This paper explores the conceptual limitations of a behavioral approach to
the understanding of couples' interaction. It notes the lack of longitudina
l evidence for the ability of behavioral observational coding systems to pr
edict divorce and to identify processes leading to it. It demonstrates the
incompatibility of behaviorial and linguistic analysis, and points to numer
ous inconsistencies and inaccuracies in the two major behavioral observatio
nal coding systems: the CISS (Couples' Interaction Scoring System) and the
RCISS (Rapid Couples' Interaction Scoring System). The paper highlights the
inability of these behavioral coding systems, and others, to deal adequate
ly with issues of gender, race, subtext, positivity and negativity, and mis
understanding.