Abstract
How behaviors vary among individuals and covary with other behaviors has been a major topic of interest over the last two decades, particularly in research on animal personality, behavioral syndromes, and trade-offs with life-history traits. Unfortunately, proposed theoretical and conceptual frameworks explaining the seemingly ubiquitous observation of behavioral (co)variation have rarely successfully generalized. For example, the “pace-of-life syndrome hypothesis” proposes that behaviors, life-history, and physiological traits should be correlated in a predictable manner. However, these predictions are not consistently upheld. Two observations perhaps explain this failure: First, phenotypic correlations between behaviors are more strongly influenced by correlated and reversible plastic changes in behavior than by among-individual correlations which stem from the joint effects of genetics and developmental plasticity. Second, while trait correlations are frequently assumed to arise via trade-offs, the observed pattern of correlations is not consistent with simple pair-wise trade-offs. A possible resolution to the apparent inconsistency between observed correlations and a role for trade-offs is provided by state-behavior feedbacks. This is critical because the inconsistency between data and theory represents a major failure in our understanding of behavioral evolution. These two primary observations emphasize the importance of an increased research focus on correlated reversible plasticity in behavior—frequently estimated and then disregarded as within-individual covariances.