A Tove on the Table

Ludwig Wittgenstein’​s Tractatus Logico-Philosophicus, the only book he published during his lifetime, is one of the greatest philosophical works of the 20th century. It might have been expected, when it first appeared in 1921, to have limited appeal. It is very much the work of a philosophers’ philosopher, forbiddingly technical in places and esoteric throughout. Yet it has gone on to capture the public imagination as few other philosophical classics have.

It consists of 525 sections, or ‘propositions’, ranging in length from four words to about a page and a half of text and diagrams. Each is given a decimal number, with the numbers indicating subordination and interconnection. Thus propositions 2.21 and 2.22 are comments on proposition 2.2, which is itself a comment on proposition 2, which is one of the seven top-level propositions. The propositions have an aphoristic quality. They are written with great compression, hardly any examples, and little explicit argument. They are for the most part general and abstract. Wittgenstein makes few concessions to his reader. But there is something undeniably awe-inspiring about their cumulative effect and about the concision with which they encapsulate his elaborate system of thought.

Part of the aim of the book is to indicate what it is about the world that makes it possible for us to represent it, in thought or in language. Wittgenstein is led to a vision of crystalline purity. The world is the totality of facts. Facts are determined by states of affairs. States of affairs, each of which is independent of every other, are configurations of objects. These objects would have existed however the facts had been. If the facts had been different, it would have been because the objects had been configured differently, not because there had been different objects. Representation itself consists of facts. Thus a thought or a statement is a fact, determined by a configuration of ‘signs’. In the most elementary case the signs stand for objects, and the fact that they are configured in the way they are represents that the corresponding objects are configured in the same way. The thought or statement in question thereby serves as a ‘picture’ of the corresponding fact. It is true if the objects are configured in that way, and it is false if they are not. In a less elementary case, for example in the case of a conjunction of two statements, truth or falsity is determined by the truth or falsity of its constituents: a conjunction of two statements is true if both its constituents are true, false otherwise.

The opening propositions present the first part of this vision, concerning the division of the world into facts and the constitution of these facts. As the book progresses, the second part of the vision emerges, concerning representation, along with reflections on the nature of logic, philosophy, mathematics and natural science. The book then approaches its climactic conclusion with a sequence of laconic remarks about value, death, God, the meaning of life, and the inexpressible. The whole thing has the air of a metaphysical disquisition on the fundamental character of reality and our engagement with it. The culmination, however, consists of two remarkable propositions that cast doubt on this impression: the penultimate proposition, numbered 6.54, in which Wittgenstein says that anyone who understands him will eventually recognise what he has been saying as nonsensical (by ‘nonsensical’ he does not mean absurd or foolish, but quite literally lacking in meaning); and the final proposition, numbered 7, in which, as if in explanation of the propensity to produce such nonsense, he says that we must keep silent about what we cannot speak about.

The material towards the end of the book, including that final proposition, bears on what Wittgenstein wrote to Ludwig von Ficker, editor of an Austrian literary magazine, while he was trying to get the work published:

The book’s point is an ethical one … [It] consists of two parts: the one presented here plus all that I have not written. And it is precisely this second part that is the important one. My book draws limits to the sphere of the ethical from the inside as it were … I have managed in my book to put everything firmly into place by being silent about it.

The suggestion seems to be that, by drawing the limits of what can be represented in thought or language, Wittgenstein has also indicated what cannot be represented in thought or language; what lies outside the world of facts; what is of value.

But to what extent has he succeeded in drawing the limits of what can be represented in thought or language? Strikingly, some of the most serious reservations about what he achieved can be found in his own later work, and especially in his second great masterpiece, Philosophical Investigations, published posthumously in 1953. Even more strikingly, and notoriously, some of the most serious reservations about what he has achieved are to be found in the Tractatus itself. Or so it seems. I have in mind that penultimate proposition, in which he renounces what has gone before as nonsensical. We shall need to come back to this.

Wittgenstein​ wrote much of the Tractatus during military service at the front in the First World War. He continued to work on it for a while in the aftermath of the war, the first nine months of which he spent in prisoner-of-war camps after being captured by the Italians. It was a struggle for him to get the text published. But it finally appeared in the journal Annalen der Naturphilosophie, with an introduction by Bertrand Russell. Russell was a major public intellectual at the time, and it was his introduction that effectively secured publication (though Wittgenstein abhorred it).

The first English translation appeared in 1922, alongside the original German and again with Russell’s introduction, slightly revised. This was for a series edited by C.K. Ogden, the International Library of Psychology, Philosophy and Scientific Method. The translation appeared under Ogden’s name, though it was mainly undertaken by Frank Ramsey, then still just a precocious mathematics undergraduate. It also included some modifications by Wittgenstein himself. A revised edition, with further modifications by him, appeared in 1933.

Wittgenstein’s changes were prompted by what struck him as excessive faithfulness to the original German. They were designed to preserve, as he put it in a letter to Ogden, ‘the sense (not the words)’. We do well to remind ourselves, however, that Wittgenstein was not a native English speaker. Even with his modifications, the translation is often clunky. Its chief drawback, as Wittgenstein’s remark to Ogden intimates, and as Michael Morris has marvellously put it, is that it is ‘dog-literal’. Moreover, it is insensitive to some philosophically critical features of the German. A well-known example is its failure to heed the distinction that Wittgenstein draws between what is unsinnig (‘nonsensical’) and what is sinnlos (‘senseless’) – where an empty tautology such as ‘What will be will be’ counts as the latter but not as the former. Brian McGuinness, in his 1988 biography of Wittgenstein, wrote that a ‘whole generation of English-speaking philosophers came to know the [Tractatus] through a translation which seems to have been … shackled by the presence of the German on the opposite page. It reads as if made from a dead language.’

It was McGuinness who, in collaboration with David Pears, produced the second English translation, in 1961. But the correspondence with Ogden, in which Wittgenstein commented on the first draft of the first translation, had not yet come to light (it was published only in 1973), so there was little appreciation at the time of Wittgenstein’s own input, however questionable, into Ogden and Ramsey’s text. After the correspondence emerged, Pears and McGuinness published a revised version of their own translation. It is much more fluent than Ogden/Ramsey, though occasionally the fluency is achieved by adding ideas – not just words – to which nothing corresponds in the original German. And there are places where the original German has a sonorous, almost biblical quality which the fluency fails to capture.

We now have three new English translations: by Michael Beaney for Oxford, Alexander Booth for Penguin and Damion Searls for Norton. (A fourth, by David Stern, Katia Saporiti and Joachim Schulte for Cambridge, is forthcoming.) The book came out of copyright in 2021 (seventy years after Wittgenstein’s death), which is the reason new translation is possible. But it’s another matter whether such a thing is desirable. Do the infelicities of the two older translations perhaps show that there is no way of achieving an appropriate level of fluency that does not involve taking at least as much liberty with the content as Pears and McGuinness do, or of maintaining an appropriate level of respect for the content that does not involve sacrificing at least as much fluency as Ogden and Ramsey do?

The three translators must have thought there was something new and worthwhile that they were capable of providing. One thing each was capable of providing, of course, was one more version of the text. Beaney, in his introductory material, observes that new versions can shake up the complacency that sets in when an existing translation is taken as standard, and that multiple translations can help us triangulate on the original. Booth, in his preface, similarly says that his translation ‘is meant to complement existing versions, not replace them’. Searls is more bullish. He laments the flaws in the earlier translations and spends much of his introduction explaining the ways he has tried to improve on them. But whether or not any of these new translations is an improvement on Ogden/Ramsey or McGuinness/Pears, their value is obviously to some extent dependent on their own intrinsic merits.

Unsurprisingly, in view of its concern with language, the Tractatus itself includes some remarks on translation. Might these be relevant here? Not if what we are looking for is a guide to good translation. For one thing, Wittgenstein’s remarks relate solely to preservation of meaning. A good translation typically involves more than that: a rhyming couplet, for instance, may need to be rendered by a rhyming couplet. And in any case, Wittgenstein’s concerns are purely theoretical. If one took them as guidelines, one might get the impression that the art of translation involves nothing more than devising a way of mapping the words of one language onto the words of another, whereafter the analysis of any given text in the source language, followed by application of the mapping, followed by construction of a corresponding text in the target language, do their algorithmic work. This is not a total travesty of his view. But whatever accuracy it harbours concerns operations that apply so far below the surface features of language as to be of no relevance whatsoever to anyone engaged in actual translation. Here we should note the fundamental distinction on which Wittgenstein insists between the outward form of clothing and the form of the clothed body.

Translators of the Tractatus must therefore rely on whatever general principles of good translation they already have at their disposal. They must pursue all the familiar desiderata: preservation of meaning, consistency, sensitivity to the various non-semantic associations of words and so on. Often they will be forced to choose between two or more imperfect renderings of a given piece of text – though, as Beaney’s edition illustrates, the addition of supplementary notes explaining the translator’s choices can go a long way towards mitigating the imperfections.

Where a philosophical work such as the Tractatus is concerned, there is a further, crucial desideratum: fidelity to the author’s philosophical intentions. This has a critical bearing on my comparative assessment of these new translations, but before I turn to it, I want to mention three distinctive challenges that any translator of the Tractatus faces.

First,​ this book is not just a work of philosophy, it is a work of art. (In a notebook from the early 1930s Wittgenstein wrote: ‘Philosophy ought really to be written only as poetry.’) All three translators are sensitive to this. Booth writes in his preface: ‘We are in the realm of craft, shape and, now and then, possibly even song.’ And Jan Zwicky, in her introduction to Booth’s translation, details the way the stress and play of vowel sounds in the final proposition of the book give it a certain musicality (‘Wovon man nicht sprechen kann, darüber muss man schweigen’). But, by the same token, the Tractatus is not just a work of art, it is a work of philosophy. There is a risk that translators will try so hard to preserve its aesthetic merits that they fail to remain faithful to Wittgenstein’s philosophical intentions. I am not so confident that all three translators are sensitive to this. Certainly Zwicky, in her introduction, and Marjorie Perloff, in her foreword to Searls’s translation, seem to betray their own insensitivity to it when they write, respectively, ‘compositions that we intuitively recognise as lyric … are enactive: how they communicate is what they mean,’ and ‘the ideal translator for the Tractatus is perhaps not a professional philosopher at all, but what we call a creative writer. From a literary perspective, what is said is never as important as how it is said in the translating language.’ Whatever truth there may be in these claims, it is of considerably less significance, as far as translating the Tractatus is concerned, than the error in them.

Second, thanks to the concern that the Tractatus has with language, it is a contribution to its own subject matter. Its translators must therefore ensure that their English version does not stand in overt tension with whatever messages about language Wittgenstein is attempting to convey. One could be forgiven for thinking that this challenge, though important, is not especially difficult to meet – were it not for the fact that it involves facing the third and most blatant of the challenges, which threatens to wreck the entire enterprise.

Here we return to the notorious fact that one of the messages about language Wittgenstein is attempting to convey is that material such as we find in the Tractatus itself is nonsensical. The third challenge is how to translate a work which, by its own lights and in a quite literal sense, leaves the translator with nothing to translate. (Booth, in his acknowledgments, says that his editor, Donald Futers, has saved him from various errors, among which he lists ‘just plain nonsense’. He then adds the standard caveat that any remaining errors ‘sadly, are entirely my own’. Not if he has done his job properly! Such errors are then largely Wittgenstein’s.) Any translator trying to meet the second challenge must render the original German material with what can ultimately be recognised as English nonsense – if indeed ‘German’ and ‘English’ are appropriate epithets here.

This casting of Wittgenstein’s own work as nonsensical on his part presents all manner of exegetical puzzles, quite apart from its challenge to the translator. What are we to make of a work whose author does that? Not that a text should be dismissed merely on the ground that it is nonsensical. Texts are produced for all sorts of purposes, some of which might well be served by nonsense – entertainment, for example, or parody. Even so, how can it not be an indictment of this text to say that it is nonsensical? Has Wittgenstein not written what he has written in an attempt to communicate something? And how can the text be nonsensical by – as I put it earlier – ‘its own lights’? If it is nonsensical, then surely it has no lights by which anything can be anything.

There is a yet more fundamental issue: what Wittgenstein actually characterises as nonsensical are what he calls ‘my propositions’, but what exactly is the scope of ‘my propositions’? All of them? All but that one? All but that one and a few other similar meta-propositions about what is at stake in the rest of the book? Just whichever are of a metaphysical cast (and not, for example, the references to what other philosophers have said, or the description at 5.5423 of the two ways of seeing the Necker cube)? Does Wittgenstein perhaps mean that there is no way of making sense of the book as a whole, though there may be of any sufficiently small part of it, rather like an Escher drawing? In any case, do we have to accept what Wittgenstein says in that penultimate proposition? Maybe that is the problem. Here we confront one of the great ironies of the Tractatus. Many books include passages which, despite their authors’ best efforts, simply do not make sense. Wittgenstein may be involved in a mirror image of this: that is, the Tractatus may include many passages which, despite its author’s best efforts, do make sense! The effect that the passages in question have on the reader, and that Wittgenstein seems to intend them to have on the reader, may make them, whether he likes it or not, successful acts of communication. Or are we to draw yet another distinction here, between that which succeeds as an act of communication and that which makes sense?

One thing seems clear. To whatever extent it is correct to say that what we are dealing with in the Tractatus is nonsense, it is very carefully crafted nonsense that does whatever philosophical work it does by appearing, initially at least, to make sense. And it is that appearance of sense that is critical for the translator. True, it is entirely possible to translate even blatant nonsense. There are ‘French’ and ‘German’ versions of Lewis Carroll’s ‘Jabberwocky’, for example (beginning, respectively, ‘Il brilgue: les tôves lubricilleux …’ and ‘Es brillig war. Die schlichte Toven …’). But a translation of the Tractatus is not like that. However illusory any appearance of sense in the book may be, it is that which the translator must try, first and foremost, to capture.

This is not, incidentally, a rejection of the ‘new’ reading of the Tractatus which has recently come to prominence, according to which whatever counts as nonsensical in the book counts as nonsensical in precisely the same way as ‘’Twas brillig, and the slithy toves …’, that is to say by virtue of the sheer lack of meaning of some of the words in it. Advocates of the new reading are dissenting from a more traditional reading whereby some of the nonsense in the book, if not most of it, arises from violations of grammar in which meaningful words are put together in non-meaningful ways, as in ‘I saw a big on the table.’ Here, it seems, there is an adjective where a noun should be. But advocates of the new reading deny that there can be an adjective where a noun should be. Being where a noun should be is already enough to prevent a word from counting as an adjective. What there is here, where a noun should be, is a homonym of an adjective, purporting to be a noun – but only purporting to be one, since it has no meaning in that role. The sentence as a whole is thus of a piece with ‘I saw a tove on the table.’ I am sympathetic to this view. But I am not now defending it, merely noting that I am not rejecting it, the point being that even words that are straightforwardly lacking in meaning can (because of their meaningful homonyms) contribute to the illusion of sense.

There is an associated issue concerning what Wittgenstein takes himself to be doing with his nonsense. On the more traditional reading, he believes that some things are inexpressible and he takes himself to be conveying some of these things, by putting meaningful words together in suitably evocative non-meaningful ways. On the new reading, Wittgenstein does not believe that anything is inexpressible; what he believes is that there are temptations to see sense where it is lacking and he is using nonsense in a therapeutic role, to expose these temptations, so that the reader will eventually be able to overcome them and recognise the nonsense for what it is, namely gibberish that conveys nothing at all.

There is also an attractive hybrid reading. On the hybrid reading, Wittgenstein does not believe that any truths are inexpressible – the only truths there are being truths about how objects are configured, which are ipso facto susceptible of expression. Nor therefore does Wittgenstein believe that he is conveying any inexpressible truths. But he does take himself to be conveying inexpressible practical insights. These include insights into how to recognise the nonsense that he is using to convey these very insights for the sheer nonsense that it is. But they include more besides. (They had better. Otherwise the nonsense would be like the plinth whose sole purpose is to support a sign reading ‘Mind the plinth.’) Notably, they include insights into how to face the world ethically.

But these issues, critical though they are to an understanding of the Tractatus, are really orthogonal to any questions about how it should be translated. Not so its appearance of sense; nor, therefore, the philosophical issues that inform that appearance of sense. I have laboured this point because a translator should ideally have some basic grasp of these issues, and it is this, above all, that accounts for the comparative merits of these three translations. Only Beaney is a professional philosopher – and it shows. Booth, to his great credit, manages an extremely elegant and successful translation despite his lack of expertise. Nevertheless, he is occasionally led astray in ways Beaney never would be. Searls’s lack of expertise often proves disastrous.

Before I amplify, I will mention some advantages that each of the three editions has over the other two. The Norton edition helpfully includes the original German text. The Oxford edition nicely accentuates the book’s tree structure by printing the seven main propositions and those at the next level down in bold, while giving subsidiary propositions different degrees of indentation corresponding to their respective levels in the structure. All three editions include fascinating supplementary material about the history, content and aesthetics of the book.

Each of the actual translations also has advantages over the other two. Beaney’s may be the best, but Booth’s is the most stylish, and Searls’s has a fluency which sometimes brings the ideas to life in a way that neither of the other two, nor either of the two older translations, does. Of the five different renderings of the book’s opening proposition, only Booth’s – ‘The world is all that happens to be the case’ – uses the construction ‘happens to be’, which captures the connotations of contingency in ‘der Fall’. (Ironically, it is Searls, in his introduction, who does most to highlight these connotations, before nevertheless rejecting any translation of the kind that Booth provides in favour of his own terrible alternative: ‘The world is everything there is.’) Of the five different ways of dealing with the last word of the book, ‘schweigen’, it is Searls’s ‘keep silent’ – in contrast, for example, to Beaney’s ‘be silent’ – that does greatest justice to the degree of effort that the word suggests.

Nearly always, however, Beaney seems to me to make the best decisions, and even if I am wrong about that, his accompanying explanatory notes give an extraordinarily detailed and helpful gloss on what is at stake in those decisions. But what of his superior grasp of what is at stake philosophically, and the corresponding insufficiencies of the other two translations?

Iwill begin​ with Booth. There is material in 4.113 and 4.114 about how philosophy sets limits to what can be represented in thought or in language. Wittgenstein uses two verbs in connection with this: ‘abgrenzen’ and ‘begrenzen’. As Beaney explains in a note, ‘it is important to translate [these] differently but connectedly … “abgrenzen” has more the meaning of “demarcate” – of one thing from something different – whereas “begrenzen” just means drawing a … limit … without specifying the “other side”.’ He further explains that this difference is important to Wittgenstein because it connects with the issues of what, if anything, cannot be represented in thought or language – of what is unsayable – and of how to indicate it, which is what Wittgenstein is alluding to when he uses ‘abgrenzen’. Beaney translates the two verbs as ‘delimit’ and ‘limit’ respectively. Booth translates them both as ‘delimit’.

This issue of how to indicate the unsayable is also pertinent to the following proposition, 4.115, in which Wittgenstein says that philosophy indicates the unsayable by clearly representing the sayable. The verb that Wittgenstein uses for ‘indicate’, and that Beaney translates as such, is ‘bedeuten’. Elsewhere in the Tractatus, ‘bedeuten’ has to be translated as ‘mean’, which Beaney does. He is normally very careful about maintaining a strict correspondence between key German expressions and their English translations. But he is also sensitive to cases that have to be treated as exceptions. This is one such. To say that philosophy ‘means’ the unsayable, which is what Booth’s translation gives us, and which in any case sounds odd, is too much of an affront to Wittgenstein’s insistence in 4.112 that philosophy is not, as Beaney puts it, ‘a set of teachings’ but an activity. (Searls, incidentally, has philosophy ‘referring to’ the unsayable. Here as elsewhere he is trying to respect Gottlob Frege’s technical use of ‘bedeuten’, to which Wittgenstein himself alludes at 6.232. But a little learning has proved a dangerous thing. Whatever else Wittgenstein is doing at 4.115, he is not suggesting that philosophy ‘refers to’ anything in Frege’s sense.)

In 5.62 Wittgenstein emphasises the word ‘der’ in the expression ‘die Grenzen d e r Sprache’. This is an attempt to accentuate the uniqueness of the language whose limits are at issue. Booth, bizarrely, emphasises the preposition in his translation: ‘the limits of language’.

One final point in connection with Booth’s translation. The reader needs to be wary of Zwicky’s introduction. She makes some excellent points, but she also includes some howlers, as when she writes: ‘I believe that when Wittgenstein speaks of thought in the Tractatus, he means whatever goes on when we grasp that there are an infinite number of primes.’ Emphatically not. Mathematical statements, for Wittgenstein, are ‘pseudo-propositions’. He is quite explicit at 6.21 that they do not express thoughts. Thoughts can only be about what ‘happens to be’ the case.

As for Searls, while the fluency of his translation sometimes pays off, often he achieves it at too high a price. Reconsider the opening proposition of the book. Searls avoids the awkwardness of talk in English about what is ‘the case’, or what happens to be ‘the case’, by rendering the proposition simply as ‘The world is everything there is.’ Perloff enthuses about this in her foreword. But something has already gone badly wrong. Wittgenstein is making a point about what kind of entity, or what logical category of entity, constitutes the world as he conceives it. In the very next proposition, 1.1, he tells us, in Searls’s own translation, that ‘the world is the sum total of all facts, not all things.’ The fact that blood circulates, for example, is part of the world, but neither blood nor circulation is. Searls’s rendering of the opening proposition completely obscures this, if indeed it does not flatly contradict it.

In his introduction, Searls raises the related question of whether the world, as Wittgenstein conceives it, is made up of nouns or of verbs. This is a bad way to put what is, in its own way, a good question. I shall not pause to consider what a good way to put the question would be. (Even a good way to put it would no doubt have to be strictly nonsensical.) The point is this. If we prescind from the badness of this way of putting the question, then the answer is: neither nouns nor verbs, but whole sentences.

Elsewhere the price that Searls pays for his fluency is excessive interpretation, sometimes downright misinterpretation. The most egregious examples concern Wittgenstein’s doctrine that thoughts and statements are pictures, and that pictures in turn are facts. There are two problems. First, and less seriously, Searls varies his translation of ‘Bild’. ‘Picture’ is the standard English translation; it is also the word used in almost all anglophone discussions of Wittgenstein’s doctrine. Searls sometimes uses it. But he also sometimes uses ‘image’, in order, as he puts it in his introduction, ‘to pull the translation a bit closer to where the German lands’. In justification he insists that a single German word does not always have to be translated by the same English word. Indeed it does not (reconsider ‘bedeuten’). But it does sometimes, otherwise needless confusion accrues. I do not see how anything but needless confusion could accrue in this case – especially when the switch from ‘picture’ to ‘image’ occurs in consecutive propositions and the second is a comment on the first, as happens at 2.17 and 2.171.

Later in Searls’s introduction – this is the second problem – he shows that he simply does not understand Wittgenstein’s doctrine. He criticises the book for being ‘confusing’ about facts. What Searls finds confusing is that Wittgenstein says both that the world consists of facts and that pictures in general, linguistic pictures in particular, are themselves facts. Whether or not this is confusing, it is certainly pivotal. That a (linguistic) picture is itself a fact is precisely what, according to Wittgenstein, enables it to represent the world. A statement comprising the elementary signs a and b, for example, is not a complex sign having a and b as two of its constituents: it is a fact about a and b, say the fact that a is put before b, and it thereby says that two corresponding objects stand in some corresponding relation to each other.

Searls not only finds the idea that facts both constitute the world and are sometimes pictures confusing; he struggles with the idea that they are sometimes pictures. ‘“How things are is a fact” makes sense in English,’ he writes in his introduction, ‘in a way that “a picture of how things are is a fact” does not.’ But Wittgenstein is unequivocal about this. ‘Das Bild ist eine Tatsache,’ he says at 2.141: ‘A picture is a fact.’

So Searls has a problem. ‘I solve this problem,’ he tells us, ‘by sometimes translating “ist eine Tatsache” (“is a fact”) as “states a fact”.’ In particular, he does this at 2.141. He justifies this as follows: ‘Varying the verb stays truer to what Wittgenstein means by the relationship between world and picture than it would be to say in English that a picture is a fact.’ This is calamitous. Some pictures, and some linguistic pictures in particular, do indeed state facts. (Only the true ones, mind.) But that is not what is at issue here. What is at issue is something that is supposed to explain how any picture states a fact. At a stroke Searls has succeeded in preventing the reader from understanding a cardinal doctrine, some might say the cardinal doctrine, of the entire book.

The offence is compounded elsewhere. At 3.143, this time sowing confusion of his own, Searls has Wittgenstein telling us that standard modes of expression in handwriting or print obscure that a sentence states a fact – why on earth would anybody think they obscure that? – not, what they certainly do obscure, that it is a fact.

Will any of these new translations become the standard? I suspect not, not least because, as two of the translators acknowledge, there are benefits in possessing more than one translation. What is more, the Pears/McGuinness translation has one compelling claim to retain its status as the standard, namely (even in these days of online search facilities) its wonderful index. That said, I strongly recommend that anglophone students of this work get hold of Beaney’s and Booth’s translations too – and maybe Searls’s, but they will need to treat the last with a great deal of caution.