Comments on: Rats check their own knowledge before taking a test

By: Erik

Erik — Tue, 03 Feb 2009 15:56:00 +0000

I think the interpretation of this experiment is wishful on the part of the scientists. It seems far likelier that the rats have learned to classify sounds into short, medium, and long, where short is mapped to right door + left lever, long is mapped to right door + right lever, and medium is mapped to left door.
.. and I can’t prove that my position is right either. There is no unambiguous conclusion that can be drawn from this experiment.

By: wesele

wesele — Wed, 29 Oct 2008 17:26:40 +0000

By assuming that the rats are evaluating difficulty of discrimination, the authors bias their conclusions. Metacognition and the assumption that the rats are evaluating difficulty of discrimination aren’t necessary to explain the results of the experiment.

By: Ian

Ian — Tue, 28 Oct 2008 00:27:37 +0000

“It would certainly seem to have survival benefits to an individual if they could recognize when they didn’t know something.”
Clearly that’s something the creationists haven’t quite grasped yet!

By: Capa Dost

Capa Dost — Mon, 27 Oct 2008 00:53:07 +0000

From the paper’s Summary:”If rats possess knowledge regarding whether they know the answer to the test, they would be expected to decline most frequently on difficult tests and show lowest accuracy on difficult tests that cannot be declined.”
The same thing would happen if the rats learned that hearing an intermediate length sound and sticking their noses in the Accept hole sometimes resulted in not getting food, and hearing the same sound and sticking their noses in the Decline hole always resulted in food. That doesn’t require the rats to engage in metacognition. They may just be declining the test more often in situations where accepting the test increases the risk of not getting food. In other words, they may be evaluating the external situation to determine risk, rather evaluating their internal state to determine knowledge. Possibly what this experiment demontrates is not metacognition, but mixing risky choices with a potentially higher pay off with non-risky choices with a guaranteed low pay off. That would seem to increase the chances of survival, as long as they weren’t risking their lives with the risky choices.
From the paper’s Discussion section: “The rate of declining to take the test increased as the difficulty of the discrimination increased.”
The authors have to assume that the rats are evaluating the difficulty of discrimination in order to justify their conclusions. But if the rats are evaluating risk, where long and short durations are less risky and intermediate durations are riskier, then the authors’ conclusion about metacognition doesn’t follow.
If the rats are engaging in metacognition, then they are doing it as part of risk evaluation (evaluating the risk of not getting food). If they are not engaging in metacognition, then they are still engaging in risk evaluation. By assuming that the rats are evaluating difficulty of discrimination, the authors bias their conclusions. Metacognition and the assumption that the rats are evaluating difficulty of discrimination aren’t necessary to explain the results of the experiment. The simplest explanation is that the rats have evolved the ability to evaluate which behavior is most likely to result in food, along with a tendency to sometimes retest behavior that sometimes results in food.

By: Paul Sunstone

Paul Sunstone — Sun, 26 Oct 2008 14:26:03 +0000

It would certainly seem to have survival benefits to an individual if they could recognize when they didn’t know something.