ArticlePDF Available
Outline of a theory of diglossia
The 3,000 or so entries in Mauro Ferna
´ndez’s bibliography on diglossia
(1993) might well lead one to the conclusion that everything that needed
to be said on the subject had already been said. Yet, forty years after
the publication of the late Charles Ferguson’s historic paper on diglossia
(Ferguson 1959), a coherent and generally accepted theory of diglossia
remains to be formulated. In his last major statement on the subject,
Ferguson himself was moved to comment on how research over the
course of the preceding three decades had failed to adopt a comparative
or typological approach to the study of diglossia, having been pre-
occupied instead with individual case studies focused largely upon the
issue of whether a particular language situation was or was not an
instance of diglossia (1991: 219). This assessment of the situation may
perhaps be seen as a particular example of Chambers’s criticism of
sociolinguistic research more generally, to the effect that ‘‘the social
science content of sociolinguistics has overshadowed its theoretical
implications’’ (1995: 11).
It is abundantly clear that Ferguson’s original contribution was intended
to reach beyond the purely descriptive and classificatory levels of analysis to
the theoretical. The abstract notion of diglossia is derived from the specifics
of four prototypical cases, and its synchronic, typological aspect is
intimately related to its diachronic, evolutionary one. In Ferguson’s own
words, ‘‘what I tried to do was to characterize this situation in the hopes that
here we would have discovered one possible element in a general typology
of socio-linguistic situations’’ (1963a: 163). This typology was to have led
to a ‘‘set of principles or frame of reference,’’ in terms of which patterns
of language use and the evolution of verbal repertoires might be concep-
tualized (1991: 215). Again in Ferguson’s words, ‘‘My goals, in ascending
order were: clear case, taxonomy, principles, theory’’ (1991: 215).
0165–2516/02/0157–0001 Int’l. J. Soc. Lang. 157 (2002), pp. 1–48
#Walter de Gruyter
Consideration of a score or more of sociolinguistic situations that bear
some significant resemblance to those discussed by Ferguson, and in
particular to the Arabic, Greek, and Swiss German cases, suggests that
these situations may form a coherent, bounded, and theoretically
motivated subset of all possible sociolinguistic situations, the shared
social and linguistic characteristics of which may be causally attributable
to common underlying sociohistorical circumstances. In terms of their
synchronic characteristics as well as what may be common patterns in
the natural histories of these situations, diglossic situations may usefully
be compared to and contrasted with other instances of standards-with-
dialects or with instances of societal bilingualism in the most generally
understood use of these terms. The resulting classification is not as
categorical as we might like; there are examples of societal bilingualism
that bear some resemblance to diglossic situations, just as there are, or
have been, examples of diglossia that, in certain aspects of their social
evolution, resemble societal bilingualism. Finally, and as Ferguson
himself has noted (1991: 219), it becomes clear upon a close examination
of a number of case studies of diglossia that this subset itself is not
completely homogeneous, and that even the most unequivocal cases of
diglossia, if such there be, present contrasts in detail that are at times
considerable, yet always instructive for a theory of diglossia and for
sociolinguistic theory more generally.
These caveats aside, one of the purposes of this study, in furtherance
of Ferguson’s interest in ‘‘the sources and outcomes of different language
situations’’ (1991: 223), is to attempt to draw a workable distinction
between the phenomenon of diglossia in the strict sense of the term
and that of diglossia in some broader sense, and to motivate this dis-
tinction theoretically in terms of the sociogenesis of each, the course of
development of each, and the long-term consequences of each for the
verbal repertoires that they constitute, in whole or in part. I will argue,
in the end, for restricting the scope of the term ‘‘diglossia’’ essentially to
that first envisioned by Charles Ferguson in 1959, on the grounds that
diglossia and societal bilingualism, two major types of sociolinguistic
arrangement often regarded as surface variants of the same underlying
phenomenon, are, in fact, fundamentally different in their social origins,
evolutionary courses of development, and resolutions over the long term,
and, furthermore, that inclusion of these two phenomena under a single
rubric obscures rather than clarifies sociolinguistic theory.
There are, in fact, two principal reasons for attempting to forge the
distinction between the type of sociolinguistic situation described by
Ferguson and its counterpart in societal bilingualism. The first of
these concerns the potential relationship between the nature of the
2A. Hudson
complementary functional distribution of linguistic varieties, on the one
hand, and either stability or direction of displacement of codes in the
event of shift, on the other hand. The second concerns the relationship
between linguistic form, or the formal structure of linguistic repertoires,
on the one hand, and language function, as well as change in language
function, on the other hand.
Functional complementarity of codes, acquisition, and stability
The Arabic, Greek, and Swiss German situations are clearly distinct from
typical cases of standard-with-dialects and from societal bilingualism
in that the former are instances of unalloyed register variation, which is
to say, linguistic variation that is stratified by context of use only and
not by the social identity of the user (Halliday 1968). What may have been
left implicit in Ferguson’s first formulation of diglossia is stated with
absolute clarity in his retrospect on the subject: ‘‘If we assume that there
are two basic dimensions of variation in language, dialect variation
correlating with the place of the speaker in the community and register
variation correlating with occasions of use, then the H and L varieties of
diglossias are register variants, not dialect variants’’ (Ferguson 1991: 222).
Such variation has also been characterized as diatypic, rather than
dialectal, or as use-oriented, versus user-oriented (Britto 1986: 37–39).
In other words, stratification of variation in diglossia should then show
sensitivity to differences in situational context without much, or indeed
any, sensitivity to differences in social class (see Bell 1984: 153, figure 3c;
Walters 1996: 175, figure 2d).
The fact that in Ferguson’s canonical cases no section of the community
uses the H variety for ordinary conversation is arguably ‘‘the most
important factor in a diglossic situation and one that makes for relative
stability’’ (Keller 1982: 90). Taking German-speaking Switzerland as
prototype, no native Schwyzertu
¨tsch speaker of any social background
would ever use Standard German for personal, informal interaction
with another. Students of the Swiss case have insisted, for example,
‘‘dass das Verha
¨ltnis von Hochdeutsch und Mundarten nicht _ein
soziolektales Verha
¨ltnis ist’’ (Ha
¨cki Buhofer and Burger 1993: 21), ‘‘dass
in der deutschen Schweiz Dialektsprechen gerade nicht schichtspezifisch
[ist]’’ (Werlen 1993: 9), and, finally, ‘‘[dass] alle Einheimischen aller
sozialen Gruppen Dialekt sprechen, der Arbeiter wie der Hochschul-
professor, der Bauer wie der Bundesrat’’ (Tru
¨b 1992: 21). As a result,
there exists no reference group within the speech community that uses
H for conversational purposes, and therefore no social motivation
Focus article 3
whatsoever for others to imitate this practice in their everyday con-
versation. Parents never speak H to each other (Keller 1982: 91), children
have no opportunity to acquire H as their native variety (Keller 1982: 91),
and, as Ferguson himself has pointed out (1959: 331), no one is ever
comfortable in H to the extent that they are in L. As the younger
generation reaches child-bearing age, the cycle repeats itself. L maintains
itself in this case, due to the fact that the only native speakers of H are
members of another speech community (or tourists or immigrants
therefrom) that do not serve as social role models for the diglossic speech
communities of German-speaking Switzerland. In the case of Arabic, too,
most, if not all, speakers of H and L select between the elevated and
vernacular speech styles based upon situational context rather than social
status of the speaker (Kaye 1994: 58), and ‘‘even Classical Arabic
literature and grammar professors _go home and speak their colloquial
dialects with their children, families, and friends’’ (Kaye 1994: 60). As
Keller (1982) correctly notes, this is a critical factor in the definition
of diglossia and in differentiating it, therefore, from other types of
sociolinguistic situations such as the better-known cases of societal
bilingualism. The general principle for sociolinguistic theory that emerges
from contrasting the two types of situation seems to be that the social
motivation that drives shift from a lower-status language or variety to a
higher one derives from the prestige accorded those speakers who use the
higher variety as a vernacular, and not from the prestige of the social
contexts in which H is employed, provided the same norms of functional
allocation obtain throughout the entire speech community.
It should also be recorded that, just as the realization of H versus L
does not vary as a function of the social identity of the speaker, H
as opposed to L may not be exchanged asymmetrically between inter-
locutors in order to signal social distance or inequality of social standing
in a given interactive event, as may, for instance, traditional Javanese
address styles. In particular, the latter have been differentiated quite
explicitly from diglossia by Errington, who notes that ‘‘Javanese address
styles can be exchanged symmetrically like L and H varieties [in diglossia],
but, unlike L and H varieties, they can also be exchanged asymmetrically’’
(1991: 200), and that this distinction reflects a ‘‘fundamental interactional
difference’’ between honorific language use and the use of H and L in
diglossia (1991: 203). Others would seem to take issue with this position,
as, for instance does Schiffman (1997: 213), when he claims that ‘‘the use
of L may be an expression of solidarity and may not be offered to speakers
whose social position is superior or distant’’ or that ‘‘H may be the only
variety appropriate in a given situation because the use of L would imply
a solidarity that is reserved only for members of a particular in-group.’’
4A. Hudson
Ferguson himself, indeed, has envisaged the possibility that in certain
‘‘idealized’’ speech communities, where everyone knows both varieties,
‘‘choice of the H or L variety can be used in the same way that the
choice of pronouns of address is used in other communities _[and that]
there could be times when it would be appropriate for one person to use
the H variety and for the adressee to use the L variety to mark power
relationships’’ (Ferguson 1991: 228). Switzerland seems close to just such
an idealized situation, yet nonreciprocal use of H would be considered
nothing short of condescending, if not absurd. In this connection, it has
been observed that ‘‘it is psychologically impossible for any two Swiss
of any class or occupation ever to address each other privately in anything
but the ‘Low’ variety’’ (Keller 1982: 91). Exceptions to the reciprocity
principle may be the use by Sanskrit dramatists of H as the language of
kings and priests, and L as the language of the common people, although
in real life ‘‘everyone spoke in one or another L variety and used H only
for religious, literary, and official documents and certain public occa-
sions’’ (Ferguson 1991: 228; see also Lee 1986: 151). Similarly, ‘‘in the
early twentieth century an Arab playwright (Mikhail Nu’aimeh) _had
the educated people speak H and the less educated people speak L,
although in real life everyone spoke L in the situations presented in his
dramas’’ (Ferguson 1991: 228). It seems clear enough, however, that
in both of these examples, the nonreciprocal use of H by upper-class
characters, and of L by lower-class characters, is a dramatic device that
does not accord with contemporary social reality, and that it is, in fact,
a type of metaphorical switching that transposes the horizontal, situ-
ational stratification of H and L into a vertical one intended to highlight
class differentiation.
Social-class stratification of the use of H and L (or its absence) must
be distinguished from social stratification of the opportunity to acquire
H, and from social stratification of access to those social contexts in
which H is appropriate a distinction not explicitly recognized by
Ferguson in his original contribution (but see Ferguson 1991: 227). The
typical case, indeed, seems to be that ‘‘la diglossie ne touche pas toujours
les meˆ mes couches de la population au meˆ me moment’’ (Mackey
1989: 16), and that only ‘‘some speakers possess the requisite linguistic
versatility to be able to use H under one set of conditions and L under
another’’ (Ferguson 1959: 325). As Schiffman has noted, ‘‘in some lin-
guistic cultures, all speakers exhibit diglossic behavior (i.e., use both
H and L varieties in complementary distribution), while in others, only
some members of the society do’’ (1997: 212), and ‘‘in many diglossic
situations, only a minority or elite control the H domain successfully’’
(1997: 206). It goes without saying that whereas social access to the
Focus article 5
informal situations in which L is appropriate is universal, access to those
more formal situations in which H is appropriate is asymmetrically dis-
tributed in favor of those educationally privileged, literate, or otherwise
specialized classes in society most likely to have had the opportunity to
acquire H formally. It is only in this sense that Ferguson’s position can
be reconciled with that of Henry and Rene
´e Kahane, who view diglossic
systems as ‘‘the linguistic representation of a class system,’’ where ‘‘H, the
prestige language, is used by, and therefore becomes the mark of, a sector
of society which excels through power, education, manners, and/or
heritage,’’ and where ‘‘L, the everyday idiom, is the language of the
others, and is often used by H speakers in their non-H roles’’ (Kahane and
Kahane 1979: 183). Britto has aptly formulated this distinction as follows:
‘‘Total superposition does not imply that every member of a diglossic
community knows H and uses H, but merely that there is no portion of
the community which actually knows H and uses H without also knowing
L and using L’’ (1991: 61).
In sum, diglossia, in its ideal form, may be conceived of as the
quintessential example of linguistic variation where linguistic realization
as opposed to language acquisition — here, grossly oversimplified, the
use of H or L — is a function solely of social context, and not of social
identity of the speaker. In diglossia, it is context, not class, or other group
membership, that controls use. Sociolinguistic situations therefore may be
compared to each other in terms of the degree to which the variation
between two or more alternants in their respective code matrices is
determined by social context, as opposed to the social identity of the
speaker, and in situations such as those described for Switzerland, Greece,
most Arabic-speaking countries, and numerous other cases, the bulk,
if not all, of the variance in the use of H and L appears to be explained
by situational context.
Communicative systems may also be compared with regard to the
degree of functional compartmentalization of their linguistic alternants,
and it is clear that Ferguson’s view of diglossia is one where the elevated
and vernacular codes are in relatively sharp complementary distribution
(Ferguson 1959: 328). Others have proposed that Ferguson’s version of
diglossia is simply the extreme end of a continuum ranging from ‘‘rigid
diglossia,’’ where ‘‘there is minimal functional overlapping between
the codes,’’ and ‘‘fluid diglossia,’’ where ‘‘several functions are less rigidly
attached to a particular code’’ (Pauwels 1986: 15). It is a commonplace
that in every speech community, some degree of differential functional
allocation of linguistic varieties is to be expected, whether categorical, as
in the prototypical diglossic case, or statistical, and viewed from a surface
synchronic perspective alone, diglossia may not appear to differ, except
6A. Hudson
in degree, from other instances of situational alternation either between
two languages or between two varieties of the same language.
What distinguishes diglossia from other instances of interlingual
or intralingual situational alternation — even more than the sharp
complementary distribution of linguistic varieties per se — is that the
functional distribution of codes in the diglossic case is one that specif-
ically protects the role of the L variety as a natively learned variety.
Like standard varieties generally, H ‘‘is not ‘native’ to anyone, being a
higher cultural endowment with functions that cannot be mastered until
after the period of normal first-language acquisition’’ (Joseph 1987: 17).
Emphatically, then, diglossia is a special case of sharp functional dif-
ferentiation of registers in which the H variety, or set of varieties, is
nobody’s mother tongue. According to Ure, diglossia is a special case
of separate sets of registers ‘‘in which the marked set _is not the
mother tongue of any members of the community’’ (1982: 16). Coulmas
too has noted that diglossia generally ‘‘is characterized by a long-term
coexistence of native spoken varieties of a language, on the one hand, and
a written variety which is the mother tongue of nobody, on the other’’
(1987: 117). Stated differently, since adults ‘‘invariably use L in speaking
to children and children use L in speaking to one another, _Lis
invariably learned by children in what may be regarded as the ‘normal’
way of learning one’s mother tongue,’’ and the learning of H ‘‘is chiefly
accomplished by the means of formal education’’ (Ferguson 1959: 331).
Finally, as Schiffman observes, ‘‘In diglossia no one speaks the H variety
as a mother tongue, only the L variety. In the standard with dialects
situation, some speakers speak H as a mother tongue, while others speak
L varieties as a mother tongue and acquire H as a second system’’
(1997: 207). Given the express, widely held view that only L is acquired
as the natural mother tongue in a diglossic speech community, it is
remarkable that time after time in the sociolinguistic literature this
critical feature of diglossia is disregarded, as, for instance, in the case of
Paraguay, where Spanish and Guaranı
´are in fact the mother tongues
of two distinct segments of the community.
It is scarcely imaginable, as Ferguson himself has noted, that any
change toward full utilization of H could take place without the will-
ingness on the part of adults to speak H for conversational purposes,
especially with their children (Ferguson 1959: 331). True, in other
instances of language or dialect contact, the vernacular role of the L
variety is also relatively protected in the initial stages, but what further
distinguishes diglossia from these latter is the absence of any prestige
group that employs H as its vernacular and could, therefore, provide the
social impulse for shift away from L as the vernacular. Language shift,
Focus article 7
like internal language change, requires social motivation, but, as noted
above, this motivation derives from the adoption as role models of high-
prestige individuals who themselves already employ H as a native,
vernacular variety, and not from its association with prestigious, high-
culture domains of interaction. In the absence of such a reference group,
and given the functional protection afforded L as a natively acquired
variety, stability in that role is to be expected. Short of stability, then
shift from H to L in formal domains of interaction, or shift from H to a
new vernacular-based standard, is to be anticipated, as opposed to shift
from L to H in informal domains.
It was Ferguson himself who pointed out that the strictly complemen-
tary functional allocation of the elevated and vernacular varieties in the
Arabic, Greek, Haitian, and Swiss German cases had existed for periods
of no less than several centuries in each case. From this observation, it
has been tempting to conclude that situations like these are inherently
stable, and therefore that the attainment and maintenance of sharp
functional compartmentalization of codes may be the key to minority-
language maintenance and even reversal of language shift within multi-
lingual or multilectal speech communities (see Fishman 1991; Paulston
1994). To argue this is in essence to beg the question, however, since
the means whereby such functional compartmentalization entailing
the eradication of the ‘‘H’’ variety from all domains of personal, informal
interaction and therefore from all first-language-acquisition contexts —
might be motivated, attained, and maintained are not suggested by the
model. Moreover, stability is a relative matter, and the fact remains that
Colloquial Arabic, Dhimotikı
´, and Schwyzertu
¨tsch have all gained
ground in the twentieth century, to one degree or another, at the expense
of their coterritorial elevated counterparts. It is clear, then, that functional
compartmentalization at a given point in time is in and of itself no
warranty against eventual shift. Whether the long-term maintenance of a
vernacular variety necessarily follows from functional compartmentaliza-
tion of codes, or whether its displacement by a more prestigious rival
follows from the absence of such functional compartmentalization, is
far from obvious. Nonetheless, the fact that in the cases just mentioned,
as in numerous others, it is precisely the elevated or culturally prestigious
variety, and not the vernacular, that has been displaced to some extent is
reason enough to distinguish these from other cases where it is typically
the vernacular that has been displaced, often to the point of exinction,
by a prestigious competitor.
Although most often described, and problematically so, in terms
of synchronic functional distribution of linguistic varieties, it is the
diachronic aspect of diglossia — its sociogenesis, course of development
8A. Hudson
across time, and ultimate resolution that is of greatest significance
where sociolinguistic theory is concerned. The variables discussed by
Ferguson may in fact be viewed as contextual, linguistic, or temporal in
nature. Function, prestige, acquisition, literary heritage, and standard-
ization are all in one way or another contextual aspects of diglossia.
Grammar, lexicon, and phonology are linguistic aspects. Stability, clearly,
refers to the temporal axis along which the various contextual and
linguistic variables may vary. In the main, scholars to date have treated
Ferguson’s nine characteristics of diglossia as a potentially open-
ended checklist of features or parameters of a sociolinguistic typology
or have tended to focus their attentions on one feature only, or a smaller
number of such features, most notably the functional compartmen-
talization of codes and the structural distance between codes. The position
advanced here, however, is (1) that Ferguson’s features are not an open-
ended list, but rather a closed, tightly interdependent set of features,
(2) that functional compartmentalization, acquisition, and stability of L
as vernacular are causally related, and (3) that the abstract social factor
underlying these latter three features is in turn related in a nonrandom
way to the structural relationships between the diglossic codes.
Structural relationships between codes in diglossia
The four cases described by Ferguson are remarkable for the fact that
the elevated and vernacular varieties in each bear a close structural
resemblance to one another not so close as to be readily mutually
intelligible, but also not so distant as to be unhesitatingly regarded as
separate languages. While acknowledging the difficulty, on the strength
of linguistic criteria alone, of identifying different language varieties as
varieties of the same or of different languages, and setting aside too the
vexed question of the nature of the creole–standard relationship in
the Haitian situation, it is hardly contestable that the linguistic affinities
between Classical or Modern Standard Arabic and Colloquial Arabic,
between Standard German and Schwyzertu
¨tsch, or between
Katharevousa and Dhimotikı
´, are of an order quite apart from that of
the affinities between, say, Spanish and Guaranı
´in Paraguay, Spanish
and Nahuatl in Mexico, or even Spanish and English in the American
Southwest. Other pairs, such as Dutch and Frisian in the Netherlands,
French and Provenc¸ al in France, and Catala
´n and Castillian in Spain,
may, admittedly, be more problematic to place linguistically between the
extremes of linguistic affinity, but the question that nonetheless presents
itself is whether proximity of linguistic relationship between varieties,
Focus article 9
however measured, is causally associated with the potential for
displacement of one by the other.
Ferguson sets his initial discussion of diglossia solidly within the
context of language situations where ‘‘two or more varieties of the same
language are used by some speakers under different conditions’’ (1959: 325
[emphasis added]) and further asserts that ‘‘no attempt is made _to
examine the analogous situation where two distinct (related or unrelated)
languages are used side by side throughout a speech community, each
with a clearly defined role’’ (1959: 325, note 2). He later recognizes
explicitly the difficulty of measuring linguistic distance and of identifying
two varieties as members of the same or different languages (Ferguson
1991: 220). Elsewhere, too, he adverts to the possibility that some future
sociolinguistic typology might extend the term diglossia to include certain
situations where the codes in question might be structurally unrelated
(Ferguson 1963b: 174), and, more particularly, to the possibility that the
relationship between Spanish and Guaranı
´in Paraguay might also
reasonably be described as diglossic (Ferguson 1973: 41– 42). Never-
theless, his intention, by his own admission, had been ‘‘that the users
would always view the two [varieties] as the same language’’ and that
‘‘diglossia’’ should not extend to cases ‘‘where superposed on an ordinary
conversational language is a totally unrelated language used for formal
purposes, as in the often-cited case of Spanish and Guaranı
´in Paraguay’’
(Ferguson 1991: 223).
Over time, opinion with regard to the theoretical significance of the
relatedness between codes in diglossia has varied in the extreme. Some
scholars, following Ferguson, have maintained the distinction between
related and unrelated language varieties. In proposing the term
‘‘diglossia’’ for situations where two relatively closely related varieties
are in complementary functional distribution, and the term ‘‘biglossia’’
for circumstances where the varieties in question are varieties of different
languages, Fellman has implied a fundamental distinction between the two
(1975a: 39). More recently, Daltas too has proposed that ‘‘Fergusonian
diglossia should be preserved as a useful notion to refer to a specific type
of variation involving an H and an L variety of (what is societally seen as)
the same language’’ (1993: 348).
Others have taken a view of the relationship between codes in
diglossia as being intermediate between that involved in multilingual
situations, on the one hand, and stylistic shifting on the other (Daltas
1993: 341). An important issue here is whether structural intermediacy
in this sense is directly or indirectly asserted to connote sociological
intermediacy as well. Wexler, for instance, seems not to be concerned
with the latter, but rather, notwithstanding the methodological dilemmas
10 A. Hudson
involved, with establishing the minimum distance between codes in
diglossia, so that the latter may be distinguished from more common
cases of standards-with-dialects by the existence of ‘‘a broad structural
gap between the standardized written norm and the unstandardized (as
a rule) spoken dialects’’ (1971: 336). Coulmas, likewise, assumes ‘‘that a
large structural gap between the spoken norm and the written norm of
a speech community is the most salient feature of diglossia,’’ although
‘‘it is far from obvious how the distance between spoken and written norm
can be measured, and how the gap in one language can be compared with
that in another’’ (1991: 126).
For other scholars, too, the difference between diglossia and societal
bilingualism appears largely to reduce to the degree of linguistic difference
between varieties in the code repertoire. Fasold, for instance, views the
diglossia of the Arabic-speaking world, Greece, and German-speaking
Switzerland as ‘‘a convenient midpoint in the possible range of relatedness
to be found in broad diglossia’’ (1984: 54), thus distinguishing the former
from so-called superposed bilingualism, where the prestige and vernacular
varieties are more distantly related to each other, and from style-shifting,
where the two varieties are more related structurally. Such a categori-
zation, however, is at one time overinclusive and underinclusive. It is
overinclusive in that it admits to the category of diglossia sociolinguistic
situations such as Dutch–Frisian in the Netherlands and French–
Provenc¸ al in France, neither of which, in all probability, properly
belongs. Likewise, it excludes situations such as Yiddish–Hebrew in
Jewish communities in Europe prior to World War II and among certain
groups of Ultra-Orthodox Jews in Israel today, all of which deserve
serious consideration as candidates for the designation of ‘‘diglossia.’’ The
difficulty with such approaches to the identification of diglossia is that
they focus almost exclusively upon structural relationships between codes
in the community verbal repertoire at the expense of social and functional
Along much the same lines, Britto regards the intermediate distance
presumed to exist between H and L in diglossia as ‘‘optimal,’’ so that these
linguistic varieties may be distinguished ‘‘from ‘languages’ (which would
be Super-Optimal varieties), and ‘‘styles,’’ ‘‘accents,’’ etc. (which would be
Sub-Optimal varieties)’’ (1986: 10). Recognizing that ‘‘it is not always
easy to judge when a pair of varieties are optimally distant’’ (Britto 1986:
19) and that ‘‘there are at present no adequate techniques to measure it
[optimal distance] empirically’’ (1986: 22), Britto proposes nevertheless
that ‘‘if such clear norms are established to identify the Optimal vari-
eties from other varieties which are too closely or too distantly related,
then it may be possible to apply these norms to a given situation and
Focus article 11
evaluate whether the situation is diglossic or not diglossic’’ (1986: 20).
This reasoning borders on tautology, in that if there are no empirical
measures by which the structural relationship between two codes may
be identified as ‘‘optimal’’ other than that the sociolinguistic situation in
which two codes coexist is already known to be diglossic, then it is clear
that optimal distance between codes cannot be used as a diagnostic
in sociolinguistic situations, the status of which is not known by
independent criteria. The best that can be said of this approach is that
it identifies the linguistic distance between codes in the Arabic, Greek,
Swiss German, and Tamil cases as prototypical for diglossia, assuming
them to be commensurate in the first instance. In the end, establishing a
certain distance between codes as criterial for diglossia proves no less
of a conundrum, and perhaps no more theoretically relevant, than
ascertaining whether two linguistic varieties should be identified as the
‘‘same’’ language. Perhaps all that can be said, then, along with Ferguson,
is that ‘‘in diglossia there are always extensive differences between the
grammatical structures of H and L’’ (Ferguson 1959: 333 [emphasis in
Other attempts to formulate the relevant linguistic questions expose a
lingering preoccupation with structural distance as a defining character-
istic of diglossia. Winford, for example, has asked, ‘‘What are the outer
limits of the structural gap, beyond which a situation becomes bilingual
rather than diglossic in Ferguson’s sense’’ and ‘‘what is the minimal
structural gap between varieties that will qualify a situation for diglossic
status?’’ (1985: 351). Ironically, even Ferguson himself regarded ‘‘the
failure to make clear how far apart (or how close together) the high
and low varieties have to be for a language situation to be character-
ized as diglossia’’ as a shortcoming of his original article (1991: 223).
Paolillo, more recently, has proposed that examination of a variety
of different sociolinguistic situations ‘‘would allow us to establish what
degree of difference constitutes _the structural difference between
diasystems necessary and/or sufficient for a situation to be diglossic’’
(1997: 292). Upon reflection, these turn out to be curious, linguistically
deterministic ways of framing the problem, implying either that diglossia
is arbitrarily defined by what appears as theoretically unmotivated
reference to surface grammatical characteristics of the diglossic codes,
or, alternatively, that structural discrepancies of a given magnitude
between the elevated and vernacular varieties in diglossia could be held
to account theoretically for the remaining features of diglossia, par-
ticularly those concerning functional complementarity of codes, the
protection of L as the sole variety acquired as a native variety, and
the stability of diglossic situations or anticipated direction of shift
12 A. Hudson
upon their dissolution. This last proposition seems on the face of it
implausible, casting linguistic structure as the principal determinant of
At the other extreme, Joshua Fishman has implicitly dismissed the
degree of structural proximity between codes as irrelevant to the definition
of diglossia, asserting that ‘‘diglossia exists not only in multilingual
societies which officially recognize several ‘languages’ but, also, in
societies which are multilingual in the sense that they employ separate
dialects, registers or functionally differentiated language varieties of
whatever kind’’ (1967: 30), and provided too that ‘‘without schooling the
written/formal-spoken [variety] cannot be understood’’ by speakers of
the vernacular (1980: 4). This view is also championed by Pauwels, who
employs the term diglossia for functional differentiation of codes within
speech repertoires generally but distinguishes between interlingual
diglossia, where the codes are not varieties of the same language, and
intralingual diglossia, where they are (1986: 15). This is as much as to
suggest, for example, that the asymmetric functional allocation of
Nahuatl and Spanish in Mexico is to be distinguished from that of
Katharevousa and Dhimotikı
´in Greece primarily on the basis of the
structural discrepancies between the codes in each case. Clearly the two
situations are worlds apart, however, not only in their surface linguistic
dimensions, but, more significantly, in their sociohistorical origins,
evolutionary courses of development, and ultimate resolutions.
Fishman was among the first to propose that the role of linguistic
relatedness between codes in diglossia be minimized, on the grounds
that it is social consensus rather than structural affinity that confers the
status of distinct languages on two or more linguistic varieties (1967: 33,
note 4). A similar theme has been sounded more recently by Berger, who
adds that ‘‘individual languages which are clearly distinguishable from
a structural point of view are not necessarily distinguishable for the
speaker,’’ and, conversely, that ‘‘speakers may consider their idiom as a
separate language although this is not tenable from a linguistic point of
view’’ (1990: 290). Mackey is almost certainly correct in his suggestion
that the theoretical difficulty inherent in establishing language identity has
been responsible for the extension of the notion of diglossia to include
any set of linguistic varieties, related or not, and that this extension in turn
has resulted in the massive multiplication of the number of situations
throughout the world that might be called diglossic (1986: 239).
Fishman’s original extension of the scope of diglossia may also be
considered in the light of an early thrust in sociolinguistic thought toward
the theoretical integration of dialect variation, diglossia, and societal
bilingualism as surface variants of the same underlying phenomenon.
Focus article 13
John Gumperz had argued that the distinction between bilingualism
and bidialectalism was not always a significant one in the social
characterization of speech communities (1968 [1962]: 463), and that
linguistic difference need not necessarily correspond to difference in
social function (1968 [1962]: 464). Fishman himself felt that a ‘‘single
theory _enabling us to understand, predict and interrelate both of
these phenomena [diglossia and bilingualism] is an instance of enviable
parsimony in the behavioral sciences’’ (1967: 32–33), and, in the same
volume, Dell Hymes took the position that ‘‘cases of bilingualism par
excellence (as for example French and English in Canada, Welsh and
English in Wales, Russian and French among pre-revolutionary Russian
nobility) are salient, special cases of the general phenomenon of variety
in code repertoire and switching among codes’’ (1967: 9). Later still,
Sankoff states that the shifting or switching among various codes
typical of individuals in multilingual societies ‘‘does not differ qualita-
tively from the behaviour of monolinguals (shifting of style or level)’’
(1972: 33). Upon reflection, however, it is not at all clear that the French–
English situation in Que
´bec and the Welsh–English situation in Wales
are of a piece with the Russian–French situation in prerevolutionary
Russia, nor that the Nahuatl–Spanish situation in Mexico might not
have a great deal more in common with the Occitan–French situation
in southern France than either has with the Schwyzertu
situation in Switzerland. The emphasis on asymmetric functional dif-
ferentiation of codes as the common denominator of all these situations
may, rather than disclosing them to be surface variants of the same social
phenomenon, instead have resulted in the surface merger of funda-
mentally distinct sociolinguistic situations. What is actually called for
in the further study of diglossia is the establishment of a balance, and an
interaction, between two complementary approaches: one, a universalist
approach such as that discussed above, which emphasizes the similarities
between diglossia and other sociolinguistic situations, and another,
particularistic approach, which focuses upon the differences.
In fact, too much has been made of the degree of structural proximity
between constituent codes in a verbal repertoire as a diagnostic of
diglossia. Defining the codes in diglossia a priori as varieties of the same
language or otherwise is an arbitrary gesture and in itself contributes
nothing of value to sociolinguistic theory. As others have noted, the issue
is an empirical one, amenable to investigation via comparison of a variety
of linguistic situations (Paolillo 1997: 294, note 11). The position taken
here is that comparative analysis of a sufficiently broad spectrum of
linguistic situations is likely to reveal that the communicative arrange-
ment characterized by Ferguson for Greece, Switzerland, and the Arab
14 A. Hudson
world is a social arrangement more fundamentally than it is a linguistic
one, and that there is, therefore, no absolute causal connection between
linguistic relatedness of the codes and the direction of eventual or
potential shift. This point of view situates diglossia within the more
general framework of prestige languages, as this term has been employed
by Henry and Rene
´e Kahane, and shares the view expressed by these
scholars with regard to structural relatedness between codes:
‘‘Genetically _H may be a variety of L, either similar to it (say, early
Medieval Latin vs. early Romance in the Carolingian age); or it may be
dissimilar (late Medieval Latin vs. French in the 15th c.); or it may be
a truly foreign language to speakers of L (Norman French vs. the English
of the Conquest period)’’ (Kahane and Kahane 1979: 183).
This much said, however, if the structural difference between codes
in diglossia is viewed as an outcome of the social circumstances giving
rise to diglossia in the first place, rather than as a defining feature of
diglossia, there is ample reason to suppose that language varieties in
diglossia will in fact show a strong statistical tendency to be varieties
of the same language, or, given the difficulties inherent in this notion, to
evince structural relationships on the order of those obtaining in the
Arabic, Greek, and Swiss-German cases. While it is possible to recall
instances where comparable social relationships exist between decidedly
distinct languages, such as Yiddish and Hebrew in certain Ultra-
Orthodox Jewish communities, this appears to be more the exception
than the rule. Rather, it is in the very nature of a diglossic accomoda-
tion between codes that the linguistic varieties involved tend more often
than not to bear a relatively close linguistic relationship to one another,
although the precise nature of this relationship and the specifics of
its history may vary from case to case, as will be demonstrated below.
Should this be the case, definitions of diglossia that are predicated
upon linguistic distance between the constituent codes will prove to be
essentially beside the point. The formulation of the problem adopted
here is, given a particular pattern of functional complementarity of
codes, a particular pattern of first-language (or dialect) acquisition, and
a particular diachronic pattern whereby H tends either to converge upon
or to be displaced by L, what structural characteristics are to be expected
in such a code matrix and what structural characteristics are actually
found? The position taken here is that the linguistic arrangement char-
acterized by Ferguson for Greece, Switzerland, and the Arab world
is fundamentally a sociological rather than a grammatical one, and that
there is, therefore, no direct causal connection between linguistic distance
between the component codes in such a code matrix and direction of
eventual or potential shift.
Focus article 15
While degree of linguistic affinity between codes may not be criterial
to the definition of diglossia, it is highly relevant, as Walters (1996: 158)
has remarked with respect to Arabic, to inquire as to the nature of the
variation found in the preponderance of cases where H and L are indeed
relatively closely related varieties. Ervin-Tripp has suggested that code
use in diglossia may be accompanied ‘‘by more cooccurrence restriction
than is style shifting where the common features of the styles may
outweigh their differences’’ (1971: 44 45). In cases similar to that of
traditional Javanese linguistic etiquette, cooccurrence restrictions appear
to be absolute: selection of a given item from one particular lexical
set absolutely controls selection of the appropriate item from a second,
independent, lexical set (Errington 1991; Geertz 1968), and these patterns
of selection, or ‘‘linked conjugates’’ or ‘‘stylemes,’’ as Geertz refers to
them (1968: 287), clearly identify the social level of the language variety
in question. Ferguson’s claim that the use of the H or the L member of
a lexical doublet immediately stamps a text as H or L suggests that the
lexical cooccurrence restrictions in diglossia are more akin in their
stringency to those operative in the situation described by Geertz
(although the social dynamics involved are very different) than they are,
say, to the rather looser cooccurrence restrictions obtaining in English
between illumination,purchase, and children, on the one hand, and light,
buy, and kids, respectively, on the other (Ferguson 1959: 334). In the case
of the English doublets, ‘‘both words may be written and both may
be used in ordinary conversation: the gap is not so great as for the
corresponding doublets in diglossia’’ (Ferguson 1959: 334). Further, ‘‘the
formal–informal dimension in languages like English is a continuum in
which the boundary between the two items in different pairs may not
come at the same point’’ (Ferguson 1959: 334). Biber (1988) and others
have developed a factor-analytic technique for the cross-linguistic study
of cooccurrence restrictions in written and spoken language, which,
applied to diglossic code matrices, might help to provide a more clinical
assessment of the extent and stringency of the cooccurrence restrictions
governing the use of H and L in various diglossic contexts. To date, the
only study of a diglossic code matrix to approximate this approach,
though without employing factor analysis as a statistical procedure, has
been Paolillo’s (1997) study of continuity versus discreteness in Sinhala
diglossia, to be discussed further below.
Where the realization of one variable does not completely control the
realization of another, other kinds of cooccurrence constrains are pos-
sible. The most obvious alternative to strict cooccurrence restrictions are
the implicational scalings of the type proposed by DeCamp for Jamaican
Creole English (DeCamp 1971) or by Bickerton for Guyanese Creole
16 A. Hudson
English (Bickerton 1971). It should be noted, however, that Paolillo’s
attempt to scale H and L variables in Sinhala has met with only limited
success, compared to those reported for various creole situations: out of
sixteen Sinhala variables investigated, only in the case of five, all of these
within the colloquial cluster of texts, ‘‘did anything like a scale emerge’’
(Paolillo 1997: 287 [no coefficient of scalability given]). On the strength of
this, Paolillo concludes that there seems ‘‘to be a broader tendency toward
covariation of H and L features, with a moderate degree of hybridization
being tolerated’’ (Paolillo 1997: 288). Whether Paolillo’s findings in the
case of Sinhala will be confirmed in other cases of diglossia remains to
be empirically tested. In any case, it is clearly possible to examine the
cooccurrence restrictions between linguistic variants in any given code
matrix and to establish whether such restrictions are categorical,
implicational, or correlational in nature, and further to explore whether
the variation documented in cases of diglossia spans the full range of such
possibilities or only a significantly restricted segment of it.
To the extent that more formal and less formal varieties in a code
matrix may be clearly distinguished, the general question has been raised
whether the more elevated varieties are less tolerant of variation than are
the vernacular varieties. In that the H variety or varieties in diglossia are
standardized, according to Ferguson, ‘‘there is an established norm for
pronunciation, grammar, and vocabulary which allows variation only
within certain limits’’ (1959: 332). By contrast, where the L variety
or varieties are concerned, ‘‘there is no settled orthography and there is
wide variation in pronunciation, grammar, and vocabulary’’ (1959: 332).
Substitution or switching has long been described as typical in casual
utterance generally (Voegelin 1960: 62), and, conversely, code consistency
has been proposed as a universal characteristic of more formal com-
municative events (Irvine 1984: 215). It may be, therefore, that the H
variety or range of varieties in diglossia presents nothing more than an
instance of a general principle of relative invariance in formal codes.
However, in postdiglossic speech communities, as the standard gradually
accommodates to the influence of the vernacular in writing, and where
the standard is no one’s native variety, it may be that the relatively
‘‘ill-defined’’ standard evinces a greater degree of variation than the
relatively ‘‘well-defined’’ vernacular (Kaye 1970, 1972).
Related to the question of cooccurrence restrictions in diglossia is the
question of discreteness as opposed to continuity in linguistic variation.
Stated simply, do the individual styles found in a diglossic code matrix,
like those found in certain postcreole continua, arrange themselves in a
unidimensional series of minimally different varieties, reaching from an
extreme basilectal vernacular variety at one pole to an extreme acrolectal
Focus article 17
standard at the other? Or is this sequence of minimally different lects
punctuated at one or more points along its progression? One view, at
least, is that diglossia is ‘‘the most striking example’’ of a divided register
range, or of multiple sets of registers within the same language (Ure
1982: 16). Ferguson himself, while conceding that a continuum of
variation can be documented in every one of his canonical cases,
nonetheless believes that ‘‘in the diglossia case the analyst finds two poles
in terms of which the intermediate varieties can be described; there is
no third pole’’ (1991: 226). Paolillo’s Sinhala data in fact suggest
‘‘a broader tendency toward covariation of H and L features’’ (1997: 288).
Although there is no absolute separation of texts into those characterized
exclusively by H features and those characterized exclusively by L
features, some 52 text segments group themselves empirically into two
broad clusters, such that the texts within each cluster are more closely
related to each other than are any two texts from different clusters. These
two clusters appear to represent texts in which Literary Sinhala features
are relatively prominent, on the one hand, as opposed to those in which
Colloquial Sinhala features are relatively prominent, on the other
(Paolillo 1997: 288).
Yet another question concerning the grammatical relationships
between codes in diglossia is, to the extent that the codes are indeed
varieties of the same language in some real or putative sense, does the
nature of the grammatical relationships between H and L tend to be
consistent across cases? For example, is the morphology of H con-
sistently more complex than L in the operational sense employed by
Ferguson, to wit, that morphemes in L have fewer alternants and
morphemic alternation is more regular, fewer obligatory categories are
morphologically realized, paradigms are more symmetrical, and/or there
is stricter concord and rection (1959: 333–334)? The Arabic, Greek, and,
arguably, Swiss-German cases seem to be consistent with this hypothesis
(Ferguson 1959: 334), but to the extent that this may be generally true
of diglossia, how can these relationships be explained? Are they, for
example, the normally expected consequence of discourse-pragmatic
constraints upon the structure of formal registers (Givo
´n 1979), are they
the consequence of universal principles governing the cultural, but
pragmatically arbitrary, expression of formality in communicative events
generally (Irvine 1984), or, last, are they due to differing rates of change
in formal and informal linguistic structures, such that H invariably reflects
an earlier form of L, and L, therefore, a later stage upon an expected,
perhaps universal, path of grammatical development (Bybee et al. 1994)?
Literary Kannada, for example, is viewed by one scholar as ‘‘an archaism,
a stage which the language reached some centuries ago, when it became
18 A. Hudson
‘frozen’ by social convention’’ (Bright 1976: 66). The colloquial dialects,
on the other hand, continued to change, with the result that ‘‘modern
literary Kannada represents, to a large extent, an earlier historical stage
of the modern colloquial dialects’’ (Bright 1976: 66).
Social origins of diglossia
Beyond the synchronic description of the social and linguistic dimensions
of diglossia, of greater significance to sociolinguistic theory is an under-
standing of the evolutionary relationship between linguistic form and
social function. In diglossia research, if not in sociolinguistic research
more generally, it is remarkable how little attention has been paid to the
role of macrosocietal and broader cultural processes in the differentiation
of various types of verbal repertoires. The position of diglossia within
an evolutionary taxonomy of speech repertoires may, perhaps, be gauged
with reference to Gumperz’s typology of linguistic communities
(Gumperz 1968 [1962]). While even small bands of hunters and gatherers
reveal differences between ‘‘casual every-day speech and non-casual styles
used in singing, recitation, myth-telling and similar ritually defined
situations’’ (Gumperz 1968 [1962]: 466), the opposition, central to
diglossia, between a vernacular variety learned in the home and a super-
posed variety learned after childhood becomes salient in intermediate
societies constituted of peasant, herder, or even tribal populations
integrated to varying degrees into the dominant society and exhibiting
a high degree of social stratification and occupational specialization
(Gumperz 1968 [1962]: 467–468). Such societies are likely to develop
special sacred and administrative codes, in addition to other special
parlances, which are ‘‘characterized by extreme codification,’’ and which
require, therefore, ‘‘a large investment of time in the study of grammar
and rhetoric,’’ as well as schools, with their complements of scholars,
for the pursuit of such study (Gumperz 1968 [1962]: 469). These codes
‘‘serve as the language of special administrative and priestly classes’’ and
function, at least in part, ‘‘to maintain group exclusiveness’’ (Gumperz
1968 [1962]: 469). When government ‘‘remains in the hands of a small
ruling group,’’ it is possible, even in the face of increasing mobilization
of the population, to maintain considerable language distance between
the administrative and sacred H-codes, on the one hand, and the rest
of the code matrix on the other (Gumperz 1968 [1962]: 469). It may
be reasonably hypothesized, therefore, that intermediate societies as
described by Gumperz have offered singularly fertile ground for the
emergence of classical diglossia.
Focus article 19
Gumperz’s intermediate societies resemble quite closely Sjoberg’s
‘‘preindustrialized civilized societies,’’ in which the bulk of the written
tradition consists mainly of the society’s sacred writings, and where
writing is restricted to, and is perpetuated by, a small, educated, priestly
group (Sjoberg 1964: 892). It is this educated elite that, ‘‘in formal
situations of various kinds, speaks in the traditionally high-status
manner,’’ while ‘‘in other situations, especially to communicate effectively
with less educated or uneducated persons, they must use a more informal
speech style, one that is more akin to the colloquials of the area’’ (Sjoberg
1964: 893). As a result, ‘‘the upper status, educated group typically
employs at least two speech styles, in some cases more,’’ all of which
‘‘differ from the speech of the common man — in the lexicon and often
the phonology and grammar’’ (Sjoberg 1964: 893). Furthermore, the
formal speech style ‘‘tends to be perpetuated over centuries with rela-
tively little change, a phenomenon that results from the high prestige
accorded it and its close tie with the written language’’ (Sjoberg 1964:
893). While neither Gumperz nor Sjoberg explicitly mentions diglossia
as such, their social descriptions of ‘‘intermediate’’ and ‘‘preindustrialized
civilized’’ societies, and their corresponding verbal repertoires, bear
more than a passing resemblance to many cases of diglossia up to
recent times.
A distinction of potential significance for evolutionary studies of
speech repertoires is that drawn by Besnier between determinants
of linguistic form that have their basis in the ‘‘physical and psychological
characteristics surrounding the situation of use,’’ on the one hand, and
those that have their basis in ‘‘the norms of communication at play in
each context,’’ or ‘‘the cultural ‘value’ of a communicative context’’
(1988: 731), on the other hand. It has been asserted that ‘‘functional’’
variation, where the variation is intrinsically related to the means or the
purpose of communication, as opposed to the cultural value of a
communicative context, is more characteristic of modern societies, while
‘‘nonfunctional’’ variation, in which the choice of variable lexical and
syntactic features is not determined by strictly pragmatic demands, is
more characteristic of premodern societies (Neustupny 1974: 39– 40). In
the process of modernization, the nonfunctional opposition between
classical and colloquial varieties in premodern diglossic speech com-
munities is replaced by the functional variation involved in an ever more
differentiated repertoire of scientific and technical varieties (Neustupny
1974: 40). With particular reference to diglossia, ‘‘most of the variation
between a Classical and a Modern standard in the case of a premodern
diglossia is non-functional in the described sense,’’ and it is, therefore,
‘‘not a matter of chance that the diglossia, a typical case of non-functional
20 A. Hudson
variation, is most often removed at an early stage of modernization’’
(Neustupny 1974: 40).
Ferguson himself assumes a polygenetic stance when it comes to the
emergence and development of diglossia: ‘‘Diglossia is not assumed to be
a stage which occurs always and only at a certain point in some kind
of evolution, _[but] may develop from various origins and eventuate
in different language situations’’ (Ferguson 1959: 326–327). If, as is argued
elsewhere in this paper, the critical distinction between the canonical
cases of diglossia and the more typical cases of societal bilingualism indeed
resides in the presence or absence of a prestige group of native H-speakers,
then it makes sense to look for the social origins of diglossia in circum-
stances where cultural–linguistic traditions develop or acquire new regis ters
without also acquiring native speakers of these new varieties who might
become prestige role models in their host speech communities and thereby
provide the social motivation for language or dialect shift. At least three
sets of circumstances suggest themselves as having this potential. In
the first, the H-variety in question historically has never been used as a
vernacular by any native speakers. In the second, the H-variety may have
been used as a native vernacular at one time, but the population of
speakers so using it has become extinct, whether through the process
of intergenerational language shift or internal language change. Third,
the H-variety still enjoys a vernacular speaker base in a separate speech
community, the members of which do not regard themselves, and are not
regarded as, members of the community that employs H for nonvernacular
purposes only.
Examples of these three sets of circumstances are not as difficult to
come by as might be supposed. Thus, although there is some controversy
surrounding the matter, some have contended that Classical Chinese
was an artificial language that was never spoken by anybody (Li and
Thompson 1982: 84; Rosemont 1974). Likewise, in Japan, kanbun was
nobody’s mother tongue, regardless of whether one’s native language was
Chinese or Japanese (Coulmas 1991: 130). Modern Standard Arabic
has also been characterized as a ‘‘somewhat artificial medium’’ (Kaye
1994: 49), which never had any native speakers (Kaye 1994: 51, 59). The
same might also be said of Old Church Slavic, which was never a
native vernacular but rather the written version of a Bulgaro-Macedonian
dialect of Slavic developed in the ninth century by the Christian
missionaries Constantine and Methodius (Comrie 1991: 160). Finally,
Katharevousa, the official standard of Greece prior to 1976, ‘‘was never
a spoken language of any historical period,’’ but was instead ‘‘an artificial
compromise between archaism and colloquialism, created in the early
nineteenth century and imposed as the official language of the Greek State
Focus article 21
in the first Greek constitution of 1834’’ (Alexiou 1982: 158). While every
standard language is of necessity to some extent artificial, the more typical
case, unlike diglossia, is that the written and spoken standards are more
firmly anchored in actual spoken varieties, often the careful spoken
varieties of the educated urban classes.
The second scenario alluded to above, the extinction of the population
of speakers upon whose vernacular the contemporary written standard
is historically based, in its most familiar guise, is the process whereby
change in written language proceeds at a significantly slower pace than
change in spoken language. The process in its most general form has been
described as follows: ‘‘The two varieties of language involved become
increasingly dissimilar, the variety of the original body of texts remaining
largely unchanged (although likely codified) while the spoken variety,
which does not undergo processes of linguistic standardization, continues
to change in ‘natural’ ways’’ (Walters 1996: 161). With particular
reference to Egyptian, Pulgram observes that ‘‘while the spoken language
of the masses goes along its undisturbed linear development, each classical
language, from its inception, is held to a level standard, without major
changes, as long as the society which employs it remains stable’’
(1950: 461). Historically, the widening discrepancy between Sanskrit
and the Prakrit dialects has been attributed to the fact that, ‘‘while
Prakrits went on changing, Sanskrit remained unchanging’’ (Deshpande
1991: 38). Literary Kannada, too, has been described as a frozen, archaic
stage of the language reached some centuries ago, while ‘‘the colloquial
dialects, which are spoken as everyone’s first language, continued to
change’ (Bright 1976: 66). In Persia, by the end of the Sassanid Empire
in the seventh century, literary Middle Persian ‘‘had become markedly
distinct from contemporary spoken Persian (dari) in Western Iran’’
´s 1984: 273). In the case of Modern Standard Arabic, ‘‘classicisms
of all shapes have worked to combine with the forces of linguistic
evolution to _slow down the rate of the linguistic change of this
somewhat artificial medium’’ (Kaye 1994: 49).
A general model of the acceleration of the vernacular away from the
standard, punctuated at socially significant intervals by the correction
of the latter back toward the former, has been proposed for the historical
relationship between spoken and classical Egyptian (Sethe 1925: 316),
for written and spoken Latin (Pulgram 1950: 462, and figure 2), and
for spoken and various varieties of literary Hebrew (Rendsburg 1990: 31,
174–175). Finally, whereas in the cases just discussed the vernacular might
be said to have been changed over time via replacement by its own
linguistic descendant, the case of Ge’ez and Amharic in Ethiopia suggests
that a vernacular may also be replaced by other than its own descendant
22 A. Hudson
and still result in a case of diglossia. Although Ge’ez was replaced as a
vernacular by Amharic and other languages between the ninth and the
twelfth centuries (Cooper 1978: 460), it survived as a literary language
into the nineteenth century, and, vestigially, as the liturgical language of
the Ethiopian Copts up to the present time (Cooper 1978: 461). While
Ge’ez and Amharic are related languages within the Ethio-Semitic group,
Ge’ez is not a direct linguistic ancestor of Amharic.
A third scenario for the emergence of diglossia is the acquisition of
an H-variety from an external speech community, unaccompanied by
any significant in-migration of native speakers of H. One of the most
interesting of these cases is the adoption of the Chinese language as the
means of writing in the earliest documents in Japan, dating from the early
seventh century (Coulmas 1991: 129). Although this written variety was
subsequently nativized and was ultimately to evolve into the literary
Japanese of the nineteenth and early twentieth centuries (Coulmas
1991: 128), it does not seem amiss to regard the differential functional
allocation of Chinese to the written sphere and of Japanese to the spoken
in the earliest stages of this development as an instance of diglossia
involving two unrelated languages.
In many of their particulars, the individual cases of diglossia referred
to in the course of this paper differ strikingly from one another. On the
surface of things, it might justifiably be asked whether, for example,
the opposition between Classical and vernacular Japanese up until
the beginning of the twentieth century should be considered of a piece
with the Swiss-German diglossia of today, or whether the traditional
opposition between literary and liturgical Ge’ez, on the one hand,
and vernacular Amharic, on the other, should be considered socio-
linguistically akin to that between Katharevousa and Dhimotikı
Greece prior to 1976, or to the centuries-old opposition between Hebrew
and Yiddish among Ashkenazic Jews. As Ferguson himself noted, ‘‘The
four cases I described are not identical; each one is quite different in
some respects from the other three, though they have many features
in common’’ (Ferguson 1991: 219). The realization, however, that the
essential characteristic of diglossia is the coexistence between an elevated
code, which has no native speakers within the speech community in
question, and an everyday vernacular introduces a higher order of
homogeneity into what on the surface may appear as heterogeneity. This
point of view liberates the notion of diglossia from the demand that the
two codes in question be varieties of the same language, although for
reasons discussed elsewhere, this is to be expected as the usual state of
affairs. It further removes the requirement that diglossia be limited
to literate speech communities, although, as discussed in the following
Focus article 23
paragraphs, the development of writing, when combined with long-term
restricted literacy, is particularly conducive to the emergence of diglossia.
The development of writing produces an unprecedented alteration
in the communicative ecology of any speech community, and although
it clearly does not in every instance lead to the emergence of diglossia,
it creates a particularly hospitable environment for it, when combined
with other supporting factors. As I have noted elsewhere (Hudson
1994: 308), four principal mechanisms account for the manner in which
the development of writing may contribute to the linguistic diversity of a
speech community, and therefore to the potential emergence of diglossia.
In the first instance, the sociocultural norms operative in contexts where
writing is appropriate commonly dictate that the grammatical structure
of written text be less casual and in some sense more elevated than
the grammatical structure of spoken utterances. Second, the different
pragmatic constraints imposed upon the realizations of written and
spoken text by the real-time limitations on production and processing
and by the physical, psychological, and social immediacy of the audience
determine to some degree the relative frequencies of occurrence of certain
grammatical forms and structures in the two modalities. Third, to the
extent that the opportunity for the acquisition and use of literacy skills
is asymmetrically allocated across the various segments of a speech
community for any significant period of time, the grammatical features
of written or literate text may come to signal membership in those,
typically more privileged, social groups within the community that have
access to the occasions on which written or literate text is appropriate.
Finally, writing as a medium acts as a fixative agent for linguistic
structure. Except for memorized oral text transmitted intact from gen-
eration to generation, the structure of oral language is in general much
more subject to diachronic change, due to phonetic erosion, than is the
structure of written language. To the extent, therefore, that written
language in general, or a particular genre of written language, is not
directly influenced by developments in the structure of oral language, the
potential arises for major disparity between the two and, therefore, for the
emergence of diglossia.
Coulmas has pointed to the need for ‘‘a sociolinguistic theory of writing
and written language which accounts for the nexus between literacy,
writing system, and diglossia’’ (1987: 122). While allowing that writing
may be neither a necessary nor a sufficient condition for the emergence
of diglossia, he maintains, along with Ferguson, that ‘‘writing introduces
the possibility of a permanent rift between characteristically different
varieties into every speech community’’ (Coulmas 1987: 122), thereby
leading to ‘‘potential diglossia in every speech community that becomes
24 A. Hudson
literate’’ (Coulmas 1987: 114). As Ferguson, too, has noted, graphization,
or the reduction of a language to writing, ‘‘adds another variety of
language to the community’s repertory’’ (1968: 29), and ‘‘communities,
as they begin the regular use of writing, generally do not feel that
ordinary, everyday speech is appropriate for written use’’ (1968: 29–30).
In nineteenth-century Japan, for example, this view was succinctly
expressed as follows by one of the leading intellectuals of the time: ‘‘In
our letters at present it is inappropriate to write as we speak as well
as inappropriate to speak as we write, because the grammars of speech
and writing in our tongue are different’’ (Nishi 1874, quoted in Coulmas
1991: 128). So pervasive was this view that Japanese scholars and
government bureaucrats travelling to the Netherlands at the time were
surprised to find that the language of writing there was very much closer
to the language of everyday conversation than was the case with Japanese,
‘‘because the idea of committing anything of serious import to writing in
a colloquial variety was alien to them’’ (Coulmas 1991: 134).
According to Ferguson, diglossia is likely to emerge as a consequence
of the long-term monopoly of a small elite on literacy and, therefore,
on direct access to the literary heritage of a speech community (1959: 338).
In particular, diglossia is likely to come into existence where (1) ‘‘there
is a sizable body of literature in a language closely related to (or even
identical with) the natural language of the community, and this literature
embodies _some of the fundamental values of the community,’’
(2) ‘‘literacy in the community is limited to a small elite,’’ and (3) ‘‘a
suitable period of time, on the order of several centuries, passes from
the establishment of (1) and (2)’’ (Ferguson 1959: 338). In Coulmas’s
version of this theory, alluded to above, the degree of difference between
the written and spoken norms, and therefore the prospect for the
emergence of diglossia, is a function of (1) the degree of association
between the literary tradition and other great cultural achievements of
a religious or artistic nature, (2) the period of time over which the literary
tradition flourishes, (3) the extent to which the written language is
cultivated by a small caste of scribes and is prevented from adjusting to
changes in the spoken language, (4) the literacy rate of the speech
community, (5) the acceptance of the written language by the illiterate
mass of the population as the only valid manifestation of their language,
and (6) the relative fit of writing system and orthography for the language
in question (1987: 121–122, 1989: 13). Others, too, have made the con-
nection between the extent of popular literacy and the likely distance
between the written and the spoken codes: Parker supposes that ‘‘the
varying and uncertain distance between spoken language and written
language in any time and place _may be inversely associated with the
Focus article 25
level and extent of literacy’’ (1983: 334 –335). As explained by Walters,
‘‘the existence of diglossia hinges on a tradition of restricted literacy
involving the written variety of a language that becomes increasingly
distant (and therefore distinct) from the native variety of language spoken
in a speech community that is overwhelmingly illiterate’’ (Walters
1996: 161–162). While Walters’s account may give the impression that
this process is the only one that gives rise to diglossia, Ferguson, as noted
elsewhere, claims that ‘‘diglossia may develop from various origins’’
(1959: 327). Of these various origins, however, it is likely that the pro-
cesses described by Coulmas and Walters are indeed those that most
typically give rise to diglossia.
The association between diglossia and literacy, in the sense of both the
incidence of individual literacy skills and the existence of a literary
tradition, is not an accidental one, since social stratification of literacy
may give rise to the independent development of two or more increas-
ingly divergent language varieties. As the linguistic varieties become
more divergent, more extensive training is required for the mastery of
the literary variety, a development that in turn confers additional
prestige upon it: ‘‘With the development of writing and a complex and
introspective literature, the language variety so employed will often be
accorded such high value because of the recorded nature of the medium
and the need to be trained to read and write it’’ (Abrahams 1972: 15).
The creation of literary varieties may in turn serve to reinforce the
exclusiveness of a class structure based on literacy or, more precisely,
on the control of knowledge via literacy. In preindustrial civilized
societies, traditional writings ‘‘have provided the basis for standardizing
thought and action for the literati who are the leaders in the society
over time and space’’ (Sjoberg 1964: 893). Thus, ‘‘a substantial body of
writing furnishes a language with a depersonalized standard that can be
fixed and, to some extent, manipulated, that is, cultivated in a conscious
way’’ (Coulmas 1987: 121).
A handful of cases serves to illustrate the intimate relationship between
literacy, literary tradition, restricted literacy, and diglossia. Arabic
diglossia in large measure has derived its stability from the fact that
the H-variety is associated with pre-Islamic poetry, the sacred texts of
Islam, later commentary on both of these, and the works of medieval
Arabic philologists, among them Sibawaihi of Basra, who based their
linguistic analyses on these earlier texts (Rabin 1955: 20; Walters
1996: 161). In the case of Chinese, the linguistic differences between the
literati and the general population were probably due to the lengthy
accumulation of literary tradition among the former, to the early fixation
of the form of writing by the printing press in China (printing from wood
26 A. Hudson
blocks since at least the beginning of the eighth century and from movable
type since the eleventh [Morton 1982: 86]), and extremely low literacy
(DeFrancis 1972: 8). Vernacular Hebrew is generally not attested in the
Biblical literature, particularly of the First Temple period, ‘‘most likely
due to the influence of the prophetic schools _which acted as a check or
brake on the written language’’ (Fellman 1977: 108). The maintenance of
Latin as a prestige language vis-a
`-vis the emergent Romance vernaculars
has been attributed to a ‘‘virtual monopoly of knowledge’’ on the part
of the Church, acquired as a result of the emphasis on clerical literacy,
monastic control of manuscript production and reproduction, and the
decline in literacy among the laity between the fourth and the tenth
centuries (Parker 1983: 336 –337). In the case of Sanskrit, it was the
traditional grammarians’ concern with correct ritual speech that led to
the development of grammars such as that written by Panini (Hock
and Pandharipande 1978: 21–22), and, in general, over the course of the
linguistic history of India, Sanskrit has retained its high status ‘‘primarily
because literacy was highly restricted to the Brahmanical classes’’
(Deshpande 1991: 38). Finally, the separation of court and literary
Turkish from the vernacular, which began during the sixteenth century,
was accelerated by the social cleavage between the vast ruling order
composed of bureaucrats, poets, the religious establishment, merchants,
and other community leaders on the one hand, and the masses on the
other, as well as by the lack of an educational system, which ‘‘prevented
the dissemination of high class Turkish among the masses’’ (Karpat
1984: 189).
While it is clear that the development of writing is not in itself a
sufficient condition for the emergence of diglossia, the question arises
as to whether it is a necessary one. Ferguson himself has suggested the
possibility that it is not: ‘‘All clearly documented instances [of diglossia]
known to me are in literate communities, but it seems at least possible that
a somewhat similar situation could exist in a non-literate community
where a body of oral literature could play the same role as the body of
written literature in the examples cited’’ (Ferguson 1959: 337, note 18).
The distinction between written and spoken language in literate traditions
has been recognized as resembling the distinction between the high and
the low varieties in diglossia (see Tannen 1982: 15, note 3), and the point
has frequently been made also that ritual or artful genres of expression
in oral traditions differ in linguistic form and social function from every-
day talk in much the same way that written and other noncasual types
of discourse differ from colloquial speech in literate cultures (Akinnaso
1982: 8; Chafe 1982: 49–50; Feldman 1991: 47– 48). Thus, Akinnaso
claims that ‘‘in nonliterate (traditional) societies, ritual communication
Focus article 27
(for example, ritual chants and divination verses) is different from every-
day talk in much the same way that written language differs from ordinary
conversational language in literate societies’’ (1982: 8), or, more explicitly,
that ‘‘there are no formal, functional, or structural differences between
written language in literate societies and oral ritual communication in
nonliterate societies’’ (1982: 29, note 3). Others, such as Chafe (1982:
49–50), have demonstrated that the same features that differentiate
spoken and written language may also differentiate colloquial and ritual
speech. Also, no less than written codes, ‘‘any speaking code used cere-
moniously will accumulate the sense of power inherent in the occasions
of its use’’ (Abrahams 1972: 15), and, in terms of functional allocation,
opportunity for acquisition, and potential for linguistic divergence from
the vernacular, will also offer the prospect of the emergence of diglossia.
More than the development of writing as such, then, it seems that an
important precursor to diglossia is the existence of a body of literature
in the sense of ‘‘that body of discourses or texts which, within any society,
is considered worthy of dissemination, transmission, and preservation in
essentially constant form’’ (Bright 1982: 272; see also Martin Joos, cited
in Voegelin 1960: 60, note 4). Such a body of literature, though typically
associated with the written medium, ‘‘may also be composed orally, and
regularly performed in that same medium’’ (Bright 1982: 171). The oral
literature of ancient India offers a remarkable case in point. The need to
preserve the religious effectiveness of the orally composed Vedic hymns
by transmission in their exact original form, and the desire to preserve the
special status and distinctness of the language of the educated classes,
culminated in the Sanskrit grammar of Panini (c. 500 BC), itself in all
likelihood an oral composition also (Bright 1982: 273; Hock and
Pandharipande 1978: 21–22).
Rise and decline of diglossia
Diglossia has been described by Ferguson as being very stable: ‘‘Diglossia
typically persists at least several centuries, and evidence in some cases
seems to show that it can last well over a thousand years’’ (Ferguson
1959: 332). Coulmas, too, sees stability as ‘‘one of the most remarkable
characteristics of diglossia’’ (1987: 117), particularly where ‘‘linguistic
differences are not aggravated by political or religious differences’’
(1987: 118). Others, however, have questioned whether ‘‘the stability
factor is so crucial to diglossia _since societal changes such as
modernisation, urbanisation, the breakdown of rigid class barriers, etc.,
have made diglossic situations as described by Ferguson (1959) rather
28 A. Hudson
rare’’ (Pauwels 1986: 16). Finally, as if to proclaim the entire discussion
moot, Mackey has declared, ‘‘contrairement aux mode
`les de Ferguson
et de Gumperz, les situations diglossiques ne sont pas stables; elles ont
chacune leur dynamique’’ (1989: 16). This small sampling of opinion as to
the stability of diglossic situations points as clearly as any discussion
could to the general failure of sociolinguistic theory in dealing with
diglossia as a distinct communicative arrangement. For further discussion
of this and related points, see Hudson-Edwards (1984: 6).
Scholars clearly reach differing conclusions on this subject for a variety
of reasons: their characterizations of the facts with which they are dealing
differ; they treat different sets of facts as though they were the same; they
differ as to their interpretations and definitions of stability; or, finally,
they differ as to their understandings of the universe of sociolinguistic
situations that might be regarded as diglossic. It is most likely this last,
combined perhaps with inevitable differences of opinion as to what in
fact constitutes stability in the first place, that accounts for the varying
perspectives on the subject of the long-term survivability of diglossic
arrangements. This particular conundrum corresponds very closely, in
fact, to what Labov refers to as a ‘‘paradox of evidence’’ (1994: 14). Such
dilemmas may perhaps be resolved by reexamination of the internal
evidence for the validity of the conflicting viewpoints, or by bringing
evidence to bear ‘‘from neighboring fields of expertise _which may show
that the contradictions were really the result of a lack of homogeneity
in the object being studied’’ (Labov 1994: 14). It is past time, indeed, for
such a reconciliation of perspectives in order to advance understanding of
the essence of diglossia and of its natural course of evolution, and thereby
to advance a significant part of sociolinguistic theory more generally.
In the end, however, the notion of stability is at worst ill-conceived
and at best relative, and whether diglossic situations are in any sense
more stable than other types of sociolinguistic situations, particularly
those involving two or more distinct languages, remains to be determined.
Some cases of societal bilingualism will almost certainly be found to have
endured for longer periods of time than some cases of so-called diglossia,
and, on the other hand, comparing instances of long-term diglossia
with instances of short-term bilingualism makes no case for the relative
stability of diglossia. In the end, too, even the most persistent cases of
diglossia eventually show change in the functional distribution of codes,
as well as in the linguistic substance of the codes themselves, thus evolving
into new and different types of speech economies. Whereas the socio-
linguistic situation in German-speaking Switzerland has been described
as being ‘‘relatively stable’’ (Keller 1982: 75), at least with regard to the
functional status of the two varieties involved if not with regard to their
Focus article 29
substance (Keller 1982: 76–77), Arabic diglossia has been described as
‘‘not a relatively stable situation’’ (Kaye 1970: 391), and, in the case of
Greek, as of the late 1970s at least, ‘‘functional and situational differences
between High and Low are far less stable and clear-cut than in Swiss
German or Arabic’’ (Alexiou 1982: 165).
What seems closer to the theoretical mark than stability, in fact, is
the likely direction of change once the inevitable realignment of codes and
their social functions finally occurs. In cases of societal bilingualism, the
general tendency appears to be for the higher-prestige language eventually
to invade the domain of the home, ultimately displacing the language of
lesser prestige as a first language in the community. This diachronic
process has been documented in strikingly similar terms for Nahuatl–
Spanish language contact by Hill (1983) and for French–Occitan language
contact by Eckert (1980), but also in innumerable other sources for many
other cases throughout the sociolinguistic literature. It is also the situation
described by Fishman under the designation of ‘‘bilingualism without
diglossia,’’ the outcome of which he describes as follows: ‘‘Without
separate though complementary norms and values to establish and
maintain functional separation of the speech varieties, that language or
variety which is fortunate enough to be associated with the predominant
drift of social forces tends to displace the other(s)’’ (1967: 36, see also
1980: 9). In diglossic contexts, on the other hand, it is H that tends to be
displaced by L through a process of structural convergence resulting in
the emergence of a new standard more closely related to certain educated
varieties of the vernacular. On the very face of it, this outcome, which
appears to be nonrandomly, though not categorically, associated with
diglossia in Ferguson’s narrow sense, reveals the fundamental socio-
linguistic distinction between true diglossia situations and situations of
societal bilingualism, including those cases of pseudodiglossia, such as
´n–Castillian, Frisian–Dutch, Galician–Portuguese, and Occitan–
French, where the appearance is presented of relatively close structural
affinity between the codes in question.
The potential decline of diglossia may be anticipated by the admission
of the vernacular into domains formerly reserved exclusively for the high
variety, and it is reasonable to ask, therefore, whether the course of
functional expansion of L follows a fixed pattern once instability in the
functional assignments of H and L has set in. Repeatedly, where popular
command of H has declined, this has provided the impetus for the
adoption of L in religious or secular texts, whether written or oral, aimed
primarily at the largely illiterate populace. In India, by the second century
BC, ‘‘the Prakrit languages were widely used for secular literature, royal
inscriptions, and especially for the propagation of _Buddhism and
30 A. Hudson
Jainism’’ (Deshpande 1991: 24). In China, too, literature intended to
entertain or edify the uneducated masses had been written in gudai baihua,
‘‘a vernacular Chinese close to the spoken language,’’ since the Tang
dynasty (618–907) (Peyraube 1991: 110; see Barnes 1982: 262), while in
modern times, a purportedly declassicized Chinese had been mandated
for use in political propaganda and other literature intended for popular
consumption in the People’s Republic of China (Li and Thompson
1982: 85). Hall reports that the authorities of the Western Church had
become so concerned over the failure of the common people to under-
stand the Latin sermons delivered at Mass that the Council of Tours
in 813 decreed ‘‘that priests should thenceforth translate their sermons
in rusticam romanam linguam aut theotiscam ‘into the rustic Roman or
German tongue’ ’’ (1974: 105; see also Parker 1983: 336). As Pulgram
aptly states, ‘‘what lies behind the reform is a resigned recognition that
you cannot talk to people in a language they have long since ceased to
understand, nor thereby save their souls’’ (1950: 461). In thirteenth-
century Japan, the use of hentai kanbun for official documents during
the Kamakura shogunate signalled a declining command of jun-kanbun
among the military elite (Coulmas 1991: 129–130). In the sixteenth
century, European Jesuit missionaries in Ethiopia used Amharic rather
than Ge’ez for written purposes and thus were able to reach a great
proportion of the Ethiopian populace ‘‘for whom Ge’ez remained a sealed
priestly mystery’’ (Fellman 1975b: 180).
Other factors, too, motivate the spread of L. In Tamil, the realistic
portrayal of natural conversation between the characters in novels,
plays, and films introduced the use of Tamil L into a domain that
had formerly been the exclusive preserve of Tamil H (Britto 1991: 69–70),
and, as a result of this, Tamil L came to be regarded as appropriate for
all monologic discourse, such as sermons, political speeches, and
lectures, where the speaker wished to create the impression of ‘‘actually
conversing with the audience’’ (Britto 1991: 70). Medieval and modern
Romance standard languages, as opposed to Latin, ‘‘were typically
first used, in the Middle Ages, for compositions which did not threaten
established intellectual or professional interests’’ as, for example,
the recitation of epics, the singing of troubadours, and romances, while
‘‘the use of the vernacular in fields such as law, medicine or religion _
was strongly resisted’’ (Hall 1978: 111). The spread of vernacular
Greek from poetry to nonfictional prose and thence to creative literature
more broadly, from the latter to serious critical writing, and thence
to scholarly, technical, scientific, and official writing has also been
well documented (Alexiou 1982: 159; Mirambel 1964 [cited in Alexiou
1982: 174]).
Focus article 31
According to Ferguson, certain social developments betoken the
eventual decline of diglossia. These developments may be broadly char-
acterized as tendencies toward mobilization and democratization and
are described by Ferguson as follows: (1) more widespread literacy,
(2) broader communication among different regional and social segments
of the community, and (3) the desire for a standard national language
as an attribute of autonomy or sovereignty (1959: 338). Enlarging some-
what upon this list, the decline of diglossia is repeatedly (though not
inevitably, as the case of Switzerland clearly indicates) associated with the
processes of modernization, urbanization, mercantilism, and industrial-
ization, and the demands that these create for a literate labor force.
Related to these forces also are the disestablishment of small ruling
groups, the breakdown of rigid class barriers and increased fluidity of
role relationships, and the democratization of education, literacy, and
knowledge that tend to accompany these. Finally, nationalism, in the
form of the desire for autonomy, sovereignty, and political unity, often
engenders the desire for, and loyalty to, a standard national language.
Such trends help forge ‘‘firmer bonds between the spoken and written
forms of the language,’’ so that it becomes ‘‘increasingly difficult to
separate educated from non-educated speech’’ (Sjoberg 1964: 894).
Modernization and its various components, culminating in mass
education and literacy, have been repeatedly associated with the decline
of diglossia. It has been argued, for instance, that the elevation of
the status of the vernacular, and the resulting adjustment of the writing
system to this, is one of the main types of linguistic reform in modern-
izing societies (Sjoberg 1964: 897). It has also been suggested, as noted
previously, that diglossia ‘‘is most often removed at an early stage of
modernization’’ (Neustupny 1974: 40). Further, the urbanization of
intermediate societies is reported to result in structural convergence
between the subcodes in their speech repertoires, to the point where, in
some highly urbanized communities, the discrepancy between standard
and local dialects diminishes virtually to the point of extinction (Gumperz
1968 [1962]: 469).
Examples abound. The perceptible growth in urban commercial
economic activity by the thirteenth century was accompanied by
‘‘a significant increase in the demand for literate individuals to meet
the growing requirements of administration, law, and commerce’’ (Parker
1983: 338), and, in turn, by the emergence of lay schools in many of
which ‘‘at least some of the instruction was in the vernacular’’ (Parker
1983: 338–339). In the eighteenth century, partly motivated by the
modernizing efforts of Peter the Great (reigned 1682–1725), there
emerged a growing desire in intellectual circles to bring the linguistic
32 A. Hudson
situation in Russia more into conformity with that in France, particularly
with regard to the perceived uniformity of the written and spoken
language, thereby jeopardizing the status of Old Church Slavic (Comrie
1991: 168). In the course of the transformation of Japanese society from
a feudal into a capitalist society, the need to translate Western books,
the desire to disseminate knowledge broadly throughout the popula-
tion, and the recognition of the importance to industrial and military
development of a unified written standard language that was close to the
spoken language all led to a restructuring of the relationship between
the written and the spoken language such that diglossia between Classical
and colloquial Japanese was essentially eliminated by the end of World
War II (Coulmas 1991: 135, 140 –141; Neustupny 1974: 36). With the
entry of new administrative, legal, educational, and military terms into the
Ottoman world during the nineteenth century, ‘‘the Ottomans began to
realize that their language [Osmanlica] was insufficient to the task of
keeping up with European advances’’ (Gallagher 1971: 161), and that
‘‘Ottoman linguistic obscurantism was a barrier to the political reform
and increased freedom they were seeking’’ (Gallagher 1971: 162). Further,
a major stimulus for the development of Modern Standard Arabic was
the discussion of modern philosophical, literary, and historical topics
that took place in various newspapers and journals in Cairo, Beirut, and
Baghdad, as Western-influenced Arab intellectuals found Classical Arabic
unable to cope with such topics as rationalism, liberalism, socialism,
democracy, and nationalism, thus creating the problem of ‘‘developing
within a few decades _a new and refined Arabic capable of expressing
a material and intellectual civilization that had evolved over centuries
in Europe’’ (Abdulaziz 1986: 16).
In conclusion, however, it should be noted that these and/or similar
processes, while seemingly contributing to the decline of diglossia, do
not always and without exception favor the elevation of the low variety to
high status. Thus, the emergence of Standard German at the expense
of the local dialects in Germany over a period of more than 500 years
may be attributed to the invention of the printing press, Luther’s trans-
lation of the Bible into Eastern Middle German, the gradual acceptance of
Luther’s German in the Catholic south, political unification in 1871, and
universal education in the twentieth century (Mackenzie 1994: 249).
The complex web of processes entailed in the establishment of a new
social order inevitably poses a threat to the maintenance of diglossia.
It has been claimed that verbal repertoires become more homogeneous
and that old administrative codes tend to be replaced by more vernacular
varieties as local populations are absorbed into dominant groups and
as increasing proportions of the population are drawn into national
Focus article 33
life (Gumperz 1968 [1962]: 469). Like prestige languages generally, the
H-variety in diglossia ‘‘is seriously weakened by a new class structure
which gives increased power to groups previously at the margins of,
or below the range of, elite society’’ (Kahane and Kahane 1979: 190).
Also like prestige languages, the H-variety ‘‘comes in with status and
elitism _[and] goes out under the pressures of popular developments
and movements which we may call nativist rebellions’’ (Kahane
1986: 498). In fact, the disestablishment of ruling elites has been
portrayed as a necessary condition for the elimination of diglossia:
‘‘Unless the ruling class is replaced by another, there is no loosening
of the diglossic control’’ (Sotiropoulos 1982: 19). As class barriers
crumble and formerly well-defined social strata are less sharply dif-
ferentiated, it becomes less clear ‘‘who sets the standards for either the
written or the spoken forms of the language’’ (Sjoberg 1964: 894). The
general principle appears to be that the code matrix becomes less and less
diverse as role distinctness decreases (Gumperz 1968 [1962]: 469), and
‘‘that shallow linguistic contrast in styles is a direct correlate of the fluidity
of roles symbolized by the distinction between caste and class’’ (Gumperz
1968 [1962]: 470).
Decline of a classical variety is often accompanied by catastrophic
political events involving the breakdown of classical society itself:
‘‘A breakdown of this society involves the breakdown of its classical
language; and the new socio-historical structure creates a new literary
language out of the spoken language then current’’ (Pulgram 1950:
461–462). In India, between 500 BC and the turn of the millennium,
the rise of anti-Vedic religions such as Buddhism and Jainism, both
founded by warrior-princes who defied the traditional religious authority
of the Brahmins, resulted in the elevation of local vernacular Prakrit
languages to the status of languages worthy for religious instruction
(Deshpande 1991: 25–26). In the case of Hebrew, it was with the success
of the popular-based Maccabaean revolution in 167 BCE that the
vernacular began to rise in importance, later to come into its own with the
compilation and codification of the Oral Law in the Mishna and related
works, roughly between 100 BCE and 200 AD (Fellman 1977: 108). As
described by Rabin, the Maccabaean revolution ‘‘presents all those
conditions in which we might expect a change-over of linguistic habits:
a protracted period of cultural anarchy and dislocation of education on
the one hand, and the rise of a new intellectual elite to take the place of
a thoroughly discredited upper class on the other’’ (Rabin 1958: 156–157
[quoted in Fellman 1977: 108]). In Ethiopia, it was the influence of the
clergy that sustained the position of Ge’ez vis-a
`-vis Amharic; when
the clergy lost its political sway, as during the decline of the Axumite
34 A. Hudson
Empire in the tenth century, and again following the political and eco-
nomic reforms of Menelik II in the late nineteenth and twentieth centuries,
Ge’ez lost its superior position relative to Amharic (Fellman 1975b).
In Europe, the decline in the relative importance of the clerical orders
in secular administration contributed significantly to the weakening of
the Church’s monopoly of knowledge by the fourteenth century, and
probably ‘‘reduced barriers to, and ultimately increased the demand for,
subsequent introduction of the vernaculars in certain spheres of official
language use, and thence generally’’ (Parker 1983: 339). Conversely,
among other considerations, the growth of an educated European public,
particularly in the fifteenth and sixteenth centuries, ‘‘replaced the
medieval class system of the learned elite and the non-educated masses,
symbolized by the dichotomy of Latin vs. the vernacular, with a system of
the new, educated (upper-?) middle class and the non-educated masses,
symbolized by the dichotomy of the new standard language vs. the
vernacular’’ (Kahane and Kahane 1979: 188). In Japan, the collapse of
the shogunate in the 1860s and the subsequent opening of the Japanese
ports led to the influx of cultural influence from the West and thus
provided the initial stimulus for the ultimate unification of the written
and spoken languages (Coulmas 1991: 134 –135). In the case of Chinese
language reform, it was only when the old social order of the Qing
Dynasty began to crumble that the Chinese script was subjected to critical
scrutiny (DeFrancis 1972: 10), and the eventual victory of the vernacular
baihua movement appears to have been associated with the replacement
of the traditional scholar-bureaucracy by a republican form of govern-
ment in 1911 (Barnes 1982: 261). The decline of Katharevousa in Greece,
foreshadowed in the rise of ‘‘a new and pragmatic middle class’’ toward
the end of the nineteenth century, was to reach its climax, after more
than a century of political vicissitudes, with the ouster of the military
junta and the restoration of official democracy in 1974, and with the
reestablishment of Dhimotikı
´as the national language two years later
(Kahane and Kahane 1979: 190). The emergence of Educated Standard
Arabic has been attributed to the rise of a new elite with different com-
municative needs from those of the 1950s and 1960s, who rejected Modern
Standard Arabic ‘‘either because they did not master it, or because it
was not modernised enough to adequately express the technological
and educational advances that _[had] shaped their lives’’ (Mahmoud
1986: 245).
The various processes of modernization as well as the replacement of
older elites or traditional social structures by newer ones often converge,
at least in modern times, upon the phenomenon of democratization, in its
narrower political sense as well as in the sense of the democratization of
Focus article 35
education, literacy, and knowledge. Thus, the language unification, or
genbun itchi, movement in Japan was not only stimulated by the collapse
of the shogunate and by the modernization of Japanese society, as noted
above, but, as an aspect of the latter, coincided with the people’s rights
movement (minken jiy
uu und
oo) and the movement for a popularly elected
assembly (minsen giin und
oo). As a result, ‘‘the new colloquial standard
which the proponents of genbun itchi created and which the society as
a whole accepted, _is to be understood as the linguistic counterpart of
the societal changes heralded by these movements’’ (Coulmas 1991: 139).
Within the realms of education and literacy more specifically, the higher-
than-average rates of both in Japan have ‘‘turned the written norm into
a variety that also serves as a supradialectal speech form’’ (Coulmas
1991: 126, also 126–127).
The decline of Latin at the end of the medieval period may be attributed
to ‘‘the broadened participation of the middle classes in a general,
vernacular-based, education’’ (Kahane 1986: 498499). Associated with
the official vernacularization of the Chinese literary language were the
initiation of public education in 1903 and the abolition of the Confucian
civil service examination system in 1905 (Barnes 1982: 262). Likewise,
across the Arabic-speaking world, the massive growth in educational
participation by people from all walks of life has brought about the end
of a tradition of restricted literacy, more widespread access to Classical
Arabic or Modern Standard Arabic on the part of the general population,
and a measure of stylistic differentiation of Arabic appropriate to modern
communicative needs (Abdulaziz 1986: 13–14; Ibrahim and Jernudd
1986: 6; Walters 1996: 166). Finally, the diglossic situation in Greece,
which persisted until the 1976 reform, originated in the last two decades
of the nineteenth century and the first two of the twentieth, during which
period important social and economic changes marked the economy’s
passage into capitalism and the definitive taking of power by the
bourgeois forces (Frangoudaki 1992: 367–368). ‘‘The demoticists
represented at the time economic and social progress, industrialization,
a free parliament, and aggressive nationalism,’’ and demoticist leaders
fought for compulsory education, technical and vocational education,
and the linguistic integration of minorities (Frangoudaki 1992: 368).
Katharevousa, on the other hand, was defended by the precapitalist
social forces of the time as the high variety, the language of the educated
minority, used in all formal contexts (Frangoudaki 1992: 368). The use
of demotic in formal contexts carried strong connotations of revolt and
social unrest (Frangoudaki 1992: 368).
Closely related to the effect of mass literacy is that of the mass media.
Throughout the Arabic-speaking lands, radio, television, and cinema in
36 A. Hudson
the wealthier states, and even newspapers and magazines, which are
often transmitted orally to the illiterate, ‘‘have greatly helped to spread
the knowledge of MSA and the urban forms of spoken Arabic to such
an extent that it is claimed that even the peasants in places like Egypt,
Syria, and Iraq can ‘comprehend’ news in MSA’’ (Abdulaziz 1986: 15–16).
In Japan, the mass media, as well as universal education, have worked
to ensure that ‘‘the entire population has access to the standard’’
(Coulmas 1991: 126–127). According to one source, the spread of literacy,
the emergence of the mass media, and the growth of modern literature
in the vernacular variety have slowly but steadily eroded the social
prestige and functions of the elevated variety in Telugu, resulting in
the gradual elimination of diglossia (Radhakrishna 1980: 238).
The relationship between national identity, language loyalty, and
diglossia is hinted at in Gumperz’s typology of linguistic communities.
Whereas in intermediate societies — those most likely to develop some
form of diglossia — language loyalty may attach to ‘‘codes which may
be quite distinct from the vernacular’’ (Gumperz 1968 [1962]: 469), in
highly urbanized communities ‘‘language loyalty _is bestowed on the
standard, which now closely reflects the majority speech’’ (Gumperz 1968
[1962]: 470). The Kahanes formulate the relationship as follows: ‘‘Where
the dichotomy of prestige language and vernacular exists, national pride
is symbolized by the latter if the former is not deeply rooted in a
prestigious tradition’’ (Kahane and Kahane 1979: 192). Thus, the decline
of Medieval Latin was linked to the rise of the modern European nation-
states, as loyalty to these was symbolized by consciousness of and loyalty
to their respective evolving standard languages (Kahane 1986: 499;
Kahane and Kahane 1979: 193). The elevation of vernacular Armenian
to the status of a new standard literary language in place of the earlier
standard, which was based on a religious argot and was incomprehensible
to the largely illiterate peasant population, was accomplished during
the latter half of the nineteenth century in the face of resistance from the
Armenian Apostolic church and its conservative supporters (Nercissians
1987: 635–636). While national conservatives warned against under-
mining the role of the church in national affairs, liberal and progressive
thinkers of the day recognized that ‘‘in the absence of the other factors
which unify a nation, language must replace religious affiliation as
the main national identity marker’’ (Nercissians 1987: 636). The ultimate
success of the vernacular language movement in China ‘‘was made
possible by its identification with the cause of Chinese nationalism’’
(DeFrancis 1972: 11), or, more specifically, by the replacement of the
premodern Sinocentric view of Chinese domestic institutions and inter-
national relations with a national consciousness ‘‘committed to the
Focus article 37
eradication of political and territorial encroachments and to the
organization of a modern nation-state capable of guaranteeing China’s
sovereignty’’ (Barnes 1982: 261). In Ethiopia, at the turn of the twentieth
century, Menelik II succeeded in making Amharic ‘‘the linguistic symbol
of national unity and sovereignty,’’ relegating Ge’ez to liturgical and
ceremonial functions only (Fellman 1975b: 181–182). Across the Arab
world, in recent decades, ‘‘governments and pan-national organizations
like ALESCO have sought _to promote the use of CA/MSA [Classical
Arabic/Modern Standard Arabic] as a unifying tool and symbol across
the Arab world’’ (Walters 1996: 162–163).
According to Ferguson, H is likely to remain as the standard ‘‘only if
it is already serving as a standard language in some other community
and the diglossia community, for reasons linguistic and non-linguistic,
tends to merge with the other community’’ (1959: 339). As noted pre-
viously, however, change toward full utilization of H cannot be expected
to take place without a radical change in the diglossic pattern of first-
language acquisition, which is to say that shift to H is unlikely unless
parents evince a willingness to speak H to their children (Ferguson
1959: 331). Absent these conditions, ‘‘H fades away and becomes a learned
or liturgical language studied only by scholars or specialists and not
used actively in the community’’ (Ferguson 1959: 339). Such a description
of things seems to be incontrovertibly the case for Classical Chinese,
Egyptian, Ge’ez, Classical Irish, Japanese, Latin, Old Church Slavic,
Sanskrit, and Classical Welsh. In other cases, however, it may not be
appropriate to speak of the complete displacement of the H variety by
L, but rather of the emergence of a third norm, ‘‘which represents a
merger of the original two norms’’ (Wexler 1971: 345–346, note 22). In
Ferguson’s terms, ‘‘if there is a single communication center in the whole
speech community, or if there are several such centers all in one dialect
area, the L variety of the center(s) will be the basis of the new standard,
whether relatively pure L or considerably mixed with H’’ (1959: 339). This
scenario seems closer to the facts in the cases of Arabic, Greek, Sinhala,
and Turkish, and perhaps also in those of Tamil and the other Dravidian
languages affected to any degree by diglossia. According to Britto, ‘‘most
Tamilists say that there exists, or is evolving, a Standard Colloquial,
which is based on the speech of the educated, non-Brahmin, middle class,
and which has currency all over Tamil Nadu’’ (Britto 1991: 65). In the
case of Turkish, ‘‘Ottoman mandarin style’’ was completely moribund by
the end of the war in 1918, and, in its place, ‘‘there was coming into being
a flexible, living language arising in good part from the spoken language
of the educated classes of Istanbul and the larger cities’’ (Gallagher
1971: 163).
38 A. Hudson
Whether H is completely replaced or merges with the vernacular to
produce a new standard, the lexicon, like those of prestige languages
generally, survives in the new standard in the form of a large-scale transfer
of terminology characteristic of upper-class civilization, abstractions, and
professional technologies (Kahane and Kahane 1979: 194). The process
has been similarly formulated by other students of diglossia: ‘‘The
phonological, morphological, and syntactic structures of the new lan-
guage prevail. However, one element of the H form that survives in _the
new standard, is the lexicon, especially words for abstract notions and
professional and scientific terminologies’’ (Sotiropoulos 1982: 19). This is
evident in the pervasive influence of Russian Church Slavic on Modern
Standard Russian, where ‘‘Slavonisms tend to concentrate in religious
vocabulary and in more learned and abstract spheres, while vernacular
elements tend to concentrate in more concrete and everyday vocabulary’’
(Comrie 1991: 169; see pp. 169–171 for examples). It is evident too in the
imprint of Katharevousa upon the emerging Demotic standard in Greece:
‘‘Puristic is receding, but its lexicon of intellectualism, science, and
technology has been channeled into the evolving standard language of
the educated’’ (Kahane and Kahane 1979: 195). In each of the Arabic-
speaking countries, ‘‘a spoken variety of growing prestige that takes the
dialect as the matrix or basis and borrows lexicon, set expressions, and
discourse markers _from CA/MSA [Classical Arabic/Modern Standard
Arabic] exists, and its use continues to expand even to the extent that it
serves s a spoken standard of some sort’’ (Walters 1996: 169). Last, in the
case of Bengali in India, although the apparently diglossic relationship that
has existed between sadhu bhasa (H) and colit bhasa (L) is now virtu-
ally dissolved, the lexicon of L ‘‘has undergone a degree of convergence
with the norms of the H, as more and more lexemes formerly confined
to H are incorporated into everyday usage in L’’ (Klaiman 1993: 162).
In a final scenario, Ferguson proposes that ‘‘if there are several _
[communication] centers in different dialect areas with no one center
paramount, then it is likely that several L varieties will become standard as
separate languages’ (1959: 339). The rise of the vernacular-based Romance
standards in Europe and of the Apabhramsa languages in northern
India vis-a
`-vis their classical ancestors, Latin and Sanskrit, respectively,
are probably the most striking examples available of this process at work.
Far too much has been made of complementary distribution of codes
and of the degree of structural affinity existing between codes as criterial
Focus article 39
attributes of diglossia. The situations described by Ferguson (excluding,
perhaps, that of Haitian Creole), and the many other potentially com-
parable situations alluded to in the course of this paper, do not con-
stitute a meaningful conceptual or theoretical class of cases based solely
upon asymmetric functional distribution of codes and/or upon the
linguistic distance between them. Their conceptual unity inheres in a
quite specific set of relationships between functional compartmental-
ization of codes, the lack of opportunity for the acquisition of H as
a native variety, the resulting absence of native speakers of H, and the
stability in the use of L for vernacular purposes. In principle at least,
the codes involved in this configuration might be varieties of totally
unrelated languages as readily as they might be minimally distinct
isolects of the same language. In practice, however, it is no accident that
these codes tend to be closely related structurally and generally to be
regarded as varieties of the same language, albeit significantly different
In cases where, over time, a single cultural-linguistic tradition dif-
ferentiates its endogenous linguistic resources according to function, it
is obvious that new forms, and new varieties of language, develop (or
may be borrowed) to accommodate new (or imported) social functions.
Insofar as these functions generally pertain to the more formal and
ontogenetically later aspects of the culture, their associated varieties tend
to serve precisely those functions that are not the main channels for first-
language acquisition (see Joseph 1987: 17, quoted above). Accordingly,
there typically does not develop a prestige community of native speakers
of H that might serve as a reference group for native speakers of L,
thereby providing the social impulse for shift from L to H as a native
variety. Exogenous language contact, on the other hand, is an entirely
different matter. Here, two separate cultural-linguistic traditions come
into contact and therefore into competition with each other for at least
some of the same socioecological niches. The linguistic varieties involved
may be relatively closely related, as in the cases of Frisian and Dutch,
´n and Castillian, or Occitan and French, or they may be essentially
unrelated languages such as Nahuatl and Spanish, English and Navajo,
or Arabic and French. In either case, however, these varieties must
be viewed socially, politically, historically, and phenomenologically as
different languages, representing different group identities, if not dif-
ferent cultures. In contrast with diglossia, it is axiomatic to say that
both cultural-linguistic traditions in the case of exogenous language
contact already have their own established functions as vernaculars,
although they may differ to varying degrees as to the number and nature
of high-culture functions for which appropriate registers have been
40 A. Hudson
developed and deployed. Under such circumstances, then, there is always
the potential for one community of speakers typically, if not by
definition, the speakers of H — to come to serve as a linguistic reference
group for the speakers of L, and thus to furnish the social motivation
necessary for the eventual shift from L to H as the native language. It
is therefore the particular social etiology of diglossia that provides the
explanation for the security of L as the vernacular over the long term,
as well as for the generally expected degree of relatedness between the
constituent codes. For this reason, it may not be realistic to think of
diglossia, in any narrow sense of the term, as providing an appropriate
model for the stabilization of endangered languages or for their eventual
revitalization as fully functioning native languages.
In a recent essay on diglossia, Schiffman comments as follows: ‘‘It
remains to be seen whether the same kind of imbalance of power exhibited
in nongenetic diglossia can be said to exist with regard to classical or
genetic diglossia’’ (1997: 206). This is surely one of the most fundamental
questions that can, and should, be asked about any sociolinguistic
condition vis-a
`-vis another. The argument that has been advanced in this
paper is that, allowing for some exceptions in both directions, there is
a general tendency for the imbalance of power found in cases of so-called
‘‘nongenetic diglossia’’ not to be found, or to be found with the opposite
polarity, in cases of so-called ‘‘genetic diglossia,’’ and that this very fact
is reason enough on its face to distinguish between the two types of
verbal repertoire at a deeper level than that of genetic relatedness between
the codes in the respective code matrices. It is, in other words, the most
fundamental reason for distinguishing between diglossia and other
sociolinguistic situations often referred to as societal bilingualism.
Surface phenomena, in this case the differential functional allocation of
codes within a speech economy, may be interesting in and of themselves,
but they attain far greater theoretical significance when understood in
relation to the underlying social processes of which they are a product,
all the more so, perhaps, when apparently similar surface patterns reflect
radically different underlying processes.
Anyone who is familiar at first hand with the situation in German-
speaking Switzerland, for example, knows full well that this situation
differs radically from the type of federal multilingual relationship existing
between French, German, Italian, and Romansch at the national level
in that country, and from the sociolinguistic relationship obtaining
between Romansch and German in Graubu
¨nden, where Romansch,
despite its recently attained constitutional status as the fourth official
language of Switzerland, is demographically very much on the defensive
against Swiss German in the vernacular domains and Standard German
Focus article 41
in the high-culture domains. Those familiar with other situations
resembling the Swiss-German one will presumably also be compelled
by their experience to differentiate those situations from others mentioned
in the course of this paper, such as Frisian and Dutch in The Netherlands,
French and Occitan in France, Irish and English in Ireland, Spanish
and Nahuatl in Mexico, Navajo and English in the United States, and,
not improbably, even Spanish and Guaranı
´in Paraguay. All of these
situations also may be characterized by a greater or lesser degree of
asymmetric functional distribution of codes among their respective com-
munities of bilingual speakers, and not a few of them, if not perhaps all,
have been treated at one time or another as instances of diglossia in
the sociolinguistic literature. Sociolinguistic theory is not well served by
the treatment of all of these situations as though they were simply surface
manifestations of the same underlying phenomenon.
Some time ago, a Swiss-German–speaking acquaintance of mine in
Basel, failing to comprehend my interest in the local language situation,
remarked to me that diglossia was such a quotidian affair as to seem
undeserving of the attention that it received from sociolinguists such as
myself. Indeed, in the German-speaking cantons of Switzerland, diglossia
is as ubiquitous, and therefore as unremarkable, as the air people breathe.
In actuality, however, the type of sociolinguistic situation found in
German-speaking Switzerland, in the Arabic-speaking world, and at least
until recently in Greece is anything but commonplace today. On the
contrary, it is worth remembering, as noted above, that ‘‘societal changes
such as modernisation, urbanisation, [and] the breakdown of rigid class
barriers _have made diglossic situations as described by Ferguson
(1959) rather rare’’ (Pauwels 1986: 16; note also Baker and Prys Jones
1998: 121; Scotton 1986: 408–409). Indeed, it is the rarity, not the
ubiquity, of such situations that confers upon them their considerable
theoretical interest for the sociologist of language. The remarkable, yet
still not widely understood, facts of the sociolinguistic situations in
German-speaking Switzerland, in the Arabic-speaking world, in Greece,
in the Tamil-speaking regions of India and perhaps other areas of South
Asia, in medieval Ireland and Wales, in Jewish Palestine during the late
Second Temple Period, in China from Confucian times up to the present
century, in Japan from the seventh century on, and in medieval Latin-
and Romance-speaking Europe present a wealth of data from which
potentially to construct a concept of diglossia and a theory of its
functioning, and it is worth examining whether these situations, and
others besides, may be sufficiently alike to warrant their inclusion under a
single designation distinct from those of other sociolinguistic situations.
In the final analysis, what the term ‘‘diglossia’’ should be understood
42 A. Hudson
as designating is a matter of linguistic convention within the discipline of
sociolinguistics. The meaning is not given in the term itself, nor is the
meaning intended by Ferguson in 1959 necessarily immutable for all time.
Fishman’s (1967) argument that the term ‘‘bilingualism,’’ referring to
individual linguistic versatility, should be distinguished from the term
‘‘diglossia,’’ referring to societally held norms governing differential
functional allocation of codes, is a crucial one, although, as I have
remarked elsewhere, it may be unfortunate that the term ‘‘diglossia’’
should have been coopted to refer to patterned situational variation more
generally. At the least, this paper has argued for a uniform use, whatever
that may be, of the term ‘‘diglossia’’ in sociolinguistic research and theory
construction. Equally important, it also argues that the similarities that
exist between the sociolinguistic situations described by Ferguson are
sufficiently different from other types of sociolinguistic situation to
warrant their recognition as a category distinct from these other types,
such as standard-with-dialects and societal bilingualism.
University of New Mexico
Abdulaziz, M. (1986). Factors in the development of modern Arabic usage. International
Journal of the Sociology of Language 62, 11–24.
Abrahams, R. (1972). The training of the man of words in talking sweet. Language in Society
1, 15–29.
Akinnaso, F. (1982). The literate writes and the nonliterate chants: written language and
ritual communication in sociolinguistic perspective. In Linguistics and Literacy,
W. Frawley (ed.), 7–36. New York: Plenum.
Alexiou, M. (1982). Diglossia in Greece. In Standard Languages: Spoken and Written,
W. Haas (ed.), 156–192. Manchester: Manchester University Press.
Baker, C.; and Prys Jones, S. (1998). Encyclopedia of Bilingualism and Bilingual Education.
Clevedon, UK: Multilingual Matters.
Barnes, D. (1982). Nationalism and the Mandarin movement: the first half-century. In
Language Spread: Studies in Diffusion and Social Change, R. Cooper (ed.), 260–290.
Bloomington and Washington, D.C.: Indiana University Press and Center for Applied
Bell, A. (1984). Language style as audience design. Language in Society 13, 145–204.
Berger, M. (1990). Diglossia within a general theoretical perspective: Charles Ferguson’s
concept 30 years later. Multilingua 9, 285–295.
Besnier, N. (1988). The linguistic relationships of spoken and written Nukulaelae registers.
Language 64, 707–736.
Biber, D. (1988). Variation Across Speech and Writing. Cambridge: Cambridge University
Focus article 43
Bickerton, D. (1971). Inherent variability and variable rules. Foundations of Language 7,
457– 492.
Bright, W. (1976). Phonological rules in literary and colloquial Kannada. In Variation and
Change in Language: Essays by William Bright, 65 –73. Stanford, CA: Stanford
University Press.
(1982). Literature: written and oral. In Analyzing Discourse: Text and Talk, D. Tannen
(ed.), 271–283. Georgetown University Round Table on Languages and Linguistics 1981.
Washington, D.C.: Georgetown University Press.
Britto, F. (1986). Diglossia: A Study of the Theory with Application to Tamil. Washington,
D.C.: Georgetown University Press.
— (1991). Tamil diglossia: an interpretation. Southwest Journal of Linguistics 10(1), 60– 84.
Bybee, J.; Perkins, R.; and Pagliuca, W. (1994). The Evolution of Grammar: Tense,
Aspect, and Modality in the Languages of the World. Chicago: University of Chicago
Chafe, W. (1982). Integration and involvement in speaking, writing, and oral literature. In
Spoken and Written Language: Exploring Orality and Literacy, D. Tannen (ed.), 35–53.
Norwood, NJ: Ablex.
Chambers, J. (1995). Sociolinguistic theory. Oxford: Blackwell.
Comrie, B. (1991). Diglossia in the Old Russian period. Southwest Journal of Linguistics
10(1), 160–172.
Cooper, R. (1978). The spread of Amharic in Ethiopia. In Advances in the Study of Societal
Multilingualism, J. Fishman (ed.), 459–476. The Hague: Mouton.
Coulmas, F. (1987). What writing can do to language: some preliminary remarks. In
Developments in Linguistics and Semiotics, Language Teaching and Learning, Com-
munication Across Cultures, S. Battestini (ed.), 107–129. Georgetown University Round
Table on Languages and Linguistics, 1986. Washington, D.C.: Georgetown University
— (1989). Language adaptation. In Language Adaptation, F. Coulmas (ed.), 1–25.
Cambridge: Cambridge University Press.
(1991). Does the notion of diglossia apply to Japanese? Some thoughts and some
documentation. Southwest Journal of Linguistics 10(1), 125–142.
Daltas, P. (1993). The concept of diglossia from Ferguson to Fishman to Fasold. In Themes
in Greek Linguistics: Papers from the First International Conference on Greek Linguistics,
Reading, September 1993, I. Philippaki-Warburton, K. Nicolaidis, and M. Sifianou
(eds.), 341–348. Amsterdam: Benjamins.
DeCamp, D. (1971). Toward a generative analysis of a post-creole speech continuum. In
Pidginization and Creolization of Languages, D. Hymes (ed.), 349–370. New York:
Cambridge University Press.
DeFrancis, J. (1972). Nationalism and Language Reform in China. Princeton: Princeton
University Press.
Deshpande, M. (1991). Conceptions of diglossia in the writings of the Sanskrit grammarians.
Southwest Journal of Linguistics 10(1), 23– 40.
Eckert, P. (1980). Diglossia: separate and unequal. Linguistics 18, 1053–1064.
Errington, J. (1991). A muddle for the model: diglossia and the case of Javanese. Southwest
Journal of Linguistics 10(1), 189–213.
Ervin-Tripp, S. (1971). Sociolinguistics. In Advances in the Sociology of Language, vol. 1,
J. Fishman (ed.), 15–91. The Hague: Mouton.
Fasold, R. (1984). The Sociolinguistics of Society. Oxford: Blackwell.
Feldman, C. (1991). Oral metalanguage. In Literacy and Orality, D. Olson and N. Torrance
(eds.), 47–65. Cambridge: Cambridge University Press.
44 A. Hudson
Fellman, J. (1975a). On ‘‘diglossia.’’ Language Sciences 34, 38–39.
(1975b). The resolution of diglossia in an African setting: the Ethiopian case. La Monda
Lingvo-Problemo 5, 178–183.
— (1977). Diglossia: the Hebrew case. La Monda Lingvo-Problemo 6, 107–111.
Ferguson, C. (1959). Diglossia. Word 15, 325–340.
— (1963a). Introduction [to National Languages and Diglossia]. In Report of the Thirteenth
Annual Round Table Meeting on Linguistics and Language Studies, E. Woodworth
and R. Di Pietro (eds.), 163. Monograph Series on Languages and Linguistics 15.
Washington, D.C.: Georgetown University Press.
(1963b). Problems of teaching languages with diglossia. In Report of the Thirteenth
Annual Round Table Meeting on Linguistics and Language Studies, E. Woodworth and
R. Di Pietro (eds.), 165–177. Monograph Series on Languages and Linguistics 15.
Washington, D.C.: Georgetown University Press.
— (1968). Language development. In Language Problems of Developing Nations, J. Fishman,
C. Ferguson, and J. Das Gupta (eds.), 27–35. New York: Wiley.
(1973). Language problems of variation and repertoire. Daedalus 102(3), 37–46.
— (1991). Diglossia revisited. Southwest Journal of Linguistics 10(1), 214 –234.
´ndez, M. (1993). Diglossia: A Comprehensive Bibliography 1960 –1990. Amsterdam:
Fishman, J. (1967). Bilingualism with and without diglossia; diglossia with and without
bilingualism. Journal of Social Issues 23(2), 29–38.
(1980). Bilingualism and biculturism as individual and as societal phenomena. Journal
of Multilingual and Multicultural Development 1, 3–15.
— (1991). Reversing Language Shift: Theoretical and Empirical Foundations of Assistance
to Threatened Languages. Clevedon, UK: Multilingual Matters.
Frangoudaki, A. (1992). Diglossia and the present language situation in Greece: a
sociological approach to the interpretation of diglossia and some hypotheses on today’s
linguistic reality. Language in Society 21, 365–381.
Gallagher, C. (1971). Language reform and social modernization in Turkey. In Can
Language Be Planned? Sociolinguistic Theory and Practice for Developing Nations,
J. Rubin and B. Jernudd (eds.), 159–178. Honolulu: University Press of Hawaii.
Geertz, C. (1968). Linguistic etiquette. In Readings in the Sociology of Language, J. Fishman
(ed.), 282–295. The Hague: Mouton.
´n, T. (1979). On Understanding Grammar. New York: Academic Press.
Gumperz, J. (1968 [1962]). Types of linguistic communities. In Readings in the Sociology
of Language, J. Fishman (ed.), 460 – 472. The Hague: Mouton.
¨cki Buhofer, A.; and Burger, H. (1993). Hochdeutsch bei sechs- bis achtja
Deutschschweizer Kindern: Verstehen—Einstellungen. In Alemannisch in der Regio:
¨ge zur 10. Jahrestagung alemannischer Dialektologen, Volker Schupp (ed.), 11–23.
¨ppinger Arbeiten zur Germanistik 593. Go
¨ppingen: Ku
Hall, R., Jr. (1974). External History of the Romance Languages. New York: American
(1978). Bi-(Multi-)lingualism and diglossia in Latin and Romance. Forum Linguisticum
3(2), 107–119.
Halliday, M. (1968). The users and uses of language. In Readings in the Sociology of
Language, J. Fishman (ed.), 139–169. The Hague: Mouton.
Hill, J. (1983). Language death in Uto-Aztecan. International Journal of American Linguistics
49, 258–276.
Hock, H.; and Pandharipande, R. (1978). Sanskrit in the pre-Islamic sociolinguistic context
of South Asia. International Journal of the Sociology of Language 16, 11–25.
Focus article 45
Hudson, A. (1994). Diglossia as a special case of register variation. In Sociolinguistic
Perspectives on Register, D. Biber and E. Finegan (eds.), 294–314. Oxford: Oxford
University Press.
Hudson-Edwards, A. (1984). Rediscovering diglossia. Southwest Journal of Linguistics
7, 5–15.
Hymes, D. (1967). Models of the interaction of language and social setting. Journal of Social
Issues 23(2), 8–28.
Ibrahim, M.; and Jernudd, B. (1986). Introduction. International Journal of the Sociology
of Language 61: Aspects of Arabic Sociolinguistics (special issue), 5–6.
Irvine, J. (1984). Formality and informality in communicative events. In Language in Use:
Readings in Sociolinguistics, J. Baugh and J. Sherzer (eds.), 211–228. Englewood Cliffs,
NJ: Prentice-Hall.
´s, E
´. (1984). Diglossia in Persian. Acta Linguistica Academiae Scientiarum
Hungaricae 34(3– 4), 271–287.
Joseph, J. (1987). Eloquence and Power: The Rise of Language Standards and Standard
Languages. New York: Blackwell.
Kahane, H. (1986). A typology of the prestige language. Language 62, 495–508.
— ; and Kahane, R. (1979). Decline and survival of Western prestige languages. Language 55,
Karpat, K. (1984). A language in search of a nation: Turkish in the nation-state. In The
Emergence of National Languages, A. Scaglione (ed.), 175–208. Ravenna: Longo.
Kaye, A. (1970). Modern Standard Arabic and the colloquials. Lingua 24, 374 –391, 412.
— (1972). Remarks on diglossia in Arabic: well-defined vs. ill-defined. Linguistics 81, 32–48.
(1994). Formal vs. informal in Arabic: diglossia, triglossia, tetraglossia, etc.,
multiglossia–polyglossia viewed as a continuum. Journal of Arabic Linguistics 27,
47– 66.
Keller, R. (1982). Diglossia in German-speaking Switzerland. In Standard Languages:
Spoken and Written, W. Haas (ed.), 70 –93. Manchester: Manchester University Press.
Klaiman, H. (1993). The demise of diglossia in Bengali. In Papers from the Fifteenth South
Asian Language Analysis Roundtable Conference 1993, A. Davison and F. Smith (eds.),
151–166. Iowa City: South Asian Studies Program, University of Iowa.
Labov, W. (1994). Principles of Linguistic Change, vol. 1: Internal Factors. Oxford:
Lee, G. (1986). Diglossia in Ancient India. In Studies on Language Change, B. Joseph (ed.),
151–164. Working Papers in Linguistics 34. Columbus: Ohio State University.
Li, C.; and Thompson, S. (1982). The gulf between spoken and written language: a case study
in Chinese. In Spoken and Written Language: Exploring Orality and Literacy, D. Tannen
(ed.), 77–88. Norwood, NJ: Ablex.
Mackenzie, J. (1994). Western Europe. In Atlas of the World’s Languages, C. Moseley and
R. Asher (eds.), 247–252. New York: Routledge.
Mackey, W. (1986). The polyglossic spectrum. In The Fergusonian Impact: In Honor of
Charles A. Ferguson on the Occasion of his 65th Birthday, vol. 2, J. Fishman, A. Tabouret-
Keller, M. Clyne, B. Krishnamurti, and M. Abdulaziz (eds.), 237–243. Berlin: Mouton de
— (1989). La gene
`se d’une typologie de la diglossie. Revue Que
´coise de Linguistique
´orique et Applique
´e8(2), 11–28.
Mahmoud, Y. (1986). Arabic after diglossia. In The Fergusonian Impact: In Honor of Charles
A. Ferguson on the Occasion of his 65th Birthday, vol. 1, J. Fishman, A. Tabouret-Keller,
M. Clyne, B. Krishnamurti, and M. Abdulaziz (eds.), 239–251. Berlin: Mouton de
46 A. Hudson
Mirambel, A. (1964). Les aspects psychologiques du purisme dans la Gre
`ce moderne.
Journal de Psychologie Normale et Pathologique 4, 405– 436.
Morton, W. (1982). China: Its History and Culture. New York: McGraw-Hill.
Nercissians, E. (1987). The development of standard literary Armenian and Persian: a
comparative study. Journal of Pragmatics 11, 623–639.
Neustupny, J. (1974). The modernization of the Japanese system of communication.
Language in Society 3, 33–50.
Nishi, A. (1874). Y ji o motte kokugo o shosuru no ron [A plea for writing the national
language with Western letters]. Meiroku zasshi 1.
Paolillo, J. (1997). Sinhala diglossia: discrete or continuous variation? Language in Society
26, 269–296.
Parker, I. (1983). The rise of the vernaculars in early modern Europe: an essay in the political
economy of language. In The Sociogenesis of Language and Human Conduct, B. Bain
(ed.), 323–351. New York: Plenum.
Paulston, C. (1994). Linguistic Minorities in Multilingual Settings: Implications for Language
Policies. Studies in Bilingualism 4. Amsterdam: Benjamins.
Pauwels, A. (1986). Diglossia, immigrant dialects and language maintenance in Australia:
the case of Limburgs and Swabian. Journal of Multilingual and Multicultural Development
7, 13–30.
Peyraube, A. (1991). Some diachronic aspects of diglossia/triglossia in Chinese. Southwest
Journal of Linguistics 10(1), 105–124.
Pulgram, E. (1950). Spoken and written Latin. Language 26, 458–466.
Rabin, C. (1955). The beginnings of Classical Arabic. Studia Islamica 4, 19–37.
— (1958). The historical background of Qumram Hebrew. Scripta Hierosolymitana 4,
144 –161.
Radhakrishna, B. (1980). Study of diglossia and the Telugu situation. In Sociolinguistics
and Dialectology, S. Agesthialingom and K. Karunakan (eds.), 238–241. Annamalaigar,
India: Annamalai University.
Rendsburg, G. (1990). Diglossia in Ancient Hebrew. New Haven: American Oriental Society.
Rosemont, H., Jr. (1974). On representing abstractions in Archaic Chinese. Philosophy East
and West 24, 71–88.
Sankoff, G. (1972). Language use in multilingual societies: some alternative approaches.
In Sociolinguistics: Selected Readings, J. Pride and J. Holmes (eds.), 33–51.
Harmondsworth: Penguin.
Schiffman, H. (1997). Diglossia as a sociolinguistic situation. In The Handbook of
Sociolinguistics, F. Coulmas (ed.), 205–216. Oxford: Blackwell.
Scotton, C. (1986). Diglossia and code switching. In The Fergusonian Impact: In
Honor of Charles A. Ferguson on the Occasion of his 65th birthday, vol. 2, Socio-
linguistics and the Sociology of Language, J. Fishman, A. Tabouret-Keller, M. Clyne,
B. Krishnamurti, and M. Abdulaziz (eds.), 403– 415. Berlin: Mouton de Gruyter.
Sethe, K. (1925). Das Verha
¨ltnis zwischen Demotisch und Koptisch und seine Lehren fu
die Geschichte der a
¨gyptischen Sprache. Zeitschrift der Deutschen Morgenla
Gesellschaft 79, 290–316.
Sjoberg, A. (1964). Writing, speech, and society: some changing interrelationships. In
Proceedings of the Ninth International Congress of Linguists, H. Lunt (ed.), 892–898.
Janua Linguarum Series Major 12. The Hague: Mouton.
Sotiropoulos, D. (1982). The social roots of modern Greek diglossia. Language Problems
and Language Planning 6, 1–28.
Tannen, D. (1982). The oral/literate continuum in discourse. In Spoken and Written
Language: Exploring Orality and Literacy, D. Tannen (ed.), 1–16. Norwood, NJ: Ablex.
Focus article 47
¨b, R. (1992). Das Schweizerdeutsche Wo
¨rterbuch und der Sprachatlas. Alemannisch
dunkt u
¨s guet 3/4, 16–27.
Ure, J. (1982). Introduction: approaches to the study of register range. International Journal
of the Sociology of Language 35, 5–23.
Voegelin, C. (1960). Casual and noncasual utterances within unified structure. In Style
in Language, T. Sebeok (ed.), 57–68. Cambridge, MA and New York: Technology Press
of the Massachusetts Institute of Technology and Wiley.
Walters, K. (1996). Diglossia, linguistic variation, and language change in Arabic. In
Perspectives on Arabic linguistics VIII: Papers from the Eighth Annual Symposium on
Arabic Linguistics, M. Eid (ed.), 157–197. Amsterdam: Benjamins.
Werlen, I. (1993). Schweizer Soziolinguistik — Soziolinguistik der Schweiz: Zur Einfu
Bulletin CILA 58, 7–15.
Wexler, P. (1971). Diglossia, language standardization, and purism: parameters for a
typology of literary languages. Lingua 27, 330 –354.
Winford, D. (1985). The concept of ‘‘diglossia’’ in Caribbean creole situations. Language
in Society 14, 345–356.
48 A. Hudson
... Sociolinguistic theory distinguishes two broadly defined dialectal contexts: Diglossia and Standard-with-Dialects contexts, both different from Societal Bilingualism (Hudson, 2002). The term diglossia first emerged in sociolinguistic theory to describe a situation where, in a given society, there is more than one language variety in use. ...
... The result is that high German is The Sociolinguistics of Diglossia in Switzerland 38 subordinated to dialect and serves as the written language and the lingua franca for communication with other Swiss who do not know the Alemannic dialects and with foreigners who can speak German (Dabène, 1994: 55). No Swiss German speaker would ever use Standard German in personal contact with another Swiss from a German-speaking canton (Hudson, 2002: 3, Keller, 1982. In the opinion of some authors (Hogg et al., 1984, Weil & Schneider, 1997, Rash, 1998, Schläpfer, 1994, the expanded use of Swiss German renders the original concept of diglossia less applicable to the German-speaking part of Switzerland. ...
... Moreover, unlike bilinguals, Arabic speakers never achieve comparable levels of proficiency in their two language varieties even after years of exposure to the standard variety, neither do they ever become more competent in the standard variety than in their first acquired spoken dialect. Hudson (2002) argues that the fact that the functional distribution of codes in a diglossic context protects the role of the spoken variety as the only natively learned variety is what distinguishes diglossia from other interlingual or intralingual situational alternation. ...
Full-text available
It is generally assumed that learning to read involves a straightforward learning of the mappings from speech to spelling. Yet, the majority of the world’s children learn to read frst in a language or dialect that is not what they speak at home or in the neighborhood with their families and friends. Some of these children must learn to read frst in a foreign language (e.g., Bemba speakers in Zambia learning to read in English), but many are learning to read in a dialect that shares some similarities with the formal written word but that also differs substantially from it. Positioned within an extended ecological approach to literacy development (McBride, 2016), this Handbook highlights some of the theoretical and practical issues that a mismatch between dialect and literacy requirements involves. These include a variety of lin�guistic aspects, but they also affect individuals demonstrably at many levels, includ�ing psycholinguistics, neurolinguistics, education, and many aspects of social interactions. A broad understanding of the interface between dialects and literacy acquisition is fundamental for all research that highlights interactions among lan�guage, literacy, and society. The current Handbook studies literacy acquisition at the intersection of sociolin�guistics and psycholinguistics by addressing literacy acquisition in diglossia and in dialectal contexts. The Handbook emanates from an international conference orga�nized in 2018 by Elinor Saiegh-Haddad and Lior Laks at Bar-Ilan University, Israel, under the auspices of the Israel Science Foundation (Grant number 2346/17) and the Arabic Language Academy in Israel. The conference brought together researchers from various regions across the world including Asia, Europe, the Far East, and North America to share research questions, methods, and fndings on literacy devel�opment in diglossic and in dialectal contexts. The Handbook at hand features some of the talks presented at the conference in 2018 and additional chapters addressing similar questions in various other regions and languages in the world. Some of the questions that are addressed include the following: How does dialect factor into literacy development and disorder in children? Which sociolinguistic features of dialectal contexts affect literacy acquisition? Is the role of sociolinguistic features of dialectal variation similar or different in different contexts and languages? Do dif�ferent dialectal contexts differ in the settings and functions of language, and do the differences yield different literacy outcomes for children? What are the milestones of literacy development in different dialectal contexts? How should educational assessment of language and literacy address sociolinguistic features of dialectal contexts? What are the most benefcial instructional practices for children raised in dialectal contexts. Given that the majority of the world’s children learn to read in a context that includes diglossia, dialectalism, and multilingualism, much more of an understand�ing of the complexities of these contexts is needed. The various perspectives offered in this Handbook underscore some of the most important issues in the feld of lit�eracy in relation to diglossia and dialectalism. We hope that this Handbook contrib�utes critically to inquiry that will beneft our children’s well-being and their prospects for success.
... The low variety is utilized in familiar and everyday contexts, whereas the high variety is spoken in formal contexts and in writing. This functions in a protective manner toward the low variety (Hudson, 2002). In other words, in diglossic contexts the high variety (which is no one's mother tongue, see, Coulmas, 1987;Schiffman, 1997;Ure, 1982) has its own distinct functional role. ...
... However, Ferguson (1959) clarifies that, in diglossia, a community utilizes (a) a language of the past, or (b) the language of some other community, for formal purposes. Concerning the second case, Hudson (2002) most helpfully clarifies that the members of the community that employs the H(igh) variety as vernacular "do not regard themselves, and are not regarded as, members of the community that employs H for nonvernacular purposes only" (p. 21); by implication, then, ethnic ties do not connect the members of the two communities. ...
... In my view, all these complexities are taken into account in Hudson's (2002) significant position that diglossic situations are neither exclusively about the structure gap of codes, nor only about their functional distribution. They are about the complex and causal correlations between functional distribution, acquisition and the stability of the low variety (as it was presented and explained above). ...
The present article aims to scrutinize the widely expressed assertion that the Greek Cypriot sociolinguistic situation is diglossic. Rather than thoroughly examining whether code-switching depends on the formality of the context, and whether Greek Cypriots only acquire the Cypriot Greek dialect, the relevant literature takes these for granted, uses them as points of departure and defines the relationship between the Cypriot Greek dialect and the Standard Greek as diglossic. Hence, it is no wonder that the relevant scholarship leaves unseen and unanswered important questions about the Greek Cypriot context. At the same time, the uncritical assertion that the sociolinguistic situation in Cyprus reflects diglossia is framed by an unjustifiable and sweeping incrimination of the Greek ethnic identity of Greek Cypriots. These lead the scholars who adhere to the assumption of diglossia in Cyprus to perform a “one-way” transfer of theory to the Greek Cypriot context by unwittingly elevating this theory to a grand narrative applicable to just any seemingly fitting situation. Thus, while advancing a current metanarrative function of diglossia and jumping on the bandwagon that its theoretical dispersal creates the relevant scholarship renders the term of diglossia ideological and tailors the Greek Cypriot sociolinguistic context to its parameters.
... In realtà, rimanendo nell'ambito delle due varietà, è possibile distinguere due tipi di bilinguismo: uno riguarda il livello del singolo parlante, e in tal caso si parlerà di 'bilinguismo individuale', l'altro riguarda il livello di tutta la comunità di parlanti, e in quest'ultimo caso si parlerà più propriamente di 'bilinguismo sociale ' (cf. Hudson 1980;2002). È quest'ultimo tipo di bilinguismo quello di maggior interesse per il presente volume. ...
... 45 Ancora, tipici del siciliano sono una serie di costrutti che esprimono la proposizione limitativa (cf. Trovato 1988a;2002), i quali si compongono di una prima parte formata da un infinito preceduto dai focalizzatori pi (cf. (41a)), probabilmente il tipo più antico, o di (cf. ...
The present volume aims at contributing to the ongoing discussion on the relationship between Italian and the Sicilian dialects, by updating it to the new communication dimension represented by Social Media, and the related increase of the written domain in daily digital interactions. The volume starts with a chapter on the technical terminology that is needed to address the question ‘ lingua vs. dialetto ’, which will justify the use of the term dilalico , à la Berruto, to describe the Italian sociolinguistic scenario. Chapter 2 offers an overview of the main European linguistic settings and discusses both the legal choices made by the Italian State regarding what varieties spoken in the Peninsula should be considered as minority languages to protect, and the normative reaction by the Sicilian regional government. Chapter 3 discusses those features that differentiate the Sicilian dialects from Standard Italian, sometimes characterising the regional Italian spoken in Sicily. It is crucial that these features are carefully taken into account in order to foster metalinguistic abilities and activate those cognitive advantages documented in different bilingual contexts. Chapter 4 deals with the issue of writing in dialect. Texting on Social Media is almost exclusively the domain of Italian, considering the persisting difficulties in promoting a written code that could be shared by most users. This is detrimental for dilalic speakers whose overall opportunities to use their dialects in such an important context diminish remarkably. Chapter 5 provides a discussion of the fact that awareness of the linguistic properties of the dialects could contribute to better control the competence in Italian, and to learn foreign languages more easily by capitalising on what dilalic speakers know about their dialects. The volume ends with some educational considerations that also hold for the other dilalic situations in Italy.
... Ideologically, both were nationalist and aspired to the irredentist concept of the Megali Idea (Great Idea), subscribing to the notions of unity and continuity which prevailed in the Greek national identity at the time. But, as far as the language question was concerned, Hatzidakis sided with the conservative view, which considered the puristic form as the sole "natural" medium of written communication (1903), whilst, according to Psycharis, spoken Greek (actually, a "vernacular standard" 15 ) 14. Based on the typology of diglossic situations proposed by Hudson (2002), Frangoudaki (2002 posits that the Greek sociolinguistic phenomenon can be considered as typically diglossic only for the period ending at the beginning of the 20th century. More precisely, she claims that the period starting in the 1880s and until the 1920s fits better into Fishman's "transitional diglossia", whereas the interwar period onwards should be seen as a case of "societal bilingualism", in the sense that both Katharevousa and Dimotiki functioned as High varieties for the state and the intellectual elite, respectively (2002: 101). ...
Full-text available
La production et la transmission des savoirs scientifiques s’effectuent par des instances dédiées (exemplairement l’Université), mais aussi par la médiation de multiples structures, qui se sont agrégées autour d’une thématique (académies, sociétés savantes), d’une affiliation (dans le cas notamment des « écoles »), ou encore d’un projet (comme la grammatisation – l’individuation et la description – d’une langue vernaculaire). Quoique bien identifiées dans leurs aspects sociaux, ces structures sont moins souvent appréhendées dans toute la complexité de leurs apports spécifiques. Tel était donc l’enjeu du colloque consacré à « la linguistique et ses formes historiques d’organisation et de production » (Paris, 24-26 janvier 2019), à l’occasion du quarantième anniversaire de la Société d’histoire et d’épistémologie des sciences du langage. La plupart des contributions au présent volume portent sur des périodes contemporaines de l’organisation moderne du savoir et montrent comment ces structures se sont positionnées par rapport aux institutions déjà existantes. Mais le lecteur y puise également des perspectives plus larges, grâce à des analyses consacrées à d’autres espaces culturels et aussi aux empans temporels abordés, de la microhistoire aux temporalités longues. Sous la diversité des contextes et des situations, certains mécanismes se révèlent alors étonnamment récurrents, qu’ils concernent l’émergence et la disciplinarisation de nouveaux objets d’étude ou les comportements cognitifs des acteurs.
... Moreover, unlike bilinguals, Arabic speakers never achieve comparable levels of proficiency in their two language varieties even after years of exposure to the standard variety, neither do they ever become more competent in the standard variety than in their first acquired spoken dialect. Hudson (2002) argues that the fact that the functional distribution of codes in a diglossic context protects the role of the spoken variety as the only natively learned variety is what distinguishes diglossia from other interlingual or intralingual situational alternation. ...
To study reading acquisition in diglossia is to study reading at the intersection of psycholinguistics and sociolinguistics. In a diglossic context, sociolinguistic variables imbue the process of first language literacy acquisition with a host of variables that are not pertinent to other first language reading contexts. In this chapter, we discuss some major sociolinguistic features of Arabic diglossia that might have direct psycholinguistic implications for literacy development. Then, we present the basic tenets of our psycholinguistic-developmental approach to the study of reading acquisition in Arabic diglossia, including concepts, assumptions, methods, and findings. Five questions are addressed: (a) Does diglossia impact acquisition of basic reading skills? (b) Is the impact of diglossia the same across dialects? (c) Does the impact of diglossia decrease with increased exposure to the standard? (d) Does the impact of diglossia interact with developmental and environmental risk factors? (e) Does reading development in diglossia show a cross-lectal transfer of skills? Educational implications and directions for future research are outlined.KeywordsArabicDialectDiglossiaLinguistic DistanceReading Acquisition
... SA and LA are used in a complementary fashion; each is strictly used in its specified context (Ferguson, 1959;Maamouri, 1998). Context, rather than the speaker, dictates which variety should be used (Hudson, 2002): SA is never used in formal education and LA is never used for everyday communication regardless of how literate the speaker is (Saiegh-Haddad, 2022). Acquiring the two varieties in such a diglossic context makes a native speaker of Arabic very similar to a coordinate bilingual, "a person who acquires the two languages in different context(s), for instance, one at home and the other at school" (D' Acierno, 1990, p. 12). ...
Full-text available
Native speakers of Arabic acquire language in a diglossic context that requires them to use different varieties for different purposes: spoken Arabic (SA) is the dialect they use informally in daily oral communications; literary Arabic (LA) is the variety they use mainly for reading, writing, and formal communications. In general, Arabic native speakers perform differently across different tasks and modalities­­—performance tends to be better when the task requires SA or LA in the same way it is normally used. In this study, an LA speech production task was performed by two groups of Arabic native speakers who varied significantly in their amount of practice with LA. Although both groups acquired LA under the same conditions, the group with more practice was more fluent. Practice-dependent differences are interpreted within a memory-based automaticity framework. Such a framework, it is argued, is able to account for differences both in general performance patterns among the Arab population as well as specific, practice-dependent patterns such as those observed in the present study.
... This formal representation of Arabic is usually more sophisticated and complex (Kaye, 2001;Saiegh-Haddad & Henkin-Roitfarb, 2014), and from a social perspective of this diglossic situation, the two language varieties are used complementarily rather than interchangeably, and serve different functions (Ferguson, 1959;Maamouri, 1998). Language variety use in diglossia is governed by context rather than speakers (Hudson, 2002). All Arabic native speakers, regardless of age, education, or socioeconomic status, use a spoken dialect for everyday communication, opposed to LA (Saiegh-Haddad, 2022). ...
Arabic-speaking children acquire literacy in a diglossic manner in which the variety of language engaged for reading and writing at school—referred to here as literary Arabic (LA)—differs from the variety they engage when speaking at home. Literacy acquisition in such a context necessitates teaching practices take into consideration the differences between spoken and literary varieties of language to assist children to bridge the gap between them. This study explores how Saudi teachers of kindergarten-level students perceive the effect of diglossia on the initial stages of literacy acquisition by their students, and which practices they follow to minimize its effect. The study participants took part in a focus group in which they shared their experiences of teaching literacy to Saudi kindergarten students, and reflected on their perceptions and practices as kindergarten teachers. Overall, the participants showed an awareness of how diglossia could generally affect literacy acquisition, as well as an awareness of how different spoken Arabic dialects work with and against LA to varying degrees, causing fluctuations in the diglossic effect across spoken varieties. In their context, however, teachers seemed to find children at a lesser disadvantage and would, therefore, prioritize remediating the challenges children experience as a result of orthographic characteristics of Arabic over the challenges posed by diglossia. Teachers indicated that they still follow certain practices to increase children’s exposure to LA and reinforce their LA knowledge, but without pinpointing any specific diglossia-based instructions—interestingly, they believe this could compromise the orthographic-based instructions they believed essential. Such reflections are discussed in light of the current empirical investigations of Arabic literacy and diglossia and the pedagogical practices they suggest.
This chapter offers a sociolinguistic overview of the diglossic situation in Switzerland. Since diglossia plays a role in shaping language attitudes and policies, it is a key element in understanding the mechanisms of language contact and change. The chapter which is a theoretical one, aims to explicate the features of the Swiss context by addressing the unique aspects of diglossia that fall within the scope of sociolinguistics as they relate to the functional allocation of two language varieties within one society. Charles Ferguson mentioned diglossia in Switzerland in his seminal article to illustrate his original formulation of the phenomenon. Ferguson’s article attracted justified attention but also provoked some criticism, particularly as regards the degree of matching of the theoretical criteria with his Swiss example of the diglossic context. The analysis of the diglossia focuses on German-speaking Switzerland, particularly by pinpointing the circumstances of language acquisition and the differences between standard language usage and dialect. As part of the diglossic communication, there is also the matter of literacy, which involves the relationship of reading and writing to spoken language. Literacy activities are always peculiar to a specific community. In German-speaking Switzerland the choice between written and spoken language goes beyond the basic choice of the medium and generates clear implications for the language’s function, prestige and acquisition, which are realized in a variety of domains. It is concluded that the language situation in German-speaking Switzerland presents a sociolinguistic challenge and defies clear-cut classifications.KeywordsDiglossiaBilingualismDialectSwitzerlandSecond language literacy
The linguistic situation in the Arab world is in an important state of transition, with the “spoken” vernaculars increasingly functioning as written languages as well. While this fact is widely acknowledged and the subject of a growing body of qualitative literature, there is little quantitative research detailing the process in action. The current project examines this development as it is occurring in Tunisia: I present the findings from a corpus study comparing the frequency of Tunisian Arabic–Standard Arabic equivalent pairs in online forum posts from 2010 with those from 2021. The findings show that the proportion of Tunisian lexical items, compared to their Standard Arabic equivalents, increased from a minority (19.7%) to a majority (69.9%) over this period. At the same time, metalinguistic comments on the forum reveal that, although its status is still contentious, Tunisian has become unmarked as a written language. These changes can be attributed to major developments in Tunisian society over the period of study – including internet access and the 2011 revolution. These findings suggest destabilization of the diglossic language situation in Tunisia and a privileging of national identity vis-à-vis the rest of the Arab world.