As often as not a major textual problem needs, in addition to all the regular perspiration and frustration, a eureka moment, a sudden inspiration or intuition when the final piece of the puzzle falls into place; and that is a rather mysterious and unpredictable process that has its own provenance and takes its own time, usually weeks or months. In the case of Lucretius 5.1442, although I always felt that progress was possible, that moment refused to come. Until now: over forty years later, this paper offers my solution.Footnote ¹

The crux requires us to consider Lucretius’ whole paragraph (lines 1440–7) where, following his detailed exposition at the heart of Book 5, he is summarising the evolutionary process of human civilisation:

The above text follows the manuscript tradition, and I have restored tum, which Lachmann altered to iam and which has been followed by many editors since, but I see no need to alter the MS consensus and a perfectly appropriate text in favour of a synonym. Some scholars make the argument of a threefold anaphora of iam in these four lines, but it is just as likely that the tum prepares for the following cum, and there is also the fact that the third iam, buried in the middle of the line, is something of an anticlimax compared with the other two in pole position at the head of the line. It may also be added that the precursor ‘prehistoric’ lines corresponding to this passage (lines 999-1010), to which we shall return, talk of the ‘wicked art of navigation’ being ‘then blind’ or unknown (tum caeca).

As to propter odores Bailey called it ‘the most difficult textual crux’ in Lucretius, a description which Murgia (Reference Murgia2000)Footnote ² adopted as the title of his comprehensive and well-argued paper in July 2000. The MSS are in agreement but Servius, citing florebat in commenting on Aen. 7.804, referred to ʻLucretiusʼ (but without any specific textual reference) thus: florebat nauibus pontus. That ancient reference has led some moderns to conjecture accordingly, notably Housman's maris … nauibus pontus.Footnote ³ Such eminence cannot be easily dismissed, but there are clearly difficulties with this approach, whether the need to emend mare or, more significantly, the fact that, as is well known (and well demonstrated by Murgia in the article cited), Servius, with his tendency to misquote or paraphrase, cannot be trusted in such matters. But if Servius’ comment is not persuasive overall, it nevertheless has the merit of expecting a substantive noun to accompany ueliuolis. While the view of those who argue that ueliuolis is sufficiently clear to stand on its own cannot be conclusively rebutted, the lack of a word for ‘ships’ remains a major weakness in their conjectures.

A very different approach is to argue that the copying process has somehow incorporated propter odores from the end of line 2.417 (where propter is an adverb) into 5.1442 in substitution for the original, which is accordingly, at least as far as known manuscript evidence is concerned, irrecoverably lost. That is possible, of course, and if true, that is the end of the matter. But it is just as likely, if not more likely, that the end of 5.1442 became corrupt and could not be understood and corrected, but that the corruption looked sufficiently like propter odores for a scribe or scholar, especially one with knowledge of 2.417, to ‘see’ and copy those words.

That leaves essentially two more lines of enquiry. The first is that propter odores is correct. The words give a reason for man's adoption of seafaring, ‘spices’ being used poetically, it is argued, as a specific for the more appropriate wider term ‘trade’ or ‘commerce’. Again, there are difficulties, and for some of us very serious difficulties, for example about whether odores can reasonably, even in poetry, mean what it must mean for the line to make sense, and then whether Lucretius has given, or could or should give, this as the sole reason for the development of shipping; and indeed whether simple common sense rather than scholarly complication is all that is required for its removal. Many scholars dismiss the reading out of hand, and even its most fervent adherents seem to concede – surely rightly and inevitably – that it is an odd, awkward or jarring phrase to use in this context.

Of course, if nothing better or more certain is available, the MSS reading ought to stand if it can reasonably do so. But propter odores fails the test of reasonableness on a number of other grounds too. First, as referred to earlier, there is no noun for ueliuolis. Secondly, the phrase is difficult to construe, except as a last resort. But most important of all – and, oddly, not a factor in the consideration of this crux to date – the issue is not whether the reason given by Lucretius is correct or not, or the most important reason or not, but that it is here neither necessary nor appropriate to give any reason at all. In these lines in which Lucretius refers to several stages in man's social evolution, no reasons are given, and that is not his purpose here. He deals in detail earlier in the book with social evolution including causation, but the essence of this paragraph by way of summary and conclusion at the end of the book is not causation but time, the gradual and incremental process of civilisation.

We are left with the last category of approach, that the phrase propter odores is a corruption, at least to some extent, of what was originally there. Here the choice is rich, and there is little point in listing all the conjectures that have been made, which can be found in the works referred to in this paper, but a summary might be useful. In this category there are essentially three sub-categories. First is the ships-and-sea-type solution, which conjectures a ‘ships’ noun for ueliuolis and then completes the line with something to do with the sea. Examples include Büchner's nauibus altum (the modern Loeb reading of Smith, who maintains that it was first proposed by Merrill in 1902 in CR 16), followed by Richter (but adding another line) and Rouse's earlier Loeb reading nauibus omne; and, with the meaning of the line spilling over into the next, Lachmann's puppibus, et res and Munro's puppibus, urbes. We should also add here, I think, Krokiewicz's prorisque decoris.

The second approach is to do with the shore, producing a meaning that the sea ‘was flourishing with sail-flying [ships] near to the shore[s]’: hence Lord's propter ad oras, though he adds an extra line to allow for a nauibus and argues that people then did not venture out into the deep sea, but hugged the shore. That seems altogether to miss the point, since what Lucretius might have been expected to say (if he said anything like this at all here) would be that the sea flourished with ships everywhere, including ‘right up to the shores’, and the problem then is that propter ad oras does not quite say that. Ernout's conjecture litora propter is worth noticing in this respect.Footnote ⁴

The final line of attack is to seek to get still closer to the MS lettering propterodores, most conjectures of this kind retaining propter in its causative meaning and then seeking a better word (with a better point) than odores, others making causation even more significant and linking this line with the next. I have indicated above my firm view that causation is not in issue here, but there are nevertheless some fine conjectures, of which perhaps the best is Wikarjak's propter honores. Others include Colin's excellent propter adora and Deshayes’ propter olores. Other conjectures based on the letters in the MSS connect 1442 with the next line: MacKay suggested propterea res and Smith propterea quod. These are commendable inasmuch as they are reasonably close textually, but they exacerbate the inappropriate issue of causation in this passage.Footnote ⁵ Furthermore, Smith's causally ‘correct sequence’ is not correct since, for propterea quod to work, he is impelled to argue that allies and alliances not only preceded but were the reason for navigation, and he is surely wrong in claiming that Thucydides implies it.Footnote ⁶

A reasonable summary of all these efforts might be that those which have the merit of adhering most closely to the letters of propterodores are in general less convincing in meaning, while those which are more convincing in meaning have the weakness of little or no match with the manuscript evidence. The conjecture offered in this paper seeks to incorporate the strengths of both and to avoid their weaknesses, and also to demonstrate how the corruption may have arisen. It is based essentially on two factors: that ueliuolis ought to have an accompanying noun; and that causation, and therefore propter in this sense, are inappropriate in this context.

Let us return to Lucretius’ paragraph containing the crux. It is clear that the first four lines illustrate stages of human development prior to the beginnings of poetry and the first written historical records. It is clear, too, that these four lines are intended to present stages of man's development which are not causally linked, as noted above, and also not strictly chronological. They simply and briefly describe a selection of four discrete stages of the evolutionary process which Lucretius had already described in detail earlier in the book: men living within fortified walls (turribus), land cultivated and divided, seafaring, and men making allies and treaties. It can be clearly demonstrated that strict chronology is not intended here by simply looking five lines further on, where Lucretius actually and precisely inverts this order (1448–9):

He lists navigation, agriculture and walls, but does not mention allies and treaties at all. Furthermore, as we saw earlier, his detailed account (lines 1019–27) makes it clear that alliances (tunc et amicitiem coeperunt iungere, 1019) and treaties (sed bona magnaque pars seruabat foedera caste, 1025) far predated all of the other things listed in 1440–3. He does not mention the building of cities and citadels (condere coeperunt urbis arcemque locare | praesidium) and the division of land (et pecus atque agros diuisere) until lines 1108–10, and then he follows this with the discovery of gold (1113) and without any mention of navigation.

Lines 1440–3 are, then, neither causative nor strictly chronological, but there is nevertheless a distinction that can be made. The first two lines (protective walls and division of land) are clearly much earlier stages of development than the navigation and treaties in the next two lines, and the break comes at the beginning of line 1442, where the MSS read tum in contrast with iam in lines 1440 and 1443. It is another reason, in my submission, for regarding the MS reading as correct, because there is a clear chronological shift between the two pairs of lines.

To return now to ueliuolis and the lack of a noun, Ennius uses ueliuolis twice in the fragmentary text that we possess, both times with nauibus. Some scholars argue, as we have seen, that ueliuolis is so obvious in meaning that the noun can be taken for granted. That does not seem to me to be right, and clearly Ennius felt the need for a noun on both occasions. A noun is necessary, in my view, but it does not have to be nauibus: any noun connoting ‘ships’ would serve the purpose. My proposed reading of 5.1442 is therefore this:

It is instructive now to turn to the earlier, more detailed section of the fifth book dealing with mankind's social evolution (from line 925 onwards) and consider in particular what Lucretius has to say about navigation (lines 999–1008):

Lucretius emphasises right at the start the incremental evolution of mankind: at non … una dies dabat exitio. War and seafaring did not come about in a day.Footnote ⁷ In those days the wicked art of navigation was hidden, and the dangers of the sea were unknown to man. In those days, too, lack of food was the problem, whereas nowadays, says Lucretius, the problem is abundance. We must once again return to tum at the beginning of line 1442, and we may now compare the conjecture tum … postmodo with the parallel passage above, in which we have tum in line 1006 in specific connection with the ‘wicked art of navigation’ and then tum … deinde in line 1007.

The conjecture in this paper is merely one more conjecture among so many. But I believe that it is significant both for its match with the letters of the manuscripts and because it is relatively easy to see how the corruption occurred. If we take postmodoproris as the original reading of the MSS, it is entirely possible that by a kind of haplography (technically, the so-called saut du même au même) a scribe's eye jumped from the p of postmodo to the p of proris. If he wrote pro before realising his mistake as the end of the line beckoned, he would then (as scribes often did) proceed by inserting the letters he had missed out and then return to the correct text, ending the line with the final three letters -ris. What he had written would be propostmodoris. Our careless scribe's erroneous text now produces a perfect match with odoris, especially when we consider the interchangeability in Lucretius of the accusative plural endings -es and īs.Footnote ⁸

A scribe or scholar confronted with propostmodoris would surely see odorīs, and would expect to see a preposition for the accusative, and propter jumps out. It is well known, and routinely exemplified by editors of Lucretius (see, for example, the Prolegomena of Bailey (Reference Bailey1947) 37–8), that in manuscripts written in minuscule o was liable to be misread as e and s misread as r, but there can be little doubt that a scribe looking at propostmodoris would be very likely to construe propterodoris in any event.

What, then, of postmodo? It is poetic (though used in prose, including by Livy) but, it might be objected, not used very often. Lucretius does not use it elsewhere, but he might have used it once – here.Footnote ⁹ (He only uses posthac once.) Terence uses it once (Hec. 208). Catullus uses it once (30.12). Horace uses it twice, but arguably only once in poetry (Od. 1.28.31), since the other use is in his Satires (2.6.27), which he expressly regarded as more like ordinary speech (sermo) than poetry proper. Ovid uses it more than twenty times, including (intriguingly) in a context specifically relating to seafaring and sails: utilior uelis postmodo uentus erit (Rem. 280). The word is commonly used in association with almost every temporal adverb, including tum and tunc.Footnote ¹⁰

But significantly for our purpose here, a key characteristic of postmodo is its use in more extended time-sequences, for example in pursuit of greater accuracy or elaboration (ordo temporis accuratius illustratur and in struct. trium uel plurium membrorum, as TLL s.v. puts it). A good example, and a striking parallel to the paragraph of Lucretius with which we are dealing (1440–7), is to be found in Ovid's celebration of the games of Ceres, where he describes the development of mankind's food supply (Fasti 4.395–404):

Whether or not Ovid is directly imitating Lucretius here – and it certainly looks like it – what is significant is that, though his focus is narrower (the development of man's food rather than mankind itself), his context of the earth and man's discoveries (tellus, reperta) and his mood and language all have Lucretian echoes. And it is in the sequence and language of time above all that the most striking echoes are to be found: Lucretius in his lines has (if postmodo is right) iam, tum, postmodo, iam, prius, priu[s], and Ovid primis, modo, nunc, postmodo, iam, prima, tum primum. In each case, furthermore, postmodo marks a clear stage in the overall sequence.

The line as conjectured looks Lucretian enough. Alliteration of the last two words (or more) of the line is remarkably common in Lucretius, and p-alliteration is one of his commonest kinds, and perhaps the most effective, as in the famous phrase, already quoted above, in 5.1004: placidi pellacia ponti. That phrase had already appeared at 2.559, a passage (lines 552–9) which clearly (coortis, mare/maris, dolumque, ridet placidi pellacia ponti) prefigures lines 999–1008 in Book 5 (mare, coortum, placidi pellacia ponti, subdola, ridentibus undis), which in turn prefigures the passage later in the book (1440–7) containing propter odores. Interestingly, too, that passage in Book 2 contains the only other use of prora in Lucretius: antemnas proram malos tonsasque natantis, 554.