THE HYPOTHESIS TESTED DERIVES FROM THE MORE GENERAL HYPOTHESIS THAT RESPONSIVENESS TO SOCIAL REINFORCEMENT IS ACQUIRED IN THE EARLY RELATIONSHIPS BETWEEN THE INFANT AND HIS CARETAKERS, THE SOCIAL CUES GAINING SECONDARY REINFORCEMENT VALUE DUE TO THEIR PAIRING WITH OTHER REINFORCEMENT. EVIDENCE ABOUT THE EARLY EXPERIENCE OF DELINQUENTS SUGGESTS THAT THE LEARNING CONDITIONS FOR ESTABLISHING STRONG POSITIVE AND NEGATIVE CONDITIONED SOCIAL CUES MAY HAVE BEEN LESS EFFECTIVE FOR THEM THAN FOR NON-DELINQUENTS, ALTHOUGH SUCH EVIDENCE HAS THE USUAL DEFECT OF BEING GATHERED RETROSPECTIVELY. TO TEST THIS HYPOTHESIS, A SAMPLE GROUP OF DELINQUENTS WAS ASKED TO PERFORM A SIMPLE TASK FOR WHICH THEY WOULD RECEIVE SOCIAL REINFORCEMENT. SEVERAL DIFFERENT TYPES OF REINFORCEMENT WERE TESTED, INCLUDING SOCIAL APPROVAL, MONEY, NO REINFORCEMENT, PEER REINFORCEMENT, APPROVAL, VARIED REINFORCEMENT, AND PUNISHMENT. THE HYPOTHESIS RECEIVED SUPPORT WHEN A FEMALE ADULT OR PEER ADMINISTERED REINFORCEMENT; WHEN THE CONTINGENT SOCIAL CUES WERE REWARDING ('GOOD' OR 'FINE'); AND WHEN THEY WERE PUNISHING ('NO'). THE HYPOTHESIS WAS NOT SUPPORTED WHEN THE VERBAL REINFORCEMENT WAS VARIED. CONTROL STUDIES USING NO REINFORCEMENT AND MONEY REINFORCEMENT INDICATED THAT THE SOCIAL REINFORCEMENT DID INDEED HAVE A SPECIFIC REINFORCEMENT EFFECT AND THAT DELINQUENTS DID NOT SHOW A GENERAL LEARNING DEFICIENCY. (AUTHOR ABSTRACT MODIFIED)