2. Searching for the first language

What was the original language? For a long time, it was thought that, if left uninfluenced, infants might somehow come out with it. The first pharaoh of the twenty-sixth Egyptian dynasty, Psammetichus I (663–610 BC), arranged for two babies to be nurtured without hearing any language. Writing about the experiment two centuries later, in the second book of his Histories, Herodotus reported that, at the age of two, the infants apparently said becos, a Phrygian word meaning ‘bread’:

Frederic II, the thirteenth-century Holy Roman Emperor, attempted a similar experiment, but without success, for it was found that ‘the children could not live without clappings of the hands, and gestures, and gladness of countenance, and blandishments’ (Salimbene 1857). To put it more bluntly, the children all died. And when James IV of Scotland put two infants with a dumb woman, Lindesay the historian noted (Reference Lindesay1728) that ‘some say they spoke good Hebrew, but as to myself I know not but by hearsay’. A few years after the Scottish experiment, the Mogul Emperor Aqbar repeated it (see Ahmad ’Alí Reference Ahmad ’Alí1864; Beveridge Reference Beveridge1897). He believed, however, that children isolated from human speech would not speak, and in his experiment, some twenty children sequestered for four years in a gang mahal (a house of the dumb) supported this belief:

Like the unfortunate children in Frederic's experiment, many of these children, too, became – as Badā’uni put it in his monumental history – ‘nurselings of mother earth’ (Ahmad ’Alí Reference Ahmad ’Alí1864: 296).

At about the same time, Michel de Montaigne (Reference de Montaigne1580: 181−182) wrote in (cautious) support of the earlier views:

Naturally-occurring ‘experiments’, however, support Aqbar's belief. ‘Wolf-children’, ‘bear-children’ and other such feral youngsters have been widely reported (if rather sparsely documented) for hundreds of years. None could speak or understand and, indeed, most efforts to teach them language were failures. Victor, the ‘Wild Boy of Aveyron’ discovered at the turn of the nineteenth century, aged about eleven, is perhaps the best-described case (see Malson Reference Malson1964; Lane Reference Lane1976; Shattuck Reference Shattuck1980); although developing some powers of comprehension, he too was unable to speak. All of this suggests what is now obvious to most people – that although human beings may be ‘prewired’ and ready, in an evolutionary sense, to learn language, the manifestation and the specific form of this capacity are determined by social influence.

It is in Plato's Cratylus that we find the first formal treatment of language origins; its subtitle is On the Correctness of Names. In the first dialogue, Hermogenes argues that language is ‘conventional’, that names are arbitrary and do not somehow contain the essence of what is named. Socrates counters by suggesting, in some lingua humana fashion, that names ‘imitate’ reality, and that the original ‘name-givers’ must have possessed the knowledge of nature necessary for the coining of ‘correct’ names. At first, he takes the part of Cratylus who, in contrast to his friend Hermogenes, holds that ‘names belong to things by nature’. Socrates describe a sort of phonetic symbolism, arguing that certain letters are ‘naturally’ associated with certain things and ideas; thus, rho suggests motion, lambda suggests gliding or sleekness, delta and tau connote binding and stasis, and so on. But while it would be nice, he concludes, if names were essentially like the things denoted, he admits that this theory will not take us very far, and that ‘we are compelled to employ in addition this commonplace expedient, convention, to establish the correctness of names’. In his De Interpretatione (Chapter 2, pp. 43−44), Aristotle agreed. Words and names ‘work’ because of arbitrary agreement, not through some intrinsic mystical connectivity between them and the objects to which they refer. ‘A name,’ he pointed out, ‘is a spoken sound significant by convention. I say “by convention” because no name is a name naturally, but only when it has become a symbol’.

Socrates, Aristotle and other scholars show us, in other words, that the force of semantic ‘convention’ has been understood for a very long time. Nonetheless, the Bible suggests that – at least for a while – Cratylus was right, after all. Genesis II:19 tells us that God formed all the birds and beasts, ‘and brought them unto Adam to see what he would call them: and whatsoever Adam called every living creature, that was the name thereof’. And Adam says (in the eighth book of Milton's Paradise Lost) that at his creation he ‘readily could name whate'er I saw.’ As God sent all the birds and beasts to him, Adam ‘nam'd them, as they pass'd, and understood their nature, with such knowledge God endu'd my sudden apprehension’. The original language of Eden – unlike all natural languages since – encapsulated a perfect correspondence between words and things. What could this first, divine language have been?

From the earliest times, this question was of the first importance; for, as Rubin (Reference Rubin1998) has pointed out, those who speak that language – or, at least, those who can claim them as ancestors – have a special, intimate, or ‘chosen’ relationship with divinity. And it is clearly a case of the medium and not the message. Little wonder, then, that claim and counter-claim were so important; after all, conceptions of social and political group identity would be immeasurably (quite literally) strengthened if their linguistic, cultural and religious components had been fashioned by God himself. Rubin considers the matter from early Jewish, Christian and Muslim perspectives.

From about the second century BC, Jewish literature described Hebrew as the first language: the medium of revelation, the language in which God spoke to Adam, the language used by all creatures until the fall, and by all people until Babel. Hebrew had the support of most of the Christian community, too. Virtually all of the Greek, Latin and (later) Byzantine patriarchs supported its primary status, as did most of the Arabic-speaking Christian theologians. Still, Hebrew did not have it all its own way: a pagan hellenistic response argued that it was a construction that had only come into being with the exodus from Egypt; it could hardly, then, be the language of creation. Aramaic, a variety widely spoken throughout the Middle East generally, and within the Jewish community specifically, was a rival for top honours. Interesting arguments ensued about the particular Aramaic/Syriac dialect that was the medium of Eden.

Among early Muslim scholars, the claims of Hebrew and Aramaic were both seen to be strong – with the latter predominating. In fact, given the seemingly powerful claims of Aramaic, an interesting accommodation was made: Adam's language was Arabic, he spoke Syriac after his expulsion from the garden – but, upon repenting, he returned to speaking Arabic (Kister Reference Kister1993). It was inevitable, after all, given the all-important linguistic buttressing of group identity, that primacy for Arabic, the language of the Qur'ān, would be claimed. As Kister (Reference Kister1993: 140) describes the argument: ‘Adam descended from Paradise speaking Arabic because Arabic was the language of God, of the angels, and the people of Paradise’.

Of course, linguistic survival was also important here. Did the first language – Hebrew perhaps, or Aramaic, or Arabic – survive in some form or other? Even if Hebrew were the Adamic language, its contemporary varieties had surely lost that essential ‘character’ that permitted a perfect fit between words and things. As Bishop Wilkins (soon to be important in our story) pointed out:

In fact, Wilkins was clearly sceptical, for elsewhere he says that the deficiencies in the known Hebrew meant that it was not ‘the same which was con-created with our first Parents, and spoken by Adam in Paradise’ (p. 5). Like other scholars of the time, Wilkins assumed that some ‘primitive’ Hebrew was the Adamic language but, equally, that a search for that lingua humana could not possibly succeed. Besides, one had to take Babel into account: Hobbes pointed out (in his Leviathan) that all of Adam's first linguistic invention was ‘lost at the tower of Babel, when by the hand of God, every man was stricken for his rebellion, with an oblivion of his former language’ (Reference Hobbes1651: 13).

Nonetheless, as I mentioned in the Introduction, conjectures about the ‘first’ language continued throughout the sixteenth and seventeenth centuries. The wild surmises were derided by important thinkers of the time. Vico, for instance, wrote of ‘opinions so uncertain, inept, frivolous, pretentious or ridiculous, and so numerous, that we need not relate them’ (Reference Vico1725: 430). Eco, who discusses Vico's ‘acid’ comments, also cites a letter written by Leibniz in 1699 in which he too ridicules those wishing to ‘draw out everything from their own language’ (Reference Eco1993: 100). Eco also points out, however, that Leibniz himself was not above making a nationalistic language claim, supporting a ‘Celto-Scythian’ hypothesis that would embrace German. After all, ‘we Germans have a peculiar touchstone for thoughts, which is unknown to others; and, when [others] are eager to know something about this, I tell them that it is our very language’ (Leibniz Reference Leibniz1838, I: 452−453).

Language-origin theories remained important, and, as we have just seen, seductive. Those scholars who first began to think about linguistic classification in any sort of modern sense remained heavily influenced by religious texts – they were, of course, very often churchmen themselves – and, with hindsight, Müller (Reference Müller1862) lamented the amount of scholarship that had been devoted to a fruitless quest for the lingua Adamica, a quest that had seriously retarded the progress of linguistic science. Matters seemed very different at the time, of course.

But earlier, in the sixteenth century, with the emergence of the ‘new science’, the story begins to change direction. Babel and the confusio linguarum remained of interest, as did the hope of somehow regaining the language of Eden. More and more, however, scholars (such as Bishop Wilkins, mentioned above) began to believe that, while a more perfect language would be highly desirable, such a medium would have to be constructed, a human contrivance to replicate the pre-Babel universality, as it were. Language projects thus began to emerge, and soon to multiply. Their most obvious motivation was a scientific one, but the religious angle was still important. Eco mentions the enhancement of scientific communication across countries and cultures, and the facilitation of taxonomic efforts – but he also writes that some new, man-made lingua perfetta could help to convert the infidel, to unify the three great religions and, especially, to mitigate the dissension between Catholics and Protestants. In other words, as we leave behind the quest for the language of Eden, and enter the age of the language ‘projectors’, as they called themselves, we must not imagine that we have replaced the original religious impulses with the new scientific ones. (Nor, as we shall see, has a generally spiritual motivation every really faded away at all.)

3. Making a more perfect language

The great Czech educator-theologian Comenius (Jan Komenský) argued for a new universal language – an a priori medium, built from first principles, with a new set of ‘real’ characters. In a work written much earlier but published in 1668, he advocated a completely regular medium, which – like the lingua humana – would ‘express all concepts and all things in all their special characteristics. . .its names will be made to match the numbers, measurements and weights of things themselves’ (p. 192). Comenius pointed out that such a lingua franca was not meant to displace anyone's mother tongue – an observation that has been generally endorsed by language makers ever since.

John Wilkins was one of the founders of the Royal Society and its first (co-) secretary, a fact that underlines the intertwining of the new Baconian thrust with the desire that a better language be available to support and guide scientific classification and discovery.Footnote ² In fact, Bishop Wilkins was, as Stillman (Reference Stillman2002: 308) notes, ‘indisputably the most accomplished of the language projectors of seventeenth-century Europe’. With Comenius and others encouraging a new and better language, Wilkins published his Essay Towards a Real Character, and a Philosophical Language (Reference Comenius1668: the word ‘philosophical’ here signifies ‘logical’ or ‘scientific’). The idea was to link an ordered knowledge of the natural world with a universal grammar to describe it. A simple example illustrates the idea:

Readers may already be thinking that such a scheme is more easily described than accomplished to any satisfactory degree, and we'll see shortly that it is, indeed, unworkable, and that all language projectors of the time were misguided. But we should also bear in mind, with Firth (Reference Firth1970: 63), that their productions were not merely ‘interesting linguistic freaks’. Why? Simply because ‘all subsequent standardization of scientific nomenclature according to a reasoned system is in continuation of the need then felt and expressed’. Salmon makes a related observation (Reference Salmon1975: 140): ‘Wilkins aimed at correlating language and reality so that, as Bacon had desired, their components should be isomorphic, i.e. so that every element of reality should be represented by a single concept in the mind and by a single character in the written language’. It is one thing to criticise such a scheme as utopian, but it would be quite another to see its author as inane.

Wilkins's Essay is presented in four main sections: an introduction; a ‘universal’ classification of things and ideas; the grammar of the language; the ‘real character’ – that is, the new orthography – and the language itself. His introduction is more interesting than most such prolegomena, in that he discusses the earliest languages and their ‘several offsprings’ (p. 1), language loss and change over time – as well as the ‘corruptions to which all vulgar languages are obnoxious’ (p. 6) – and writing systems and their imperfections. He ends with the observation that, preliminary to any scientific system of language, ‘the first thing to be considered and enquired into is, concerning a just enumeration and description of such things or notions as are to have marks or names assigned to them’ (p. 20).

This leads naturally into the second, the lengthiest and the most vital section of the Essay, in which Wilkins outlines his classification plans. These cover everything from God and the spiritual world, to stones, plants and animals, to matters of quantity (magnitude, space and measure) and quality (judgement, conscience, habits and manners), to the human senses, to the categorisation of disease, to actions, operations and motions, to ‘oeconomical’ relationships, possessions and provisions, to public and social relationships (civil, judicial, military and ecclesiastical). Wilkins provides page after page of classification and sub-classification, by which all these things may be logically characterised. Stones – to give a physical example – are divided, first, into ‘vulgar’ (‘and of no price’), ‘middle-prized’ and ‘precious’. The vulgar ones then sub-divide according to their magnitude, consistency and use (large ones are used for building walls and houses, smaller ones comprise sand and gravel). There are also sub-categories for materials used for the sharpening, cutting and polishing of stones. ‘Qualitative’ classification is exemplified by the arrangements made for ‘understanding’, which is broken down into components such as thinking, believing and reasoning, but also esteeming, approving and trusting.

With this grand typology in place, Wilkins moves on – in the third section of the book – to the ‘philosophical’ grammar which is to reflect and communicate it. As he puts it:

In this section, Wilkins discusses grammar in terms of structure and classification. Thus, he notes that all words are either integrals (‘principal words. . .such as signifie some entire thing or notion’: p. 298) or particles (‘less principal words, which may be said to consignifie, serving to circumstantiate and modifie those integral words’: p. 304). Important words in the first class are nouns and verb, and, in the second, prepositions and conjunctions – but this does little justice to Wilkins's complicated grammatical arrangements. He has a category for ‘transcendental particles’, in which (for example) the endings ‘-like’ or ‘-ish’ are seen to expand original senses, often in some metaphorical way (‘childlike’ or ‘devilish’); Wilkins himself calls transcendental those particles ‘which do circumstantiate words in respect of some metaphysical notion’ (p. 318). He also has many things to say about vowels and consonants and diphthongs, the physical production of different sounds, and other matters of grammar and syntax.

In the fourth part of his Essay, Wilkins gives us his ‘real character’. Here, an extremely complicated set of ‘marks’ – lines, loops, hooks, points, circles – are to represent, in dizzying permutations, all the categories previously outlined in the preceding grammatical section, and in such a way that these categories map logically on to the opening section, in which the world is itself classified. Thus, Wilkins denotes all pronouns by arrangements of points, all prepositions by ‘small curved figures’, all interjections by ‘upright lines, streight or hooked’, and so on (p. 388). He illustrates his ‘real character’ system by presenting the Lord's Prayer and the Apostles’ Creed ‘written’ in this curious fashion, followed by a very full explication of the use of the characters and their representations.

But all this is still only to reproduce things and ideas by a commonly understandable set of marks or characters, to represent them in ‘writing’. After all, it is the essence of the ‘real character’ idea to achieve an accurate representation of reality through a universally understood orthography or set of markings. The markings in themselves give no guide to pronunciation. How are they to sound?

The general procedure in this last part of the Essay is to assign particular sounds and letters for the initial categories. For a simple example, we can return to the categorisation for stones, noted above. The consonant-and-vowel combination ‘di’ represents the category, and further letters will specify further sub-division. Since the first consonant is ‘b’, the term ‘dib’ will then signify the first sub-category of stone (i.e. ‘vulgar’ stone). Since the first vowel is ‘a’, the extension ‘diba’ will mean a vulgar stone of the ragstone variety; and so on (see Knowlson Reference Knowlson1975 for a brief discussion of Wilkins's model, and of other universal-language schemes). To return to Wilkins's Lord's Prayer example, ‘Our father’ is ‘Hαι coba’ – the explanation of which is as follows:

Wilkins is the doyen of all the a priori language ‘projectors’ (his was not, of course, the only such attempt – only the best known and the most thoroughly carried out). At the time, his classification scheme for the ‘things and notions’ of the natural world was seen as a great advance upon existing categorisations of flora and fauna, and it made a contribution to the evolution of a fuller and more scientific nomenclature. ‘What Wilkins proposed to do with words is precisely what Linnaeus was later to do with plants’, Emery (Reference Emery1948: 176) notes, and the fact that the language project failed ought not to blind us to the fact that his scheme was at once a contribution to, and a reflection of, the great classificatory thrust of the time (see Cram Reference Cram and Hüllen1990, Reference Cram, Blaicher and Glaser1994).

At the end, however, Wilkins's scheme foundered on a common enough rock: to make a ‘philosophical’ language of ‘real characters’ sufficiently comprehensive to warrant the name is to make it too difficult to learn – despite the best efforts at internal coherence and mnemonic arrangements. Wilkins himself admitted, in the dedication of his scheme to the Royal Society, that he had only ‘slender expectations’ of it, although perhaps this only reflects the self-deprecation that is typical of such front matter. The initial enthusiasms of his Society colleagues and other scientists were generally not sustained, and Wilkins's attempts to interest others in furthering his work were not very fruitful.

Eco (Reference Eco1993) suggests that the passion for ‘philosophical’ languages had pretty well run its course by the end of the nineteenth century; however, as Slaughter (Reference Slaughter1982) and Large (Reference Large1985) have argued, the game was really up much earlier. The over-elaborate and often bizarre arrangements of various sorts of ‘characters’ – and the virtual impossibility of any normal person being able to memorise the necessary basic elements – came to appeal less and less. The a priori schemes also fell into disfavour because their strong religious underpinnings came into growing conflict with the claims – and the essentially secular demands – of the new science. Thus, some of the later attempts provided fuel for studies of scholars who were more obsessive than objective. Eco also notes that, apart from the continuing productions of these eccentrics – ‘the special province of visionaries of all lands’ (p. 308) – a priori language-making became a game, as in the creation of imaginary varieties in science-fiction novels. (The languages that Tolkien invented for The Lord of the Rings trilogy – notably Quenya and Sindarin – are not completely a priori in construction, however, since the Finnish and Welsh that Tolkien loved were formative influences; at the same time, they are not so obviously a posteriori productions as those to which we now turn.)

4. Contemporary constructed languages

From efforts to discover the lingua humana, to those dedicated to the production of a ‘universal’ a priori language, we come to the third chapter in our story. It involves the rise of the a posteriori languages, of which Esperanto has been far and away the most successful.Footnote ³ In one respect, the ‘new’ languages of this kind complete a circle that takes us back to the seventeenth century. In their attempts to come up with simple and spartan and regularised lexicons, we are reminded that – before they turned to the idea of ‘real characters’ and a ‘philosophical’ language – Wilkins and other like-minded ‘projectors’ had imagined a language whose components were words: new words, new ‘primitive’ or ‘root’ vocabularies, perhaps, but words of a familiar form, nonetheless. It was only when ideas of a quasi-hieroglyphic nature came to the fore that these initial ideas were set aside in favour of an entirely new system of markings that linked the names of things with their essential characters: pasigraphies. And, just as the idea of an a priori universal language came to fruition at a time when Latin was in decline as the European lingua franca, so the resurgence of interest in constructed languages occurred in the late nineteenth and early twentieth centuries, when French (in particular) was losing ground as the most common link language, and when revolution and war served to stoke what Eco calls the ‘pacifist illusion’ associated with Esperanto and its epigones.

There have been hundreds of universal-language schemes over the centuries and, in modern times, perhaps as many as 200 a posteriori schemes (see Monnerot-Dumaine Reference Monnerot-Dumaine1960; Porset Reference Porset1979). Many treatments of modern constructed languages have focused upon the century following Zamenhof's publication of Esperanto; this is both expedient and reasonable, given that broadly similar aims have always animated language-makers. The basic hope in all cases has been to produce some neutral and auxiliary variety that would facilitate global communication. Although some constructions have been modified or simplified versions of existing ‘natural’ languages, the desire for neutrality has typically left powerful existing languages out of the lists: they are always tinged by history, and often by imperialism. Thus the way is theoretically clear for a constructed language to fill what is perceived as a yawning and receptive gap. The idea that an easy-to-learn auxiliary language could contribute significantly to world peace may indeed be an illusion – the bitterest of conflicts, after all, have been civil wars – but it is not an ignoble conceit.

Ludwik Łazarz Zamenhof, a Polish oculist, had already interested himself in constructed languages when he published his own scheme in 1887, under the pseudonym of Dr Esperanto (‘the hoping one’). Encouraged by the initial response, Zamenhof soon published a second book on the language; further publications followed and the first international congress on Esperanto took place in Boulogne in 1905. It is clear that Zamenhof, like virtually all other makers and supporters of constructed languages, hoped that Esperanto would provide more than a universal second language to supplement, but not supplant, mother tongues. He believed it could contribute greatly to some ‘trans-national identity’, an apt goal for one who observed that ‘if the nationalism of the strong is ignoble, the nationalism of the weak is imprudent’ (Lieberman Reference Lieberman1979: 96). To dilute the former and to obviate the latter must have seemed a pressing need when Zamenhof said this – in 1914.

Like all contemporary constructed languages, the guiding principle behind Esperanto is simplicity and regularity. The core of the language lies in its famous ‘sixteen rules’ of grammar: thus, all nouns end in o (nominative case), with an n added for the accusative; the definite article la serves for all cases, numbers and sexes; all adjectives end in a; verb forms are the same, regardless of person or number; accent is always on the penultimate syllable; and so forth.

One of the enduring difficulties with a posteriori languages is that, despite their claims to a desirable neutrality, they are not equally easily accessible to all. Esperanto, for example, is essentially a romance language, with some eastern European overtones. For example, ‘Jen la domo de la familio Lang. En la domo estas ses ĉambroj’ (Here is the house of the Lang family. In the house there are six rooms). Or ‘Ekster la domo estas hundejo . . . kaj aŭtomobilejo’ (Outside the house is the doghouse and the garage). When we consider that ‘knabo’ is boy and ‘knabino’ is girl, or that ‘patro’ is father and ‘patrino’ is mother, we see how easy it has been to mount feminist critiques of another type of non-neutrality (the examples here are from Cresswell & Hartley's ‘Teach Yourself’ book: Reference Cresswell and Hartley1968).

An interesting objection to constructed languages is that, if successful, they would eventually begin to show the same internally-divisive tendencies that have historically affected ‘natural’ languages: that is, one would begin to see regional variations, perhaps even marked dialectal diversity. Applied to a priori languages, the point was made a very long time ago – see Destutt de Tracy (Reference Destutt de Tracy1801), for instance – and, more recently and specifically, van Deth (Reference van Deth1983: 27) has written that ‘si l'espéranto réussissait. . .il se créerait nécessairement de nouvelles langues, une nouvelle famille de langues’ [were Esperanto to succeed, it would naturally lead to the creation of new languages]. So success would mean the demise of the universal-auxiliary idea: success would mean failure. If a language is spoken, it changes; if some central agency tries to arrest or control change, then, as Eco put it, ‘ce n'est plus une langue naturelle’ [it is no longer a ‘natural’ language] (Ertl & Lo Jacomo Reference Ertl and Jacomo1994: 92). Well, even a globally-used Esperanto need not be another German, French or English, but one sees Eco's point. A counter-claim – one that goes back all the way to Zamenhof himself – is that, once launched, Esperanto has evolved in ways not unlike those of natural languages. Some have argued, too, that a purely auxiliary language might not be so prone to dialectal break-up, that it might not come under the same pressures for change that affect ‘natural’ varieties (see Martinet Reference Martinet1991; Eco Reference Eco1993). An Esperantist told Fallows (Reference Fallows1986) that an easily-learned and essentially constant medium is required for universality, but this neglects the existence of a formal linguistic apparatus within the language community – dictionaries and grammars, but also planning bodies and councils – that makes linguistic evolution possible, allows for the creation of necessary neologisms, and so on. It may be that change from the street, as it were, is very restricted in a constructed variety, and it may be that this is very undesirable, but the scholarly apparatus available to natural languages can be brought to bear upon constructed ones, too (for a brief overview, see Jansen Reference Jansen2003).

The greatest impediment to the spread of planned languages of all types has been the construction of a rationale powerful enough to appeal to large numbers of people. There is, indeed, a sort of Catch-22 here. Esperanto may be very easy to learn compared to ‘natural’ languages, but it does of course take effort. Most language learners invest their time and effort with some instrumental use as the goal – but no constructed language has as yet reached a level at which broad practical application can be claimed for it. Consequently, language-learning motivation is weak. Yet how can a sufficiently attractive degree of utility be achieved unless the speech community grows, unless more people agree to begin learning the language?

There has always been a debate about the breadth and popularity of Esperanto, about the number of its speakers: estimates for the latter range from several hundred thousand to fifteen million. One problem is that most of the information comes either from committed apologists or from detractors – and neither group has always had a sterling reputation for accuracy. Large (Reference Large1985: 95) cites one critic as follows: ‘the leading Esperantists constitute an enthusiastic group with a fondness for large numbers and an astonishing forgetfulness of the fact that a majority of those who purchase their textbooks and leaflets fail to make any particular study or use of the language after the first interest has worn away’. Piron (Reference Piron and Schubert1989) argues for about three million worldwide users, of whom about three-quarters are European. There are apparently more male than female Esperantists (perhaps of the order of 3:1), and Piron suggests that teachers constitute the largest single category (of the order of 20% to 25%), joined near the top of the list by clerical workers and some professionals (engineers, draughtsmen and technicians of various stripes figure prominently here). Forster (Reference Forster1982) adds to the profile a left-leaning political bias and low levels of orthodox religiosity.

Piron, Forster and other commentators have noted that the typical Esperantist is more highly educated than the population at large. However, a university study by Edwards & MacPherson (Reference Edwards and MacPherson1987) found that, while they knew more about constructed languages, faculty members were much less enthusiastic about them than were students. It was the faculty who most downplayed the usefulness of languages like Esperanto, who saw them as impractical, who viewed the whole enterprise as naïve. If we accept that Forster and others are right in relating favourable attitudes to constructed languages with a relatively high degree of education, might it also be the case that those with very high levels of education are less favourably disposed, precisely because they know more about such languages and the disadvantages and impracticalities associated with them?

If one attempts to summarise actual research findings concerning Esperanto and other constructed languages – ‘hard data’, so to say – one is struck immediately by their paucity. While a few famous linguists – Sapir, Firth and Jespersen (see below) come immediately to mind – have interested themselves in the area, their interests have generally been peripheral and have not translated into research. Meillet (Reference Meillet1918) was sympathetic to the idea of an international auxiliary language, while pointing out that it had never achieved any substantial realisation. (His famous observation was that theoretical discussions paled beside the fact that Esperanto worked. This part is often cited by Esperantists, but not Meillet's immediate qualification: what works must still be implemented.) Zellig Harris (Reference Harris1970), who was Chomsky's teacher, published a short but cautiously favourable paper, in which he pointed out some of the obvious advantages of an international auxiliary language – and went on to say that past failures were not in themselves ‘an argument against the possibilities of the future’ (p. 797). Harris noted that, since research into the transformational basis of language suggests that all languages are cut from the same basic cloth, ‘a single world language. . .would not represent a great loss for human culture’ (p. 798): this seems a little perverse, in that the makers of constructed languages have never meant their creations to oust ‘natural’ varieties anyway. Chomsky (Reference Chomsky1977) himself suggested that constructed languages are essentially games or puzzles; they stand apart from real varieties because they are not cut from the cloth of the ‘universal grammar’ that – in Chomsky's view – underpins all natural languages. We learn the latter without formal instruction, simply through immersion from infancy, but, says Chomsky, the learning of constructed languages is a more self-conscious task. (It would be interesting to know Chomsky's opinion about those infrequent situations in which children are brought up with Esperanto as a first language.)

It is at least an interesting aside to note that – as Yaguello (Reference Yaguello1984) reminds us – every constructed-language scheme has been the product of the male mind; she sees ‘logophilia’ as a ‘delirium of naming. . .a taxonomic madness’, something that only infects men. Yaguello clearly believes that an urge for completeness, for unity, for what Eco subsequently called ‘la lingua perfetta’, is a male disorder. Given the religious underpinnings of many constructed varieties, this quest for an ideal language may be likened to a quest for some pre-Babel paradise. And perhaps, too, this idée fixe may be related to the urge for collecting things – for the psychological literature (e.g. Muensterberger Reference Muensterberger1994) suggests that this generally springs from a desire for completeness: think of the philatelist driven to round out his collection of the Canadian Small Queens, or the bibliophile obsessed with every edition and every state of Burns's poems. Yaguello notes, early in her fascinating but very uneven treatment, that the making of languages is a curious business: grandiose schemes, huge amounts of invested energy, meagre results.

While Esperanto is far and away the most successful of all the constructed languages, success is a relative quantity. Indeed, in comparison with existing ‘natural’ lingua francas – French, German and English, to name only those important during Esperanto's life-span – it has been a dismal failure. Furthermore, the future looks quite bleak, since every indication is that English will continue to expand around the world, displacing ‘smaller’ varieties and certainly leaving very few toe-holds for any constructed contender. In this sense, the lack of large-scale success of constructed languages is not unlike the failure to revive flagging minority languages. That is, the motivation necessary for resuscitation – or, in the case of constructed languages, for widespread adoption and use – is hard to elicit in societies in which other powerful languages already hold sway, and particularly when these other languages are, like English, contenders for global lingua franca status. It may well be that, as Ó Riain (Reference Ó Riain2003) has pointed out, Esperanto is the only constructed variety to have emerged as a fully functional and utilitarian language, but this sort of formality obviously counts for very little in the real world. In the same way, proponents may argue all they like about the unfairness inherent in natural-language dominance, about the regularity and ease of learning of a constructed form, about all the many advantages, in fact – none of this matters very much, it seems, in the world as it is.

None of this means, however, that constructed a posteriori languages like Esperanto are unworthy of serious attention. As Lieberman (Reference Lieberman1979: 100) noted, researchers in sociolinguistics and the sociology of language have largely ignored an intriguing area of study, one with its ‘facts, texts and living subjects readily available’. Michéa (Reference Michéa1983: 30) adds that ‘la communauté espérantiste offre en miniature un modèle intéressant de ce qu'est un ensemble social’ [the Esperanto community is an interesting model in miniature of a social group] – which suggests, I think, that much of the research interest might lie under the rubrics of sociology, of political science, of psychology. Thus, Fettes (Reference Fettes1996: 57) writes: ‘here is a field of study which could provide exciting insights into the relationships between [sic] language, culture, ethnicity and identity – exciting precisely because the case is atypical, and therefore a test for theories derived from studies on other, more familiar social groups.’ In a rather neat adaptation of Terence's famous observation, Calvet (Reference Calvet1983: 35) simply said ‘je suis linguiste, tout ce que touche aux langues m'intéresse’ [I am a linguist, and everything about language is of interest to me] – surely this is the right attitude. Nonetheless, apart from the work undertaken by Esperantists themselves, the situation has hardly changed since Lieberman's observations (Forster's study being a notable exception). The whole area – constructed languages, their developers, and their speakers – has largely been prejudged and dismissed.

As noted, there has been little scholarly interest in constructed (or ‘artificial’, ‘planned’ or ‘international’) languages – linguists have tended to see them as well-intended but pointless anomalies, and their proponents as naïve and often crankish enthusiasts. But there have been some notable exceptions. The great Danish grammarian and Anglicist Otto Jespersen was intrigued by the idea of a universal and easily learned auxiliary language, was involved in the development of Ido (an Esperanto offshoot), and eventually produced his own language, which he called Novial. It is noteworthy that Jespersen devoted two chapters of his autobiography to international languages: the first describes his early involvement with the politics of Esperanto and Ido, up to the time of the First World War; the second deals with the inter-war years, and Jespersen's attention to Novial. It is also noteworthy that, despite thinking and writing about constructed languages for seven decades, Jespersen says that his interests here

I apologise for the lengthy quotation, and for more details about Jespersen's work than might at first seem to be necessary. But we have here, in microcosm as it were, something of the tensions that have always surrounded constructed languages. On the one hand, there is an obvious appeal to the idea of promoting a simple auxiliary variety that could be a universal second language – logical, functional, no threat to the maternal tongue: a dependable maid-of-all-work who never interferes above stairs. It is hard to see how the global sharing of a common medium could be anything but beneficial. On the other hand, though, there is – as Jespersen states – a sense of wasted time, a feeling that any investment in an auxiliary language represents a lapse of rationality.