inferred that other planets, from the same relation, may be inhabited. Or, when it is inferred, from the fact that virtue affects our well-being, that vice must likewise; virtue and vice being both moral habits or dispositions, and the relation being the same-both alike affecting condition.

§ 154. Analogical reasoning is COMPLEX when two different relations are introduced.

Thus it may be argued from the fact that virtue tends to happiness, that vice must tend to misery. In this case, the whole analogical proof rests on the similarity of relation between both virtue and vice, and welfare. This is the generic relation. Another specific relation is introduced as belonging to each of the terms-that of virtue to happiness, and of vice to misery. These are dissimilar relations. It is by another principle of proof that the tendency to affect welfare common to virtue and vice is believed to be in the one case salutary, in the other pernicious. This is an instance of Aristotle's argument from contraries- vaνTIV.

a complex analogical argument, however, it is not necessary that the second relations should be to opposites. As from the relation of a seed to the plant we may argue in respect to the relation of an egg to the fowl. The rela tions of a germ to the parent and to the living product are common to the seed and to the egg. These are the generic relations. The specific relations of the egg to the fowl and of the seed to the plant are dissimilar, but are not proper opposites. The force of the analogy reaches only to the similarity or resemblance of the relations.

§ 155. Examples may be REAL OR INVENTED. Real examples, or such as are taken from actual observation or experience, carry with them their own evidence.

Invented examples must possess intrinsic probability or be credible in themselves; otherwise they evidently can have no weight as arguments.

Aristotle instances as an invented example that employed by Socrates, of the mariners choosing their steersman by lot. The case, probably, never in fact occurred; but it clearly might occur, and it well illustrates the possibility of the lot falling upon an unskillful person; and, therefore, was a valid argument as used by Socrates against the practice, then common, of appointing magistrates by lot.

Dr. Whately has well observed that while a fictitious case which has not this intrinsic probability has absolutely no weight whatever, any matter of fact, on the other hand, however unaccountable it may seem, has some degree of weight in reference to a parallel case. "No satisfactory reason," he proceeds to remark," has yet been assigned for a connection between the absence of upper cutting teeth, or of the presence of horns, and rumination; but the instances. are so numerous and constant of this connection, that no Naturalist would hesitate, if on examination of a new species he found those teeth absent and the head horned, to pronounce the animal a ruminant.”

§ 156. As the points of resemblance between different objects are diverse, and things most unlike may yet have some resemblance to each other, and therefore be embraced under the same class, it becomes important in the use of this kind of argument, on the one hand, carefully to set forth the particular point of resemblance on which the argument rests; and, on the other, in estimating the weight of the argument to reject from the estimate those points in which there is no resemblance

While those arguments which rest on resemblances in objects most unlike are generally in themselves more striking and forcible, they are yet often sophistically invalidated and rejected, because in most respects they are so dissimilar. On the other hand, no sophistry, perhaps, is more common than that of assuming a resemblance in all points where there is such resemblance in many. In the use of this species of argument, it becomes, then, of the utmost importance to bear in mind both that the most similar things differ in some respects, and perhaps in that very point on which the argument in a given case depends; and, also, that the most dissimilar things may have some properties or relations in common, and may therefore furnish foundations for valid reasoning.

The decisive test of the soundness of all arguments founded on resemblance, is furnished in the inquiry: do the particulars of resemblance owe their existence to the same cause; or, where the cause is not known, to the same law? As the whole force of examples as arguments rests on the sameness of the cause, or of the law which has given origin to the resemblances on which the classification depends, the detection of this cause or law, where possible, will ever discover the validity or invalidity of the example as an argument. Just so far as there remains a doubt of the sameness of the cause or law, so far must there be weakness in the argument.

§ 157. While all simple arguments may be referred to some one of the foregoing classes, many complex arguments partake of the nature of two or more; their force in reasoning is consequently modified in reference to the respective character of the classes of arguments of the nature of which they partake.

What is often called a priori reasoning not unfrequently includes in itself not only an antecedent probability argument, but also a sign, or an example. From the falling of the barometer, we infer a priori that there will be a change of the weather; not because we suppose the fall of the mercury to be the cause of the change, but because it is the sign of the existence of the cause. We in this case, in truth, first argue by a sign, to the existence of a cause, and then by an antecedent probability argument, to its effect, viz: a change of the weather. In the argument in "the Goodridge case," before referred to, § 144, several circumstances are advanced as signs in proof of a cause or motive to feign a robbery; from which cause, thus proved, the inference was that the prosecution was groundless.

Lord Chatham in his speech "on removing the troops from Boston," argues the continued and determined resistance of the Americans to an arbitrary system of taxation from the spirit of liberty which animated them in common with all Englishmen; and the existence of this spirit is proved by an example--the proceedings of the General Congress at Philadelphia. This would ordinarily be called an a priori argument, inasmuch as the force of it rests mainly on the existing cause to produce the continued resistance. But an "example," which is of the nature of an a posteriori argument, is introduced to prove the existence of the cause, and the intermediate step of the argument, the cause itself, is not expressed but only implied.

In the same speech we have another form of the combination of the antecedent probability argument with the example. The example is introduced, not as in the other case, to prove the antecedent probability argument itself, but to confirm it as proof of the main proposition. The speaker exemplifies the working of that spirit of liberty in the effec

tual opposition to "loans, benevolences, and ship-money in England," in the procuring of "the bill of rights," &c. The reasoning, as a whole, is a priori; but is complex, consisting of an antecedent probability argument and examples.

By an a priori argument, the fact of a revelation from heaven is inferred from the general corruption of the human race. The argument consists of an antecedent probability argument-the determination of God to do all that is necessary to effect the recovery of the race; and of a sign-the corruption of the race, to prove the necessity of such an interposition by revelation.

A posteriori reasoning, also, often includes arguments of different classes. From the migration of birds to the north, we infer that some of the various effects of spring have appeared in the place of their hibernation. From the migration of birds, as a sign, we infer the return of warm weather as its cause; and from this we infer again, by an antecedent probability argument, the usual effects of the return of spring.

While both a priori and a posteriori reasoning thus often contain arguments of two or more classes, there is yet an obvious distinction between them. In the former, the antecedent probability argument is the one on which the force of the reasoning mainly depends; in the latter, the sign or the example is the prominent argument.

The analysis of complex arguments will often discover the precise amount of validity due to them. It will disclose also the point where the sophistry of a suspected proof en

ters.

Testimony and authority, also, often combine arguments of different species, and are themselves frequently combined together in the same process of reasoning.

What is often called reasoning from experience, is dis