The purpose of this study is to identify what sociolinguistic variables are problematic for learners of Japanese and English as second languages and how those variables may affect the types of Compliment Responses (Crs) the learners choose in conversation. To investigate interlanguage CRs, this study compares the learner data with baseline data taken from native Japanese and American English speakers. The sociolinguistic variables studied were the targets (Family versus Self) and topics (External versus Internal) of compliments. The variables were combined into four target-topic categories: Family External, Family Internal, Self External, and Self Internal. A total of 60 informants participated in the study, of which 12 were American native speakers of English (AAs), 17 were native speakers of Japanese (JJs), 14 were Japanese learners of English (ESLs); and 17 were American learners of Japanese (JSLs). Each informants had a 30- to 60-minute conversation session with a female friend, or conversation leader, who was a native speaker of the language in which the conversation was conducted. The data was analyzed both quantitatively and qualitatively. All data was tape-recorded, narrowly transcribed, and coded by politeness type. The politeness types are those from Brown and Levinson's theory of politeness: Positive Politeness, Negative Politeness, Off-Record, and Do-not-Do-FTA (Face Threatening Act). In politeness type, the compliment responses of the learners were found to approximate those of the native speakers of the target languages. A detailed examination of the sociolinguistic variables of CRs and the dispreferredness of certain discourse structures, however, revealed that the learners use pragmatic transfer from their L1 and overgeneralize politeness types of their L2. The result of a chi-square test revealed that the use of Negative Politeness in Family External CRs was most significantly different across groups, whereas its use in Self External CRs was found to be the most similar across groups. The family variable in CRs was the most difficult for the learner groups, especially for the JSLs, who are not familiar with using Negative Politeness when responding to compliments about their family members. The study also investigated the correlation between politeness type and the intensity of the compliments and the correlation between politeness type and the different sequences of complements and their responses. Only among AAs did Positive Politeness CRs correlate with the intensity of the compliments, while ESLs seemed to be bothered by increased intensity.