American Scientist

To the Editors:

Read Montague's introduction to his terrific experiment, as described in "Addicts Bet on the Present" by Fenella Saunders (Science Observer, July-August) needs some modest correction. Dr. Montague says "There's no other animal on the planet that can do that," referring to suicide by hunger strike.

Around 1958 in the Memorial Hall Basement at Harvard, B. F. Skinner demonstrated that a pigeon trained first on a long forced delay of reinforcement (DRL) in the presence of, say, a green response key, would, when presented with both the green key and a red key that extinguished the green key, respond by pressing the red key. In this way it deprived itself of food. It was kept alive only by the good graces of the animal caretaker, Mrs. Papp, who secretly fed the about-to-expire pigeon that had lost about 40 to 60 percent of its normal body weight.

We all called the experiment the Hungarian Revolution Experiment for obvious reasons. Skinner's lab had demonstrated that even when known positive reinforcers with the power to shape the structure of a response are available, the noxious properties of the positive DRL schedule can result in life-threatening consequential behavior.

I know of no reference to this study. In those days many findings were in lab notebooks, but never reached a journal. We respected the outcomes, but many of us thought these grace notes to the primary research would deflect attention from the real studies, as exemplified by Skinner's Baby-Box work with Julie Skinner.

Eugene Galanter
Columbia University

To the Editors:

Read Montague says that the risk-accepting behavior of smokers may be due to "fictive error," which amounts to an uncoupling of an idea from its consequences. There are other possible interpretations of the experimental findings. Is life like a game of craps or a game of blackjack? In craps, the odds are selfrestoring with every roll. The dice do not remember what happened on the last time out. In blackjack, the odds on the next card depend on what is left in the residual pack. The smoker may operate on the assumption that market behavior in the future will not depend on how much he won or lost on the last round. By the way, who won the most money, the smoker or the nonsmoker?

M. David Tilson
Columbia University

Dr. Montague responds:

The game was designed so that no one wins more money than anyone else. The smokers and nonsmokers made the same amount of money. This design feature allowed us to disentangle money obtained from the variables that control choice. The game was also designed so that the history of choices had meaning—that is, it is unlike dice where the probabilities "reset" or are independent of one another.