François Tonneau

Equivalence Relations: A Critical Analysis

In terms of empirical output and number of published articles, equivalence-class research seems remarkably healthy. However, closer examination reveals systematic confusions around the most fundamental concepts of the field: those of “relations” and “equivalence.” Through the invocation of undefined “relations,” no less than four different concepts of equivalence are regularly confused with one another. In this article I document the existing confusions and suggest a reorientation of behavioral research on symbolic functioning. This reorientation involves a shift from “stimulus equivalence” to functional equivalence, and from the matching-to-sample procedure to stimulus correlations.

Who Can Understand Relational Frame Theory? A Reply to Barnes-Holmes and Hayes

95 Before replying to Barnes-Holmes and S. C. Hayes’ questions (this issue), some background may help the readers understand the nature of the current debate. Among the commentaries on Equivalence relations: A critical analysis (Tonneau, 2001) that were published in this journal, some defended relational frame theory (RFT) as a promising alternative to a correlation-based approach to symbolic behavior. Barnes-Holmes, S. C. Hayes, and Roche (2001), in particular, defended RFT and criticized the approach I proposed on two main issues. A first issue concerned possible difficulties with symmetry in matching to sample and with A-B, A-C function transfer when the training trials are mixed. After suggesting that a Pavlovian process could not operate in the latter conditions “unless one wishes to invoke the problematic concept of backward associative conditioning,” Barnes-Holmes, S. C. Hayes, and Roche (2001, p. 37) concluded that I had “completely failed” to address the emergence of bidirectional behavioral phenomena. A second challenge raised by Barnes-Holmes, S. C. Hayes, and Roche (2001) involved experimental results such as those of Barnes, Hegarty, and Smeets (1997) and Dymond and Barnes (1995). Unfortunately for these arguments, explanations of matching symmetry in terms of forward pairings have been available since 1992 (L. J. Hayes, 1992: cited in my target article on p. 19; more on this below), function transfer after mixed A-B, A-C trials does not require backward conWho Can Understand Relational Frame Theory?

Equivalence Relations: A Reply

Some commentaries on Equivalence relations: A critical analysis (this issue) have questioned the consistency and generality of a correlation-based alternative to equivalence-class research, whereas others defend the use of matching-equivalence concepts in behavior theory. In this reply I reiterate the most important points of the target article, provide further clarifications, and discuss various misunderstandings. In contrast to equivalenceclass notions, the concept of function transfer is clear, simple, and coherent; and it necessarily plays a crucial role in the behavioral analysis of complex psychological functioning. A molar view based on environmental networks is well qualified to explain function transfer and thus provide insights into a variety of complex behavioral phenomena.

Human group behavior: the ideal free distribution in a three-patch situation

A group of 15 college students was exposed to repeated trials of a task in which money was available for choosing among three colors (blue, red, and green). The amount of winning tokens for each color was varied across phases to test whether group distribution would track the ratio of winning tokens between patches. Confirming previous reports on ideal free performance in humans, group choice proved sensitive to the available resources but tended to undermatch the ratio of winning tokens. The difference-equalization rule of Sokolowski, Tonneau, and Freixa i Baqué [Psychonom. Bull. Rev. 6 (1999) 157] gave a satisfactory fit to the data.

Early-session increases in responding during extinction

After training under a variable-interval 60-s schedule of reinforcement, four rats were exposed to 30-min extinction tests, which occurred either at the start or at the end of the session (each session being 50-min long). Response rate in extinction decreased when the extinction test occurred at the end of the session, but first increased and then decreased when the extinction test occurred at the start of the session. Consistent with other recent results, this finding suggests that some variable, other than reinforcement, contributes to early-session increases in responding.

Pitfalls of Behavioral Selectionism

Numerous analogies have been offered to relate evolutionary and behavioral phenomena. B. F. Skinner (1981/1988), in particular, has suggested that similar processes of selection operate at the levels of evolution, behavior, and culture. This essay examines Skinner’s proposed analogy between natural selection and operant reinforcement. We argue that the analogy fails to characterize a shared causal structure. Operant reinforcement mimics superficial aspects of natural selection through entirely different mechanisms; hence no detailed understanding of operant behavior can be expected from selectionist notions. Selection analogies with respect to reinforced behavior are generally useless and probably misleading. The failure of Skinner’s selection metaphor reflects on the current relations of some parts of psychology to evolutionary theory.

Behaviorism in 1942: A Precis of Tilquin's Le Behaviorisme: Origine et Developpement de la Psychologie De Reaction en Amerique.

Book-length treatments of behaviorism from a philosophical and historical perspective are few in number. Tilquin’s (1942) is one of these, but its publication in French during World War II and the limited number of available copies make for difficult access. In this paper, I summarize the contents of the book for a general audience of behavior analysts. Tilquin’s work is a useful tour of the behaviorism of its time, and most of the topics discussed in it remain relevant to behavior analysis.

The Observational Analysis of Behavior

Dillenburger and Keenan’s use of non-experimental methods represents a significant departure from the scientific orthodoxy of behavior analysis. However, field studies cannot be by-passed if behavior analysis is to provide serious accounts of natural phenomena such as bereavement. Here I explore some theoretical implications of a non-experimental approach to behavior.