Reflections on Diglossia

Coby Lubliner

Diglossia, dialect and standard

It’s a warm Friday mid-afternoon in a village in the Peruvian Highlands, not far from Cuzco. In the main plaza, a small crowd consisting mostly of women and young children – some of elementary-school age, some pre-school – is chatting in Quechua, seemingly waiting for something to arrive. And , sure enough, something does arrive: a brightly painted school bus, which stops near the crowd and disgorges a dozen or so neatly uniformed high-school students. It’s their families who have been waiting for them. But before these older boys and girls join in the Quechua chatter with their mothers and younger siblings, they bid farewell to one another and to their schoolmates who stay on the bus to go on to the next village. They talk about plans for the weekend – soccer, movies – and for the following week. How do I know this? Because this conversation is in Spanish, not Quechua.

There are many societies where the school language (usually an official language of the state) is significantly different from the home language (often called a “dialect” or “vernacular”). Such a situation is referred to by linguists as diglossia, a term that was introduced in 1959 in a paper with that title by the linguist Charles Ferguson, “modeled,” as he wrote, “on the French diglossie, which has been applied to this situation, since there seems to be no word in regular use for this in English; other languages of Europe generally use the word for ‘bilingualism’ in this sense as well.”(1)

Ferguson applied the term he coined to situations where the vernacular (which he called ‘low’ or L) and the formal language (called ‘high’ or H) could be regarded as, in some sense, variants of the same language; the examples he gave were those of colloquial and classical Arabic (his area of expertise), demotic and katharevousa Greek, Creole and French in Haiti, and Schwyzertütsch (Swiss-German) and German in Switzerland. In the 1960s, however, research by Joan Rubin in Paraguay showed that the relation between the linguistically unrelated Guaraní and Spanish in that country was, socially, of the same nature, and so the use of term was expanded.

I use the term ‘diglossia’ somewhat grudgingly. A few facts, which I consider important, are missing from Ferguson’s definition. One is that the French diglossie is an adaptation, introduced by the Greek-French writer Jean Psichari (Ioannis Psikharis), of the Modern Greek διγλωσσία, which simply means ‘bilingualism’, and is typically defined in Modern Greek dictionaries as ‘the use by a community of two languages or of two varieties of the same language (emphasis mine). The French philologist Auguste Dozon already commented in 1889, in reviewing Psichari’s work, that the only possible rendition of διγλωσσία would be bilinguisme, which he however characterized as a mot barbare, and which did not become current in French until the linguist Antoine Meillet used it in 1917.

Another missing fact is that ‘bilingualism’ was in fact used “in this sense as well” in English, as Ferguson would have discovered by consulting the Oxford English Dictionary, where he would have found citations to this effect from the 1940s and 1950s; the latest edition of the OED shows that the term was so used in the 1960s as well.

And then there is the fact that when English words are modeled on French words ending in -ie which in turn derive from Greek words ending in -ία or -εία, then, with extremely few exceptions, the English ending becomes -y (monarchy, therapy, theology, philosophy, and so on ad infinitum), unless the word describes a specific medical condition, in which case the Latin-like ending -ia is adopted (anemia, neuralgia, myasthenia, etc.); English even has the doublet melancholy/melancholia. The only exceptions that come to mind are ‘mania’ (along with its compounds) and ‘nostalgia’, both of which have medical analogues (and, of course, ‘many’ would not have worked). An unusual case is that of ‘anomie,’ where the French ending has been kept.

In fact, in 1959 the word ‘diglossia’ was present in unabridged dictionaries (such as Webster’s Third International) with ‘condition of the tongue being bifid’ as its only meaning, while the linguistic term ‘diglossy’ was being used at least by linguists specializing in Greek, for example P. C. Costas,(2) and continued to be so used after the publication of Ferguson’s paper, for example by Robert Browning.(3) It is only in dictionaries published in 2000 or later that the sociolinguistic meaning of ‘diglossia’ can be found.

So much for the word itself. Its choice shows, if nothing else, that even an outstanding linguist can be insensitive to the nuances of his native language.

One point that Ferguson insisted on was a distinction between diglossia as he defined it and the more common “dialect-standard” dichotomy, the difference being that while in the latter situation there are people who actually speak “standard” (as Ferguson no doubt thought of himself as doing), under diglossia no one speaks H colloquially. But I have been told by a Greek friend about an uncle of hers, a priest, for whom katharevousa was the natural medium of expression. And I firmly believe that no one really speaks “standard” as taught in school, with all the grammatical and syntactic rules that one is expected to follow in expository writing. People who think that they speak standard (“incorrectly,” they often believe) actually speak a dialect that is a colloquial, regionally colored variant thereof; I will call such a dialect, for lack of a better term, a parastandard (I would welcome suggestions for a better term). In German it is referred to as Umgangssprache; Italians, French and others speak of italiano regionale, français régional, and the like. In any parastandard there are, of course, degrees by which any idiolect differs from the standard; people who read a lot – academics, for example – are more likely to have their natural speech influenced by the written language. But, as the Israeli linguist Menachem Dagut(4) has written, “[e]very living language displays a greater or lesser degree of diglossia between its spoken and written use.”

Another characteristic of diglossia that Ferguson considered essential was stability. I don’t consider the fact that katharevousa has virtually vanished in Greece a refutation of this assumption; it is rather an illustration of Ferguson’s misinterpretation of the Greek situation, where the polarization between katharevousa and demotic has always been more of a political right-left opposition than a high-low prestige relation. In fact, Ferguson’s characterization of H as being more codified than L was belied by Greek, because a written norm for demotic evolved naturally on the basis of the speech of Athens, while there were – as the poet Cavafis is reported to have said – “a thousand katharevousas.”

It is my observation – typified by the Peruvian scene depicted at the beginning – that diglossia (in the more general sense) typically remains stable in a society as long as most of its children undergo only minimal schooling, limited to a few years of teaching the three Rs; the “dialect” is then likely to persist as the main vehicle of ordinary conversation in the community. But when a significant segment of youth of both sexes attends school well into adolescence, there is a fairly general tendency for a parastandard based on the school language to become the medium of peer conversation, and, after a greater or lesser number of generations, to displace the dialect altogether. The process is of course accelerated if the authorities discourage or suppress the use of the dialect at school, even during recess; this is how, in the course of the twentieth century, most of the regional dialects and patois died out in France, and how (Cajun) French was lost in Southwest Louisiana. Similar cases abound all over the world.

It takes strong counterpressure to oppose this tendency. One example is German Switzerland, where colloquial use of Swiss-German is maintained at all social levels as a kind of symbol of Swiss identity. Similarly, it’s Paraguayan nationalism that keeps even the educated classes of Paraguay speaking Guaraní alongside Spanish.

In the last third of the twentieth century the promotion of linguistic diversity and the preservation of threatened linguistic species became causes associated with left-leaning, “progressive” movements (though such causes are of course inherently conservative, in the literal, nonpolitical sense). At least in part their adoption may have been a reaction to the “one-nation-one-language” policies that were a hallmark of the fascist and quasi-fascist regimes of the century’s middle third, in contrast to the promotion of regional languages practiced by communist regimes. Even today, opposition to such policies as bilingual education is most likely to come from the right. But it wasn’t always thus. The French Revolution saw the dialects spoken by the masses of France (roughly three-quarters of them knew no French, according to a survey undertaken by Abbé Grégoire, the “patriot priest”) as obstacles to social progress; Grégoire’s report to the National Assembly was pointedly titled “On the need and the means to wipe out the patois and to universalize the use of the French language.” A century and a half later, the French linguist Albert Dauzat,(5) writing about attitudes toward the use of Breton in Brittany in the 1920s, reported that “the socialists... are rather lukewarm if not hostile, not to mention that their internationalist views, in general, mesh poorly with regionalism. Conversely, all conservatives are ardent supporters of Breton among the peasants, whether to keep the latter tied to the land or for the sake of social conservation.” And as recently as the year 2000, Jean-Pierre Chevènement, the leader of the leftmost faction of French socialism, resigned as interior minister in order to protest Prime Minister Lionel Jospin’s plan of autonomy for Corsica that would include the teaching of Corsican in the island’s public schools.

Church and language

Indeed, most of the nineteenth-century movements that promoted the revival of previously suppressed languages (and of the national or regional identities underlying them) were conservative in nature. In much of Europe, the Catholic Church played no small part in them.

The Catholic Church has a very old tradition of fostering regional dialects. While, for the almost two millennia preceding Vatican II, the formal liturgy as well as the official writings of the Church (canon law, bulls, encyclicals) were in Latin, for a good deal of that time local vernaculars were used in the informal, mainly oral proceedings such as preaching and communal singing, but sometimes also including written uses such as parish records and individual pastoral letters (in Roussillon, Catalan was used by the Church as late as the 1890s, a century after the Revolution proclaimed the universality of French). It was already in the days of Charlemagne that, while the study of good Latin was promoted, clergy were enjoined by several Church Councils to preach “in the vulgar tongue, in order that the people may understand,” since the common people were no longer expected to understand Latin (the Council of Tours of 813 specifies “in Romance or German,” in rusticam romanam linguam aut theotiscam). While the boundaries of ecclesiastic provinces usually followed historical lines, dioceses were more often than not organized along linguistic ones; the “Leys d’Amor,” a 14th-century codification of Occitan poetry, specifically states that there are as many dialects (parladuras) as dioceses, except that the large diocese of Toulouse has several. Some relics of this organization have persisted until very recently; for example, the diocese of Lleida, in Catalonia (Spain), until the 1990s included a strip of land that has always belonged politically to Aragon, but where the spoken dialect is Catalan.

It was only when, in the wake of Vatican II, Latin was replaced by the “national” languages in the liturgy that the local dialects had to give way to the official standards. In part this was a recognition of the fact that, in the modern world, the laity in general could no longer be assumed to be illiterate, while literacy no longer implied knowing Latin.

The history of Protestant churches is quite different in this regard (a difference that seemed to be lost on Ferguson when he discussed the relation between religion and diglossia). The doctrine of justification by faith and the central place of the Bible as the sole source of truth and salvation – common to the Lutheran and Calvinist churches – meant that “[i]n order that all could have access to the Holy Writ, Protestantism soon came to stand for elementary education for all.”(6) The result was that the language of the Bible version used by a given church was necessarily the language of the church and of schooling.

Now, a Bible translation was, and still is, a major undertaking, and obviously not every regional language variety could have one of its own.

Martin Luther is often given credit for creating modern German by means of his Bible translation. He deserves this credit only in the sense that the language of his translation immediately became the school language in those parts of historic Germany that turned Protestant, except for some northern states that, for a few generations, clung to Low German (which in the Middle Ages had a prestigious written standard). German as Luther wrote it is, in fact, not greatly different from that used by scribes and writers throughout middle and southern Germany since the early 15th century, when a kind of amalgam of Middle and High German began to be used in the Imperial chancery (at the time located in Prague) and in the chanceries of many states and cities.

When Denmark turned Lutheran, it imposed the Danish Bible in Norway, which was under its rule, on the not unreasonable assumption that the Norwegian dialects were close enough to Danish so that Norwegian children could learn the written language without too much difficulty. Swedish children could no doubt do the same, and probably would have if Sweden had not shaken off Danish rule a short time before the Reformation. As the American linguist Einar Haugen(7) has written, “After their political separation the Swedes were not inclined to tolerate Da[nish] features in their writing, and chose to emphasize their differences. The [Bible] translators ... developed a fairly consistent norm, and one that on many points was consciously different from Da[nish].”

But the political motivation for a separate versus a common Bible language was the exception rather than the rule; usually, the motivation was pragmatic. A Bible translation was, as I said, a major project, to be undertaken if linguistically necessary; while the Norwegians did not get their own Bible under Danish rule, the Icelanders did. Similarly Finland, ruled by Sweden at Reformation time, got a Finnish Bible.

Scotland was an independent kingdom when John Knox, after many years in England and among the English in Geneva, returned to lead the Church of Scotland, bringing with him the Geneva Bible and a habit of preaching and writing in English. But even before this event, the Tyndale and Coverdale Bibles, as well as the Book of Common Prayer, were used by Scottish Protestants. The Lowland Scots did not perceive written English as a foreign language; they generally referred to their own language as Inglis, calling it Scottis mainly when they wanted to contrast it with that of England, which they called Southroun or Sudrun, that is, southern (to knap Southroun was to affect Anglicisms in speech). And so it was natural that English became the language of Scottish schools.

Further examples abound of areas where a Bible translation was adopted in a related language of a neighboring area, which then became the school language. Friesland, a self-governing province, adopted the Dutch Staten Bible; for Protestant Slovaks it was the Czech Bible of the Moravian Brothers. The Huguenots of southern France who spoke some variety or other of Occitan (including Gascon and Provençal) got their Bible in French. This was true even in Bearn and Lower Navarre, which, ruled by the House of Albret, were politically independent of France and thus unaffected by Francis I’s ordinance of 1539 making French the only legal language; indeed, a Bearnese variant of Gascon was the official language there long after the rest of southern France switched to French. Queen Jeanne d’Albret made Calvinism the official religion of her realm, and while she commissioned rhymed Psalm translations (for singing purposes) in Bearnese, as well as a bilingual (French-Bearnese) catechism for children, that was about as far as it went: the clergy were made to pronounce Au Diu bibent (Gascon for “by the living God”) when swearing upon the Bible, but the Bible on which they swore was in French, and so were church services. Protestantism was eventually suppressed in most of southern France, but it retained a stronghold in the Cévennes region, and as late as the 18th century travelers passing through it were astounded to find peasants in the Midi who actually knew French.

Jeanne understood, however, that for those of her subjects who spoke Basque – a language totally unrelated to any other – learning to read the Bible in French would be no simple matter, and therefore she commissioned a translation into Basque; while the project was not completed (only the New Testament was done before Jeanne’s son and successor Henri renounced Protestantism in order to become Henri IV of France), it marked the beginning of Basque as a written language.

Like the Basques and the Finns, other linguistic minorities whose language was significantly different from that of a given state’s majority got Bible translations of their own: the Slovenes inhabiting the Austrian province of Carniola (essentially present-day Slovenia) got one; the Sorbs or Wends of Lusatia in eastern Germany got not one but two, since they live in two non-contiguous communities, with the dialect of one (Lower Lusatia, in Brandenburg) closer to Polish and the other (Upper Lusatia, in Saxony) to Czech. And of the three separate valleys in Graubünden where Rhaeto-Romance (Romansch) dialects are spoken, each one got its own – a situation that still complicates language politics in Switzerland, where Romansch has been officially recognized as a national language since 1937, but only recently has a common standard been (grudgingly) agreed on.

In the kingdom of Hungary, however, those Protestants whose speech was Slovak were content with reading the Bible in the closely related Czech, which naturally became their standard (several of the foremost figures of nineteenth-century Czech literature were Slovaks). The pressure for a distinctly Slovak standard came, just as naturally, from the Catholic Church, though it was finally the Protestant Štúr who formulated the modern standard.

Slavs, incidentally, seem to be more sensitive to diglossia than peoples of other language families. The difference between Czech and Slovak is often less than that between dialects of neighboring regions of Germany or Italy, but it was enough for the two peoples to regard themselves as distinct nations and to justify the breakup of Czechoslovakia. In Serbocroat, there are separate standards in Serbia, Montenegro, Croatia and Bosnia, and their users often like to think of them as “different languages.” Macedonian Slavs – who before 1910 considered themselves to be Bulgarians – developed, in the Yugoslavian (now independent) Republic of Macedonia, a standard as different from Bulgarian as they could possibly make it.

Strong and weak diglossia

Accepting, then, the universality of diglossia as proposed by Dagut, I would like to posit a distinction between two kinds, which I will call strong and weak.

Strong diglossia is what has up to know been called diglossia tout court: a situation where dialect and standard are different enough for their users to be fully aware of their being two distinct codes, with different names for them. To designate the dialect in this situation I will use the traditional term patois. This word has traditionally been used in this sense in the northern half of France, as well as neighboring areas of Belgium, Switzerland and Italy (Aosta Valley), in contrast to standard (or parastandard) French; a regional designation (picard, champenois and the like) might be used to contrast one patois with another. In the non-Romance parts of northern France (western Brittany, Alsace), however, the regional designation (breton, alsacien) is what is generally used (when not subsumed under the generic term dialecte), and this is also the norm in Italy and in southern and central Germany (napoletano, Schwäbisch, etc.), while in northern Germany and the Low Countries the generic platt – comparable to the English broad – is usual. In northern Spain, bable and fabla, both meaning, essentially, “speech,” have been used to designate the patois of Asturias-León and Aragón, respectively.

Weak diglossia, by contrast, occurs when the dialect is what I have called a parastandard, and in this case speakers usually think only in terms of “the language” (“English,” “French” or whatever) which may be used more or less “correctly” (the model of “correctness” being of course the standard), and discrepancies are perceived as “faults,” or at the very least as “colloquialisms,” but not as indications of distinct language varieties. To cite a sociolinguistic research example, “the French spoken by the indigenous inhabitants [of Brussels] developed into a separate variety, different from Belgian French ... Belgian French displays a number of typical features that distinguish it from standard French... For many speakers, it was not entirely clear what was meant with [sic] Brussels French. To them, there was only one variety of French.”(8)

The difference in thinking is especially marked during periods of transition from strong to weak diglossia (which occurs, as I have remarked, as secondary schooling becomes universal). To use French again as an example: in Corsica, the generation (those born before 1945, more or less) that grew up speaking Corsican while learning French in school believes that it speaks “[true] French, with all the grammar and all the rules,” while young people “don’t speak it so well,” that is, they speak it “more naturally.”(9)

This transition was once a slow process. In the Protestant regions of Europe where the Bible language was noticeably different from the vernacular, it is generally attested that the change didn’t occur in any great measure, even among the more educated urban classes, until the 18th century, a good two centuries after the Reformation. Examples include the displacement of Norwegian by a Norwegianized form of Danish (Dano-Norwegian, the basis of what is now called bokmål), of Scots by a Scottish-inflected English, of Frisian by a Frisianized Dutch (called Stadfries or “city Frisian”). In the aforementioned Cévennes region of southern France, while knowledge of standard French was common, everyday speech was in Occitan until the nineteenth century.

In northern Germany, it appears that in Hanover – perhaps because of the presence of the electoral (later royal) court – a parastandard High German was spoken by the 18th century as well, at least among the educated, with the curious result that Hanover speech – though non-native – became the model of German pronunciation on the stage (Bühnendeutsch), since everywhere else in Germany dialects were still spoken by everyone. Other capitals (Berlin, Dresden, Munich, Vienna) eventually developed their own Umgangssprachen, but the Hanover model remained the ideal. In the Hansa cities, however, even the merchant class spoke Low German till the end of the nineteenth century, as attested (for Lübeck) in Thomas Mann’s Buddenbrooks.

In the twentieth century the process has been greatly accelerated by the mass media, which carry the parastandard of the cultural centers to once isolated areas where formerly the standard was learned only in school. This doesn’t mean that the parastandard heard on television, on radio or in the movies is necessarily the one adopted, except perhaps by a minority that wants to appear up-to-date. In France, Spain and the United States, the peculiar southern modalities of speech seem to be resisting the dominance of Parisian French, Madrid Spanish and “General American” English, respectively. (In the United Kingdom a kind of reversal seems to be taking place: British people no longer feel the need to lose their regional accents as they rise socially, and the BBC is no longer dominated by speakers of Received Pronunciation [RP], so that this variety has lost its former designation of ‘BBC English.’) What it does mean is that massive exposure to parastandard changes the perception of the “language” from that of a rigid, school-taught and rule-bound medium to that of a living, spoken one – since people don’t consciously differentiate between standard and parastandard – and makes it easier for them to form their own parastandard.

The displacement of a dialect by a parastandard is, however, but one possible way of weakening diglossia. The opposite can happen as well: an old standard is replaced by a new one that is based on the dialect of a prestigious locality (such as Beijing for modern standard Chinese).

The standardization of dialects for literary purposes is usually driven by regional or ethnic consciousness, not purely literary needs. When individual authors have chosen to write in dialect, typically in order to represent the speech of common folk, they have usually created their own versions (examples in American literature include Mark Twain’s Huckleberry Finn and Alice Walker’s The Color Purple). It is only in the context of a collective effort that a standardized written form for a dialect is created. In Europe this has, historically, been limited at first to the needs of poetry, narrative and other popular forms, with science, history, law and the like being left in the domain of the standard.

I will call such a more-or-less standardized version of a dialect a semistandard. Many modern standards began life as semistandards. In medieval Western and Central Europe, for example, Latin was the universal standard, and the written forms of the vernacular were used mainly for popular poetry and stories. It is significant that Dante, who may be said to have created literary Italian, wrote his famous essay on the subject, De vulgari eloquentia, in Latin, just as Pompeu Fabra, the codifier of modern Catalan, wrote his first Catalan grammar in Spanish.

A semistandard may acquire official status and become the language of elementary (but usually not secondary) education, or of parliamentary debate (but not bills); this is the case of Lëtzebuergesch in Luxembourg (which in addition to this semistandard has two standards, French and German) and of Haitian Creole. There can thus be situations where a patois, a semistandard based on it, a standard and a parastandard exist side by side, as is the case of the Peruvian highlands, where a Quechua semistandard (based on Cuzco speech) now has a measure of official status.

Standard Catalan began its ascent from semistandard to full standard in the 1930s, but this process was rudely interrupted by the Franco regime. It was quickly completed, however, in constitutional Spain after 1979.

In Catalonia today, the population is split about evenly between those who use Spanish and those who use Catalan as their primary vernacular (though there are many families, especially in Barcelona, where both languages are spoken interchangeably). In addition, standard Spanish and standard Catalan are both official at all levels of government and education. Only in the courts does Spanish have an advantage, due to Spain’s centralized judiciary in which judges may be assigned to positions all over the country. In the autonomous parliament and local councils, by contrast, Catalan predominates, as it does in the realm of “high” culture: it is the only language of supertitles at the opera, while at lectures, book presentations and poetry readings, even if the text presented is itself in Spanish (or any other language), the introductions are likely to be in Catalan.

Technically, then, those who speak Catalan as their primary dialect and use Spanish as their preferred standard (as well as the reverse, though in this case the numbers are probably small) are in a state of strong diglossia, while those whose linguistic lives are spent entirely (or mostly) in one or the other are weakly diglossic. There are, in fact, people in Catalonia (both in the working class that has immigrated from elsewhere in Spain and the native aristocracy that looked to Madrid for inspiration) who have never bothered to learn Catalan, spoken or written. (The converse, though it exists, is rare.) But for most Catalonians(10) the situation is rather muddled. In particular, younger people of the Barcelona middle class are equally fluent in Catalan and Spanish oral and written use, so that their situation is the peculiar one of bilingual weak diglossia – weak by definition, because they don’t think of the colloquial and formal versions of either language as distinct language varieties.

Teaching “the language”

The tendency, in societies with more-or-less universal schooling, of diglossia to metamorphose from strong to weak (by either of the two processes outlined above) illustrates something like a social analogue to the second law of thermodynamics: weak diglossia represents greater uniformity, and hence higher entropy, than strong diglossia.

I don’t know enough economic theory to know if there is a corresponding economic law, but it also appears that in terms of educational resource allocation, weak diglossia is cheaper. Teaching the standard in a situation of strong diglossia is, at least at first, like teaching a foreign language, and requires a corresponding effort, including teachers with special qualifications, and the recognition that a whole new code has to be taught at the same time that literacy is first introduced. It certainly seems easier to teach children how to read and write words that they already know.

Of course, appearances can deceive. I recall that, when I lived in France over forty years ago, it was reported that the highest scores on the baccalauréat exams (largely measured by the ability to write formal French) were attained by Alsatians and Corsicans, who at that time were still strongly diglossic. I remember thinking at the time that this was precisely because they learned standard French as a quasi-foreign language, with little chance of its being contaminated through a prior use of colloquial French.(11) I compared this process to my own experience of learning English (which is now my primary language) from scratch at the age of ten; I learned every word simultaneously in its spoken and written form, and consequently I make no spelling errors. I had also mastered the rules of standard English grammar before I learned to speak the vernacular in colloquial fashion, and I daresay that I use standard English more “grammatically” than most native speakers (Chomsky’s theories to the contrary notwithstanding).

The previously quoted comments by older Corsicans, comparing their own mastery of (standard) French with that of their juniors, are another case in point, as is the fact that “the French of some people from Haiti or the French West Indies [whose vernacular is Creole] appears stilted or excessively formal to speakers from France.”(12)

In French-speaking societies, the differences between formal and informal language are more sharply marked than in most other weakly diglossic ones, through the concept of “registers.” In the vernacular certain verb forms of standard French – the simple past and the imperfect subjunctive – are simply not used, the interrogative is constructed differently, the negative particle ne is often skipped, and so on. But, even if such standard features are heard as “stilted or excessively formal,” they are still judged as “more correct” or “better” French. It may even be, as Leonard Bloomfield wrote in 1927, that the notion that “some persons are felt to be better models of conduct and speech than others... may be a generally human state of affairs, true in every group and applicable to all languages, and the factor of Standard and Literary Language versus dialect may be a superadded secondary one.”(13)

The great pitfall in teaching the standard to children who speak a parastandard is the generalized belief, even (especially) on the part of teachers, that they are not distinct language varieties but that the former is just a “better” and the latter a “worse” form of “the language.” Rather than introducing the standard as a code to be mastered, with special emphasis on the ways in which it differs from the vernacular – which, however, would be acknowledged as a valid medium in its own right – teachers set out to “correct” what are called “faulty speech habits.” Many such corrections may be based on grammarians’ outdated notions of correctness; an example in English is the proscription of ‘they’ (‘them,’ ‘their’) in the sense of ‘he or she’ (‘him or her,’ ‘his or her’) – a usage that goes back (in writing!) to Middle English and includes Shakespeare, Byron and most notably Jane Austen.

Among non-immigrant English-speakers of North America, diglossia is generally weak except among those whose dialect is AAVE (African-American Vernacular English, also known as Black English and – unfortunately – as Ebonics). The notorious Ebonics controversy of 1997 resulted from an attempt (however clumsy) to teach standard English to AAVE-speaking children in the way that such teaching is done in strong diglossia, implying, of course, an acknowledgment that the dialect (in this case a patois) is a legitimate language variety. But this kind of acknowledgment is difficult for a society accustomed to think in terms of “good English” versus “bad English,” with Black English a particular case of the latter.

Another perverse effect of teaching the standard as a “better” form of the language is that it sets up, in those of a rebellious or simply an intellectually lazy nature, a negative reaction to authority in the form of a refusal to adopt the norms of the standard even under circumstances that clearly call for it. I can’t think of a better example of this effect than the diction of the current President of the United States.


1. "Diglossia” in Word, Vol. 15 (1959).

2. An Outline of the History of the Greek Language with Particular Emphasis on the Koine and the Subsequent Stages (1936).

3. Medieval and Modern Greek (1969).

4. "The revival of Hebrew and language planning” in J. D. Woods (ed.), Language Standards and their Codification: Process and Application (1985).

5. Les patois; évolution-classification-étude (1946).

6. H. P. Douglas and E. deS. Brunner, The Protestant Church as a Social Institution (1935).

7. The Scandinavian Languages (1976).

8. Jeannine Treffers-Daller, Mixing Two Languages: French-Dutch Contact in a Comparative Perspective (1994). The difference between the French-speaking inhabitants of Brussels and those of the rest of Belgium is that for the former French has displaced Flemish, while for the latter it has displaced the Walloon patois.

9. Marie-José Dalbera-Stefanaggi, “Les corses et leurs langues: science et conscience” in J.-C. Bouvier (ed.), Les Français et leurs langues (1991).

10. Here I am taking advantage of the distinction that English allows between “Catalonian” (referring to the country of Catalonia) and “Catalan” (referring to the language and culture), analogous to “Serbian” versus “Serb.”

11. If there is any linguistic contamination in strong diglossia, it is usually the opposite: the vernacular contaminated by the standard. “Spanglish” is a well-known example; another is the extent to which colloquial Catalan has been affected by Spanish during the hundreds of years that the latter was the only school language, to the extent that many Catalan-speakers name the letters of the alphabet in Spanish, including (with sounds that don’t exist in Catalan) [θe] for C and [χe] for G.

12. Gertrud Aub-Buscher, “French and French-based Creoles: the case of the French Caribbean” in C. Sanders (ed.), French Today (1993).

13. “Literate and illiterate speech” in C. F. Hockett (ed.), A Leonard Bloomfield Anthology (1970).

November 12, 2002

Revised July 20, 2004

© 2002 by Jacob Lubliner

Return to Essays index

Return to Personal Page

Return to Home Page

E-mail Coby Lubliner