What is a Goat-Stag?: Aristotle on Definitions and Non-Existent Kinds
In II.7-10 of the Posterior Analytics, Aristotle discusses various characteristics of definitions, and a question which has garnered attention is over the possible place and function of “nominal definitions.” Some, such as Robert Bolton, argue that Aristotle does include nominal definitions. This comes largely from the view that the kind of definition Aristotle lists at the beginning of II.10 (93b29), which is described as “an account of what a thing is,” (Ὁρισμὸς δʼ ἐπειδὴ λέγεται εἶναι λόγος τοῦ τί ἐστι), refers to the third kind of definition Aristotle gives at 94a14 which is “the conclusion of a syllogism which demonstrates the essence.”
This view leads to the conclusion that there cannot be a definition of kinds which do not exist. My intention is to argue against this interpretation by showing that, at 93b29, Aristotle is speaking of “definition” in a broader sense than he does at 94a14.
I will first give a brief account of the relevant sections of II.7-10. Next, I will directly address the question of whether or not Aristotle describes nominal definitions in II.10, and if so, how the kind of definition which he speaks of at 93b29 relates to the three kinds he gives at 94a11-14. Then, I will consider how Aristotle would use a nominal definition in his discussion of discovering the definition of both an eclipse and thunder. Finally, I will argue that, due to the interpretation of Aristotle already given, Aristotle would allow that there be a type of definition for non-existent kinds that is merely the meaning of a word. This type of definition, however, is not found in the stricter kind of definition Aristotle gives at 94a11-14. To begin, let us examine the relevant texts in Posterior Analytics II.7-10.
Overview of Posterior Analytics II.7-10
One of the first points relevant for our purposes is Aristotle’s statement that “Anyone who knows what ‘man’ or any other thing is must also know that it is; because no one knows what a non-existent thing is.” Thus we see the beginning of this notion that in order to know the definition (or τι εστι) of a thing, it must first exist. This point is immediately followed by the statement that one, “may know the meaning of a phrase, or of a name if, e.g., I speak of a [goat-stag].” The distinction that Aristotle clearly makes is between how we refer to kinds of things which exist, and those which do not. We see that it is impossible for us to come to know what a non-existent kind is, simply because it does not exist. In II.7, Aristotle also states that it is by a demonstration that we come to know the existent of a kind but that definitions are not discovered in this way. We can say that we come to know that something is by a demonstration, but what itis must be found in some other way.
It is in II.8 that Aristotle gives examples of coming to know the existence of a thing. He states that we can come to know, “a thing’s existence accidentally or because we have some grasp of the thing itself . . ..” As an example of when someone has this “grasp of the thing itself,” Aristotle gives two ways in which someone becomes aware of eclipses. He states first the three terms: A is eclipse, B is obstruction by the Earth, and C is Moon. He then states that if we find that there is a B, we come to know that there is an eclipse. This happens by applying A to B, and then B to C. As Aristotle states, this syllogism relies on us finding B. The second syllogism takes A to also refer to eclipse, C to moon, and this time B is “the inability of the moon at its full to cast a shadow, there being nothing visible in the way.” In the same way as the previous example, when we apply A to B, and B to C, we know that there is an eclipse. We also learn that in the first syllogism, the cause of an eclipse is discovered and therefore we know that there is an eclipse and we know why. In the second instance, however, we only know that there is an eclipse, because the inability to cast a shadow is actually an effect of an eclipse, and not the eclipse itself. In the first case, the obstruction by the Earth is the reason why there is an eclipse, and thus we know the cause of the eclipse as well as that there is an eclipse. Aristotle then applies this same treatment to thunder. He comes to know what thunder is and why thunder is by using a syllogism consisting of three terms: cloud, thunder, and extinction of fire. In another way, Aristotle uses a syllogism consisting of cloud, thunder, and noise instead of extinction of fire. It is clear that, in the first case, Aristotle is able to show what thunder is (extinction of fire) and why (fire is extinguished in clouds)it thunders, whereas in the second case, one can only know that there is thunder (a noise in the clouds). As Aristotle summarizes, it is in this way that “the essence is apprehended, and becomes known to us . . .” In II.9, Aristotle restates the view that although we cannot demonstrate the nature of a thing (τί ἐστί), the nature is shown through a demonstration.
It is in II.10 that Aristotle states what kinds of definitions one can have of a thing. This is one of the most critical passages for our purposes. At 93b29, Aristotle states that, “Since definition means ‘an account of what a thing is,’ obviously one kind of definition will be an explanation of the meaning of the name . . . .” Mure translates this opening phrase: “Since definition is said to be the statement of a thing’s nature, obviously one kind of definition will be a statement of the meaning of the name . . . .” Barnes has a similar rendering: “Since a definition is said to be an account of what something is, it is clear that one type will be an account of what its name, or some other name-like account, means . . . .” Charles has a slightly different translation: “Since a definition is said to be an account given in reply to the “What is—?” question, it is clear that one kind of definition will be an account given in replay to the question “What is it that a name or other name-like expression signifies.’” Although this translation is rather awkward compared to the other two, the important section for my purpose is the beginning line of “Since a definition is said to be . . . .” Bolton does not seem to give an explicit translation of this text, but he does offer a rendering of this opening statement as “one sort of account of what something is.” I will later take up the significance of this statement, but for now it is enough to point to this particular phrase and how these various scholars translate it. I especially take note of the respective rendering of Bolton and Charles, since it is by means of the latter’s interpretation that I will attempt to argue against the interpretation of the former. At 94a11-14, Aristotle summarizes the different kinds of definitions by stating “in one sense definition is an indemonstrable account of the essence; in another is a logical inference of the essence, differing from demonstration in grammatical form; and in a third is the conclusion of the syllogism which demonstrates the essence.” As I will show later, it is this third kind of definition (“the conclusion of a demonstration”) that Bolton relates to the kind of definition (which I will henceforth refer to as “nominal definition”) listed at 93b29, and it is this specific connection that I intend to dispute. I have now given a brief overview of the important texts from which I will draw. Now I will present Bolton’s interpretation of this passage and his argument that nominal definitions are connected with the kind of definition referred to as a “conclusion of a demonstration” listed at 94a14.
Nominal Definitions and II.10
Bolton, as was said, connects nominal definitions listed at 93b29 (“meaning of a name”) with the kind of definition listed at 94a14 (“conclusion of a syllogism”). He states that, “At 94a7-9 the first of them [propositions whose grasp involves a non-accidental knowledge of the existence of something] is offered as a specimen of the type of definition which can function as the conclusion of a demonstration. It is evident from the summary account of the types of definitions given at 94a11ff. that this is the nominal definition.” In a footnote, Bolton states that the conclusion of a syllogism is the only kind of definition listed at 94a11-14 that could correspond to a nominal definition. He acknowledges that some have instead posited four kinds of definitions. Bolton disagrees with this interpretation because it “has against it the fact that in his summary of the three types of definition discussed in the chapter (ἄρα, 94a11ff.) Aristotle lists only three, and at 75b31-32 he claims that there are only three types.” Thus, we can see that one fact motivating Bolton to connect nominal definitions with conclusions of demonstrations is the fact that in II.10 as well as I.VIII (75b31-32) of the Posterior Analytics, Aristotle only lists three kinds of definitions. He does not list four, so why would we add kinds of definitions to Aristotle’s list when he only gives us three?
Bolton also argues that non-existent kinds cannot have nominal definitions. He states, “ the characteristics which we use to nominally define some entity are ones which can, eventually at least, be demonstrated to belong to it. Given this we can only have nominal definitions of the names of entities which are genuine kinds whose nature we can uncover . . . .” If we simply consider this point apart from Aristotle’s own words, it seems reasonable. If something does not exist, we will be unable to find its nature, because non-existent things do not have natures. Bolton adds further support to this point by saying that “syllogisms which have nominal definitions as conclusions give the reason for and, hence, prove the existence of the kinds defined, unless positing the existence of the kinds has already been required without proof.” Thus, if we follow Bolton’s line of interpretation, we can see that a non-existent kind will not have any cause to its existence or nature because it does not exist and therefore has no nature. Bolton, then, is consistent in maintaining that there is no nominal definition of things which do not exist. He then also concludes that in all of chapter 10, there are only three kinds of definitions listed and not four.
It is important also to see how Bolton addresses Aristotle’s words at 92b5-8 in which he states that we can know the meaning of a name such as goat-stag. Bolton’s interpretation of this is important. Concerning this passage, Bolton states that, “Aristotle does not in fact say there that we can have a definition of what ‘goat-stag’ signifies but only that we can know what it signifies. This leaves open what the other evidence requires, that the account of what it signifies is not even a nominal definition.” Let us suppose that we attempt to define goat-stag, or centaur, or Thor. In the first case we might say, “A goat-stag is a creature with the body of a goat and the head of deer.” We might also say that “a centaur is a creature with the upper torso of a man and the lower body of a horse,” or that “Thor is a god who wields a mighty hammer.” On Bolton’s interpretation, we would be forced to say that, according to Aristotle, none of these definitions determine the meaning of the name “goat-stag,” “centaur,” or “Thor.” This seems a rather odd conclusion. It is awkward to argue that we cannot have a definition of the word “centaur.” Is this in fact the case? Does Aristotle really believe that we cannot have even a definition of a name without any thought as to whether or not they exist? If Bolton’s interpretation is correct, this would be the case. If Aristotle does mean to connect nominal definitions found at 93b29 with the “conclusion of a syllogism” at 94a14, and we cannot have definitions of things which do not exist (92b5), then we cannot call “god who wields a mighty hammer” as the nominal definition of “Thor,” or “creature with the upper torso of a man and the lower body of a horse” as the nominal definition of “centaur.” This, however, is unreasonable. In order to expose the error, I will now turn to Charles and his more accurate interpretation of II.10.
Charles, in his Aristotle on Meaning and Essence, argues for a reading of Aristotle that includes definitions in a strict sense (which can only be of things which exist), and in a broader sense (which would allow definitions of names which signify non-existent kinds). Charles states:
The first sentences of B.10 contain some indication that Aristotle is making the liberalizing moves required if he is to take answers to the “What does ‘F’ signify?” question as definitions and as proper answers to the “What is . . . ?” question. To do so, he needs to introduce some relevant wider use of the term “definition” and that what is . . . ? question. But this is precisely what is suggested in the first sentence of B.10, where Aristotle notes that definitions and answers to the “What is . . . ?” question are said to be coterminous. For, in this context, the use of “is said” may be taken to signal a relaxation of the standard constraints on definition. Indeed, it may invoke some popular view about what definition is.
Whereas Bolton argued that non-existent kinds do not have even a nominal definition because Aristotle states that there cannot be a definition of a thing which does not exist, Charles, differentiates between Aristotle speaking of definitions in the strict sense and where he may allow for a broader understanding of the term “definition.” Using definition in a broader sense at 93b29 would allow for a definition of “goat-stag,” “centaur,” and “Thor.” Since this connection, or disconnection, is so vital to the interpretations of both Bolton and Charles, we must carefully examine this specific portion of Aristotle’s text and ask whether Charles is correct in saying that Aristotle is referring to definitions that are said to be or called definitions in a looser sense.
Aristotle states in the first line of II.10,
Ὁρισμὸς δʼ ἐπειδὴ λέγεται εἶναι λόγος τοῦ τί ἐστι, φανερὸν ὅτι ὁ μέν τις ἔσται λόγος τοῦ τί σημαίνει τὸ ὄνομα ἢ λόγος ἕτερος ὀνοματωδης, οἷον τὸ τί σημαίνει [τί ἐστι] τρίγωνον.
This is the critical section which Charles argues relaxes the strict sense of definition and allows for a looser sense of “definition” of non-existent kinds. I earlier gave the various ways that this first line has been rendered into English, and it would be appropriate to return to those now to notice the critical differences in translation. Charles translates this opening line: “Since a definition is said to be an account given in reply to the ‘What is—?’ question, it is clear that one kind of definition will be an account given in reply to the question ‘What is it that a name or other name-like expression signifies?’” Bolton does not offer an explicit translation of this section, but it appears that he would translate this opening line as “one sort of account of what something is.” Mure has a similar way of translating the opening line of II.10 by rendering it as: “Since a definition is said to be the statement of a thing’s nature, obviously one kind of definition will be a statement of the meaning of the name,” and Barnes as: “Since a definition is said to be an account of what something is . . . .” Tredennick offers a slightly different translation than the others by rendering this opening phrase as: “Since definition means ‘an account of what a thing is,’ obviously one kind of definition will be an explanation of the meaning of the name . . . .”
The critical phrase in Aristotle’s passage is “Ὁρισμὸς δ̓ ἐπειδὴ λέγεται εἶναι.” As we saw, Charles’ argument for a looser understanding of definition relies heavily on the phrase “said to be.” This is a very accurate way of translating the phrase λέγεται εἶναι, which could be translated as “said to be” or “called.” The other scholars seem to agree in this translation. Barnes and Mure both translate this phrase as “Since a definition is said to be.” Tredennick offered a slightly different wording by saying that “Since definition means.” This use of “means” instead of “said to be” changes the interpretation of this passage to favor Bolton’s interpretation that Aristotle is still following a strict understanding of definition. However, the better translation is “said to be” or “called,” rather than “means.” Once again, I cannot find anywhere that Bolton explicitly renders this phrase into English. The closest he seems to come is when he states that, “A nominal definition is, furthermore, one sort of account of what something is.” In order for Bolton to maintain that this nominal definition is connected with the “conclusion of a syllogism” (94a14), he would have to place a great deal more emphasis on the “what something is” (τί ἐστι), and less on the “sort of account” (which seems to be the closest he comes to explicitly rendering λέγεται εἶναι, and may have less of a sense of broadening the scope of “definition”).
Since the translation of λέγεται εἶναι is best rendered as “said to be” or “called,” it is correct to conclude with Charles that in this text Aristotle allows for a broader sense of “definition.” This sense would allow us to call certain accounts “definitions” which would not be so-called in the strict sense which Aristotle uses implying“definitions” of non-existent things in a certain sense. Charles also states that,
While it is correct that there are only three types of definitions of kinds in B.10, there are more than three types of definition. [The mistake of those who only see three types of definitions] was to fail to see that the first type of definition (the one connected with signification) is of a fundamentally different type from those that follow.
The three types of definitions listed at 94a11-14 are of a different type than that which is listed at 93b29, falling under the more strict type of definition that Aristotle gives of kinds which exist, as Aristotle states earlier at 92b3-5 when he posits that no one can know the nature of a non-existent kind. This separation broadens the text so that the name of a non-existent kind, e.g. “goat-stag,” “centaur,” and “Thor,” would fall under nominal definitions at 93b29. In order to get a deeper insight into how Aristotle might use nominal definitions and what kinds of things are included under them, it would be helpful to look at examples of when Aristotle might use nominal definitions in discovering the existence and nature of kinds. In order to do this, I shall turn to Aristotle’s treatment of how one might discover the nature of an eclipse and of thunder.
Function of Nominal Definitions in II.8-10
Aristotle gives two examples of finding the existence of an eclipse at 93a29-93b8. He states that these are instances when we come to find “something of whose essence we have some grasp . . ..” In the first example, Aristotle gives three terms: eclipse, moon, and “obstruction by the earth.”42 He then says that we seek to know if there is an “obstruction by the earth” when we seek if there is an eclipse. A syllogism of this could be rendered:
An obstruction by the earth is an eclipse.
The moon is obstructed by the earth.
The moon is eclipsed.
In the second syllogism, Aristotle keeps eclipse and moon, but instead of “obstruction,” uses “the inability of the moon to cast a shadow, there being nothing visible in the way.”43 Thus, the second syllogism could be written in this way:
Inability to cast a shadow is an eclipse.
The moon cannot cast a shadow with nothing in the way.
The moon is eclipsed.
Bolton does not address these arguments in great detail, except to say “All of the examples are taken in 93a29ff. To have this defining feature of nominal definitions. So the grasping of these nominal definitions involves the apprehension of the existence of something conforming to them.” Bolton refers to the nominal definition of “eclipse,” which in the first case is “obstruction by the earth,” and in the second “inability to cast a shadow.” From Bolton’s interpretation, we would say that when we “grasp” these nominal definitions of an eclipse, then we will know that eclipses exist. Bolton would have to argue that giving the above definitions as explanations of what we mean by “eclipse,” at least before we know that an eclipse exists, would be an example of us having a nominal definition of an eclipse but not knowing that this is a nominal definition.46 Either this is the case, or our once “non-nominal definition” becomes a nominal definition when we are aware that the kind exists. What would help is to notice how Bolton interprets Aristotle’s arguments regarding thunder.
Aristotle gives a syllogism concerning thunder at 93b8-14. Once again, he offers three terms: A stands for thunder, B for “extinction of fire,” and C for cloud. Aristotle formulates this syllogism in the following way:
Extinction of fire is in the clouds.
Thunder is the extinction of fire.
Thunder is in the clouds.
J.L. Akrill sees another syllogism at 94a7-9:
Noise in the clouds is the extinction of fire.
Extinction of fire in the clouds is thunder.
Thunder is a noise in the clouds.
This second syllogism takes place in the context of Aristotle distinguishing between knowing what thunder is and why there is thunder. Aquinas comments on this section by saying that:
When we say that it thunders because fire is extinguished in a cloud, we signify in the manner of a continuous demonstration, i.e., not dividing the demonstration into distinct propositions, but taking all its terms continuously, whereas when we say that the thunder is the sound of fire being extinguished in a cloud, we signify in the manner of a definition. But if we say that thunder is a sound in the clouds, making no mention of the extinction of fire, we will have a definition signifying what thunder is. This definition will be only the conclusion of a demonstration.
Aquinas shows that saying thunder is a “sound in the clouds” is a definition. This kind of definition could be found in the conclusion of a demonstration. This should immediately remind us of Aristotle’s list of definitions at 94a11-14, and specifically at 94a14, where he describes the third kind of definitions as a “conclusion of the syllogism which demonstrates the essence.” The Greek phrase of Aristotle at 94a14 τῆς τοῦ τί ἐστιν ἀποδείξεως συμπέρασμα (which Tredennick translates as “conclusion of the syllogism which demonstrates the essence”) is identical to the phrase he writes at 94a8. In the latter case, Tredennick translates this as “whereas thunder can be defined as a noise in the clouds, which is the conclusion of the syllogism that demonstrates the essence.”
This seems to support Bolton when he argues that, “A nominal definition is the conclusion of a demonstration which displays the essential nature.” Bolton further argues that, “Since the definition ‘Thunder is a noise in the clouds’ tells us what the term ‘thunder’ signifies, it must both point to the essence of thunder and specify the reference of the term ‘thunder.’” For Bolton, in order for us to “grasp” the meaning of the term “thunder,” we must have some knowledge of the essence of thunder and thus be able to know that thunder exists. This would certainly require more in the way of demonstration. For how can we be said to have anything of the essence of a kind if we merely know the meaning of a name? Does defining the word “thunder,” however, mean the same thing as using a demonstration which gives us something of the essence? Clearly, “noise in the clouds” seems very much like a nominal definition of thunder. Does this mean that the conclusion of a syllogism at 94a14 corresponds to the meaning of a term at 93b29?
Charles addresses this specific point. He asks, “How is this [definition as the conclusion of a syllogism] related to the account of what the name ‘thunder’ signifies? Are they the same? They certainly appear to use the same terms.” Charles’s ultimate response to this question is that a major difference between these two instances of definitions—nominal definitions and the kind which can be the conclusion of syllogism—is that nominal definitions deal merely with the meaning of a particular word, whereas the conclusion of a syllogism serves to define an actual existing kind. He states that, “the two types of definition cannot be identical because their aims and relations are different: the first concerns the significance of ‘thunder,’ the second the nature of thunder.” Charles argues that nominal definitions are concerned merely with what a particular word means. So that the person who has only a nominal definition of thunder knows what “thunder” means but does not yet know if there are actual kinds which correspond to that word. Nominal definitions, unlike those definitions that are conclusions of syllogisms, are not connected to a demonstration.
Aristotle states: “No demonstration can prove that a given name has a given meaning.” The distinction between definitions, which signifies the meaning of words, and those that deal with actual kinds finds support in Aristotle’s text. After discussing the account of definition which is the signification of a word (nominal definition), Aristotle states that, “The above is one definition of definition; but in another sense definition is a form of words which explains why a thing exists.” Thus, we see a real change from the one kind of definition to another kind. This allows us to conclude that there are two categories of definitions: the first is merely the signification of a word, while the second is concerned with a real existing kind: this second type of definition is then broken down into three subsets of definition, one of which can be the conclusion of a demonstration. What if the two kinds of definitions use the same words? Charles states that, “Even if some of the same terms are used (‘noise in the clouds’) . . . they are used [in nominal definitions] to make an indefinite claim about a type of noise (tis), while in the second they uniquely identify the phenomenon in question.”
Thus, nominal definitions allow a starting point for a science. Clearly, a science seeks to gain scientific knowledge about kinds. In doing this, we come to find definitions of kinds. As we previously noted, there cannot be a definition of a non-existent kind because definitions are meant to be part of a science, and we cannot have scientific knowledge of something that does not exist. If we wish to have knowledge of real things, we must first know that they exist, and nominal definitions give us the ability to begin a search for the particular kinds we wish to investigate. As Bolton rightly states, “the chief function of nominal definitions is to enable scientific inquiry to get off the ground.” The function of a nominal definition is to help the investigation of kinds. Once we know that a particular kind exists, however, we move beyond the realm of nominal definitions, and are then working with what may be called a “real definition.”
Before concluding, I should like once more to speak about the relationship between nominal definitions and non-existent kinds in Aristotle. I do this because I wish to emphasize the awkward conclusion that we are forced to draw if we accept Bolton’s interpretation of nominal definitions in Aristotle.
Nominal Definitions and Non-Existent Kinds
One of the first difficulties that we encounter, should we take Bolton’s interpretation of Aristotle, is the problem of how to interpret Aristotle’s own words at 92b5-8. At this point, Aristotle states that one, “may know the meaning of a phrase, or of a name, e.g., . . . a [goat-stag]; but it is impossible to know what a [goat-stag] is.” Aristotle is clearly describing some kind of understanding of the term “goat-stag.” Is this a nominal definition? Could a statement such as “a creature with the head of a goat and the body of a stag” serve as a nominal definition of “goat-stag”? Clearly not for Bolton. He states, “Aristotle does not in fact say there that we can have a definition of what ‘goat-stag’ signifies but only that we can know what it signifies. This leaves open what the other evidence requires, that the account of what it signifies is not even a nominal definition.” Bolton is quite correct when he states that Aristotle does not call the meaning of a name such as “goat-stag” a definition. In fact, at this point, Aristotle is attempting to say the exact opposite. He is trying to show that we can only know the essence or nature (τί ἐστί) of kinds which actually exist. If we remember, however, that Aristotle calls nominal definitions a sort of definition (λέγεται εἶναι), we can allow for a broadening sense of definition, and this would include the meaning of words and thus non-existent kinds. As Charles states, “The first [kind of definition] concerns an account of what terms signify, the last three definitions of things or kinds.” Thus, the kind of “definition” that we would have of a goat-stag is not a true definition in the strict sense which Aristotle presents, but in an extended sense.
This textually based difficulty with Bolton’s interpretation leads to a second difficulty which is the more awkward consequence. When beginning a scientific investigation, we would ask what a particular type of thing is. Let us return to our previous examples of a “goat-stag,” a “centaur,” and the figure of “Thor.” If we want to discover whether any of these exist, we would naturally ask what is meant by each name. If we do not accomplish even this preliminary step, we would be unable to determine if the kind we are searching for exists because we will not recognize it when we see it. Bolton, however, denies that “a Norse god who wields a lightning-hammer” is an example of a nominal definition, i.e. a definition of the name “Thor.” Bolton does state that, “we can know [only] what it signifies.” It seems that this is the exact same thing as saying that a nominal definition is an account of the meaning of a name. Aristotle does not actually use the term “nominal definition,” so we must actually apply a nominal definition to “nominal definition.” The word “nominal” seems to suggest that we are dealing with only what a word means, independent of whether or not a sort of thing corresponding to this word actually exists in reality. This is why it seems rather awkward and strange to say that we can have no nominal definition of “goat-stag,” “centaur,” or “Thor.” The fact that “nominal definition” implies that we are working more with words and not with kinds naturally leaves out the condition that such definitions be of kinds which actually exist. Thus, we are naturally confused when we are told that we cannot have definition of the name “goat-stag.” Bolton states that Aristotle denies, “that we can have a definition of what ‘goat-stag’ signifies” (second italics are my emphasis). I would agree that Aristotle denies this, because a “goat-stag” is strictly speaking nothing. However, this does not mean that “goat-stag” cannot have some other signification. We may all be aware that “goat-stag” signifies a mythical creature made up of a goat and a deer, or that a “centaur” is a similarly mythical creature that is made of the upper torso of a man and the body of a horse. Thus, Bolton is incorrect in saying that the name “goat-stag” can signify nothing absolutely because there is no kind which exists in reality that corresponds to what “goat-stag” signifies.
To comment or respond to the above essay, please send an email to : email@example.com
1. The term “nominal definition” does not actually appear in Aristotle, but this does not mean that there is not something in his work which could be given this name.
2. My primary focus will be a response to Bolton, Robert, “Essentialism and Semantic Theory in Aristotle: Posterior Analytics, II, 7-10,” The Philosophical Review 85:4 (Oct. 1976): 514-544. For others who take this same view, cf. Demoss, David and Daniel Devereux, “Essence, Existence, and Nominal Definition in Aristotle’s ‘Posterior Analytics’ II 8-10,” Phronesis 33:2 (1988): 133-154.
3. Unless otherwise noted, all translations of Aristotle are taken from Posterior Analytics, trans. by Hugh Tredennick (Cambridge and London: Harvard University Press, 1960). Above found on 207, 209.
4. 92b4-6. The Greek text is taken from Tredennick’s edition. “Ἔτι πῶς δείξει τὸ τί εστιν; ἀνάγκη γὰρ τὸν εἰδότα τὸ τί ἐσιτν ἄνθρωπος ἢ ἄλλο ὁτιοῦν, εἰδέναι καὶ ὅτι ἔστιν.”
5. 92b5-8. “τὸ γὰρ μὴ ὂν οὐδεὶς οἶδεν ὅτι ἐστίν, ἀλλὰ τί μὲν σημαίνει ὁ λόͅγος ἢ τὸ ὄνομα, ὅταν εἴπω τραγέλαφος, τί δʼ ἐστὶ τραγέλαφος ἀδύνατον εἰδέναι.” Tredennick actually renders this term (τραγέλαφος) as “unicorn.” Bolton in his “Essentialism and Semanitc Theory” uses the term “goat-stag,” as does Jonathan Barnes in his translation Posterior Analytics (Oxford: Clarendon Press, 1994), 55, as does G.R.G Mure, Posterior Analytics, in Great Books of the Western World, ed. by Mortimer J. Adler, vol 7: Airstotle I, (Chicago: Encyclopedia Britannica, Inc: 1952), 126, and David Charles Aristotle on Meaning and Essence, (Oxford: Clarendon Press, 2000), 25. I have chosen to use the translation “goat-stag” because the two main commentators I will be drawing from—Bolton and Charles—use this term. In a note on 196 of Tredennick’s translation, he notes that the literal translation is “goat-deer.” The important point is that this is a fictional being that does not exist. Aristotle is attempting to contrast the definition of a real kind with an obviously non-existent kind.
6. 92b12, 37.
7. 93a21-22. “τὸ δʼ εἰ ἔστιν ὁτὲ μὲν κατὰ συμβεβηκὸς ἔχομεν, ὁτὲ δʼ ἔχοντές τι αὐτοῦ τοῦ πράγματος . . . ”
8. 93a29-31. “Ὧν οὖν ἔχομέν τι τοῦ τί ἐστιν, ἔστω πρῶτον μὲν ὧδε. ἔκλειψις ἐφ̓ οὗ τὸ Α, σελήνη ἐφ̓ οὗ Γ, ἀντίφραξις γῆς ἐφ̓ οὗ Β.”
10. 93a37-39. “σελήνη Γ, ἔκλειψις Α, τὸ πανσελήνου σκιὰν μὴ δύνασθαι ποιεῖν μηδενὸς ἡμῶν μεταξὺ ὄντος φανεροῦ, ἐφ̓ οὑ̄ Β.”
12. St. Thomas Aquinas Commentary on Aristotle’s Posterior Analytics trans by Richard Berquist (Notre Dame, Indiana: Dumb Ox Book, 2007), II.7.
15. 93b15. “Ὡς μὲν τοίνυν λαμβάνεται τὸ τί ἐστι καὶ γίγνεται γνώριμον, εἴρηται, ὥστε συλλοψισμὸς μὲν τοῦ τί ἐστιν οὐ γίγνεται οὐδ̓ ἀπόδειξις, δῆλον μέντοι διὰ συλλογισμοῦ καὶ δἰ ἀποδείξεως.”
17. 93b29. “Ὁρισμὸς δ̓ ἐπειδὴ λέγεται εἶναι λόγος τοῦ τί ἐστι, φανερὸν ὅτι ὁ μέν τις ἔσται λόγος τοῦ τί σημαίνει τὸ ὄνομα . . . ”
18. Mure, 128.
19. Barnes, Posterior Analytics, 58.
20. Charles, Meaning and Essence, 23.
21. Bolton, “Essentialism,” 524.
22. 94a11-14. “̓́Εστιν ἄρα ὁρισμὸς εἷς μὲν λόγος τοῦ τί ἐστιν ἀναπόδεικτος, εἷς δὲ συλλογισμὸς τοῦ τί ἐστι πτώσει διαφέρων τῆς ἀπόδειξεως, τρίτος δὲ τῆς τοῦ τί ἐστιν ἀποδείξεως συμπέρασμα.”
23. Bolton, “Essentialism,” 521.
24. Ibid, 522.
25. Ibid, n. 14.
27. Ibid, 523.
28. Ibid, 524. Demoss and Devereux make this same claim in “Essence, Existence, and Nominal Definition.” They state that, “Aristotle’s argument [at 92b26-30] has the form of a reduction: if a definition is in no way of the ‘what it is,’ then it would be possible to have definitions of non-essences and non-existents, and this is absurd. So a definition must be an account of what something is; it cannot be simply an account of what a name means or signifies. Later in chapter 10 we are told that one kind of definition of what something is is an account of what a name signifies. It is possible, then, to have an account of what a name signifies which is not merely that—it also tells us what something is. Such an account, as we have seen, is a nominal definition,” 140. We see that Demoss and Devereux follow the same opinion with Bolton in stating that nominal definitions are only of kinds which exist.
29. Bolton, “Essentialism,” 524.
30. Charles, Meaning and Essence, 32-33.
31. 93b29-31, taken from Tredennick, 206.
32. Charles, Meaning and Essence, 23.
33. Bolton, “Essentialism,” 523-524.
34. Mure, Posterior Analytics, 128.
35. Barnes, Posterior Analytics, 58.
36. Tredennick, Posterior Analytics, 207.
37. Italics are my emphasis.
38. Tredennick, Posterior Analytics, 207.
39. Bolton, “Essentialism,” 523-524.
40. Charles, Meaning and Essence, 33-34.
41. Ibid, 47. For a similar view of this, cf. McKirahan, Richard D, Principles and Proofs: Aristotle’s Theory of Demonstrative Sciences (Princeton, NJ: Princeton University Press, 1992), 198-208.
42. 93a29-30. “Ὧ οὖν ἔχομέν τι τοῦ τί ἐστιν, ἔστω πρῶτον μέν ὧδε.”
45. Bolton, 522.
46. Barnes argues that, at this point, “Aristotle’s examples . . . can scarcely be interpreted as ‘nominal definitions’; and B 10, 93b32-5, seems to set possessors of nominal definitions in the opposite camp, among ‘incidental’ knowers.” Aristotle’s Posterior Analytics, 218. Barnes states earlier that having a nominal definition is that, “we know that there is an eclipse, knowing what the word ‘eclipse’ means.” For Barnes, then, we would know a thing exists when we know its nominal definition. This is a very plausible interpretation of Bolton, although I see it more as coming from the other direction: we come to know the nominal definition when we come to know that the kind defined exists. As stated above, our disagreement is the consequence of Bolton’s interpretation.
47. The text in Bolton which implies this argument as a consequence of his thought is in “Essentialism” p.523, where he states in reference to nominal definitions having existential import, “By this [Aristotle] means that the characteristics which we use to nominally define some entity are ones which can, eventually at least, be demonstrated to belong to it.” I have already quoted from this passage, but it is worth repeating. For a similar view, cf. Sorabji, Richard, “Definitions: Why Necessary and in What Way?” in Aristotle on Science: The Posterior Analytics, edited by Enrico Berti, (Padova: Editrice Antenore, 1981), 213-219.
48. St. Thomas Aquinas, Commentary, II.7.
49. Ackrill, J.L. “Aristotle’s Theory of Definition: Some Questions on Posterior Analytics II 8-10,” in Aristotle on Science: The Posterior Analytics, ed. by Enrico Berti (Padova: Editrice Antenore, 1981), 360.
50. St. Thomas Aquinas, Commentary, II.8. “Nam cum dicitur tonat propter hoc quod extinguitur ignis in nube, significatur per modum demonstrationis continuae, idest non distinctae per diversas propositiones; accipiuntur tamen continue omnes termini demonstrationis. Cum autem dicitur quod tonitruum est sonus extincti ignis in nubibus, significatur per modum definitionis. Sed si dicamus quod tonitruum est sonus in nubiibus, nulla mentione facta de extinctione ignis, erit definition significans quid est, et erit solum demonstrationis conclusio.” Opera Omnia, Leonine, vol. 1, Posteriorum Analyticorum (Rome, 1882).
51. 94a14. “τρίτος δὲ τῆς τοῦ τί ἐστιν ἀποδείξεως συμπέρασμα.”
52. Tredennick, 209.
53. 94a8. “ἔτι ἐστίν ὅρος βροντῆς ψόφος ἐν νέφεσι. τοῦτο δ̓ ἐστὶ τῆς τοῦ τί ἐστις ἀποδείξεως συμπέρασμα.”
54. Bolton, “Essentialism,” 521.
55. Ibid, 529.
56. Charles, Meaning and Essence, 45.
57. Ibid, 44.
58. 92b33. “ἔτι οὐδεμία ἀπόδεξις ἀποδείξειεν ἂν ὅτι τοῦτο τοὔνομα τουτὶ δελοῖ.”
59. 94a1-3. “Εἷς μὲν δὴ ὅρος ἐστὶν ὅρου ὁ εἰρημένος, ἄλλος δ̓ ἐστίν ὅρος λόγος ὁ δηλῶν διὰ τί ἔστιν.”
60. Barnes, in his commentary, writes that, “The older commentators explain that the sort of definition distinguished at 93b29-37 is nominal definition, and that Aristotle deliberately omitted this scientifically uninteresting notion from the enumeration at 94a11-14 (222)”. It does not seem that this is Barnes’s personal interpretation, but it does deserve comment. It is not that Aristotle omitted nominal definitions from the list at 94a11-14 because they are “scientifically uninteresting,” but that he was listing his view of definition in the strict sense. This sort of definition can only be of a kind which exists. Nominal definitions, listed at 93b29, are very important for the scientific process (as Bolton, 521and Aquinas, II.8 both state) to even begin. They do not, however, fit the strict criteria of definition, and for this reason Aristotle did not include them at 94a11-14.
61. Charles, Meaning and Essence, 45.
62. Bolton, “Essentialism,” 521.
64. 92b5-8. “τὸ γὰρ μὴ ὂν οὐδεὶς οἶδεν ὅτι ἐστίν, ἀλλὰ τί μὲν εἰ δείξει τί μέν σημαίνει ὁ λόγος ἢ τὸ ὄνομα, ὅταν εἴπω τραγέλαφος, τί δ̓ ἐστὶ τραγέλαφος ἀδύνατον εἰδέναι.”
65. Bolton, “Essentialism,” 524.
66. Charles, Meaning and Essence, 47. Charles gives a very fascinating analysis of the process he believes that Aristotle is presenting in Posterior Analytics II.8-10. Charles argues that Aristotle presents a “Three-Stage View” of investigating a particular kind (24).
Stage 1: Understanding the meaning of a name.
Stage 2: Knowing that a kind exists which corresponds to that name.
Stage 3: Discovering the essence of the kind which corresponds to the name.
I do not comment on whether or not Charles is correct in his analysis and model, but his views do seem to have merit. It seems accurate to state that Aristotle would first begin with the meaning of a name, then discovering if a kind with that name exists, and finally investigation into the essence or nature of that kind. I do believe that this model fits the spirit of scientific investigation that Aristotle proposes in the AnPo.
67. Bolton, “Essentialism,” 524.
Ackrill, J. L. "Aristotle’s Theory of Definition: Some Questions on Posterior Analytics II 8-10." In Aristotle on Science: The Posterior Analytics, edited by Enrico Berti, 359-384. Padova: Editrice Antenore, 1981.
Aristotle. Aristotle's Posterior Analytics. Translated by Jonathan Barnes. Oxford: Clarendon Press, 1975.
———. Posterior Analytics. Loeb Classical Library. Translated by Hugh Tredennick. Cambridge, Massachusets: Harvard University Press, 1960.
———. "Posterior Analytics." In Great Books of the Western World, edited by Mortimer J. Adler, translated by G.R.G Mure. Vol. 7. Chicago: Encyclopedia Britannica, Inc., 1952.
Bolton, Robert. "Essentialism and Semantic Theory in Aristotle." Philosophical Review 85 (10/01, 1976): 514-44.
Charles, David. Aristotle on Meaning and Essence. Clarendon Press: Oxford, 2000.
Demoss, David and Daniel Devereux. "Essence, Existence, and Nominal Definition in Aristotle's 'Posterior Analytics' II 8-10." Phronesis: A Journal of Ancient Philosophy 33 (01/01, 1988): 133-54.
McKirahan, Richard D. Principles and Proofs: Aristotle’s Theory of Demonstrative Sciences. Princeton, N.J.: Princeton University Press, 1992.
Sorabji, Richard. "Definitions: Why Necessary and in what Way?" In Aristotle on Science: The Posterior Analytics, edited by Enrico Berti, 205-244. Padova: Editrice Antenore, 1981.
St. Thomas Aquinas. Commentary on Aristotle's Posterior Analytics. Translated by Richard Berquist. Notre Dame, Indiana: Dumb Ox Books, 2007.
———. Opera Omnia. Leonine. Vol. 1. Posteriorum Analyticorum. Rome, 1882.