In the target article, Murayama and Jach aim at the opacity of “higher-order” motivational constructs such as needs for competence, relatedness, and autonomy, and contend that “such high-level motivation is a subjective construal or emergent property of underlying mental computational processes which determine behavior.” Their point is well taken, and a useful reminder that (just like psychometrically derived personality traits) constructs such as needs and motives should be temporary placeholders for the operations of yet-to-be-identified psychological mechanisms.

My goal in this commentary is to bring some good news: there is an entire tradition of research on motivational systems, rooted in ethology and flourished within psychology thanks to the work of John Bowlby (Reference Bowlby1982) and others, that has been peering into the black box for decades with very interesting results (e.g., Kenrick, Griskevicius, Neuberg, and Schaller, Reference Kenrick, Griskevicius, Neuberg and Schaller2010). I recently contributed to this tradition with an architectural account of the mechanisms involved in motivation and their interplay (Fig. 1; Del Giudice, Reference Del Giudice2023a, Reference Del Giudice2023b, Reference Del Giudice, Al-Shawaf and Shackelford2024). The General Architecture of Motivation (GAM) clarifies the respective roles of different kinds of mechanisms, and seamlessly connects the two main functions of motivation – the “energization” and “direction” of behavior – through the goal-setting activity of emotions (see below).

Figure 1.

Schematic diagram of the general architecture of motivation (GAM). Reproduced with permission from Del Giudice (Reference Del Giudice2023a).

Motivational systems are not conceptualized as amorphous “internal variables,” but as specialized control systems that regulate behavior in fitness-critical domains such as mating, attachment, affiliation, caregiving, social status, as well as physical safety ad exploration (see Del Giudice, Reference Del Giudice, Al-Shawaf and Shackelford2024; Kenrick et al., Reference Kenrick, Griskevicius, Neuberg and Schaller2010, Reference Kenrick and Lundberg-Kenrick2022). They are cognitively impenetrable but experience- and context-sensitive, are typically regulated by feedback processes (though feedforward, anticipatory regulation is likely to be important for at least some of them), and orchestrate the onset of specific emotions when they are activated (or deactivated) by cues that signal domain-specific threats and opportunities. Motivational systems are amenable to computational modeling, as demonstrated by the various simulations of the attachment system proposed over the years (see Cittern and Edalat, Reference Cittern and Edalat2014; Petters and Beaudoin, Reference Petters, Beaudoin, Érdi, Bhattacharya and Cochran2017; Schneider, Reference Schneider and Wright2001). They energize and orient the individual in the pursuit of evolved goals, from more “basic” to more complex, such as obtaining protection and security, learning about the environment, defending and enhancing one's status, or caring for one's offspring and kin (see Del Giudice, Reference Del Giudice, Al-Shawaf and Shackelford2024). Their neurobiological substrates include functionally specialized “hubs” that collect and integrate cues relevant to a particular domain to orchestrate behavior and physiology on a broad scale (for a striking example, see the work on parenting circuits by Kohl and colleagues [Reference Kohl, Babayan, Rubinstein, Autry, Marin-Rodriguez, Kapoor and Dulac2018; Kohl, Reference Kohl2020; Kohl and Dulac, Reference Kohl and Dulac2018]).

What this approach does not explain is how individuals pursue instrumental goals – the explicitly represented, hierarchically organized goals that guide moment-to-moment actions throughout daily life (typically associated with the direction of behavior), and that are linked only indirectly to the unrepresented, innate goals embodied by motivational systems (associated with the energization of behavior). Historically, these two kinds of goals have been addressed by different, largely non-overlapping research communities. The GAM integrates them with the inclusion of a “programmable,” general purpose control system tasked with managing hierarchies of goals (each with its own importance/urgency, abstraction, and location in time); pursuing currently active goals by generating appropriate sub-goals and monitoring their success or failure; and sending concrete actionable goals to downstream systems for action selection and motor control. This Instrumental Goal Pursuit System (IGPS) is the natural computational substrate for “higher-order” motivations related to competence and mastery, which are not well accounted for by classic models of motivational systems.

But how can motivational systems regulate behavior, if moment-to-moment instrumental goals are under the control of the IGPS? One of the key insights of the GAM is that motivational systems control behavior indirectly by activating emotions, which in turn provide the IGPS with urgent, abstract goals and/or “stop signals” that instruct the IGPS to suspend or terminate currently active goals. The idea that emotions generate abstract, high-priority goals for the individual (e.g., “avoiding danger” in the case of fear; “cleansing oneself” in the case of disgust; “reaching proximity to the caregiver” in the case of attachment anxiety) is consistent with motivational theories of emotions, which speak of relational goals (Scarantino, Reference Scarantino, D'Arms and Jacobson2014) and emotivational goals (Roseman, Reference Roseman2011). The IGPS evaluates these goals according to their importance/urgency (likely based on the intensity of the corresponding emotions) and their compatibility with the existing goal structure; as a result, the goal hierarchy may be rearranged to include the new emotion-generated goals, derive concrete sub-goals, etc. In this way, emotions bridge the gap between qualitatively different kinds of goals, and serve as the “glue” that binds together multiple control mechanisms into a coordinated whole.

At the same time, the pursuit of instrumental goals by the IGPS is regulated by a set of procedural emotions that signal success and failure across domains and help regulate the allocation of the individual's effort – emotions like frustration, satisfaction, disappointment, and anxious indecision in response to unresolved conflicts between goals. This means that the control of goal-directed behavior can be represented by two nested loops, an outer loop managed by motivational systems and an inner loop managed by the IGPS. Crucially, both loops involve emotions, further underscoring the fact that energization and direction are not separate but intermixed functions. (A related implication is that goal pursuit always has an affective component, even in the case of instrumental goals, although there are differences in the specific kinds of emotions involved.)

There are other aspects of the GAM than I cannot discuss in this brief commentary, including a theory of moods as higher-order coordination mechanisms and the conceptual tools to describe individual differences in motivation (and personality) as differences in the operating parameters of motivational systems (e.g., activation and deactivation thresholds) and the IGPS (e.g., depth of the goal hierarchy, rigidity of goal priorities, persistence of striving in the face of failure, stringency of criteria for determining success). The latter are especially relevant in light of Murayama and Jach's call for “a theory of mental computational processes that explicitly addresses how intra-individual processes translates into long-term development” of stable individual differences (see Del Giudice, Reference Del Giudice2023a, Reference Del Giudice2023b). In short, I believe that this framework dovetails perfectly with the renovation project advocated in the target article, and offers researchers an evolutionarily grounded map of what used to be the inscrutable black box of motivation.