This report contains two experimental studies. These studies were undertaken for the purpose of verifying the hypothesis that the recognition of a letter located in different contexts shows the different results in recognition.
In the first study, we used higher and lower contextually constrained sequences of letters as the stimuli, and in these stimuli the number of preceding letters were both four. In the second study, as the preceding letters we used 0, 2 and 4 letters respectively, and in all these experiments the response to the succeeding letter was examined.
In the first study, the stimulus materials of lower contextually constrained sequences were obtained by selecting letters from the seventy-one Hira-liana letters using a table of random number, and those of the higher contextually constrained sequences were obtained by selecting letters according to their relative frequencies of occurence in sequences of five letters in Japanese. In the second study, the stimulus materials were obtained by the similar method as the first study.
The findings resulting from these studies were as follows:
(1) The recognition of the letter is influenced by the context in which it is located. Namely, when the higher contextually constrained sequences are exposed, the letter is recognized easily, but when the contextual constraint is lower, it is difficult to recognize the letter.
(2) When the contextual constraint is higher, there are few errors on the recognition of letters, but when the contextual constraint is lower, there are many chances to respond erroneously.
(3) The recognition of a letter is more influenced by the context, and the structural characteristics of a letter itself may be regarded as a secondary factor on it.