A count of the relative incidence of letters in 431 pleasant (P) words and 702 unpleasant (U) words revealed that some letters tend to occur more frequently in the initial position of P words and other letters more frequently in the initial position of U words. A task requiring Ss to guess for each letter whether it occurred in a P word or in a U word showed that people are able to approximate these objective probabilities of initial-letter occurrence. These findings can explain how it is possible to identify the probable affective meaning of a word seen in a tachistoscope prior to its complete recognition. Faster recognition of P words in tachistoscopic experiments was accounted for in terms of response probability. {\textcopyright} 1969 Academic Press Inc. All rights reserved.