THE AUTHOR POINTS OUT PROBLEMS IN TEST RESEARCH AND INTERPRETATION, SOME OF WHICH ARE DUE TO CONFLICTS BETWEEN THE FINDINGS OF THE DATA ANALYST WHO IS RESTRICTED TO BASING HIS DECISIONS ON SELECTED DATA ONLY, AND THE TEST INTERPRETER WHO IS AWARE OF VARIABLE VALIDITIES OF SUCH UNTESTED FACTORS AS SITUATIONAL ANXIETY, PERSONALITY, MOTHER-TONGUE INFLUENCES, CULTURAL CLASH, AND SENSE OF COMMUNICATION. HOWEVER, THE AUTHOR FEELS IN SPITE OF THESE AND OTHER SHORTCOMINGS, THERE ARE A NUMBER OF REASONS FOR CONTINUING TO DO FACTOR ANALYSIS OF TEST RESULTS. ONE FACTOR, "FEEDFORWARD," BASED ON THE PSYCHOLOGY OF PERCEPTUAL EXPECTANCE, DEALS WITH SETS OF THE CATEGORIES THAT INDIVIDUALS HAVE AVAILABLE FOR THE PROCESSING OF ANY INTERNAL AND EXTERNAL PERCEPTIONS INCLUDING THOSE FOR LANGUAGE RECEPTION AND PRODUCTION. A VALID TEST OF LANGUAGE PROFICIENCY WOULD BE A TEST OF THE CATEGORIES THAT THE SUBJECT BRINGS TO ANY PROCESSING OF THE LANGUAGE. THE AUTHOR REVIEWS RECENT AND CURRENT RESEARCH WHICH IS CONCERNED WITH THE FACTOR OF "REDUNDANCY UTILIZATION", THE ABILITY OF THE NATIVE SPEAKER TO PREDICT SEQUENTIAL LANGUAGE SIGNALS AS CONTRASTED WITH THE NON-NATIVE SPEAKER'S DEPENDENCY ON INTERPRETING EACH WORD ON THE BASIS OF THE SIGNAL ITSELF. THIS WORKPAPER WAS PRESENTED AT THE ATESL SEMINAR IN AUSTIN, TEXAS, APRIL 27, 1967. (AM)