Austin, Lectures on Jurisprudence, 5th ed., vol. 2, essays on interpretation and analogy

The following Essays or Notes are referred to in the thirty-third Lecture (p. 577, ante), in which the same subjects are more succinctly handled. They were not found with the Lectures, but were doubtless intended by the Author to be incorporated in the great work which he meditated.

In a note in the page above-mentioned, I spoke of the ‘Essay on Interpretation’ as complete. This, unfortunately, is a mistake; nor have I been able to find any trace of the conclusion.

The original of the ‘Excursus on Analogy’ consists, in great parts of unarranged and almost illegible fragments, amongst which it was extremely difficult to establish anything like order and coherence. I hesitated for some time whether to submit what the Author left in so imperfect a state to the public eye. Nor should I have ventured to do so, had I not been encouraged by the opinion of several persons of high authority in such a matter, to whom it has been submitted. They have exhorted me by no means to suppress this essay; ‘Since,’ to use the words of one of those most qualified to decide, ‘though, from the fragmentary form in which it must necessarily appear, its excellences will probably be hidden from most readers, its great philosophical value will be apparent to those who study it with attention.’

IN my thirty-third Lecture I have tried to contrast Interpretation (in the proper acceptation of the term) and the induction of a rule of law from a judicial decision. In (the earlier portion of) the present note, I shall try to distinguish Interpretation (in the proper acceptation of the term) from the various modes of judicial legislation to which the name of interpretation is not unfrequently misapplied.

The discovery of the law which the lawgiver intended to establish, is the object of genuine interpretation: or (changing the phrase), its object is the discovery of the intention with which he constructed the statute, or of the sense which he attached to the words wherein the statute is expressed. For the reasons which I have given in the text, the literal meaning of 990the words wherein the statute is expressed, is the primary index or clue to the intention or sense of its author.

Now the literal meaning of words (or the grammatical meaning of words) is the meaning which custom has annexed to them. It is the meaning attached to them commonly by all or most of the persons who use habitually the given language: Or, if the words be technical, it is the meaning attached to them commonly by all or most of the persons who are specially conversant or occupied with the given art (or science). Generally, the customary meaning of the words wherein the statute is expressed, is obvious or easily assignable; and generally, therefore, the interpreter assumes it tacitly, and without hesitation and inquiry. But, occasionally, the customary meaning of the worth is indeterminate and dubious. What is the meaning which custom has annexed to the words, is, therefore, an inquiry which the interpreter may be called upon to institute. Consequently, the interpretation of a statute by the literal meaning of the words may possibly consist of a twofold process: namely, an inquiry after the meaning which custom has annexed to the words, and a use of that literal meaning as a clue to the sense of the legislature. The interpreter seeking the meaning annexed to the words by custom, may not be able to determine it; or he may not be able to find in it, when he has determined or assumed it, any determinate sense that the legislature may have attached to them: And, on either of these suppositions, he may seek in other indicia, the intention which the legislature held. Or, when he has determined or assumed the customary meaning of the words, the interpreter may be able to discover in their customary or literal meaning, a determinate or definite intention that the legislature may have entertained: And, on this supposition, he ought to presume strongly that the possible intention which he finds is the very intention or purpose with which the statute was made.

The intention, however, of the legislature, as shewn by that literal meaning, may differ from the intention of the legislature, as shewn by other indicia; and the presumption in favour of the intention which that literal meaning suggests, may be fainter than the evidence for the intention which other indicia point at. On which supposition, the last of these possible intentions ought to be taken by the interpreter, as and for the intention which the legislature actually held. For the literal meaning of the words, though it offers a strong presumption, is not conclusive of the purpose with which the statute was made.

991It appears, then, from what has foregone, that the subjects of the science of interpretation are principally the following; namely, the natures of the various indices to the customary meaning of the words in which the statute is expressed; the natures of the various indices, other than that literal meaning, to the intention or sense of the lawgiver; the cases wherein the intention which that literal meaning may suggest, should bend and yield to the intention which other indicia may point at.

Having stated the object or purpose of genuine interpretation, and pointed at the subjects of the science which is conversant about it, we will touch upon the interpretation, ex ratione legis, through which an unequivocal statute is extended or restricted. It may happen that the author of a statute, when he is making the statute, conceives and expresses exactly the intention with which he is making it, but conceives imperfectly and confusedly the end which determines him to make it. Now, since he conceives its scope inadequately and indistinctly, he scarcely pursues its scope with logical completeness, or he scarcely adheres to its scope with logical consistency. Consequently, though he conceives and expresses exactly the intention with which he is making it, the statute, in respect of its reason, is defective or excessive. Some class of cases which the reason of the statute embraces is not embraced by the statute itself; or the statute itself embraces some class of cases which a logical adherence to its reason, would determine its author to exclude from it.

But, in pursuance of a power which often is exercised by judges (and, where they are subordinate to the State, with its express or tacit authority), the judge who finds that a statute is thus defective or excessive, usually fills the chasm, or cuts away the excrescence. In order to the accomplishment of the end for which the statute was established, the judge completes or corrects the faulty or exorbitant intention with which it actually was made. He enlarges the defective, or reduces the excessive statute, and adjusts it to the reach of its ground. For he applies it to a case of a class which it surely does not embrace, but to which its reason or scope should have made the lawgiver extend it; or he withholds it from a case of a class which it embraces indisputably, but which its reason or scope should have made the lawgiver exclude from it.

Now, according to a notion or phrase which is current with writers on law, the judge who thus enlarges, or thus reduces the statute, ‘interprets the statute by its reason:’ or his extension 992or restriction of the defective or excessive statute is ‘extensive or restrictive interpretation ex ratione legis.’ His adjustment, however, of the statute to the reach or extent of its ground, is a palpable act of judicial legislation, and is not interpretation or construction (in the proper acceptation of the term). The discovery of the intention with which the statute was made, is the object of genuine interpretation; and, of the various clues to the actual intention of the lawgiver, the reason of the statute is one.

But where a statute is extended or restricted in the manner which we now are considering, the actual intention of the lawgiver is not doubted by the judge. Instead of unaffectedly seeking the actual intention of the lawgiver, and using the reason of the statute as one of the various clues to it, the judge rejects an actual (though faulty or exorbitant) intention which the lawgiver palpably held. Instead of interpreting a statute obscurely and dubiously worded, the judge modifies a statute clearly and precisely expressed: putting in the place of the law which the lawgiver indisputably made, the law which the reason of the statute should have determined the lawgiver to make. Consequently, where the judge in show interprets the statute restrictively, he abrogates or annuls it partially. And where the judge in show interprets the statute extensively, he makes of its reason a judiciary rule by which its defect is supplied. He makes of the reason of the statute a general ground of decision which provides for the class of cases overlooked and omitted by the lawgiver: For, as a ratio decidendi, though not as a ratio legis, the reason of a statute may perform the functions of a law.

In the following passages from the Pandects (Lib. I. tit 3, ll. 10, 11, 12, 13), the rationale of the process of extension, which the judge performs upon the statute, is stated unaffectedly and frankly. ‘Neque leges, neque senatusconsulta ita scribi possunt, ut omnes casus, qui quandoque inciderint, comprehendantur; et ideo de his, quæ primo constituuntur, interpretatione (aut constitutione principis) certius statuendum est.’ ‘Non possunt omnes articuli singillatim aut legibus aut senatusconsultis comprehendi; sed cum in aliquâ causâ sententia (sive ratio) eorum manifesta est, is, qui jurisdictioni præest, ad similia procedere, atque ita jus dicere debet.’ ‘Quotiens lege aliquid unum vel alterum introductum est, bona occasio est, cetera, quæ tendunt ad eandem utilitatem, interpretatione, vel certe jurisdictione suppleri.’

There is, it is true, an extensive, or restrictive interpretation 993which is properly interpretation or construction. For the literal meaning of the words wherein the statute is expressed, may not correspond to the purpose wherewith it was actually made; and the interpreter of the statute, guided by another index to the actual purpose of the statute, may abandon the meaning which custom has annexed to the words, for the meaning which the lawgiver attached to them.

Now, if the meaning annexed to the words by custom be narrower than the meaning attached to the words by the lawgiver, the interpreter (it is commonly said) interprets the statute extensively: If the former of the meanings be broader than the latter, the interpreter (it is commonly said) interprets the statute restrictively. But, manifestly, the statute itself is not extended or restricted by the process which we now are considering. The very law which actually was made by the lawgiver, is also the very law which is sought and stuck to by the interpreter; who merely proportions the grammatical meaning of the words to the broader or narrower meaning with which the lawgiver used them. The interpreter extends or restricts, not the statute itself, but the literal meaning of the words wherein the statute is expressed.

Having tried to distinguish genuine interpretation from the modes of judicial legislation which often are styled interpretation, we proffer a few remarks on some of the leading terms which are connected with the distinction in question.

Although the intention with which a statute is made often differs from the end which moves the lawgiver to make it, the reason of a statute and the actual intention of the lawgiver oftener coincide or tally. Accordingly, they often are opposed together, or contradistinguished jointly, to the literal meaning of the words in which the statute is expressed. Now, as contradistinguished jointly to the literal meaning of the words, the reason of a statute with the actual intention of the lawgiver are commonly styled by the moderns ‘the spirit of the law:’ by the Roman jurists, and the moderns who adopt their language, ‘the sentence of the law.’

According to most of the writers who have treated of interpretation, it is either grammatical or logical. The interpretation of a statute bears the name of grammatical in so far as it seeks the meaning which custom has annexed to the words, or seeks in that meaning exclusively the actual intention of the lawgiver. As looking for other indices to the actual intention of the lawgiver, or as seeking his actual intention through such other 994indicia, the interpretation of a statute assumes the name of logical. But as every process of interpretation involves a logical process, the contradistinguished epithets scarcely suggest the distinction which they are employed to express. The extension or restriction, ex ratione legis, of a statute unequivocally worded, is not interpretation or construction (in the proper acceptation of the term). According, however, to most of the writers who have treated of interpretation, this process of extension or restriction belongs to the kind of interpretation which they mark with the name of logical.

Of the numerous equivocal terms which the language of jurisprudence comprises, equity perhaps is the most equivocal and perplexing.¹⁵ Now, of the many and disparate meanings given to this slippery expression, some are connected inseparably with a kind of spurious interpretation: namely, with the so-called extensive interpretation, ex ratione legis, of a statute unequivocally worded. Where a defective statute is thus adjusted to its reason, there lies an equality or æquity (or a parity, analogy, or likeness) between the cases which the statute includes and the cases to which it is extended. A case to which it is extended, as well as a case which it includes, is embraced by its reason; and the two cases, therefore, in their common relation to its reason, are equal, analogous, or like.

¹⁵ See Lecture XXXIII. ante.

Accordingly, equity or æquity signifies the objects following (besides a multitude of others):—1. The parity between a case to which the statute is extended and a case which the statute includes.—2. The spurious extensive interpretation ex ratione legis: that is to say, the extension of the statute agreeably to the parity of the cases, or the process of extending the statute agreeably to the parity of the cases.—3. A personified abstract name which is moved by the parity of the cases to call for the extension of the statute. (‘Quod in re pari valet (says Cicero) valeat in hac quæ par est: valeat æquitas, quæ paribus in causis paria jura desiderat.’)—4. The reason of the statute to which the extension is given, or the reason of any statute which needs a similar extension. (It often is said, for example, of such or such a case, that the case is within the equity of such or such a statute, though the case is not included by the statute itself.)

The spurious extensive interpretation ex ratione legis, is 995styled analogy as well as equity. And it is said of analogy, as it is said of equity, that she is moved by the parity of the cases to call for the extension of the statute. It is said moreover of the pretended interpreter, that he interprets the statute analogically. But it would seem that the term analogy, like the expression equity, signifies most properly the parity between the cases.

[‘Equity’ is not applicable to restrictive interpretation ex ratione legis.]

The extensive or restrictive interpretation ex ratione legis, of a statute unequivocally worded, are not the only modes of judicial or oblique legislation to which the name of interpretation is often misapplied. E.g.: Entire or partial abrogation of a statute (with or without substitution of a new rule), without regard even to the reason of the statute. The grounds for which are, the judge’s own views of utility, or of that consequence and analogy (in legislation) to which we shall advert hereafter.

(He is to be blamed commonly, not for innovation, but for working it under false pretences, and without system.)

So, also, the creation of judiciary law (independently of the application of any statute) has been styled interpretation. E.g.: The law devised by prudentes, and adopted by tribunals, was said to be devised by interpreting the old law: such interpretation consisting, partly, in forming new rules, by consequence and analogy, on anterior law (statute or judiciary) and hence, probably, the name of interpretation; and, partly, in forming new rules, without regard to consequence or analogy, according to general utility, or any other standard of ethics (or legislation).¹⁶

¹⁶ ‘Interpretis officium, quod proprie in legis sententia explicanda versatur, per se quidem facile discernitur ab eorum munere, quorum est, ad causas applicare leges; unde etiam recte de utroque genere seorsim multa præcipiuntur. Est tamen genus quoddam præceptorum velut promiscuum atque in medio positum, eorum scilicet, quæ ad leges ex earum ratione, aut ad similia producendas aut restringendas spectant.’—Mühlenbruch, Doct. Pand. vol. i. lib. i. cap. 3.

As I shall shew hereafter, authentic interpretation is also genuine or spurious. A declaratory law being truly such, or introducing the new law under the guise of interpreting the purpose with which the old was made.¹⁷

¹⁷ Interpretation by règlement. See French Code and Bentham’s Judicial Establishments.

From the two processes¹⁸ which I have endeavoured to contrast, I pass to the celebrated phrase which is closely connected 996with them; though it regards the application of law (and also the creation of law) rather than the discovery of law by interpretation or induction. ‘After all the certainty (says Paley) that can be given to points of law, either by the interposition of the legislature, or the authority of precedents, one principal source of disputation, and into which, indeed, the greater part of legal controversies may be resolved, will remain still: namely, “the competition of opposite analogies.”’ The nature of the difficulties denoted by this celebrated phrase he attempts to state in the following passage. ‘When a point of law has once been adjudged, neither that question, nor any which completely, and in all its circumstances, corresponds with that, can be brought a second time into dispute. But questions arise which resemble that only indirectly, and in part, and in certain views and circumstances, and which seem to bear an equal or greater affinity to other adjudged cases: questions which can be brought within any fixed rule only by analogy, and which hold an analogy by relation to different rules. It is by the urging of the different analogies that the contention of the Bar is carried on: And it is in the comparison, adjustment, and reconciliation of them with one another; and in the discerning of such distinctions, and the framing of such a determination, as may save the various rules alleged in the cause, or, if that be impossible, as may give up the weaker analogy to the stronger, that the sagacity and wisdom of the Court are seen and exercised.’ Now, like all or most of the phrases into which ‘analogy’ enters, the celebrated phrase ‘the competition of opposite analogies’ is often used indeterminately. Accordingly, it has darkened the nature of the difficulties which it was contrived to express; and, therewith, it has obscured that pregnant distinction between statute and judiciary law with which we are presently occupied. It needs, therefore, the exact examination which I now shall bestow upon it.

¹⁸ See p. 632, ante.

Two distinct difficulties, incident respectively to two distinct processes, are denoted by the phrase as it is usually applied. Of these distinct difficulties, one is incident to the application of statute or judiciary law already obtaining or existing: the other is incident to the decision of a specific or particular case for which the existing law affords no applicable rule. The former may perplex the judge in his purely judicial character, or as properly exercising his properly judicial functions: the latter may embarrass the judge in his quality of judicial legislator, or as virtually making a rule for cases of a new description. As it is usually applied, the phrase is confined to the difficulty 997which is incident to the application of law, with the difficulty which is incident to judicial legislation. But a difficulty resembling these is incident to direct legislation, or the process of creating a statute. And this difficulty may be styled, as properly as the two others, a competition of opposite analogies.

I first shall consider the difficulty which is incident to the application of law already obtaining or existing; though the difficulty which is incident to judicial legislation is probably the difficulty that Paley particularly contemplated. Having considered the difficulty which is incident to the application of law, I shall proceed to consider the difficulty which is incident to judicial legislation, and also the resembling difficulty which is incident to the creation of a statute.

The system of positive law obtaining in any nation (or the complexion or collective whole of its positive law) is a body or aggregate (methodised or unmethodised) of various but connected rules. Now every rule which is definite or precise is applicable to cases of a class (or governs cases of a class) which also is definite or precise. For the rule is shaped exactly to the essence or nature of the class, or to the essentials, positive and negative, possessed by a case of the class: meaning by the positive and negative essentials of the case the properties which it necessarily has, and the properties which it necessarily wants, in so far as it belongs to the class, and in so far as the rule will apply to it. And as the rule is shaped exactly to the essence or nature of the class, so is the essence of the class exactly marked by the rule: the determinate purpose of the rule, and the determinate import of the rule, fixing the class of cases which its author intended it to govern. Consequently, if every rule in a system of law were perfectly definite or precise, every specific case that the whole of the system embraced would belong to a kind or sort perfectly definite or precise.

On the appearance of any case that the whole of the system embraced, the class to which it belonged, and the rule by which it was governed, might be known and assigned with certainty; or (at the least) the class and the rule might be known and assigned with certainty, by every discerning lawyer who had mastered the system thoroughly. But the ideal completeness and correctness which I now have imagined is not attainable in fact. More or fewer of the rules which constitute a system of law, and more or fewer of the cases which the whole of the system extends to, are inevitably framed and classed more or less indefinitely. What exactly are the cases which a given 998rule applies to, or what exactly is the rule which governs a given case, is a doubt that would arise occasionally, and would not be soluble always, though the system had been built and ordered with matchless solicitude and skill. And hence arises, or hence arises mainly, the difficulty which I now am considering: a difficulty incident to the application of law, and not to the creation of law by judicial or direct legislation.

In order to an analysis of the difficulty which I now am considering, we will suppose that the rule A is not perfectly definite; or (what is the same thing expressed in a different form) that the essence of the class of cases which A was intended to govern is not marked by A with perfect exactness. Further, we will suppose that the rule B, with the essence of the class of cases which B was intended to govern, is in the same plight of uncertainty. Moreover, we will suppose that the case x demands judicial decision: that x in certain respects bears a likeness to the cases which seem to be governed by A; but that x in other respects bears a likeness to the cases which seem to be governed by B. Lastly, I will suppose that the judge, distracted by the two likenesses, doubts whether A or B be the rule applicable to x.

Now, the difficulty which stays the judge from applying the law to the fact, may be called (in metaphorical language) ‘a competition of opposite analogies.’ For the likeness of x to the cases which seemingly are governed by A, and the likeness of x to the cases which seemingly are governed by B, may be deemed two opposite suitors contending for the preference of the judge: the former entreating the judge to decide x by A, and the latter imploring the judge to resolve x by B.

The difficulty arises, however, from the indefiniteness of the two rules. For if the rule A be the rule applicable to x (or if the rule B be the rule applicable to x), x, in respect of A (or x, in respect of B), is in the following double predicament. It has all the positive essentials possessed by cases of the class which the rule was intended to govern: And, moreover, it has no property or character through which it differs from those cases essentially or materially: that is to say, through which it differs from those cases in such wise and degree as render a common rule inapplicable to it and them.

Consequently, if the rules A and B be perfectly definite, and the classes of cases which they govern be therefore perfectly definite, the judge can arrive with certainty at one of the following conclusions: namely, that x is in that predicament in respect of the rule A, and therefore must be solved by A; or that x is 999in that predicament in respect of the rule B, and therefore must be solved by B; or that x is not in that predicament in respect of either of those rules, and therefore must be solved by a third, if a third that applies may be found. The difficulty arises, therefore, from the indefiniteness of the two rules; and is rather a competition or conflict of those indefinite rules than of the opposite analogies of x to the cases of the indefinite classes.

To the foregoing analysis of the difficulty which I now am considering, I append the following explanations:—To render my supposed example as simple as possible, I have imagined that the indefinite rules which strive for the preference of the judge are only two. But, though a greater number of such rules strive for the preference of the judge, the hindrance to the application of the law is substantially the one which I have analysed. The greater, however, is the number of the indefinite and conflicting rules, the greater, of course, is his difficulty. The greater, of course, is his difficulty in subsuming the case before him under the appropriate rule; or in finding that the case before him is embraced by none of the rules of which the law that he administers is actually composed.

Although it arises more frequently from a conflict of indefinite rules, the difficulty which I now am considering (or a difficulty essentially like it) may arise, without such a conflict, from a single indefinite rule. For it probably has happened, where a rule or principle is indefinite, that some judges have applied it to certain specific cases, whilst others have withheld it from cases essentially similar to the former. In other words, the rule has been applied to some, and withheld from other cases, though all the adjudged cases are of one description or category. Now if x, a case in controversy, be of that description or category, it bears a likeness to the cases to which the rule has been applied, and also a likeness to the cases from which the rule has been withheld. And though the likenesses are identical in relation to the various cases, they yet are opposed and contending in relation to the indefinite rule: one of them inviting the judge to apply the rule to x and the other suggesting to the judge that the rule is inapplicable to it.

But the difficulty which stays the judge springs from the indefiniteness of the rule. Supposing that the rule is judiciary, the difficulty implies that the existence of the rule is questionable; or, at least, has been disputed by the judges who have refused to apply it. For the rule itself is made by decisions (if it exist at all). Supposing that the rule is perfectly definite, the 1000judge may determine certainly (if not easily and quickly) how he should dispose of the case which awaits his judicial solution. It surely is, or it surely is not, of the class which the rule was intended to govern.

We have assumed tacitly, up to the present point, that the competition of opposite analogies which is incident to the application of law arises exclusively from the indefiniteness of a rule or rules. But it may possibly arise from a somewhat different cause: namely, the inconsistency inter se of several definite rules, or the intrinsic or self-inconsistency of one definite rule. Or (what, in effect, is exactly the same thing) it may possibly spring from the inconsistency with which such rules or rule have been administered or applied. Of two cases, for example, which belong to one category, and which therefore should have been adjudged by one and the same rule, the one may have been decided by a definite or precise rule and the other by a definite rule essentially different from the former. Or, supposing a single rule which is perfectly definite or precise, it may have been applied to one and withheld from another case, though the two adjudged cases are of one description or class.

Now, on either of these suppositions, it may happen that a case in controversy is essentially similar to the cases which have been resolved inconsistently. But, assuming that a case in controversy is essentially similar to those cases, it bears a likeness to one of them, and the same likeness to the other; which respective likenesses, though identical in relation to the cases, are opposed and contending analogies in relation to the rules or rule.¹⁹ Consequently the difficulty may spring from the inconsistency of several definite rules, or from the intrinsic inconsistency of a single definite rule. But, though it may spring from inconsistency which is not an effect of indefiniteness, it commonly springs from inconsistency of which indefiniteness is the cause: that is to say, it commonly springs from the inconsistency of several indefinite rules, or from the intrinsic inconsistency of a single indefinite rule. For the inconsistency in rules which is not an effect of their indefiniteness, is generally an evil that is easily corrigible.

¹⁹ If two cases essentially different have been decided by a common rule, a competition of opposite analogies cannot arise from the inconsistency. For if a case in controversy be essentially similar to either, it is essentially different from the other, and not essentially like it.

Generally, therefore, a system or body of law is kept passably free from it by direct or judicial legislation. But the inconsistency in rules which is caused by their want of precision, 1001is often an invincible malady, or a malady difficult to heal.²⁰ It often inheres in the purpose of a rule, and therefore is simply incurable. And where it is susceptible of cure (which far more often is the fact), it can seldom be expelled from the system without a solicitude and skill which lawgivers, direct or judicial, have rarely felt and attained to.

²⁰ Rules involving degrees: e.g. libel, lunacy, prodigality, reasonable time, reasonable notice. These are hotbeds of competing analogies. The indefiniteness is incorrigible. A discretion is left to the judge. Questions arising on them (and all competitions of analogies) are questions of law: e.g. they regard the applicability of an uncertain or inconsistent rule or rules to a given and known fact. They are hardly questions of interpretation or induction, for though the rule were explored and known as far as possible, doubt would remain.

It appears, then, from the foregoing analysis, that the competition of opposite analogies which is incident to the application of law, arises from this: that a rule in the system of law which the judge is engaged in administering, is inconsistent with itself; or that two or more of the rules which actually compose the system, are inconsistent with one another. It appears, also, from the same analysis, that the rule or rules are commonly thus inconsistent, because, inevitably or otherwise, it or they are indefinite; but that the rule or rules are occasionally thus inconsistent, although it or they are perfectly precise.

From the competition of opposite analogies which is incident to the application of law, I turn to the similar difficulty which is incident to judicial legislation.

Having analysed the equivocal phrase of which we have treated in the text,²¹ we will here review the meanings (or rather the principal meanings) of the equivocal term which it involves: namely, the term Analogy.

²¹ See Lecture XXXVII. ante.

1. As taken with the largest, and perhaps the most current of its meanings, the term analogy is coextensive with the term likeness: In other words, it signifies likeness or resemblance of any nature or degree. The process, for example, of reasoning, which we shall scrutinise hereafter, may be grounded on a likeness or resemblance of any nature or degree; and whatever be 1002the nature or degree of the given likeness or resemblance, the reasoning which is grounded upon it is styled reasoning by analogy. Moreover, the term analogy, as borrowed by the Romans from the Greeks, often signifies a reasoning which is grounded on a likeness or resemblance, instead of the likeness or resemblance whereon the reasoning is grounded. And analogy, as meaning such reasoning (or the consequent yielded by such reasoning) receives from the Roman Varro (treating of its etymon and value), the following extensive definition: ‘Veritas et ratio quæ a similitudine oritur.’

2. Between the species or sorts which are parts of a genus or kind, there obtains a resemblance or likeness whereon the genus is built; or (changing the phrase) between individuals of any, and individuals of any other of those sorts, there obtains a resemblance or likeness by which they are determined to the kind. Moreover, between individuals or singulars as parts of any one of those species, there obtains another resemblance, which is the ground or basis of the sort. For example: Between the various species which are parts of the genus animal (or between individuals of any, and individuals of any other of those species) there obtains a resemblance or likeness which determines them to the genus animal. And between individual men, as belonging to a sort of animals, there obtains another resemblance which determines them to the species man.

The resemblance between the species which are parts of a genus or kind (or between individuals of any, and individuals of any other of those species) is styled, in the language of logicians, generic. The resemblance between individuals as parts of any one of those species is styled, in the language of logicians, specific.

Now the likeness between any of the sorts which are parts or members of a kind (or between individuals of any, and individuals of any other of those sorts) is commonly contradistinguished, under the name of analogy, to the likeness between individuals as parts of any one of those sorts. For example: The likeness of a man to any of the lower animals, as distinguished from the likeness of a man to any of the human species, is commonly called analogy. Again: The term intellect, when it is used emphatically, denotes the human intellect. But between the intellect of men and the intellect of the lower animals there obtains a generic resemblance. Accordingly, the intellect of the lower animals is styled an ‘analogon intellectûs:’ that is to say, a something which re1003sembles generically the intellect of the human species, or the peculiar and pre-eminent intellect which is called intellect emphatically. Again: In relation to the several titles from which they respectively arise, an obligation ex contractu and an obligation quasi ex contractu are obligations of different species. But these two different species are parts of a common genus: namely, the genus of obligationes (in the sense of the Roman lawyers). Accordingly, an obligation quasi ex contractu (as the adverb quasi imports) is an analogon of an obligation annexed by the law to a contract.

It follows from what has preceded, that when we denote by the term analogy a generic and remoter resemblance between individuals or singulars, we make the following suppositions concerning the compared objects: First: That one or some of those objects are parcel of a given class, which, for the purpose in hand, we deem a species: that is to say, a class consisting exclusively of mere individuals or singulars, and not containing or comprising lower or narrower classes. Secondly: That the other or rest of those objects are not of that given species. Thirdly: That all the compared objects are parcel of a given genus by which that species is embraced.

3. Two or more objects may bear to another object (or they may bear respectively to several other objects), similar though several relations. Thus, A may be related to x, as B is related to x; or A may be related to c (or to c, d, and e), as B is related to x (or to x, y, and z). Now where several objects are thus related similarly, they bear to one another a likeness lying in a likeness of their relations. And this resemblance of objects lying in a resemblance of their relations, has been named by Greek and Latin, and also by modern writers, analogy: by Latin writers, translating from Greek, proportio.

Where objects are allied by a likeness lying in a likeness of their relations, the objects may be connected, or the objects may not be connected by a likeness of a different nature. On either, however, of these suppositions, the objects are parts of an actual, or a possible genus or species, built on the likeness of their relations. For a likeness of objects which lies in a likeness of their relations, as well as any other likeness connecting the objects with one another, may form a reason or ground for putting them together in a class.

Of the analogy or likeness of objects which lies in a likeness of their relations, the following examples are specimens.—The fin of a fish and the wing of a bird are analogous objects: 1004the fin being to the fish, in respect of its movements through the water, as the wing is to the bird, in respect of its movements through the air. There also obtains an analogy between an egg and a seed: for the egg is related to the mother, and to the incipient bird, as the seed is related to the generating, and to the inchoate plant. Where several legal cases are included by a law or principle, their similar though several relations to the law or principle which includes them, make them analogous cases. Thus, the several specific cases actually comprised by a statute; or the several specific cases comprised by a judiciary rule, carry a mutual analogy, or a mutual parity or æquity, in respect of their like relations to the statute or rule of law.

Again: We may suppose that the author of a statute, when he is making the statute, omits some class of cases falling within its reason. Now, on this supposition, a case which is thus omitted and a case which the statute includes bear to the reason of the statute similar though several relations. And in respect of their similar though several relations to the reason, the omitted and included cases are analogous, æqual, or pares.

But here, to prevent misconception, I must add the following remarks:—The several specific cases which are actually comprised by a statute, or the several specific cases which are comprised by a judiciary rule, are therefore analogous or pares: for the respective relations of the cases to the statute or rule which includes them are similar though several relations.

But when it is said of a litigated case, that it bears an analogy or parity to another case or cases, it commonly is not intended that the doubtful and litigated case is surely and indisputably included by a statute or rule in question. It commonly is meant that the litigated bears to the other case a specific or generic likeness; and that the former ought to be decided on account of the alleged resemblance, by a statute or rule in question on account of the resemblance alleged. Or else it is meant that the litigated bears to the other case a specific or generic likeness; but that the former ought not to be decided, although the resemblance is admitted, by a statute or rule in question, notwithstanding the admitted resemblance. Consequently, the asserted or admitted analogy of the litigated case to the other is not an analogy or parity lying in a likeness of their relations; or, at the least, it is not an analogy or parity lying in a likeness of their relations to a given statute or rule indisputably including both.

1005In truth, when it is said that a litigated case is analogous to another case, one of the following meanings is commonly imported by the phrase. It is meant that the litigated case bears to the other case a specific and proximate resemblance; and that the former ought to be decided on account of the alleged resemblance, by a given statute or rule in which the latter is included. Or else it is meant that the litigated bears to the other case a generic and remoter resemblance; and that the former should be brought or forced, on account of the alleged resemblance, within a statute or rule by which the latter is comprised: that is to say, that a new rule of judiciary law, resembling a statute or rule by which the latter is comprised, ought to be made by the Court, and applied to the case in controversy.

In any of the meanings which we have reviewed above, the term analogy signifies likeness: namely, likeness or resemblance of any nature or degree; generic and remoter likeness, as opposed to specific and proximate; or a likeness of several objects lying in a likeness of their relations. In any of the meanings which we shall review below, the term analogy denotes an intellectual process: a process which is caused or grounded by or upon an analogy (in one or another of the meanings which we have reviewed above).

1. Analogy denotes the analogising of several analogous objects: that is to say, the considering the several objects as connected by the analogy between them.

2. Analogy denotes an inference, or a reasoning or argumentation, whereof an analogy of objects is mainly the cause or ground.

The nature of the inference, or reasoning, may be stated or suggested thus. Two or more individuals, or individual or singular objects, are allied by a given analogy. It is known (or, at least, is assumed), before and without the reasoning, that a given something is true of one or some of these objects. But it is not known, before and without the reasoning, that the given something is true of the other or rest of these objects. From the given analogy, however, which connects these objects with one another, the following inference is drawn: Namely, that the given something which is true of one or some of these objects, is true of the other or rest. Or the nature of the inference or reasoning may be stated or suggested thus:—A is x and y and z. An analogy or parity 1006obtains between A and B; for B as well as A is x and y. We know (or, at least, we assume), before and without the reasoning, that A is also z. We do not, however, know, before and without the reasoning, that B is also z. But, since B as well as A is x and y, and we know, the reasoning apart, that A is also z, we infer, by analogy, or parity, that B is also z. In short, from two antecedents or data which are known independently of the reasoning, we argue or proceed to a consequent which is unknown without it. From our knowledge that several objects are connected by a given analogy, and our knowledge that a further something is true of one or some, we infer that the further something is true of the other or others.

From the following passage in Quinctilian, it seems that the Roman writers, as borrowing a term from the Greek, styled the reasoning in question, analogy; and that the Roman writers styled such reasoning proportio (sometimes also comparatio), when they turned the Greek expression into its nearest Latin equivalent. ‘Analogiæ, quam proxime ex Græco transferentes in Latinum proportionem vocaverunt, hæc vis est: Ut id, quod dubium est, ad aliquid simile, de quo non quæritur, referat; ut incerta certis probet.’—Inst. Orat. l. i. c. 6. We have already quoted the definition of the term, given by the etymologist Varro: agreeably to his definition, analogy signifies the consequent yielded by the reasoning in question, rather than the reasoning itself.

But though the reasoning in question is styled analogy or proportion, it is styled more commonly reasoning by analogy or else analogical reasoning: ratiocinatio per analogiam or argumentatio analogica. It is styled also reasoning by parity, and since equality or æquity is tantamount to analogy or parity, it might be styled moreover reasoning by equality or æquity.

And here I may remark that the homely phrase to liken or to liken one thing to another, is equivalent to the finer phrase to reason by analogy or parity. We know that several objects are like in certain respects: We know that a something is true of some of these objects, although we do not know that the same is true of the others: From our knowledge that the former and the latter are like in certain respects, and our knowledge that the given something may be truly affirmed of the former, we infer that the given something may be truly affirmed of the latter. Now here in homespun English, we liken the latter to the former: that is to say, we argue them like the former in one 1007or more respects, since we know them, the inference apart, like the former in others.

In a case of analogical reasoning, the analogy of the compared objects may lie in a resemblance of their relations, or in another resemblance. On either of the two suppositions, the resemblance may be specific, and comparatively close or strong, or the resemblance may be generic, and comparatively distant or faint. In other words the objects from which we infer, and the objects to which we argue, may be parts of one species: or the former may belong to a species, parcel of a given genus, and the latter to another species, parcel of the same genus. Whatever be the nature of the likeness on which the reasoning is grounded, or whatever be the degree of the likeness on which the reasoning is grounded, the reasoning may be called with propriety, as it commonly is called in practice, reasoning by analogy or parity.

But such analogical arguments as are grounded on specific resemblance, and such analogical arguments as are grounded on generic resemblances, are not unfrequently distinguished by the following antithesis of phrases. An inference from singulars of a sort is grounded, it is said, on experience. An inference from singulars of a sort which is parcel of a given kind, to singulars of another sort which is parcel of the same kind, is grounded, it is said, on analogy.

For example: It is said that a physician would reason from experience, in case he reasoned thus: ‘Some men have died or have suffered some other harm, on taking a certain drug; therefore, other men will die, or will suffer a like harm, if the drug be taken by them.’ It is said that he would reason from analogy, in case he reasoned thus: ‘Some animals of one of the lower sorts have died of a certain drug; therefore men will die of the drug, if it be taken by them.’

But such analogical arguments as are grounded on specific resemblances and such analogical arguments as are grounded on generic resemblances, are vaguely divided by a difference which is merely a difference of degrees. There is not the opposition of natures between the two classes of arguments, which seems to be expressed or intimated by the foregoing antithesis of phrases. For whether the likeness be specific or generic, an argument raised on a likeness rests upon two antecedents: first, the likeness between the objects from which we reason or infer, and the objects to which we argue; secondly, the given and further 1008something which we know or assume of the former, and which, by analogy or parity, we impute or ascribe to the latter.

Now, whether the likeness be specific or generic, each antecedent is known to us (either directly or mediately) by or through an experience which has occurred to ourselves or others. Consequently, whether the likeness be specific or generic, the reasoning is built on experience, and also is built on analogy. In respect of either antecedents the reasoning is built on experience. In respect of that antecedent which consists in the likeness of the objects, the reasoning is built on analogy with which experience presents us. We know, indeed, through experience (helped by analogical reasoning) that such analogical arguments as rest on specific resemblances, are commonly worthier of trust than such as rest on generic. But this extremely vague, though extremely important difference, is merely a difference of degrees. Although an analogical argument raised on a stronger resemblance may be surer than a similar argument raised on a weaker resemblance, the natures of the several arguments are essentially alike or identical.

As concerned with matter of a nature which we shall endeavour to explain, analogical argumentation (supposing it justly conducted) is only contingently true: Or (changing the phrase) the something which is true of the object from which we reason or infer, is only probably true, or only contingently true, of the objects to which we argue.

As concerned with matter of another nature, which we shall also endeavour to explain, analogical argumentation (supposing it justly conducted) is necessarily or certainly true: Or (changing the phrase) the something which is true of the objects from which we reason or infer, is certainly or necessarily true of the objects to which we argue.

Analogical reasoning of the former kind we will call analogical reasoning concerning contingent matter. Analogical reasoning of the latter kind we will call analogical reasoning concerning necessary matter. We incline to believe that the latter is not commonly called analogical reasoning, and it certainly differs essentially from analogical reasoning concerning contingent matter. Accordingly, we have hitherto assumed, in treating of reasoning by analogy, that all analogical reasoning concerns contingent matter.

1009[Analyse analogical reasoning of the first kind, and compare it with syllogism and perfect induction, and with analogical reasoning (if such it can be called) which is concerned with necessary matter.]

Contingent Truth is truth not inseparable from a notion of the object: Our belief resting, not on a necessary connection of truth with the object, but on experience and observation (one’s own or others’) of their invaluable or customary conjunction: such as the experience of pleasure or pain as connected with objects of a given class: Our belief or expectation of future conjunction being (at least after experience) proportioned to the degree in which, in times past, conjunction has approached to invariableness. E.g. The death of men is expected with perfect confidence; but the effect of a drug on the human body, or of an object on the human mind, in the way of pleasure or pain, is expected with less confidence. The nature of the belief or expectation derived from past conjunctions, is not within the province of the logician, who takes the facts from the philosophy of the human mind. The belief or expectation seems to be confident before experience, and to be afterwards reduced to experience; widened by analogical reasoning founded on generic resemblance. It will be admitted by all, that our belief ought to be commensurate with the experience; e.g. with the proportion borne by individual instances in which the conjunction has been experienced, to individual instances in which it has not. Whether our belief is first absolute, and then proportioned to experience, or is first hesitating and gradually proportioned to experience, is not a question failing within the range of the logician.

In ordinary language, contingent truths are certain or probable. As opposed to necessary, all are contingent in the sense above explained; but, according as truth is accordant with all past experience, or only accords with past experience generally, it is certain or probable.

A contingent or probable truth does not necessarily belong to the object, although in fact no object of the sort has been known without it.

Necessary Truth is that which is true of all objects, like the objects argued from, by reason of their having that wherein they are analogous. A necessary efflux of that; a something without which the object cannot be conceived as having that. 1010[e.g. Triangles, as such. (Hobbes.) Law case of a class, as such: i.e. as abstracted from its individual peculiarities. Legal consequence true of it, etc.]

In all these cases the truth seems to be proprium or property, strictly so called. It flows necessarily from the essence of the object: i.e. from the properties (positive or negative or both) which make the object to be of the class to which it belongs: Although it is not itself of the essence: e.g. equality of the three angles of any triangle, etc.

In pursuance of the order suggested above we shall proceed to analyse analogical reasoning concerning contingent matter.

1°. Induction; i.e. analogical reasoning extended to all other objects (or other objects generally and indeterminately), having the given analogy to the object or objects argued from. (Call it, at present, imperfect induction, or induction simply; being totally different in nature from what is commonly called perfect induction, and to which we shall advert below in conjunction with syllogism.)

2°. Analogical argument not involving any such universal or general illation; but regarding one or some individually determined singulars, having the given analogy to the objects argued from.

This latter, as opposed to induction, may be called particular reasoning; and may be drawn without adverting to others generally. But in so far as it will hold, it supposes that the truth applies universally or generally, and indeterminately.

The degree of assurance with which the particular conclusion may be embraced is proportioned to the approach to universality. E.g.: A has x and y. A has also z. B has x and y. Ergo, B has z. But why? Because all objects having x and y, or objects generally having x and y, have z: insomuch, that B is, certainly or probably, one of a number of objects having z.

This is what we do when we attempt to state the grounds of our inferences. Also, when, in confuting others, we suggest a contradictory case or cases. Immense importance of the habit: Most people being apt to assume from a few cases universally, and then to syllogise. This leads me to compare 1011syllogism as concerned with contingent matter, and particular analogical reasoning as concerned with the same.

[Distinguish analogical reasoning which is concerned with contingent matter, from syllogism and perfect induction, of which the matter is also contingent.

Analyse analogical reasoning (induction and particular) concerning necessary matter, and compare it with induction (necessarily perfect) and syllogism (necessarily concluding absolutely) as concerned with the same.]

In analogical reasoning as concerned with contingent truths, the truth or probability of the inference depends on two causes: namely, the truth of the two antecedents, and the invariable or customary connection of the second antecedents with other singulars like those from which we argue. If the antecedents are true and the conjunction invariable, the inference is certain: i.e. has the certainty which alone can belong to contingent truth (as explained above).

If the antecedents are true and the conjunction not invariable, the inference recedes from certainty in proportion to the recess from invariableness.

All analogical reasoning proceeds from a singular to a singular, or singulars, or from singulars to singulars, or a singular. But when we infer from singulars or a singular to one or some, it is usually called ‘reasoning,’ or ‘particular reasoning.’ When from a singular or singulars to the rest of the actual or possible class, ‘induction.’

But in every case, the process is essentially the same. For the confidence in a particular conclusion depends upon the approach to invariableness of conjunction: i.e. upon the possibility of an induction approaching to truth. Many inductions are founded directly on an analogy: e.g. What is true of one of a species, is true of other individuals. But this again rests ultimately on experience.

It appears from what has preceded, that reasoning by analogy or likeness (of any nature or degree) is grounded on two antecedents: first, the likeness between the objects from which we reason or infer, and the objects to which we argue; 1012secondly, the something, which (the inference apart) we know to be true of the former, and which, by analogy or parity, we impute or ascribe to the latter. But though these two antecedents are immediately the ground of the inference, the inference reposes also on a further or ulterior basis. For why do we argue from the likeness between the compared objects and the something which we know to be true of one, or some, of the objects, that the something is true moreover of the other, or rest of the objects? The nature of the ulterior basis on which the inference reposes is determined by the nature of the matter with which the inference is concerned.

In [some] cases of reasoning by analogy the truth of the analogical inference (supposing it deduced justly) is contingent, or probable: that is to say, the something that is true of the objects which the just inference is brought from, is contingently true of the objects to which the inference is carried. In other cases of reasoning by analogy, the truth of the analogical inference (supposing it deduced justly) is necessary or certain: that is to say, the something that is true of the objects which are justly argued from, is necessarily true of the objects which are justly argued to.

Analogical Reasoning as concerned with contingent Matter, distinguished from Syllogism and perfect Induction as concerned with the same.

Whatever there has been of reasoning, as meaning process from known to unknown, has been performed by an analogical argument (an induction), by which we obtained the major premise.

And, moreover, in contingent matter, syllogism is apt to mislead. It rarely happens that the major premiss can be universal, conformably with material truth, though the formal truth of the conclusion depends on assuming such material universality.

Since then syllogism can give us no new truth, and since it may mislead, what is its use?

I incline to think that the important part is not syllogism. But terms, propositions, definitions, divisions (abstracted from 1013all particular matter) are all-important. It is a great error of most logicians to consider these as merely subordinate to syllogism, which is the most futile part. From my friend John Mill, who is a metaphysician, I expect that these, and analogical reasoning and induction, abstracted from particular matter (which are the really practical parts), will receive that light which none but a philosopher can give. For though logic is a formal science, and takes its truth from others, none but a metaphysician can determine its boundaries or explicate it properly.

[Necessity for illustrations from numerous sciences.—Many of the methods seemingly peculiar, would be found universal or general.]

A something equivalent, or nearly approaching, to syllogism, always happens when we state in our own minds the grounds for a conclusion in a particular reasoning: e.g.:

A is x and y. A is also z. B is x and y. B, ergo, is also z. But why? Because all singulars being x and y are also z, or singulars being x and y are generally and indeterminately z. In other words, we can only infer from A to B, on the supposition that A is the representative of a whole class, or of singulars generally contained in a class. The argument, therefore, must be put thus: All singulars being x and y, or the singulars generally which are x and y, have z. The singular or singulars forming the subject of the inquiry are z. Therefore, certainly or probably, B is z. And this would be much more convenient than the ordinary syllogism, which assumes in the major premiss a universality commonly false in fact, and which, therefore, must be denied again in the conclusion. For the conclusion, in fact or materially, cannot be absolute, unless the universality assumed in the major premise be materially true.

[Use of syllogism (or analogous process), in leading us to review grounds: In confutation:—reminding antagonist that he has assumed something not tenable.]

But, in fact, we never syllogise, though we perform an analogous process. We run the mental eye along the analogous objects, and if we find them contradictory, etc., we conclude probably, or reject, unless we find special reason. Hence Locke’s sarcasm.²²

²² See ‘Essay on the Human Understanding,’ vol. ii. c. xvii. § 4.

In all particular analogical reasoning which is concerned with contingent matter, the truth of the inference (considered 1014as such), depends on the truth of the antecedents (and on something else). And what I have said of syllogism, as to the dependence on terms and propositions, applies to perfect induction. As in syllogism, true of all, true of every, so in induction, true of every, true of all: Vi materiæ, vi formæ [materially or formally]. It follows that by syllogism we can arrive at no new truth, the conclusion being involved in the major premiss. We merely affirm of one what we had affirmed of all, including one. Or we merely affirm that the subject of the conclusion is one of the all, of which, in the major premiss, we had affirmed the predicate of conclusion.

Analogical Reasoning (Induction and Particular) as concerned with Necessary Matter. The Induction necessarily perfect. Syllogism as employed about the came Matter.

Now here we merely reason from a singular to a singular, as in the case of contingent truths. But the argument carries with it all the apodictic certainty which belongs to a syllogistic inference.

For A is a and b. A is also x: and A has x in such wise that every singular like it, in having a and b, must have x. B is analogous to A, in having a and b. Therefore B is of necessity x.

It is manifest that this is equivalent to the following syllogism. Every singular which is x is also z. B is a singular which is x. Therefore, B is also z.

But still there is this difference, that though, like a syllogism, the inference follows formaliter, it also follows materialiter. So that the cogency lies in the truth of the antecedent, and not in the relation and disposition of the terms. And, on the other hand, it differs from an analogical argument concerning a contingent truth. For the antecedent necessarily imports the consequent Analogical reasoning is generally considered as being conversant about contingent matter, and therefore I have so considered it.

Much of the certainty ascribed to mathematical reasoning lies in the truths with which it deals being of this class. Or at least, in approaching so near them that the deviations may be thrown aside, and afterwards allowed for in the way of limitation 1015to the inference. This is also the case with many of the truths with which lawyers have to do. And hence Law (teste Leibnitz) much like mathematics.^[*] In either case, the cogency arises from the nature of the premiss.

[^* The 3^rd edition has this note: “Leibnitz, Epist. ad Kestnerum. See quotation, post.” The quotation it mentions appears on p. 1087 below.]

[Eulogy on Law, from being Connected, on the one hand, with Ethics and Religion, and on the other, not less fitter to form the mind to habits of close thinking than the most abstruse of the mathematical sciences. Also, to exercise the mind in evaluation of evidence regarding contingent truths.]

For example: Interpretation of part of a statute by another part: interpretatio secundum analogiam scripturæ (so called as applied to a statute, or to any other written document). Interpretation of a statute by a statute made by the same legislature in pari materiâ: interpretatio secundum analogiam scriptoris (genuine interpretation).²³

²³ For the Analogy of Grammarians, see Stewart, 249, 250. Johnson, ‘Analogy.’

… The last, an inference resting in speculation. But often, a practical consequence built upon a perception and comparison of analogous objects; e.g. similis similium declinatio.

The extension of a statute, etc. ex ratione legis, is an example of analogical interpretation (genuine).

Lower animals reason,—how? The process of inference which they employ ought to be called reasoning. They also compare and abstract, as a necessary forerunner to inference.

It follows not from the form, but by reason of the matter. For because A and B are x and y, and A is always z, it is not true that B is also z. Whatever truth there is, therefore lies, not in the form of the reasoning, but in the intrinsic truth of the antecedents: i.e. because the antecedents are intrinsically true, we infer the truth of the consequent.

In syllogism and perfect induction, that is, in formal argumentation, the conclusion follows, ratione formæ.

Endeavouring to suggest an answer to this pregnant and 1016difficult question,²⁴ we begin with discriminating, as sharply and clearly as we can, formal and material reasoning. By ‘formal reasoning’ (the propriety of which expression will appear hereafter), we signify the process of syllogism, with the process of formal induction. The nature of these processes (to which we shall revert below) may be indicated briefly in the following manner. In the process of syllogism, a narrower proposition is extracted, by a formal and necessary influence, from a larger and universal proposition which contains the narrower implicitly: In the process of formal induction, a universal proposition is collected, by a formal and necessary influence, from all the singular propositions of which the universal is the sum.

²⁴ The question asked by one of the class (apparently Mr. J. S. Mill), viz., ‘What, then, is the use of syllogism?’—S. A.

By ‘material reasoning’ (the propriety of which expression will also appear hereafter), we denote analogical reasoning in each of its principal kinds: namely, the reasoning which yields a consequent that is either singular or partial, and the reasoning which yields a consequent that is either universal or general. Of the difference between these processes (to which we shall revert below) the following is a brief description.—In every reasoning raised on a likeness or analogy, the analogical inference proceeds from an assumed singular or singulars to another singular or singulars: that is to say, it proceeds to one or more of all those various singulars which are connected by the given analogy with the singular or singulars assumed. But where it yields a consequent which is either singular or partial, it proceeds to one or a few of all those various singulars. Where it yields a consequent which is either universal or general, it proceeds, without exception, to all those various singulars, or to all those various singulars with more or less of exception.

We venture to name the inference which is merely singular or partial (or which yields a conclusion or consequent that is merely singular or partial), reasoning by example. For it seems identical with the process which logicians denominate exemplum, and which they describe usually in some such words as the following:—‘Argumentatio in quâ unum singulare ex alio colligitur.’ The universal or general inference is called emphatically induction, and is usually described by logicians in some such words as the following:—‘A singulari ad universale progressio.’ To distinguish it from the formal induction which is a necessary induction or inference, we style it material induction. And here we must remark, that, in treating of argumentation of any of 1017the above-mentioned sorts, we always assume (unless we express the contrary) that the inference which we are considering is perfectly good or legitimate: that is to say, that the consequent has all the truth, in nature or in degree, which the natures of the reasoning and the case will allow the reasoner to reach.

With this remark we pursue our attempted discrimination of formal and material reasoning: of the process of syllogism with the process of formal induction, and the process of reasoning by example, with the process of material induction.

In any legitimate syllogism of any figure or mode, the process of argumentation is virtually this. In the major proposition, or major premiss, we affirm or deny a something of all the individuals or singulars which constitute a given class. In the minor proposition, or minor premise, we assume and affirm of a number of individuals, that they are some of the individuals which constitute the given class; or we assume and affirm of a single individual, that it is one of the individuals which constitute the given class. In the consequent proposition, or conclusion, we affirm or deny of the subject of the minor, what we affirmed or denied of the subject of the major.

Or the process of affirmation or negation which we perform in the conclusion, may be stated more clearly thus: What, in the major, we affirmed or denied of the all, we affirm or deny of the singulars or singular, which, by the assumption in the minor, are some or one of the all. Where a syllogism is affirmative, the process of argumentation runs in the following manner:

‘Every A (all A’s constituting the given class) is x. But every B is an A: that is to say, all the singulars of the narrower class constituted by all B’s, are some of the singulars of the larger class constituted by all A’s. Therefore, every B is x.’

‘Every A is x. But some B’s (of which the quantity or number is not determined) are A’s; or some B’s (of which the quantity or number is determined, but which are not determined individually or singularly) are A’s: or one B (not determined individually) is an A. Therefore, such some, or such one, are, or is, x.’

‘Every A is x. But these or those (individually determined) B’s are A’s; or this or that (individually determined) B is an A. Therefore, these or those B’s, or this or that B, are, or is, x.’

Where the syllogism is negative, the process of argumentation pursues the following course:—‘No A is x. But every B is an A: that is to say, all, etc. Therefore, no B is x.’ ‘No A is x. But some B’s (of which, etc.) are A’s; or some B’s (of 1018which, etc., but which, etc.) are A’s; or one B (not, etc.) is an A. Therefore, such some, or such one, are not, or is not, x.’

‘No A is x. But these or those (individually, etc.) B’s are A’s; or this or that (individually, etc.) B is an A. Therefore, these or those B’s, or this or that B, are not, or is not, x.’

It may be gathered from the foregoing exposition, that the conclusion of every syllogism lies implicitly in the premisses; or that what is asserted by that, is asserted implicitly by these. In the process, therefore, of syllogising, there is not really an illation or inference. Inasmuch as the truth in the conclusion is parcel of the truth in the premisses, there is not a progression to a consequent really distinct from the antecedents. Really (though not formally), the process consists exclusively of two assertions: first, that a given something may be said truly of every of a given all; secondly, that every of the individual objects which form the subject of the minor (or the single individual object which forms the subject of the minor) is one of the given all.

It also may be gathered from the foregoing exposition, that the consequent or concluding proposition, as being the consequent of the premisses, follows from the premisses by reason of their form: that is to say, independently of any truth which the premisses themselves may contain, and even of any of the meanings which their subjects and predicates may import. Though each of the premisses asserts a falsity, or though its subject and predicate signify anything or nothing, the conclusion or consequent proposition, as being the consequent of the premisses, is deduced or deducible from these by a formal and necessary illation. ‘Conclusio a premissis colligitur, per necessariam et formalem consequentiam, propter legitimam præmissorum in mode et figurâ dispositionem.’ That the conclusion follows from the premisses, independently of any of the meanings which their subjects and predicates may import, is shewn by the foregoing examples; wherein A, B, and x, may signify anything or nothing.

That the conclusion follows from the premisses, independently of any truth which the premisses themselves may contain, is shewn by the examples following. ‘Every animal is a stone. But every man is an animal. Therefore, every man is a stone.’ ‘No animal is sentient. But every stone is an animal. Therefore, no stone is sentient.’ Now in each of these syllogisms, the consequent proposition, as being the consequent of its premisses, is necessarily true; or (speaking 1019more accurately), it follows from its premisses by a legitimate and necessary inference. In the former syllogism, however, the major premiss is false; and the conclusion inferred from the premisses, as not being such conclusion, is false also. In the latter syllogism, each of the premisses is false; whilst the conclusion inferred from the premisses, as not being such conclusion, is true; but since the premisses are false, the truth of the conclusion, as not being such conclusion, has no connection with its truth in its quality of a true consequent.

In short, the rationale of the process of syllogising may be expressed by the following formula:—‘A something may be said or predicated of every of a given all: Every of a number of individuals, or one single individual, is one of the given all: What may be said or predicated of every of the given all, may be said of the subsumed every, or of the subsumed one, which, according to the subsumption, is one of the given all.’ It is manifest from this formula, that the truth or falsity of either of the premisses, or the significance or insignificance, of the subject or predicate of either, neither affect the consequence, nor the consequent, to which it leads. The validity of the consequence or inference depends exclusively upon two data: first, the unlimited universality of the affirmative or negative proposition which constitutes the major premise; secondly, the assumption that the singulars or singular which form the subject of the minor, are, or is, of the singulars which form the subject of the major. These being granted, the consequent, as the consequent, follows by a necessary consequence.

Various singular objects are connected by a common resemblance which shall be called z: and by reason of this common resemblance they constitute, or might constitute, a given species or genus. But of every of these various singulars, when considered singly and severally, and also without respect to the actual or possible class, a given something, which shall be called x, may be affirmed or denied. Now what may be affirmed or denied of every of these various singulars, when considered severally, may also be affirmed or denied of every of these various singulars if they be considered collectively, and as forming or constituting the class.

Major premiss: Various singulars, including A and B, are connected by y. But A is, or is not, x; B is, or is not, x: And every other of the various singulars, as considered singly and severally, is, or is not, x.

1020Minor premiss: All these various objects, as considered jointly and collectively, constitute, or might constitute, the species or genus Z.

Conclusion: Every singular constituent of the actual or possible class is, or is not, y.

It is manifest that there is no illation. That what is true of every of the objects as considered singly and severally, is true of every of the same as considered jointly and collectively, and as being the constituent parts of an actual or possible class.

It is manifest that it follows by reason of the form. For let the major or minor be what it may, what is true of every when the objects are taken severally, must equally be true of every when the objects are taken collectively, and considered as bound together by a class or common name. Or that which is true of every unit of twelve when not considered as forming a dozen, is true of every of twelve considered as forming a dozen.