Behavioral inhibition is one of the most frequently studied traits of child temperament. Although consistent with theoretical views of temperament, the biological and physiological underpinnings of behavioral inhibition have often been highlighted to the exclusion or minimization of environmental factors. Nevertheless, a growing literature suggests that environment plays a role in the development and expression of temperamental traits. Further, given a developmental psychopathology framework, the processes of normal development can be elucidated by the examination of maladaptive or otherwise atypical developmental patterns or environments (Cicchetti, 1984; 1990; Sroufe, 1990). The current study addresses the above issues in two ways—by examining the rates and stability of traditional behavioral inhibition across the second year of life in children from maltreating and nonmaltreating families and by proposing an alternate method of classifying behavioral inhibition for possible greater precision in non-normative, higher-risk samples. Behavioral inhibition was assessed through a standard laboratory paradigm when children were 14 and 26-months of age. Trained raters coded data for both fear behaviors (i.e., hesitancy to interact with stimuli, distress vocalizations, retreat, and clinging to the mother) and approach behaviors (i.e., positive affect, vocalization, and exploration of the stimuli). Behavioral inhibition was then defined by two methods, a traditional method based on total number of fear behaviors and an alternative method utilizing both fear and approach behaviors. Results indicate that when a traditional definition of behavioral inhibition was utilized, there were no differences in rates of classification between the maltreated and nonmaltreated samples; however, when categories were defined by both fear and approach, children from maltreating families were found to demonstrate greater rates of “lack of behavior,” characterized by a combination of low fear and low approach. Regarding stability of behavioral inhibition classification across the second year of life, there was little stability found for either method of categorization. Significant differences were found, however, in the stability of approach behaviors between the maltreated and nonmaltreated groups, with only the nonmaltreated participants displaying significant stability. Additionally, fear and approach behaviors showed predictive utility for traditionally-defined behavioral inhibition classification, but only for nonmaltreated children. Finally, differences in the expression of positive or negative affect between the two groups were minimal and thus did not appear to account for other differences between the maltreated and nonmaltreated participants. The findings of the present study suggest that both fear and approach are integral in the determination of behavioral inhibition and may be particularly useful in elucidating differences between normative and atypical samples. The differences observed between the maltreated and nonmaltreated participants suggest that the experience of maltreatment altered the expected patterns of behavior given past research on behavioral inhibition. Thus, it is possible that exposure to maltreatment within the first year of life modifies the biological propensity for the expression of fear or approach when confronted with novel stimuli.