Jason Mitchell, a Harvard social neuroscientist, gives an argument against scientific replication, and a defense of unreplicated science, here. His argument, in a nutshell, is that all we ever learn from the failure to replicate an experiment is that the attempted replicator is a lousy experimenter. Quoting Mitchell:
- Recent hand-wringing over failed replications in social psychology is largely pointless, because unsuccessful experiments have no meaningful scientific value.
- Because experiments can be undermined by a vast number of practical mistakes, the likeliest explanation for any failed replication will always be that the replicator bungled something along the way. Unless direct replications are conducted by flawless experimenters, nothing interesting can be learned from them.
Mitchell is giving, pretty much out of a bottle, what Harry Collins and Trevor Pinch call "The Golden hands" argument. "If you can't replicate what I did in the lab, that shows I have golden hands and you have iron claws." Interestingly, C&P generally believe that this is always a rationally defensible claim that any scientist can make. Of course, unlike Mitchell, C&P understand that was is sauce for the goose is sauce for the gander. If the above is, as both M and C&P maitain, always a defensible claims, then so is its converse: "If I got the opposite result in the lab and you can't replicate it, then that shows I have golden hands and you have iron claws."
The problem is, if all the above is true, then this would vindicate C&P's claim that all scientific reasoning is circular. This is, after all, and for example, what they think the famous case of Pasteur and Pouchet and spontaneous generation shows. Pasteur can always claim that Pouchet has iron hands, and Pouchet can always claim that Pasteur has iron hands, and this explains, from each one's point of view, why the other is getting results that don't accord with the first one's theoretical views.
C&P are absolutely right to think that it follows from this that all scientific reasoning is circular. It becomes a matter of definition, on this conception, that you have iron hands just in case you get results that don't accord with my theoretical views. But scientific reasoning isn't circular. And hence there has to be something wrong with this conception. And what's wrong is that pace C&P, there are often, in the long run (and hopefully not such a long run that we are all dead, thought that may have been the case in the Spontaneous Generation case) independent means of deciding whose experiment is the valid one. And that means that pace Mitchell, scientific replication, and the process of figuring out who has the iron claws and who has the golden hands, is a crucial part of the scientific enterprise.
Recent Comments