Dik ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 2014 Helma Dik ‘Most likely to succeed’: Degree Adverbs Specifying Probability in Classical Greek Helma Dik HIS PAPER ARGUES that in Classical Greek a number of degree adverbs (µάλιστα, ἥκιστα, and µᾶλλον, among others) can play a role in specifying the likelihood of potential events, acquiring the meaning ‘most likely’, ‘least likely’, and ‘more likely’, respectively, in certain contexts. Introduction: adverbs and modal expressions in Classical Greek Since the role of adverbs, other than the modal ‘particle’ ἄν, in Ancient Greek modal expressions has received little attention in general, I will start with the best-known case of interaction of an adverb with the modal force of the predicate. It is docu- mented in grammars of Ancient Greek that the potential optative (‘it is possible that X’) accompanied by a negation expresses what Gildersleeve calls ‘total negation’,1 in other words, it should be paraphrased with ‘it is not possible that X’ rather than with ‘it is possible that not-X’.2 Consider (1), from a speech by the Corinthians at Sparta (Thuc. 1.121.4):3 1 B. L. Gildersleeve, Syntax of Classical Greek (New York 1900) §442. 2 The axiom δὶς ἐς τὸν αὐτὸν ποταµὸν οὐκ ἂν ἐµβαίης ‘you can’t step into the same river twice’ is perhaps the most generally known example. Here too it is obvious that Heraclitus would not have wanted us to understand, ‘you might possibly not step into the same river twice’. Gildersleeve (§442) and A. Rijksbaron, The Syntax and Semantics of the Verb in Classical Greek3 (Chicago 2002) 41, 48 (“emphatic negation”), associate this construction primarily with the aorist, but compare A. Lanski, “Emphatic Negation and the Potential Optative,” Mnemosyne 66 (2013) 777–783. 3 Unless otherwise indicated, examples are from the OCT Thucydides. Translations are taken from the Loeb (C. F. Smith, 1919). T 600 ‘MOST LIKELY TO SUCCEED’ ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 (1) ὃ γὰρ ἡµεῖς ἔχοµεν φύσει ἀγαθόν, ἐκείνοις οὐκ ἂν γένοιτο διδαχῇ. The excellence that nature has given us cannot become theirs through instruction. The claim made by the speaker is not that instruction could possibly not yield results; rather, the claim is that instruction could not possibly yield results. Formally, one would analyze the combination of adverb and modal as one in which the adverb has scope over (is hierarchically superior to) the modal. This phenomenon practically exhausts the discussion of ad- verbs in modal expressions by the standard reference gram- mars, with two exceptions. Goodwin, who does not in fact discuss the negative, does mention a further adverb as especially frequent in combination with the modal particle ἄν,4 namely τάχα ‘perhaps’ (Thuc. 1.81.1): (2) τάχ’ ἄν τις θαρσοίη ὅτι τοῖς ὅπλοις αὐτῶν καὶ τῷ πλήθει ὑπερ- φέροµεν, ὥστε τὴν γῆν δῃοῦν ἐπιφοιτῶντες. Perhaps some of us are emboldened by our superiority in arms and numbers, which enable us freely to invade and lay waste their ter- ritory. It is understandable that it is only τάχα that was singled out by Goodwin, since the adverb is so frequently found with ἄν in the classical period and τάχα ἄν can even be used as a stand-alone expression.5 Often the adverb ἴσως is found in combination with τάχ’ ἄν (6.78.3): 4 W. W. Goodwin, Syntax of the Moods and Tenses of the Greek Verb (Phil- adelphia 1890) §221. In the classical period, the predominant meaning of τάχα is ‘perhaps’. I leave τάχα ‘quickly’, which barely survives in the classical period (its place having been taken by ταχέως), out of consideration here. 5 Gildersleeve §458; LSJ s.v. τάχα. Since ἄν cannot normally be used in the absence of a verb with which it is construed, this is strong evidence for τάχ’ ἄν being felt as a fixed combination. Furthermore, in Thucydides we find one example (6.2.4) of τάχα ἂν δέ, instead of the expected ordering τάχα δ’ ἄν, which again is evidence for the combination being treated as a unit. HELMA DIK 601 ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 (3) καὶ εἰ γνώµῃ ἁµάρτοι, τοῖς αὑτοῦ κακοῖς ὀλοφυρθεὶς τάχ’ ἂν ἴσως καὶ τοῖς ἐµοῖς ἀγαθοῖς ποτὲ βουληθείη αὖθις φθονῆσαι. And if he should err in judgment, when he has to lament his own ills he may perhaps some day wish once more to become envious of our good fortune. While the potential optative can range from requests, to polite assertions of fact, to actual expression that an event expressed by the verb may possibly take place, τάχα explicitly attributes at least some degree of probability to the event.6 In the case of the negative, the composite meaning (‘not possible’ vs. ‘possible…that not’) demonstrates the hierarchy of adverb and predicate. As for adverbs meaning ‘perhaps’ (ἴσως or τάχα), these cannot be used to indicate the manner in which a state of affairs takes place; their usage is confined to this same high level of ‘attitudinal disjunct’, that is, they are adverbs which express the speaker’s attitude to the state of affairs.7 There is one further comment in Schwyzer-Debrunner, un- accompanied by illustration, which reads, in full, “Der Wahr- scheinlichkeitsgrad kann durch Adverbia wie ἴσως, ep. ῥεῖα näher bestimmt werden.”8 In a nutshell, this is the point I want to make in this paper, since this interpretation still seems to be overlooked too often as a possibility. I will argue that there are additional adverbs, namely πλεῖστα, µάλιστα, ἥκιστα, and ἐλάχιστα, that should be interpreted as further specifying the likelihood of the potential event. Unlike τάχα, these adverbs 6 This degree of probability is, however, not clearly specified further by τάχα, as noted in LSJ s.v.: “to express any contingency from a probability to bare possibility.” For an example of the former one might propose Thuc. 1.77.6 (if you Spartans should come to power, you would τάχα become un- popular as well); for an example of the latter, Nicias’s thought in 6.19.2 that he might τάχα change people’s minds by giving a high estimate of troop levels needed for the expedition. 7 On disjuncts, compare R. S. Quirk, S. Greenbaum, G. Leech, and J. Svartvik, A Comprehensive Grammar of the English Language (London 1985) 52, 612–613. 8 E. Schwyzer and A. Debrunner, Griechische Grammatik II (Munich 1959) 324. 602 ‘MOST LIKELY TO SUCCEED’ ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 place the likelihood of events at an extreme: either extremely high, or extremely low.9 But before I turn to the additional adverbs for which I claim this use, I should point out an important distinction between the two adverbs mentioned in Schwyzer-Debrunner. As a result of their meaning ‘perhaps’, the use of ἴσως and τάχα with the potential optative, which is widespread, is unlikely to lead to in- terpretative difficulty. With ῥεῖα, the situation is different. The English adverb ‘easily’ is most readily associated with manner (an ‘adjunct’) as in, ‘I could easily see the match from my court- side seat’ (answering, ‘how was your view at the French Open?’) but can be used as an attitudinal disjunct as well (‘On clay, I can easily see Djokovic beat Federer’, answering, ‘Do you think he can beat him?’), where the speaker asserts her belief about the odds of Djokovic beating Federer, not about the ease with which the spectator follows the match. In both uses, the adverb ῥεῖα can be translated by the same English word, so while it is potentially ambiguous, nevertheless, like ἴσως and τάχα, it is unlikely to confuse a modern reader. First consider Homer Il. 17.70–71:10 (4) ἔνθά κε ῥεῖα φέροι κλυτὰ τεύχεα Πανθοΐδαο Ἀτρεΐδης, εἰ µή οἱ ἀγάσσατο Φοῖβος Ἀπόλλων, Easily then would Atreus’ son have carried off the glorious armor of the son of Panthous, if Phoebus Apollo had not begrudged it him, … My preference is to read this instance along the lines suggested by Schwyzer-Debrunner, with ῥεῖα indicating a high degree of likelihood. One could paraphrase with ‘it could easily have happened 9 As lower-level adjuncts, the most common use, these adverbs specify the degree to which a predicate holds true, e.g., ‘least trustworthy’, ‘the news- paper people trust most ’. Two examples from Thucydides are 1.3.1 δηλοῖ δέ µοι καὶ τόδε τῶν παλαιῶν ἀσθένειαν οὐχ ἥκιστα (specifying δηλοῖ) and 1.10.4 ἔξω τῶν βασιλέων καὶ τῶν µάλιστα ἐν τέλει (specifying ἐν τέλει, “most powerful”). 10 Translations are taken from Wyatt’s revised edition of Murray’s Loeb translation (1999). HELMA DIK 603 ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 that Menelaos…’, indicating that the narrator presents this as a likely scenario before the divine interference of the εἰ µή clause.11 This seems preferable to an interpretation as a Manner Adjunct, roughly ‘Menelaos would now have carried off the armor with ease’, but the latter interpretation is not impossible.12 In (5), on the other hand, where Nestor has asked if the horses he admires were given by a god, Odysseus denies this, but ac- knowledges that it would be easy for gods to give such a gift, or rather an even better one (Il. 10.556–557): (5) ῥεῖα θεός γ’ ἐθέλων καὶ ἀµείνονας ἠέ περ οἵδε� ἵππους δωρήσαιτ’, ἐπεὶ ἢ πολὺ φέρτεροί εἰσιν. Easily might a god if he wished bestow even better horses than these, since the gods are mightier far. In this case, the adverb ‘easily’ is more likely a comment on the ease with which gods can accomplish anything (as in ῥεῖα µάλ’ ὥς τε θεός, Il. 3.381; θεοὶ ῥεῖα ζώοντες, 6.138) rather than on the likelihood of this gift imminently happening. In later Greek, I would argue that in its classical form, ῥᾳδίως can also be found to express a higher likelihood, as in this state- ment in Xenophon about Socrates (Mem. 4.4.4): (6) ἀλλὰ ῥᾳδίως ἂν ἀφεθεὶς ὑπὸ τῶν δικαστῶν, εἰ καὶ µετρίως τι τούτων ἐποίησε, προείλετο µᾶλλον τοῖς νόµοις ἐµµένων ἀπο- 11 On the mood of φέροι and for further references see M. Edwards, The Iliad: A Commentary V (Cambridge 1990) 70. Chantraine (quoted by Edwards) translates “on eût vu…” (one could have seen…) which makes explicit that we should read this as authorial comment by the narrator. 12 As a further test, one can try to substitute ‘certainly’ for ‘easily’ and con- sider whether the sense of the sentence fundamentally changes as a result. I would argue that ‘certainly’ fits better in (4) and (6) than in (5). This test is inspired by Donatus’ comment on Ter. Andr. 720, cited in TLL s.v. facilis (VI.1 70.76–77), ‘facile’ adverbium confirmantis est, id est ‘liquido’ et ‘manifesto’. veteres ‘facile’ dicebant pro ‘certo’. The TLL itself describes this use of the adverb facile as “magis de ‘opinione’, fere i. q. ‘certo’, ‘sine dubio’.” Quirk et al., Comprehensive Grammar 589, compare ‘easily’ in modal expressions with ‘well’: ‘They might easily have been arrested’ ≈ ‘They might well have been ar- rested’. 604 ‘MOST LIKELY TO SUCCEED’ ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 θανεῖν. but even though he could easily have been acquitted by the jury, if only he had made even a slight compromise, he preferred to be true to the laws and die. As was the case with ῥεῖα in (4), the adverb ‘easily’ does not describe the manner in which the event of acquitting is carried out, but the likelihood of it happening. In conclusion to this section, a quick note on the position of the adverbs is perhaps in order. Instances (4) through (6), in which I argue for a different status of the adverbs, as adjunct (5) or as disjunct ((4) and (6)), all feature the adverbs at the be- ginning or as close to the beginning of the clause as they can be following a Setting (ἔνθα) or connector (ἀλλά). Whereas there is a strong tendency in English for style disjuncts (‘frankly’, ‘briefly’) to be placed at the periphery of the sentence, and one might see this as a direct reflection of the wide semantic scope of these adverbs, this tendency is less strong for attitudinal dis- juncts (compare the examples in n.12). For the adverbs studied below, I see insufficient evidence as yet that adjunct versus dis- junct status matters to their position in the clause. This lack of differentiation may simply be due to the fact that superlatives generally will present highly salient information, and are there- fore likely to appear in preverbal Focus position, which will often be clause-initial.13 µάλιστα, πλεῖστα, ἥκιστα, ἐλάχιστα as expressions of likelihood Having discussed the sparse references in the literature to these expressions of probability, I now turn to further adverbs which, like ῥεῖα and ῥᾳδίως, can play a variety of roles in clauses, but which, unlike these, must be translated differently depending on their role in the clause. Translators and com- 13 This statement assumes a general clause pattern in classical Greek as proposed in H. Dik, Word Order in Ancient Greek: A Pragmatic Account of Word Order Variation in Herodotus (Amsterdam 1995), and Word Order in Greek Tragic Dialogue (Oxford 2007). Refinements and modifications have since been pro- posed, among others by D. Matić, “Topic, Focus, and Discourse Structure: Ancient Greek Word Order,” Studies in Language 27 (2003) 573–633. HELMA DIK 605 ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 mentators have often overlooked that a number of quite fre- quent adverbs can be used to express a very high or low degree of likelihood. I will primarily discuss instances from Thucydides’ Histories, in which I first noticed the phenomenon, but this is by no means in order to claim that this construction is exclusively Thu- cydidean. It is, however, a staple of Thucydidean discourse to speculate about an opponent’s strategy and it is such contexts where, I argue, we can see these adverbs used to specify very high or very low probability. Since I have been able to find as little discussion in commentaries as in the reference grammars, translations are practically my only evidence for earlier scholars’ interpretations, and I will therefore quote from these trans- lations quite extensively. My aim in doing so is twofold: obvi- ously, translations that do use expresssions of likelihood support the present argument; on the other hand, translations that do not use such expressions are often quite unsatisfactory in their treatment of the adverbs in question, showing, one hopes, that the present paper is filling a lacuna in the scholarship. In any case, the presence of third-party translations should make it easier for readers to make up their own minds. The adverbs I will concentrate on are µάλιστα and πλεῖστα to express high likelihood, and ἥκιστα and ἐλάχιστα to express low likelihood; I will include a few examples of the compara- tives µᾶλλον and ἧσσον in the notes. In contrast to τάχα and ἴσως, these adverbs are all quite frequent in Thucydides, and they are found in many more contexts than statements of potential events. As a heuristic for finding such statements, I have looked for combinations of the adverbs with ἄν.14 To be sure, the modal particle ἄν is not a necessary ingredient for 14 Since the presence of ἄν is indeed only a heuristic, one would expect this use of the adverbs to extend beyond this particular environment. Indeed, I would argue that despite the absence of explicit modal expression Hdt. 8.136.2 τοὺς γὰρ Ἀθηναίους οὕτω ἐδόκεε µάλιστα προσκτήσεσθαι is also best translated as “he thought he would be most likely to win over the Athen- ians in this way.” 606 ‘MOST LIKELY TO SUCCEED’ ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 statements of potential or likely events. Likelihood can be ex- pressed lexically by means of εἰκός, for instance, or it can be left implicit. In Aristotle’s Rhetoric, we find statements not containing explicit modals explained by statements that do contain these, as at 1397b12–13:15 ἄλλος ἐκ τοῦ µᾶλλον καὶ ἧττον, οἷον “εἰ µηδ’ οἱ θεοὶ πάντα ἴσασιν, σχολῇ οἵ γε ἄνθρωποι”· τοῦτο γάρ ἐστιν “εἰ ᾧ µᾶλλον ἂν ὑπάρχοι µὴ ὑπάρχει, δῆλον ὅτι οὐδ’ ᾧ ἧττον”. τὸ δ’ ὅτι τοὺς πλησίον τύπτει ὅς γε καὶ τὸν πατέρα ἐκ τοῦ “εἰ τὸ ἧττον <ὑπάρχον> ὑπάρχει, καὶ τὸ µᾶλλον ὑπάρχει”· τοὺς γὰρ πατέρας ἧττον τύπτουσιν ἢ τοὺς πλησίον· Another topic is derived from the more and less. For instance, if not even the gods know everything, hardly [can] men; for this amounts to saying that if a predicate, which is more probably affirmable of one thing, does not belong to it, it is clear that it does not belong to another of which it is less probably affirmable. And to say that a man who beats his father also beats his neigh- bors, is an instance of the rule that, if the less exists, the more also exists. For people are less likely to hit their father than their neigh- bors. Despite the absence of explicit modals, we can easily under- stand them following an adverb such as σχολῇ ‘scarcely’; similarly, I translate ἧττον τύπτουσιν with ‘are less likely to hit’, not ‘hit less’, despite the indicative mood. Turning, however, to the collocations with ἄν now, only about one in fifteen occurrences of µάλιστα in Thucydides co- occurs with ἄν within a clause that is either a potential optative, or an infinitive in indirect statement. In the table below, I give the number of occurrences of these words (and of τάχα16 and 15 The last sentence of this passage is included in the Ross’s OCT edition, but not found in all mss. The translation of the passage up to that point is that of the Loeb edition, by J. H. Freese (1926). 16 To the few instances of τάχα could be added ταχύ in 6.17.4. The one instance in which τάχα is not accompanied by ἄν (8.94) does not speculate about an eventuality but about the reason for an action that actually took place: τάχα µέν τι καὶ ἀπὸ ξυγκειµένου λόγου, “perhaps because of a prior arrangement.” HELMA DIK 607 ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 ἴσως for the sake of comparison) immediately adjacent to ἄν, separated at most by δέ or γάρ, as well as their total number of occurrences. These are the instances where the two words will most likely be felt as a unit. There are also instances of further separation, in which I would nevertheless argue for a likelihood interpretation, but most examples where I would argue for a likelihood interpretation have immediate adjacency.17 Adverb Adverb + ἄν Total occurrences τάχα 10 11 ἴσως 4 10 πλεῖστα 7 34 µάλιστα 19 316 ἥκιστα 8 57 ἐλάχιστα 2 10 µᾶλλον 8 363 ἧσσον 2 72 It will be immediately apparent that the proposed interpre- tation will hold for only a small minority of the total instances of use of these adverbs, and the question arises how one would go about demonstrating that other possible translations should be discarded. Clearly, it would be necessary to show that more frequent interpretations such as degree expressions at the level of the predicate itself (‘they were most fortunate’), or internal object interpretations (‘they won the fewest victories’) are excluded. In making the case, then, I will focus on clauses in which degree expressions or internal objects are incompatible 17 Giving an ‘exact’ count is not as straightforward as it may seem. There is in fact only one case of ἴσως ἄν, in Thuc. 2.20.2. The other three counted also feature τάχα, as in τάχ’ ἂν ἴσως. Then, beyond the four passages counted, two more that feature ἴσως and ἄν in the same sentence have not been counted because they are not immediately adjacent: 1.82.3 ἴσως … µᾶλλον ἂν εἴκοιεν and 6.11.3 κἂν ἔλθοιεν ἴσως. In sum, six out of ten in- stances of ἴσως are found within the same clause as ἄν, while one could argue that ἴσως ἄν functions as a unit only once. 608 ‘MOST LIKELY TO SUCCEED’ ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 with the predicate, and on those where other Manner expres- sions or internal objects are present. First, however, I will illu- strate the conundrum one faces when the particular predicate is compatible with various interpretations of the adverb. In fact, the title of this paper is my proposed translation of a phrase that Thucydides uses four times in all: πλεῖστ’ ἂν ὀρθοῖ(ν)το. Below is the first occurrence of the phrase in the Histories (3.30.4), and the various translations by Smith (Loeb, 1919–1923), Warner (Penguin, 1954), and Crawley (Every- man’s Library, 1874): (7) καὶ µὴ ἀποκνήσωµεν τὸν κίνδυνον, νοµίσαντες οὐκ ἄλλο τι εἶναι τὸ κενὸν τοῦ πολέµου ἢ τὸ τοιοῦτον, ὃ εἴ τις στρατηγὸς ἔν τε αὑτῷ φυλάσσοιτο καὶ τοῖς πολεµίοις ἐνορῶν ἐπιχειροίη, πλεῖστ’ ἂν ὀρθοῖτο.” Let us not shrink from the danger, remembering that the element of surprise in warfare is precisely of this nature. And if a general guards against such surprises in his own case, and, whenever he sees an opportunity to employ them in the case of the enemy, makes the attempt, he will win the greatest success (Smith) … a good general is the one who … (Warner) … is what makes a successful general (Crawley) For πλεῖστα, an interpretation as internal object is not impos- sible here. The use of the present tense (φυλάσσοιτο, ἐπιχει- ροίη, ὀρθοῖτο) suggests a habitual attitude which can lead to success on multiple occasions. A translation as ‘greatest success’, however, is problematic, since ὀρθόω, unlike English ‘prosper’ or ‘succeed’, is not compatible with degree expressions (one stays upright, or, in the negative, one fails to do so). All four instances of the phrase (repeated at 3.42.4, 5.9.4, and 5.111.4) allow for an interpretation of πλεῖστ’ ἂν ὀρθοῖτο as ‘would be successful on most occasions’.18 I would argue, however, that they would also lend themselves to an interpretation as ‘would be 18 A similar argument can be made for 3.37.4, ὀρθοῦνται τὰ πλείω, again in a generalizing context, with frequency and probability two sides of the same coin. See below. HELMA DIK 609 ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 most likely to succeed’. The following passage is of interest, in that the antonym ἐλάχιστα follows in the next sentence, and one of the translators, Smith, in fact adopts an interpretation of this second adverb as an expression of likelihood.19 However, it will be seen that Crawley (‘fewer blunders’) follows Classen- Steup’s lead in construing ἐλάχιστα as the internal object of ἁµαρτάνειν, which is equally defensible. 3.42.4: (8) καὶ πλεῖστ’ ἂν ὀρθοῖτο ἀδυνάτους λέγειν ἔχουσα τοὺς τοιού- τους τῶν πολιτῶν· ἐλάχιστα γὰρ ἂν πεισθεῖεν ἁµαρτάνειν. indeed, [the city] would prosper most if its citizens of this stamp had no eloquence at all, for then the people would be least likely to blunder through their influence (Smith) it would be better for the country if they could not speak at all, as we should then make fewer blunders (Crawley) it would be the best possible thing for the city (Warner) In the remaining two instances of the phrase, Smith adopts ‘most likely’ in his translation of πλεῖστ’ ἂν.20 Questions of translations aside, however, I should point out that in the case of a description of repeated, or repeatable, situations, the dis- tinction between a frequency expression and a likelihood ex- pression is really a distinction without a difference: ‘will succeed at nine out of ten occasions’ is equivalent to ‘is ninety percent likely to succeed’. 21 19 For a similar pairing compare Pl. Prm. 137B τίς οὖν, εἰπεῖν, µοὶ ἀποκρι- νεῖται; ἢ ὁ νεώτατος; ἥκιστα γὰρ ἂν πολυπραγµονοῖ, καὶ ἃ οἴεται µάλιστα ἂν ἀποκρίνοιτο, “Who then,” said he, “is to answer my questions? Shall we say the youngest? He would be least likely to be over-curious and most likely to say what he thinks” (transl. Fowler, Loeb, 1926). 20 5.9.4 ὅστις δὲ τὰς τοιαύτας ἁµαρτίας τῶν ἐναντίων κάλλιστα ἰδὼν καὶ ἅµα πρὸς τὴν ἑαυτοῦ δύναµιν τὴν ἐπιχείρησιν ποιεῖται µὴ ἀπὸ τοῦ προφα- νοῦς µᾶλλον καὶ ἀντιπαραταχθέντος ἢ ἐκ τοῦ πρὸς τὸ παρὸν ξυµφέροντος, πλεῖστ’ ἂν ὀρθοῖτο, “would be most likely to succeed”; 5.111.4 ὡς οἵτινες τοῖς µὲν ἴσοις µὴ εἴκουσι, τοῖς δὲ κρείσσοσι καλῶς προσφέρονται, πρὸς δὲ τοὺς ἥσσους µέτριοί εἰσι, πλεῖστ’ ἂν ὀρθοῖντο, “are most likely to prosper.” 21 A further quite similar example is in 6.9.2, νοµίζων ὁµοίως ἀγαθὸν πολίτην εἶναι ὃς ἂν καὶ τοῦ σώµατός τι καὶ τῆς οὐσίας προνοῆται· µάλιστα γὰρ ἂν ὁ τοιοῦτος καὶ τὰ τῆς πόλεως δι’ ἑαυτὸν βούλοιτο ὀρθοῦσθαι. 610 ‘MOST LIKELY TO SUCCEED’ ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 In order to make the case, then, that an expression of likeli- hood is indeed a good reading for which support can be found elsewhere in Thucydides, the following discussion will consider examples in which both a degree expression and an internal object seem problematic, due to the semantics of the verb or the presence of other internal objects or degree expressions in the sentence. Compatibility of the adverb with the predicate In (9), the potential statement happens to be in indirect dis- course (ἡγούµενοι … µάλιστ’ ἂν φέρεσθαι), but this is im- material to the argument that the adverb µάλιστα in this case expresses a high degree of likelihood that the event will come to pass (3.53.1): (9) τὴν µὲν παράδοσιν τῆς πόλεως, ὦ Λακεδαιµόνιοι, πιστεύσαν- τες ὑµῖν ἐποιησάµεθα, οὐ τοιάνδε δίκην οἰόµενοι ὑφέξειν, νοµιµωτέραν δέ τινα ἔσεσθαι, καὶ ἐν δικασταῖς οὐκ ἂν ἄλλοις δεξάµενοι, ὥσπερ καὶ ἐσµέν, γενέσθαι ἢ ὑµῖν, ἡγούµενοι τὸ ἴσον µάλιστ’ ἂν φέρεσθαι. When we surrendered our city, Lacedaemonians, trusting in your good faith, we had no thought that we should have to undergo a trial like this, but supposed it would be a more regular procedure; and when we consented to be on trial before you, and you alone as judges, as we now are, we believed that we should be most likely to obtain fair treatment (Smith). The presence of the definite τὸ ἴσον ‘justice’ makes it hard to interpret µάλιστα in the normal sense of ‘to the highest degree’. Smith as well as Crawley and Warner translate µάλιστα with ‘most likely’.22 Classen-Steup do not offer comment on µάλι- στα, but translate with ‘sicherlich’, ‘certainly’.23 This, however, obscures the fact that comparison still plays a role: The Spar- tans, compared to others, are considered most likely to do the right thing. 22 Unlike πλεῖστα, of course, µάλιστα can only be an adverb, so that the internal object interpretation is excluded as well. 23 J. Classen and J. Steup, Thukydides. Erklärt von J. Classen. Bearbeitet von J. Steup III (Berlin 1892) 107. HELMA DIK 611 ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 Demosthenes offers a parallel, in which the verb and its complement, ‘obtain the peace treaty’, again would not seem to be compatible with an expression of degree or extent (Dem. 19.317): (10) συνέγραψε δ’ ἐπιστολὴν ὡς ὑµᾶς, ᾗ µάλιστ’ ἂν ᾤετο τῆς εἰ- ρήνης τυχεῖν. He [Philip] then composed a letter to you, as the best means of ob- taining the peace he desired (transl. Vince, Loeb 1926). The translator appears to construe µάλιστα closely with the rel- ative pronoun (‘the best means’), but what this means is rather unclear.24 Instead, I propose: “a letter, with which he thought he was most likely to obtain the peace treaty.” A fairly close parallel is offered by clauses with the phrasing οὕτως ἂν µάλιστα, as in (11), again from Demosthenes. An adoption is deemed most likely to lead to the preservation of a family (43.12): (11) ἡγούµενος, ὦ ἄνδρες δικασταί, ἐκ τῶν ὑπολοίπων τοῦτον εἶναι ἑαυτῷ οἰκειότατον, καὶ οὕτως ἂν µάλιστα τὸν οἶκον τὸν ἑαυτῶν διασῴζεσθαι καὶ οὐκ ἂν ἐξερηµωθῆναι. for he thought, men of the jury, that of his surviving relatives this boy was nearest to him, and that in this way their house would best be preserved and kept from extinction (Vince). While the translator inserts ‘best’, I believe διασῴζεσθαι is another instance of a verb that is not readily compatible with degree expressions: survival is what is at stake here. The full string used here, οὕτως ἂν µάλιστα, is a frequent collocation.25 24 One of the things ‘the best means’ could mean is the means most likely to be successful (as in, ‘email is the best way to reach me’ is ‘you are most likely to reach me if you use email’). But it is presumably not the medium, but the content that has been selected for its persuasiveness. 25 Cf. Lys. 2.1, διὰ τοῦτο καὶ ἡ πόλις µοι δοκεῖ, προνοουµένη τῶν ἐνθάδε λεγόντων, ἐξ ὀλίγου τὴν πρόσταξιν ποιεῖσθαι, ἡγουµένη οὕτως ἂν µάλιστα συγγνώµης αὐτοὺς παρὰ τῶν ἀκουσάντων τυγχάνειν. Here again I would understand µάλιστα as ‘most likely’ rather than adding a specification of ‘degree’ of συγγνώµη (regularly expressed by forms of πολύς rather than adverbially). 612 ‘MOST LIKELY TO SUCCEED’ ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 Returning to the Thucydidean material, in 3.71.1 the po- tential optative is a form of δουλόω ‘enslave’; ἥκιστα, I would argue, is used in close conjunction with ἄν to express a low degree of likelihood (in this case, Smith’s translation in fact uses such wording): (12) δράσαντες δὲ τοῦτο καὶ ξυγκαλέσαντες Κερκυραίους εἶπον ὅτι ταῦτα καὶ βέλτιστα εἴη καὶ ἥκιστ’ ἂν δουλωθεῖεν ὑπ’ Ἀθηναίων, τό τε λοιπὸν µηδετέρους δέχεσθαι ἀλλ’ ἢ µιᾷ νηὶ ἡσυχάζοντας, τὸ δὲ πλέον πολέµιον ἡγεῖσθαι. After they had taken these measures the conspirators called the Corcyraeans together and told them that it was all for the best, and that now they would be least likely to be enslaved by the Athenians; and in future they should remain neutral and receive neither party if they came with more than one ship, regarding any larger number as hostile (Smith) … what they had done was all for the best and would prevent the island being enslaved by Athens … (Warner) An expression of degree like ἥκιστα ‘to the least extent’ would seem incompatible with this verb, as with ὀρθόω.26 ἥκιστα is sometimes equated with a strong negation (‘minime’);27 however, an actual negation here would lead to a reading ‘that it was not possible that they would be enslaved’, which is not wanted here.28 26 Note, however, that ἥκιστα is also combined with a future indicative in an indirect question in 6.82.2: ἐσκεψάµεθα ὅτῳ τρόπῳ ἥκιστα αὐτῶν ὑπακουσόµεθα, “how to be least subservient,” i.e., how to avoid, but perhaps also, how to be least likely to… 27 One instance of ἥκιστα ἄν in Plato I find hard to decide, but I lean toward a ‘minime’ as opposed to a ‘least likely’ interpretation: “σκώπτεις, ὦ Σώκρατες.” “οὐ µὰ τὸν Φίλιον τὸν ἐµόν τε καὶ σόν, ὃν ἐγὼ ἥκιστ’ ἂν ἐπιορκήσαιµι”: “You are joking, Socrates.” “No, I swear by our common God of Friendship, whose name I would never (be least likely to) take in vain” (Alc. 1 109D). 28 6.82.3 is quite similar: αὐτοὶ δὲ τῶν ὑπὸ βασιλεῖ πρότερον ὄντων ἡγεµό- νες καταστάντες οἰκοῦµεν, νοµίσαντες ἥκιστ’ ἂν ὑπὸ Πελοποννησίοις οὕτως εἶναι, δύναµιν ἔχοντες ᾗ ἀµυνούµεθα, καὶ ἐς τὸ ἀκριβὲς εἰπεῖν οὐδὲ ἀδίκως καταστρεψάµενοι τούς τε Ἴωνας καὶ νησιώτας, οὓς ξυγγενεῖς φασὶν ὄντας ἡµᾶς Συρακόσιοι δεδουλῶσθαι, “Having ourselves become leaders of those HELMA DIK 613 ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 An example from extended indirect discourse is at 5.36.1: (13) οὕτω γὰρ ἥκιστ’ ἂν ἀναγκασθῆναι Βοιωτοὺς ἐς τὰς Ἀττικὰς σπονδὰς ἐσελθεῖν· ἑλέσθαι γὰρ Λακεδαιµονίους πρὸ τῆς Ἀθη- ναίων ἔχθρας καὶ διαλύσεως τῶν σπονδῶν Ἀργείους σφίσι φίλους καὶ ξυµµάχους γενέσθαι. For in this way the Boeotians would be least likely to be forced to come into the treaty with Athens, since the Lacedaemonians would prefer gaining the friendship and alliance of the Argives, counting that more important than the enmity of the Athenians and the disruption of the treaty (Smith). Again, with the verb ἀναγκάζω a degree reading seems to make little sense—the Boeotians would not be forced to a lesser de- gree; rather, the proposed strategy (οὕτως) makes their alliance with Athens less likely.29 Presence of another adverb A further factor that points in the direction of interpreting ___ who were before subject to the King, we so continue, thinking that we should in this way be least subject to the Peloponnesians, because we have power with which to defend ourselves” (Smith). I propose to translate in- stead: “that we should in this way be least likely to be subject…” Also 3.24.1: ἐν δεξιᾷ ἔχοντες τὸ τοῦ Ἀνδροκράτους ἡρῷον, νοµίζοντες ἥκιστ’ <ἂν> σφᾶς ταύτην αὐτοὺς ὑποτοπῆσαι τραπέσθαι τὴν ἐς τοὺς πολεµίους, “[The Pla- taeans] keeping the sanctuary of the hero Androcrates on the right, believing that they [the Peloponnesians] would be least likely to suspect them of taking that route, which led to the enemy.” 29 In 6.49.2 µάλιστα seems equally incompatible with περιγενέσθαι ‘sur- vive; come out on top’: αἰφνίδιοι δὲ ἢν προσπέσωσιν, ἕως ἔτι περιδεεῖς προσδέχονται, µάλιστ’ ἂν σφεῖς περιγενέσθαι. 4.80.3 is a similar case, with a degree expression incompatible with ἐπιθέσθαι: προεῖπον αὐτῶν ὅσοι ἀξιοῦ- σιν ἐν τοῖς πολέµοις γεγενῆσθαι σφίσιν ἄριστοι, κρίνεσθαι, ὡς ἐλευθερώ- σοντες, πεῖραν ποιούµενοι καὶ ἡγούµενοι τούτους σφίσιν ὑπὸ φρονήµατος, οἵπερ καὶ ἠξίωσαν πρῶτος ἕκαστος ἐλευθεροῦσθαι, µάλιστα ἂν καὶ ἐπι- θέσθαι, “They made proclamation that all Helots who claimed to have rendered the Lacedaemonians the best service in war should be set apart, ostensibly to be set free. They were, in fact, merely testing them, thinking that those who claimed, each for himself, the first right to be set free would be precisely the men of high spirit who would be the most likely to attack their masters.” 614 ‘MOST LIKELY TO SUCCEED’ ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 µάλιστα as a disjunct rather than as an adjunct is the presence of other adverbs in the clause (4.19.2): (14) νοµίζοµέν τε τὰς µεγάλας ἔχθρας µάλιστ’ ἂν διαλύεσθαι βεβαίως, οὐκ ἢν ἀνταµυνόµενός τις καὶ ἐπικρατήσας τὰ πλείω τοῦ πολέµου κατ’ ἀνάγκην ὅρκοις ἐγκαταλαµβάνων µὴ ἀπὸ τοῦ ἴσου ξυµβῇ, ἀλλ’ ἢν παρὸν τὸ αὐτὸ δρᾶσαι πρὸς τὸ ἐπιεικὲς καὶ ἀρετῇ αὐτὸν νικήσας παρὰ ἃ προσεδέχετο µετρίως ξυναλλαγῇ. We believe, too, that a permanent reconciliation of bitter en- mities is more likely to be secured, not when one party seeks revenge and, because he has gained a decided mastery in the war, tries to bind his opponent by compulsory oaths and thus makes peace with him on unequal terms… (Smith) Indeed if great enmities are ever to be really settled, we think it will be, not by the system of revenge and military success… (Craw- ley) As Smith saw, it is unlikely that µάλιστα would be used to fur- ther modify βεβαίως (the superlative βεβαιότατα occurs at 1.138.3 and 6.91.6);30 similarly, in 6.24.1 ἀσφαλῶς is already present as a modifier of ἐκπλεῦσαι: (15) ὁ µὲν Νικίας τοσαῦτα εἶπε νοµίζων τοὺς Ἀθηναίους τῷ πλήθει τῶν πραγµάτων ἢ ἀποτρέψειν ἤ, εἰ ἀναγκάζοιτο στρατεύεσθαι, µάλιστ’ <ἂν> οὕτως ἀσφαλῶς ἐκπλεῦσαι. So much Nicias said, thinking that he would deter the Athenians by the multitude of his requirements, or, if he should be forced to make the expedition, he would in this way set out most safely. I propose to translate instead: “in this way, he would be most likely to set out safely.”31 Conclusion More than has been practiced in translations so far, I believe 30 The superlative ἀσφαλέστατα is found in 1.107.4 and 6.18.7. 31 Further examples with both µάλιστα and οὕτως (‘do such and such, for in this way it would be most likely that…’) in 6.38.5 µάλιστα γὰρ δοκῶ ἄν µοι οὕτως ἀποτρέπειν τῆς κακουργίας and 7.8.2 ἔγραψεν ἐπιστολήν, νοµίζων οὕτως ἂν µάλιστα τὴν αὑτοῦ γνώµην µηδὲν ἐν τῷ ἀγγέλῳ ἀφανισθεῖσαν µα- θόντας τοὺς Ἀθηναίους βουλεύσασθαι περὶ τῆς ἀληθείας. HELMA DIK 615 ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 µάλιστα and the other adverbs discussed here should be in- terpreted as expressions of likelihood. In this paper I have not exhausted the possible examples from Thucydides and barely touched other authors. In conclusion, I should therefore per- haps state explicitly that I do not consider this use of degree adverbs to specify likelihood a feature of Thucydides’ style or a marker of elevated prose style more generally. In Xenophon’s manual on horsemanship we find an example as part of his advice on selecting a good horse, where the point is not finding the best possible horse (Eq. 1.17): (16) εἶδος µὲν δὴ πώλου οὕτω δοκιµάζοντες µάλιστ’ ἂν ἡµῖν δοκοῦσι τυγχάνειν εὔποδος καὶ ἰσχυροῦ καὶ εὐσάρκου καὶ εὐσχήµονος καὶ εὐµεγέθους. He who applies these tests to a colt’s shape is sure, in my opinion, to get a beast with good feet, strong, muscular, of the right look and the right size. (transl. Marchant, Loeb, 1925) Similarly, in Aristophanes’ Lysistrata Calonice asks whether ab- stinence will increase the likelihood of peace (146–148): (17) εἰ δ’ ὡς µάλιστ’ ἀπεχοίµεθ’ οὗ σὺ δὴ λέγεις, – ὃ µὴ γένοιτο, – µᾶλλον ἂν διὰ τουτογὶ γένοιτ’ ἂν εἰρήνη; Well, what if we did abstain from, uh, what you say, which heaven forbid: would peace be likelier to come on that account? (transl. Henderson, Loeb, 2000) Neither is it the case that I am proposing that every instance of µάλιστ’ ἂν be interpreted the same way. In Thuc. 5.9.5, a superlative is needed with ἀπατήσας to balance καλλίστην and µέγιστα: (18) καὶ τὰ κλέµµατα ταῦτα καλλίστην δόξαν ἔχει ἃ τὸν πολέµιον µάλιστ’ ἄν τις ἀπατήσας τοὺς φίλους µέγιστ’ ἂν ὠφελήσειεν. And these stratagems, which do the greatest service to our friends by most completely deceiving our enemies, have the most bril- liant name in war. (Crawley) In quite a few cases, the interpretation is not certain. Is one most likely to terrify the enemy, or could one terrify them most (6.34.4, µάλιστ’ ἂν αὐτοὺς ἐκπλήξαιµεν)? I will conclude with a familiar passage to illustrate the problem. I believe Thucydi- 616 ‘MOST LIKELY TO SUCCEED’ ————— Greek, Roman, and Byzantine Studies 54 (2014) 599–616 des to say in 2.48.3 that someone with prior knowledge would be the most likely to be able to recognize a return of the plague, but translators and commentators offer a variety of transla- tions:32 (19) ἐγὼ δὲ οἷόν τε ἐγίγνετο λέξω, καὶ ἀφ’ ὧν ἄν τις σκοπῶν, εἴ ποτε καὶ αὖθις ἐπιπέσοι, µάλιστ’ ἂν ἔχοι τι προειδὼς µὴ ἀγνοεῖν, ταῦτα δηλώσω… I shall reveal the things [symptoms] which, if someone should investigate [the disease] on their basis and gain some prior knowledge [of it], he would be most able to recognise it, if it should ever occur again…33 … best able (Smith) … explain the symptoms by which perhaps it may be recognized (Crawley) … knowledge of which would enable it to be recognized (War- ner) This paper will have fulfilled its purpose if it makes readers more likely to consider this further option when encountering degree adverbs.34 August, 2014 Department of Classics University of Chicago Chicago, IL 60637 helmadik@uchicago.edu 32 Classen-Steup (II 81) do not translate, but it appears that they connect µάλιστα with προειδώς: “es ist das [µὴ ἀγνοεῖν] aber möglich durch das προ- ειδώς τι, wenn man sie soviel wie möglich kennen gelernt hat.” 33 Transl. J. Rusten, Thucydides: The Peloponnesian War Book II (Cambridge 1989) 182. 34 I first became aware of this phenomenon while reading Thucydides Book 5 with my students Branden Kosch and Jonah Radding. I thank them, as well as Elizabeth Asmis, Harm Pinkster, David Sansone, Ruth Scodel, and Peter White for their encouragement and helpful comments. A version of this paper was presented at the APA in Chicago in January 2014.