Abstract
This paper, originally proposed in 1972, outlines a formal information-processing model of human motivation, shifting the discourse from vague "psychic energy" metaphors to a rigorous computational framework. By conceptualizing the human mind as a heterarchical time-sharing system, the author defines the "will" as a resource allocator—similar to a computer's operating system supervisor—that manages parallel, competing goal-processes. This model provides a structural basis for understanding how wishes translate into actions and offers a path toward applying artificial intelligence in clinical psychology and psychiatry.
Introduction
In the early 1970s, motivation was a largely neglected domain in both psychology and artificial intelligence. Psychological theories of the era often relied on "hydraulic" analogies of energy (e.g., Freud or Tinbergen), which lacked the precision necessary for formal modeling. Meanwhile, AI research focused primarily on logic and problem-solving, often ignoring the underlying mechanisms that determine which goals a system should pursue. G.R. Kiss argues that for AI to be useful in clinical contexts, it requires a formal model of personality that explains how "wishes" and "wants" are managed within a system of finite resources.
The Heterarchical Architecture
The paper proposes a heterarchical system—a non-hierarchical structure where various mental resources interact as equals rather than in a strict top-down command chain. Kiss divides the personality into five core resource types:
- Central Control: The seat of consciousness and attention.
- Knowledge and Reasoning: The repository of concepts, beliefs, and inference mechanisms.
- Motivation: The system of goals, values, and intentions.
- Affect: The domain of emotions and attitudes.
- Somatic Interfaces: The channels for sensory input and motor output.
The Will as a Resource Allocator
The core of the Kiss model is the treatment of the "Will" as a resource allocator. Drawing a direct parallel to the time-sharing computer systems of the early 1970s, the author suggests that multiple goal-processes run in parallel within the human mind. These processes compete for "computational work" (CPU time) and memory space. In this framework, the Will does not provide "energy" in the biological sense; instead, it acts as a supervisor or scheduler that determines which process receives the necessary computational resources to proceed toward execution.
Computational Definitions of Motivation
Kiss provides sharp computational distinctions for previously nebulous psychological terms:
- Wishes vs. Wants: A wish is defined as a state description with an assigned utility (value). A want is a wish that has successfully secured an allocation of resources for action.
- Mental Effort as Computational Work: The model replaces the metaphor of "psychic energy" with "computational work"—the actual number of logical operations required to perform a task.
- Concentration: This is viewed as the act of protecting a specific process from interruption or preemption by competing goals.
Goal Dynamics and Preemption
While many goals exist in parallel, the model suggests that one is usually "regnant," or in primary control of the system's output. However, other "wishes" do not simply disappear; they continue to monitor the environment reactively.
This "Reactive Motivation" allows the system to detect if environmental changes have suddenly made a dormant wish more attainable. If an obstacle to a dormant goal is removed, that goal can "preempt" the current regnant process. Kiss suggests that this mechanism provides a computational explanation for "Freudian slips" and other instances where suppressed intentions suddenly manifest in behavior.
Conclusion and Mapping
The primary contribution of this work is the formal mapping of psychological phenomena onto computer science concepts, providing a vocabulary for future implementation:
- The Will $\rightarrow$ Operating System Supervisor/Scheduler.
- Psychic Energy $\rightarrow$ Computational Work (Logical Operations).
- Intention $\rightarrow$ The adoption of a specific plan for execution.
- Satisficing $\rightarrow$ The decision to terminate a process when a goal state is "close enough."
Limitations and Future Work
As a theoretical outline, the paper acknowledges that implementation is in the early stages. The author notes that the time-sharing analogy is imperfect, as human goal-processes communicate and shift control more fluidly than the rigid OS schedulers of the 1970s. Additionally, while "Affect" is identified as a core resource, its specific mechanics were deferred for later research. The ultimate goal remains the development of computer-based systems capable of assisting in psychiatric and clinical psychological practice.
Summary updated