correlation coefficients between observed frequencies in each sequence of 10, 000 letters and those of the original tables (with zero-order material, by the goodness-of-fit test to uniformity hypothesis). The results confirmed that the sequences were properly generated. Two sets of tables are presented. The first set, Tables 1-3, gives 8 examples of 25-letter sequences of each of zero to second-order material. The other, Table 4, gives 30 10-letter sequences of each type, selected with the restriction that no letter should appear more than once in the sequence.