The insights described by Nevin & Grace may be summarized without reference to the Newtonian concepts they suggest. The metaphor to Newtonian mechanics seems dubious in three ways: (1) extensions seem to lead to paradoxes; (2) many well-known phenomena are ignored; (3) the Newtonian concepts seem difficult to reconcile with the larger framework of evolutionary theory.

Nevin & Grace's approach is an interesting and useful attempt to find ways to measure “core” effects of a history of exposure to reinforcement. The momentum analogy makes intuitive sense, and the evidence for its utility is increasing. Several questions remain, however, about how the analogy will fare in the case of concurrent rather than sequential activities, about the use of extinction as a method to test resistance to change, and about the generality of some of the effects.

Evidence from 45 early studies of resistance to extinction following reinforcement of differing amounts, taken in sum, challenges both the basic and the augmented models of Nevin & Grace. The augmented model seems too ad hoc in salvaging the analogy between persistence in behavior and concepts from physics, as my meta-analysis of these data affirms.

Metaphors and models involve correspondences between events in separate domains. They differ in the form and precision of how the correspondences are expressed. Examples include correspondences between phylogenic and ontogenic selection, and wave and particle metaphors of the mathematics of quantum physics. An implication is that the target article's metaphors of resistance to change may have heuristic advantages over those of momentum.

Nevin & Grace's behavioral-momentum model accommodates a large body of data. This commentary highlights some experimental findings that the model does not always predict. The model does not consistently predict resistance to change when response-independent food is delivered simultaneously with response-contingent food, when drugs are used as response disrupters, and when responding is reinforced under single rather than multiple schedules of reinforcement.

Nevin & Grace identify a difference between the predictions of delay reduction theory and the contingency-ratio account underlying behavioral momentum approaches to choice. This is shown not to be a true difference. The role of the overall context of reinforcement must be carefully incorporated by any theory of choice.

Nevin & Grace invoke a behavioral metaphor from the physics of momentum. The idealized assumptions they invoke are argued to translate to behavior only in the limited case of steady-state, constant-probability VI responding. Rather than further refine this limit case, mathematical models should be applied to generalizations of the limit case itself, broadening our understanding of behavioral processes.

Nevin & Grace present a fresh and thoroughly empirical argument undermining the traditional view of learning based on response contingent reinforcement. We endorse the main points in the target article and relate them to an ethological feed forward principle that complements the principle of behavioral momentum.

Nevin & Grace (N&G) buttress their metaphor with some good props. However, it is still not clear what momentum is analogous to. If momentum is a measure of strength, then the authors should say so and tell us how to calculate it. Furthermore, if “other” behavior can be introduced into the equation (and N&G's foray into the applied world suggests that it can), it is unclear when the masses are accrued and how much is accrued to each behavior.

Although the momentum metaphor is successful in many ways, there remain problems with the adequacy of the notion of a force in the behavioral sense and the question of whether the conditions used to apply force can truly be separated from the conditions that establish and maintain behavioral mass.

We briefly discuss some potential contributions of behavioral momentum research to the study of the behavioral effects of abused drugs. Contributions to the study of the direct effects of drugs on operant responding and to the study of drugs as reinforcers are addressed. Too little empirical evidence is available to thoroughly evaluate the relevance of behavioral momentum concepts to the study of drugs and behavior, but we note several reasons for optimism regarding its potential to make positive contributions.

Despite corrigible details, Nevin & Grace forge a clearer place for persistence as a fundamental attribute of motivated behavior and assay converging experimental operations in its measurement.

Although conditioned suppression has face validity as a technique for assessing resistance to change of operant behaviour, it is not discussed by Nevin & Grace. However, application of their approach to the results of a conditioned suppression study that varied food deprivation and reinforcement magnitude (Leslie 1977) produces paradoxical results.

The goal of therapy is often to increase the momentum (persistence) of self-control behaviors. Determining how best to accomplish this goal necessitates conducting behavioral momentum research under a wider variety of conditions.

An important measure of the validity and utility of basic behavioral research is the extent to which it can be applied in real life. Basic research on behavioral momentum and the model unifying choice and resistance to change (Nevin & Grace 1999) has stimulated the development of behavioral technologies aimed at increasing the persistence of adaptive behavior and decreasing maladaptive choices.

This commentary considers factors that may account for the inconsistency between the behavioral momentum formulation and the partial reinforcement extinction effect. The method of testing, the variability of the schedule, the nature of the response-contingency, and response effort are considered. Some applications to real-world problems are also discussed.

Behavioral momentum is a part of the larger field of behavioral dynamics concerned with modeling conditions controlling changes in behavior. The analogy of behavioral momentum to Newtonian and Einsteinian dynamics is briefly treated along with additional physical intuitions related to resistance to behavior change and preference.

We have analyzed many discrimination learning difficulties as reflecting multiple stimulus control topographies (SCTs). Nevin & Grace's analysis offers new variables to consider in the design of stimulus-control shaping procedures and cross-setting generalization of newly established behavior. A multiple-SCT perspective also suggests that fixed-trial discrimination procedures may offer advantages for reconciling momentum theory and partial reinforcement extinction effects.

Grace's contextual-choice model can account for the results from many studies on choice under concurrent-chain schedules. However, other models, including one that I call the “hyperbolic value-added model,” can also account for these results. Preference and resistance to change may indeed be related, but the best model of preference remains to be determined.

The concept of behavioural mass provides one avenue for justifying (or making rational) the phenomenon of declining impatience, according to which decision makers put more value on delays that will occur in the near future than on those that will occur later.