The debate over the extensive use of paired and group oral assessments has yielded an enriched body of theoretical inquiries and empirical investigations, marking a shift from viewing second language (L2) speaking ability as residing in the individual to emphasizing joint construction distributed among interlocutors within local contexts. This paper aims to review this line of research to provide an enhanced understanding of peer-to-peer interaction so as to better inform and advance L2 speaking construct conceptualization, test design, rating scales development, and rater training. This paper also provides an account of the theoretical background underlying paired and group orals, and then related empirical studies are reviewed category by category. Reviewed research findings are discussed, critiqued, and synthesized. The paper concludes with implications and recommendations for future research.