1) Pavlov - Reinforcement is the US because when paired with the CS it produces the CR, whereas presenting the CS without the US leads to extinction of the CR

2) Thorndike - Reinforcement is a "satisfying state of affairs," defined behaviorally: "By a satisfying state of affairs is meant one which the animal does nothing to avoid, often doing such things as attain and preserve it."

3) Watson - Reinforcement ensures that the desired response is the one most frequently paired with the stimulus, because it's the response that must happen on every trial - each trial continues until the desired response is made, and reinforcement ends the trial, e.g., cats in the puzzle box don't have to scratch the wall, vocalize, move around, etc. on every trial, but they do have to press the paddle on every trial, because the reinforcement doesn't happen and the trial doesn't end until they do that.

4) Guthrie - Reinforcement works not because the animal has some goal or desire but because it changes the stimulus situation, thus protecting the successful response (the most recent one) from being replaced by a new response: no new response can be attached to those stimuli if they're no longer present (e.g., still inside the puzzle box after the cat has gotten out).

5 & 6) Hull - Reinforcement was initially defined as "drive reduction", i.e., anything that reduces drive, which linked it to biology and animals' drive to meet their biological needs. It was later redefined as "drive stimulus reduction" in the sense of merely seeming to meet those biological needs by providing the stimuli associated with them (chewing, swallowing, tasting, etc), which brought the concept back into the realm of the psychological ("seeming" is perceptual or cognitive, not biological).

(Note that Mazur (pp. 210 and 211) refers to these respectively as "Need Reduction" (= Hull's "drive reduction") and "Drive Reduction" (= Hull's "drive stimulus reduction"). That second pair of terms isn't exactly equivalent but they're close. On the exam I'll refer to the "earlier" or "later" version of Hull's definition and will make it perfectly clear which one I mean.)

7) Tolman - Reinforcement doesn't make learning happen, but is instead the motivation for performance

8) Skinner - Reinforcement (whether positive or negative) is anything that increases the rate of responding

9) James Olds showed that electrical stimulation of the brain in certain regions (Mazur p. 206) was apparently a very strong reinforcer for rats, even being preferred to eating, which suggested there was an underlying neural basis to all reinforcement; this led to some unreplicable findings and sometimes uninterpretable attempts to extend the work to humans, and was eventually abandoned as a universal explanation of how reinforcement might work biologically.

10) Premack's principle (Premack 1959): Reinforcers may be behaviors rather than stimuli, and they are relative in the sense that a more probable behavior in the animal's normal behavioral repertoire will reinforce a less probable behavior.

11) Response Deprivation Theory (Timberlake and Allison, 1974), the best modern view of reinforcement: the opportunity to engage in any behavior that is restricted relative to its baseline frequency will become reinforcing for any other behavior, as the animal attempts to restore its natural balance of all the variety of behaviors it prefers to engage in. That's different from Premack's principle, because it means even a low-probability behavior can reinforce a high-probability behavior -- as long as that low-probability one is restricted to be even less available than it would normally be.