Murayama and Jach (M&J) raise an important point by highlighting that some constructs in motivation science are underdeveloped: They are used as an end point of research and not as its beginning. We agree that some researchers and theorists do not show enough curiosity to fully specify the constructs that they use and do not show sufficient vigor in detailing specific motivational mechanisms or do not make them explicit enough. However, the conclusion that this is a general problem of motivation science is not warranted. Motivation science has always looked at filling the “black box” by specifying constructs and explaining how behavioral tendencies are generated. Computational models of motivation have been around for decades – even if they have not been labelled “computational.”

For instance, building on Lewin, Dembo, Festinger, and Sears's (Reference Lewin, Dembo, Festinger, Sears and Hunt1944) formal theory of resulting valence, Atkinson's risk-taking model (Reference Atkinson1957) suggested that the direction of achievement behavior – whether one approaches or avoids a specific task – depends on the relative strength of two competing motivational forces: The motivation to achieve success and the motivation to avoid failure. These two high-level constructs were further specified by postulating that they are determined by the subjective probabilities and incentive values of success and failure and weighted by individuals’ motives to achieve success and to avoid failure. Atkinson and Lewin et al.'s models thus did not only suggest high-level constructs that determine the direction of behavior, but also elaborated on the mechanisms underlying these constructs using an approach that would nowadays be called “computational.” Another early motivation theory specifying high level constructs – drive and habit – and offering an explicit computational model outlining the mechanisms determining the direction and intensity of behavior is Hull's (Reference Hull1943) drive reduction theory. More recently, Kruglanski et al.'s (Reference Kruglanski, Bélanger, Chen, Köpetz, Pierro and Mannetti2012) model of cognitive energetics has used the higher-order constructs potential driving force, restraining force, and effective driving force to explain the energization of behavior, the selection of goals, and the likelihood of goal attainment. This model also elaborates on the mechanisms underlying the postulated high-level constructs: Potential driving force is suggested to be a function of goal importance and the amount of available resources, and restraining force is predicted to be determined by resource conservation tendency, task difficulty, and the salience and importance of alternative goals. Importantly, Kruglanski et al.'s model also includes a process perspective by suggesting that the strength of potential driving force and restraining force are computed and compared before a decision about whether to engage in goal pursuit and how much effort to invest is taken. Another example is motivational intensity theory (Brehm & Self, Reference Brehm and Self1989), which has been explicitly acknowledged by M&J as a model that elaborates on the mechanisms underlying motivated behavior. Motivational intensity theory suggests that task difficulty and success importance – which are postulated to be a function of need state, level of instrumentality, and incentive value – jointly determine the effort that is invested in goal pursuit. The specific process by which difficulty and success importance determine effort is predicted to be a function of the clarity of task difficulty (Richter, Reference Richter2013). Like Kruglanski et al.'s model, motivational intensity theory suggests a specific sequence in which the computations are executed: Clarity of task difficulty information is processed first, followed by an assessment and comparison of task difficulty and success importance that determines whether one engages in the task at hand, and a final decision about how much effort is exerted.

The models described in the preceding paragraph constitute only a subset of the motivation-related theories that have already done what M&J ask for. Carver and Scheier's (Reference Carver and Scheier1981) control theory, Lewin's (Reference Lewin1939) field theory, Locke and Latham's (Reference Locke and Latham1990) goal setting theory, Kukla's (Reference Kukla1972) attributional theory of performance, or Vroom's (Reference Vroom1964) valence-instrumentality-expectancy theory constitute further examples of models that are not limited to high-level constructs but unpack the “black box” by describing specific mechanisms that underlie the high-level constructs and clarify how they motivate behavior. Moreover, in many of these models, motivation is not considered as the initial cause of behavior but the result of a multitude of processes. It is also of note that some models of motived behavior that seem to offer only high-level constructs often implicitly postulate more complex mechanisms underlying motivated behavior. For instance, self-determination theory's (Ryan & Deci, Reference Ryan and Deci2017) high-level concept of autonomous motivation seems to constitute at first sight one of the high-level, “black box” concepts that M&J criticize. However, even if it is not frequently explained in work on self-determination, autonomous motivation is not considered to be a direct determinant of behavior. For instance, autonomous motivation is supposed to influence performance via the intervening variables perceived locus of causality, perceived volition, and freedom of choice (Cerasoli, Nicklin, & Nassrelgrgawi, Reference Cerasoli, Nicklin and Nassrelgrgawi2016; Reeve, Reference Reeve2009). Based on this theorizing, one could even argue that autonomous motivation is not considered to be the initial driving force of performance but only one of many variables that are used as input for the computational mechanisms underlying performance.

The preceding paragraphs demonstrate that motivation science has always been concerned with mechanisms underlying motivated behavior. It is certainly true that in some work the underlying mechanisms did not get the attention that they deserve. It is also true that focusing on the mechanisms underlying high-level motivation constructs provides a great opportunity to advance our understanding of how the direction and intensity of behavior are determined. However, we disagree with M&J's position that motivation science in general avoids specifying what is inside the “black box” of high-level motivation constructs. Considering mental computational processes in motivation theories is neither new, nor something that motivation scientists need to begin to focus on. It has been an integral part of motivation science for decades. We therefore consider it of lesser relevance to remind motivation scientists that they should examine specific mechanisms underlying motivated behavior. The more important question to us is why not all motivation scientists focus on these mechanisms and why some researchers seem to be satisfied with not looking into the “black box.”