Anomalies And Experimental Errors
In the Lecture Jason made the claim that it is impossible to determine what is currently an experimental error and what is an Anomaly that our current scientific theory cannot explain.
I do not accept this view. In principle an experiment may be repeated whilst the initial conditions change. That is, if a finding contradicts the current theory, then the experiment may be run again and again to deduce whether the result was likely to be an error or instead is an anomaly which contradicts the theory. This is for example the process in by which the physicists at Cern are speculating over the findings of faster than light neutrinos. If the experiment cannot be repeated with the same results, then an error has occured (assuming that the law being measured is congruent throughout time). If the Physicists at CERN repeated the experiment after fixing any believed errors, and the neutrinos were still detected then a stronger case for an anomaly can be made. So in principle it is possible to determine between an anomaly and a error simply by re-running the experiment, which has nothing to do with our current theory (the theory may be falseified or be upheld by discovering the nature of the error).
Paul Rosenthal
—
Hi Paul, I think your point can be challenged by looking closer at what in fact an experiment is.
Any experiment always has to do something with some theoretical assumptions (that is, the claim that an experiment ‘has nothing to do with our current theory’ is false). When you design an experiment, you have to make sure that it either confirms or falsifies some theoretical hypothesis - otherwise the experiment would be pointless. Therefore, the hypothesis you’re trying to test necessarily shapes the scope and details of the experiment - it dictates what you want to observe and how, what data you want to collect or ignore, and defines the framework within which the data will be interpreted. Moreover, while conducting the experiment and making observations, you rely on various sophisticated tools and equipment, the use of which assumes the correctness of a whole bunch of theories (e.g. optics for using a microscope). Thus, any experimental data is heavily ‘theory-laden’ - the data is not an ‘impartial’ snapshot of the ‘objective’ reality (assuming there is one), but a product of our current science.
This means that whatever experimental results you get, they are utterly contaminated by possible (and likely) errors in our theoretical knowledge, by the narrowness of the experimenter’s perspective, and, in addition, by various social factors. For these reasons, no experimental results can be conclusive, and however many times you try to replicate an experiment, you won’t be able to tell an experimental error from an anomaly.
(Max Fedoseev)
I think I’d lean towards Pauls perspective on this one, although your points are good Max, and true for some anomalies.
It is not impossible to tell the difference between experimental error and theory-crushing anomalies because of the nature of the scientific method; methods and procedures are written down and experiments repeated by different experimenters, with nuanced changes. Feeding into this are the errors recorded by the original experimenter, which can be scrutinised by other scientists, and intentionally tested by repeated, or further experiment.
Of course, it’s not a perfect system, prone to repeated human error, and assumptions of technology as you said, but for all those scientific revolutions caused by long unsolved anomalies, I dare say there are just as many anomalies that crop up and are resolved quickly, before a new theory is set in stone.
Experimental errors can be eliminated by further experiment; anomalies require a shift in thinking.
—
I’m with Max here. I agree that it is true that you can repeat experiments and change conditions to try and remove sources of error, so often what may appear to be anomalies will disappear. However, consider the case where an experiment is repeated many times by different experimenters, under different conditions and the same anomaly persists. It seems there are two choices here - to accept the presence of a truly theory-refuting anomaly and make a large-scale change in thinking; or to assume that there is some way in which our present equipment/understanding just hasn’t allowed this experiment to accurately reflect the world. It appears that over the course of history, there have been numerous examples of both. So how do we, standing here and now, know for certain which is the correct choice in a particular situation? It seems to me that we don’t.
Heather B.
— Making a judgement is a problem persistent with human error (unfortunately or fortunately). But i question is there the possibility that because an experiment is changed that more anomalies come into existence which complicates things even further. Therefore i think that sometimes with should just accept anomalies if they are too readily apparent in the outcome of experiments.
Bernadette
—
Joshua, are you enrolled in this course? I don’t know who you are: you haven’t been to any tutorials or submitted an assignment, or, judging from your comment above, been to the lectures (at least, not the lectures on this topic). You’re very welcome here whether you’re enrolled or not, but you might like to come to the classes on the topics you want to discuss.
Heather, yes, great points, and Bernadette, also great points if I understand you correctly.
Jason
