WHEN ASKED TO INDICATE WHETHER EACH OF 200 ORALLY PRESENTED WORDS HAD APPEARED BEFORE (+) OR NOT (-), 28 STUDENTS GAVE MORE PLUSSES TO COMMON ASSOCIATES AND SYNONYMS OF PRECEDING WORDS THAN TO CONTROL WORDS. IN EXP. II, FALSE RECOGNITION ERRORS WERE OBTAINED WHEN THE PRECEDING WORDS ASSOCIATIVELY ELICITED THE TEST WORDS AND WHEN THE ASSOCIATIVE RELATION WAS BIDIRECTIONAL BUT NOT WHEN ONLY THE TEST WORDS ELICITED THE PRECEDING WORDS. RESULTS WERE TAKEN AS AN INDICATION THAT INITIAL CODING OF WORDS CONTRIBUTES TO FALSE RECOGNITION AND THAT THE PHENOMENON IS NOT MERELY AN ARTIFACT OF TESTING FOR IT. THE OCCURRENCE OF FALSE RECOGNITION ERRORS WAS TAKEN AS SUPPORT FOR A CHARACTERIZATION OF WORDS AS COMPLEXES OF ATTRIBUTES OR FEATURES. (17 REF.) (PsycINFO Database Record (c) 2006 APA, all rights reserved)