In the latest issue:

Consider the Hermit Crab

Katherine Rundell

Emigrés on the Make

Sheila Fitzpatrick

Autopsy of an Election

James Butler

Short Cuts: Harry Goes Rogue

Jonathan Parry

‘Cosmo’ for Capitalists

Stefan Collini

Kara Walker’s ‘Fons Americanus’

Cora Gilroy-Ware

So many ships and fleets and armies

N.A.M. Rodger

British Sea Power

Paul Rogers

Richard Holbrooke

Samuel Moyn

Four poems after Callimachus

Stephanie Burt

‘Your Duck Is My Duck’

Christian Lorentzen

On Paul Muldoon

Clair Wills

Leanne Shapton

Namara Smith

Antigone on Your Knee

Terry Eagleton


Michael Wood

Walter Pater

Elizabeth Prettejohn

Two Poems

Rae Armantrout

Diary: In Monrovia

Adewale Maja-Pearce

Horsey, HorseyJohn Sturrock

Terms and Conditions

These terms and conditions of use refer to the London Review of Books and the London Review Bookshop website ( — hereafter ‘LRB Website’). These terms and conditions apply to all users of the LRB Website ("you"), including individual subscribers to the print edition of the LRB who wish to take advantage of our free 'subscriber only' access to archived material ("individual users") and users who are authorised to access the LRB Website by subscribing institutions ("institutional users").

Each time you use the LRB Website you signify your acceptance of these terms and conditions. If you do not agree, or are not comfortable with any part of this document, your only remedy is not to use the LRB Website.

  1. By registering for access to the LRB Website and/or entering the LRB Website by whatever route of access, you agree to be bound by the terms and conditions currently prevailing.
  2. The London Review of Books ("LRB") reserves the right to change these terms and conditions at any time and you should check for any alterations regularly. Continued usage of the LRB Website subsequent to a change in the terms and conditions constitutes acceptance of the current terms and conditions.
  3. The terms and conditions of any subscription agreements which educational and other institutions have entered into with the LRB apply in addition to these terms and conditions.
  4. You undertake to indemnify the LRB fully for all losses damages and costs incurred as a result of your breaching these terms and conditions.
  5. The information you supply on registration to the LRB Website shall be accurate and complete. You will notify the LRB promptly of any changes of relevant details by emailing the registrar. You will not assist a non-registered person to gain access to the LRB Website by supplying them with your password. In the event that the LRB considers that you have breached the requirements governing registration, that you are in breach of these terms and conditions or that your or your institution's subscription to the LRB lapses, your registration to the LRB Website will be terminated.
  6. Each individual subscriber to the LRB (whether a person or organisation) is entitled to the registration of one person to use the 'subscriber only' content on the web site. This user is an 'individual user'.
  7. The London Review of Books operates a ‘no questions asked’ cancellation policy in accordance with UK legislation. Please contact us to cancel your subscription and receive a full refund for the cost of all unposted issues.
  8. Use of the 'subscriber only' content on the LRB Website is strictly for the personal use of each individual user who may read the content on the screen, download, store or print single copies for their own personal private non-commercial use only, and is not to be made available to or used by any other person for any purpose.
  9. Each institution which subscribes to the LRB is entitled to grant access to persons to register on and use the 'subscriber only' content on the web site under the terms and conditions of its subscription agreement with the LRB. These users are 'institutional users'.
  10. Each institutional user of the LRB may access and search the LRB database and view its entire contents, and may also reproduce insubstantial extracts from individual articles or other works in the database to which their institution's subscription provides access, including in academic assignments and theses, online and/or in print. All quotations must be credited to the author and the LRB. Institutional users are not permitted to reproduce any entire article or other work, or to make any commercial use of any LRB material (including sale, licensing or publication) without the LRB's prior written permission. Institutions may notify institutional users of any additional or different conditions of use which they have agreed with the LRB.
  11. Users may use any one computer to access the LRB web site 'subscriber only' content at any time, so long as that connection does not allow any other computer, networked or otherwise connected, to access 'subscriber only' content.
  12. The LRB Website and its contents are protected by copyright and other intellectual property rights. You acknowledge that all intellectual property rights including copyright in the LRB Website and its contents belong to or have been licensed to the LRB or are otherwise used by the LRB as permitted by applicable law.
  13. All intellectual property rights in articles, reviews and essays originally published in the print edition of the LRB and subsequently included on the LRB Website belong to or have been licensed to the LRB. This material is made available to you for use as set out in paragraph 8 (if you are an individual user) or paragraph 10 (if you are an institutional user) only. Save for such permitted use, you may not download, store, disseminate, republish, post, reproduce, translate or adapt such material in whole or in part in any form without the prior written permission of the LRB. To obtain such permission and the terms and conditions applying, contact the Rights and Permissions department.
  14. All intellectual property rights in images on the LRB Website are owned by the LRB except where another copyright holder is specifically attributed or credited. Save for such material taken for permitted use set out above, you may not download, store, disseminate, republish, post, reproduce, translate or adapt LRB’s images in whole or in part in any form without the prior written permission of the LRB. To obtain such permission and the terms and conditions applying, contact the Rights and Permissions department. Where another copyright holder is specifically attributed or credited you may not download, store, disseminate, republish, reproduce or translate such images in whole or in part in any form without the prior written permission of the copyright holder. The LRB will not undertake to supply contact details of any attributed or credited copyright holder.
  15. The LRB Website is provided on an 'as is' basis and the LRB gives no warranty that the LRB Website will be accessible by any particular browser, operating system or device.
  16. The LRB makes no express or implied representation and gives no warranty of any kind in relation to any content available on the LRB Website including as to the accuracy or reliability of any information either in its articles, essays and reviews or in the letters printed in its letter page or material supplied by third parties. The LRB excludes to the fullest extent permitted by law all liability of any kind (including liability for any losses, damages or costs) arising from the publication of any materials on the LRB Website or incurred as a consequence of using or relying on such materials.
  17. The LRB excludes to the fullest extent permitted by law all liability of any kind (including liability for any losses, damages or costs) for any legal or other consequences (including infringement of third party rights) of any links made to the LRB Website.
  18. The LRB is not responsible for the content of any material you encounter after leaving the LRB Website site via a link in it or otherwise. The LRB gives no warranty as to the accuracy or reliability of any such material and to the fullest extent permitted by law excludes all liability that may arise in respect of or as a consequence of using or relying on such material.
  19. This site may be used only for lawful purposes and in a manner which does not infringe the rights of, or restrict the use and enjoyment of the site by, any third party. In the event of a chat room, message board, forum and/or news group being set up on the LRB Website, the LRB will not undertake to monitor any material supplied and will give no warranty as to its accuracy, reliability, originality or decency. By posting any material you agree that you are solely responsible for ensuring that it is accurate and not obscene, defamatory, plagiarised or in breach of copyright, confidentiality or any other right of any person, and you undertake to indemnify the LRB against all claims, losses, damages and costs incurred in consequence of your posting of such material. The LRB will reserve the right to remove any such material posted at any time and without notice or explanation. The LRB will reserve the right to disclose the provenance of such material, republish it in any form it deems fit or edit or censor it. The LRB will reserve the right to terminate the registration of any person it considers to abuse access to any chat room, message board, forum or news group provided by the LRB.
  20. Any e-mail services supplied via the LRB Website are subject to these terms and conditions.
  21. You will not knowingly transmit any virus, malware, trojan or other harmful matter to the LRB Website. The LRB gives no warranty that the LRB Website is free from contaminating matter, viruses or other malicious software and to the fullest extent permitted by law disclaims all liability of any kind including liability for any damages, losses or costs resulting from damage to your computer or other property arising from access to the LRB Website, use of it or downloading material from it.
  22. The LRB does not warrant that the use of the LRB Website will be uninterrupted, and disclaims all liability to the fullest extent permitted by law for any damages, losses or costs incurred as a result of access to the LRB Website being interrupted, modified or discontinued.
  23. The LRB Website contains advertisements and promotional links to websites and other resources operated by third parties. While we would never knowingly link to a site which we believed to be trading in bad faith, the LRB makes no express or implied representations or warranties of any kind in respect of any third party websites or resources or their contents, and we take no responsibility for the content, privacy practices, goods or services offered by these websites and resources. The LRB excludes to the fullest extent permitted by law all liability for any damages or losses arising from access to such websites and resources. Any transaction effected with such a third party contacted via the LRB Website are subject to the terms and conditions imposed by the third party involved and the LRB accepts no responsibility or liability resulting from such transactions.
  24. The LRB disclaims liability to the fullest extent permitted by law for any damages, losses or costs incurred for unauthorised access or alterations of transmissions or data by third parties as consequence of visit to the LRB Website.
  25. While 'subscriber only' content on the LRB Website is currently provided free to subscribers to the print edition of the LRB, the LRB reserves the right to impose a charge for access to some or all areas of the LRB Website without notice.
  26. These terms and conditions are governed by and will be interpreted in accordance with English law and any disputes relating to these terms and conditions will be subject to the non-exclusive jurisdiction of the courts of England and Wales.
  27. The various provisions of these terms and conditions are severable and if any provision is held to be invalid or unenforceable by any court of competent jurisdiction then such invalidity or unenforceability shall not affect the remaining provisions.
  28. If these terms and conditions are not accepted in full, use of the LRB Website must be terminated immediately.
The Search for the Perfect Language 
by Umberto Eco, translated by James Fentress.
Blackwell, 385 pp., £24.95, September 1995, 0 631 17465 6
Show More
by Gérard Genette, translated by Thaïs Morgan.
Nebraska, 446 pp., £23.95, September 1995, 0 8032 2129 0
Show More
Show More

Anyone who has ever felt drawn to the remote but seductive question of what form the first human language may have taken will have been stirred the other day by Gillian Shephard’s announcement that the Government is going to spend (a very little) money on coaching our young inarticulates so that they stop ‘grunting’ and start using words. This looks rather like an attempt to recapitulate on the cheap the slow linguistic evolution of the species, as Trevor McDonald and his fellow therapists educate the grunters out of the animal and into the human state. Except, of course, that the grunts complained of are not natural phenomena but already signs, an admittedly crude but still authentic element of culture. All grunts are not identical, either in the way they sound or in what they may be taken to mean. They depend for their interpretation on how grunters grunt, in response to what, and who they grunt to (or at). They are not to be so easily dismissed as prehistoric intruders in our otherwise eloquent midst.

What they do, however, is to hark topically back to the big idea that binds these two learned, sceptical, on occasions witty, books together. This big idea concerns the relation we believe to hold between the words that we use and whatever we use them to refer to: between so-called ‘natural’ language and the extra-verbal world beyond it. It is an irony that we should have now come to call English, say, a ‘natural’ language when from one point of view that is exactly what it isn’t, since the form that it takes has been determined not by directions given to Anglophones by nature but by agreement among the generations of those who have used and elaborated it.

The deep question of how ‘natural’ human language is or isn’t goes a long way back, by scholarly tradition if not in historical fact to Plato and the dialogue known as the Cratylus. In this, two incompatible views of language are set in opposition. According to one view, argued for by Hermogenes, language is an institution founded on convention and on that alone. The forms that any particular language contains have arisen and have evolved through time, and bear only an arbitrary relationship to the world beyond language that they are used to refer to. There’s nothing inherently horsey, in short, about the word horse, or cheval, or Pferd, or any of the many other terms to be found in the world’s languages which may serve to identify this particular natural item. This is the doctrine of the ‘arbitrariness of the sign’ redefined early this century by Saussure and agreed to, I imagine, give or take some quite small refinements, by everyone today who has ever given the matter any thought. Hermogenes’ opponent in Plato is the Cratylus for whom the dialogue is named, though he cannot be said to have argued his way to victory. Cratylus’ view is that the relation of language to reality is not arbitrary or conventional but mimetic, that words imitate things, which have the names that they do because these are the ‘right’ names for them, imposed of necessity by the actual properties of what it is they name.

It might look as though any rational compromise would be hard to find between these two positions, but Socrates advances one, as Gérard Genette brings out in a typically rigorous analysis of the dialogue. Socrates emerges from it in the end as a ‘disappointed Cratylist’: his thesis is that whoever first gave things their names did a poor job on the whole, since so many of the names – though not all: this is something to come back to in a moment – seem to be wrong, or inappropriate, appropriateness being held to lie in a self-evident match between their sound and their sense. For Socrates, and for those who have thought like him in the many centuries since, the arbitrariness of the sign is an affliction, brought about by the incompetence of the first namer(s).

Mimologics (now very well and carefully translated, twenty years after it first appeared in French) gives a masterly account of Cratylism, and of the various ways in which its basic notions have been reformulated, stretched, narrowed or modernised by the succeeding waves of those anxious in one degree or another to re-assert the claims of a natural ‘motivation’ in language over those of a conventionalism damnable for being the mark of our cosmic alienation. In this sense, Genette’s book, like Eco’s, too, in part, is the history of a doomed but often laudably ingenious movement to go against the linguistic grain and rediscover a truly natural language: a language of Nature or of God as it were, the appropriateness of whose signs there could be no denying.

There is overlap between Mimologics and The Search for the Perfect Language, though Eco’s book moves the faster and has in any case a subject that extends beyond Cratylism. As Eco points out, a perfect language doesn’t in fact have to be Cratylist, or not in any usual understanding of that term. What he calls the ‘a priori’ languages, thought up by such as the 17th-century Bishop of Chester, John Wilkins, or by Leibniz, in order to remedy the perceived illogicalities of existing languages, are mimetic in one sense, because the verbal forms they contain are derived from an analysis of the perceived properties of the things or ideas they stand for; but that analysis is, as Eco makes clear in some of his best pages, in itself arbitrary and very much culture-bound, so that any necessary connection between the linguistic and the natural has been broken. Does anyone need reminding, at this late stage, of Borges’s exquisite parody of Wilkins’s arbitrary method of classification in his essay on that muddled prelate – a joke influentially recycled by Foucault at the start of Les Mots et les choses?

The Search for the Perfect Language is a brisk, chronological account of the many thinkers about language, from antiquity onwards, who have conceived programmes for undoing the effects of time and either recovering the ur-language that they believed must once have existed only later to be lost, or else inventing a replacement for it. One tradition which survived for centuries thought that, for good Biblical reasons, the proto-language must have been Hebrew, the language that came before the building work of Babel and the divinely ordered confusion of tongues that followed. Ethnic presumption having always had much to say in questions of this kind, rival candidates were later proposed as deserving of the priority, the strangest of all these surely being the dialect of Dutch spoken around Antwerp, which was ur because, asserted Goropius Becanus of Antwerp, the Antwerpers’ ancestors had had the unique good fortune not to be present when the Tower of Babel was thrown down.

The ideals of the language unifiers, like their presumption, were frequently of the highest, since they assumed that if there were only one language in it, the world would become a better, happier place. ‘Very early on, I had dreamt of plans for perfecting grammar and achieving unity in the language system, from which I quite naturally thought a great amelioration of society would ensue ... eternal peace and the universal confraternity of peoples.’ This was one of the most resourceful and engaging of Genette’s Cratylists, the poet and linguist Charles Nodier, writing in 1828 about the hopes he had felt as an 18-year-old, fired as he then was no doubt by the propaganda of the French Revolution. For the dream of linguistic unity could hardly not have a politics to go with it. This could as well be a dream of imposing a single ‘perfect’ language on all the world’s speakers in the service of territorial imperialism (Stalin, let’s remember, took up the idea obsequiously fed to him by the infamous and deluded Marr that his native Georgian was the original language), as one of installing a millennium of peace and goodwill, founded on the unreasonable assumption that if we all spoke the same language we would never want to go to war with one another. Not for nothing did the originator of the most successful of the 20th century’s ‘international’ language, Dr Ledger Ludwik Zamenhof, take the pseudonym of Dr Esperanto (=Dr Hopeful).

The Cratylist dream is in its simplest form onomatopoeic. It wants words to sound right, so that we would all allow they accord with what they mean. That some words of our language sound right is not in dispute, and to that extent Socrates had a case: there are times when phonetics and semantics seem to go hand in hand. The English words ‘crash’, ‘bang’, ‘wallop’ all sound as if they might well have begun as attempts at a mimeticism of audible, tending to violent events – rather as S.J. Perelman once memorably said that in the names of the Madison Avenue advertising agents, Batten, Barton, Durstein and Osborn, you could but hear a tin trunk falling downstairs. That spoken words seem to fit perfectly with their sense is no sort of proof, however, that they are onomatopoeic; or better, were onomatopoeic, because, as Saussure famously observed, the difficulty with onomatopoeic words is that, once they have entered the language, they are subject to phonetic change just like other words. In my dictionary (Chambers), the etymology of crash is indeed given as ‘imit.’, though how long ago the imit. occurred is not stated. Bang, on the other hand, is said to come from Old Norse banga, ‘to hammer’, and wallop from Old French waloper, ‘to gallop’. This perhaps lets the Anglophone onomatopoeticist off the hook. Did Old Norsemen decide that banga was just right for the sound of a hammer hitting a nail, and Old Frenchmen that waloper was ideal for the sound of a galloping horse? To suppose so is to assume that onomatopoeia holds across languages that have different phonetic values, a universalising assumption that would be hard to defend. Mostly onomatopoeia is an agreeable fancy, but no less significant or productive, as Genette shows, for being fanciful.

The Cratylism of the Cratylus is of the simplest kind, like the examples offered above. Plato is concerned there with a one-to-one relation between this particular word and that particular thing, as if a language consisted purely of a nomenclature or a collection exclusively of nouns. Moreover, not even Socrates gives any attention to languages other than Greek, although the existence of a multiplicity of languages poses a large problem for a Cratylist, who is faced with trying to establish the evidence for mimeticism in all of them, instead of in just the one he knows best, his own. The way round this difficulty is to take a rather broader view of what actually counts as mimeticism, so that the different terms that exist in the world’s languages for the ‘same’ thing may be argued as mimicking different aspects of that thing. The lines of communication between the real and the verbal can then be made more complex, and the determination of the second by the first made looser, allowing room for the idea that the signifiers of a language may be metonymic or metaphorical of their referents.

More promising as a line of defence, however, is that pursued by a number of French writers discussed in the second half of Mimologics, who elaborated on earlier attempts to onomatopoeticise not whole words of the language but some at least of its phonemes. Nodier was remarkably adept at this, and so above all was Mallarmé, on whom, as one would expect, Genette spends some profitable and incisive pages. Mallarmé descends in a direct line from the Socrates of the dialogue inasmuch as he, too, is a disappointed Cratylist, but one unprepared to rest in his disappointment. On the contrary, his dissatisfaction with the inappropriate sounds of which the French vocabulary is full lies at the heart of his vocation as a poet. ‘Verse remunerates the failing of natural languages, being their superior complement.’ What this means is that the task of the poet, according to Mallarmé, is to so use his defective natural language as to counteract its downward tendency towards prose. In a celebrated passage Mallarmé wrote:

Languages are imperfect because multiple, the supreme one is missing ... the diversity of idioms on earth prevents anyone from uttering words which, otherwise, were they to appear in a single flash, would be truth itself incarnate ... from the aesthetic perspective, I regret to see how discourse fails to express objects by means of keys that would correspond to them in colouring or in aspect – keys that do exist in the instrument of the voice, among languages and sometimes in one language. When compared to the opacity of the word ombre, the word ténèbres does not seem very dark; how disappointing is the perversity that contradictorily assigns dark tones to jour, bright tones to nuit.

This perversity of French is also invigorating, since it challenges the frustrated Mallarméan poet to deploy the phonemes of the language in such a way as to make verse into a peculiarly privileged because Cratylist medium, in semi-independence from the ‘ordinary’, merely conventionalist language that others are prepared to live with. With Mallarmé you could say that Cratylism learns how to imitate itself.

Not that it needs to be preoccupied only, or at all, with the sounds of a language; there have been graphic Cratylists also, well described by Eco, who were nostalgic for a script so constructed that it could represent the real world pictorially. When Egyptian hieroglyphics were rediscovered, in the 15th and 16th centuries, it was assumed, understandably, that all its signs were iconic, or determined visually by the things they stood for. Alas no: many of them are not pictures as such but ideograms or, worse still, are determined phonetically by a form of punning, so that a hieroglyph may represent a sound of the language by displaying the image of something whose name sounds similarly. This contamination of the iconic by the quasi-alphabetic was bad news for Cratylists.

Even with a fully phonetic alphabet like our own all is not lost, however, for the possibility remains for what Genette labels, precisely if cumbrously, ‘articulatory phonomimesis’. This is the belief that the graphic forms taken by the letters of the alphabet represent the physical disposition of the speech organs pictured in the act of articulating them. In the case of one or two letters – vowels, inevitably – this can seem almost plausible, as when the great Julius Caesar Scaliger compares the phonomimetics of O and I and decides that ‘the figure [O] comes from a representation of the circle of the mouth, whereas the I, which notates the shrillest sound, appears without either hump or belly.’ But even if we are persuaded by this, we need only to move away from the relative straightforwardness of O and I to the oral gymnastics required by H and Z for the phonomimetic edifice to totter and fall.

More rewarding and suggestive by far is that version of Cratylism that chooses to operate not at the level of individual words but of the sentence, and in doing so moves decisively into the present. It moves also away from the sensual and towards the grammatical. The Cratylist claim now becomes that the structure of a sentence reflects the structure of reality. Eco, the happy medievalist, introduces this way of thinking early on in his story, with the ‘Modist’ grammarians of the 13th century who ‘asserted a relation of specular correspondence between language, thought and the nature of things’. Their thesis seems to have been much the same as that of the early Wittgenstein in the so-called ‘picture theory’ of language that he later gave up, and which had proposed a formal equivalence between verbal propositions and their subject-matter.

This was an advanced and defensible version of an argument that surfaces several times in Eco’s book, and not always so defensibly. Cratylists and linguistic perfectionists were often led astray by a fierce desire to promote their own language above all others, if need be by going to the limit, à la Goropius Becanus of Antwerp, and declaring that we need look no further for the perfect language because it is all around us, it is English or French or whatever. This claim may extend beyond vocabulary to grammar. Count Antoine de Rivarol, to take a characteristic example, writing in 1784, averred that the French language was ideal in every conceivable way but above all because of its word order, which followed that of a ‘natural logic’. Many of us met with curious vestiges of this belief when we were small and were told by one or other teacher of Latin that by learning that language we would be ‘taught to think’ – the unspoken premise of which assurance was that Latin was closer to nature than English, that its, to us, erratic word order better represented the order to be found inscribed once and for all in the human mind.

Where Rivarol and the rest went adrift was in maintaining that one word order is superior to another for being more natural. Had they gone in for comparatism rather than chauvinism, they might have found the far from negligible argument that any word order reflects one possible realisation of underlying grammatical structures that are common to all languages and to all thought, if not necessarily to the reality we think about. Thus, if you accept that thought and language are separable, it is possible even now to preserve a modicum of Cratylism by arguing that the sentences we speak or write are mimetic of the universal ‘mentalese’ or Chomskyan ur-language that functions deep inside every human brain. In this attenuated, as yet secret, form Cratylism still lives.

Send Letters To:

The Editor
London Review of Books,
28 Little Russell Street
London, WC1A 2HN

Please include name, address, and a telephone number.

Read anywhere with the London Review of Books app, available now from the App Store for Apple devices, Google Play for Android devices and Amazon for your Kindle Fire.

Sign up to our newsletter

For highlights from the latest issue, our archive and the blog, as well as news, events and exclusive promotions.