tion, to its extension over the entire class. The experiments of chemistry are of this nature. When the cause of the phenomena in one portion of air, or one globule of water-in one piece of a metal, an earth, or a salt-is discovered, it is at once extended, by an inductive argument, to all other individuals of the class. When Newton had analysed one ray of light into the prismatic colours, he concluded that the same analysis would apply to all other rays of light. When the cause of one glacier is discovered, the origin of other glaciers can be no longer doubtful. If the cause of a single hail-storm, of a single volcanic eruption, of a single earthquake, of a single aërolite, could be discovered with certainty, the explanation would probably apply to all the other phenomena of the same class.

Even in physics, however, great natural sagacity, combined with a practised scientific mind, is often necessary for perceiving the extent and limits of the class to which the generalization is to apply, and for avoiding inductive extensions founded upon fanciful and casual analogies. The early stages of every science are marked by rash and unfounded generalizations, and by attempts to extend principles of causation into regions with which they have no real affinity. Thus, Aristotle adopts an opinion, formed by scientific inquirers before his time, that animals which have not a gall-bladder are long-lived; and that there is a causal relation between the gall and shortness of life.()

Lord Bacon, in his Historia Vitæ et Mortis, lays it down, that persons of fair complexion are short-lived, and that those of black or red complexion are long-lived; that hard and bristly hair is a sign of long life, soft and curling hair a sign of short life; that a small head, wide nostrils, a large mouth, a gristly ear, and strong teeth, are signs of long life, with many other

(2) See De Part. Anim. iv. 2. Compare Anal. Prior. ii. 25, § 4; Anal. Post. ii. 14, § 11. The short explanation of induction given by Aristotle, in the Topics, shows clearly that the sense which he attributed to it agrees with that in which it is now generally understood. • Induction (he says) is an advance from singulars to generals; thus, if a skilful steersman is the best, and also a skilful charioteer; then, generally, he who is skilful in each department is the best.-i. 10. Compare Barth. SaintHilaire, Logique d'Aristote, tom. ii. p. 34. With respect to the sufficiency of single instances in physics, see above, ch. ix. § 17; ch. xi. § 2.

Now,

similar signs, indicating or contra-indicating longevity.(3) if these were really signs of long and short life respectively, they must be connected with it, either as cause or effect, or as inseparable concomitants. But although cases may easily be found in which persons of fair complexion are short-lived, and persons of dark complexion are long-lived, yet the physiologist would not generalize those instances, without first instituting a wide observation of facts; and when that observation had been made, it would probably appear that the colour of the complexion was wholly unconnected with longevity.

It would be useless to repeat any of the numerous instances of hasty generalizations in physics which have been collected by Dr. Whewell, and other writers on inductive reasoning in the physical sciences. The belief that the moon influences the weather, because it influences the tides, may serve to exemplify the class of inductive errors to which we allude.

In politics, the inductive argument by which the efficacy of a cause in one instance is extended to other instances, is sometimes nearly as simple as in those physical sciences, where one instance serves as a perfect specimen of the entire class. In general, however, this extension cannot be made without much preliminary inquiry, and many restrictions and qualifications. One man is not as exactly similar to another man, one race of men is not as exactly similar to another race of men, one political community is not as exactly similar to another political community, as one piece of platinum is to another piece of platinum, or as one vial of oxygen is to another vial of oxygen. Hence, when, in politics, a principle of causation is extended by an inductive argument from a single case, or a few cases, to a large class, it is necessary to avoid the danger of carrying it too far, and of including cases to which it is not properly applicable.”

As a simple example of the process by which an inductive conclusion is confirmed and established, we will take the physiological law, that the life of man never exceeds a certain limited

(3) Works, vol. x. p. 172-3.

period. In the first place, we know that, since the commencement of history, all men have died within that period; we know that no man is now living who was born more than a certain number of years ago; we further know, that the physiological structure of the living generation of men is similar to that of the previous generations who died within the fixed period; and that this structure is so constituted as to be subject to disturbance, decay, and disorganization. We observe, likewise, that living men are subject to the same diseases and bodily injuries which proved ultimately fatal to their predecessors; and that many living persons have, in consequence of diseases and wounds, been in danger of death. These physical arguments are strengthened by the considerations derived from the structure of human society, and from the principle of population. If the present generation were to be immortal, or if the period of their life were considerably enlarged, while their powers of increase by propagation remained unchanged, the conditions of society would be deranged, by the introduction of a new element in the ratio between population and food. All these arguments, again, might be confirmed by the analogy of the mammalia and the other races of animals. Putting together all these considerations, we draw, without the smallest doubt, the general conclusion, that every living man will die within a certain number of years from his birth; and looking to the general permanence of the order of nature, in all matters of animal physiology, we are equally certain that all future men, yet unborn, will hold their life upon similar conditions. Hence we perceive that, in applying to the mortality of man the universal hypothetical premise, that if some individuals of a class possess a certain property, all the individuals of that class possess the property,' we proceed upon sure grounds; and that the inference, in this case, is fortified by so many considerations as to be placed out of all doubt.

On the other hand, many cases will readily occur, where, if we inferred that a property possessed by some individuals of a class was common to the whole class, without inquiring into the

nature of the property, or the grounds of the generalization, we should commit grievous errors. Such an inference is only rendered safe by a preparatory and collateral process of inquiry. The generality and vagueness of the hypothetical premise must be corrected and limited by the considerations specially affecting each case.

§ 2 Having made these introductory remarks, let us now consider how far universal principles of causation, and, generally, how far propositions of universal application, or theories of scientific universality, can be established in politics.

I. Propositions universally true can be laid down in positive politics. Taking human nature as it is, it can be affirmed that there are certain conditions with which every government, all laws, all legal rights and obligations, and various other subjects of jurisprudence, must comply. These may be defined in abstract terms, and such definitions, if correctly framed, are universally true. They are analogous to the descriptions of the human body, and of its functions, given by the physiologist. Positive politics are the descriptive anatomy and physiology of the state. They describe the state as it is necessarily formed and constituted, and the necessary mode of its action. Like the body natural, the body politic has a certain invariable form, and certain members furnished with certain organs; and these have certain functions, exhibiting themselves in certain results. Thus, Locke, remarks that the following propositions, Where there is no property, there is no injustice,' and 'no government allows absolute liberty,' are as certain as any propositions in the mathematics.(*) If, by property,' we mean a legal right to certain external objects; and if, by injustice,' we mean the violation of such a legal right, then it is clear that injustice necessarily implies the existence of property. In like manner, if by 'absolute liberty' is meant an exemption from all legal restraint, it is clear that absolute liberty is incompatible with the existence of political government.

(4) Essay on the Human Understanding, b. iv. c. 3, § 18.

?

୮

The principal business of positive politics is to deal with necessary co-existences, and not to consider necessary sequence or causation. It decides what are the circumstances which constitute a supreme government, a law, a right, an obligation, a legal status, an administrative or judicial act, a treaty, an armistice, a blockade, a colony, a dependency; what is the distinction between freedom and slavery, between a state of war and a state of peace, and the like. It may, indeed, consider what is the general origin or cause of the existence of government, or how any class of rights or obligations, or any political institution, originates or ceases, such origin or cessation being considered as common to all nations and systems of law. But it does not teach what are the tendencies of this or that form of government or class of laws. Positive politics, therefore, is only partially concerned with causation. Owing to this limitation of its province, it has a certainty, precision, and generality, which no department of politics, not purely descriptive, can attain. notes whatever is common to all forms of government and all systems of law, and describes their essential characteristics. Whether a government produce good or bad effects, whether a law exist in Europe, Asia, or America, it must, in order to be a government or a law, possess certain properties; and these properties it is the province of positive politics to ascertain.(5)

It

Every science of an art, every system of theoretic principles intended to serve as the substratum of a set of practical rules, involves a series of descriptive propositions similar to those of positive politics, and possessing equal generality. For example,

(5) See above, ch. iii. § 3; ch. v. § 7. Mr. Bentham lays it down, that if there are any books which can, properly speaking, be styled books of universal jurisprudence, they must be looked for within very narrow limits. . . . . To be susceptible of an universal application, all that a book of the expository kind can have to treat of is the import of words; to be, strictly speaking, universal, it must confine itself to terminology.'Introduction to the Principles of Morals and Legislation, vol. ii. p. 259; ed. 1823. Mr. Stewart (First Diss. p. 92) observes, on this passage, that Mr. Bentham goes too far in limiting universal jurisprudence to the import of words. His proposition, however, may be considered true, if understood in the sense indicated in the text.