Wednesday, December 1, 2010

More absolutes

Yesterday I wrote on the significance of relative and absolute measurements, concluding that relative measurements were in some sense less arbitrary than absolute ones because they do not depend on the definition of a physical constant. I claimed that this dependence is the practical problem with absolute measurements.

Further reflection has led me to believe that this is not the practical problem that has presented itself many times over during the course of my studies. Rather, the practical problem is that an absolute measurement is incredibly sensitive to the manner in which it is performed. Many parameters that are found in physical theories simply do not take account of the limitations in a measurement, such as integration times and nonlinearities in measuring devices. It is extremely difficult to extract a number from a measurement performed under realistic constraints.

Fortunately, any condition that ruins the agreement between a parameter obtained from a measurement and one obtained from calculation will be present amongst many different measurements. The effects of these conditions will essentially "divide out" under comparison, leaving only the signature of any independent variables that were changed between the measurements. Therein lies the strength of conclusions drawn from relative measurements.