Automaticity is a concept in psychology with a long history, used in various domains such as perception, memory, social cognition, learning, motivation, and emotion. Despite its central role, there is no consensus on its definition. This article provides an in-depth analysis of the concept and its features, discussing implications for future research. Several theoretical views of automaticity are discussed, focusing on features such as unintentional, uncontrolled, autonomous, unconscious, efficient, and fast. Contemporary views suggest these features should be investigated separately. The authors examine whether these features can be disentangled conceptually, as this would justify separate investigation. They conclude that conceptual analysis is largely feasible, though some researchers disagree, arguing that feature overlap is determined by their views of automaticity and information processing models. Automaticity is often viewed as processing with minimal attention, with early models considering attention as a structural property of the information-processing system. Later models, however, viewed attention as a flexible resource. The capacity view of automaticity suggests that processes become automatic with practice, requiring less attention. This view has been challenged, with studies showing that features of automaticity do not always co-occur, leading to the rejection of an all-or-none view. The dual-mode model of information processing, proposed by Shiffrin and Schneider, suggests that automatic and nonautomatic processes represent two opposite modes of processing. However, this model has been criticized for promoting an all-or-none view, which is not supported by empirical evidence showing that features of automaticity can overlap. The gradual view of automaticity suggests that automaticity is a continuum, with processes becoming more automatic with practice. Logan proposed that automaticity is determined by the amount of training, with features evolving over time. This view challenges the idea of a fixed, feature-based definition of automaticity. The capacity view of automaticity has been challenged for its assumptions, including the idea that automatic processes are unconstrained by attentional capacity and the lack of an explicit learning mechanism. Alternative accounts, such as the algorithm strengthening view, suggest that automatization involves the strengthening of algorithms, leading to faster and more efficient processing. The article also discusses the features of automaticity, including goal-related features such as (un)intentional, goal directed, goal (in)dependent, (un)controlled, and autonomous. These features are often conceptually overlapping, making it difficult to separate them. The authors argue that a decompositional approach is viable only if features are conceptually separable. In conclusion, the article emphasizes the importance of investigating features of automaticity separately, as this approach provides a more nuanced understanding of the concept. The gradual view of automaticity is favored, as it allows for a more flexible and comprehensive understanding of the concept. The authors argue that a feature-based approach is the most cautious and informative for diagnosing automaticity.Automaticity is a concept in psychology with a long history, used in various domains such as perception, memory, social cognition, learning, motivation, and emotion. Despite its central role, there is no consensus on its definition. This article provides an in-depth analysis of the concept and its features, discussing implications for future research. Several theoretical views of automaticity are discussed, focusing on features such as unintentional, uncontrolled, autonomous, unconscious, efficient, and fast. Contemporary views suggest these features should be investigated separately. The authors examine whether these features can be disentangled conceptually, as this would justify separate investigation. They conclude that conceptual analysis is largely feasible, though some researchers disagree, arguing that feature overlap is determined by their views of automaticity and information processing models. Automaticity is often viewed as processing with minimal attention, with early models considering attention as a structural property of the information-processing system. Later models, however, viewed attention as a flexible resource. The capacity view of automaticity suggests that processes become automatic with practice, requiring less attention. This view has been challenged, with studies showing that features of automaticity do not always co-occur, leading to the rejection of an all-or-none view. The dual-mode model of information processing, proposed by Shiffrin and Schneider, suggests that automatic and nonautomatic processes represent two opposite modes of processing. However, this model has been criticized for promoting an all-or-none view, which is not supported by empirical evidence showing that features of automaticity can overlap. The gradual view of automaticity suggests that automaticity is a continuum, with processes becoming more automatic with practice. Logan proposed that automaticity is determined by the amount of training, with features evolving over time. This view challenges the idea of a fixed, feature-based definition of automaticity. The capacity view of automaticity has been challenged for its assumptions, including the idea that automatic processes are unconstrained by attentional capacity and the lack of an explicit learning mechanism. Alternative accounts, such as the algorithm strengthening view, suggest that automatization involves the strengthening of algorithms, leading to faster and more efficient processing. The article also discusses the features of automaticity, including goal-related features such as (un)intentional, goal directed, goal (in)dependent, (un)controlled, and autonomous. These features are often conceptually overlapping, making it difficult to separate them. The authors argue that a decompositional approach is viable only if features are conceptually separable. In conclusion, the article emphasizes the importance of investigating features of automaticity separately, as this approach provides a more nuanced understanding of the concept. The gradual view of automaticity is favored, as it allows for a more flexible and comprehensive understanding of the concept. The authors argue that a feature-based approach is the most cautious and informative for diagnosing automaticity.