I

The common rendering ‘and when once they have heard’ in Heraclitus B1.(i) produces a peculiar and unproductive communicative situation: if human beings turn out to lack understanding both before and after they have heard the logos,⁵ then why should Heraclitus bother to go about his task of describing how the world is? If neither first-hand experience nor listening ever provides a path to understanding, why should his audience bother to keep listening to him?⁶ Indeed, if Heraclitus affirms here without qualification that we quite simply never acquire understanding, then it would be difficult not to come away with the impression that this is because understanding is altogether beyond our reach. At stake here, then, is how Heraclitus sees what people can achieve by engaging properly with his book.

On a stylistic level, Heraclitus is not one for superfluity or prolixity, but with the translation ‘once’ τὸ πρῶτον is otiose. The pairing of before and after hearing, clearly signalled by the correlative pairing of καί … καί, could have been adequately expressed by the simple aorist participle ἀκούσαντες alone (cf. B34: ἀξύνετοι ἀκούσαντες κωφοῖσιν ἐοίκασι, ‘people without understanding, after they have heard, resemble deaf people’). In effect, the rendering ‘once they have heard’ erases a qualification to the contrast that Heraclitus himself foregrounds in B1. Indeed, the placement of τὸ πρῶτον at the end of its sentence might suggest that these words carry a certain degree of emphasis. Idiom hardly demands that πρόσθεν ἤ … τὸ πρῶτον be taken together to mean ‘before X and once X has occurred’; indeed, we are unable to offer a good linguistic parallel for such a construction.⁷

On a semantic level, it is very far from clear whether τὸ πρῶτον can have a sense so weak as it has in the translations quoted above. None of these scholars adduces specific parallels, but Kirk, one of the few to justify his choice, writes that this expression ‘means “once”, “at all”, as frequently in Homer’.⁸ He cites LSJ⁹ s.v. πρότερος B.III.3.e, which begins: ‘πρῶτον, πρῶτα are used after the relat[ive] Pron[oun] and after relat[ive] Adv[er]bs, like Engl[ish] once (= at all)’. In Heraclitus B1, however, the relative adverb πρόσθεν does not govern τὸ πρῶτον.⁹ In any event, the recent Cambridge Greek Lexicon s.v. πρῶτον, πρῶτα does not recognise any such sense, presumably subsuming the relevant cases under the heading of ‘in the first place, at the outset’ (3), and with reason.¹⁰ Markwald’s entry in the Lexikon des frühgriechischen Epos s.v. πρῶτος, πρώτιστος 3 systematically catalogues every adverbial occurrence in traditional epic (1599–1604). The basic meaning, according to Markwald, is ‘zuerst, am Anfang’ (1599). In conjunction with a subordinate clause, the adverb can mean ‘1a sobald, 1b seitdem, wenn erst einmal, 1c nachdem (wenn) erst einmal, da einmal, 1d bevor’ (1604). Cunliffe s.v. πρῶτος offers the most detailed analysis of adverbial Homeric usage, distinguishing no fewer than 12 distinct senses; none of these correspond to the entry in LSJ⁹ quoted above.¹¹ The sort of usage that LSJ⁹ seem to have in mind is exemplified by passages like Hesiod, Theogony 740–1: χάσμα μέγ’, οὐδέ κε πάντα τελεσφόρον εἰς ἐνιαυτόν | οὖδας ἵκοιτ’, εἰ πρῶτα πυλέων ἔντοσθε γένοιτο, ‘a great chasm, whose base a man could not reach even in a whole year if once he got inside the gates’. But here εἰ πρῶτα means ‘“if once”, literally “if he began by” entering’ (West): as elsewhere in such cases, the adverb marks the beginning of a process that cannot be, or will not be, undone or reversed once it has begun.¹² Heraclitus B1 is self-evidently different.

We advocate an alternative understanding of ἀκούσαντες τὸ πρῶτον in Heraclitus B1 that affords these words due significance, obviates the peculiar communicative situation described above, and also produces a more plausible and interesting overall interpretation of the fragment: ‘after they have heard it for the first time’.

At the outset it is worthwhile to disambiguate this rendering—(A)—from another variation of the alternative construal: ‘after they first have heard it’ (B). Reading (A) more strongly suggests subsequent discrete iterations of some process taken as a whole (as in: ‘when I shaved my face for the first time’). Somewhat differently, reading (B) gestures towards the initial phase of some activity, process, event or sequence of events in a less determinate way that, depending on the context, could suggest subsequent discrete iterations (as in: ‘when I first shaved my face’) but need not do so (as in: ‘when we first started painting this room’).

It is a fine distinction between these two variations of the alternative construal of τὸ πρῶτον. Indeed, these two closely related senses of the same Greek words may in practice blend into each other, and there is no strong reason to suppose that Heraclitus sharply distinguished between them. Either variant avoids the difficulties with the rendering ‘once/after they have heard’. Still, we submit that the sense ‘for the first time’ best captures the qualified warning that Heraclitus issues to us as we embark on our first reading of his book. The cultural practice of repeated reading (to which we relate Heraclitus below) suggests this. Furthermore, the idea of multiple readings of Heraclitus’ book, and of the individual remarks of which this book is comprised, will pay interpretive dividends (as we will see in Section III below). Most likely, then, Heraclitus is telling us that we will not comprehend his book when we hear or read it ‘for the first time’. To accept this is not, however, to exclude that, in line with sense (B), Heraclitus could also be conveying to us (perhaps as a secondary connotation) that things may be especially difficult in the early goings (‘at first’) but then get easier as we proceed.¹³

Both variations of the alternative construal are easy to parallel. Passages that are most plausibly rendered in line with sense (A) are common in Greek generally (LSJ⁹ s.v. πρότερος III.3.d), in Homer (Cunliffe 1924, s.v. πρῶτος 6d), and in Herodotus, who is our best comparandum for Heraclitus’ early Ionic prose.¹⁴ Consider, for example, Odyssey 4.158–9 (νεμεσσᾶται δ’ ἐνὶ θυμῷ | ὧδ’ ἐλθὼν τὸ πρῶτον ἐπεσβολίας ἀναφαίνειν, ‘he feels it wrong in his heart, thus having come before you for the first time, to bring forth uninvited words’), Pindar, Pythian 9.41 (ἀμφανδὸν ἁδείας τυχεῖν τὸ πρῶτον εὐνᾶς, ‘to openly engage in sweet sex for the first time’) or Herodotus 1.60.1 (οὕτω μὲν Πεισίστρατος ἔσχε τὸ πρῶτον Ἀθήνας, ‘thus Peisistratus took Athens for the first time’). As these examples show, when τὸ πρῶτον means ‘for the first time’, there need not be any explicit indication of subsequent iterations; context is decisive in each case.

Passages that suggest sense (B) are also common in Greek generally (LSJ⁹ s.v. πρότερος BIII.3.a), in Homer (Cunliffe 1924, s.v. πρῶτος 6b), and in Herodotus (Powell 1938, s.v. πρῶτος 327). Consider, for example, the Homeric Hymn to Apollo 71 (ὁπότ’ ἂν τὸ πρῶτον ἴδῃ φάος ἠελίοιο, ‘whenever he first sees the light of the sun’), Pindar, Pythian 4.29–31 (φιλίων δ’ ἐπέων | ἄρχετο, ξείνοις ἅτ’ ἐλθόντεσσιν εὐεργέται | δεῖπν’ ἐπαγγέλλοντι πρῶτον, ‘he made a beginning of friendly words, the sort which generous men first announce when offering dinner to strangers who have arrived’) or Herodotus 2.125.2 (τοιαύτην τὸ πρῶτον ἐπείτε ἐποίησαν αὐτήν, ἤειρον τοὺς ἐπιλοίπους λίθους μηχανῇσι ξύλων βραχέων πεποιημένῃσι, ‘after they first made such a step [sc. as the first part in the process of building a pyramid], they raised the rest of the stones with contrivances made of short wooden logs’).

Hereafter we give priority in our translations and glosses to our preferred variation, according to which Heraclitus is telling us that people ever fail to understand both before they hear and when they have heard ‘for the first time’ (A). But we take our suggestions to be also compatible with—even if, in some cases, not quite as sharply captured by—what is in the end a slightly different version of broadly the same construal, according to which people fail to understand both before they hear and when ‘first’ or ‘at first’ they have heard (B).

On the common interpretation ‘once’, Heraclitus’ book opens with a remarkable repudiation of its readership, which forecloses their efforts at understanding him as inevitably doomed to failure. On the alternative interpretation, his work begins with a more plausible and effective didactic gesture which is at once a warning, a challenge, and an invitation: readers are cautioned that they will not grasp this text (or its subject-matter, the cosmic logos) immediately or on their first hearing, but they are implicitly encouraged to keep trying to understand it. Compare, for example, Hesiod’s famous words to Perses (Works and Days 290–2): μακρὸς δὲ καὶ ὄρθιος οἶμος ἐς αὐτήν | καὶ τρηχὺς τὸ πρῶτον· ἐπὴν δ’ εἰς ἄκρον ἵκηται, | ῥηιδίη δὴ ἔπειτα πέλει, χαλεπή περ ἐοῦσα, ‘long and steep is the path to [aretē] and rough at first; but once one reaches the summit, the road becomes easy, difficult though it is’.

On this reading of B1, Heraclitus implicitly recommends a special sort of sustained engagement with his text. Compare Empedocles DK31 B110:

If you firmly place these things beneath your crowded mind and behold them graciously with immaculate practice, all these things will be present to you throughout your life, and you will acquire many other things from these.

Pausanias should store up Empedocles’ teaching in his mind and then subsequently reflect upon it (ἐρείσας … ἐποπτεύσῃς). As Wright notes, μελέτῃσιν connotes ‘constant practice and effort, as in athletic training, military duty, or rehearsing’.¹⁵ Despite the distinctive Empedoclean features of this passage, its basic message is not alien to quotidian, common-sense epistemology: repeated listening and sustained contemplation lead to better understanding.¹⁶

‘In ancient Greek ἀκούειν commonly refers to reading’;¹⁷ listening and reading are thus often hard to separate neatly. On our interpretation, Heraclitus implicitly enjoins his audience to pursue understanding by continued and repeated engagement with his book by explicitly saying that people fail to understand when they listen or read ‘for the first time’. In the early Greek world, such re-reading was not the exception but the norm. The literate minority did not have access to extensive libraries, public or private; like readers in many other pre-modern cultures known to modern scholarship, they practised ‘intensive’ rather than ‘extensive’ reading, paying careful and sustained attention to a relatively small group of texts.¹⁸ In the eponymous Platonic dialogue, Theaetetus, for example, has read Protagoras’ words often (ἀνέγνωκα καὶ πολλάκις, 152a5). Aristotle assumes that some people have engaged with their Empedocles intensively enough to have (some of) it memorised (Nicomachean Ethics, 1147^a20, 1147^b12). Ancient readers generally practised intensive reading, and the same will have held all the more true in Heraclitus’ time, when a culture of private literary reading was in its early days and the number of written texts in circulation was relatively low.

If Heraclitus implicitly recommends repeated and intensive engagement with his own logos, then this approach contrasts with the extensive and undiscriminating pursuit of wisdom that he elsewhere derides. B40 attacks four recognised panhellenic figures of wisdom: πολυμαθίη νόον ἔχειν οὐ διδάσκει· Ἡσίοδον γὰρ ἂν ἐδίδαξε καὶ Πυθαγόρην αὖτίς τε Ξενοφάνεά τε καὶ Ἑκαταῖον, ‘much learning does not teach intelligence; for otherwise it would have taught Hesiod and Pythagoras and again Xenophanes and Hecataeus’ (B40). Pythagoras in particular excelled all mankind in historia (ἱστορίην ἤσκησεν ἀνθρώπων μάλιστα πάντων) and drew on many books (ταύτας τὰς συγγραφάς), only to produce something pernicious: ἑαυτοῦ σοφίην, πολυμαθίην, κακοτεχνίην, ‘his own wisdom, much learning, evil artifice’ (B129).¹⁹

At least some people in antiquity did indeed take an intensive approach to Heraclitus’ book. The ancient reception provides abundant testimony to his famous, and even perhaps intentional, obscurity;²⁰ nonetheless, the extensive tradition of Heraclitean exegesis reflects the perception that careful engagement with his book has value.²¹ In the most optimistic expressions of this attitude, such engagement is the path to enlightenment. A case in point is Anthologia Palatina 9.540, which echoes the famous Hesiodic passage quoted above:

Don’t quickly unroll to its end the book of Heraclitus the Ephesian. The path is very difficult, I tell you; it is darkness and lightless gloom. But if an initiate inducts you, things are brighter than the radiant sun.

Here understanding Heraclitus demands both careful reading and expert help, but understanding is both achievable and worth the considerable effort.

II

Tarán provides a rare argument against the reading advocated here and in favour of the translation ‘once’: ‘[t]his latter meaning [sc. ‘for the first time’] would leave open the possibility of men’s understanding the Logos when they hear it for the second or third time, etc., and such a meaning is precluded by the rest of this fragment as well as by the rest of the related evidence.’²² Tarán does not spell out his reasoning. In a footnote, though, he expands: ‘[b]oth in this fragment and elsewhere … Heraclitus speaks of men’s failure to understand as characteristic of the human condition.’²³ It will be helpful to address Tarán’s objection to our reading of τὸ πρῶτον, not only in order to defuse his objection, but also because doing so will bring to light further, philosophical reasons to prefer this reading and help us to get clearer about the progression of thought within B1.

A key point to underline at the outset is that Tarán’s claim, in the footnote, that, according to Heraclitus, failure to understand is ‘characteristic’ of the human condition does not support his stronger, modal claim, in the main text, that there is no ‘possibility’ for humans to understand. It is only the modal claim that would pose a problem for our interpretation, but only the weaker one that B1 or other fragments support.

On our reading, when we reach the final two words of B1.(i), τὸ πρῶτον, we come upon a crucial qualification to Heraclitus’ initial assertion that people (ἄνθρωποι) ever (ἀεί) fail to understand. Once those final two words are read, Heraclitus’ claim becomes not the unqualified one, that people simply never succeed in acquiring understanding, but the importantly qualified one, that they never succeed before they hear or when they have heard for the first time (or: when they have first heard).²⁴ In (ii), Heraclitus then lends support to the claim he advanced in (i), or at least amplifies this claim (‘for’: γάρ), by observing how a person, despite having pertinent experiences of the account and realities in question (‘such words and deeds’: καὶ ἐπέων καὶ ἔργων τοιούτων), may still be equivalent for all intents and purposes to one who has not experienced them. In (ii), then, Heraclitus claims that, by failing to digest and reflect on their experiences properly, people are—as things currently stand—equivalent to agents who never had those experiences to begin with.²⁵ Importantly, however, Heraclitus does not thereby imply or even suggest the modal claim that people could not go on to cultivate a better and more illuminating mode of engagement with his text and (thereby) with their experiences. Instead, what emerges more and more clearly as we read through the passage is that people who acquiesce in that mode and level of engagement with their experiences are presented to us as a negative exemplar of the wrong ways to engage with this text (Heraclitus’ logos) as well as with the world to which it relates (the cosmic logos). By the end of (ii), we also first encounter a contrary, positive exemplar: the authorial ‘I’ of Heraclitus himself, who delineates and shows things in accordance with nature. In (iii), finally, Heraclitus enriches his picture of the negative exemplar, likening the lack of awareness that characterises an uncritical engagement with one’s experiences (and with his text) to dreaming and forgetting.

Our interpretation of this final component of B1 is complicated by the question of precisely how to understand the contrast between Heraclitus (‘I’: ἐγώ) and ‘the other people’ (τοὺς δὲ ἄλλους ἀνθρώπους). We will return to this question in a moment. For now, let us assume (with Tarán) the strongest possible interpretation of this contrast: Heraclitus pits himself on one side and the rest of humanity on the other; in other words, ‘the other people’ here means ‘all other people’.²⁶ Even on this maximal interpretation, however, Heraclitus is still diagnosing in (iii) our current sorry state, using indicative verbs (λανθάνει … ποιοῦσιν), and not our capacities and limitations. Put differently, Heraclitus is still discussing here only what we are and not what we could be. For us, as Heraclitus’ auditors and readers, the force of this dramatic division of humanity is thus not despair but provocation: the message is not that in this sad state we must sadly remain, but that here is our chance to leave behind the benighted mass of humanity and join an enlightened minority.

At this point, however, Tarán might wish to press further the worry about the contrast between Heraclitus and ‘the other people’. He might object that, if Heraclitus says in (i) that people always fail to understand when they have heard ‘for the first time’ (or: ‘at first’), then he implicitly allows that people do, as things are, succeed in understanding later on, on the second (or third, etc.) go, thus contradicting his remark in (iii) that, as things are, all other people (τοὺς δὲ ἄλλους ἀνθρώπους) lack awareness of what they do.

A first response to this worry is that, by saying that people always fail to understand both before they hear and when they hear for the first time, Heraclitus need not commit himself to any definite claims about the measure of understanding that, as things are, some people manage to acquire on further hearing. At least, Heraclitus need not be implying that, on further hearing, people do, as things currently are, manage to get as far as to acquire that kind and measure of understanding that ‘the other people’ of (iii) fail to exhibit.

A second response, however, is that Tarán’s initial assumption, that the expression ‘the other people’ in (iii) refers to all human beings bar Heraclitus, is far from obvious. It is indeed difficult to pin down the precise reference and scope of ‘the other people’, and the expectation that this expression has a precise remit may be misplaced. On reflection, however, it would be somewhat bizarre for Heraclitus to claim that all other human beings lack a type of awareness that he takes himself to exhibit, when he can only have subjected a small handful of other human beings to autopsy. It is no less intuitive, and it is on second thought more plausible, to approach the scope of the expression ‘the other people’ in the light of the content of the contrast that is being drawn (note the contrastive particle ‘but’ (δέ) with which Heraclitus introduces this expression and links it to his previous statement). On this approach, we will take ‘the other people’ to refer to all those people—whatever portion of humanity this group represents—that, by contrast with Heraclitus himself (ἐγώ, (ii)), cannot make proper critical use of their experiences (ii), and lack the requisite awareness of their actions (iii). Put differently, what defines this subset of humanity as ‘the other people’ is not simply that they are non-identical with Heraclitus, but more specifically that, unlike Heraclitus, they fail to exhibit the appropriate attitudes towards their experiences and their actions. In saying this we do not doubt that Heraclitus diagnoses in B1 a general and widespread epistemic and exegetical failure among humans, and that he gives us to understand that ‘the other people’ in (iii) represent the great, unenlightened majority of humanity. We only insist that the interpretation of ‘the other people’ as referring to, literally, all human beings but Heraclitus is neither obvious nor even, on reflection, very plausible. If our alternative, context-sensitive interpretation of ‘the other people’ is on the right lines, however, then Heraclitus need not be affirming in (iii) that all human beings bar him lack understanding. If so, then the prospect of a possible conflict with (i) disappears. In (i), Heraclitus will be identifying what he takes to be a universal but qualified weakness in human beings: humans always fail to understand the logos before hearing it and when they hear it for the first time. The qualified nature of this diagnosis (‘for the first time’) may indeed create the impression that at least some persistent and critical-minded people succeed in acquiring at least some measure of understanding upon further (second, third, etc.) hearing of the logos. This impression, however, would be quite consistent with the contrast in (ii)–(iii) between Heraclitus and ‘other’ people, where this indicates not all other human beings bar Heraclitus, but rather those other human beings who, unlike Heraclitus, lack the proper engagement with their experiences and their actions.²⁷

When we look beyond B1, we are left with the same impression: that the failure to understand is a widespread characteristic of humanity, but not that progress is hopeless or that knowledge and understanding are beyond our reach.²⁸ The most obvious instance of successful human understanding is none other than Heraclitus himself, who highlights his attempt to convey something worth listening to and so, by implication, capable of being understood. From B1 onwards, Heraclitus makes himself familiar to us as a positive exemplar. In B50, he envisions the possibility that after listening, not to him, but to the logos (οὐκ ἐμοῦ, ἀλλὰ τοῦ λόγου ἀκούσαντας) his auditors can come to a wise agreement (literally, to share in the same logos: ὁμολογεῖν) that all things are one. Again, by encouraging us to think about Delphic Apollo, who ‘neither says nor conceals but gives a sign’ (οὔτε λέγει οὔτε κρύπτει ἀλλὰ σημαίνει, B93), Heraclitus is, we take it, gesturing towards those positive exegetical attitudes that we should, and no doubt can, cultivate. And indeed, Heraclitus’ deliberate style poses distinctive interpretive challenges that precisely impel us to develop and hone the exegetical skills required for hearkening in the right ways to those ‘signs’ that Heraclitus’ text issues to us: with ears attuned towards nuances, unobvious meanings, and multiple, simultaneously available interpretations (we return to this point below). There is, then, much to be said for the familiar view in the scholarship that learning how to listen properly to Heraclitus’ logos is a preparatory exercise for learning how to listen properly to the cosmic logos. In other words, proper, critical engagement with the challenges of interpreting Heraclitus’ remarks is supposed to prepare and condition us for proper, critical engagement with our experiences and the insights that those experiences can afford us into the hidden nature (B123) and unseen attunement (B54) of things.²⁹ If, in B50, logos refers to Heraclitus’ book, then the fragment is telling us that a skilled auditor or reader of this book will ultimately agree with it, not because of Heraclitus’ personal authority (‘listening not to me’), but because of the ‘account’ or ‘argument’ (the logos) itself. If, alternatively, logos refers here to the cosmic principle, then B50 tells us that, in the end, our proper engagement with Heraclitus’ book (‘listening not to me’, i.e. not to my book) is itself only a preparatory exercise for something else: our proper engagement with the cosmic logos. Both kinds of logos should be kept in view.³⁰ It might even be the case that, for those of us fortunate enough to have it, Heraclitus’ book (his logos) represents our best chance of attaining understanding of the hidden nature and unseen attunement of things (of the cosmic logos). It would be going too far, however, to say that learning from Heraclitus’ book is a necessary condition for attaining such understanding: Heraclitus’ own example proves otherwise.

By the same token, we are not doomed to remain among those who ‘do not know how to listen’ (ἀκοῦσαι οὐκ ἐπιστάμενοι, B19; cf. B34), presumably both to the cosmic logos and to Heraclitus’ own words.³¹ In general, Heraclitus’ critical remarks about the failures of most humans to understand are not designed to depress us with the sad lot that we could never escape, but rather to provoke and orient us through negative exemplars. Thus, for example, the remark that ‘many’ people (πολλοί) fail to comprehend what they encounter and learn (B17) tellingly falls short of the claim that this is true of all people (compare B104: ‘the many are bad, but few the good’; cf. also B2).³² Again, the qualified remark ‘poor witnesses for people are the eyes and ears of those who have barbarian souls’ (B107) suggests, by contrast, that, for those with the proper sort of souls, eyes and ears can be valuable witnesses.³³ Indeed, his general pessimism about people notwithstanding, Heraclitus also insists that there is a share for ‘all people’ (ἀνθρώποισι πᾶσι) in knowing themselves and being of sound mind (γινώσκειν ἑωυτοὺς καὶ σωφρονεῖν, B116; cf. B113).³⁴

In sum, in B1 and elsewhere Heraclitus is telling us that acquiring at least some measure of understanding of his text and of the world would be an arduous and protracted affair for us, but not an impossible one. Once we appreciate this point, we see that the sequence of thought across the three components of B1, and the force of the fragment in the broader context of the book as a whole, in fact present us with a further reason to avoid the common rendering ‘once’. After all, it would be, at a minimum, misleading for Heraclitus to affirm, quite generally and without qualification, that people always fail to understand, both before and after hearing the logos, if, in different fragments, he gives us to understand that it is not beyond the reach of humans—as, if nothing else, the case of Heraclitus himself proves—to gain at least some measure of understanding of his words and of the realities underpinning our experiences. Instead, the claim Heraclitus makes in (i) is the more restricted one, that people always fail to grasp this account both before they hear it and ‘after they have heard it for the first time’. To be sure, this claim grabs our attention and pulls us up short. But it leaves it quite open that at least some people will gain at least some measure of understanding of Heraclitus’ words and of their own experiences, if they apply themselves to the former and then to the latter with the requisite persistence and critical mindedness.

III

On our interpretation of B1, when hearing or reading the opening of Heraclitus’ work, the audience at first learns that they will not at first understand; Heraclitus thus implicitly encourages his audience toward sustained, repeated engagement with his text. On the most fine-grained level, Heraclitus’ syntax, which has long given readers pause, is itself a mechanism to provoke such engagement. Already Aristotle faults Heraclitus’ writing in general, and in particular the programmatic first sentence of B1: it is neither easy to read nor easy to understand (εὐανάγνωστον … εὔφραστον, Rhetoric 1407^b11–12) because it is unclear how ἀεί is to be construed (see note 6 above and cf. A4 = Demetrius, De elocutione 191–2). Such ambiguity—better perhaps ‘linguistic density’ (Kahn 1979, 91)—is manifestly a conscious stylistic trait assiduously cultivated across Heraclitus’ book. One might interpret this deliberate strategy not as a barrier to understanding but rather as a spur to repeated engagement with the text in the light of its multiple possible interpretations.³⁵ Heraclitus’ verbal style, in other words, seems calibrated to prevent his audience from supposing that they have understood him at first sight or on just one single hearing.

On our interpretation of τὸ πρῶτον, Heraclitus himself articulates something like the exegetical principle that Charles Kahn has convincingly and influentially identified: the same remark, read again and in the light of all the other ones, tells you something new. Our understanding of some particular Heraclitean remark is thus continually changed and deepened by our progressive encounter with the other Heraclitean remarks that surround it and by our reflections on all those remarks and on the interrelations between them.³⁶ We thus understand a given remark later on in a way that we could not understand it at first. On the view defended here, Heraclitus himself concisely and subtly conveys as much in the beginning of his book.

Indeed, the qualification τὸ πρῶτον may itself be one whose significance we can only grasp fully upon further and repeated reading: it will only be in retrospect that we can appreciate for ourselves why we indeed could not have understood on first reading. In these respects, our interpretation fits well with a pattern observed by Jaap Mansfeld in an article informatively entitled ‘Insight by Hindsight: Intentional Unclarity in Presocratic Proems’ (1995). Mansfeld reads Heraclitus’ proem as ‘what we may call a declaration of obscurity’ and maintains that Heraclitus (like Parmenides and Empedocles) introduced his work ‘in such a way that one can only understand, or hope to understand, what is meant after having studied the whole’.³⁷

Heraclitus’ very term in B1 for the failure of people to understand— ἀξύνετοι—is one particular word that will command renewed and newly informed attention on a second hearing or reading. At first, we might hear simply that people turn out not to understand. Subsequently, however, we will know to connect this particular term for incomprehension with Heraclitus’ category of ‘common’ (inclusive, broad, synoptic) vs ‘private’ (selective, narrow, myopic) understanding (a very similar connection is drawn unmistakably in B114: ξὺν νόῳ … τῷ ξυνῷ). We will know, also, to read this particular term in the light of Heraclitus’ overarching preoccupation with the (as he sees it) literal and etymological meaning of words (e.g. B32, B48, B50, B114). Read in this light, ἀξύνετοι is, on our view of τὸ πρῶτον, a highly apposite term for Heraclitus to use at this programmatic juncture: what auditors and readers will indeed not be able to do on a first hearing or reading is to bring together (συνίημι) a term (such as, for example, ἀξύνετοι) or a statement (such as, for example, B1) with the terms and statements to which they will be exposed later on in the book. On subsequent hearing and reading, however, this possibility now opens up for the properly attuned and astute re-reader. In this respect, our newly informed and deeper understanding of Heraclitus’ very term for incomprehension (ἀξύνετοι) itself exemplifies his programmatic point about what is possible on a further—but not on a first—hearing.³⁸