When is a law probabilistic?

Daniel Gile

November 27, 2005


Probabilistic laws lie somewhere between determinism, where a phenomenon is certain, and total uncertainty, where one has no idea as to whether the phenomenon will materialize and how. Probabilistic laws attempt to quantify uncertainty. For instance, if it is known that the mean height of males in a population is 180 cm, it is possible to quantify the probability that in a random sample of 30 males, the mean height will be between 170 cm and 190 cm.

            Uncertainty in empirical observation of phenomena may be due to the probabilistic nature of a phenomenon (i.e. variation is an intrinsic characteristic of the phenomenon). It may also be due to environmental reasons: the underlying entity itself may be regular in its nature and form, but its manifestations are influenced by external factors which are themselves probabilistic in nature or too complicated in their interactions with the phenomenon at hand to be predictable. Finally, uncertainty may be due to limitations or weaknesses in the detection and measurement of the phenomenon by man.

            In many cases, researchers investigate the existence of a trend, not its quantitative contours. For instance, in TS, the explicitation hypothesis assumes the existence of a tendency, not how strong it is or to what extent a TT will be more explicit than the ST. In such a case, researchers are interested not in quantitative variation, but in its occurrence as such.

            If such occurrence is irregular, this does not necessarily mean that the law is not true, or that it is probabilistic in nature. Other factors may have prevented it from being manifest. Lunar eclipses do not become probabilistic just because cloudy skies sometimes prevent people from seeing them. The trend to make TTs more explicit than STs may be universal without necessarily being manifest everywhere, for instance if cultural conventions or the client’s brief or time pressure etc. inhibit its expression in a particular set of translations.

             Assuming that a phenomenon is probabilistic or “conditional” just because in empirical studies, it is found not to occur regularly or is found to occur only under certain conditions is the same as assuming that lunar eclipses are probabilistic or conditional upon the absence of clouds in the sky. When irregularities occur, the natural procedure in research is to try to find what causes this irregularity, starting with the removal of environmental factors which may interfere, for instance through experimental research. Only when it is thought that all interfering environmental factors and observer-related factors have been removed and there is still irregularity in the occurrence of the phenomenon will one conclude that this irregularity suggests that the law is probabilistic.