Europe’s Linguistic Challenge[1]

Philippe Van Parijs

(in European Journal of Sociology 45 (1), 2004, pp. 111-152.)

 

 

 

Along with much of the world, Europe is converging rapidly towards adopting English as its sole lingua franca. This process is, in a sense to be clarified, inevitable. It is also, in a sense to be qualified, desirable. But it raises a number of serious questions, at least four of which can be formulated as issues of fairness. In this article, I shall consider these in turn, and indicate what I believe to be the best way of handling each of them. But before doing so, I need to spell out the nature of the fundamental mechanism that underlies the core of the dynamics of secondary language learning and multilingual interaction, not only in today’s Europe, but throughout history and throughout the world.

 

1. Why we are moving towards one lingua franca

Most of the trends in linguistic competence we can observe can be understood as the product of the (sometimes explosive) interaction of two micro-mechanisms.

 

1.1. Probability-sensitive learning

What I shall call probability-sensitive learning simply captures the following two-dimensional fact. The extent to which people maintain and improve their linguistic competence in some particular language is strongly affected by the probability with which they can expect to have to function in that language. This is in the first place a matter of motivation: the more likely it is that competence in a particular language will be useful to communication, the greater the effort one will decide, individually or institutionally, to invest in learning it. But this differential learning is also in the second place a matter of opportunity: the more often one finds oneself in a context in which a particular language is actually being used, the smaller the effort required to learn it. Moreover, these two dimensions of the mechanism feed into each other: the motivation easily induces the creation of more opportunities than those that offer themselves spontaneously, and the actual enjoyment of the opportunities (over and above the expectation of further opportunities) may nurture the motivation to learn by enabling people to experience what difference it makes to possess the linguistic competence required to understand what is being said and to take an active part in the conversation.

I am of course not denying that other factors — for example, how aesthetically attractive one finds the language one considers learning, or how close it is to one’s native tongue, or how large a literary corpus it give access to — may affect significantly either the motivation to learn a language or the ease with which one will learn it or both. All I am asserting is that, through these two channels of motivation and opportunity to learn (and retain), the probability of having to interact in a particular language will be a massively important determinant of the extent to which average competence in a particular non-native language tends to expand or shrink in a particular population. A greater probability means both a larger expected benefit from any given level of linguistic proficiency in the language concerned and a lower cost of acquiring or preserving it. More sophisticated measures of the communicative value of a language have been offered, for example Abram de Swaan’s (2001) stimulating notion of Q-value. But none of them offers the promise of improving much, if at all, upon the simple probability of interaction in that language, as a predictor of differential learning and retention. (See Appendix 1.)

 

1.2. Maximin communication

What I shall call the maximin law of communication captures a distinct, somewhat less obvious but hardly less general mechanism, which can be sketched as follows. Suppose you have to address simultaneously a set of people who each know to various extents a number of languages and by all of whom you want to be understood. When deciding which language among those you know you should pick, the question you will spontaneously tend to ask yourself will not be which is your own best language, or which language is the best language of the majority, or which language is best known on average by your audience, but rather which language is best known by the member of your audience who knows it least. In other words, you will systematically tend to ask yourself whether there is any language that is known to some extent by all. If, to the best of your knowledge, there is one and only one, you will choose it. If there is none, you will tend to choose the language which is known to some extent by most. And if there is more than one, you will make a guess for each of them about the level of competence achieved by the person least competent in it, and you will choose the language for which this level of competence is highest.

This “maximin” criterion amounts to maximizing the minimum competence. It can also be described as a criterion of minimal exclusion. It has a number of direct corollaries, such as the systematic victory, in linguistically mixed marriages, of the language of the “worst linguist”, i.e. of the partner who knows least well the language of the other and systematically tends to be the speaker of the more widely spread of the two languages.

Again, I am not claiming that this maximin law operates without exception. To start with, deviation from it happens on a massive scale for pedagogical reasons. In foreign language classes, for example, teachers often know the mother tongue of their pupils (which may well be their own) far better that the pupils know the language they are learning, but the mutually accepted rules of the teaching game will frequently entail the partial or total banning of the maximin language. For analogous reasons, some people choose (as I did) to speak their mother tongues with their children, even though their children have been all along and they have themselves become significantly more fluent in at least one other shared language.

On a less massive scale but often in a highly sensitive way, deviation may also occur, even in informal contexts, for what could be called expressive reasons. This may happen in negative fashion, for example when post-1989 East-Europeans struggle to communicate with one another in English, even when it would be (linguistically) far easier for them to do so in Russian. It may also happen in positive fashion. For example, it happened to me more than once to address a Brussels audience in Dutch rather than French, despite the fact that all would have understood me adequately in French, while some did not in Dutch, for example because the organisers felt that some fair time sharing between French and Dutch needed to be kept. At least in part for an analogous reason, a vade mecum dispatched by the French foreign ministry insistently instructs France’s representatives in all European institutions that, even at informal meetings or after the interpreters have gone home, “les Français parlent leur langue” (Ministère des affaires étrangères 2002).[2] Providing the number of mother tongues involved does not exceed two or three, this sort of consideration may also lead one to operate, as often done at interpretation-free federal-level meetings in Belgium and Switzerland, according to the rule “Each speaks his/her own language”.[3]

However, as soon as efficiency in communication prevails over pedagogical or expressive concerns, perceptible inequalities in the minimum knowledge of the various languages involved will generate a hardly resistible pressure for all to adopt the maximin language: What’s the point of uttering beautiful sentences with carefully chosen words if my audience would understand me far better were I to express myself more clumsily in a language far more familiar to them. Hence, although didactic effectiveness and symbolic impact may sometimes strongly constrain language choice, this will not prevent the maximin criterion from running the show whenever communication is the prime concern, i.e. in the bulk of spoken and written language use.

 

1.3. An explosive interaction

Needless to say, these two mechanisms powerfully interact with each other. The more a particular language is being learned in some section of the world population, the more likely that language is to be the maximin language in contexts of interaction involving members of that section of the population. And the more often a particular language is picked as the language of interaction, the stronger the motivation for learning it and the more frequent the opportunity to learn it. It is worth noting that this positive feedback loop would also exist if the speech partners systematically tended to pick the language for which the average knowledge is greater (call it the maxi-mean language), or even the best language of the majority, but it would then operate at a considerably slower pace.

To illustrate this difference, take the situation that used to prevail before the Swedes and the Finns joined the EU. Both the maxi-mean and the maximin language in contexts of informal interaction between multilinguals within and around the European institutions then tended to be English and French in varying proportions (with German far more often maxi-mean than maximin). Given how small a percentage of the total population of speech partners they represented, the arrival of the Scandinavians did not change much in terms of maxi-mean. But it made a big difference in terms of maximin. For while the second best language for most British and Irish people was and is French, the Scandinavians’ average competence in French was far poorer, and therefore tended to make English a clear winner in terms of maximin (though only marginally better than before in terms of maxi-mean) in any context in which they turned up. It is obviously far easier for newcomers to upset the prevailing choice of a language under maximin than under maxi-mean: it suffices for them to be about totally ignorant of the prevailing language, while everyone else knows at least some more of at least one of the languages they know better. And once the switch is done, language learning is accordingly redirected for both incentive and opportunity reasons, leading further contexts to do the switch, and so on.

Undoubtedly, this analysis is very rough. Its basic assumptions need to be qualified and its implications should be modelled out in detail to provide precise answers to questions such as the following. Under what conditions does this twofold mechanism lead to a convergence to a single lingua franca? Under what conditions is it on the contrary consistent with the lasting coexistence of two or more linking languages? Under what conditions does it imply the decline of multilingualism (as opposed to bilingualism), and indeed of any bilingualism that does not consist in combining a mother tongue and the lingua franca? Under what conditions is it consistent with stable universal diglossia — competence of all members of a community in both their mother tongue and the lingua franca — or does it imply a long-term threat to the very survival of linguistic diversity? I do not know the answers to these questions (and would be interested in finding out). But I have been, from the day of my birth, a participant observer in enough thousands of situations of multilingual interaction, and seen enough figures, tables and graphs depicting existing trends, to feel confident about the nature and power of the twofold mechanism outlined above.

 

2. Why we need one lingua franca

Whatever the power of the mechanism just described, one may want to pause to ask whether we really need the one lingua franca which it tends to bring about. Given the values to which I am committed, there is not the slightest doubt in my mind that we need a way of communicating directly and intensively across the borders drawn by the differences of our mother tongues, without the extremely expensive and constraining mediation of competent interpreters. We need it in particular if we do not want Europeanisation, and beyond it globalisation, to be the exclusive preserve of the wealthy and the powerful who can afford quality interpretation. If we want all sorts of workers’, women’s, young people’s, old people’s, sick people’s, poor people’s associations to organise on the ever higher scale required for effective action, we must equip them with the means of talking to one another without interpreting boxes and highly skilled and paid professionals in them. One way of putting this is by saying that we need to meet the linguistic preconditions for turning Europe, and ultimately the world, into one demos, without this needing to mean that Europe, or the world, is thereby turned into a single ethnos: a forum can be shared thanks to a common language without the culture, including the language, becoming one. It cannot be taken for granted, however, that the linguistic preconditions for the existence of a single demos involve the adoption of one lingua franca.

 

2.1. Clever softwares?

Firstly, one can try to imagine a situation in which technological development will have made informal communication possible between different language groups without requiring the learning of a single common language. It suffices to combine the best of voice recognition and translation softwares to convey instantaneously through earphones in any chosen language what is being said in any other. Both kinds of software, we are told, are making fast progress. But those who have experienced some of the oddities generated by translation software even when having to cope with only slightly casual style, and by voice recognition software even under favourable acoustic conditions, can imagine how stilted and contrived a spontaneous interchange would need to become in order for its participants to feel reasonably confident of being understood. [4]

Moreover, any interacting group soon develops a small culture of its own, with words being used between inverted commas, as it were, or proper names turned into nouns, or short-lived imports from another language. Even very imperfect mastery of a common language would provide for a far better medium than beautiful mastery of one’s own language constantly threatened by ridiculous stiffness on one side (if one bears the technology in mind) and ridiculous misunderstandings on the other (if one does not). Techno-freaks can keep dreaming about it. But there is no salvation to be hoped from these quarters in my view.

 

2.2. Esperanto?

If technology does not enable us to dispense with a common language, why not opt for a neutral one? This second solution is less fanciful. It is vigorously defended on grounds of neutrality and simplicity. But these two advantages look far greater than they really are. And in addition, the spreading of Esperanto faces a prohibitive hurdle.

Take neutrality first. Esperanto is of course far from being neutral in the sense of equidistant from all existing languages. It belongs unambiguously to the Western group of Indo-European languages, with identifiable Latin, Slavic and Germanic ingredients. Even within Europe with Finnish, Estonian, Hungarian, Basque and Maltese as part of the picture, it cannot make any claim to “neutrality”. Moreover, it does not stand a chance as a European lingua franca if in addition English needs to be learned as a worldwide lingua franca. But when offered on a world scale, it must lose all hopes of being sold on grounds of fairness, especially to the millions in India, Nigeria, South Africa or even now Japan and China, who have already invested massively in the learning of another Western language and will understandably show little patience for this new Eurocentric gimmick which they are enthusiastically invited to absorb.

It is true that, unlike English, Esperanto would be a learned language for everyone, and hence its adoption as a lingua franca would definitely be more egalitarian than that of English or French. This advantage, however, would only be transitional. Just assume that Esperanto successfully spreads and starts being used in a growing number of contexts, including by mixed couples in the upbringing of their children. Nothing would then prevent it, after some generations, from thickening from a lingua franca into the mother tongue of some — as happened to Swahili, for example —, with the consequence that once again neutrality would be lost and the whole process of designing a neutral language, in the modest sense of being the mother tongue of no one, would need to be relaunched from scratch.

Secondly, consider the ease of learning. Syntactic and morphological rules are exceptionless in Esperanto, and therefore undoubtedly far simpler to learn from a grammar book than are those of natural languages. Moreover, compared to languages like English and French whose spelling was established long ago and very conservatively managed, Esperanto offers a sensible relationship between oral and written forms. However (usually shorter) irregular forms are rarely, if ever, sheer irrational nuisance. Natural languages operate complex trade offs between least effort in memorising and least effort in pronouncing, and the more intensively a language is used, the more the latter matters relative to the former. Esperanto turned into a really living language would soon be subjected to such pressures. Moreover, Esperanto made relevant for all contexts would need to beef up its lexical stock massively through imports from other languages. Like all other languages today, it would import massively from English, and probably more than others because of its smaller initial stock. Hence, it would not take long for it to start looking like a bulky language, with a slim Esperanto component that can be learned in a couple of days and a huge English vocabulary with subtle nuances which it could take years to master. Most fundamentally, however, the notion that Esperanto unambiguously has the upper hand over English in terms of learning rests on a very schoolish picture of language learning. As pointed out earlier, the learning of a language is essentially a matter of having the opportunity to play, whisper and quarrel, listen to music, watch TV and scan the web in that language, and a matter of being motivated to do all these things, especially at an early stage in one’s life. If this more realistic picture of large-scale language learning replaces that of enthusiastic but lonely xenophiles confined to grammar books and vocabulary lists, it is no longer that clear that Esperanto is, in the relevant sense, easier to learn.

With its claims to neutrality and simplicity thus drastically qualified, Esperanto is hardly in a strong position to face the formidable hurdle it faces, precisely as long as it is not the mother tongue of a significant group. Investing in the learning of such a language is definitely cheapened by the exceptional simplicity of its morphological and syntactic rules. But as long as speech partners, films, music and TV broadcasts in that language are not all over the place, it still comes at a significant cost for someone with average learning skills. In the case of widespread natural languages, there is a secure minimum return to the learning investment, thanks to the tens or even hundreds of millions of people with whom one can be sure one acquires the capacity to communicate. Even in the case of Esperanto, the most widespread among the artificial languages currently advocated, this minimum return is not guaranteed, as all depends on whether a sufficient number of people will be willing to make and keep making the deliberate effort of learning the language, which is itself dependent on which language learning choices they expect others to do.[5]

The size of this handicap, relative to English, keeps growing as English keeps expanding in terms of native speakers, mainly thanks to continued net migration into English-speaking countries, and, at a much faster rate, in terms of the total number of people competent in it: English is probably the only natural language today, and certainly the only major language, with (far) less native speakers than people who learned it as a second language.[6] Given the twofold mechanism sketched at the start — probability-sensitive learning and maximin —, dislodging it from this position will become an ever more impossible task. Esperanto is a wonderful way of linking up a fantastic bunch of generous and hospitable people around the world, but it is no more hopeful than clever software as an alternative Europe-wide or world-wide medium of communication.

 

2.3. Lingua franca pluralism?

Having granted that we need a natural language, perhaps we should not rush into asserting that we need only one. To avoid the drawbacks and dangers of the dominance of a single language, many (especially, but not exclusively, among those whose language stands a chance of being picked as one of the lingua francas in the event that the proposed formula were implemented) have proposed that there should be two or three lingua francas side by side, with an identical status.

A first way of understanding this lingua franca pluralism consists in viewing the selected lingua francas as alternatives to one another: each person learns one of them and only one. But reflection on some very modest arithmetic exercises should suffice to make us quickly discard this version of the idea. In a population of six people with three distinct mother tongues, a balanced choice between two lingua francas provides a common medium to randomly grouped people far more frequently than a random choice of a second language. It does so, however, with a decreasing frequency as the size of the grouping rises. Moreover, this frequency decreases sharply as the number of distinct mother tongues increases. By contrast, the learning by all of the same lingua franca provides a common medium in 100% of the cases, whatever the size of the groupings and whatever the number of mother tongues. Moreover this can be achieved with a sizeable discount, as those whose mother tongue is being learned can be exempted from the learning of any second language. (See Appendix 2.)

There is, however, another interpretation of lingua franca pluralism which performs just as well as the single-lingua-franca option in ensuring inter-communication in all groupings. It consists in viewing the two or more languages granted lingua franca status not as alternatives but as supplements. In other words, the rule is no longer that each individual is supposed to learn one of the lingua francas, but that he must know them all. This may sound at first sight like a wasteful overkill: two or three times more learning without any gain in inter-communication, as one lingua franca is sufficient. It is, however, reasonable to conjecture that the passive knowledge of a language is easier to acquire (and even easier to be believed to be acquired) than an active knowledge of it. Therefore, the cost reduction that flows from the fact that natives of a language promoted to lingua franca status no longer need to acquire an active knowledge of the original lingua franca must be matched against the cost increase flowing from the fact that others need to acquire a passive knowledge of that language. In Belgium or Canada, therefore, quite apart from being perceived as fairer (an issue to which I return shortly), an “each his own language” regime may compete reasonably well, in terms of cost-effectiveness, with a single-lingua-franca regime.

The trouble comes again when the number of native languages increases. In the European context, French is keen to share the lingua franca status with English. But this would mean that Germans, for example, still need to acquire an active knowledge of either French or English, while in addition having to acquire a passive knowledge of the other. As they form the largest native language group, they find this understandably hard to accept. If francophones are to have a chance of winning their case, they therefore realise that they need to broaden their alliance by proposing to further share this lingua franca status with German. [7] But how will the Spanish, the Italians, and all the rest feel? Making life more comfortable for the Germans and the French by exempting them from acquiring an active knowledge of English makes things considerably worse for all others, now forced to acquire a passive knowledge of two more languages without being exempted from acquiring an active knowledge of one of them.

What may make sense in the presence of two native tongues, possibly even three, does not make the slightest sense when there are many. Any attempt to press for the adoption of one’s native tongue as a second lingua franca in this supplementary sense will immediately be perceived for what it is: trying to get greater comfort for oneself at the expense of increasing the burden on those who enjoy the privilege of having had their native tongue picked as the only lingua franca so far — which is defensible enough —, but also on all other language groups who are not better situated than oneself — which is indefensible. Any attempt to assuage some of these by offering to extend the lingua franca status to them at the same time risks turning the net benefit to oneself into a net cost, unavoidably further increases the burden on any group still left out, and further boosts the global cost of the whole scheme.

Hence, for quite a different reason, lingua franca pluralism is no more promising in the supplementary sense than in the alternative sense. Whatever the language historical fate happens to have picked, we definitely need convergence to a single lingua franca. Those saddened by the fact that it is not the one they learned as infants will have to come to terms with it. Their narcissism should not jeopardise the satisfaction of our urgent communicative needs, in Europe and in the world.

 

3. Unfairness as unequal access to linguistic advantage

The twofold mechanism sketched in section 1 and the feedback loop between its two components enable us to understand what is now leading to the dominance of English. No hidden conspiracy by the Brits, let alone the Americans, but the spontaneous outcome of a huge set of decentralised decisions, mainly by non-anglophones, about which language to learn and which language to use. Next, our exploration of imaginable alternatives in section 2 led to the conclusion that the increasing dominance of one natural language as the single lingua franca simply makes a lot of sense: to communicate with one another, we need one and only one idiom, and it will need to be a natural language. Both unavoidable and wise then? Undoubtedly. Fair too? Certainly not. I now turn to four possible characterisations of the nature of the unfairness involved and briefly indicate in each case how I believe it is best to respond to the problem, as characterised.

 

3.1. Undeserved linguistic rents

First, convergence towards a lingua franca that is the mother tongue of a subset of the population concerned unavoidably provides the members of this subset with undeserved advantages over the rest. They can express themselves with more ease and therefore tend to be more active and more persuasive in discussions conducted in the lingua franca, whether of a business, political or social nature. In addition, some jobs restricted to native speakers of the lingua franca — such as a far more than proportional number of language teachers and language editors paid for by non-natives, a more than proportional number of translators and interpreters into that language paid for by international organisations — would not exist without the privileged status enjoyed by that language. Moreover, a large number of jobs that are not specifically linguistic in content are explicitly or implicitly restricted, or far more easily accessible, to native speakers of that language, because of the central importance of being able to communicate in that language.[8] This booming demand for people proficient in the lingua franca unavoidably tends to boost the relative pay of people with native competence in that language, whether, for example, through being able to ask for high fees for private language tuition or through faster promotion in inter- or supranational organisations.

The undeserved inequalities thus created are by no means restricted to inequalities between natives and non-natives of the lingua franca. Among non-natives, there are also huge and increasingly consequential social inequalities in terms of the extent to which the family environment provides children and adolescents with both the opportunity and the motivation to learn the lingua franca. There is a big difference between children whose parents have both a rich set of foreign connections with whom the lingua franca is commonly spoken and a purse large enough to fund Summer courses in Oxford, and those who have never taken part in any English conversation and whose parents would not know how to start to give them the chance of however modest an immersion.

 

3.2. Stepping back

Thus, growing unfairness there appears to be. Before considering what can and must be done about it, it is worth pausing briefly to get a sense of perspective.

First, the problem thus characterised is far from being unprecedented. In most nation states, the majority of the population had a mother tongue, typically labelled a “dialect”, that differed notably from the national language, as used in the media and the educational system, in high culture and political life, and in business transactions beyond the local level. Indeed, in many places, there is still a big discrepancy between the home language of many families, especially rural ones, and the nationally imposed lingua franca. In most cases, linguistic distance was not as great as between most European languages and English, but in some cases it was, and in all cases it involved forms of disadvantage in economic and political life, often even forms of blatant discrimination, closely parallel to those now encountered as English becomes just as much of a trans-national must as the dialect of the capital was a national must.

In the national context, the task of drastically reducing the resulting inequality of opportunities was (regarded as) accomplished through compulsory schooling in the national language. In the case of a trans-national lingua franca that no one would dare to try to impose as the main language of the various national populations concerned (and rightly so, as I shall argue later), the job looks far trickier. But let us bear in mind that the average number of years spent at school and the resources devoted to education in today’s European context are enormous in both a historical and a comparative perspective. For example, when we are demanding that a country like the Congo, whose educational system is in shambles and whose formal political life is entirely conducted in an alien language mastered by no more than an estimated 7% of the population, should operate democratically, and hence at the very least enable a majority of the citizens to more or less follow what is going on, we are demanding something incomparably more utopian, in terms of its linguistic preconditions, than universal competence in English throughout Europe.

Moreover, as reflected in recent surveys, the process is well under way. In Belgium, for example, in which there are two national languages on the same footing, average competence in English for the younger cohorts of adults is considerably higher than average competence in the second national language has ever been in the history of the country.[9] It is true, however, that even in these younger cohorts it remains a minority feature, and on average still a very long distance from the competence of native speakers. But there is one simple and cheap measure which, if taken vigorously throughout Europe, can be expected to have a dramatic impact both in reducing this distance and in spreading competence in English in all layers of the population.

 

3.3. Ban dubbing!

To see what this could be, just reflect for a while on the distribution of competence in English across European countries, as revealed by Eurobarometer (see Table 3). If we leave out the UK and Ireland because they are essentially anglophone, and Belgium and Luxemburg because they are multilingual, we are left with eleven countries, five with a Germanic language, four with a Latin language and two others. To no one’s surprise, the five Germanic countries score better in terms of self-assessed knowledge of English (with an unweighted average of 65%), than the four Latin countries (with an unweighted average of 38%). This seems to provide strong support for the common wisdom that this sizeable inequality is rooted in the fact that English is an (admittedly quite latinised) Germanic language, and hence intrinsically easier to learn for the average citizen of the former set of countries than for the average citizen of the latter.

 

Table 3. Percentage who say they “know” English (EU 15)

 

Age group                     15 or more 15 to 39

United Kingdom              99                        98              

Ireland                            95                        96

Sweden                          79                        94

Denmark                        75                        84

Netherlands                    70                        80

Finland                           61                        87

Germany                        54                        74

Greece                           47                        71

Austria                           46                        66

France                            42                        63

Belgium                          40                        55

Italy                               39                        63

Spain                              36                        61

Portugal                         35                        59

Luxemburg                     19                        27

Source: Eurobarometer 54.[10]

 

There is however, a second conjecture that turns out to be far more consistent with the data as soon as some attention is paid to the two remaining countries. Greek and even more Finnish are uncontroversially far more remote from English than either the Latin or the Germanic languages. As one moves from Finnish to Greek and next to the Latin and the Germanic group, one would therefore expect competence in English to rise monotonically. Yet for the population as a whole, the profile yielded by the data is 61% for Finland, 47% for Greece, 38% for the average Latin country and 65% for the average Germanic country. Even worse, for the younger generation (under 40), we find 87% for Finland, 71% for Greece, 61.5% for the average Latin country and 79.5% for the average Germanic country. On closer inspection, therefore, linguistic distance looks like a very bad — and worsening — predictor of competence in English.

To find a better predictor, let us partition our eleven countries according to the number of native speakers of their official language worldwide: less than 10 million (Denmark, Finland, Sweden), between 10 and 50 million (Greece, the Netherlands), between 50 and 100 million (Italy, France, Germany, Austria) and over 100 million (Portugal, Spain). The average proportion of people who say they know English now drops quite sharply and consistently from one category to the next: 72%, 58.5%, 45% and 35.5%, respectively, for all age groups together; 88%, 75.5%, 66.5% and 60%, respectively, for those under 40. Why?

My conjecture is that the key intermediate variable is the relative frequency of dubbing versus subtitling in the broadcasting of English-language series, films and other programmes. It is estimated that the average cost of one hour of dubbing is about fifteen times the cost of one hour of subtitling (Luyken & al. 1991). Hence, the threshold, in terms of number of viewers, as from which it starts making sense to incur the cost of translation is far lower in the case of subtitling that it is in the case of dubbing, which a majority of viewers seems to prefer.[11] Consequently, the extent to which English-language productions are dubbed, rather than subtitled, can safely be expected to rise steadily as one moves from countries whose language is spoken by comparatively few people, such as Denmark, Sweden, Finland, Greece and the Netherlands, to countries populated by the members of larger linguistic groups.[12] As revealed by incipient research, the watching of undubbed foreign programmes provides, under appropriate conditions, a powerful way in which children learn foreign languages. [13] No wonder, therefore, that we should find a strong negative correlation between size of the language group and competence in English.

In order to motivate the proposal I am about to make, I do not need to assert that no other factor plays a significant role. It is sufficient for me to be able to assume, as the available evidence strongly suggests I can, that the learning mechanism in question is a powerful one. We cannot do much about linguistic distance between languages, nor about the numbers of native speakers of the various languages, nor therefore about the relative profitability of subtitling and dubbing. But we can outlaw dubbing. And if we do so, while providing supportive language teaching and letting MTV music, web chats and other less virtual trans-national contacts do the rest of the job, competence in English will become all over, in the space of one generation, even less of a problem than it now is in the most English-literate parts of the European continent. [14]

Refusing to ban dubbing in those countries in which it is currently common practice amounts to unnecessarily inflicting a linguistic handicap to the most disadvantaged layers of the populations concerned and therefore strengthening the privilege enjoyed by the elite whose access to competence in English is far easier through quality schooling and foreign contacts. It also amounts to perpetuating an increasingly costly disadvantage for many members of their populations who are any likely to be involved in the global economy, in supra-national organisations or in the trans-national civil society. If we want to be serious about fighting linguistic injustice in the sense of unequal access to linguistic advantage, therefore, my recipe is simple and inexpensive: Ban dubbing![15]

 

3.4. Three objections

One possible objection is that such a ban would violate the fundamental freedom of expression. Note, however, that it applies indiscriminately to all languages, that it involves no restriction whatever on the content of what is being subtitled or dubbed, and that it does not deprive anyone from addressing directly through the media an audience with whom it has no language in common.[16] This objection is therefore bound to be regarded as ludicrously formalistic, especially if the ban it incriminates can persuasively be shown, along the lines sketched above, to better equip a large proportion of the population to express themselves in a language in which it will be increasingly crucial for them to be able to express themselves if they are to be heard by those by whom they will need to be heard.

A second objection is that the ban would hurt head on the interests of professionial actors, who use dubbing as a way of securing more regular incomes than film or theatre contracts can provide. There will undoubtedly be an effect of this kind, but it will be buffered, if not offset, by a significant increase in the demand for local production if it remains the case, beyond the transition period, that a majority of people prefer dubbing to subtitling. A residual net negative effect on professional actors taken as a whole cannot be ruled out. But the vested interest of a tiny minority cannot legitimately block a move that would massively benefit a large, comparatively disadvantaged majority.

Finally, there is the risk that the ban would be by-passed as a result of people going to the cinema or watching videos and DVDs not subjected to the same ban. While the ban seems easy enough to extend to cinemas, it seems more difficult for videos and DVDs. But the fall in demand from TV channels and cinema distribution may in itself be sufficient to make quality dubbing unprofitable for videos and DVDs alone, even if some boosting of the demand for dubbed videos is triggered by the ban. Needless to say, if the effect of the ban were that people would shift entirely to programmes in the native language, or to cartoons with a sound track in the native language, or to dubbed videos, or to a combination of these, the intended effect would not be achieved. But although some shift in each of these three directions can be expected, it is most doubtful that it would inhibit a lasting and expanding impact, especially as tolerance for subtitling develops through practice and as teletext technology makes it possible to open a wide range of individual choices for subtitle languages and to optionally get rid of subtitling altogether as competence in English (or any other non-native language) makes it superfluous for a growing number of non-native people and for a widening range of programmes.

 

 

4. Unfairness as the unequal sharing of the burden of language learning

4.1. Free riding

To phrase as sharply as possible our second problem of linguistic injustice, let us next assume that competence in English has spread massively to the non-native speakers, and pretty equally among them, so that the adoption of English as a lingua franca no longer gives a great advantage in discussion or competition to English natives or to non-natives with a privileged access to English. Language-related injustice has not disappeared. For unlike the community of natives, the non-natives have had to spend a considerable amount of time and resources to the learning of a foreign language.[17] It is estimated that the average time required to master adequately a foreign language is 10.000 hours — compared to a standard school year totalling less than 1.000 hours in the classroom.[18] However speculative such estimates, it is clear that the cost in time and resources of acquiring proficiency in a foreign language is huge. This heavy effort obviously benefits the community that performs it ­— otherwise it would not bother —, but also, in some cases possibly even to a larger extent, the community whose language is being learned.

In other words, there is a public good— the creation of a lingua franca — being enjoyed by all linguistic groups throughout the world involved in global communication, but produced only by those groups whose language has not been picked as the lingua franca, with the lucky ones whose native language happens to have been picked enjoying a free ride. This is the second sense in which “linguistic injustice” can be said to be involved. What can be done about it?

 

4.2. Cost sharing

Scandinavians, who speak the least widely spread of the EU’s current eleven official languages have tended to be pretty blunt about admitting the dominance of English, while being quite imaginative about suggesting how the induced fairness could be reduced. Thus, the first Danish delegation to the European Parliament is said to have made the following proposal. They conceded upfront that they could not expect others to understand Danish and agreed to speak English, but only on condition that the others, including the French, did the same with the sole exception of the British, who would have to speak French. The British were no doubt quick to point out that this would be grossly unfair to them as they would be forced to express themselves in a language only a minority would grasp, while everyone else could be understood by all. As they were still allowed to speak in English at that stage, the others must have understood they had a point, and the idea was dropped.[19]

More recently, the Swedish Prime Minister, somewhat scared at the prospect of a near doubling of the number of official EU languages, made a distinct proposal. Instead of having all countries paying jointly (roughly according to their wealth) for the translating of everything into every language, why not have a system in which the cost of language services would be systematically shared equally between the countries whose language is being used and the countries into whose language the translation is being made. As an ever greater majority of texts is being produced in English and as the Swedes are competent enough in English not to need a translation for most documents, the rule would end up practically exempting the Swedes from any contribution. Fairly, it might be said, as this counts as a compensation for their investment in the learning of English. Efficiently too, it may be added, as this would provide other countries with an incentive to follow suit, thereby facilitating massive savings in translation costs. At the limit, all translation costs would be eliminated as a result of all countries conforming to the Swedish pattern. However, while translation costs may then be down to zero, unfairness would not, as one linguistic community would still get away with not learning a foreign language, while all others would need to.

 

4.3. Proportionality between cost and benefit

As argued persuasively by Jonathan Pool (1991), the only real solution to this problem, the only real way of reconciling communicative efficiency and linguistic fairness in this second sense, consists in introducing a subsidy from the linguistic group whose language is being learned to those who do the learning. How high should this subsidy be? Various criteria are worth discussing. For example, Pool (1991) proposes that each language group should contribute to the cost of the learning of the lingua franca according to its numerical size, while David Gauthier’s (1986) general conception of co-operative justice as maximin relative benefit would amount, in this case, to requiring equality between the per capita benefits derived from the existence of the lingua franca by the various language groups.[20]

I argue elsewhere (Van Parijs 2002) that neither of these prima facie attractive criteria is defensible and that a distinct one is to be preferred: the equalisation of cost-benefit ratios across language groups or, put differently, proportionality between (total or per capita) contribution of each group to the cost of the existence of the lingua franca and (total or per capita) benefit derived from it. Suppose we measure roughly the benefit to a language group of the existence of the lingua franca by the number of people with whom the latter enables members of the group to communicate. And suppose we measure the contribution of a language group by the amount (in money and time) its members spend on acquiring the lingua franca, if any, plus the taxes paid in this connection to other language groups, if any, minus the subsidies received in this connection from other language groups, if any. What the proposed criterion requires is that, across all language groups involved, the total contribution (taking taxes and subsidies into account) be proportional to the number of people the language group can communicate with thanks to the lingua franca.

In all circumstances, this criterion will require a net transfer from the linguistic group whose language is being learned to the groups who do the learning, and the per capita size of this language tax will grow, other things being equal, as more and more people learn the language. The size of the transfer will never exceed the benefit to lingua franca natives, as the criterion requires the ratio of contribution to benefit to be the same for all, and the learning only makes sense if the cost exceeds the benefit. Yet, it is clear that the criterion justifies massive transfers from those countries in which the bulk of the English natives live — in particular the United States, home to 70% of them— towards the rest of the world.

 

4.4. Four qualifications

This conclusion needs to be qualified in four ways. Firstly, as the biggest language groups in the world — the Mandarinophones and the Hispanophones — fully join the global game, the English natives will not be the only ones from whom fairness will require a subsidy. For as one moves from a larger to smaller language group that learns the lingua franca, it is not just that the total amount of the subsidy justified by our criterion shrinks, but also its per capita level, because smaller groups unavoidably gain more speech partners than large ones thanks to the lingua franca. So much so that for small language groups learning the lingua franca along far bigger ones, this subsidy may be negative. Thus, overall equality of cost-benefit ratios may require small learners such as the Danes, the Dutch and even the French, to pay, along with the Americans and the Brits, for part of the learning of English by such potential big learners as the natives of Mandarin, Castillan, Hindi and Bengali. [21]

Secondly, as English spreads as a world lingua franca, the quantity of learning may be rising, but its per-unit cost is bound to fall at some point for two reasons. One is that there are more and more opportunities to speak English as the number of (non-native) English speech partners expands, and the expansion of costless opportunities to speak is the surest way of cheapening language learning. The second reason is that the local spread of competence in English makes it possible to provide prospective learners far more cheaply with the competent teachers they need — it is no longer necessary to import natives at high cost or to send children to immersion courses in native territory. For this reason the swelling of the global cost of lingua franca learning is bound to be far less than proportional to the swelling of its quantity. At the limit, if it ever became as easy and natural to learn the lingua franca as it is to learns one’s mother tongue, i.e. if our first problem of linguistic unfairness had become solvable at no cost, our second problem of linguistic unfairness would vanish altogether.[22]

Thirdly, one has to draw the full implications of the fact that talking to some willing native speakers of a language in a context in which it is natural to speak that language is one of the most widespread and most effective ways of improving one’s knowledge of a language — this is precisely the opportunity side of the probability-sensitive learning mechanism at the core of the language dynamics sketched in section 1. But it is its reverse side which I now want to draw attention to. As competence in English spreads worldwide, there are ever less circumstances, because of the maximin dynamics sketched earlier, in which English natives will have a natural opportunity to speak another language and improve their knowledge of it. The advantage of being able to use one’s own language in an ever growing number of contexts therefore has the side effect of making it increasing difficult to learn other languages. Even though the importance of knowing other languages for communication purposes decreases accordingly, this is a genuine disadvantage. One way of putting it is that language learning is to a large extent made up of free riding on patient speech partners. As English spreads, interaction between English natives and others happen more and more, soon nearly exclusively, in English. Consequently, this type of free-riding of English natives on others will get reduced to very little, relative to the symmetric free-riding of these others on English natives (even though an ever growing majority of the people with whom non-English natives will be talking English will be other non-English natives).[23] This growing asymmetry in learning assistance may be far from offsetting the growing asymmetry in exemption from learning, but it does qualify the assessment of the size of the unfairness involved.

 

4.5. Poach the web!

Finally, one must be aware of the fact that both the incentive and opportunity to learn any foreign language but English will decrease as English increasingly suffices to get by wherever one is. As a consequence, English will become more and more a globally public language, while other languages will remain or increasingly become globally private languages. Having no private language means being far more liable to give away information to any outsider who cares to listen or read. This may take some minor forms: whatever your mother tongue, you may benefit from overhearing two American tourists telling each other, in the queue to the museum, that the door to the toilet is locked. Had they been Finnish, you would have lost in vain your position in the queue. Trivial asymmetric benefiting of this sort may seem hardly worth mentioning. But as more and more information gets loaded onto the web, easy to access, copy and use worldwide, this asymmetry is taking gigantic proportions. Whatever is being made available in this way to the 350 million English natives is being made available simultaneously to the 700 million or so non-natives who bothered to learn English (and are massively over-represented among web users from their respective countries). By contrast, practically nothing of the information that these 700 million put on the web in their native languages can be “overheard” by English natives, because so few of these know other l