Page:Popular Science Monthly Volume 29.djvu/556

540 of certain stars at the birth of a child is, by the superstitious, said to bode good or evil. If a new piece of work be commenced while the moon is on the wane, or on a Friday, the undertaking is doomed to fail. The belief in good and evil omens has survived thousands of years, and has come down to the present day; in fact, the influence which this belief has on the mind can only be shaken off by calm reasoning and self-training. Many other instances of superstition, still in vogue in our enlightened times, might readily be given.

All of these are false conclusions derived in the same manner: post hoc, ergo propter hoc (after it, therefore because of it). A careful comparison of different cases is not undertaken, no close scrutiny or investigation is attempted, no distinction made between essential and non-essential conditions. In each case a general assertion is based on a few separate, consecutive facts; the relation between cause and effect can not be proved in any instance. In fact, if we except the example of quack-medicines cited, in all other cases, even the most vivid imagination will fail to cast a bridge—be it ever so frail—over the chasm that separates what has preceded from the seeming effect. In short, many prophecies that can be found and met with every day among the people, in newspaper advertisements, etc., are replete with error, and wholly unreliable. It is, then, not surprising that one comes to regard all predictions skeptically; in fact, one is entirely justified in looking upon at least nine tenths of them with suspicion.

The true observer will not rest content with the mere word "experiment," a term so universally used. If one comes to look into matters closely, it will nearly always be found to refer to mere enumeration. Rarely has a word been more misused than this one, "experiment." Science has found a more adequate expression, and terms it "induction." Induction is the means of discovering and proving general propositions. This simple definition should be remembered.

The best-known form of induction is Bacon's method by simple enumeration. Can this method be successfully applied to formulate predictions? Many scholars consider this way of going to work entirely useless for the ascertaining of truths. "Of what use can it be," they say, "to know that a certain phenomenon has taken place a hundred times? Does that afford any guarantee that it must take place again? Or, even granting that it may happen once more, can not the time come when it will not occur?"

Hence, induction by simple enumeration does not seem to be adapted to the finding of general truths, such as science demands, and in consequence does not seem serviceable as a means of securing definite predictions. In fact, induction applied without the necessary caution is the most crude and deceptive means of arriving at general truths, and gives rise to innumerable false conclusions; and yet we owe to this inadequate method some important empirical generalizations.