Reply to both Cogito and lmkg since they are similar counterpoints: >Cogito: "maximum number of marshmallows"* >lmkg: "the bigger number signals that one answ

Reply to both Cogito and lmkg since they are similar counterpoints: >Cogito: "maximum number of marshmallows"* >lmkg: "the bigger number signals that one answer is correct," If the child doesn't trust or believe the researcher (a stranger), then in his mind, the 1 marshmallow is a bigger number than zero marshmallows promised later. Because "later" could translate to "never", and "2 marshmallows" could translate to "0 marshmallows". Is it possible for a few, or some of the very young 4-year old children to meta-analyze the situation and conclude "you can't trick me, this is really a test!" ?? Sure, but Smaug123 contended that this reasoning process generalized to the test population. If we want to entertain the fact that children want to "please" adults to "pass a test", we also have to accept that they can interpret the situation as "evaluating promises of future rewards being kept or broken" and rationally choose 1 marshmallow as the optimum decision. Choosing 1 now is also "passing a test" if we insist on framing it as a "test" in the vein of Hercules 12 labors. ("See mom, you told me to never trust strangers and so I got the 1 marshmallow right away.") If we still want to hang on to the idea that 4-year olds keep reinterpreting life interactions as a "test to pass", a lot of frustrated parents are going to wonder... "why can't Johnny see that keeping to himself instead of pulling his sisters hair to get a piece of candy later is a TEST TO PASS?!? He just keeps misbehaving!" <-- If that situation is happening, it means overlaying observed behavior (both wanted & unwanted behaviors) with a framework of "desire to pass a test" is overestimating its explanatory power.