Science

Aristotle's account of science in the Posterior Analytics states the terms in which the tradition afterward considers the question. Scientific knowledge, or episteme, is distinguished both from opinion and from art or skill by the necessity of what it knows. To possess unqualified knowledge of a thing is to know the cause upon which the fact depends, and to know further that the fact could not be otherwise than it is. Knowledge of this sort is acquired by demonstration, and demonstration requires premises that are true, primary, and better known than the conclusions drawn from them. Where the premises are merely probable, the reasoning may be dialectical or rhetorical, but on Aristotle's criteria it is not scientific. This restriction of science to the universal and the necessary raises at once the question how the premises themselves are secured, since they cannot in turn be demonstrated without regress.

The answer is that first principles are reached by induction from experience, and that each science possesses principles of its own which cannot be borrowed from a more general discipline. Physics, mathematics, and the study of living things proceed from different starting points, and the demonstrations proper to one are not transferable to another. It should be observed, however, that Aristotle does not place all the sciences at the same distance from observation. The particular inquiries, such as astronomy or the study of animals, mingle demonstration from principles with the testing of theories against what is seen, and they are accordingly qualified by a tentativeness that the more philosophical disciplines do not exhibit. Aristotle rejects astronomical hypotheses which fail to account for the appearances, and on the generation of bees he holds that credit must be given to observation rather than to theories, and to theories only if what they affirm agrees with the observed facts.

A further requirement attaches to the study of nature, namely that explanation must exhibit the cause. Aristotle distinguishes four senses in which a thing may be said to be caused, and the scientific account of a natural process is incomplete until the relevant causes are given: the matter out of which a thing comes, the form it takes, the agency that brings the change about, and the end toward which the change tends. The last of these, the final cause, has been the most disputed, and the modern reduction of natural explanation to matter and motion turns largely upon its rejection. The several senses of cause are treated more fully under the idea of Cause, and the standing of physics among the speculative disciplines under the idea of Physics.

Aristotle's conception of science as demonstrative knowledge of necessary causes governs the discussion until the seventeenth century, and it is taken over with little alteration by Aquinas and the schools. The objections raised against it are of two sorts. Some are logical, and ask why knowledge should be counted scientific only when it is demonstrative. Others are drawn from the practice of the experimental disciplines, whose conclusions are probable, revisable, and framed as correlations rather than as necessities. Bacon presses the second sort of objection; the first is renewed in different ways by Hume and by Mill.

Aquinas takes over the Aristotelian account of science and applies it to a question Aristotle did not raise, namely whether sacred doctrine can be called a science at all. The difficulty is that the first principles of theology are articles of faith rather than propositions evident to the natural light of intelligence. Aquinas meets it by way of the distinction between sciences which proceed from principles known in themselves and sciences which take their principles from a higher discipline. Music, he observes, accepts its principles from arithmetic and remains a science notwithstanding. Sacred doctrine stands in a like relation to the knowledge which God and the blessed possess of divine things. On this view the theologian does not demonstrate his premises but argues from them, and the resulting body of doctrine has the form, if not the self-sufficiency, of any other science.

What this permits is a mode of procedure rather than a new source of truth. Once the revealed principles are granted, inferences may be drawn from them, objections met, and the whole ordered into a single discipline. Aquinas holds sacred doctrine to be one science, since it considers all its objects under a single formal aspect, that of being divinely revealed; and he holds it to be both speculative and practical, though more speculative than practical, since it treats chiefly of God and only secondarily of human acts. The Summa Theologica exhibits the method throughout, in the movement from objections to a determination and thence to replies.

The division of the speculative disciplines occupies him as well. Natural philosophy considers things that are changeable and material; mathematics considers quantity abstracted from motion and matter; the divine science considers what exists apart from matter altogether. Since the mode of knowing follows the mode of the object, the method proper to one of these is not proper to another, and observation of sensible things, which suffices in natural philosophy, cannot be the way to knowledge of what transcends sense. Whether the sciences should be divided in this manner, by subject and mode of abstraction, or whether a single method extends to every subject, is a question raised again by Descartes and, from the opposite side, by Bacon. The related issues are treated under the ideas of Knowledge and Metaphysics.

The union of Aristotelian science with Christian theology which Aquinas effects governs the curriculum of the medieval universities, and it is against this settlement, quite as much as against Aristotle himself, that the reformers of the seventeenth century direct their criticism. When Bacon complains that the schools multiply disputation in place of inquiry, and when Hume proposes to consign school metaphysics to the flames, the doctrine under attack is in large part the one Aquinas states.

Bacon's proposal for the reform of the sciences is at once a criticism of method and a redefinition of the end which knowledge serves. On the first head he argues that the received sciences have made no progress because their general notions are abstracted too hastily from a narrow experience and are afterward defended by disputation rather than corrected by fresh recourse to nature. On the second he holds that the office of knowledge is not contemplation alone but the relief of man's estate, so that the sciences are to be judged in part by what they enable men to do. It may be noted that this last position is not necessarily opposed to the older view, provided that what Bacon calls the philosophy of nature is understood in the experimental rather than the philosophical sense.

The failure of the received learning he traces to the idols, which are the standing dispositions of the mind to misrepresent what it examines. The idols of the tribe belong to human nature as such, among them the disposition to suppose more order and regularity in things than is actually found. The idols of the cave arise from the temper and education of the individual inquirer. The idols of the market place proceed from words, which are imposed according to the apprehension of the vulgar and thereafter obstruct the understanding. The idols of the theatre are the received systems of philosophy, accepted as if they were stage plays of the inquirer's own devising.

The remedy has a negative and a positive part. The idols are first to be recognized and set aside; then a method of gradual induction, resting upon natural histories collected without prejudice and arranged in tables of presence, absence, and degree, is to replace both the syllogism of the schools and the casual gathering of instances. Bacon further supposes that such inquiry cannot be the work of one man, and he asks for instruments, endowments, and an organized division of labor. Since the knowledge sought is experimental in origin, it is by its nature capable of application, and Bacon accordingly treats mechanics as the operative part of physics and looks for a productive counterpart to every speculative discipline. The New Atlantis presents a society ordered upon that expectation. The relation of science to the useful arts is discussed further under the ideas of Art and Knowledge.

Bacon displaces the authority of Aristotle and puts method in its place, on the supposition that observation conducted under rule will of itself converge upon truth. Whether that supposition is warranted becomes one of the persistent questions of the tradition, and it is raised in its sharpest form by Hume's examination of induction. Descartes agrees that the old learning must be abandoned, but he judges the accumulation of natural histories too weak a foundation and looks instead to principles which the mind can perceive clearly and distinctly. The two programs together set the alternatives within which the seventeenth century conducts its argument about the method of the sciences, an argument pursued further under the ideas of Induction and Hypothesis.

Descartes shares Bacon's dissatisfaction with the sciences he inherits, but he draws the opposite conclusion about the remedy. The defect, in his view, is not that the schools reason too much and observe too little, but that they reason from premises which have never been secured. Science in its entirety, he holds, is true and evident cognition, and it has been the mathematicians alone who have succeeded in producing reasons that are evident and certain. What is wanted, therefore, is not the abandonment of demonstration but its extension: a body of knowledge which begins from principles that cannot be doubted and advances by steps each of which is seen as clearly as the last.

The method of doubt is the instrument by which such principles are sought. Whatever can be called in question is set aside, until there remains the thinker's own existence as a thing that thinks. From this he argues to the existence of God, from the divine veracity to the reliability of what is clearly and distinctly perceived, and from that to the laws of motion and to a philosophy of nature framed in terms of extension and its modes. Philosophy as a whole he compares to a tree whose roots are metaphysics, whose trunk is physics, and whose branches are the several particular sciences.

The unity so obtained differs from Aristotle's arrangement chiefly in the claim that one method suffices for all subjects, whereas Aristotle had divided the sciences by their objects and assigned to each its own principles. Yet Descartes qualifies the claim when he turns to the study of nature. There, he acknowledges, experiments become the more necessary the further one has advanced, since particular effects may be deduced from the principles in several different ways, and the only means of determining which deduction holds is to devise an experiment whose outcome would differ according as one explanation or the other were true. The place of experiment in natural inquiry is treated at greater length under the ideas of Experience and Physics.

The mechanical philosophy which Descartes frames sets the terms of natural inquiry for much of the seventeenth century, and the ambition to make metaphysics a science after the model of mathematics is renewed by Spinoza. The demand for certainty at the outset, however, sits uneasily with the way the natural sciences are in fact conducted, and Newton's synthesis is achieved by propositions gathered from the phenomena rather than by deduction from principles held to be indubitable. Whether the sciences can be founded upon definitions and axioms, or must in the study of nature submit to experiment, remains in dispute, as the chapters on Hypothesis and Induction show.

Newton joins mathematical demonstration to experiment in a manner which later writers take as the pattern of what a natural science ought to be. He calls his undertaking experimental philosophy, and he opens the Optics by saying that his design is not to explain the properties of light by hypotheses but to propose and prove them by reason and experiments. The same temper appears in the Principia, where the laws of motion and the law of gravitation are presented as propositions gathered from the phenomena and afterward extended by mathematical reasoning to bodies at any distance.

The Rules of Reasoning in Philosophy state the conditions under which he is willing to generalize. No more causes are to be admitted than suffice to explain the appearances; effects of the same kind are to be assigned the same causes; qualities found in all bodies within the reach of experiment are to be taken as qualities of bodies universally; and propositions collected by induction are to be held as accurately or very nearly true until other phenomena occur by which they may be corrected or restricted. His refusal of hypotheses is directed not against mathematical reasoning but against conjectures about hidden mechanisms which observation cannot reach. That gravity acts according to the inverse square of the distance he takes the phenomena to establish; what gravity is, whether material or immaterial, he declines to determine.

The result is a body of natural knowledge which is exact in its mathematical form and cautious in its metaphysical commitments. The laws state relations among measurable quantities, and their interpretation in terms of underlying causes is left open. This settlement proved fruitful in practice, yet it leaves unanswered the question of what entitles anyone to pass from observed regularities, however numerous, to a law asserted of all bodies whatsoever. The warrant for such inference is examined under the idea of Induction, and it is the question Hume presses.

In proportion as Newtonian physics becomes the model of what a science is, the word tends to be reserved for disciplines which are experimental, or at least empirical, in their procedure, and the non-experimental disciplines are questioned when they claim the name. Hume and Mill both take the Principia as the standard which the study of man should imitate. The questions Newton leaves open pass into the philosophy of the following century, where they appear as Hume's examination of causal inference and as Kant's account of the conditions under which experience is possible.

Hume's division of the objects of human reason yields a criterion by which the claims of any discipline to be a science may be tried. Relations of ideas, of which the propositions of mathematics are the examples, are discoverable by the operation of thought alone and are demonstratively certain, but they assert nothing about what exists. Matters of fact are established only by experience, and the contrary of any of them remains conceivable, so that conclusions about them are never more than probable. No third sort of knowledge is allowed. Whatever is neither demonstrable in the manner of the first nor testable in the manner of the second is on this account not knowledge at all, and the reasonings of the schools concerning substance, essence, and the causes of things fall under the exclusion.

The concluding passage of the Enquiry states the test in its plainest form. Any volume is to be asked whether it contains abstract reasoning concerning quantity or number, and whether it contains experimental reasoning concerning matter of fact and existence; if it contains neither, it is to be committed to the flames. The rule bears upon much that Aquinas and Descartes had offered as science, and it anticipates the positivist doctrine which later writers state more systematically.

The criticism is accompanied by a positive proposal. Hume undertakes a science of human nature conducted on the model of natural philosophy, which is to say by observation and experiment rather than from principles assumed at the outset. The passions, the association of ideas, the formation of belief, and the conventions of justice and property are treated as phenomena admitting of general laws. Such a science he regards as fundamental to the rest, since every other inquiry is carried on by faculties whose reach and limits it alone can determine. His own account of causal inference belongs to this undertaking: the connection between cause and effect is not perceived but is produced by custom following repeated conjunction, a result examined further under the ideas of Cause and Induction.

A difficulty which Hume himself raises remains after his criterion has been applied. If all knowledge of matters of fact rests upon inference from experience, and if that inference presupposes a uniformity in nature which experience cannot establish without arguing in a circle, then the natural sciences appear unable to satisfy the standard proposed for them. Kant, who reports that this problem interrupted his dogmatic slumber, argues that the principle of causality is neither a habit acquired from observation nor an assumption without warrant, but a condition without which experience of an ordered world would not be possible. Mill takes the other course, holding that the uniformity of nature is itself the widest of inductions.

Kant tries the claims of any discipline to be a science by the marks of necessity and universality. Generalizations drawn from observation, however often confirmed, remain contingent, since what has always happened might yet happen otherwise. The propositions of mathematics are necessary, but Kant holds them to be synthetic as well, in that they extend knowledge beyond what analysis of the concepts would yield. Knowledge properly called scientific must possess both characters at once, being necessary and yet informative about objects, and the question which the Critique of Pure Reason sets itself is how judgments of this kind, which he names synthetic a priori, are possible.

The answer offered to Hume is that the understanding supplies the form under which anything can be an object of experience at all, and does not simply register such regularities as observation happens to present. The categories, of which causality is one, are not abstracted from what is given but are the conditions of a unified experience, and nature, considered as the sum of appearances subject to law, must conform to them. On this view the laws of nature in their most general shape are not read off from the phenomena but prescribed to them, while the particular laws are still to be discovered by observation and experiment.

Kant compares the change he proposes to the one Copernicus introduced in astronomy: instead of supposing that cognition must conform to objects, it is supposed that objects, as known, conform to the constitution of our cognition. Necessity is thereby secured for the principles of natural science, at the price of confining knowledge to the field of possible experience. Of things as they are in themselves, apart from the forms of sensibility and the categories, no science is possible. Kant adds a further restriction in the Metaphysical Foundations of Natural Science, where he holds that a doctrine of nature contains science properly so called only in the measure that mathematics can be applied within it, a criterion which places physics on one side and the descriptive study of living things on the other. The bearing of these doctrines on the older speculative disciplines is treated under the idea of Metaphysics.

Two lines of subsequent thought take their departure from this account. Hegel retains the view that the categories constitute their object but sets them in motion, treating them as moments of a development rather than as a fixed table. Mill rejects the a priori altogether and undertakes to show that even the axioms of geometry and the principle of causation are generalizations from experience. The question which both must answer, and which the chapter on Knowledge pursues further, is in what sense the conclusions of science hold of things and not merely of appearances.

Mill states the empiricist position in its most extended nineteenth century form. Against Kant he denies that any knowledge whatever is a priori, holding that the axioms of geometry and even the principles of logic are generalizations from experience, distinguished from other generalizations by the number and uniformity of the instances rather than by their source. Against the older rationalism he maintains that the procedures of the sciences are continuous with the reasoning men employ in ordinary affairs, differing in rigor and in self-consciousness but not in kind.

The System of Logic accordingly treats induction as the whole of inference concerning matters of fact. The five canons, of agreement, of difference, of the joint method, of residues, and of concomitant variations, set out in explicit form the ways in which an antecedent may be shown to be the cause of a consequent. They are offered less as new instruments than as the articulation of what careful inquirers have always done, made explicit so that inferences may be criticized. The syllogism, which the older tradition placed at the center of scientific demonstration, Mill reinterprets: the general proposition is a record of observed particulars, and the real inference runs from particulars to particulars, the major premise serving as a memorandum of the warrant. Whether induction rests upon a principle of the uniformity of nature which is itself inductively established remains, on his own account, a question of some difficulty.

The sixth book extends the same logic to the study of man and society. Mill holds the moral sciences to be inexact rather than unscientific, comparable in this respect to the study of the tides or of the weather, where the general laws are known but the circumstances are too many and too variable for exact prediction of particular events. He asks for a science of the formation of character, which he calls ethology, to stand between the laws of individual psychology and the generalizations of sociology, and he considers whether the study of society should proceed directly by observation or by deduction from the laws of human nature verified against history. The bearing of these questions upon human freedom is discussed under the ideas of Man and Necessity.

Mill's proposal to bring human character under inductive law raises a difficulty which he acknowledges without wholly disposing of it. The individual whose liberty is defended in the essay on that subject is, in the System of Logic, formed by circumstances according to laws which would render his conduct predictable in principle were the circumstances fully known. Mill answers that a character formed by circumstances may include the desire to form itself, so that determination by causes does not amount to compulsion. Freud takes up the same problem in a clinical setting, treating the mental life as a natural order subject to investigation while addressing the patient as capable of change. The general question is treated under the idea of Necessity and Contingency.

The last of the New Introductory Lectures turns to the question of a Weltanschauung, by which Freud understands a comprehensive view of the world constructed upon a single hypothesis. Science, he holds, offers the only one that can be maintained, since there is no source of knowledge of the universe other than the intellectual manipulation of carefully verified observations. He states the implications without qualification. It is inadmissible, on his view, to hold that science, religion, and philosophy are separate fields of intellectual activity with equal claims to truth, or that a man is free to choose among them; scientific research looks upon the whole field of human activity as its own and must be uncompromising toward any other power that would claim a part of it.

Religion he regards as the only serious opponent, since it exerts its power over the strongest emotions of men, whereas philosophy has little influence upon the great majority. The two might be reconciled if religion would confine itself to what consoles and ennobles and leave the question of fact to science, but this Freud thinks it cannot do without forfeiting its hold. Philosophy is not in his view necessarily opposed to science, and at times it behaves as if it were one and employs some of the same methods; the objection to it is that it clings to the expectation of a complete and coherent picture of the universe, which must fall to pieces with every advance in knowledge.

Something should be added about his conception of scientific procedure, since it bears on the older question whether a science begins from definitions. Freud denies that it does. No science, not even the most exact, starts from clear and sharply defined basal concepts; inquiry begins rather in describing phenomena and then grouping, classifying, and correlating them, with the aid of abstract ideas which at first possess some measure of uncertainty. Definitions come afterward, and even then must retain a certain elasticity, as the history of physics shows. On the arrangement of the disciplines he is equally brief: there are strictly only two sciences, psychology pure and applied, and natural science, sociology being nothing other than applied psychology. Whether psychology can be a natural science, and where its limits lie, is discussed under the ideas of Mind and Soul.

Freud's statement gives one of the plainest expressions to the attitude which converges from many sources under the name of positivism, and which identifies science with knowledge of fact obtained and verified empirically. William James, writing at nearly the same time, draws the line between science and philosophy in a similar place while allowing metaphysics a task of its own. The criterion is afterward systematized by the logical positivists, and afterward disputed by writers who hold that observation is itself directed by theory. Whether philosophy and religion are thereby excluded from knowledge, or only distinguished from one kind of it, is a question pursued under the ideas of Philosophy, Religion, and Truth.