WITH THE INTENTION OF APPLYING A MATHEMATICAL THEORY OF LEARNING TO SECOND LANGUAGE ACQUISITION, A TWO-PART EXPERIMENT USING TWO SETS OF 20 STUDENTS WAS CONDUCTED AT STANFORD UNIVERSITY. IN EXPERIMENT I THE SS (SUBJECTS) WERE ASKED TO DISTINGUISH DIFFERENCE OR SAMENESS IN PAIRS OF INITIAL RUSSIAN SYLLABLES. THESE 144 SYLLABLES WERE GROUPED INTO CONTRASTING PAIRS DIFFERING ONLY IN THE INITIAL CONSONANT PHONEME. THESE CONTRASTS WERE THEN CLASSIFIED ACCORDING TO EXPECTED DIFFICULTY OF DISCRIMINATION AND PRODUCTION FOR AN AMERICAN WITH NO KNOWLEDGE OF RUSSIAN. THE PAIRS OF SYLLABLES WERE PLAYED TO SS OVER A LOUDSPEAKER, AND THE SS WERE TOLD TO GIVE AN IMMEDIATE RESPONSE OF EITHER "SAME" OR "DIFFERENT" TO EACH SET. EXPERIMENT II WAS BASICALLY THE SAME AS I, WITH SLIGHT CHANGES IN THE SELECTION OF SUBJECTS AND STIMULUS MATERIAL. A LINGUISTIC INTERPRETATION OF THE RESULTS INDICATES THAT--(1) AN ADDITIONAL PRESENTATION OF THE VOICED-VOICELESS CONTRASTS MIGHT RESULT IN NEARLY PERFECT IDENTIFICATION, AND THAT PEDAGOGICALLY, REPEATED TRIALS ON THE SAME CONCEPT COULD BE MORE EFFECTIVE THAN RANDOM PRESENTATION, (2) GENERALLY, FRICATIVES ARE MORE EASILY DISTINGUISHED THAN STOPS, (3) SUBJECTS HAD DIFFICULTY DISTINGUISHING BETWEEN VOICED AND VOICELESS STOPS IN RUSSIAN, AND (4) MOST OF THE FINDINGS FULFILL THE EXPECTATIONS RESULTING FROM A CONTRASTING PHONEMIC ANALYSIS OF RUSSIAN AND ENGLISH. (FB)