Bryn Mawr Classical Review 2000.08.11
Wolfgang-Rainer Mann, The Discovery of Things: Aristotle's Categories and Their Context. Princeton: Princeton University Press, 2000. Pp. xii + 231. ISBN 0-691-01020-X. $39.50.
Reviewed by Malcolm Wilson, University of Oregon (email@example.com)
Word count: 2377 words
This book seeks to reveal and emphasize Aristotle's extraordinary accomplishment in providing a theory of ordinary objects. According to Mann (M.) Aristotle was the first in the Greek philosophical tradition to give expression to what has become a commonsense view of the world, that everyday objects, men, dogs, buildings, etc., are stable subjects for changing and shifting predications. These subjects are what M. calls 'things', and he claims that Aristotle discovered them. Behind this deliberately provocative claim stand some fairly familiar ideas. For we learn that the discovery of things means that Aristotle distinguished between accidental and essential predication, that he bestowed ontological primacy on particular substances, that he was the first philosopher to describe particular objects in terms of said-of and being-in relations, and that he did so against the background of preSocratic and Academic, specifically Platonic, philosophy. M.'s claim to originality lies in the novel and unusual interpretations he gives to several aspects of Platonic and Aristotelian philosophy, specifically, the significance of homonymy, synonymy and paronymy for Aristotle's theory of predication, the influence of Anaxagorean homoiomeries on Plato's theory of Forms, and the importance of the Late-Learners in the Sophist.
The book is generally clear and well written. M.'s historical approach is admirable: Aristotle was not writing in an intellectual vacuum; he was the greatest of the Platonists (after Plato, perhaps), and his work is best understood in a Platonic context. And yet the argument is not as compelling as one might wish, and the connections between the sub-arguments are not always clear and convincing.
The book consists of a lengthy discussion of some aspects of Platonic metaphysics set between two sections on Aristotle's Categories. As such, the subtitle "Aristotle's Categories and Their Context" is somewhat misleading since well over half of the book is devoted to Plato, including an extended discussion of self-predication and the participation relationship, and much of this material seems to be pursued for its own sake and not just as means to explaining Aristotle's Categories. M. rightly claims to be unusual in focusing on Platonic particulars and what becomes of them at the hands of Aristotle. Basically his argument is this: when Aristotle comes to deal with particular objects, he effects a compromise between Plato's position that they are bundles of accidental attributes with no essential nature and the position of the Late-Learners represented in the Sophist (251b-d) that things are only essence and have no accidental attributes. Aristotle distinguishes between essence and accident, and is thereby able to provide stability to particular objects while allowing them to don and doff transitory features: the 'thing' is discovered. M. further claims that the significance of Aristotle's discovery has not been made clear because of a misinterpretation of Plato's particulars: the prevailing interpretation has failed to see that Plato's sensible objects are bundles of properties, and instead gives them an Aristotelian interpretation as objects which have properties predicated of them. He is certainly right that the Late-Learners, a peculiar and intriguing group, are usually overlooked.
M. begins Part I with an analysis of Categories 1, Aristotle's discussion of homonymy, synonymy and paronymy, and shows how these so-called -onymies grow out of Platonic eponymy (the naming of particulars after the Forms they instantiate). He then attempts to relate the -onymies to the forms of predication (said-of and being-in) discussed in Cat. 2, thereby tackling the difficult problem of explaining the significance of the first chapter of the Categories to the rest of the treatise. The middle term in his explanation is Plato's eponymy relation between particulars and forms (e.g., the particular beautiful thing and the Beautiful). The particulars are named eponymously after the Forms. M argues ingeniously that if we interpret Aristotle's -onymies as forms of Platonic eponymy we can see that paronymy and homonymy correspond to the Aristotelian being-in relation. Synonymy corresponds to the said-of relations.
This argument suffers from some serious weaknesses. First, M. begins by trying to show that the Aristotelian -onymies originate in Platonic eponymy. He does so by reformulating, for example, Aristotle's homonymy as "x is called P homonymously after y". The result is a very odd origin for Aristotelian -onymies, and indeed Aristotle shows no signs of deriving his distinctions from the eponymy relation in this way. For calling one thing after another is only appropriate for paronymy (among Aristotelian -onymies). Instead we should look to Speusippus' analysis of tautonymy and heteronymy and bypass eponymy altogether. Second, M. claims that the -onymies constitute a test for whether a given predicate is said of or is in a given subject. The test, however, is imperfect at best. The test for the said-of relation is: "If x is called N synonymously after y, then y is said of x" (54). M. requires us to restrict the scope of synonymy to cases where, e.g., man is said of Socrates and man is said of the species man, where the term 'man' is used synonymously. If we accept this restriction, M.'s condition does constitute a test. However, it will fail to capture the vast majority of the said-of predications, e.g., animal is said of man. Moreover, Aristotle probably did not have this test in mind when he composed Cat. 1, since he does not provide a said-of predicate and its subject as an example of synonymy. Instead he says that animal is said synonymously of both man and cow. Indeed, Aristotle does not seem to be interested in identity relations in Cat. 1, that is, he does not make use of statements like, man is said of the species man. Nor does he formulate synonymy in terms of the called-after relation. So if synonymy can be used as a test in this way, Aristotle shows no awareness of the fact. Third, for the being-in predication it would be convenient if the test were the contradictory of the test for said-of predication: if x is properly predicated of y, and is not called synonymously after y, then it is in y. But this test would admit many said-of predications (e.g. animal said of man). Instead, M. points out that predicates that are in a subject may be homonymous, paronymous or heteronymous with their subject. One need only construct a Venn diagram to see that there are few useful correlations within the set, 'two-term combinations', where the subsets are 'said-of', 'being-in', 'synonymous', 'called-synonymously-after', 'paronymous', 'homonymous' and 'heteronymous'. Arguably, only 'called-synonymously-after' is wholly contained in 'said-of', and 'paronymous' is wholly contained in 'being-in'. M's test is quite inadequate, and it is presumably for this reason that Aristotle uses instead the logos test (if the logos of the predicate can be predicated of the subject, the predicate is said of the subject; if not, the predicate is in the subject), a vastly more reliable and logically perspicuous test.
In an appendix, M. dismisses Speusippus as unimportant for this debate. I think that he was premature to do so for two reasons. First, Speusippus developed the notions of heteronymy and tautonymy within a system of dichotomous division and these lie behind some of Aristotle's distinctions. Second, Speusippus may have had a very interesting theory of the particular, a relational theory in which the being of a thing is the sum of its sameness and difference from all other things. As such it makes an interesting precursor of Aristotle's thing.
The central section (Part II) of the book is devoted to Platonic metaphysics and especially to the nature of the particular (as opposed to the Form). M. argues that Plato's sensible particulars are not things in Aristotle's sense and have no nature, but rather are bundles of participant features. He begins by distinguishing this view from the traditional Eleatic opposition of permanence and change, for he argues that Platonic "becoming" does not entail actual change. I found his discussion of the verb γίγνομαι to be the most interesting section of the book, since the 'funerary' use, which he identifies as Platonic, is rarely discussed and is not covered in LSJ. In this use γίγνεσθαι x means "to act in an x way" or "to manifest x quality." γίγνεσθαι καλός means "to act nobly" and does not imply that the agent is noble in his nature. We then have a sense of γίγνεσθαι which is different from coming to be through a process of change, a common interpretation of Platonic becoming. The significance of this point for M.'s larger argument is rather subtle. M. is trying to establish a sense for Platonic "becoming" that does not necessitate change and alteration, and that as such will provide a contrast with Aristotle's logical (rather than physical) notions of said-of and being-in. Perhaps such subtlety is unnecessary: there is no need to separate logic from physics in either Plato or Aristotle.
M. deepens his investigation of the Platonic particular with a discussion of Anaxagoras' homoiomeries. His basic point here is that Anaxagoras serves as a model for Platonic metaphysics in the specific areas of participation and self-predication. M. argues first that Anaxagoras was innocent of the distinction between stuffs (e.g. gold) and properties (e.g. hot), and shares this innocence of proto- thinghood with Plato. Second, homoiomerous gold, pure and unalloyed, has gold predicated of it. But an ordinary gold ring is called gold in virtue of the predominant participation of gold in it (in spite of all the other ingredients). Both are called gold in different but related ways. M. claims that Plato adopted this model, stripped the homoiomeries of their sensible and corporeal nature, but somehow was still thinking in sensible terms in the following way. When in the middle dialogues he adapts the homoiomeries to the formal realm, Plato maintains that the Forms must be purely what they are just as homoiomerous gold is purely what it is. But, as he learns later in the Sophist, this is impossible since there must be conceptual connections or mixture between one Form and another. Forms cannot be homoiomerous. M. notes that the Anaxagorean distinction between homoiomery and mixture can map onto the distinction between being [εἶναι] and manifesting [γίγνεσθαι]. Moreover, the model shows how Platonic particulars are bundles of features rather than things with natures and that they are called what they are called by eponymous synonymy. Again, M. draws some interesting observations here but seems to underestimate the force of logical Eleaticism when he claims that its paradoxes arise merely from a confusing disanalogy between the sensible and the intelligible realms. And once again, this particular interpretation of Platonic Forms and particulars is not essential for sustaining the Aristotelian part of the argument and therefore is really a digression.
At the end of Part II M. treats the Late-Learners, who according to M. hold that there is nothing predicated in a subject. Here M. admits that he is extending their view, but the result of his extension is a significantly different position from the one Plato puts in their mouth. For M. makes them accept said-of predications (predications of genus or differentia) in addition to mere identity statements. There is little hint that the Late-Learners' position is to be so interpreted. Plato feels that he must force them to admit such statements as "movement is being", and they explicitly accept only identity statements, e.g., man is man, good is good. True, the predications that the Late-Learners find objectionable (Sophist 251a8-b1) are all Aristotelian being-in predications, but their objection is that they involve multiplicity, and it is clear that they will resist even essential predication statements like "movement is being" for the same reason (251d5-e1). Now, nothing in the text explicitly contradicts M.'s interpretation, but Plato is not making the point M. attributes to him. This makes difficulties for M.'s argument since he sees Plato and the Late-Learners as advocating two antithetical positions regarding particulars, positions which Aristotle then resolves with his discovery of things.
In Part III M. tries to show how the -onymies help Aristotle to solve his predecessors' problems. Here M. refines his linguistic test which serves as a criterion for deciding whether any given predicate should be said of or in a subject: if an abstract paronym of the predicate can be created, then it can only be said in a subject, e.g. beautiful can only be said in a subject because it has an abstract paronym, beauty. On M.'s own admission the criterion must face the difficulty of the differentia, which is said of its subject but may also be paronymized into an abstract (e.g. rational into rationality). And how does one paronymize predicates like 'to the left of the man in the yellow hat?' Nevertheless, M. argues, it is by means of paronymy that Aristotle can show the Late-Learners how a name can be predicated of a different name: it is not true that man is beauty, but paronymously man is beautiful. M. argues that Aristotle replaces Plato's undifferentiated eponymy with a two-fold distinction between synonymy on the one hand, which describes the said-of relationship, and the paronymy and homonymy relationship on the other, which describes the being-in relationship. His reason for joining paronymy and homonymy is that he views homonymy as a degenerate case of paronymy: whereas the courageous man is so-called paronymously from courage (the words differ), the white (thing) is called homonymously from the white (as general quality). In the latter case white (subject and predicate) is the same word, but has different accounts. Though white is a homonym in this case, it does the same job as a paronym. Synonymy describes the said-of relation because the name and the account of the particular and the general that is said of it are the same: this man is man and man is man; the predicate is the same in both cases. The weakness of this section is similar to that of Part I: M.'s forces the -onymies into increasingly idiosyncratic and less Aristotelian usage in order to make them seem relevant to the forms of predication. One cannot help thinking that the story of the discovery of the thing could be told more lucidly without introducing the -onymies.
On the technical side there are a fair number (26 by my count) of typographical errors, some quite glaring, both in English and Greek; as well as obscure marginal lines (51) and misdirected arrows (189).