mlevin@jade.tufts.edu (Mike Levin) (05/01/91)
Does anyone remember (can give me a reference to) a paper that described the following experiment: you take a bunch of mice, run them through a maze, until they learn it, then take the stupidest mice, breed them, and keep doing it over and over again, always selecting for the stupidest mice. I think this particular experiment went on for several years. Anyways, it originally started out as an experiment to show inheritance of acquired characteristics, but what they found was that even though they kept selecting for the stupidest mice, the overall speed of maze running (the mice's intelligence, as measured here) went singificantly *up* over the course of the experiment. Has anyone read this paper? Any criticisms of it? Any ideas on what could be going on in such a case, provided the study was well done? Mike Levin
ronald@uhunix1.uhcc.Hawaii.Edu (Ronald A. Amundson) (05/03/91)
In article <1991May1.154706.16453@athena.mit.edu> mlevin@jade.tufts.edu (Mike Levin) asks: > > Does anyone remember (can give me a reference to) a paper that >described the following experiment: you take a bunch of mice, run them >through a maze, until they learn it, then take the stupidest mice, >breed them, and keep doing it over and over again, always selecting >for the stupidest mice. I think this particular experiment went on for >several years. Anyways, it originally started out as an experiment to >show inheritance of acquired characteristics, but what they found was >that even though they kept selecting for the stupidest mice, the >overall speed of maze running (the mice's intelligence, as measured >here) went singificantly *up* over the course of the experiment. Has >anyone read this paper? Any criticisms of it? Any ideas on what could >be going on in such a case, provided the study was well done? > >Mike Levin Mike -- The experiments you're alluding to were performed by William McDougall and later assisted by J. B. Rhine (yes, the ESP afficionado) at Harvard and later Duke from 1927 to 1938. McDougall and Rhine were Lamarckians of a somewhat spiritualistic bent (McDougall, at least). The experiments involved running successive generations of rats through a simple water maze and recording their success rates. There was no selection for slow learners. McDougall found that as generations went by, the rats learned the maze faster and faster. He took this as a demonstration of Lamarckian inheritance. Some people who tried to reproduce the experiment found high rates of learning right away. On this and other design and experimental bases, almost everyone promptly rejected McDougall's interpretation. The real fun started when Rupert Sheldrake (author of _A New Science of Life_) got hold of the case. Sheldrake has a mystical theory involving a non-physical "formative causation" which is said to be responsible for pretty much everything not-yet-explained. (Crystal formation, e.g.) Sheldrake is fun to read, and the New Age loves him. And he does know his history of biology -- he knows who the spiritualistic biologists were, and reads up on all their material, including the stuff rejected and forgotten by mainstream science. Sheldrake admits that the Lamarckian theory was refuted by the attempts to replicate, but claims that the high rates of learning in the newer experiments showed that "formative causation" magically transmitted McDougall's rats' learning to the replicators' rats' little minds. Spiritualism wins again! The facts are these: McDougall and Rhine's experiments were _incredibly_ poorly designed and controlled. There were massive opportunities for experimenter bias, and McDougall didn't even keep pedigrees of his rat populations. Since it was pretty well established that fecundity goes down over generations in inbred populations like McDougall's (a fact confirmed by those who attempted replication), chances are fairly high that the 1938 30th generation rats were all descended from one 1926 pair. This, among several other blunders, makes McDougall's experiments nigh onto worthless for anyone but a collector of pathological science (a person such as your present correspondent). A good short description of the episode is in Norman Munn, _Handbook of Psychological Research on the Rat_, (1950, Houghton Mifflin, p. 38-39). The gory details can only be gotten from the original articles, referenced in Munn. Sheldrake's version of the experiments is in his book, but it is not to be trusted. Has anyone seen Elvis recently? Ron Amundson Dept. of Philosophy University of Hawaii at Hilo