[psych, video] Much Cognition, So Wow
I am not quite sure what this shows, but I am fascinated by it. (Note, almost entirely visual.)
(If anybody would like to contribute a description of the video for the visually impaired in the comments, that would be great!)
Things I find notable:
Object rotational exercises have a history in IQ tests.
Right before figuring the solution, the dog has a moment where the log winds up momentarily on the diagonal, which may have given him the idea. But then when he implements the solution, he turns the log the other way.
That's notable because it's not the "reinforced" behavior. At the very least it's a geometric permutation on it, which is not, to my knowledge, accounted for in the woefully misnamed "learning theory" of the behaviorists.
He backs up, thinks about it a moment, and then just implements the right solution. There's no gradual approximation of the right answer. He does the same wrong thing a bunch of times, and then does something radically different which is correct.
So this really looks like what we mean by "thinking", colloquially, but very much not what behaviorists mean by "learning". Or at least it looks very much not like what behaviorists mean by learning; maybe they have an explanation for it that is congruent with their models.
It doesn't even look like Bandura's social learning. The dog is not, as far as we're shown, emulating someone else.
It looks very much like what behaviorism has been trying to explain away for about 100 years.
Good doggie! Whoda supa-smart doggie?! Whoda supa-smart doggie?! You are, yes, you are! You got that stick through there all by yourself, yes, you did and falsified Skinner too! Good boy! You are such a good doggie!!
Click to view(If anybody would like to contribute a description of the video for the visually impaired in the comments, that would be great!)
Things I find notable:
Object rotational exercises have a history in IQ tests.
Right before figuring the solution, the dog has a moment where the log winds up momentarily on the diagonal, which may have given him the idea. But then when he implements the solution, he turns the log the other way.
That's notable because it's not the "reinforced" behavior. At the very least it's a geometric permutation on it, which is not, to my knowledge, accounted for in the woefully misnamed "learning theory" of the behaviorists.
He backs up, thinks about it a moment, and then just implements the right solution. There's no gradual approximation of the right answer. He does the same wrong thing a bunch of times, and then does something radically different which is correct.
So this really looks like what we mean by "thinking", colloquially, but very much not what behaviorists mean by "learning". Or at least it looks very much not like what behaviorists mean by learning; maybe they have an explanation for it that is congruent with their models.
It doesn't even look like Bandura's social learning. The dog is not, as far as we're shown, emulating someone else.
It looks very much like what behaviorism has been trying to explain away for about 100 years.
Good doggie! Whoda supa-smart doggie?! Whoda supa-smart doggie?! You are, yes, you are! You got that stick through there all by yourself, yes, you did and falsified Skinner too! Good boy! You are such a good doggie!!