Cursus Philosophicus Thomisticus, Secunda Pars Artis Logicae

To explain fully what pertains to knowing the premises, we must set out their conditions, which premises serving demonstration must satisfy.

Thus, the first thing to consider is a proposition known per se, or immediate. For although not every demonstration is formed from propositions known per se, every demonstration nevertheless depends on them and is resolved into them. And since every proposition known per se is immediate, it is explained negatively insofar as it is immediate, and positively insofar as it is known per se. A proposition is defined as immediate by the Philosopher, as also by St. Thomas, Posterior Analytics I, lecture 5¹: “a proposition lacking a middle, or a proposition before which there is no other.” It is understood to lack a probative middle in an intrinsic and a priori proof, as will be explained further in the following disputation, although it can admit proof a posteriori and through extrinsic causes. Hence St. Thomas says, De Veritate, q. 10, a. 12², that something is known per se when the predicate is “of the account of the subject.” Here one must observe that this definition is to be understood formally: that is, when a predicate belongs to the account of the subject immediately and per se. Otherwise, if a predicate belongs to something through another, it is not incompatible with its being proved and made manifest through some middle, however essential it may be. Thus it can be proved that a human being is sentient because he is an animal, and that a line is not a substance because quantity is not a substance. The Philosopher mentions this in Posterior Analytics I, ch. 11³; St. Thomas may be consulted there, lecture 26⁴. And this is how the divine Thomas is to be understood in Summa theologiae I, q. 14, a. 1, when he proves that God is intelligent because he is immaterial: intellectuality is included in immateriality and converts with it, just as Peter is proved rational because he is a human being. These are proofs from what includes to what is included.

If, however, a proposition known per se is considered positively, it is defined as “one that becomes known from the terms themselves, or from an explanation of the terms.” Hence it is clear that a proposition known per se is not the same as one known intuitively or through sensory experience. For what is known by sense is not known as a proposition; nor does it become known either as a simple object apprehended or from the explanation of the terms alone, but because it is reached by external experience. Thus, although snow’s being white is known by experience in sense, in the intellect it is not a proposition known from terms connected per se, but rather concerns contingent matter. Finally, a proposition known per se is not the same as one known without reasoning. For through faith we know without reasoning, on account of the authority of the speaker, and yet this is not known per se; and an angel knows without reasoning in first principles things that become known through proof and causes.

We explained in the text, ch. 2, how a proposition known per se, or a first principle, is divided into axioms and propositions—that is, into most common and less common principles. It will be discussed further in the following disputation.

There is, moreover, the celebrated division of a proposition known per se into one known per se in itself and one known per se to us, which St. Thomas gives in Summa theologiae I, q. 2, a. 1, and De Veritate, q. 10, a. 12⁵. Scotus and his school generally do not accept it, however, and some more recent authors say that St. Thomas spoke improperly by dividing a proposition known per se in this way. For what is known per se in itself and not to us is, strictly and formally, not a proposition but the thing signified by the proposition. Yet this thing is not said to be known per se or knowable except insofar as it is subjected to terms, whether formal concepts or words, by which it is made known. If, then, the terms that make it known are not themselves known per se but are proved through something else, neither will the proposition itself be known per se as a proposition. But if the very thing signified is subjected to other terms known per se, it will be a proposition known per se—not, however, insofar as it is subjected to those confused terms that must be proved through something else. The whole opposing argument, which denies that there is a proposition known per se in itself, is reduced nearly to this.

Nevertheless, St. Thomas, considering a proposition known per se more deeply, spoke not only of it according to the formal signification of a proposition, on the side of the concept or word that signifies, but also on the side of the foundation found in the very thing signified. For whether a proposition is demonstrable through another middle or lacks the middle by which it is demonstrated, and is thus immediate and known per se, arises not from the signification of the terms themselves but from the very connection of the things signified. For where a predicate belongs immediately to a subject, it lacks a middle by which that predicate might be proved of that subject; consequently, so far as it itself is concerned, it grounds a proposition known per se. Yet a proposition known per se cannot always be formed with respect to us, because the extremes or terms of which that proposition consists are not immediately known to us, so that their connection with one another is evident to us through the explanation of the terms alone. Rather, it becomes known presupposing a reasoning process that makes the terms manifest and consequently their connection. Therefore it is not known per se to us absolutely, but presupposing a reasoning process; that is what it is for something to be known per se to the wise.

There is no impropriety in taking a proposition objectively, not formally, when we speak of a proposition, but from the side of the thing signified, which is formally signified through concepts and terms. St. Thomas did this all the more aptly because the present concern is more to explain the truth of a proposition than the proposition itself. For the subject is whether a truth is demonstrable or indemonstrable, that is, known per se; this had to be explained also from the side of its foundation and of the thing signified.

Hence the opposing argument is answered. We use “proposition known per se” not only for a formal proposition but also for its objective content and for the foundation of its truth. It is known in itself when, on the side of the object, it has an immediate connection; it is known per se to us when that connection is constituted and explained by such terms that the connection of the extremes becomes known without any reasoning, through the explanation of the terms alone.

Concerning the predicate of every, to understand what it is and how it is required for the premises of a demonstration, one must note, following St. Thomas, Posterior Analytics I, lecture 9⁶, that “being predicated of every” is taken one way in the books of the Prior Analytics and another way here. There, attention is given only to the form of universal predication of every subject; here, to its matter or necessity. Thus there the predicate is said to be of every subject when it belongs to each thing contained under the subject, whether necessarily or contingently. In a word, “being predicated of every” according to the Prior Analytics form means any predicate qualified by a mark of universality and distribution, whether probable or necessary; for syllogistic form does not consider this, since it serves demonstrative as well as opinionative syllogism. By contrast, the Posterior Analytics sense of “of every” includes two conditions, as St. Thomas says in the cited place. First, nothing can be subsumed under the subject to which the predicate does not belong; in this it agrees with “of every” in the Prior Analytics sense. Second, there must be no time at which the predicate does not belong to it. Thus the Philosopher designates certain and necessary truth, and embraces the whole definition in these words: “By ‘of every’ I mean that it belongs to one and does not belong to another, applies at one time and does not apply at another, but applies to every one and always.”

Accordingly, two kinds of universality are understood here: of time and of supposits. Universality of supposits is either positive or negative. Positive universality actually has several inferiors; negative universality is such that, in itself, it is not incompatible—or is not conceived as incompatible—with several things being taken under it, whether it has several in reality or is incompatible with having them, provided that its formal concept does not present it as incompatible. This second mode of universality suffices for demonstration concerning God and the angels, although in reality they have no multiple inferiors, because they can be conceived after the manner of a common thing, as was said above, q. 8⁷. It does not apply to a singular term, such as Peter or Socrates, because their very mode of being conceived expresses singularity, which is incompatible with universality.

Similarly, universality of time is spoken of in two ways: in one way, simply and absolutely. “Absolutely” means throughout every difference of time, or perpetually; “restrictedly” means what belongs to something given its existence, either continuously—as blackness to a raven—or at certain times—as flowering to a tree.

Concerning the predicate per se, one must note that, speaking generally, it is what is opposed to a predicate per accidens. “Per accidens” sometimes means not essentially but accidentally and contingently. Likewise, per se sometimes means immediately, or not through another; at other times it means necessarily and essentially.

The Philosopher divides the predicate per se into four modes of per se predication, which St. Thomas gathers in Posterior Analytics I, lecture 10⁸, because the particle “through” signifies two things: either a relation of cause or a relation of position. A predicate is said to be per se when it is without another, or alone. When “through” signifies a relation of cause, it can signify an intrinsic cause, such as a formal or material cause: for example, when a body is said to live through the soul as through its form, or to be colored through a surface as through its material cause. Or it designates an extrinsic cause, especially an efficient cause, as when water is said to be heated through fire.

The first mode of per se, then, occurs when something is attributed to something through its form. Since a definition is maximally formal in every thing, when a definition or some quidditative predicate belonging to the definition is predicated of something, it is said to be predicated in the first mode per se.

The second mode of per se, St. Thomas says, is taken when the preposition “through” designates a material cause, that is, the proper subject of some form, which belongs to the form’s definition. This is the same as saying that when a proper attribute is predicated of its subject, it is the second mode per se. For although the predicate is not of the essence of the subject, the subject nevertheless belongs to the essence of the predicate, insofar as the proper subject is placed in the definition of the attribute.

The third mode per se is not a mode of predicating but of existing: something is said to be per se, that is, alone, as primary substance is said to exist per se. Concrete accidents, such as white or musical, are not understood to exist alone, but with the form that denominates and inheres in another, namely, a subject.

Finally, the fourth mode of saying per se is not a mode of predicating or of being, but of causing. Aristotle defines it: “When something belongs because of itself”; that is, when the proper rationale by which it is caused is signified. For example, “A builder builds,” but not “A builder walks,” because the art of building does not contribute to walking. Thus, in this fourth mode of saying per se, there is, as it were, an appeal to the proper form and rationale for causing something: what contributes to causing is signified, not what does not contribute, even if the latter is found there identically and materially. Hence it can well be that something belongs in the first mode per se but not in the fourth. Essential identity suffices for it to belong in the first mode, whereas the fourth requires the formal relation under which something belongs or is caused. Thus animal and rational are identified, and belong to a human being in the first mode per se; yet “Animal reasons” is not per se in the fourth mode, because “animal” does not signify the formal rationale for reasoning.

From these points you will gather that a predicate per se in the first or second mode must be predicated of every subject, because it belongs necessarily, and hence always and to all. The third mode per se applies only to a singular; the fourth applies to singulars and universals alike wherever the formal rationale of causing is signified.

BUT YOU WILL OBJECT: This proposition, “The risible is a human being,” is not in the first mode, yet the predicate is part of the definition of the subject. Therefore Aristotle wrongly reduces to the first mode per se what is placed in the definition, and then to the second mode when the subject belongs to the definition of the predicate.

The same argument can be made about indirect predications, which are not “of every” and consequently are not per se, yet one thing can be placed in the definition of another. For example, when I say, “Animal is human or rational,” “animal” is placed in the definition of the predicate; and likewise when one correlate is defined through another. Therefore it is not right to say that the first mode per se occurs when a definition or part of a definition is predicated of something.

Third: All contingent predicates will be per se in the second mode, because the subject is placed in the definition of any accident, even the fifth predicable, as is gathered from St. Thomas, De ente et essentia⁹. Therefore the definition of the second mode per se is not sound.

Finally, contingent predicates are placed in the fourth mode, as when one says, “Fire heats,” or “A builder builds.” For actions that belong contingently to agent causes and can be impeded are predicated. Therefore they are not predicates per se, since what belongs contingently belongs per accidens.

The answer to the first objection is that authors are divided concerning the proposition in which the subject of an accident is predicated of the accident itself. Some think it is absolutely not per se in the first mode, because the subject is not an intrinsic part of the accident and the proposition itself is indirect, and consequently is neither “of every” nor per se; for example, “White is a body,” “The risible is a human being.” Thus Cajetan, Posterior Analytics II, ch. 18, judges absolutely that these are propositions per accidens, although they are improperly called per se. He thus explains his own position, having said in ch. 3 that this proposition was per se because the subject is part of the accident’s definition, whether in the direct or oblique case.

Others say that the subject is predicated per se in the first mode of the accident, not absolutely, but by addition and extrinsically, because the quiddity of such an accident depends on its subject as on something extrinsic. Thus it is predicated per se in the first mode in the way it belongs to the definition. Hence, among examples of the first mode per se, Aristotle places a triangle, which consists of lines, and a line, which consists of points. A point, however, is predicated of a line only obliquely, as the term of a relation and the object of a power. When Aristotle says that a subject is predicated per accidens of an accidental form, “per accidens” is understood to mean through another, not accidentally. Thus, when a snub nose is called a nose, “nose” belongs to it through another, namely, through the human being who is its subject. Or, better, Aristotle is not speaking of a proper subject: a foreign and common subject in relation to the accident makes the proposition accidental. This is certain from the Philosopher’s text, as is clear in this book I, ch. 22¹⁰, lecture 33 in St. Thomas¹¹, where Aristotle gives as an example the statement, “White is wood.” The case is not similar when a proper subject is predicated of its attribute because of the necessary connection between them. This last solution seems sufficiently probable, although the first is not without probability. The answer to the argument, then, is that Aristotle placed in the second mode per se those propositions in which the predicate is the proper attribute of its proper subject, because such a subject belongs to the account of the predicate. If, however, the subject to which the proper attribute belongs is predicated of the attribute itself, the Philosopher did not say that this belongs to the second mode.

The answer to the second objection is that indirect predications are not excluded from the second mode per se simply because they are indirect, as some think, on the ground that an indirect proposition is not “of every,” which is required for a proposition to be per se. An objection is raised, however, from those indirect propositions in which the subject converts with the predicate, so that the predicate belongs to every subject, as when one says, “The risible is a human being,” or “The rational is a human being.”

Therefore one must say that among these indirect propositions, some are essentially identical, as when one says, “The rational is a human being.” In that case the proposition also belongs to the first mode by identity, because one extreme belongs to the essence of the other by identity. At other times they make a contingent proposition, as when I say, “Animal is a human being”; then it is in no mode per se and is not “of every,” because it is contingent. At other times the subject is predicated of its proper attribute, as in “The risible is a human being”; then it is in the first mode per se, because the predicate “human being” belongs to the definition and quiddity of the risible, at least by addition, just as a proper subject is placed in the definition of an accident, as is said in Metaphysics VII¹². In the same way, when a correlate is predicated of its correlate in the oblique case, as when one says, “A father is the father of a son,” it also belongs to the first mode per se obliquely and by addition, because in this way it belongs to its definition. This proposition, however, “Animal is rational or irrational,” is necessary and “of every” by reason of the disjunction, and is reduced to the second mode per se. For although rational and irrational are contingent differences, that disjunct functions as a proper attribute, since it is proper and per se to animal to be contracted by one of that disjunction.

The answer to the third objection is that common accidents, if they can be reduced to a proper subject—heat to fire, cold to water, for example—are then defined as proper attributes through such subjects, in the first mode per se by addition, as explained. If, however, they cannot be reduced to some proper subject, then they are not defined through such a subject insofar as concerns the rationale of their species, because they do not derive their species from that subject. The inherence of such an accident, however, or the accident insofar as it is capable of inhering, is defined through the subject, because aptitude for inhering is essential to an accident, though it belongs to it generically, not specifically.

The answer to the fourth objection is that a predicate per se in the fourth mode can be accidental in its predication, but it is per se in causing. Thus, “What is hot heats” expresses predication of the fifth predicable in the order of being, but causality per se in causing.

Concerning the predicate called secundum quod ipsum or universal, one must note, following the divine Thomas, Posterior Analytics I, lecture 11¹³, that “universal predicate” is not said in the way the five predicables are called universals in predication, but according to a certain totality and adequacy: neither is the subject found without it, nor it without the subject. This is to say that they convert with one another. For something can be predicated universally but not secundum quod ipsum, as when one says, “Every stone is colored.” The predication is universal, but not secundum quod ipsum, because a stone is not colored insofar as it is a stone, but insofar as it is a mixed body.

Therefore three conditions are required for a universal predicate, or one secundum quod ipsum: First, that it be “of every.” Second, that it be per se in the first or second mode, that is, belong necessarily. Third, that it be said convertibly, or insofar as it is such. For this reason the ultimate difference and its proper attribute belong per se in the first mode and secundum quod ipsum. Thus “A human being is rational or risible” belongs per se in the first mode and by reason of the human being himself. If, however, one says, “A human being is an animal or sentient,” it does not belong per se in the first mode, although it belongs essentially, because it also belongs to others and consequently by reason of a common nature—not convertibly, but secundum quod ipsum.

This predicate is therefore defined as “what is predicated of every subject, per se, and as that which it itself is.” By per se, I mean in the first or second mode of per se predication, which alone are modes of predication. Thus “of every,” per se, and secundum quod ipsum stand as higher and lower: every predicate secundum quod ipsum is per se, but not conversely; every predicate per se is “of every,” but not conversely.