Response-independent Events In The Behavior Stream.
Pigeons choose by totaling the next three or four reinforcers, not just the closest one.
01Research in Context
What this study did
Adkins et al. (1997) watched pigeons choose between two keys on concurrent-chain schedules. Each key led to a different string of upcoming reinforcers.
The team asked: do birds care only about the next reinforcer, or do they look further ahead? They tested math models that weighed one, two, three, or four future reinforcers.
What they found
The best fit came from a sums-of-reciprocals model that counted the next three or four reinforcers. Models that used only the next reinforcer or simple averages lost.
In plain words, pigeons act as if they read a short menu of future payoffs, not just the next item.
How this fits with other research
Hawkes et al. (1974) and Mellitz et al. (1983) showed pigeons track immediate reinforcement odds. K et al. extend those ideas by showing birds also sum several upcoming rewards.
Coe et al. (1997) ran the same title and setup but saw mixed, smaller effects. The difference: A et al. inserted extra free food that was not part of the choice keys. That noise blurred the choice pattern, so the same birds looked less strategic.
Rilling et al. (1969) proved time allocation matches reinforcement ratios. K et al. keep the matching spirit but swap simple ratios for a short mental ledger of future events.
Why it matters
When you write token boards, response-cost systems, or chained schedules, think beyond the next reinforcer. Build three to four visible steps so clients can "see" the upcoming payoff string. If you add surprise freebies, know they may temporarily weaken the choice pattern you are trying to teach.
Get CEUs on This Topic — Free
The ABA Clubhouse has 60+ on-demand CEUs including ethics, supervision, and clinical topics like this one. Plus a new live CEU every Wednesday.
Sketch the next four tokens or rewards on a mini-schedule card and show it to the learner before the session starts.
02At a glance
03Original abstract
When pigeons choose between situations that provide access to food reinforcers after a delay, choice is better predicted by computations based upon sums‐of‐reciprocals distances from the point of choice to each of the next three or four reinforcers in series than by computations of optimality based upon mean rates of reinforcement. The present experiments were designed to examine the generality of this finding. Pigeons were exposed to concurrent‐chains schedules in which one brief initial link led to a fixed‐ratio schedule (either 15, 30, or 60, depending on the condition), and the other link led to a geometrically increasing progressive‐ratio schedule whose rate of escalation was systematically varied across conditions. Each combination of fixed‐ratio size and escalation rate of the progressive schedule was assessed at two different levels of deprivation (75% and 80% of free‐feeding weights). Computations based upon the sums‐of‐reciprocals principle, treating ratio schedule sizes as proportional to delays, predicted and described the pigeons' median switch points better than those based on arithmetic means. Neither the distance to the next reinforcer (as implied by some molecular analyses) nor molar optimization (as described by arithmetic means) were as successful at accounting for patterns of choice in these situations. Hence, it appears that the birds' choices were most influenced by the relative proximity of a choice to several reinforcers in a series of reinforcers, with each of the less proximal reinforcers having relatively less influence over the current choice.
Journal of the experimental analysis of behavior, 1997 · doi:10.1901/jeab.1997.68-357