Operant Variability and Voluntary Action
Allen Neuringer and Greg Jensen
A behavior-based theory identified 2 characteristics of voluntary acts. The first, extensively explored in
operant-conditioning experiments, is that voluntary responses produce the reinforcers that control them.
This bidirectional relationship—in which reinforcer depends on response and response on reinforcer—
demonstrates the functional nature of the voluntary act. The present article focuses on the second
characteristic: a similar bidirectional relationship between reinforcement and the predictability/
unpredictability of voluntary acts. Support for the theory comes from 2 areas of research. The first shows
that levels of behavioral variability—from highly predictable to randomlike—are directly influenced by
reinforcers. Put another way, variability is an operant dimension, analogous to response rate and force.
The second source of support comes from psychophysical experiments in which human participants
judged the degree to which “choices” by virtual actors on a computer screen appeared to be voluntary.
The choices were intermittently reinforced according to concurrently operating schedules. The actors’
behaviors appeared to most closely approximate voluntary human choices when response distributions
matched reinforcer distributions (an indication of functionality) and when levels of variability, from
repetitive to random, changed with reinforcement contingencies. Thus, voluntary acts are characterized
by reinforcement-controlled functionality and unpredictability.
voluntary action, theories of free will, reinforced variability, concurrent schedules of
Debates have raged for thousands of years concerning voluntary
action: its characteristics, explanation, and, indeed, its very existence.
One of the more difficult issues has been how to reconcile the
apparent freedom of an individual’s actions with a lawful universe.
Many writers on this topic have claimed that we are able to initiate
actions to do—within reasonable limits—what we want to do, but
scientists and advocates of a scientific worldview have also main-
tained that we are physical/biological beings whose behaviors are
subject to the laws of nature. How to reconcile these two? In this
article we attempt to identify the types of observable behaviors that
lead to attributions of volition. Our claim is that voluntary behavior
is indicated by functional changes in reinforcement-controlled
behaviors including, importantly, levels of behavioral variability.
We provide two main lines of evidence. The first comes from
experimental studies showing that levels of response variability,
ranging from stereotyped repetitions to randomlike variations, are
controlled by reinforcement contingencies. Stated differently, vari-
ability is an operant dimension of behavior. The second is that
human participants attribute volition to agents whose behaviors,
including the variability (or predictability) of the behaviors,