time of reaction to a relevant word was that of association with a word referring to the object seen. The irrelevant stimulus words that gave abnormally long reactions were as follows: (1) 'Position—too indefinite;' here the abstract character of the stimulus word seems to have caused the delay. (2) 'Clue—detective;' the next longest r. t. was that for 'bite—scratch.' These stimulus words probably had some affective accompaniment that caused the delay. (3) 'Remark—word;' here there is no discoverable reason for the long r. t., and the same is true of (4), 'house—brown,' and (5), 'walk—parachute.' If 'walk' had been a relevant word, the far-fetched character of this association together with the fact that its time was so long, would have led to the positive conviction that the object connected with 'walk' had been seen. Unfortunately the observer's introspection on this association was not recorded. In the case of (6), 'telepathy—man,' the unusual character of the stimulus word may again have been, at least in part, responsible for the delay; with (7), 'disgrace—prison,' affective influences undoubtedly played a part. In case (8), the longest three reaction times were to irrelevant words: 'comfort—vague idea;' 'servant—play;' 'disgrace—sin.' The last two involved affective elements (the observer had recently seen "The Servant in the House"); the first gave a slow reaction probably because of its abstract character. In general it may be said that when a word referring to an object not seen gives an abnormally long reaction time this is because it (a) is of a disturbing affective character in general, like 'disgrace;' (b) refers to some emotional complex, peculiar to the observer, or (c) is abstract or unusual in character. It would seem that the rule that the longest single reaction time to a relevant word is to a word associated with the object seen may be safely followed except where some word referring to the other object comes under one of the three classes just mentioned. One could never be sure of selecting all the relevant words in a series so that none should be connected with an emotional complex peculiar to the observer, but it would be possible to avoid using, as relevant words, those of general emotional import, or those which were unusual or abstract.
