On the role of "memory" in the analysis of behavior.

Branch (1977) wrote a think-piece, not an experiment. The paper asks: do we need the word "memory" to explain why behavior changes?

The author says no. Instead, watch how behavior moves from one state to another. Track visible transitions, not invisible storage bins in the head.

The paper finds that "memory" is extra baggage. We can describe the same data by plotting what the organism does next, given what just happened.

The call: build single-subject charts that show irreversible patterns. If behavior stays changed, the past is built into the new pattern—no memory label needed.

Nevin (1982) and Morris et al. (1982) pick up the same baton. They widen the fight from "memory" to all cognitive words. The rule stays the same: if a concept does not improve prediction or control, drop it.

Bachman et al. (1988) echo the method, not the topic. Both papers beg researchers to stick with single-subject designs. Branch (1977) wants them to study transitions; E et al. want them to rescue obesity treatment.

Frankot et al. (2024) seem to push back, but they don’t. They praise big-data tools yet still slice data to see each individual. Small-N charts and large pools can live side-by-side.

When you write a behavior plan, skip phrases like "he forgot" or "she remembered." Describe the trigger and the change. Graph one client at a time and look for lasting shifts in the curve. If the new level sticks, the history is already in the line—no extra box called memory required.