I would like to take this opportunity to express my gratitude to Maddalena Bonelli, to whom I owe and with whom I share the project of collecting all extant evidence on Alexander’s lost commentary on De interpretatione. The main ideas for this paper were originally presented at a workshop she organized in Bergamo in 2019. Her feedback and all comments from students and colleagues on that occasion greatly helped me develop my thoughts on these materials. I am very grateful to the Editors of this volume for their interest in the paper; to Gweltaz Guyomarc’h for his help with the French language; to Katerina Ierodiakonou and one anonymous referee for their comments.
- 1 Ammonius states clearly at the beginning of his commentary that he is relying on Proclus’s teachin (...)
- 2 As is well known, the distinction between onoma and rhêma plays a prominent role both in Plato and (...)
1In his commentary on Aristotle’s De intepretatione (Busse 11,1 ff.), Ammonius1 raises the question of why, of the eight parts of speech distinguished by the grammarians, Aristotle only mentions onoma and rhêma.2 In this context, Ammonius refers to and criticises the view, which he ascribes to Alexander of Aphrodisias, that epirrhêmata, ‘adverbs’, fall under the rubric of onomata. Ammonius gives some hints about Alexander’s reasons for endorsing such a view and gives a series of arguments against it. Ammonius’s criticism of Alexander seems quite sensible – in fact, so sensible that it is hard to imagine that a plausible view could have prompted Ammonius’s comments. Since Alexander’s commentary on De interpretatione is not preserved, we do not have access to the text Ammonius presumably refers to. However, I intend to show that, mainly based on some evidence in Alexander’s commentary on the Topics, we can reconstruct at least some of the things Alexander could have said about adverbs in such a way that such views would be intelligible, in line with Aristotle’s writings, and compatible with Ammonius’s criticism.
2I suggest that, while Alexander’s and Ammonius’s views about why Aristotle at the beginning of DI distinguishes only onoma and rhêma may have been similar, Alexander, apparently unlike Ammonius, may have been keen on distinguishing between two classes of adverbs, both of which play some role in Aristotle’s logical writings: tropoi (i.e. the modal operators such as ‘possibly’, ‘contingently’, ‘necessarily’ Aristotle explicitly deals with in DI 12 -13 and in his modal syllogistic); and ptôseis, i.e. the ‘inflections’ (adverbs in -ôs derived from a noun – usually an abstract noun) playing a prominent role in several of the topoi Aristotle gives in the Topics. Parts of my reconstruction of Alexander’s views are inevitably speculative. However, the texts in support of the reconstruction I propose are worth investigating in their own right as witnesses of Alexander’s views on different types of adverbs, independently of whether one endorses my suggestion that they may help us draw a likely picture of the thesis (presumably in Alexander’s lost commentary on De interpretatione) Ammonius refers to.
3The paper is divided into three parts. In the first part (section 2), I present Ammonius’s criticism by setting it up within the broader context of his views on parts of speech in the commentary on De interpretatione and by specifying what we can infer from it about Alexander’s views. In the second part (section 3), I review the main evidence for Alexander’s views in the commentaries on the Topics and Prior Analytics I. In the third part (section 4), I make some further suggestions about what Alexander might have thought about what ancient grammarians (and Ammonius) took as another type of adverb, i.e. adverbs expressing an obligation, before submitting a few general conclusions.
- 3 ‘Name’ is a common translation for onoma, but one should bear in mind that the Greek grammatical c (...)
- 4 Ammonius explains how haplê phonê (discussed in the Categories), onoma, rhêma, phasis (discussed i (...)
4Ammonius’s criticism of Alexander comes at the end of the section of the commentary (Busse 11,1 ff.) in which Ammonius approaches the question why, of the eight parts of speech distinguished by the grammarians, Aristotle only mentions onoma (‘name’)3 and rhêma (‘verb’). The eight parts of speech Ammonius has in mind are onoma, pronoun, verb, participle, adverb, article, preposition, and conjunction. Ammonius’s answer to the question is that onoma and rhêma are what is necessary and sufficient to build a simple declarative sentence (apophansis), which is the topic of DI. In the simple declarative sentence made of one onoma and one rhêma, the two logical roles of subject and predicate distribute neatly over the two basic components of the sentence (Busse 11,11-14). And the two basic components of such a sentence belong, respectively, to the grammatical categories of onoma and rhêma. Importantly, the account of the declarative sentence in DI is, according to Ammonius (as, before him, according to Alexander) supposed to be functional to Aristotle’s account of the syllogistic in APr., which is in turn functional to his account of scientific knowledge in APo. So, quite conventiently, the division into two parts of the simple declarative sentence maps onto the division into two terms of syllogistic premises and conclusions.4
- 5 Presumably, some syntactical considerations play some role, too, but Ammonius’s account may sugges (...)
5As Ammonius explains (Busse 11,8 ff.), not all parts of speech can take on the role of subject or predicate in a declarative sentence, but whether an expression can take on either role is crucial to determine a basic partition between (a) parts of speech that Aristotle would regard (according to Ammonius) as genuine parts of the declarative sentence, and (b) parts of speech that, without being genuine parts of the declarative sentence, are parts of the language in which the declarative sentence is expressed. Whether an expression can take on the role of subject or predicate is mainly established on semantic grounds:5 the types of expression signifying ‘determinate natures or simply people or activities or affections or some interweaving of these’ are those that can take on the role of subject or predicate in a declarative sentence. Such expressions include, in addition to onoma and rhêma, pronouns and participles, as Ammonius’s examples show (Busse 11,12-13): ‘Socrates walks.’; ‘I walk.’; ‘The running <thing> walks.’; ‘Socrates is <a> running <thing>.’ Expressions falling under these four grammatical categories, Ammonius says (Busse 12,16-20), all fall under Aristotle’s partition of onoma and rhêma, depending on whether they do (rhêma) or do not (onoma) signify time in addition to their basic signification, and on whether they take on the role of subject (onoma) or predicate (rhêma). All expressions not complying with the requirements for falling under either onoma or rhêma are not regarded as parts of the logos strictly speaking (I take this to mean: they are not genuine parts of the declarative sentence), but are only parts of its lexis, ‘linguistic formulation’ (Busse 12,20-13,2). Their role is that of bringing about a certain ‘composition’ (sunthesis) and ‘coordination’ (suntaxis) (Busse 13,2-3) of the parts that are parts of the logos strictly speaking. Conjunctions, articles, prepositions, and adverbs all belong within the class of parts of the language (or: of the linguistic expression) that are not parts of the logos.
- 6 Translations are mine unless otherwise indicated.
6In ascribing the distinction between parts of logos and parts of lexis to Aristotle, Ammonius explicitly refers to the division of the parts of lexis Aristotle draws in Poet. 1456b20-21: ‘These are the parts of all lexis: vocal element, syllable, conjunction, name, verb, article, inflection, account.’6 (Τῆς δὲ λέξεως ἁπάσης τάδ’ ἐστὶ τὰ μέρη, στοιχεῖον συλλαβὴ σύνδεσμος ὄνομα ῥῆμα ἄρθρον πτῶσις λόγος.) More specifically, the text of the Poetics, as Ammonius emphasises, lists logos as a part of lexis, which may suggest that lexis includes more than logos. Even if Ammonius does not linger on this point, it is also clear that this text mentions parts of speech that Aristotle does not mention in DI. Ammonius emphasises the semantic criterion for distinguishing parts of logos from parts of lexis: the logos is ‘an aggregate principally of the vocal sounds that signify things’, whereas lexis is an aggregate ‘of all vocal sounds, without qualification, that are used for the sake of language.’ (Busse 13, 7-9).
7To sum up, Ammonius’s account of why Aristotle only mentions onoma and rhêma in DI is not a reductionistic account of grammatical types. On Ammonius’s account, no grammatical category is “reduced” to any other grammatical category. His account of how pronouns and participles fall under Aristotle’s distinction rather suggests that he associates the labels of the grammatical types onoma and rhêma to the logical roles of subject and predicate, roles that can be assumed also by pronouns and participles.
8Ammonius’s report and criticism of Alexander’s alleged view that adverbs are onomata should be read against this backdrop. I shall quote the passage (Busse 13,19-14,17) in full:
(a) Now, it seems to the interpreter from Aphrodisias that adverbs too are names, just like pronouns and what is called by the grammarians ‘appellation’ (prosêgoria). (b) However, this does not seem to make any sense, because he could hardly suppose that certain adverbs were names, such as the adverbs of assent, refusal, forbidding, or those which are prefixed to oaths, or very many others, (c) while others seemed to him, because they were derived from certain names, to have the same force as the names, such as ‘beautifully’ and ‘clearly’, which derive from ‘beautiful’ and ‘clear’, just like ‘justly’ derives from ‘justice’. (d) But if this were so, then we shall call gamêteon (‘must-marry’) and pleusteon (‘must-sail) ‘verbs’, and not ‘names’, since they have come to be from ‘must marry’ or ‘must sail’ (ὡς
ἐκ τοῦ δεῖν γαμεῖν ἢ πλεῖν γεγονότα).
(e) Actually, it does not make sense either to call these ‘verbs’ or to call the former ‘names’, since none of them can be either a subject or a predicate in propositions, which is what name and verb respectively must do. For you will find the name ‘clear’ predicated in the sentence : ‘This passage of De interpretatione is clear’, while you will not find that this is the case for ‘clearly’.
- 7 The translation is a slightly modified version of Blank 1996, p. 22-23.
(f) And we will also deny that a sort of compound predicate comes to be from this and from that about which ‘clearly’ is said, e.g., ‘interprets clearly’, as occurs in the case of ‘dead man’ and ‘debased coin’. If that were so, the denials of these affirmations would have to arise in the same way as those which have a simple predicate, just as occurs in the case of genuine compound predicates : just as the affirmation ‘The deceased is a man’ takes the negation ‘The deceased is not a man’, and ‘The deceased is dead’ takes the negation ‘The deceased is not dead’, in the same way ‘The deceased is a dead man’ takes the negation ‘The deceased is not a dead man’. But in the case of propositions with a mode (meta tropou protaseis), necessarily negations come to be in a different way and not just like the negations of propositions without mode (aneu tropou) : the affirmation ‘Socrates walks’ takes the negation ‘Socrates does not walk’, but ‘It is possible that Socrates walks’ does not take ‘It is possible that Socrates does not walk’, but rather ‘It is not possible that Socrates walks’, as Aristotle will instruct us quite clearly in the last section of this book.7
9In (a), Ammonius ascribes to Alexander a thesis about adverbs: adverbs are onomata. Two main points are worth stressing. Firstly, Ammonius’s formulation may suggest that this is a thesis about all adverbs, but it seems unlikely that Alexander formulated any such universal thesis. In fact, Ammonius’s remarks in (c) suggest that Alexander formulated a thesis about adverbs of a certain type, i.e., adverbs ending in -ôs and sharing their root with an onoma (a noun or an adjective).
- 8 On indeterminate sentences, see Aristotle, DI 17b7-11; Top. III 6, 120a6-20; APr. I 1, 24a19; Alex (...)
- 9 Again, the important passage is Aristotle, Top. III 6, 119b6-20; cf. Alexander, in Top. 288,13-289 (...)
- 10 Aristotle criticises this suboptimal way of reaching a universal conclusion in a cumulative way, w (...)
10It is quite difficult to imagine why Alexander would have generalized such a thesis (whatever it amounted to exactly) to be about all adverbs. But it is not too difficult to imagine how a thesis about some adverbs could have been turned by Ammonius into a thesis about all adverbs. In fact, this move can be explained in two different ways. Alexander could have expressed his thesis as an indeterminate (i.e., not quantified) thesis about adverbs: ‘Adverbs are names’. Since indeterminate sentences can be interpreted either as universal or as particular,8 Ammonius may have ascribed to Alexander the universal thesis, i.e., the least charitable interpretation of the thesis he intends to criticise. Since a universal affirmative thesis (‘All adverbs are names’) is refuted by establishing the corresponding particular negative thesis (‘Some adverbs are not names’), Ammonius’s strategy would then consist in showing that the particular negative thesis can be established in many ways, for several different types of adverbs. Alternatively, Ammonius could take the indeterminate thesis as equivalent to a particular thesis and try to refute it as such. The dialectical strategy to refute a particular thesis (‘Some adverbs are names’) and an indeterminate thesis (‘Adverbs are names’) is the same,9 and it consists in arguing for the corresponding negative universal (‘No adverb is a name’). On this reconstruction, Ammonius is supposed to show that no adverb is a name. If this is Ammonius’s strategy, however, it should be noticed that Ammonius does not manage to produce one single argument for the universal negative conclusion. Rather, he produces a list of arguments targeting specific types of adverbs, each showing that some adverbs are not onomata. The desired conclusion, that no adverb is an onoma, would be obtained by adding the conclusions of the specific arguments, provided that the specific arguments cover all types of adverbs. This way of establishing a universal thesis about all adverbs by establishing specific theses about each type of adverb would be suboptimal in a scientific context.10 However, Aristotle acknowledges that, in a dialectical exchange, the answerer is required to concede a universal conclusion if the questioner has managed to obtain all or several of the corresponding specific claims (Top. II 2, 109b13-29), unless the answerer manages to come up with a counterexample to the universal conclusion that the questioner is trying to establish. In sections 3 and 4, I submit a few suggestions about how Alexander could have replied to Ammonius’s arguments.
- 11 On the determination of a thesis, see Aristotle, Top. III 6, 120a20-31; Alexander, in Top. 290,2-2 (...)
11If the formulation Ammonius reports picks up on Alexander’s original formulation of the thesis, it is remarkable that the thesis is not quantified. This may suggest that Alexander’s thesis was a thesis about adverbs, but it was a thesis about some adverbs, and not (or not necessarily) about all adverbs. The question remains of whether Alexander’s thesis was really completely indeterminate or rather quite determinate and about one specific class of adverbs, i.e. those that Aristotle calls ptôseis in the Topics.11 If Alexander’s thesis was a specific thesis about ptôseis, the only relevant arguments in Ammonius’s criticism would be those explicitly targeting the claim about ptôseis, i.e. those in sections (d), (e), (f). I shall return to this point in sections 3 and 4 of the paper.
- 12 According to Ammonius, all genuine parts of the logos signify natures, persons, activities, affect (...)
12The second point I would like to emphasise about Ammonius’s formulation of Alexander’s thesis, is that it is unclear what relation obtains between the grammatical category of adverb and the grammatical category of onoma. As we saw, Ammonius claims that, for Aristotle, all grammatical types that can take on the role of subject or of predicate in a declarative sentence fall under onoma or rhêma. This account associates the distinction of onoma and rhêma in DI to a functional distinction, in terms of logical roles (subject and predicate) in a declarative sentence, and to a semantic distinction, in terms of whether an expression signifies or does not signify time. For the sake of clarity, I shall distinguish grammatical types (those distinguished by the grammarian), logical types (with reference to the logical function in a declarative sentence) and semantic types (depending on whether time is included in the signification12 of a term). In light of these distinctions, Ammonius’s account of Aristotle’s views seems to be that the grammatical types of pronoun and participle, in addition to the grammatical types of onoma and rhêma, can all be ordered under the logical types typically associated to onoma and rhêma in DI as well as under the semantic types of onoma and rhêma. Whether Ammonius also wants to ascribe to Aristotle a reductionist or partially reductionist account of grammatical types, based on which, e.g., the grammatical type of pronouns is just a sub-group of the grammatical type onoma, is unclear, but does not seem to be required by the text.
13With respect to adverbs, Ammonius introduces further distinctions (Busse 11,14-12,13). He initially discusses adverbs under the general label of the parts of speech which do not signify natures, persons, activities, affections, or a combination of these, but ‘rather show a certain relation of the predicate towards the subject, as most adverbs do’ (11,15-16). But he then further distinguishes (Busse 12,6-13) between adverbs that ‘by showing a certain relation of the predicate towards the subject seem to contribute something to the generation of such declarative sentences’, and adverbs that rather contribute something to other species of logos (prayers, orders, questions etc.). At least the first class of adverbs is clearly different from the parts of speech that are completely without signification on their own: article, preposition, conjunction (12,13-15).
- 13 For types and examples, cf. Dionysius Thrax, Ars grammatica, Uhlig 72,3 ff.
- 14 On this type of adverbs, see section 4.
14Adverbs that contribute to the generation of a declarative sentence can signify many different things (Busse 11,16-12,9),13 but it seems that Ammonius takes all of them to signify a relation of the predicate to the subject. For example, they can signify how, when, where, how often or in what order the predicate belongs to the subject; they can signify a comparison in the way in which the predicate belongs to different subjects, or the intensity with which the predicate belongs to the subject. But they can also signify the attitude of the subject uttering the sentence (e.g., conjectural rather than fully endorsing, complaining, admiring etc.); they can express an obligation,14 or they can signify the speaker’s assent to or denial of something said by others.
- 15 On the difficulties to reconstruct Alexander’s views on signification involved in his understandin (...)
15Ammonius’s partition of adverbs relies on the work of the grammarians, but his discussion of adverbs that contribute something to what the declarative sentence signifies raises a series of philosophical questions. One of them is whether Ammonius would have conceded that at least some adverbs (e.g., those showing how, when, where a predicate belongs to a subject) may signify items belonging to some of the accidental categories. Neither Ammonius nor Alexander addresses this issue in the texts about adverbs, but one might wonder whether one thought behind Alexander’s views on adverbs could have been that, when one must identify the simple terms that pick out simple items belonging to the different categories, at least some adverbs will have to be considered. Note that, if this was Alexander’s strategy, then he may have indeed had an argument for the thesis that several of the adverbs that Ammonius lists among those contributing to the generation of a declarative sentence are onomata in the sense that each of them ‘names’ some item in a category. I have not found any evidence that this consideration played any special role in Alexander, and I will not explore it any further in this paper, but I do not think that the hypothesis is a non-starter.15
- 16 Barnes 1991 emphasises this feature of Ammonius’s account of adverbs, but he mainly discusses Ammo (...)
- 17 The three Greek expressions are used adverbially, but I try to keep in the translation the morphol (...)
- 18 One can, of course, construe the sentence as in ‘Clearly, Socrates interprets’, which would mean t (...)
16One important feature of Ammonius’s account of adverbs is that it does not seem to make room for a distinction between adverbs that qualify the way in which the predicate belongs to the subject and adverbs that only qualify the predicate.16 As we saw, he seems to think that adverbs that ‘contribute to the generation’ of a declarative sentence do so by expressing a qualification of the way in which the predicate belongs to the subject. One passage illustrating this point is Ammonius’s account (Busse 8,4-13) of the types of proposition Aristotle discusses in DI. He distinguishes (1) those made of predicate and subject only, (2) those made of predicate, subject and copula, and (3) those having, in addition to these constituents, ‘a mode... signifying how the predicate belongs to the subject, e.g. “necessary”, “impossible”, “contingent”,17 “beautifully”, “clearly”, “justly”, as when I say “It is possible that Socrates is musical” or “Socrates interprets clearly.” [...].’ While Ammonius’s general view on adverbs as expressing a certain relation of the predicate to the subject captures the scope of modal operators, it leaves out the possibility that e.g., in the sentence ‘Socrates interprets clearly’, the adverb ‘clearly’ only qualifies the predicate ‘interprets’ and not the way in which the predicate belongs to the subject.18 As we will see, this might be one of the main differences between Ammonius’s and Alexander’s views on adverbs.
17If this is correct, one consequence of this is that Ammonius’s point in (f) about the negation of modal sentences makes perfect sense for those adverbs, like the modes, that qualify the relation of the predicate to the subject, but it is not obvious that it would do any harm to Alexander’s point about the adverbs that are supposed to be onomata. As we shall see in section 3, there is some evidence that Alexander thought of ‘inflections’ (ptôseis) in the Topics rather as indicating the mode of an activity or a disposition, i.e., of term-like items, and not as qualifications of the relation between predicate and subject. He seems to share Ammonius’s views about tropoi, but he does not seem to think that tropoi and ptôseis work in the same way. Alexander’s views on inflections in the Topics may also suggest that he could have had a response to the point Ammonius makes in (e).
- 19 I return to this point in section 4.
- 20 Apparently, the status of adverbs was an issue not only for (presumably) Alexander and Ammonius. P (...)
18One final important point to stress before moving on to the analysis of Alexander’s texts is that, if my reconstruction is correct, the argument or, perhaps, the intuition Ammonius ascribes to Alexander in support of the thesis that adverbs are names in (c), i.e. that the linguistic formulation of some adverbs derives from the linguistic formulation of a semantically cognate onoma, is very different from the arguments Ammonius gives earlier for the claim that pronouns and participles also fall within the partition of onomata or rhêmata. The argument Ammonius ascribes to Alexander in (c) is based on morphological features of linguistic expressions, but it is unclear that exactly the same strategy is used in (d) to argue that some adverbs would turn out to be verbs.19 This shows that there might have been different ways, for those who wanted to do so, to argue in support of the view that adverbs fall under either onoma or rhema, and that it is unclear how much doctrinal orthodoxy could be squared with systematic reflection on grammar and language.20
19Before taking a closer look at how Alexander’s views on adverbs may have differed from Ammonius’s, it might be worth emphasizing that such differences do not seem to be linked to radical differences in their general understanding of what the ‘genuine’ parts of a declarative sentence are. Differences might have rather concerned their views on what grammatical categories can be counted among the genuine parts of a sentence (as opposed to, in Ammonius’s terms, parts of language or linguistic expression only).
- 21 Alexander, in APr., Wallies 10,13 ff.
- 22 More generally on Alexander’s understanding of protasis with reference to his views on De interpre (...)
20We can get a sense of Alexander’s take on this point from his commentary on the initial chapter of APr. I. Alexander emphasises21 that in APr. I 1, 24a16 ff. Aristotle gives an account of protasis (‘premise’ or ‘proposition’) before introducing the (according to Alexander, unusual: 14,25-28) terminology of horos (‘term’) to indicate the parts of a protasis. In this way Alexander explicitly ties his understanding of the parts of a protasis to the two-term account of the protasis, which is also a declarative sentence on the account of DI. The two terms (horoi) are the parts ‘from which’ the protasis is delimited and out of which it is composed (14,29-15,1). In the simple protasis, such terms are onoma and rhêma (14,28-29); one term is the subject, one term is the predicate (15,1-4). Furthermore, one should not be misled by the presence of esti (‘is’) in sentences construed with a copula into thinking that esti is a third term (15,4-7). Every protasis, even those in which esti and a negation (or a mode) appear, is divided into two terms (15,11).22 These remarks certainly do not give us the full picture of Alexander’s views on parts of the protasis, parts of the declarative sentence and parts of speech, but the general impression is that Alexander’s views on this basic point may have not been radically different from Ammonius’s.
- 23 Blank 1996, p. 141, n. 75 refers to Schoemann 1862, p. 157, who “points out that Aristotle nowhere (...)
21Let us now go back to the passage in which Ammonius’s criticises Alexander’s views on adverbs. In his notes on this passage,23 David Blank, relying on Gerog F. Schoemann’s work on parts of speech in antiquity, suggests that Alexander’s views might refer to Aristotle’s mention in Top. VI 10, 148a10-13 of adverbs in -ôs as ptôseis of names. That Alexander had the Topics in mind is, I believe, certainly correct. The passage in Top. VI 10 is one of many in which adverbs in -ôs are called ptôseis, but the passage does not really say that such adverbs are, in general, ptôseis of names. In the passage Aristotle rather makes the point that, if we want to check whether a formula F that has been submitted as a definition of a subject S, that is picked out by a name N, is a definition of S, we should check whether the ptôsis of F is also the definition of the ptôsis of N:
- 24 Ἔτι ἐπὶ τῶν ὁμοίων τοῦ ὀνόματος πτώσεων αἱ ὅμοιαι τοῦ λόγου πτώσεις ἐφαρμόττουσιν· οἷον εἰ ὠφέλιμον(...)
Furthermore, similar inflections of the account correspond to similar inflections of the name : e.g., if what is productive of health is useful, ‘in a way that is productive of health’ is ‘in a useful way’, and what has produced health has been useful.24
22So, in fact, this passage rather says that there can be inflections of single onomata as well as of complex linguistic formulas. To appreciate Aristotle’s rather idiosyncratic use of ptôsis in Top., we can take a look at a passage where Aristotle actually feels the need to explain what he means by ptôsis. The context of the passage is that Aristotle has just mentioned (114a26-27) two types of topoi he often refers to in all books of the Topics, i.e., topoi from the coordinates (sustoikha) and topoi from inflections (ptôseis):
- 25 Top. II 9, 114a27-36: λέγεται δὲ σύστοιχα μὲν τὰ τοιάδε οἷον τὰ δίκαια καὶ ὁ δίκαιος τῇ δικαιοσύνῃ(...)
Coordinates are called, e.g., the just things and the just man with respect to justice, and the brave things and the brave man with respect to bravery. And similarly, the things that produce or preserve [something] are coordinates with that which they produce or preserve, e.g., healthy things with health and wholesome things with wholesomeness. And in the same way in the other cases too. Such things, then, are usually called ‘coordinates’, whereas inflections (πτώσεις) are e.g., ‘justly’ and ‘bravely’ and ‘healthily’ and all things that are said in this way. And it seems that also things said according to inflections are coordinates, e.g., ‘justly’ with justice and ‘bravely’ with bravery.25
- 26 Cf. Alexander, in Top., Wallies 104,14-16; 105,13-16.
- 27 With reference to the cases Ammonius mentions in section (d), one might wonder why those could not (...)
23In his commentary, Alexander is perfectly aware of the fact that this use of the word ptôsis can be confusing. He does not miss a chance to remind his readers of the peculiar way in which Aristotle uses this word in the Topics (see e.g.: Wallies 103,30-104,16; 105,16; 263,8-10; 282,4; 330,20-331,3; 331,5-6). He explicitly records the fact that ptôsis can be used (and is usually used) with a broader meaning (koinoteron: 103,30), to indicate any inflection (in particular: the different grammatical cases of nouns and adjectives), but in the Topics ptôsis indicates specifically (idiôs: 104,13) what takes its “shape” from something else in the way in which grammatikôs (‘grammatically’) derives from grammatikê26 (cf. 104,14-16; 105,13-16). Alexander’s repeated caution about Aristotle’s special use of this word might be due to the fact that Alexander knew that elsewhere (e.g., Poet., 1457a18-23; DI 16b1, b17, 17a10) Aristotle uses this expression to refer to all inflections of nouns and verbs.27
- 28 Castelli 2020, p. 121-122.
- 29 Cf. Alexander, in Met., Hayduck 401,14-22, where adverbs in -ôs (e.g., kalôs / phaulôs, ‘well’ / ‘ (...)
24Alexander goes out of his way to spell out what the ptôseis of the Topics signify, especially in opposition to paronyms (Wallies 103,30-104,16) and coordinates (Wallies 196,27-197,31), to which they are often associated in the Topics. One passage (Wallies 196,27-197,31, where Alexander is commenting on Aristotle, Top. II 9, 114a26 ff.) is especially significant. I have provided a full translation of the passage elsewhere28, but the important points for present purposes can be summarised in this way: coordinates ‘are pragmata’ of various types (Wallies 197,22: πράγματά τινά ἐστιν; cf. 197,1-17), ‘whereas ptôseis do not show underlying things, but rather a mode of a certain activity or a disposition’ (Wallies 197,22-24: αἱ δὲ πτώσεις οὐ πραγμάτων ὑποκειμένων εἰσὶ δηλωτικαὶ ἀλλὰ τρόπου ἐνεργείας ἢ διαθέσεώς τινος).29 This makes it quite clear that Alexander does not think that these adverbs qualify the relation between predicate and subject in the way in which Ammonius spells out in his criticism in (f). Rather, these adverbs typically qualify an activity or a disposition, both of which are indicated by term-like expressions. So, Ammonius’s criticism in (f) does not really apply to the adverbs with reference to which Alexander presumably formulated the thesis Ammonius is discussing.
25Typically, the expression signifying what a ptôsis is a ptôsis of in the Topics is indeed an onoma (e.g., ‘justly’ is an inflection of ‘justice’), and this seems to be what Alexander had in mind in formulating the view Ammonius attacks in his commentary (as David Blank suggests in his notes). But if this is so, then Alexander could have made one crucial point in responding to Ammonius’s criticism in (e): Ammonius objects that no adverb can play the role of subject or predicate, i.e., that no adverb can be a term in the two-term analysis of the proposition. But Alexander could have easily pointed out that, on Aristotle’s account of topoi from ptôseis, the ptôseis of subject and predicate in the proposition that has to be tested keep playing the role of subject and predicate in the proposition that is used to test the original formulation. This can be already seen in the passage from Top. VI 10 quoted on p. 18, but most topoi from ptôseis would do. By way of example, consider the following passage from Aristotle, Top. II 9, 114b8-13 (I provide a very literal and intentionally unintelligible translation):
- 30 καὶ εἰ ἡ δικαιοσύνη ἐπιστήμη, καὶ ἡ ἀδικία ἄγνοια· καὶ εἰ τὸ δικαίως ἐπιστημονικῶς καὶ ἐμπείρως, τ (...)
And if justice is knowledge, then injustice is ignorance ; and if *the justly* is knowledgeably and competently, then *the unjustly* is ignorantly and incompetently. And if these are not the case, then those are not the case either, as in the case just mentioned : for *the unjustly* may appear to be more competently than incompetently.30
26The expressions I indicated with *...* translate Greek expressions in which the singular neuter article to preceeds an adverb in –ôs; similarly, all adverbs in the translation translate adverbs in –ôs. Aristotle’s texts on ptôseis in the Topics are particularly challenging, to understand and to translate, because it is not fully clear whether Aristotle takes inflections to be linguistic items or extra-linguistic items. In the former case, the best translation for *the justly* would be a quoted word (e.g., ‘… “justly” …’: the singular neuter article in Greek can be used to quote the following linguistic expression). In the latter case, *the just* should rather be rendered: ‘what <is done> justly’, or, perhaps, ‘the just way of doing things’ – or something along these lines.
27These difficulties indirectly affect Alexander’s commentary, but occasionally Alexander paraphrases Aristotle’s text in such a way that it becomes clear that, in agreement with his general view about what ptôseis signify or show (a mode of an activity or a disposition), he takes the corresponding adverbs not to be quoted adverbial expressions, but rather to indicate, possibly in an elliptical way, the ‘mode’ in which something is done or something acts or comes about. One example for this is the formulation at Wallies 282,4-6, where he renders the inflection of ‘something unjust’ (τι ἄδικον) with the formulation: ‘some of the things that come about unjustly’ (τι τῶν ἀδίκως γινομένων). The paraphrase certainly testifies to Alexander’s attempt to come up with a plausible explanation of how Aristotle’s arguments from ptôseis are supposed to work, but the main point is that Alexander certainly has to deal with cases in which the role of subject and predicate (i.e., the roles of the two terms in a simple declarative sentence) is taken by ptôseis, i.e., adverbs of a certain type.
- 31 Alexander, in Top., Wallies 177,9-10; cf. 134,31-137,4.
- 32 Ibid. 177,19-27.
28It is important to stress that Alexander seems to believe that tropoi, i.e., the modes Ammonius regards as paradigmatic adverbs in (f), do qualify the way in which the predicate belongs to the subject. He repeatedly says so not only in the commentary on APr. , but also in the commentary on the Topics. Tropoi indicate the material differences in how something belongs to something else (in APr., Wallies, 27,16 ff.; cf. 28,20; 330,3-5; 411,33 ff.). Furthermore, Alexander seems to think in similar terms of the different ways in which a predicate is ascribed to a subject, depending on whether the predicate is an accident, a distinctive feature, a genus, or the definition of the subject. In his commentary on the Topics (Wallies 177,8-180,17), Alexander analyses how the distinction of the ways in which predicates belong to the subject in Top. II 6, 112b1-20 (‘necessarily’; ‘for the most part’; ‘as it happens’) interacts with the fourfold distinction of types of predicates in the Topics,31 with the modes distinguished in DI 12-13 and in the modal syllogistic in APr.32 These are all tropoi tês huparxeôs (178,1; 179,4-5), ‘modes of obtaining’ or, rather, ‘modes of belonging’: they indicate the way in which something belongs to something else (177,27-178,2).
29This evidence shows (1) that Alexander may have had an argument, based on Aristotle’s use of topoi from ptôseis in the Topics, that there are cases in which ptôseis indeed play the function of the relevant terms (i.e., subject and predicate) in a declarative sentence, and (2) that Alexander presumably distinguished the ways in which ptôseis and tropoi work in a declarative sentence. If this is correct, (1) would have given Alexander a response to Ammonius’s argument in (e), and (2) would have given him a response to Ammonius’s argument in (f). In the next section I will present some evidence on Alexander’s views on the sort of adverb Ammonius considers in (d). This evidence does not necessarily tell us much about what Alexander would have responded to (d), but it does point in the direction of some important differences in the way in which Alexander and Ammonius thought of adverbs expressing an obligation (i.e., verbal adjectives in –teon).
- 33 For some discussion on the late appearance of adjectives as a grammatical category, see Alfieri 20 (...)
- 34 Cf. Dionysius Thrax, Ars grammatica, Uhlig 85: Τὰ δὲ θετικά, οἷον γαμητέον πλευστέον.
30In section (d) of the passage in which Ammonius criticises Alexander, Ammonius refers to (what according to contemporary linguistic are) verbal adjectives in –teon.33 These expressions were regarded by ancient grammarians as adverbs expressing an obligation (thetika epirrhêmata),34 and Ammonius himself mentions such adverbs at different places in the commentary (e.g., 9,14-15; cf. 11,24-26). According to Ammonius, these adverbs, too, belong to the group of adverbs that express ‘a certain relation of the predicate towards the subject’ (11,15-16: σχέσιν δέ τινα τοῦ κατηγορουμένου πρὸς τὸ ὑποκείμενον δηλοῦσιν). More specifically, as Ammonius spells out, these are adverbs ‘showing that it is advantageous for us to choose what is predicated’ (11,24-25: ὅτι λυσιτελὲς ἡμῖν τὸ αἱρεῖσθαι τὸ κατηγορούμενον δηλοῦντες). In section (d) Ammonius suggests that, if one followed Alexander’s strategy to justify the claim that adverbs are names, one could equally well say that adverbs are verbs since adverbs expressing an obligation derive from verbs, e.g., ‘must marry’ or ‘must sail’.
31Ammonius’s remarks raise several questions. For example, in spelling out why such adverbs should be called ‘verbs’ on Alexander’s account, he says that the verbal expression from which they ‘have come to be’ is, e.g., δεῖν γαμεῖν, ‘must marry’ rather than just γαμεῖν, ‘to marry’. This may suggest that Ammonius has in mind a verbal paraphrase that renders the semantic content of the adverb rather than morphological derivation, which rather seems to be what he refers to in (c). Be this as it may, it is hard to tell whether Alexander would have been interested in tracking morphological relations in this case too. If so, he could presumably have said that there are onomata, such as e.g., gamos and plous respectively, that express the same content as the verbs, without the addition of time. But this is not the point I want to explore. The point I want to explore is rather Alexander’s account of the ‘deep grammar’ of sentences including such adverbs, which, again, may have differed from Ammonius’s account.
32As far as we can tell from the brief remarks mentioned earlier, Ammonius thinks of sentences including adverbs of obligation in the following way: the declarative sentence is constitued by a subject and a predicate, e.g. ‘we’ and ‘marry’, and the adverb expresses a relation obtaining between the predicate and the subject – in this case: that it is ‘advantageous for us to choose [what is expressed by] the predicate’. I shall leave aside the question of the extent to which this can be regarded as an analysis of the grammar of the sentence rather than a more of less free paraphrase of the meaning of the expression gamêteon. It is not obvious that there is a clear boundary between the two, in Ammonius as well as in other authors, such as Alexander himself. But one important point is that, on Ammonius’s account, the adverb that is being analysed is semantically linked to the predicate of the declarative sentence it is taken to modify. As we saw,35 Ammonius is keen on emphasizing the correspondence between verb and predicate, and this might give us a clue about Ammonius’s reasons for thinking that adverbs expressing obligations should be led back to verbs rather than to names. Of course, Ammonius does not endorse the argument he presents in (d). But it is unclear how the argument goes, and I am suggesting that Ammonius’s remarks on these adverbs elsewhere may give us a clue about his way of thinking about the argument.
- 36 Following Aristotle, for Alexander both dialectical problems and propositions (or premises: protas (...)
33In one passage of the commentary on the Topics, Alexander discusses dialectical problems that are expressed through what, in Ammonius’s classification, are adverbs of obligation. Alexander does not call them ‘adverbs’ (or anything else), but he discusses how the corresponding problems should be construed. One important point to bear in mind here is that for Alexander all declarative sentences corresponding to dialectical problems and propositions36 are supposed to have predicative structure. So, one of Alexander’s concerns in spelling out the content of questions such as πότερον γαμητέον ἢ οὔ (which we can only render: ‘whether one should get married, or not’) consist in identifying a subject and a predicate of the problem. In particular, Alexander suggests that problems expressed through this formulation can all be understood as problems in which the predicate is an accident of the subject in this way:
- 37 Alexander, in Top., Wallies 52, 19-27. For some context, see Castelli 2022, p. 24-33.
[…] the problems that pertain to deliberation (τὰ συμβουλῆς ἐχόμενα προβλήματα) are somehow themselves comparative too (καὶ αὐτὰ συγκριτικά πώς ἐστι) – for which reason they, too, would be brought under those from the accident. For example : ‘Should one get married, or not ?’ (πότερον γαμητέον ἢ οὔ), and ‘Should one have children, or not ?’ (πότερον παῖδας ποιητέον ἢ οὔ) are themeselves comparative, too : for in them the enquiry is which of the things in the problem is more choiceworthy (ζητεῖται γὰρ ἐν αὐτοῖς πότερον αἱρετώτερον τῶν ἐν τῷ προβλήματι). For either this [scil. : what is more choiceworthy] is the object of the enquiry, or whether getting married or having children, or whatever is posited in the problem, is choiceworthy (εἰ αἱρετόν ἐστι τὸ γαμεῖν ἢ τὸ παῖδας ἔχειν ἢ ὅ τι ἂν ᾖ ἐν τῷ προβλήματι κείμενον). But all such things are accidents : for the enquiry is whether being choiceworthy belongs as an accident to the subject (ζητεῖται γὰρ εἰ συμβέβηκε τῷ ὑποκειμένῳ τὸ αἱρετόν) : so that such problems, too, would fall under the accident.37
34Alexander is here discussing Aristotle’s classification of problems in Top. I 5, 192a14-19, based on which ‘all mutual comparisons that are said, in whatever way, from the accident’ are supposed to fall under the rubric of problems from the accident. Alexander is trying to explain what ‘in whatever way’ (192a15) can possibly mean, and suggests that it might be meant to include, among other things, problems relating to deliberation. As the final lines of the passage show, Alexander analyses gamêteon rather differently from Ammonius, in terms of ‘getting married is choiceworthy’ (or ‘getting married is more choiceworthy than not getting married’). I do not have the space here to give a full assessment of the advantages and disadvantages of Ammonius’s and Alexander’s analyses, but one point worth emphasizing is that Alexander links the semantic content of the adverb to what he takes to be the subject of the corresponding declarative sentence. Furthermore, he does not seem to have a problem with verbal expressions working as names and indicating the subject of a declarative sentence.
35It is unclear to me whether Ammonius’s point in (d) would have been upsetting for Alexander (if his thesis was only about some adverbs, (d) may just miss the mark). But independently of that, the evidence presented in this paper shows that, on different groups of adverbs, Alexander’s views may have been different from, but not obviously worse than, those of Ammonius.