This paper introduces the Causal Attitude Network (CAN) model, which conceptualizes attitudes as networks consisting of evaluative reactions and interactions between these reactions. Relevant evaluative reactions include beliefs, feelings, and behaviors toward the attitude object. Interactions between these reactions arise through direct causal influences (e.g., the belief that snakes are dangerous causes fear of snakes) and mechanisms that support evaluative consistency between related contents of evaluative reactions (e.g., people tend to align their belief that snakes are useful with their belief that snakes help maintain ecological balance). In the CAN model, the structure of attitude networks conforms to a small-world structure: evaluative reactions that are similar to each other form tight clusters, which are connected by a sparser set of ÒshortcutsÓ between them. We argue that the CAN model provides a realistic formalized measurement model of attitudes and therefore fills a crucial gap in the attitude literature. Furthermore, the CAN model provides testable predictions for the structure of attitudes and how they develop, remain stable, and change over time. Attitude strength is conceptualized in terms of the connectivity of attitude networks and we show that this provides a parsimonious account of the differences between strong and weak attitudes. We discuss the CAN model in relation to possible extensions, implication for the assessment of attitudes, and possibilities for further study. Key words: Network models, attitudes, tripartite model, connectionism, small-world consistency , recent connectionist modeling of attitudes (Monroe & Read, 2008) and recent advancements in applying network theory in psychology (e.g., Cramer, Waldorp, van der Maas, & Borsboom, 2010; for excellent discussions of the relevance of network analysis to the social sciences in general and psychology in particular see Borgatti, Mehra, Brass, & Labianca, 2009; Westaby, Pfaff, & Redding, 2014) to derive a set of requirements for a CAUSAL ATTITUDE NETWORK MODEL 6 realistic formalized measurement model of attitudes. Third, based on these requirements we develop the CAN model and discuss the proposed small-world structure of attitudes that underlies it. Fourth, we discuss the CAN modelÕs perspective on attitude formation and structure, attitude stability and change, and attitude strength. Fifth, we discuss possible extensions of the CAN model, the modelÕs implications for the assessment of attitudes, and some possible avenues for further study of the CAN model.