"Science" is merely a word of Latin derivation meaning "knowledge." If we except mathematics, which is not so much a science in itself as a mode of measurement employed in all the sciences, there are no exact sciences. The more definitely measurement can be employed, the more exact a science becomes. Hence, astronomy and physics may be considered as reasonably exact sciences, though even here when we approach infinite magnitude, as of distance in astronomy, or infinite smallness, as of electrons in physics, and our measuring devices are not sufficiently acute, we discover a wide margin of inexactitude. Sciences which relate wholly or in part to human nature are considered the least exact. History and biography may be exact as to dates, but in so far as they deal with human motives, the "why" of historical and personal events, they can never hope to be absolutely correct. Genealogy, as one of the sciences in which human nature is a factor, is considered to be one of the less exact sciences. As practiced by many of its devotees, it is certainly one of the least exact. Yet it is entitled to rank higher, provided only that proper scientific methods be pursued. The real object of genealogy is to establish lines of descent of human beings. Whether the motives of the inquirer be to make a study of heredity, or to join a certain society by proving descent from a qualifying ancestor, or mere curiosity to learn the identity of one''s forebears, the line of descent is the essential thing. All else is incidental. Among these incidentals are dates. These are important for purposes of identification of ancestors; they are the measuring device which helps to make genealogy an exact science. No one who lacks a mathematical mind can hope to become a genealogist of the very first rank, for it is necessary to deal with dates constantly. The dates in themselves may not be utterly exact. The family Bible may differ a day or two from the town record of birth; it may even differ by an exact year. The date of death may not be precisely known, except that it falls between the making and proving of the man''s will. Yet the dates, if ascertained and copied with meticulous care, are usually exact enough for the larger purpose of identification of persons. Biologically, the genealogist is concerned with proving a line of descent; which means, proving the parentage of one individual at a time, then the parentage of his parents, and so on, step by step. How exact is this process? We may as well concede, at the start, that the paternity of every child in a human pedigree is a matter of faith, or belief, not of proved fact. Although the present writer, like most genealogists, has excellent reasons for the assumption that an extremely high percentage of children were actually the offspring of their reputed parents, it is hardly necessary to call attention to the fact that a single infidelity on the part of an ancestress would be sufficient to invalidate the paternal ancestry back of that generation.
Hence, biologically considered, it must be granted that genealogy is not as exact a science as could be desired, since an entirely unknown margin of error always exists, at least as a theoretical possibility. The genealogist has no means of going behind the official records. The pity is, that he does not more consistently pursue the policy of depending on the official records for his conclusions. Every science must admit the possibility of a margin of error. But in most of the sciences, conclusions are reached only after the collection of all facts which might affect the matter, and after experimentation; which, in genealogy, means the setting up of hypotheses, the testing of these hypotheses by known facts, and the successful elimination of all but one hypothesis, which is then accepted as the only one which fits and explains the facts. When these scientific methods are employed, by a genealogist of sufficient knowledge and training, genealogy becomes a reasonexact science. Let us consider an example of scientific methods, to illustrate how they work. Peter Gubbins appeared, let us say, in the town of Straitsville, where his children were born between 1800 and 1820. The line has been traced back to this Peter, and his origin is sought. Using the splendid facilities that are now available to the genealogist in many of the older sections of the country, it is found that a Peter, son of John Gubbins, was born in 1775 in Freetown, some fifty miles away. The dates fit, but the identity of the two Peters is a mere assumption or guess, if we stop here. We therefore collect every atom of evidence concerning Peter which is available in the records of Freetown, or at the county seat, or at the State Library. We find that his father John died in 1798, leaving a will in which he gave specified realty to each of his sons, including Peter.