Joshua A.

Fishman

THE SOCIOLOGY OF LANGUAGE: AN INTERDISCIPLINARY

SOCIAL SCIENCE APPROACH TO LANGUAGE IN SOCIETY*

1.0 INTRODUCTION

Man is constantly using language - spoken language, written lan-

guage, printed language - and man is constantly linked to others via
shared norms of behavior. The sociology of language examines the
interaction between these two aspects of human behavior: use of
language and the social organization of behavior. Briefly put, the
sociology of language focuses upon the entire gamut of topics related
to the social organization of language behavior, including not only
language usage per se but also language attitudes, overt behaviors to-
ward language and toward language users.

1.1 Sociolinguistic Headlines

The latter concern of the sociology of language - overt behavior
toward language and toward language users - is a concern shared by
political and educational leaders in many parts of the world and is
an aspect of sociolinguistics that frequently makes headlines in the
newspapers. Many French-Canadian university students oppose the
continuation of public education in English in the Province of Que-
bec. Many Flemings in Belgium protest vociferously against any-
thing less than full equality - at the very least - for Dutch in the
Brussels area. Some Welsh nationalists daub out English signs along
the highways in Wales and many Irish revivalists seek stronger gov-
ernmental support for the restoration of Irish than that made avail-
* From Current Trends in Linguistics, 12, in press. Reprinted with permission.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
218 Joshua A. Fishman
able during half a century of Irish independence. Jews throughout
the world protest the Soviet government's extermination of Yiddish
writers and the forced closing of Yiddish schools, theaters and pub-
lications.
Swahili, Philipino, Indonesian, Malay and the various provincial
languages of India are all being consciously expanded in vocabulary
and standardized in spelling and grammar so that they can increasingly
function as the exclusive language of government and of higher cul-
ture and technology. The successful revival and modernization of
Hebrew has encouraged other smaller communities - the Catalans, the
Provencals, the Frisians, the Bretons - to strive to save their ethnic
mother tongues (or their traditional cultural tongues) from oblivion.
New and revised writing systems are being accepted - and at times,
rejected - in many parts of the world by communities that hitherto had
little interest in literacy in general or in literacy in their mother tongues
in particular.
Such examples of consciously organized behavior toward language
and toward users of particular languages can be listed almost endlessly.
The list becomes truly endless if we include examples from earlier
periods of history, such as the displacement of Latin as the language
of religion, culture and government in Western Christendom and the
successive cultivation of once lowly vernaculars - first in Western
Europe, and then, subsequently, in Central, Southern and Eastern
Europe, and, finally, in Africa and Asia as well. Instead of being
viewed (as was formerly the case) as merely fit for folksy talk and
for common folk the vernaculars have come to be viewed, used and
developed as independent languages, as languages suitable for all
higher purposes, and as languages of state-building and state-deserving
nationalities. All of these examples feed into modern sociology of
language, providing it with historical breadth and depth in addition to
its ongoing interest in current language issues throughout the world.

1.2 Subdivisions of the Sociology of Language

However, the subject matter of the sociology of language reaches far
beyond interest in case studies and very far beyond cataloging and
classifying the instances of language conflict and language planning
reported in chronicles, old and new. The ultimate quest of the socio-
logy of language is pursued diligently in many universities
throughout the United States and other parts of the world, and is
very far from dealing directly with headlines or news reports. One
part of this quest is concerned with describing the generally accep-
ted social organization of language usage within a speech community

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 219
(or, to be more exact, within speech and writing communities). This
part of the sociology of language - descriptive sociology of language -
seeks to answer the question "who speaks (or writes) what language
(or what language variety) to whom and when and to what end?"
Descriptive sociology of language tries to disclose the norms of language
usage - that is to say, the generally accepted social patterns of lan-
guage use and of behavior and attitude toward language - for parti-
cular social networks and communities, both large and small. Another
part of the sociology of language—dynamic sociology of language—seeks
to answer the question "what accounts for different rates of change in
the social organization of language use and behavior toward language?"
Dynamic sociology of language tries to explain why and how the social
organization of language use and behavior toward language can be
selectively different in the same social networks or communities on
two different occasions. Dynamic sociology of language also seeks to ex-
plain why and how once similar social networks or communities can
arrive at quite different social organizations of language use and beha-
vior toward language.
These two subdivisions taken together i.e. descriptive sociology of
language plus dynamic sociology of language constitute the sociology
of language, a whole which is greater than the mere sum of its parts.

1.3 Language IS Content; the Medium IS (at Least Partly) the

Message
Newspaper headlines with all of their stridency may serve to remind
us of a truism that is too frequently overlooked by too many Ameri-
cans, namely, that language is not merely a means of interpersonal
communication and influence. It is not merely a carrier of content,
whether latent, or manifest. Language itself is content, a referent
for loyalties and animosities, an indicator of social statuses and per-
sonal relationships, a marker of situations and topics as well as of the
societal goals and the large-scale value-laden arenas of interaction that
typify every speech community.
Any speech community of even moderate complexity reveals several
varieties of language, all of which are functionally differentiated from
each other. In some cases the varieties may represent different occupa-
tional or interest specializations ('shop talk', 'hippie talk', etc.) and,
therefore, contain vocabulary, pronunciation and phraseology which
are not generally used or even known throughout the broader speech
community. As a result, the speakers of specialized varieties may
not always employ them. Not only must they switch to other varieties
of language when they interact in less specialized (or differently spe-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
220 Joshua A. Fishman
cialized) networks within the broader speech community of which they
are a part, but most of them do not even use their specialized varieties
all of the time with one another. On some occasions, interlocutors
who can speak a particular specialized variety to one another neverthe-
less do not do so, but, instead switch to a different variety of language
which is either in wider use or which is indicative of quite a different
set of interests and relationships than is associated with their specia-
lized variety. This type of switching represents the raw data of des-
criptive sociology of language, the discipline that seeks to determine
(among other things) who speaks what variety of what language to
whom, when and concerning what.
The varieties of language that exist within a speech community
need not all represent occupational or interest specializations. Some
varieties may represent social class (economic, educational, ethnic) dis-
tinctions within coterritorial populations. 'Brooklynese' and 'Cockney'
English within New York and London, respectively, do not connote
foreignness or even a particular section of the city as much as lower
class status in terms of income, education or ethnicity. Nevertheless,
many individuals who have left lower class status behind can and do
switch back and forth between Brooklynese and more regionally
standard New York English when speaking to each other, depending
on their feelings toward each other, the topic under discussion, where
they happen to be when they are conversing and several other factors,
all of which can exhibit variation and, as a result, can be signalled by
switching from one variety of English to another.
A speech community that has available to it several varieties of
language may be said to possess a verbal repertoire. Such repertoires
may not only consist of different specialized varieties and different
social class varieties but may also reveal different regional varieties
(Boston English, Southern English, Midwestern English and other
widely, and roughly, designated dialects of American English are
regional varieties), if the speech community is sufficiently large such
that enclaves come to arise within it on a geographic basis alone.
Furthermore, multilingual speech communities may employ, for the
purpose of intragroup communication, all of the above types or
varieties of language within each of the codes that the community
recognizes as 'distinct' languages (e.g. within Yiddish and Hebrew,
among most pre-World War II Eastern European Jews; within En-
glish and Hindi, among many upper-class individuals in India
today, etc.).
Regardless of the nature of the language varieties involved in the
verbal repertoire of a speech community (occupational, social class, re-
gional, etc.) and regardless of the interaction between them (for

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 221
initially regional dialects may come to represent social varieties as
well, and vice versa) descriptive sociology of language seeks to disclose
their linguistic and functional characteristics and to determine how
much of the entire speech community's verbal repertoire is available
to various smaller interaction networks within that community since
the entire verbal repertoire of a speech community may be more ex-
tensive than the verbal repertoire controlled by sub-groups within
that community. Dynamic sociology of language on the other hand
seeks to determine how changes in the fortunes and interactions of
networks of speakers alter the ranges (complexity) of their verbal
repertoires.
All in all, the sociology of language seeks to discover not only the
societal rules or norms that explain and constrain language behavior
and the behavior toward language in speech communities but it also
seeks to determine the symbolic value of language varieties for their
speakers. That language varieties come to have symbolic or symp-
tomatic value, in and of themselves, is an inevitable consequence of
their functional differentiation. If certain varieties are indicative of
certain interests, of certain backgrounds, or of certain origins, then
they come to represent the ties and aspirations, the limitations and the
opportunities with which these interests, backgrounds and origins, in
turn, are associated. Language varieties rise and fall in symbolic value
as the status of their most characteristic or marked functions rises
and falls. Varieties come to represent intimacy and equality if they are
most typically learned and employed in interactions that stress such
bonds between interlocutors. Other varieties come to represent
educated status or national identification as a result of the attainments
associated with their use and their users and as a result of their
realization in situations and relationships that pertain to formal
learning or to particular ideologies. However, these functions are
capable of change (and of being consciously changed), just as the
linguistic features of the varieties themselves may change (and may
be consciously changed), and just as the demographic distribution of
users of a variety within a particular speech community may change.
The step-by-step elevation of most modern European vernaculars
to their current positions as languages of culture and technology is
only one example of how dramatically the operative and symbolic
functions of languages can change. Similar changes are ongoing today:
Since the preservation of adequate control over the labour force loomed
so large in the minds of the early planters, various devices have evolved, of
which the maintenance of castelike distance was perhaps the one most
significantly affecting race relations. One thinks immediately of the frequently
cited admonition in the Rabaul Times of August 8, 1926, by a veteran

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
222 Joshua A. Fishman
Territorian, "Never talk to the boys in any circumstances. Apart from your
house-boy and boss-boy, never allow any native to approach you in the
field or on the bungalow veranda." This free advice to the uninitiated
planters was, no doubt, intended to preserve "White prestige", but it was
also conceived as a protective device to "keep labour in its place". So
also the Melanesian Pidgin, which had come into being as a medium of
interchange in trade, subsequently acquired, on the plantations, the charac-
ter of a language of command by which the ruling caste "talked down" to its
subordinates and "put them in their place". A wide range of plantation
etiquette symbolizing proper deference by workers toward their masters and
expressed in expected form of address and servile conduct gave further
protection to the system and any signs of insubordination or "cheekiness"
on the part of the workers might be vigorously punished and rationalised
by the planter as a threat to the system." (Lind 1969, p. 36)

Yet today, barely half a century since Melanesian Pidgin began to

expand, it has been renamed Neo-Melanesian and is being groomed
by many New Guineans to become their country's national language,
and, as such to be used in government, education, mass media, reli-
gion and high culture more generally (Wurm 1961/62).
The sociology of language is the study of the characteristics of
language varieties, the characteristics of their functions, and the
characteristics of their speakers as these three constantly interact,
change, and change one another, both within and between speech
communities.

2.0 SOME REASONS WHY THE SOCIOLOGY OF LANGUAGE HAS

ONLY RECENTLY BEGUN TO DEVELOP

Given the obvious importance of the sociology of language, given its

apparent interest for all who are interested in either or both of its
parent disciplines (as well as for all who wish better to understand
events and processes all over the world), and, finally, given the sub-
stantial applied promise of the sociology of language for educational
and governmental use, it is quite natural to ask: why is the sociology
of languages only now coming into its own? Actually, the sociology of
language, as a field of interest within linguistics and the social sciences,
is not as new as its recent prominence may suggest. The 19th and
early 20th century witnessed many studies and publications that be-
longed to this field (many are cited in Hertzler 1965). Nevertheless, it
is quite true that the disciplinary priorities and biases of both fields
were such that those earlier attempts were prematurely set aside and
only recently has momentum been attained in this field to enable
it to attract and train specialists devoted to it per se (Ferguson 1965,
Fishman 1967b).

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 223

2.1 Invariant Behavior

Linguistics has classically been interested in completely regular or

fully predictable behavior. The p in "pun" is always aspirated by native
speakers of English. The p in "spin" is always unaspirated by
these speakers. This is the kind of entirely determined relationship
that linguistics has classically sought and found - to such an extent
that a highly respected linguist wrote a few decades ago: "if it exists
to some degree, it's not linguistics" (Joos 1950). The implication of
this view is quite clear: linguistics is not interested in 'sometime
things'. The phenomena it describes are either completely determinable
occurrences or non-occurrences. Wherever some other lesser state of
determinacy was noted, e.g. in usage, this was defined as 'exolinguistic',
as 'free variations' that was outside of the realm or the heartland
of linguistics proper.
The social sciences on the other hand, were (and remain) singularly
uninterested in apparently invariant behavior. Any such behavior
could only prompt the observation "so what?" from the social sciences
since their preserve was and is societally patterned variation in be-
havior and the locations of those factors that parsimoniously explain
and predict such variation. If one were to observe to a social scientist
that the same individuals who always wore clothing when they were
strolling on Fifth Avenue never wore any when they were bathing or
showering, his reaction to his brand of societal invariance would be "so
what?"
Given the above basic difference in orientations between its two
parent disciplines, it is not even necessary to add that linguistics was
classically too code-oriented to be concerned with societal patterns in
language usage, or that sociology, e.g., was classically too stratifica-
tionally oriented to be concerned with contextual speaking (or writing)
differences within strata. Fortunately, both fields have recently moved
beyond their classical interests (see Figure 1) and, as a result, fostered
the kinds of joint interests on which the sociology of language now
depends.

2.2 Moderately Variable Behavior

Linguistics has, in recent years, plunged further and further into

'sometimes things' in the realm of language behavior. Some of the
same speakers who say "aint" on certain occasions do not use it on
others, and some of the same cotton-pickers who have such a colorful
and unique vocabulary, phonology and grammar on occasions also

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
224 Joshua A. Fishman
Type of Behavior Linguistics Social Sciences
Invariant Classical Interest No Interest

Moderately Variable Recent Classical Interest

Growing Interest

Highly Variable Possible Future Recent Growing

Interest Interest

F i g u r e 1. T h e Changing Interests a n d E m p h a s e s of Linguistics a n d t h e Social

Sciences with Respect t o Variation in Behavior
(after L a b o v )

share other varieties with their many non-cotton-picking friends and

associates. This is the kind of societally patterned variation in behavior
that social scientists not only recognize and understand, but it is the
kind they are particularly well prepared to help linguists study and
explain. When such behavior is reported the social scientist is oriented
toward locating the smallest number of societal factors that can account
for or predict the usage variation that has been reported.

2.3 Highly Variable Behavior

Finally, and even more recently, even more complex societally pat-
terned variation in behavior has come to be of interest to the social
scientist. This behavior is so complexly patterned or determined that
a goodly number of explanatory variables must be utilized and com-
bined, with various quantitative weights and controls, in order that
their total impact as well as their separate contributions can be gauged.
This kind of highly variable and complexly patterned societal be-
havior obviously exists (and plentifully so) with respect to language
too. However, linguists generally lack the skills of study design, data
collection and data analysis that are required in order to undertake
to clarify such multiply determined language behavior. At this level,
more than at any other, the corpus of language per se is insufficient
to explain a major proportion of the variation in language behavior
that obtains. Nor are a few demographic (age, sex, education), nor a
few contextual (formality-informality in role relationships), nor a few
situational factors sufficient for this purpose. Rather, predic-
tors of all of these kinds are needed and, to the extent that this is so,
their joint or combined use will result in far greater explanatory or
predictive power than would any two or three of them alone. The
social sciences themselves have only rather recently become accustomed

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 225
to working with large numbers of complexly interrelated and differen-
tially weighted variables. This is obviously a level of analysis which
will become available to the sociology of language only if there is
genuine cooperation between linguists and social scientists.
The sociology of language is thus a by-product of a very necessary
and very recent awareness on the part of linguistics and the social scien-
ces that they do indeed need each other in order to explore their joint
interests in a productive and provocative manner. This cooperative
attitude has yielded important results in the few years that it has
been actively pursued (Grimshaw 1969, Hymes 1967a) and we may
expect even more from it in the future when a greater number of in-
dividuals who are themselves specialists in both fields simultaneously
(or in the joint field per se) will have been trained.

2.4 Variability and Predictability

While it is, of course, true that the more variable behavior is, the
more the factors that need to be located in order to account for it in
any substantial way, the less predictable the behavior is until the
proper factors have been located and combined or weighted in the
most appropriate ways. Ultimately, however, if the quest of rigorous
data collection and data analysis is successful, as high a level of predic-
tability or explainability may be attained with respect to complexly
determined and highly variable behaviors as with the far less and the
somewhat less complexly determined and variable ones. Thus, the
methodological differences that have existed between linguistics and
sociology have been primarily differences in the extent to which a
very few well chosen parameters could account substantially for the
behaviors that the respective disciplines choose to highlight. Ultimately,
all disciplines of human behavior - including linguistics and sociology
- strive to locate and to interrelate the most parsimonious set of ex-
planatory-predictive variables in order to maximally account for the
variability to which their attention is directed.
With respect to societally patterned language behavior, there is
doubtlessly variability that can be well-nigh perfectly accounted for
by a very few well selected intra-code positional factors. Social scien-
tists should recognize such behavior for it not only leads them to a
recognition of linguistics per se but to the clearer realization that the
entire world of socially patterned variability in language behavior
still remains to be explored - and to be explored by linguists and
social scientists together - after the variability explainable on the
basis of intra-code factors alone has been accounted for. However,

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
226 Joshua A. Fishman
at that level of inquiry it is not possible to simply put linguistics aside
and, turning to more exciting and difficult tasks, simply to 'do social
science'. Studies of more complexly determined and more highly
variable socially patterned language behaviors still require rigorous
descriptions and analyses of language usage per se and for such ana-
lyses the social sciences will always be dependent on linguistics.

3.0 SOME BASIC SOCIOLINGUISTIC CONCEPTS

The sociology of language deals with quite a range of topics: small

group interaction and large group membership, language use and
language attitudes, language-and-behavior norms as well as changes
in these norms. We expect to deal with all of these topics, at least
briefly, in this presentation, and, necessarily, to introduce the tech-
nical terms and concepts which specialized fields of discourse inevitably
require. However, before moving into any of these more specialized
substantive topics there are a number of basic sociolinguistic concepts
that are of such general intertopic utility that we had best pause to
consider them here, rather than to permit them to remain as primitives
any longer.

3.1 Language-Dialect-Variety

The term variety is frequently utilized in sociology of language as a non-

judgemental designation. The very fact that an objective, unemotional,
technical term is needed in order to refer to 'a kind of language' is, in
itself, an indication that the expression 'a language' Js often a judg-
mental one, a term that is indicative of emotion and opinion, as well
as a term that elicits emotion and opinion. This is a n important fact
about languages and one to which we will return repeatedly. As a
result, we will use the term 'variety' in order not to become trapped
in the very phenomena that we seek to investigate, namely, when and
by whom is a certain variety considered to be a language and when
and by whom is it considered something else.
Those varieties that initially and basically represent divergent
geographic origins are known as dialects (Ferguson and Gumperz
1960; Halliday 1964b). It is in this purely objective sense of the
word that it is used in such terms as dialectology and dialect geo-
graphy within linguistics, and it is in this sense that the sociology of
language employs it as well. However, dialects may easily come to
represent (to stand for, to connote, to symbolize) other factors than

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 227
geographic ones. If immigrants from region A come to be a large
portion of the poor, the disliked and the illiterate in region B, then
their speech variety (Dialect A) will come to stand for much more than
geographic origin alone in the minds of the inhabitants of region B.
Dialect A will come to stand for lower social status (educationally,
occupationally) than will dialect B. In this way what was once a
regional variety (in the sense that a particular time its speakers were
viewed as merely concentrated in a particular area) may come to be
viewed (and to function) much more importantly as a social variety or
sociolect (Blanc 1964). Furthermore, if the speakers of variety A are
given hardly any access into the interaction networks of region B, if
they marry primarily only each other, engage primarily in their original
regional customs and continue to value only each other's company, they
may, in time, come to consider themselves a different society, with
goals, beliefs and traditions of their own. As a result, variety A may no
longer be viewed as a social variety but, rather, as an ethnic or religious
variety and, indeed, it may come to be cultivated as such to the point of
being viewed as a separate language (Kloss 1967; Fishman 1968c).
However, within the community of A speakers there may come to
be some who have learned B as well. They may utilize A with each
other for purposes of intimacy and in-group solidarity but they may
also use B with each other for occupational and deferential purposes.
Thus, for them, A and B will be contrasted and complementary
functional varieties, with B also being (or including) a specialized
(occupational or other experiential) variety and, therefore, in some
ways different than variety B as used by others (Weinreich M. 1953).
The above theoretical sketch has more than general didactic
value. It represents the route that many varieties - regional and social
- have travelled in the past and the route on which still others are em-
barked at this very time (Haugen 1966c; Deutsch 1966). Never-
theless, it is the general point that is of particular value to us at this
juncture. Varieties may be viewed as regional at one time and social
at another. Varieties may be reacted to as regional within the speech
community of their users and as social (or ethnic) by outsiders.
Varieties may have additional functional uses for some of their users
that they do not have for others who possess fewer contrasted varieties
in their verbal repertoires. Thus, the term variety - unlike the term
dialect - indicates no particular linguistic status (other than diffe-
rence) vis-a-vis other varieties. A dialect must be a regional i«ft-unit
in relation to a language, particularly in its vernacular or spoken reali-
zation. 'Language' is a superordinate designation; 'dialect' is a sub-
ordinate designation. Both terms require that the entire taxonomy to
which they pertain be known before they themselves can be accepted.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
228 Joshua A. Fishman
The sociology of language is interested in them only in so far as mem-
bers of speech communities contend over which is which, and why.
As the result of such contention varieties hitherto considered to be
dialects may throw off their subordination and be 'promoted' by
their speakers to official and independent status, whereas formerly
independent languages may be subordinated. The term variety, on
the other hand, merely designates a member of a verbal repertoire.
Its use implies only that there are other varieties as well. These can
be specified by outsiders on the basis of the phonological, lexical and
grammatical differences that they (the varieties) manifest. Their
functional allocations, however - as languages or as dialects - are
derivable only from societal observation of their uses and users rather
than from any characteristics of the codes themselves.
Varieties change over time but varieties are also changed, either
by drift or by design. Varieties that have been used in palaces and
universities may later come to be used only by the rural and unlet-
tered. In this process their lexicons may well become impoverished
(hundreds or thousands of the terms once needed dropping into
disuse). At the same time lexicons, grammars as well as phonologies
may become much influenced by other temporarily more prestigeful
and possibly genetically unrelated varieties. Conversely, varieties that
had never been used outside of the most humble speech networks
may be elevated in function, increased in lexicon and purified or en-
riched in whatever direction their circumstantially improved speak-
ers may desire (Kloss 1952, 1967; Fishman 1968c). All varieties of all
languages are equally expandable and changeable; all are equally
contractable and interpenetrable under the influence of foreign models.
Their virtues are in the eyes (or ears) of their beholders. Their
functions depend on the norms of the speech communities that employ
them. These norms, in turn, change as speech communities change in
self-concept, in their relations with surrounding communities and in
their objective circumstances. Finally, such changes usually lead to
changes in the varieties themselves. Speech communities and their
varieties are not only interrelated systems; they are completely inter-
dependent systems as well. It is this interdependence that the sociology
of language examines.

3.2 Major Types of Attitudes and Behaviors toward Language

One of the best known societal behaviors toward language is Standar-
dization, i.e., "the codification and acceptance, within a community
of users, of a formal set of norms defining 'correct' usage " (Stewart
1968). Codification is, typically, the concern of such language 'gate-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 229
keepers' as scribes, storytellers, grammarians, teachers and writers, i.e.,
of certain groups that arise in most diversified societies and whose
use of language is professional and conscious. Given codification as
a goal, this desired 'good' is formulated and presented to all or part
of the speech community via such means as grammars, dictionaries,
spellers, style manuals, and exemplary texts, whether written or oral.
Finally, the acceptance of the formally codified (i.e. the standardized)
variety of a language is advanced via such agencies and authorities
as the government, the educational system, the mass media, the
religious institutions and the cultural 'establishment'. The standard
variety then becomes associated with such institutions, the types of
interactions that most commonly occur within them, and the values
or goals they represent (Haugen 1966a).
Note that not all languages have standard varieties. Note also, that
where a standard variety does exist it does not necessarily displace the
non-standard varieties from the linguistic repertoire of the speech
community for functions that are distinct from but complementary
to those of the standard variety. Note, additionally, that there may
be several competing standard varieties in the same speech community.
Note, finally, that hitherto non-standard varieties may themselves
undergo standardization whereas hitherto standardized varieties may
undergo de-standardization as their speakers no longer view them
as worthy of codification and cultivation. Standardization is not
a property of any language per se, but a characteristic societal treat-
ment of language given sufficient societal diversity and need for sym-
bolic elaboration.
Another common societal view of language is that which is con-
cerned with its autonomy, i.e., with the uniqueness and independence
of the linguistic system or, at least, of some variety within that system.
Autonomy is often of little concern to speech communities whose
languages differ markedly from each other. These may be said to be
autonomous by dint of sheer abstand or linguistic distance between
them (Kloss 1952; Kloss 1967). On the other hand, where languages
seem to be quite similar to each other - phonologically, lexically and
grammatically - it may be of great concern to establish their autonomy
from each other, or at least that of the weaker from the stronger. Were
such autonomy not to be established it might occur to some that one
was 'no more than' a dialect (a regional variety) of the other, a sub-
servience which may become part of a rationale for political subser-
vience as well.
A major vehicle of fostering autonomy views concerning a lan-
guage is its standardization. The availability of dictionaries and
grammars is taken as a sure sign that a particular variety is 'really a

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
230 Joshua A. Fishman
language'. However, the availability of dictionaries and grammars not
only represents autonomy, but also cultivates and increases it by
introducing new vocabulary and stressing those phonological and
grammatical alternatives that are most different from those of any
given autonomy-threatening contrast language. "Heroes are made,
not born." The same is true of the autonomy of genetically (historical-
ly) related languages. Their autonomy has to be worked on. It is not
autonomy by abstand, but, rather, by ausbau (by effort, and, often,
by fiat or decree), and pertains particularly to their standard (and
most particularly to their written standard) varieties.
It is a characteristic of the newly rich to supply their own ancestors.
In a similar vein those speech communities, the autonomy of whose
standard variety is based most completely on ausbau-activity, are
also most likely to be concerned with its historicity, that is with its
'respectable' ancestry in times long past. As a result, many speech
communities create and cultivate myths and genealogies concerning
the origin and development of their standard varieties in order to de-
emphasize the numerous components of more recent vintage that
they contain (Ferguson 1959b). As a result of the widespread prefer-
ence for historicity, currently utilized (and recently liberated or
standardized) varieties are found to be derived from ancient proto-
types that had largely been forgotten, or are found to be the language
of the gods, or to have been created by the same miraculous and
mysterious forces and processes that created the speech community
itself, etc. Thus, a variety achieves historicity by coming to be asso-
ciated with some great ideological or national movement or tradition
(Fishman 1965c). Usually, historicity provides the ex post facto ratio-
nale for functional changes that have transpired with respect to the
verbal repertoire of a speech community.
Finally, a speech community's behavior toward any one or another
of the varieties in its linguistic repertoire is likely to be determined, at
least in part, by the degree to which these varieties have visible vi-
tality, i.e., interaction networks that actually employ them natively
for one or more vital functions. The more numerous and the more
important the native speakers of a particular variety are the greater
its vitality and the greater its potential for standardization, autonomy
and historicity. Conversely, the fewer the number and the lower the
status of the native speakers of a variety, the more it may be reacted to
as if it were somehow a defective or contaminated instrument, unworthy
of serious efforts or functions, and lacking in proper parentage or
uniqueness. As usual such biased views are likely to be self-fulfilling
in that when the numbers and the resources of the users of a given
variety dwindle they are less likely to be able to protect its standardi-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 231

zation, autonomy or historicity from the inroads of other speech com-

munities and' their verbal repertoires and language-enforcing resources.

ATTRIBUTES*
1 2 3 4 VARIETY-TYPE SYMBOL

+ + + + Standard S

— + + + Vernacular V

— — + + Dialect D

— — — + Creole K

— — — — Pidgin P

+ + + — Classical C

+ + — — Artificial A

*1 = standardization, 2 = autonomy, 3 = historicity, 4 = vitality

Figure 2. Evaluations of different types of language varieties (Stewart 1968)

Given these four widespread patterns of societal belief and beha-

vior toward language, it is possible to define seven different kinds
of varieties, depending upon their absence or presence at any given
time (Figure 2). Note, however, that any speech community may
include in its repertoire a number of such varieties which are differen-
tiable on the basis of the four widespread belief-and-behavior systems
just discussed. Furthermore, occupational, social class and other ex-
periential subvarieties are likely to exist within most of the varieties
listed in Figure 2. Indeed, the members of any given community may
not agree as to whether standardization, autonomy, historicity and/
or vitality are absent or present in connection with one or more of
the varieties in their repertoire. After all, these dimensions are highly
evaluational, rather than objective characteristics of language varieties
per se, and as such, variation in evaluations may be expected both
synchronically (at any particular time) as well as diachronically (across
time).
In some speech communities deference due an interlocutor with whom
one stands in a particular role-relationship may be indicated by switch-
ing from one social class variety or from one dialect to another.
In other speech communities this very same function may be realized

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
232 Joshua A. Fishman
by switching from a dialect to the standard variety (which latter
variety, alone, may possess formal verb-forms and pronouns of res-
pect). In yet another speech community a switch from one language
to another (or from a dialect of one language to the standard variety
of another) may be the accepted and recognized realization pattern
for deferential interaction. While the precise nature of the switch will
depend on the repertoire available to the speech community, switching
as such and the differentia and concepts by means of which it may
be noted and explained are of constant interest to sociolinguistic
method and theory.

3.3 Speech Community

Speech community (a term probably translated from the German
Sprachgemeinschaft), like variety, is a neutral term. Unlike other
societal designations it does not imply any particular size nor
any particular basis of communality. A speech community is one, all
of whose members share at least a single speech variety and the norms
for its appropriate use. A speech community may be as small as a
single closed interaction network, all of whose members regard each
other in but a single capacity. Neither of these limitations, however,
is typical for speech communities throughout the world and neither
is typical for those that have been studied by sociologists of language.
Isolated bands and nomadic clans not only represent small speech
communities but speech communities that also exhaust their mem-
bers' entire network-range while providing little specialization of roles
or statuses. Such speech communities usually possess very limited
verbal repertoires in terms of different varieties, primarily becaua.
one individual's life experiences and responsibilities are pretty much
like another's. Nevertheless, such similarity is likely to be more ap-
parent than real. Even small and total societies are likely to differen-
tiate between men and women, between minors and adults, between
children and parents, between leaders and followers. Indeed, such
societies are likely to have more contact with the 'outside world' than
is commonly imagined, whether for purposes of trade or exogamy
(Owens 1965). Thus, even small, total societies reveal functionally
differentiated linguistic repertoires (and, not infrequently, intra-
group bilingualism as well) based upon behaviorally differentiated in-
teraction networks.
Such small and total (or nearly total) societies differ, of course, from
equally small or even smaller family networks, friendship networks,
interest networks, or occupational networks within such larger speech
communities as tribes, cities or countries. In the latter cases the inter-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 233
action networks are not as redundant as in the former (i.e., one
more frequently interacts with different people in one's various roles
as son, friend, work colleague, party member, etc.). However, var-
ieties are needed not only by diverse small networks but also by
large networks of individuals who rarely, if ever, interact but who
have certain interests, views and allegiances in common. Thus, not
only are network redundancy and network size attributes that charac-
terize and differentiate speech communities but so is the extent to
which their existence is experiential rather than merely referential.
One of the characteristics of large and diversified speech communities
is that some of the varieties within their verbal repertoires are primar-
ily experientially acquired and reinforced by dint of actual verbal
interaction within particular networks, while others are primarily re-
ferentially acquired and reinforced by dint of symbolic integration
within reference-networks which may rarely or never exist in any
physical sense. The 'nation' or the 'region' are likely to constitute a
speech community of this latter type and the standard ('national')
language or the regional language is likely to represent its corres-
ponding linguistic variety.
Many American cities present ample evidence of both of these
bases - verbal interaction and symbolic integration - for the func-
tioning of speech communities. Every day hundreds of thousands
of residents of Connecticut, Up-State New York and various parts
of Pennsylvania come to New York City to work and shop. In terms
of waking hours of actual face-to-face verbal interaction these
speakers of dialects that differ from New York City English may
talk more, and more frequently, to New Yorkers than they do to in-
habitants of their places of residence and to speakers of their local
dialects. How then can we explain the fact that not only do most
of them differentially utilize the markers of their local dialects (and
not only during the evenings, week-ends and holidays when they are at
home rather than at work) but the simultaneous fact that many of them
can and do also employ a more regionally neutral variety, which
is their approximation to 'Standard American', as distinct from New
York City English on the one hand and Lower Connecticut Village
English on the other? Obviously, the 'Standard American' of these
commuters to New York City cannot be based on much verbal in-
teraction with a separate network known as 'the American people'.
Nor can it be based upon any other interaction network, however
referred to, whose speakers use 'Standard American' and it alone.
There is no other alternative but to conclude that the speech community
of 'Standard American' represents a reference group for the den-
izens of Connecticut villages while 'Standard American' itself is a

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
234 Joshua A. Fishman
variety that has the functions of 'symbolic integration with the nation'
in their linguistic repertoire.
Thus, some speech communities and their linguistic repertoires are
preserved primarily by communication gaps that separate them from
other communities and their repertoires. Other speech communities and
their repertoires are preserved primarily by the force of symbolic
(attitudinal) integration even in the absence of face-to-face inter-
action. Many speech communities contain networks of both types.
Many networks contain both kinds of members. Societal norms that
define communicative appropriateness can apply with equal force
and regularity regardless of whether direct interaction or symbolic
integration underlies their implementation.
As mentioned earlier, the standard variety of a language is likely
to be that variety that stands for the nation as a whole and for its most
exalted institutions of government, education and High Culture in
general. It is this variety which comes to be associated with the
mission, glory, history and uniqueness of an entire 'people' and, in-
deed, it is this variety which helps unite individuals who do not
otherwise constitute an interaction network into a symbolic speech com-
munity or 'people'. Thus it is that standard varieties and larger-than-
face-to-face speech communities are historically and functionally
interdependent. While interaction networks of speakers of standard
varieties doubtlessly do exist (literati, scholars, social and educational
elites, etc.), these are likely to arrive at somewhat specialized usages,
on the one hand, as well as to require a non-standard variety, on the
other hand, if they are to engage in more intimate and informal kinds
of interactions as well. Thus, the standard language per se, without
further differentiation or accompaniment, is most fitted for commu-
nication across large but referential (or non-interacting) networks
such as those involving the mass media, governmental pronouncements,
legal codes and textbooks. The standard variety is the 'safest' for those
communications in which a speaker cannot know his diversified and
numerous listeners (Joos 1959). However, the more the communi-
cation is expected to live on, independently of both speaker and listen-
er (or sender and receiver), over an appreciable period of time, the
more it will be viewed as archaic (or classical) rather than merely
'standard'.
A basic definitional property of speech communities is that they
are not defined as communities of those who "speak the same lan-
guage' (notwithstanding Bloomfield 1933), but, rather, as communi-
ties set off by density of communication or/and by symbolic integra-
tion with respect to communicative competence regardless of the
number of languages or varieties employed (Gumperz 1964a). The

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 235
complexity of speech communities thus defined varies with the extent
of variation in the experiential and attitudinal networks which they
subsume. Speech communities can be so selected as to include greater
or lesser diversity on each of these grounds. In general the verbal
repertoire of a speech community is a reflection of its role repertoire
(in terms of both implemented and ideologized roles). This reflection
pertains not only to repertoire range but also to repertoire access and
fluidity.
Speech communities with a larger role repertoire reveal a larger
verbal repertoire as well (Gumperz 1962). Communities most of
whose members are restricted in daily experiences and in life aspirations,
will also tend to show little linguistic range in terms of differentiable
varieties. This tends to be the case not only in the small, total commu-
nities that were mentioned earlier but also, some suspect, in large,
democratic, industrialized communities of the most modern sort.
Actually, both kinds of speech communities show more repertoire
range (in terms of verbal repertoire and in terms of role repertoire)
than is obvious on superficial inspection. Nevertheless, they both tend
to have narrower (and less diversified) ranges than are encountered
in the stratified speech communities that exist in intermediate socie-
ties of the traditional, non-Western World. Whereas the modern, re-
latively open speech community tends to reveal several varieties of
the same language the more traditional speech community will typi-
cally reveal varieties of several languages (see Figure 3).

Speech Speech Speech Speech

Societal Community Community Community Community
Domain 1 2 3 4
Home ai Cl Cl dl
School and Culture a2 b3/C2 b2/C2 a2
Work a3 C3 d2 d2
Government a
2 bi a2 a2
Church ei b2 b2 ei
(Moscow, (Mea Shearim, (Ostropol, (Ostropol,
1960) 1966) 1905) 1905)
[Russians] [Jews] [Jews] [Ukrainians]

Some communities have more obviously diversified repertoires than others (e.g.,
SCI utilizes 3 varieties of one language and one of another, whereas SC3 utilizes
varieties of four different languages). Varieties that are related to one societal
domain in one SC (e.g., b2 in SC2) may be associated with more or different
societal domains in another SC (e.g., b2 in SC3). All speakers of varieties of a
particular language do not necessarily constitute a single speech community.
Figure 3. Speech Communities and Verbal Repertoires
(based upon concepts of Gumperz, 1964a and elsewhere)

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
236 Joshua A. Fishman
These two types of speech communities are also quite likely to differ
in the extent to which their members have access to the roles and to
the varieties available in the respective repertoires of their commu-
nities. In the more traditional speech communities access to certain
roles is severely restricted and is attained, in those cases in which
access to new roles is available, on the basis of ascription. Those
whose ancestry is inappropriate cannot attain certain new roles, re-
gardless of their personal achievement. Similarly, access to an expanded
verbal repertoire is also severely restricted, most varieties not learned
in childhood being available only to those who can afford to devote
many years of patient and painstaking formal study to their acquisition.
Both of these conditions are not nearly so likely to exist in modern,
personal-achievement-oriented societies, although their lack of com-
pletely equal and open access is evident to all students of the disadvan-
taged (including Negro non-standard speech) in the midst of America's
plenty.
In more traditional societies in which status is based on ascription
there is also likely to be more role compartmentalization. Thus, not only
are certain individuals barred from enacting certain roles but, in general,
the rights and duties thai constitute particular roles are more distinct
and the transition from one role to the next, for members of those
classes who may enter into them, are ritually governed, as are the
roles themselves. Such societies also tend to reveal marked verbal
compartmentalization as well (McCormack 1960). When an indi-
vidual speaks language or variety A he takes great care not to switch
into B and not to slip into traces of B, whether phonologically, lexi-
cally or grammatically. Each variety is kept separate and uncontam-
inated from the other just as is each role. How different such com-
partmentalization is from the fluidity of modern democratic speech
communities in which there is such frequent change from one role to
the other and from one variety to another that individuals are frequent-
ly father and pal, or teacher and colleague, simultaneously or in rapid
succession! The result of such frequent and easy role shifts is often
that the roles themselves become more similar and less distinctive or
clearcut. The same occurs in the verbal repertoire as speakers
change from one variety (or language) to another with greater fre-
quency and fluidity. The varieties too tend to become more similar
as the roles in which they are appropriate become more and more
alike. This is particularly likely to occur, as we will see below, among
lower class speakers whose mastery of the more formal roles and
varieties available to their speech communities is likely to be marginal
at best.
Thus, just as varieties are characterizable by a small number of

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 237
attributes and their combinations, so is this true of the attributes that
characterize speech communities at the most general level. The inter-
actional basis of speech communities, their symbolic-integrative basis,
their size, repertoire range, repertoire access and repertoire compart-
mentalization are all concepts that we will need to refer to again
and again in the pages that follow.

4.0 INTERACTIONAL SOCIOLOGY OF LANGUAGE:

MICRO A N D MACRO

Boss Carmen, do you have a minute?

Secretary Yes, Mr. Gonzalez.
Boss I have a letter to dictate to you.
Secretary Fine. Let me get my pen and pad. I'll be right back.
Boss Okay.
Secretary Okay.
Boss Okay, this is addressed to Mr. William Bolger.
Secretary That's B-o-r-g-e-r?
Boss B-o-1
Secretary Oh, oh, I see.
Boss Okay. His address is in the files.
Secretary Okay.
Boss Okay.. Dear Bill, Many thanks for telling me about your
work with the Science Research Project. The information
you gave me ought to prove most helpful.
Secretary That was "The information you gave me ought to prove
most helpful"?
Boss Correct.
Secretary Okay.
Boss Okay, ah. I very much appreciate the time you gave me.
Never mind, strike that out. Ah, enclosed are two of the
forms that you let me borrow. I'll be sending back the data
sheets very soon. Thanks again. I hope that your hospital
stay will be as pleasant as possible and that your back
will be soon in top shape. Will soon be in top shape. It
was nice seeing you again. Sincerely, Louis Gonzalez.
Secretary Do you have the enclosures for the letter Mr. Gonzalez?
Boss Oh yes, here they are.
Secretary Okay.
Boss Ah, this man William Bolger got his organization to
contribute a lot of money to the Puerto Rican parade.
He's very much for it.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
238 Joshua A. Fishman
¿Tú fuiste a la parada?
(Did you go to the parade?)
Secretary Sí, yo fui.
(Yes, I went.)
Boss ¿Si?
(Yes?)
Secretary Uh huh.
¿Y cómo te estuvo?
(and how did you like it?)
Secretary Ay, lo mas bonita.
(Oh, very pretty.)
Boss Sí, porque yo fui y yo nunca había participado en la parada
(Yes, because I went and I had never participated in the
parade
y este año me dió curiosidad por ir a ver como era y estuvo
eso
and this year I became curious to go and see how it was and
that was
fenómeno. Fui con mi señora y con mis nenes y a ellos
también
a phenomenon. I went with my wife and my children and
they also
le gustó mucho. Eh, y tuve un día bien agradable. Ahora
lo que
liked it very much. And I had a very pleasant day. Now
me molesta a mí es que las personas cuando viene una
cosa así,
what bothers me is that people when something like this
comes along,
la parada Puertorriqueña o la fiesta de San Juan, corren de
la
the Puerto Rican parade, or the festival of San Juan they
run from
casa a participar porque es una actividad festiva, alegre, y
sin
the house to participate because it is a festive activity,
happy, and
embargo, cuando tienen que ir a la iglesia, o la misa para
pedirle...
then, when they have to go to church or to mass, to ask . . . )
Secretary (Laughter)
Boss A Diós entonce no van
(God then they don't go.)

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 239

Secretary Si, entonces no van.

(Yes, then they don't go.)
Boss Pero, asi es la vida, caramba.
(But that's life, you know.)
Do you think that you could get this letter out today?
Secretary Oh yes, I'll have it this afternoon for you.
Boss Okay, good, fine then.
Secretary Okay.
Boss Okay.

If we carefully consider the above conversation it becomes evident

that it reveals considerable internal variation. Speaker A does not al-
ways speak in the same way nor does his interlocutor, Speaker B. Were
it possible for us to listen to the original tapes of this conversation,
several kinds of variation within each of them would become evi-
dent to us: variations in speed of speaking, variations in the extent
to which Spanish phonology creeps into English discourse, and, vice
versa, variations in the extent to which English phonology creeps into
the Spanish discourse, etc. However, even from the conventionally
(orthographically) rendered transcription available to us on the previous
pages one kind of variation remains exceedingly clear: that from
Spanish to English or from English to Spanish for each speaker. It is
precisely because bilingual code switching is often more noticeable than
other kinds of sociolinguistic variation that bilingualism is so commonly
examined in sociolinguistic theory and research. However, the concepts
and findings that derive from such examinations must be provocative
and illuminating for the sociology of language more generally, and,
indeed, that is the case, for the societal patterning of bilingual interaction
is merely an instance (hopefully a more obvious and, therefore, pedagog-
ically useful instance) of the vastly more general phenomenon of
societal patterning of variation in verbal interaction.
How shall we describe or measure the phenomenon of interest to
us: societal patterning of variation in verbal interaction? Usefully
accurate description or measurement is certainly the basic problem
of every scientific field of endeavor. Most of mankind has con-
stantly been immersed in a veritable ocean of cross-currents of talk.
Nevertheless, as with most other aspects of everyday social behavior,
it is only in very recent days that man has begun to recognize the
latent order and regularity in the manifest chaos of verbal interaction
that surrounds him.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
240 Joshua A. Fishman
4.1 How should Talk be Described Contextually?

How should 'talk' be described contextually in order to best reveal

or discover its social systematization (assuming that its 'basic' lin-
guistic description is already available)? Let us begin with some
passages of actual 'talk', making sure to preserve its verbatim form
(preferably by utilizing sensitive audio and visual recording equip-
ment) rather than merely summarizing the content of such talk. The
smallest sociolinguistic unit that will be of interest to us is a speech
act: a joke, an interjection, an opening remark (Schegloff, 1968), a
question, in general - a segment of talk that is also societally recog-
nizable and reoccurring. Speech acts are normally parts of somewhat
larger speech events, such as conversations, introductions, lectures,
prayers, arguments, etc. (Hymes 1967b), which, of course, must also
be societally recognizable and reoccurring.
If we note that a switch has occurred from variery a to variety b
- perhaps from a kind of Spanish to a kind of English, or from more
formal English to less formal English, or from regionally neutral,
informal Spanish to Jibarro (rural) informal Spanish - the first question
that presents itself is whether one variety tends to be used (or used
more often) in certain kinds of speech acts or events, whereas the other
tends to be used (or used more often) in others. Thus, were we aware
of the speech acts recognized by bilingual Puerto Rican youngsters in
New York, we might venture to explain a switch such as the fol-
lowing:

First Girl Yes, and don't tell me that the United States is the only
one that has been able to in Puerto Rico
Boy Okay so you have a couple of people like Moscoso and
Luis Ferrer.
First Girl ¡Un momento!
Boy ¡Bueno!
First Girl ¡Un Momento!
Boy Have you got people capable of starting something like . . .
like General Motors?

as being related to the act of interruption or disagreement in the

midst of a somewhat specialized argument. There may be a pro-
blem, however, when testing this interpretation, in determining the
speech acts and speech events that are to be recognized within a
speech community.
Certainly, it is not appropriate to simply apply the system of
acts and events that has been determined for one speech community

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 241
in the study of another, without determining first its appropriateness
in the second community. Similarly, it is not sufficient for the investi-
gator, no matter how much experience he has had with the verbal
behavior of a particular speech community, merely to devise as
detailed a listing of speech acts and events as he can. Such a list runs
the decided risk of being etic rather than emic, i.e., of making far
too many, as well as behaviorally inconsequential, differentiations, just
as was often the case with phonei/c vs. phonemic analysis in lin-
guistics proper. An emic set of speech acts and events must be one
that is validated as meaningful via final recourse to the native
members of a speech community, rather than via appeal to the in-
vestigator's ingenuity or intuition alone.
An emic set of speech acts and speech events is best approximated,
perhaps along a never-ending asymptote, by playing back recorded
samples of 'talk' to native speakers and by encouraging them to react
to and comment upon the reasons for the use of variety a 'here' as
contrasted with the use of variety b 'there'. The more the sensitive
investigator observes the speech community that he seeks to socio-
linguistically describe the more hunches he will have concerning
functionally different speech acts and speech events. However, even
the best hunches require verification from within the speech com-
munity. Such verification may take various shapes. The views of both
naive and skilled informants may be cited and tabulated as they
comment upon recorded instances of variation in 'talk' and as they
reply to the investigator's patient probes and queries as to "Why didn't
he say 'Just a minute!' instead of '¡Momento!'? Would it have meant
something different if he had said that instead? When is it appropriate to
say '¡Momento!' and when is it appropriate to say 'Just a Minute!'
(assuming the persons involved known both languages equally well)?",
etc. Once the investigator has demonstrated (not merely assumed or
argued) the validity of his sets of functionally different speech acts
and events he may then proceed to utilize them in the collection and
analysis of samples of talk which are independent of those already
utilized for validational purposes. Such, at least, is the rationale of
research procedure at this micro-level of sociolinguistic analysis, al-
though the field itself is still too young and too linguistically oriented
to have produced many instances of such cross-validation of its
social units selected for purposes of sociolinguistic analysis.

4.2 Micro-level Analysis in the Sociology of Language

Sociolinguistic description may merely begin - rather than end - with
the specification and the utilization of speech acts and events, depen-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
242 Joshua A. Fishman
ding on the purpose of a particular research enterprise. The more
linguistically oriented a particular study may be, the more likely it
is to remain content with micro-level analysis, since the micro-level in
the sociology of language is already a much higher (i.e., a more con-
textual and complicated) level of analysis than that traditionally em-
ployed within linguistics proper. However, the more societally orien-
ted a particular sociolinguistic study may be, the more concerned
with investigating social processes and societal organization, per se,
the more likely it is to seek successively more macro-level analyses.
Micro-level sociology of language (sometimes referred to as ethno-
methodological) constitutes one of the levels within sociolinguistic
inquiry (Garfinkel 1967; Garfinkel and Sacks, in press). The various
levels do not differ in the degree to which they are correct or accurate.
They differ in purpose and, therefore, in method. We can trace only
a few of the successive levels in this SECTION, primarily in order to
demonstrate their similarities and their differences.
One of the awarenesses to which an investigator may come
after pondering a mountain of sociolinguistic data at the level of
speech acts and events is that variation in 'talk' is more common and
differently proportioned or distributed between certain interlocutors
than it is between others (Schegloff 1968). Thus, whereas either the
boy or the girl in Conversation 2 may initiate the switch from one
language to another, it may seem from Conversation 1 that the boss
is the initiator of switching far more frequently than is the secretary.
Therefore, while a great deal of switching is functionally metaphorical,
i.e., it indicates a contrast in emphasis (from humor to seriousness,
from agreement to disagreement, from the unessential or secondary to
the essential or primary, in any interchange already underway in a
particular language variety), interlocutors may vary in the extent to
which they may appropriately initiate or engage in such switching,
depending on their role-relationship to each other. Note, however, that
it is necessary for a certain appropriateness to exist between a variety
and certain characteristics of the social setting before it is possible to
utilize another variety for metaphorical or contrastive purposes.

4.3 Role-relationships
Any two interlocutors within a given speech community (or, more
narrowly, within a given speech network within a speech community)
must recognize the role-relationship that exists between them at any
particular time. Such recognition is part of the communality of norms
and behaviors upon which the existence of speech communities depend.
Father-son, husband-wife, teacher-pupil, clergyman-layman, employer-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 243
employee, friend-friend: these are but some examples of the role-
relationships that may exist in various (but not in all) speech commu-
nities (Goodenough 1965). Role-relationships are implicitly recognized
and accepted sets of mutual rights and obligations between members
of the same socio-cultural system. One of the ways in which members
reveal such common membership to each other, as well as their recog-
nition of the rights and obligations that they owe toward each other, is
via appropriate variation (which, of course, may include appropriate non-
variation) of the way(s) they talk to each other. Perhaps children should
generally be seen and not heard, but when they are heard, most societies
insist that they talk differently to their parents than they do to their
friends (Fischer 1958). One of the frequent comments about American
travelers abroad is that they know {at most) only one variety of the
language of the country they are visiting. As a result, they speak in the
same way to a child, a professor, a bootblack and a shopkeeper, thus
revealing not only their foreignness, but also their ignorance (of the
appropriate ways of signalling local role-relationships).
It is probably not necessary, at this point, to dwell upon the kinds
of variation in talk that may be required (or prohibited) by certain
role-relationships. In addition, and this too should require no extensive
discussion at this point, whether the variation required is from one lan-
guage to another or from one geographic, social or occupational variety
to another, the functionally differential role-relationships must be
emically validated rather than merely etically enumerated. There are
certainly sociolinguistic allo-roles in most speech communities. How-
ever, two other characterizations of role-relationships do merit mention
at this point, particularly because they have proved to be useful in
sociolinguistic description and analysis.
Role-relationships vary in the extent to which their mutual rights
and obligations must or must not be continually stressed. The king-
subject role-relationship may retain more invariant stress than the
shopkeeper-customer relationship. If shopkeepers and their customers
may also interact with each other as friends, as relatives, as members
of the same political party, etc., whereas kings and their subjects (in
the same speech community) may not experience a similar degree
of role change, access and/or fluidity vis-a-vis each other, then we
would expect to encounter more variation in the 'talk' of two indi-
viduals who encounter each other as shopkeeper and customer than
we would expect between two individuals who encounter each other as
king and subject. In addition, a shopkeeper and his customer may
be able to set aside their roles entirely and interact entirely on the
basis of their individual and momentary needs and inclinations. This
may not be permissible for the king and his subjects. Thus, we would

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
244 Joshua A. Fishman
say that a shopkeeper and his customer may engage in both personal
and transactional interactions (Gumperz 1964a), whereas the king
and his subjects engage only in transactional interactions. Transactional
interactions are those which stress the mutual rights and obligations
of their participants. Personal interactions are more informal, more
fluid, more varied.
In part, speech acts and events are differentially distributed through-
out various role-relationships because personal and transactional
interactions are differentially permitted in various role-relationships.
The sociology of language is necessarily of interest to those investi-
gators who are concerned with determining the functionally different
role-relationships that exist within a given community. Micro-level
sociology of language, at least, is concerned with the validation of
such relationships, via demonstration of differential role access, role
range and role fluidity, as well as via the demonstration of differential
proportions of personal and transactional interaction, through the data
of 'talk'. Role-relationships may be used as data-organizing units both
with respect to variation in talk as well as with respect to other varia-
tions in interpersonal behavior. That is the reason why role-relations
are so frequently examined in the sociology of language.

4.4 The Situation: Congruent and Incongruent

It has probably occured to the reader that if the shopkeeper and his
customer are not to interact only as such but, rather, also as friends,
lovers, relatives, or party-members, that more than their roles are
likely to change. After all, neither the time nor the place of the store-
keeper-customer role-relationship is really ideal for any of the other
relationships mentioned. Lovers require a time and a place of their
own, and the same is true - or, at least, is typical - for other role-
relationships as well. These three ingredients (the implementation
of the rights and duties of a particular role-relationship, in the place
(locale) most appropriate or most typical for that relationship, and
at the time societally defined as appropriate for that relationship),
taken together, constitute a construct that has proven itself to be of
great value in the sociology of language: the social situation (Bock
1964; see Figure 4).
The simplest type of social situation for micro-level sociology of
language to describe and analyze is the congruent situation in which
all three ingredients 'go-together' in the culturally accepted way.
This is not to say that the investigator may assume that there is only
one place and one time appropriate for the realization of a particular
role-relationship. Quite the contrary. As with the wakes studied by Bock

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 245

SITUATION: "CLASS' Time: Class Meeting

Roles: + Teacher
Space: Classroom + Pupil
± Student-Teacher

+ indicates obligatory occurrence

± indicates optional occurrence
Figure 4. The Social Situation (Bock 1964)

on a Micmac Indian Reserve, there may be various times and various

places for the appropriate realization of particular role-relationships
(see Figure 5). Nevertheless, the total number of permissible combina-
tions is likely to be small and, small or not, there is likely to be little
ambiguity among members of the society or culture under study as to
what the situation in question is and what its requirements are with
respect to their participation in it. As a result, if there are language
usage norms with respect to situations these are likely to be most
clearly and uniformly realized in avowedly congruent situations.
However, lovers quarrel. Although they meet in the proper time
and place they do not invariably behave toward each other as lovers
should. Similarly, if a secretary and her boss are required to meet
in the office at 3:00 a.m. in order to complete an emergency report,
it may well be difficult for them to maintain the usual secretary-
boss relationship. Finally, if priest and parishioner meet at the Yonkers
Raceway during the time normally set aside for confessions this must
have some impact on the normal priest-parishioner role-relationship.
However, in all such instances of initial incongruency (wrong behavior,
wrong time, or wrong place) the resulting interaction - whether so-
ciolinguistic or otherwise - is normally far from random or chaotic.
One party to the interaction or another, if not both, reinterpret(s)
the seeming incongruency so as to yield a congruent situation, at
least phenomenologically, for that particular encounter, where one
does not exist socioculturally.
Because of incongruent behavior toward each other lovers may
reinterpret each other as employer and employee and the date
situation is reinterpreted as a dispassionate work situation. Because
of the incongruent time, secretary and boss may view the work
situation as more akin to a date than is their usual custom.
Because of the incongruent place priest and parishioner may pretend
not to recognize each other, or to treat each other as 'old pals'. In

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
246 Joshua A. Fishman
M-14 T-l T-2 T-3 T-4 T-5
S-l: Bier s-1.1:
nucleus R-l R-l R-l R-l R-l

Area s-1.2:
margin ±R-2 ±R-2
S-2: Front Area R-3 R-4 r-2.1
S-3: Audience ±R-2 r-2.2
Area R-2 R-2 ±R-4 R-4

S-4: Mar- s-4.1: r-2.1

ginal kitchen

Area s-4.2: ±r-2.2

outside r-2.2 ±R-4

Figure 5. Situation-Matrix #14: Indian Wake (Bock, 1964)

14.SC-A: Place of Wake—External distribution into 9.S-A.1: House site (usually

that occupied by deceased)
S-l: Bier Area
s-1.1: nucleus—contains coffin
s-1.2: margin—area immediately surrounding coffin
S-2: Front Area—focal region of performance during T-2, -3, and -5.
S-3: Audience Area—seating area for R-2: Mourner
S-4: Marginal Area—residual space, including
s-4.1: kitchen area
s-4.2: outside of house

14.TC-A: Time of Wake—External distribution (see discussion above).

TC-A = / / T - l / T - 2 / / : T - 3 / T - 4 : / / ± T - 5 / / : T - 3 / T - 4 : / /
T-l: Gathering Time—participants arrive at SC-A: Place of Wake
T-2: Prayer Time—saying of the Rosary by R-3: Prayer Leader
T-3: Singing Time—several hymns sung with brief pauses in between
T-4: Intermission—longer pause in singing
T-5: Meal Time—optional serving of meal (about midnight)

14.RC-A: Participant Roles—External distribution noted for each:

R-l: Corpse—from 3:RC-A: Band Member
R-2: Mourner
r-2.1: Host—member of 9.RC-A: Household Group (of deceased)
r-2.2: Other—residual category
R-3: Prayer Leader
r-3.1: Priest—from 3.R-B. 1.1: Priest
r-3.2: Other—from 14.R-4
R-4: Singer—usually from 11.R-A.4: Choir Member

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 247
short, after a bit of 'fumbling around' in which various and varying
tentative redefinitions may be tried out, a new congruent situation is
interpreted as existing and its behavioral and sociolinguistic require-
ments are implemented (Blom and Gumperz, in press; Fishman
1968b). Thus, whereas bilingual Puerto Rican parents and their
children in New York are most likely to talk to each other in Spanish
at home when conversing about family matters, they will probably
speak in English to each other in the Public School building (Fishman,
Cooper and Ma 1968). As far as they are concerned, these are two
different situations, perhaps calling for two different role-relationships
and requiring the utilization of two different languages or varieties.
Situational contrasts need not be as discontinuous as most of our
examples have thus far implied. Furthermore, within a basically
Spanish speaking situation one or another member of a bilingual
speech community may still switch to English (or, in Paraguay, to
Guarani) in the midst of a speech event for purely metaphorical (i.e.,
for emphatic or contrastive) purposes. Such metaphorical switching
would not be possible, however, if there were no general norm assign-
ing the particular situation, as one of a class of such situations, to
one language rather than to the other. However, in contrast to the
frequently unilateral and fluid back-and-forth nature of metaphorical
switching (perhaps to indicate a personal interlude in a basically
transactional interaction) there stands the frequently more reciprocal
and undirectional nature of situational switching.
More generally put, situational switching is governed by common
allocation, i.e., by widespread normative views and regulations that
commonly allocate a particular variety to a particular cluster of topics,
places, persons and purposes. Metaphorical switching, on the other
hand, is governed by uncommon or contrastive allocation. It is opera-
tive as a departure from the common allocations that are normally
operative. Without well established normative views and regulations
relative to the functional allocation of varieties within the repertoire
of a speech community neither situational nor metaphorical switching
could effectively obtain. A switch to Cockney where Received Pro-
nounciation (and grammar) is called for may elicit a brief raising of
eyebrows or a pause in the conversation - until it is clear from the
speaker's demeanor and from the fact that he has reverted to RP
that no change in situation was intended. However, such metapho-
rical switching can be risky. Someone might feel that Cockney for
the situation at hand is in poor taste. Metaphorical switching is a
luxury that can be afforded only by those that comfortably share not
only the same set of situational norms but also the same view as to
their inviolability. Since most of us are members of several speech

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
248 Joshua A. Fishman
networks, each with somewhat different sociolinguistic norms, the
chances that situational shifting and metaphorical switching will be
misunderstood and conflicted - particularly where the norms per-
taining to variety selection have few or insufficiently powerful guardians
- are obviously great.

4.5 The Transition to Macrolevel Sociology of Language

The situational analysis of language and behavior represents the
boundary area between microlevel and macrolevel sociology of lan-
guage. The very fact that a baseball conversation 'belongs' to one
speech variety and an electrical engineering lecture 'belongs' to
another speech variety is a major key to an even more generalized
description of sociolinguistic variation. The very fact that humor
during a formal lecture is realized through a metaphorical switch
to another variety must be indicative of an underlying sociolinguistic
regularity, perhaps of the view that lecture-like or formal situations
are generally associated with one language or variety whereas levity
or intimacy is tied to another (Joos 1959). The large-scale aggregative
regularities that obtain between varieties and societally recognized func-
tions are examined via the construct termed domain (Fishman 1965d;
Fishman in press).
Sociolinguistic domains are societal constructs derived from pains-
taking analysis and summarization of patently congruent situations
(See Fishman, Cooper, and Ma 1968, for many examples of the
extraction of emic domains via factor analysis as well as for
examples of the validation of initially etic domains). The macro-so-
ciologist or social psychologist may well inquire: What is the signi-
ficance of the fact that school situations and 'schoolish' situations (the
latter being initially incongruent situations reinterpreted in the di-
rection of their most salient component) are related to variety a? Fre-
quently, it is helpful to recognize a number of behaviorally separate
domains (behaviorally separate in that they are derived from dis-
continous social situations) all of which are commonly associated
with a particular variety or language. Thus, in many bilingual speech
communities such domains as school, church, professional work-
sphere and government have been verified and found to be con-
gruent with a language or variety that we will refer to as H (although
for purely labelling purposes we may refer to it as a or X or 1).
Similarly, such domains as family, neighborhood and lower work-
sphere have been validated and found to be congruent with a language
or variety that we will refer to as L (or b, or Y or 2). All in all, the
fact that a complex speech community contains various superposed

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 249
varieties - in some cases, various languages, and, in others, various
varieties of the same language - is now well documented. The exis-
tence of complementary varieties for intra-group purposes is known
as diglossia (Ferguson 1959a) and the communities in which diglossia
is encountered are referred to as diglossic. Domains are particularly
useful constructs for the macro-level (i.e., community-wide) functional
description of societally patterned variation in "talk" within large
and complex diglossic speech communities, about which more will
be said in Section 7, below.
Some members of diglossic speech communities can verbalize the re-
lationship between certain broad categories of behavior and certain
broad categories of "talk". More educated and verbally more fluent
members of speech communities can tell an investigator about such
relationships at great length and in great detail. Less educated and
verbally limited members can only grope to express a regularity
which they vaguely realize to exist. However, the fact that the for-
mulation of a regular association between language (variety) and
large scale situational behaviors may be difficult to come by is
no more indicative of a dubious relationship than is the fact that
grammatical regularities can rarely be explicitly formulated by native
speakers is to be considered as calling the abstracted rules themselves
into question.
As with all constructs (including situations, role-relationships and
speech events), domains originate in the integrative intuition of the
investigator. If the investigator notes that student-teacher interactions
in classrooms, school corridors, school auditoriums and in school lab-
oratories of elementary schools, high schools, colleges and universities
are all realized via H as long as these interactions are focused upon
educational technicality and specialization, he may begin to suspect
that these hypothetically congruent situations all belong to a single
(educational) domain. If he further finds that hypothetically incon-
gruent situations involving an educational and a non-educational in-
gredient are, by and large, predictably resolved in terms of H rather
than L if the third ingredient is an educational time, place or role-
relationship, he may feel further justified in positing an educational
domain. Finally, if informants tell him that the predicted language or
variety would be appropriate in all of the examples he can think of that
derive from his notion of the educational domain, whereas they
proclaim that it would not be appropriate for examples that he draws
from a contrasted domain, then the construct is as usefully validated
as is that of situation or event - with one major difference.
Whereas particular speech acts (and speech excerpts of an even
briefer nature) can be apportioned to the speech and social situations

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
250 Joshua A. Fishman
in which they transpire, the same cannot be done with respect to such
acts or excerpts in relationship to societal domains. Domains are
extrapolated from the data of 'talk', rather than being an actual com-
ponent of the process of talk. However, domains are as real as the
very social institutions of a speech community, and indeed they show
a marked paralleling with such major social institutions (Barker 1947).
There is an undeniable difference between the social institution, 'the
family', and any particular family, but there is no doubt that the
societal norms concerning the former must be derived from data on
many instances of the latter. Once such societal norms are formulated
they can be utilized to test predictions concerning the distributions
of societally patterned variations in talk across all instances of one
domain vs. all instances of another.
Thus, domains and social situations reveal the links that exist be-
tween microlevel and macrolevel sociology of language. The members
of diglossic speech communities can come to have certain views
concerning their varieties or languages because these varieties are asso-
ciated (in behavior and in attitude) with particular domains. The H
variety (or language) is considered to reflect certain values and relation-
ships within the speech community, whereas the L variety is considered
to reflect others. Certain individuals and groups may come to advocate
the expansion of the functions of L into additional domains. Others may
advocate the displacement of L entirely and the use of H solely.
Neither of these revisionist views could be held or advocated without
recognition of the reality of domains of language-and-behavior in the
existing norms of communicative appropriateness. The high culture
values with which certain varieties are associated and the intimacy and
folksiness values with which others are congruent are both derivable
from domain-appropriate norms governing characteristic verbal inter-
action.

4.6 On the Reality of Sociolinguistic Compositing

So little (if, indeed, any) microsociolinguistic data has been subjected
to rigorous quantitative analysis or obtained via experimentally con-
trolled variation that it is fitting that we pause to examine a study
that has attempted to do so, even if it deals only with sociolinguistic
normative views and claims. The study in question (Fishman and
Greenfield, 1970) is concerned with the relative importance of
persons, places and topics in the perception of congruent and incon-
gruent situations and with the impact of perceived congruency or
incongruency on claimed language use in different domains. Since
domains are a higher order generalization from congruent situations

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
 The Sociology of Language 251
(i.e. from situations in which individuals interact in appropriate role-
relationships with each other, in the appropriate locales for these
role-relationships, and discuss topics appropriate to their role-relation-
ships) it was first necessary to test intuitive and rather clinical estimates
of the widespread congruences that were felt to obtain. After more than
a year of participant observation and other data-gathering experiences
it seemed to Greenfield (1968) that five domains could be generalized
from the innumerable situations that he had encountered. He tentatively
labeled these "family", "friendship", "religion", "education" and "em-
ployment" and proceeded to determine whether a typical situation
could be presented for each domain as a means of collecting self-report
data on language choice. As indicated below each domain was repre-
sented by a congruent person (interlocutor), place and topic in the self-
report instrument that Greenfield constructed for high school students.

Domain Interlocutor Place Topic

Family Parent Home How to be a good son or daughter
Friendship Friend Beach How to play a certain game
Religion Priest Church How to be a good Christian
Education Teacher School How to solve an algebra problem
Employment Employer Workplace How to do your job more efficiently

Greenfield's hypothesis was that within the Puerto Rican speech

community, among individuals who knew Spanish and English equally
well, Spanish was primarily associated with family and with
friendship (the two, family and friendship constituting the intimacy
value cluster), while English was primarily associated with religion,
work and education (the three constituting the status-stressing value
cluster). In order to test this hypothesis he first presented two seeming-
ly congruent situational components and requested his subjects (a)
to select a third component in order to complete the situation, as
well as (b) to indicate their likelihood of using Spanish or English if
they were involved in such a situation and if they and their Puerto
Rican interlocutors knew Spanish and English equally well. Section I
of Table 1 shows that Greenfield's predictions were uniformly con-
firmed among those subjects who selected congruent third components.
Spanish was decreasingly reported for family, friendship, religion,
employment and education, regardless of whether the third component
selected was a person, place or topic.
However, as Blom and Gumperz (in press), Fishman (1968b) and
others have indicated, seemingly incongruent situations frequently
occur and are rendered understandable and acceptable (just as are
the seemingly ungrammatical sentences that we hear in most spon-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
252 Joshua A. Fishman
TABLE 1

Spanish and English Usage Self-Ratings in Various Situations

for Components Selected
I. Congruent Situations: Two 'congruent* components presented; S selects third
congruent component and language appropriate to situation. 1 =all Spanish,
5=all English.
Congruent Persons Selected
Parent Friend Total Priest Teacher Employer Total
Mean 2.77 3.60 3.27 4.69 4.92 4.79 4.81
S.D. 1.48 1.20 1.12 .61 .27 .41 .34
N 13 15 15 13 13 14 15

Congruent Places Selected

Work
Home Beach Total Church School Place Total
Mean 2.33 3.50 2.60 3.80 4.79 4.27 4.27
S.D. 1.07 1.26 1.10 1.51 .58 1.34 .94
N 15 6 15 15 14 15 15

Congruent Topics Selected

Friend- Employ-
Family ship Total Religion Education ment Total
Mean 1.69 3.30 2.64 3.80 4.78 4.44 4.38
S.D. .92 1.20 .95 1.47 1.53 1.12 .73
N 16 18 18 15 18 18 18

II. Incongruent Situations: Two 'incongruent' components presented; S selects

third component and language appropriate to situation. l = a l l Spanish, 5=all
English.
Persons Selected
Parent Friend Total Priest Teacher Employer Total
Mean 2.90 3.92 3.60 4.68 4.77 4.44 4.70
S.D. 1.20 .64 .70 .59 .48 .68 .52
N 16 16 16 14 15 9 15

Places Selected
Work
Home Beach Total Church School Place Total
Mean 2.63 3.86 2.77 3.71 4.39 4.42 4.10
S.D. .77 .94 .70 1.32 1.90 .96 .82
N 15 5 15 15 15 15 15

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 253
Topics Selected
Friend- Employ-
Family ship Total Religion Education ment Total
Mean 2.83 3.81 3.26 3.07 3.66 3.81 3.49
S.D. 1.04 1.13 1.02 1.00 1.20 .85 .76
N 18 16 18 18 17 18 18

taneous speech). Interlocutors reinterpret incongruences in order

to salvage some semblance of the congruency in terms of which they
understand and function within their social order. Were this not the
case then no seemingly congruent domains could arise and be main-
tained out of the incongruencies of daily life. In order to test this
assumption Greenfield proceeded to present his subjects with two
incongruent components (e.g., with a person from one hypothetical
domain and with a place from another hypothetical domain) and asked
them to select a third component in order to complete the situation
as well as to indicate their likelihood of using Spanish or English in
a situation so constituted. Greenfield found that the third component
was overwhelmingly selected from either one or the other of any two
domains from which he had selected the first two components. Further-
more, in their attempts to render a seemingly incongruous situation
somewhat more congruent his subject's language preferences left the
relationship between domains and language choice substantially
unaltered (directionally), regardless of whether persons, places or
topics were involved. Nevertheless, all domains became somewhat less
different from each other than they had been in the fully congruent
situations. Apparently, both individual indecisiveness as well as socio-
linguistic norms governing domain regularity must be combined and
compromised when incongruencies appear. Language choice is much
more clear-cut and polarized in 'usual' situations governed neatly by
sociolinguistic norms of communicative appropriateness than they are
in 'unusual' situations which must be resolved by individual interpreta-
tion.
Yet, another (and, for this presentation, final) indication of the
construct validity of domains as analytic parameters for the study
of large scale sociolinguistic patterns is yielded by Edelman's data
(1968). Here we note that when the word naming responses of bilingual
Puerto-Rican children in Jersey City were analyzed in accord with
the domains derived from Greenfield's and Fishman's data re-
ported above significant and instructive findings were obtained. The
most Spanish domain for all children was 'family' (Table 2A). The
most English domain for all children was 'education'. The analysis
of variance (Table 2B) indicates that not only did the children's res-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
254 Joshua A. Fishman
TABLE 2A
Mean number of words named by young schoolchildren (Edelman, 1968)
(N= 34)

Age Language Domain

Family Education Religion Friendship Total

6-8 English 6.2 8.2 6.6 8.3 7.3

Spanish 7.6 6.2 5.8 6.4 6.5
Total 6.9 7.2 6.2 7.4 6.9
9-11 English 11.7 12.8 5.7 10.9 11.0
Spanish 10.5 9.4 7.2 9.7 9.2
Total 11.1 11.1 7.9 10.3 10.1
Total English 9.0 10.5 7.7 9.6 9.2
Spanish 9.0 7.8 6.5 8.0 7.8
Total 9.0 9.1 7.1 9.0 8.5

TABLE 2 B
Analysis of variance of young schoolchildren's word-naming scores

Sum of Mean
Source Squares df Square F F95 F99
Between Subjects 1844.12 33
C (age) 689.30 1 689.30 19.67a 4.17 7.56
D (sex) 15.54 1 15.54 .44 4.17 7.56
CD 87.87 1 87.87 2.51 4.17 7.56
error (b) 1051.41 30 35.05
Within Subjects 1795.88 238
A (language) 123.13 1 123.13 9.73a 4.17 7.56
B (domain) 192.54 3 64.18 8.51» 2.71 4.00
AB 65.12 3 21.71 11.67a 2.71 4.00
AC 16.50 1 16.50 1.30 4.17 7.56
AD 42.08 1 42.08 3.32 4.17 7.56
BC 61.54 3 20.51 2.72 2.71 4.00
BD 2.89 3 .96 .13 2.71 4.00
ABC 23.99 3 8.00 4.30a 2.71 4.00
ABD 6.70 3 2.23 1.20 2.71 4.00
ACD 14.62 1 14.62 1.15 4.17 7.56
BCD 13.53 3 4.51 .60 2.71 4.00
ABCD 7.98 3 2.66 1.43 2.71 4.00
error (w) 1225.26 210
errori (w) 379.88 30 12.66
error2 (w) 678.31 90 7.54
errors (w) 167.07 90 1.86
Total 3640.00 271

a Significant at or above the .01 level.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 255
ponses differ significantly by age (older children giving more responses
in both languages than did younger children), by language (English
yielding more responses than does Spanish), and by domain (church
yielding fewer responses than does any other domain), but that these
three variables interact significantly as well. This means that one
language is much more associated with certain domains than is the
other and that this is differentially so by age. This is exactly the kind
of finding for which domain analysis is particularly suited. Its utility
for inter-society comparisons and for gauging language shift would
seem to be quite promising, but its major value should be in describ-
ing and demonstrating the dependence of communicative appropriate-
ness on the composing appropriateness of members of speech com-
munities, whether monolingual or bilingual.
One thing appears to be clear from the theoretical and empirical
work cited: there are classes of events recognized by each speech
network or community such that several seemingly different situations
are classed as being of the same kind. N o speech network has a linguistic
repertoire that is as differentiated as the complete list of apparently dif-
ferent role relations, topics and locales in which its members are
involved. Just where the boundaries come that do differentiate be-
tween the class of situations generally requiring one variety and
another class of situations generally requiring another variety must
be empirically determined by the investigator, and constitutes one
of the major tasks of descriptive sociology of language. Such classes of
situations are referred to as domains. The various domains and the
appropriate usage in each domain must be discovered from the data
of numerous discrete situations and the shifting or non-shifting which
they reveal. This is a central task of descriptive sociology of language,
and it can only be accomplished by painstaking research — utilizing all
of the available social science methods: participant observation,
interviews, surveys and experiments too. The compositing concerns
of some researchers in the sociology of language are thus far
from being research strategies alone. Ultimately they also seek to
reveal the behavioral parsimony of members of speech communities
all of whom inevitably come to rely on a relatively functional socio-
linguistic typology to guide them through the infinite encounters of
daily interaction.

4.7 Sociology of Language: Multilevel and Multimethod

The list of constructs utilized in the sociolinguistic description and
analysis of samples of 'talk' is far from exhausted. We have not men-
tioned several of the social units long advocated by Hymes (1962), such

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
256 Joshua A. Fishman
as participant vs. audience roles, the purposes and the outcomes of
speech events, the tone or manner of communication, the channel
of communication employed (oral, written, telegraphic), nor all of
the various parameters and components for the analysis of talk data
that he has more recently advanced (Hymes 1967b; see figure 6A);

(S) SETTING or SCENE: time and place; also, psychological setting and cul-
tural definition as a type of scene.
(P) PARTICIPANTS or PERSONNEL: e.g., addressor-addressee-audience
(E) ENDS: ends in view (goals, purposes) and ends as outcomes
(A) ART CHARACTERISTICS: the form and the content of what is said
(K) KEY: the tone, manner or spirit in which an act is done
(I) INSTRUMENTALITIES: channel (the choice of oral, written, telegraphic
or other medium) and code (Spanish, English, etc.) or subcode (dialect,
sociolect)
(N) NORMS OF INTERACTION and of INTERPRETATION: specific behav-
iors and properties that may accompany acts of speech, as well as shared
rules for understanding what occurs in speech acts
(G) GENRES: categories or types of speech acts and speech events: e.g., con-
versation, curse, prayer, lecture, etc.

Figure 6a. Components of Speech Events:

A heuristic schema
(Hymes 1967b)

we have not discussed such social psychological parameters as the

saliency of individual vs. collective needs (Herman 1961), nor the
several functions of speech so revealingly discussed by Ervin-Tripp.
Suffice it to say that there are several levels and approaches to socio-
linguistic description and a host of linguistic, socio-psychological and
societal constructs within each (see Figure 6b). One's choice from
among them depends on the particular problem at hand (Ervin-Tripp
1964). This is necessarily so. The sociology of language is of interest
to students of small societies as well as to students of national and
international integration. It must help clarify the change from one face-
to-face situation to another. It must also help clarify the different
language-related beliefs and behaviors of -entire social sectors and
classes. In some cases the variation between closely related varieties
must be highlighted. In other cases the variation between obviously
unrelated languages is of concern.
It would be foolhardy to claim that one and the same method of data
collection and data analysis be utilized for such a variety of problems
and purposes. It is one of the hallmarks of scientific social inquiry
that methods are selected as a result of problem specifications rather

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 257

Figure 6b.
Relationships Among Some Constructs Employed in Sociolinguistic Analysis1
1
From: Robert L. Cooper, "How can we measure the roles which a bilingual's
languages play in his everyday behavior?" in L. G. Kelly (ed.), The Description
and Measurement of Bilingualism (Toronto, Toronto University Press, 1969), p.
202.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
 258 Joshua A. Fishman
than independently of them. The sociology of language is neither
methodologically nor theoretically uniform. Nevertheless, it is gratifying
to note that for those who seek such ties the links between micro- and
macro-constructs and methods exist (as do a number of constructs
and methods that have wide applicability through the entire range of
the sociology of language). Just as there is no societally unencumbered
verbal interaction so are there no large scale relationships between
language and society that do not depend on individual interaction
for their realization. Although there is no mechanical part-whole re-
lationship between them microlevel and macrolevel sociology of lan-
guage are both conceptually and methodologically complementary.

5.0 SOCIETAL DIFFERENTIATION AND REPERTOIRE RANGE

Speech communities - particularly those at the citywide, regional or

national levels - obviously vary in the degrees and kinds of language
diversity that they reveal. What do such differences imply with res-
pect to the social differentiation and organization of the communities
and networks to which they apply? If we examine the varieties of
Javanese required by linguistic etiquette in the communities described
by Geertz (1960), the varieties of Baghdadi Arabic described by Blanc
(1964), the varieties of Hindi or Kannada described by Gumperz
(1958) or McCormack (1960), and the varieties of Indonesian des-
cribed by Tanner (1967) it is clear that these compose quite different
kinds of repertoires than do the varieties of Norwegian described by
Haugen (1961), or the varieties of American English described by
Labov (1963, 1964, 1965), or by Levine and Crockett (1966).
In addition, the types of speech communities in which these varieties
are encountered also differ strikingly, as do the larger national or
regional units in which the communities are imbedded. To put it
very briefly, the speech communities in the first cluster seem to be
much more stratified socially and to employ much more diversified
repertoires linguistically than do those in the second. The documented
co-occurrence of linguistic heterogeneity and societal heterogeneity
- when both are examined in intra-group perspective - is a major
contribution of the sociology of language to the study of social
organization and social change.

5.1 The Significance of Pervasive Linguistic Discontinuity

Prior to the development of the sociology of language per se, area-
dialectology had already clearly indicated that discontinuous popu-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 259
lations (i.e., populations that lived at some distance from each other
or that were impeded in their communication with each other by
physical or political barriers) frequently revealed substantial phono-
logical and morphological differences between their language systems
(see, e.g., Herzog 1965 and Kandori 1968 for examples of such work
today). Where such differences did not obtain despite the absence of
communicational frequency and socio-cultural unity, recency of settle-
ment from a single source or other similar unifying factors (conquest,
religious conversion, etc.) were assumed and encountered. Indeed, if we
view the entire world as a single geographic area we tend to find
similar (i.e., genetically related) languages clustered contiguously or
closely to each other ('language families' are normally clustered geo-
graphically, except for the confounding fact of colonization and distant
migration). Some parts of the world, of course, are famous for their
concentration of highly diversified languages found in close proximity
to each other. However, these same areas are also noted for their
mountains, jungles, deserts and rivers, i.e., for barriers that have limited
travel, commerce and common endeavor.
More difficult to explain are those variations in language and
behavior that are co-territorial. In such instances sheer physical distance
cannot be invoked as either a causal or a maintenance variable for
the variations encountered. In such cases cultural and social factors
alone must be examined and they alone must be meaningfully related
to the degree and kind of language differences noted. In reviewing co-
territorial linguistic diversity throughout history it becomes clear that
it can be maintained in an extremely stable manner. Throughout the
world - but particularly throughout the ancient and traditional world
- populations have lived side by side for centuries without learning each
other's language(s) and without significantly modifying or giving up
their distinctly discontinous repertoires. Except for the relatively few
middlemen that connect them (merchants, translators, etc.) such pop-
ulations represent distinct speech communities although they may be
citizens of the same country, of the same city, and, indeed, of the same
neighborhood. However, the maintenance of such well-nigh complete
linguistic and socio-cultural cleavage - equal in degree and kind to that
encountered between territorially discontinuous populations - is usually
indicative of population relocation some time in the past that has subse-
quently been buttressed and maintained by socio-cultural (including
ethnic and religious) differences. The former differences are responsible
for the origin of the differences noted by Blanc (1964) between the
Moslem Arabic, Christian Arabic and Jewish Arabic of Bagdad. The
latter differences are responsible for the maintenance of these cleavages
in as sharp a manner, or nearly so, as initially established.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
260 Joshua A. Fishman
While it may often be relatively difficult to overcome the cleavage
between separate but co-territorial speech communities it is not im-
possible to do so. The forced conversion of various Jewish and Christian
communities during certain periods of Islamic rule, the urban-
industrial assimilation of hitherto rural or small town immigrants
and their children in the United States (Nahirny and Fishman 1965,
Fishman 1965a, 1965e, 1966c), the very similar assimilation of tribal
populations moving to Wolof-speaking Dakar (Tabouret-Keller 1968),
the Hellenization and Romanization of many 'barbarian' elites in
ancient Rome and Alexandria, the convergence between illiterate
speakers of Marathi and Kannada in India (Gumperz 1967) - these
are all examples of the fusing into one of populations that originally
functioned as largely separate though co-territorial speech communi-
ties. Conversely, the mutual alienation of populations that originally
considered themselves to be united can create fargoing linguistic
differences between them where none, or few, existed previously.
In general, the more far-going the linguistics differences between
any two co-territorial populations (i.e., the more the differences are
basically grammatical-syntactic and morphological - rather than pri-
marily phonological or lexical), the more linguistic repertoires are
compartmentalized from each other so as to reveal little if any inter-
ference, and the more they reveal functionally different verbal reper-
toires in terms of the sociolinguistic parameters reviewed in Section 4,
above - then the greater the interactional and socio-cultural gap
between the speech communities involved.
Geertz's data (see Figures 7a, b and c) might well be examined in
the light of the above generalization concerning the social significance
of marked grammatical discontinuity between the repertoires of co-
territorial speech communities. In Geertz's case we are dealing with
co-territorial speech networks that differ greatly in verbal repertoires
but that cannot be considered to be either of separate geographic origin
or of separate cultural or religious self-definition. Here we find three
different social classes or strata within Java each differing in repertoire
range and each lacking entirely one or more speech varieties available
to at least one of the others. While the intra-network variation shown
by Geertz is probably less that which actually exists (thus, we may
assume that metaphorical switching also occurs in Java and, if it does,
level 2 (for example) may be employed on occasions which are nor-
matively viewed and regulated as being more appropriate for level lb
or la) let us consider this to be merely an artifact of the data model that
Geertz employs and ask ourselves (a) what kind(s) of variations does
it reveal and (b) what kind(s) of repertoire differences does it reveal.
Geertz's data clearly indicate that social class differences exist

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 261
(or existed at the time his field work was done) in Javanese verbal
behavior. In addition, however, the data also indicate that contextual-
situational variation also exists in Javanese verbal behavior. The very
fact that both of these types of variation regularly co-occur is an in-
dication that although stratificational differences involved are rigid
and deep, nevertheless the strata constitute a single integrated speech
community with shared normative expectations and regulations vis-a-vis
intra-strata and inter-strata communication.
The fact that networks in each stratum lack at least one variety
available to networks drawn from the other strata is a sign of far-
going discontinuity also in their respective behavioral repertoires.
Networks from certain strata are not expected to engage in certain
role relationships and as a result, lack entirely certain morpho-syn-
tactic co-occurrences available to networks from other strata. Thus,
in these latter respects, the variation that occurs is stratificational
only and not contextual at all. This stratificational discontinuity in
morpho-syntactic co-occurrences is shown graphically in Figure 8
for the forms apa, napa and menapa. The strata that do possess these
forms use them for identical contexts of interaction and with apparently
equal frequency of realization. However, there is in each case also a
stratum that lacks these forms. The graphic representation of social
and verbal discontinuity should be kept in mind for comparison with
other graphs presented further below (e.g. Figures 9a and 9b).

5.2 More Marginal but Systematic Linguistic Differences between

Social Strata
However, most co-territorial populations that differ in verbal reper-
toire cannot be considered fully separate speech communities, even
if the differences between them can be considered as basically
geographic in origin. There are very many areas today, primarily
urban in nature, where sub-populations that differ in social class,
religion or ethnic affiliation, nevertheless view themselves as sharing
many common norms and standards and where these sub-populations
interact sufficiently (or are sufficiently exposed to common educational
institutions and media) to be termed a single speech community.
It is hardly surprising, therefore, that the linguistic differences be-
tween such socio-cultural sub-populations (or networks) within the
same speech community are more linguistically marginal (i.e., lexical
and, to a lesser degree, phonological) rather than syntactic and all-
embracing. It is clear that the social class variation that exists in
New York City English is of this kind rather than of the kind that
develops between clearly separate, non-interacting and mutually

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
 262

Level are you going to eat rice and cassava now Complete sentence

c
8
a
•e
.3

a
h.
pandjenengan

a-a

CO
a—
s

?
I
a s

a 8

§ S
s -4Í
•xs-SS

S Í3
<U Q.

•O
menapa kalijan samenika
Menapa sampéjan badé neda sekul

ro
sekul kalijan kaspé samenika?

i
Napa sampéjan adjeng neda sekul

e
napa sampéjan adjeng kaspé saniki lan kaspé saniki?

Apa sampéjan arep neda sega lan

5
c

<a
kaspé saiki?

1
arep sega saiki
Apa kowé arep mangan sega lan
kowé mangan kaspé saiki?

z
e
a
u
w

o
cd

I
•a

¡5
CU

tá
O
(-I

T3
«2

c9
<ù
%
o

pu, I

J
©

8z

£ §
su
oCO a
a o
0\

r- Q
a cl<

I0 H
•O 5

&

o
«

3 s?
fi

a co
£ co
co

CO
i
s

x> Ü3
WJ OT

Authenticated
Download Date | 2/16/20 10:45 AM
Brought to you by | The University of Texas at Austin
Joshua A. Fishman
 Level are you 1 going to eat rice and cassava now Complete sentence

o
napa Napa sampéjan adjeng neda sekul
adjeng sekul saniki lati kaspé saniki?

e
•a
sampéjan
Apa sampéjan arep netfa sega lan

.2
lan kaspé kaspé saiki?
apa arep sega saiki
Apa kowé arep mangan sega lan
kowé mangan

»-H
kaspé saiki?

JÓ
a

Q
3
o

<

•o
at

B-
(2

cd
•O
cu
The Sociology of Language

"o3.
Level are you going to eat rice and cassava now Complete sentence
a
a
•c

-Q

ro
at
pandjenengan

•a
•s:
<3
S-s

a a
cc ^
§, 8

ca ~3
§ 8
1
§
&
C <

§15.• S>
"3 C
U

a
menapa

•Ci
sekul kalijan kaspé samenika
Menapa sampéjan ba$é neda sekul

e
sampéjan

•a
kalijan kaspé samenika?

Apa pandjenengan arep dahar sega

<
•e
•a
pandjenengan

<3
lan kaspé saiki?

Apa sampéjan arep netjia sega lan

a
e

«
a

apa sampéjan arep sega .2

w
saiki kaspé saiki?

kowé Apa kowé arep mangan sega lan

mangan

f—t
kaspé saiki?

Authenticated
Download Date | 2/16/20 10:45 AM
Brought to you by | The University of Texas at Austin
u
0.

¡5
•c

"3
263
264 Joshua A. Fishman

Apa Napa Menapa

frequency
of realization

formality of interaction

=Prijajis =Non-Prijaji, =Peasants

urbanized and and unedu-
somewhat edu- cated towns-
cated people
Figure 8. Verbal and Behavioral Discontinuity

alienated speech communities. One of the surest indications of this

is the fact that (if we delete features attributable to Southern Negro,
Puerto Rican and other recent, geographically derived differences)
few of the characteristic phonological features of lower class speech
in New York are entirely absent from the speech of other classes in
New York City, just as few of the characteristic phonological features
of its upper-class speech are entirely lacking from the lower class
speech of that city. What does differentiate between the social classes
in New York is the degree to which certain phonological variables
are realized in certain ways on particular occaisions, rather than their
complete absence from the repertoire of any particular class.
Labov's studies of the phonological correlates of social stratification
(1964, 1965, 1966a, 1966b, 1966c, 1968a, 1968b) illustrate this point.
In one of his studies (1964) Labov gathered four different samples of
speech (each by a different method calculated to elicit material approxi-
mating a different kind of speech situation) from four different social
classes of informants. Studying such variables as th (as in thing,
through), eh (the height of the vowel in bad, ask, half, dance), r (the
presence or absence of final and preconsonantal / r / ) and oh (the
height of the vowel in off, chocolate, all, coffee) Labov found that all
social classes yielded some values of each variable in nearly every
speech situation (see figure 9). However, the differences between the
social classes remained clear enough. Lower class speakers were less
likely to pronounce the fricative form of the [®] when saying 'thing' or

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
 100 Socioeconomic class groups
SEC 0-1 lower class
2-4 working class
y g j lower middle class
9 upper middle class
80

s \ SEC
ai
8co \0-l

X 60
U
D
g
\2-4 N.

I
u
40

u>
6 \ N \
20

A B C D
casual careful reading word
speech speech style lists
CONTEXTUAL STYLE
Figure 9a. Class stratification diagram for (th)
(Labov, 1964)

C O N T E X T U A L STYLE
Fig. 9b. Class stratification diagram for (r)
(Labov, 1964)

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
266 Joshua A. Fishman
'through' than were working class speakers; working class speakers less
likely to pronounce it than lower middle class speakers; lower middle
class speakers less likely to yield it than upper middle class speakers.
Speakers of all classes were more likely to pronounce the standard
fricative form (rather than the sub-standard affricate [t<9] or lenis stop
[t]) in reading word lists than they were when reading passages; more
likely to pronounce it when reading passages than when being inter-
viewed ( = careful speech); more likely to pronounce it when being
interviewed than when recounting "a situation where you thought you
were in serious danger of being killed" ( = casual speech).
This may be considered a hallmark of social class differences in
speech where the classes as a whole share continuous experiences,
goals and expectations, i.e., neither their role repertoires nor role
access have been fully compartmentalized. As long as individuals in
each class can differ in repertoire, depending on their personal
opportunities and experiences with respect to interaction with various
speech networks, there can be no complete discontinuity in repertoires,
no complete freezing of social class position, and no overriding alien-
ation into separate religious, ethnic or other relatively fixed and
immutable speech communities.
Of course, not all variables yield such dramatic and clearcut
social class differences as those found in connection with th in New
York. With respect to r, eh, and oh Labov's data reveals much more
similarity between the several social classes, although the differences
between contexts and between classes remain quite clear. Labov's data
also reveal a recurring reversal with respect to the lower middle
class' performance on wordlists and passage reading. This reversal,
dubbed hypercorrection, shows the lower middle class to be more
'correct' (more careful, more inclined to use the standard or cultured
pronounciation) than is the upper middle class at its most correct
or careful. Such a reversal may well indicate a variable that has be-
come a stereotype rather than merely a marker of class position. As
such it tends to be used (or overused) by those who are insecure
about their social position, i.e., by those who are striving to create a
more advantageous social position for themselves in a speech commu-
nity in which upward social mobility seems to be possible. This
explanation is not dissimilar from that which Labov utilized to explain
observed differences in centralization of / a i / and / a u / in Martha's
Vineyard (1963). Such centralization was most common among minor-
ity group members (of Portuguese and Indian extraction) who
sought to stress their positive orientation to Martha's Vineyard, rather
than among the old Yankees whose feelings toward the Vineyard
were more low-keyed and required no linguistic underscoring. Whether

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 267
consciously employed or not the 'Pygmalion effect' in language is a
striking indicator of reference group behavior and of social aspirations
more generally (Ross 1956).
Similar results to Labov's (in the sense that the proportional reali-
zations of particular variables were found to differ regularly and
smoothly both between social classes and between contexts) have been
reported by Lindenfeld 1969. Examining syntactic variation in French,
Lindenfeld found that nominalization, relativization and sentence length

_L
Formal Nominalization Informal Formal Relativization Informal

25

20

15
Class I
- X
Class II I

10

J_
Formal Subordination Informal

Formal Sentence Length Informal

Figure 10. The Social Conditioning of Syntactic Variation in French
(Lindenfeld 1969)

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
268 Joshua A. Fishman
(but not subordination) showed both types of variation, although upper
middle class speakers were much more likely to reveal contextual
variation than were lower class speakers (Figure 10). This may be taken
as a sign that the socio-economically more favored subjects had more
of a real repertoire range behaviorally so that the difference between
formal and informal interactions was very real for them. For lower
class speakers, on the other hand, this difference may be quite hypothe-
tical in that it tends to have much less functional reality associated
with it.
The demographic differentials observed in usage are as related
to the societal allocation of codes as are the more directly contextual or
functional differentials. The fact that an extensive cluster of phonologi-
cal, lexical and grammatical realizations is more widely or charac-
teristically employed by one particular social class than by another is
commonly related to the fact that the social class in question is also
more likely or characteristically engaged in particular pursuits or in-
volved in particular situations. Demographic and contextual variations
are particularly likely to be redundant in relatively closed societies
in which role access is restricted and in which roles tend to be ascribed
and compartmentalized However, the awareness of verbal and beha-
vioral repertoires - a central awareness indeed in the sociology of
language - should not keep us from realizing that even in relatively
open societies there is often appreciable redundancy between demo-
graphic and contextual differentials in usage. Both Labov's and Linden-
field's data referred to previously reveal this redundancy when they
show that for most levels of formality one social class is much more
likely to yield a particular variant than are the others, even though
repertoire continuity exists. This redundancy strengthens the normative
sense of members of speech communities and, indeed, enables them
to guide their own speech behavior more appropriately, as well as to
comment upon it validly to one another and to outside investigators
and to do so over and above the metaphorical variation that undoub-
tedly obtains round about them.

5.3 The Implications of either Contextual-Situational or Demographic

Variation
The foregoing comparisons of social class differences and contextual-
situational differences in language usage suffer in at least two ways.
Neither Geertz nor Labov nor Lindenfeld have been able to indicate
which of these two sources of language variation is the stronger for
their data. In order to answer this question a more quantitative approach
is needed to the study of social class or other demographic-group

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 269
variation in usage. In addition, neither Geertz nor Labov norLindenfeld
have asked the question "what could it mean - insofar as the overall
societal organization of language behavior — if only one or another of
these two sources of usage variation obtained?" In order to answer these
two questions let us take another look at data obtained in the study
of Bilingualism in the Barrio (Fishman, Cooper, Ma, et. al. 1968).
The data we will review was obtained as part of an interdisciplinary
project on the measurement and description of widespread and re-
latively stable bilingualism in a Puerto Rican neighborhood in the
Greater New York City area. The neighborhood studied by a team
of linguists, psychologists, and sociologists included 431 Puerto Ricans
(or individuals of Puerto Rican parentage) living in ninety households.
All these individuals were covered in a language census that obtained the
demographic data utilized for the purposes of this report at the
same time that it obtained detailed self-reports on bilingual usage and
ability (Fishman 1969d). The linguistic data utilized for this report was
obtained in the course of two to four hour interviews and testing ses-
sions with a random-stratified sample of those Puerto Ricans living in
the study neighborhood who were over the age of 12.

Speech Contexts. The interviews and testing sessions were designed

to elicit speech data in five different contexts that form a continuum
from most formal or careful to most informal or casual as follows:
Contexts D: Word Reading. Subjects were asked to read two dif-
ferent lists of separate words, one in English and one in Spanish.
The speech data obtained in this fashion was considered to be repre-
sentative of the most careful pronunciation available to the subjects.
Context C: Paragraph Reading. Subjects were asked to read four
different paragraphs, two in English and two in Spanish. The speech
data obtained in this fashion was considered to be representative of
(somewhat less) careful pronunciation.
Context WN: Word Naming. Subjects were asked to "name as many
words as come to mind that have to do with (domain-locales)." This
task was performed separately in English and in Spanish for each of the
following domain-locales: home, neighborhood, school, work, church.
The speech data obtained in this fashion was considered to be represen-
tative of intermediate pronunciation (neither markedly careful nor
casual).
Context B: Careful Conversation. Subjects were asked factual
questions concerning five taped 'playlets' to which they had just
listened. Ideally, half of the questions were asked (and answered) in
Spanish and half were asked and answered in English. The speech
data obtained in this fashion was considered to be representative of

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
270 Joshua A. Fishman
somewhat (but not completely) casual pronunciation.
Context A: Casual Conversation. Subjects were asked their
personal opinions and preferences with respect to the problems that
figured in the 'playlets' to which they had just listened. The speech
data obtained in this fashion was considered to be representative of
the most informal pronunciation that could be elicited by an inter-
viewer.
Only the last three contexts (WN, B, A) will be examined in the
discussion that follows in view of the restricted corpuses obtained in the
two reading contexts in the study population.

Linguistic Variables. The taped speech samples obtained for the

above mentioned five contexts were independently scored by two
linguists on seven Spanish and ten English variables. The reliability
of scoring varied only slightly and irregularly from context to context
and from one language to the other, the reliability coefficients ob-
tained ranged from 0.73 to 0.94 with a median of 0.90. A full report
on the contextual variation encountered for each variable as well as
on the factorial relationship between all variables is available elsewhere
(Ma and Herasimchuk 1968). The present discussion deals only with
selected values on one Spanish and one English variable in order to
illustrate a method of analysis hitherto not utilized in sociolinguistic
research. The particular linguistic values selected for presentation in
this study are further explained in the Results section below.

Demographic Variables. Four demographic factors (sex, age, edu-

cation, and birthplace) are included in the analyses presented in this
report. Social class, a variable frequently utilized in other sociolinguistic
research on phonological variables, was not utilized in the present
research due to the severe restriction in range that our overwhelmingly
lower-class Puerto Rican subjects revealed in this connection. An
extensive analysis of the demographic variation encountered in our
study neighborhood is available elsewhere (Fishman 1968c). The
reliability coefficients for the various items of obtained demographic
information are all 0.90 or higher.
Sex has consistently proved to be a non-significant demographic
variable in accounting for phonological variation in Puerto Rican
Spanish. It was included in the present study merely in order to
provide a comparison with prior studies.
Age was categorized in two separate ways. As a three-category
variable the categories employed were < 2 5 , 25-34, > 3 4 . As a two-
category variable the categories utilized were < 2 5 and > 2 5 . By categor-
izing age in two different ways we will be able to tell whether one cate-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 271
gorization is more related to linguistic variation than the other and,
at the same time, sum both age categorizations into one age variable.
Education was categorized three different ways. As a four-cate-
gory variable the categories were < 7 years, all in Puerto Rico; 7 or
more years, all in Puerto Rico; partially in Puerto Rico and partially
in continental U.S.A.; all in continental U.S.A. As a two-category
variable education was categorized in two different ways: first, all
in Puerto Rico vs. all or part in continental U.S.A. and, second, all
U.S.A. vs. all or part in Puerto Rico. Once again our analytic technique
enabled us to sum these three different ways of categorizing education
as well as to tell whether there is any difference between them in ex-
plaining linguistic variation.
Birthplace was categorized in two different ways. As a four-
category variable the categories used were highland Puerto Rico,
coastal Puerto Rico other than San Juan and suburbs, San Juan and
suburbs, and continental U.S.A. As a two-category variable the
categories utilized were highland Puerto Rico vs. all other birthplaces.
As in the other two instances of multiple categorization of demographic
variables, we will be able both to compare the effectiveness of these
two categorizations of birthplace in explaining linguistic variation and
to sum them into one birthplace variable.

Statistical Analysis. The statistical technique utilized in this report

is that of analysis of variance via multiple regression analysis. Anal-
ysis of variance is a technique designed to answer questions concern-
ing the separate significance as well as the interactional significance
of several simultaneous effects. In the context of the present study,
analyses of variance can tell us whether context, age, education, or
birthplace are separately significant in explaining variation in the
production of a particular linguistic variant or whether the inter-
action between any two of them, e.g., between context and birthplace,
has explanatory significance. Multiple regression analysis is a tech-
nique designed to answer questions concerning the value of utilizing
additional explanatory parameters beyond those already utilized at
any given stage in the explanatory process (Bottenberg and Ward 1963;
Cohen 1965,1968a, 1968b). In the context of the present study multiple
regression analysis can tell us whether or not certain explanatory
parameters (e.g., context plus age) are already so powerful in explaining
variation in the production of a particular linguistic variant that it is
not necessary or productive to add other explanatory parameters even
if the latter too are significantly related per se to the variation in
question.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
272 Joshua A. Fishman
Hypotheses. Spanish Variables. Our general hypothesis regarding
linguistic variation in Puerto Rican Spanish (PRS) in the speech com-
munity under study is that it will consist of contextual variation pri-
marily and demographic variation only secondarily. Except for re-
gionally related differences between speakers of highland origin and
speakers of coastal origin we consider our subjects as constituting a
single speech community. Our subjects have all learned the norms
of Spanish communicative competence pretty much in the same
way and at the same developmental period of their lives. These norms
incorporate contextual variation. Too few of our subjects have
had too little exposure to formal, educated Spanish to constitute an
educated network of the speech community. Such a network might
develop speech norms of its own that could significantly modify
(i.e., raise or lower) the contextual variation norms that exist for
the speech community as a whole.
Our general hypothesis is that beyond a highland-coastal difference
in a few variables no significant demographic factors will be en-
countered in explaining any linguistic variation that may exist in
Puerto Rican Spanish above and beyond contextual variation. This
hypothesis will be tested here against one illustrative Spanish variant
where a variant is described as one of the realizations that a variable
can assume.
English Variables. With respect to linguistic variation in Puerto
Rican English in the speech community under study our general hypo-
thesis is that it will exist of demographic variation primarily and
contextual variation secondarily (if at all). We do not view our
subjects as constituting a unitary English speech community with
its own contextual norms of communicative competence in that
language. In general, the English-speaking horizons and experiences of
most of our subjects are still too limited for contextual varieties of
English to have developed (or to have been adopted) and to have been
stabilized. On the other hand, there are within the speech community
those whose English has been significantly modified by substantial
influences stemming from outside the community, such as those that
derive from American education in particular and increased time
in the continental United States in general. We would expect their
English to differ from those with other demographic characteristics
who have not had these experiences. We expect these differences
between demographic groups to be pervasive in their use of English
rather than contextualized along a casualness-carefulness dimension
for intra-group purposes. This hypothesis will be tested here against
one illustrative English variant.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 273
Results. Spanish Variant SpC-O. SpC-0 refers to the dropping of
the plural marker s when the following word begins with a consonant.
An example of this realization is (los) muchacho comen as opposed
to the standard realization (los) muchachos comen (SpC-1) or the
common PRS variation (los) muchachoh comen (SpC-2). This
variable (SpC) had a very high number of occurrences, and the reali-
zation in question showed considerable contextual variation, ac-
counting for just 17 percent of the cases of SpC in the most formal
context but 62 percent in the least formal context (Ma and Herasim-
chuk 1968). 5 in this morphophonemic environment was realized
quite differently from s in other environments. For instance, s before
a consonant within a word showed zero realization only 11 percent of
the time in the least formal context. Similarly, s marking a plural
article preceding a word beginning with a consonant was realized as
zero only 23 percent of the time in the least formal context. In these
environments S-2 or [h] was the preferred realization 81 percent
and 70 percent of all times respectively in style A. Thus SpC is
definitly a favorable environment for zero realization of s, with the
further advantage, for our present purposes, that there was substan-
tial variation in the realization of SpC-0 across contexts. Under
these circumstances, then, we decided to ask whether other parameters
of a directly demographic nature might also be significantly related
to differential production of SpC-0.
If we examine the first column in Table 3 (labeled r), we will
note that only context, in each of its aspects, correlates significantly
with differential use of SpC-0. The second aspect of context (that
which differentiates between word naming and B + A) correlates
with SpC-0 as well (0.423) as do both aspects taken together (co-
lumn 3, R = 0.424).
The fact that only the two aspects of context correlate significantly
with SpC-0 is corroborated in column 8, where only the two aspects
of context yield significant F ratios. Thus we can safely conclude
that in the speech community under study demographic differences
are not significantly related to differential use of SpC-0, whereas
contextual differences are so related. However, if we are to stop
our prediction of SpC-0 with context alone, we will have accounted
for only 18 percent of the casual variance (see column 6). If we add
sex of speaker to the prediction of SpC-0, we can account for 24.4
percent of the casual variance. This increase is due to the fact that
there is a slight tendency (column 1: r = —0.240) for males to use
SpC-0 more frequently than females.
If we continue to add successive demographic variables, our
multiple prediction of SpC-0 continues to rise (see column 5) and

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
274 Joshua A. Fishman

S vi
S ri V V V V

0\ ÛS o ©
ri
' 'h, ri V V V

o O
•o '.
tt,i- vi
o
(M V V V V V V V ri

v
/oi
i
o
o o o o
Ö ö ö o

Oí o
005
Ss n
u
3 O r)
Ö ö

TT M
Ít"3 S
w © o\ OO
V
©
U
o r) -<t
00
O o o
o o o

"t o
rJ
a-6« r) r) m
O ö o

T Tfr
r) © © t-i
-t
t o00 8
o
o
o o O O o n
o <
/-
o
o o o o O © Ö
1-H VO
O •Al r) ,-H n o\
i-H r) VD VO
Crt

r) o O 1 r—t
© © n
f i

O © © © © © © ©

¡ó <
>.iyj > c/3
A« 13
1M o i >
m
-
« 5 •Öo
<u
m "3 'S Is Cd u
cd
.¡5 s00 " B
^ a > <> J3
s «5 « Pi en
C/3
•g
£ £ £ .. >
O 3
a —ed
.
ft*
X
V V Vo ai ê^ ,c3l t
»
e
d "E S
a J? O c
O o S <ÖJ) < &QB,fco •
tO
u •
a o
3h ° m
(5 3

O <J w M « O

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 275
finally reaches the appreciable figure of 0.602. A multiple correlation
of this magnitude accounts for 36.2 percent of the causal variance
in SpC-0, a substantial increase beyond that accounted for by context
alone.
Although none of the demographic variables is significantly related
to differential use of SpC-0, sex of speaker approaches such signifi-
cance. This, however, is due to the fact that in the speech community
under study more women than men are of highland origin in Puerto
Rico. The context by birthplace interaction, therefore, also approaches
significance, which indicates that some birthplace groups show more
contextual variation than do others.

Table 4 reveals the mean number of occurrences of SpC-0 in the

three different contexts for our sample as a whole and for two dif-
ferent birthplace subsamples. This table confirms that the effective
contextual difference comes between WN and the two conversational
styles. Table 4 also confirms the greater contextual sensitivity of
Highland born subjects for whom we find greater average contextual
differences than those found for other subjects.

TABLE 4
Contextual Differences in Mean Number of Occurrences of SpC-0,
for Total Sample and for Birthplace Groups

Birthplace Contexts
Total
groups
WN B A

Highland 27.13 57.27 66.58 49.17

Other 30.38 53.29 57.05 56.09

Total 29.13 54.17 59.87 54.39

English Variant EH-2. EH-2 represents the Standard American

English sound [ae], as in cat, bad, ham. Two other variants of this EH
variable were recognized: EH-1, as in New York City [kedt, b«(5d,
h£<Sm], and EH-3, as in accented English cah'nt, bahd, hahm. EH-2
serves fairly effectively to differentiate accented from native English
speakers, as the, sound is not available in Spanish phonology. Mastery
of this phone seems to imply mastery of a number of other typically
English sounds not available in Spanish.
Use of the three variants of E H changed but slightly and irre-
gularly with context (Ma and Herasimchuk 1968), which support
the hypothesis of more or less fixed usage of one sound by any given

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
276 Joshua A. Fishman
speaker. EH-2 showed an overall higher incidence of occurrence
and, for this reason, was chosen over EH-1 for testing. It is also less
ambiguously American; EH-1 can be approximately by the Spanish
[e] or [e], so a score of EH-1 does not clearly isolate the sound as
English but rather marks some form or other of dialect-realization. For
reasons both of numerical frequency and of phonological exclusiveness,
then, EH-2 is a very good variant for the statistical testing of relation-
ships between differential use of sounds and the characteristics of their
users.
Table 5 reveals quite a different picture from that shown in Table
3. The values in column 1 indicate that neither of the two aspects of
context are significantly related to differential use of EH-2. Indeed
even when both aspects of context are taken together, it is still the
least important multiple predictor of EH-2 except for sex of speaker
(column 3). If we utilize context alone, we are able to account for
only 3.6 percent of the causal variance pertaining to differential use
of EH-2 (column 6). If we add sex of speaker to context, our predic-
tion rises only to 5.8 percent. However, as soon as we consider such
demographic variables as age, education, and birthplace the picture
changes radically.
Of the three major demographic variables related to differential
use of EH-2, the most important is clearly education (column 1). If
we combine all three aspects of education, we obtain a multiple cor-
relation of 0.753 (column 3), which itself accounts for 56.7 percent
of the causal variance (column 4).
Those of our subjects who were partly or entirely educated in the
United States are more likely to utilize EH-2 than those entirely
educated in Puerto Rico (note minus correlations in column 1).
This relationship between differential use of EH-2 and education is
further clarified in Table 6 which reveals it to be consistent for each
speech context.
If education is now combined with the variables that precede it
in Table 5 (context, sex of speaker, and age), then the resulting cumu-
lative multiple correlation with EH-2 rises to 0.785 (column 5),
and we have accounted for 61.6 percent of the causal variance in
differential use of EH-2 (column 6).
Although neither age nor birthplace are as strongly related to
EH-2 as is education, their independent correlations with EH-2 are
clearly significant (columns 1 and 8). When all three of them are added
to context and sex of speaker, we arrive at a cummulative correlation
of 0.810 (column 5), which indicates that we have accounted for
65.6 percent of the causal variance in differential use of EH-2 (co-
lumn 6).

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 2 7 7

1-H rt
n
V Tf V V

rt X3 A
v
r-
Tf \o
w-i o\ tn
<N

X3 XI J2 Ta
II
II
1
.0 <4
ts 00 ©
0S
<rt
V V V Ov © vi
<N
\o
<s
c4 VO ri wi

<N

¿3
hi <s © 00 © 00
00 r- •<t ©
-s: © cs © ©
© © © © ©

c
hi
\o 00
0
00 VO

m
c ro >0 m
« O 0 NO VO VO

§
O © ©" © O* ©

o\
00
t-H r)
00
»n ©
00
3 <s «n r- 00 00
OH 0 0 © © © ©

?! VO
©
r-
VO 00
lO O *n CS
w « O © © © ©
•J c

K
H VO
0\
Ov VO 00
•2 00 W-l cs
< «1 r-
0 © 0 © ©
^

ft! 0 m 00 <ri 00
Tf
o\ I/--.
OS
m
00
f«1 © n ON
© O 0 (N m v-l 0
0 O © O © © © © O e> ©
**

" 3
•<t VO -"J- r- o\ VO o\ 00
r- i-H m VI «s 00 ©
T-H TH i-H Wi I- t- m Tl-
•S 0 © © © © © © © © ©
1
l l 1 l 1
u
c
«i M wi ^
•2
Si >> > <A
a A C3
+ «5
JC >
n 1-- >
0
> o
Ji
u
£ 5 >

£ « v
n m £ o •s
ca u
•5 »
s? I B i 3 o • a a ca
o- *o
> > a § "3.
. G

K
«o co
C/3
to
t
os
w-i 3 n

v v v .
u
rt u co x
cs ,CO
2 »' 3
,3 M
09
S 2
o i J
c 3

$ < <
o § x o
<J 6 •o
U
oi
PL,
^
O
•O
W n O

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
278 Joshua A. Fishman

TABLE 6
Contextual Differences in Mean Number of Occurrences of EH-2 for
Total Sample and for Educational Groups

Educational Contexts
Total
groups
WN B A

Educated 15.75 16.43 19.40 16.46

entirely in
Puerto Rico
Educated 60.71 64.43 65.17 63.35
partially
or entirely
in U S A

Total 35.79 38.57 57.71 40.20

Although context itself is not significantly related to differential

use of EH-2, the interaction between context and birthplace is signi-
ficantly related to such use. This implies that certain birthplace groups
show more contextual variation than do others. Whereas our sample
as a whole increasingly uses EH-2 as it proceeds from WN (35.79)
to B (38.57) to A (51.71), this variation occurs primarily between
B and A for our highland-born subjects and between WN and B
for other subjects, with the latter using EH-2 more frequently in all
contexts.
Incremental Prediction of EH-2. Not only are age and education
significant variables in accounting for differential use of EH-2 but
they are also incrementally significant in this respect. Column 10 of
Table 5 reveals that it pays to add age as a predictor of differential
use of EH-2 when one has previously used only context and sex of
speaker in this connection. Another way of saying this is that 0.338
(column 6), the cumulative prediction of EH-2 based on three variables
(context, sex of speaker, and age), is significantly better than the
cumulative prediction based on only the first two (0.058). Similarly,
Table 5 indicates that it pays to add Education as well to our prediction
of differential use of EH-2, even after context, sex of speaker, and
age have been used cumulatively in this connection. The cumulative
prediction of EH-2 based upon these four variables (0.616) is
significantly greater than that based on the first three (0.338).
The same cannot be said, however, with respect to birthplace or
the interaction between birthplace and context. Although it is true
that their cumulative addition to the prediction of differential use
of EH-2 (after context, sex of speaker, age, and education have been

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 279
cumulatively utilized for this purpose) does increase the multiple
prediction of EH-2 from 0.616 to 0.656 to 0.664, these increases,
though welcome, are not statistically significant. Thus, if birthplace
were an expensive or difficult measure to obtain, we would be justi-
fied in deciding to forego it because it does not produce a significant
increment in our efforts to account for differential use of EH-2.
There have recently been several other studies of the importance
of demographic factors in accounting for the variability of usage (see,
e.g., Ellis 1967, Huffine 1966, Jernudd 1968, McCormack 1968).
The study just reported gains considerably from the fact that it
sought to compare demographic with contextual variation, and to do
so in quantitative terms, as well as to do so separately for each
of the languages used in a functioning community (rather than by
a random sample of speakers).

Conclusions. The foregoing analysis of SpC-0 shows that its variable

realization was primarily attributable to contextual-situational varia-
tion along a continuum of formality-informality. Whereas demographic
factors (not social class in this case since our subjects were so uni-
formally of the lower class) added to the overall prediction of this
variable - as did the interaction between demographic factors and
speech context - it is clear that these are of lesser importance than the
speech-community-wide norms relating SpC-0 to informality rather
than to formality. Scores of other Spanish phonological variables
behave in this same way in the Puerto Rican neighborhood under
study. As a result we may consider it a single, relatively homogeneous
speech network as far as Spanish phonology is concerned, i.e., one
in which experiential differences have not resulted in the for-
mation of significantly different groups within the population with
substantially unique speech norms of their own. Our Puerto Rican
subjects are behaving more like Labov's Lower East Siders than like
Geertz's Javanese in this respect.
Just the opposite seems to be true vis-a-vis variability in the realiza-
tion of English phonology. In connection with EH-2 - and scores of
other English variables — no neighborhood-wide contextual-situational
variation has as yet developed. Those individuals who have spent
larger proportions of their lives in the USA and who have obtained
more formal education in the USA have a different English phonology
than do their more recently arrived and less American-educated
neighbors. Instead of a single set of speech community norms with
respect to English phonology there are several different demographic
subgroups (social classes if you like) each with their own substantially
different English phonologies used consistently in all contexts (by and

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
280 Joshua A. Fishmati
large). Our Puerto Rican subjects are behaving more like Geertz's
Javanese than like Labov's Lower East Siders in this respect. Without
common contextual norms vis-a-vis English phonology they are frag-
mented into more and less advantaged discontinuous strata insofar as
English phonology is concerned.
More generally stated in conclusion, the existence of societally shared
contextual variation is a sure indication of the existence of a speech
community or speech network. Societally shared contextual variation is
indicative of social interaction governed by common normative regu-
lations. On the other hand, demographic variation alone is not neces-
sarily indicative of the existence of a speech community or speech
network. Indeed, demographic variation in usage is, in and of itself,
ambiguous in this very respect. On the one hand, it may be merely
indicative of separate experiential groups (e.g., separate castes, social
classes, regional origin groups, etc.) that are required to interact in
marginal or limited ways. On the other hand, demographic variation
may be indicative of relatively pervasive, inflexible and compartment-
alized role-relationships within a speech community, such that members
of network X always utilize variety x, members of network Y always
utilize variety y, etc. Sorenson 1967 has described multilingual speech
communities of this kind in the Northwest Amazon region.
The co-occurrences of contextual and demographic variations must
not, therefore, be considered a necessary feature of speech com-
munities. It reflects a degree of interaction, a degree of complexity
of stratification, and a degree of shared open-network access and
repertoire fluidity that are by no means encountered everywhere.

5.4 Non-proletarians of All Regions, Unite!

In a relatively open and fluid society there will be few characteristics

of lower class speech that are not also present (albeit to a lesser
extent) in the speech of the working and lower middle classes. Whether
we look to phonological features such as those examined by Labov
or to morphological units such as those reported by Fischer (1958)
(Fischer studied the variation between -in' and -ing for the present
participle ending, i.e., runnin' vs. running - and found that the
former realization was more common when children were talking to
each other than when they were talking to him, more common among
boys than among girls and more common among 'typical boys' than
among 'model boys'), we find not a clearcut cleavage between the
social classes but a difference in rate of realization of particular
variants of particular variables for particular contexts. Even the

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 281
widely publicized distinction between the 'restricted code' of lower
class speakers and the 'elaborated code' of middle class speakers
(Bernstein 1964, 1966) is of this type, since Bernstein includes the
cocktail party and the religious service among the social situations in
which restricted codes are realized. Thus, even in the somewhat more
stratified British setting the middle class is found to share some of the
features of what is considered to be 'typically' lower class speech.
Obviously then, 'typicality', if it has any meaning at all in relatively
open societies, must refer largely to repertoire range rather than
primarily to unique features of the repertoire.
This is the most suitable point at which to observe that between
Bernstein's view that lower class speech is typically more restricted
and Labov's view that lower class speech is typically more informal
there is an implied contradiction, if "restricted" is defined as more
predictable and "informal" as less predictable. Actually, the contra-
diction is more apparent than real. In terms of speech repertoire range
both investigators would agree that the range of the lower class is typi-
cally narrower than that of the middle and upper middle classes. This is
what Bernstein is reacting to when he considers lower class speech more
restricted and, therefore, more predictable. On the other hand both
investigators would certainly agree that the phonological, lexical or
grammatical markers of lower class speech more commonly resemble
those of informal usage within the larger speech community. However,
as far as redundancy of speech is concerned, one must distinguish
between predictability between varieties and predictability within any
of them. Lower class usage may well be more predictable or redundant
when between-variety variation is considered, as Bernstein claims, and
yet be more eliptical and incomplete than middle or upper middle class
usage when within-variety variation is considered. When Joos and
others point to the greater redundancy (ritualization, predictability) of
frozen and other more formal styles they are reacting to within-variety
rather than between-variety variation. Thus, rather than being in
conflict, Bernstein and Labov, taken together, sensitize us additionally
to two different but equally important types of variation in the speech
behavior of socially variegated speech communities.
Those speech networks with the widest range of experiences, inter-
actions and interests are also those that have the greatest linguistic
repertoire range. In many speech communities these networks are
likely to be in one or another of the middle classes since some networks
within these classes are most likely to maintain direct contact with
the lower and working classes below them (in employer-employee,
teacher-pupil and other role-relationships), as well as with the upper
class above them (in educational, recreational and cultural inter-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
282 Joshua A. Fishman
actions). However, whereas the repertoire ranges of the upper and
lower classes are likely to be equally discontinuous (even if not
equally restricted) there is likely to be a very major distinction between
them if the larger speech community (the region, the country) is
considered. Lower classes tend to be regionally and occupationally
separated from each other to a far greater extent than do upper
and middle classes (Gumperz 1958). Thus, there may well be several
different lower class varieties in a country (depending on regional
and on occupational or other specializations), while at the same
time upper and upper-middle class speech may attain greater uni-
formity and greater regional neutrality. The more advantaged classes
travel more frequently, engage in joint enterprises more frequently,
control the agencies of language uniformation (schools, media, language
planning agencies and government per se). They more quickly arrive
at a common standard, at least for formal occasions, than do
the lower classes who remain fragmented and parochial. Differences

(Tanner, 1967)

Figure 11. Functional Specialization of Codes in Indonesia

and Among the Case Study Group (Tanner, 1967)

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 283
such as these are illustrated in Nancy Tanner's case study of an Indo-
nesian elite group (1967; see Figure 11). Whereas the lower classes
speak only their local ethnic language, the middle and upper classes
also speak several varieties of Indonesian (including a regionally neutral
variety that is least influenced by local characteristics) and the elites
speak English and Dutch as well. One can predict that as these elites
lose their local ties and affiliations and assume Pan-Indonesian roles,
establishing speech communities of their own in Djakarta and in a few
other large cities, their need for local languages and for locally influen-
ced and informal Indonesian will lessen and their stylistic variation will
proceed, as it has with elites in England, France, Germany, Russia and
elsewhere in the world, via contrasts with foreign tongues.
Another way of arriving at the conclusions indicated above con-
cerning the greater discontinuity between the lower class varieties
than between upper or middle class varieties in most relatively open
societies is to consider the differences referred to in Figure 12. Here
Lexical Phono- Grammat-
logical ical
NETWORKS DRAWN FROM DIVERSE" D¡^er_
Differ- Differ-
enees ences ences

EXTRA-REGIONAL-Origin Groups ++ ++ ++
RACIAL Groups ++ ++ ++
Ethnic Groups ++ + +
O C C U P A T I O N A L Groups ++ + +
Religious Groups + + +
A G E Groups ++ — —

SEX Groups ++ — —

Legend: + 4- = Substantial differences are judged to exist between categories

(e.g., between different age groups) on the diversity parameter in
question.
Legend: + = M o d e r a t e differences are judged to exist between categories on
the diversity parameter in question.
Legend: — =Negligible or no differences are judged to exist between cate-
gories on the diversity parameter in question.

a
Categories are compared on the assumption that all other bases of group-
functioning are held constant when networks are selected at the level of any
given diversity parameter. Thus, when considering networks drawn from diverse
occupational groups judges were asked to assume that racial, ethnic, religious and
other diversity parameters were held constant.
Figure 12. Extent of Linguistic Differences and Extent of Socio-
Cultural Differences Within Various Kinds of Speech Networks
(as judged by Stanford students native to the Lower Peninsula)

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
284 Joshua A. Fishman
we note that when all other factors are held constant co-territorial
groups of diverse regional origin may frequently be expected to
differ most profoundly linguistically. The lower classes are exactly
those whose regional origins are most diversified in most cities the
world over. Indeed, the lower classes are likely to be more heteroge-
neous than the upper classes in exactly those factors - whether they
be diversity of origin or diversity of experience - that are associated
with more than peripheral lexical differences between co-territorial
populations. They are far more likely to be regionally, socially, cul-
turally, occupationally and religiously diverse than are the upper
classes whose self-uniformizing tendencies and capacities have al-
ready been mentioned. Indeed, it is only in connection with sex and
age variationability that the lower classes are often more homogeneous
than the upper but these generally tend to be associated only with the
more marginal linguistic differences.
As a result of the differential experiences and opportunities vis-a-vis
uniformation to which they are exposed, social class differences in rela-
tively open societies have commonly arrived at the following state of
affairs: (a) the middle and upper middle classes have larger repertoires
in language and in social behavior than do the lower classes; (b) the
lower classes tend to remain more diverse - regionally, ethnically, reli-
giously, racially, etc. - than the upper classes and, therefore, there are
preserved more and more discontinuous varieties of lower class speech
than of upper class speech. These two tendencies are not in conflict
with each other, except as social conflict itself may exist and, therefore,
come to disturb whatever societal and usage patterns have been
stabilized. They are both due to societal differentials in normal social
class role ranges and in exposure to the uniformizing institutions of the
larger polity.

5.5 Diversification vs. Massification

One further consideration deserves at least brief attention in our review
of societal differentiation and language variation; namely, the common
view that there is a trend toward overall uniformation, in language and
in other social behavior, as industrialization progresses (Bell 1961;
Boulding 1963; Hertzler 1965; Hodges 1964). It is undeniable that life
in urbanized and industrial countries is in some ways more uniform
than is the case in countries where local and regional particularisms
remain relatively untouched. Nevertheless, it seems to be erroneous to
think of preindustrial rural heterogeneity and industrial urban homoge-
neity as either accurate or mutually exclusive designations. Both stages
of development seem to foster as well as to inhibit certain kinds of uni-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 285
formation and differentiation in language as well as in other aspects of
behavior.
Certainly, the preindustrial rural society is not as internally hetero-
geneous as is the urban society with its variety of classes, religions,
ethnic groups, and interest groups. Thus, the supposedly uniformizing
effect of urbanization and industrialization must pertain to inter-regio-
nal or inter-urban comparisons rather than to intra-urban or intra-local
ones. Nevertheless, the best available evidence indicates that no trend
toward inter-regional homogeneity in religion, politics or other genera-
lized behaviors is apparent in the United States (Glenn 1966, 1967a,
1967b), nor are such trends apparent in other countries, such as
England, France, Holland or Belgium that have been industrialized or
urbanized for the greatest length of time. There the differences in
values, tastes, social and political orientations between manual and non-
manual workers seem to be as great or greater than they are today in
the United States (Hamilton 1965; Bonjean 1966; Schnore 1966; Broom
and Glenn 1966, etc.).
At the language level both uniformation and differentiation are found
to go on simultaneously, indicative of the fact that the traditional and
the modern are frequently combined into new constellations rather than
displaced one by the other. Uniformation pressures seem to be strongest
in conjunction with only certain varieties within a speech community's
verbal repertoire as well as in conjunction with only some of the inter-
action networks of that community. The language variety associated
with school, government and industry tends to be adopted differentially,
the degree of its adoption varying with the degree of interaction in these
domains. Not only need such adoption not be displacive (particularly
when populations remain in their former places of residence) but -
even though the adoption may be quite uniform and official for an en-
tire country - it may remain an entirely passive rather than active com-
ponent in the repertoire of many interaction networks. Thus, even
though television viewing and radio listening are most frequent and
prolonged among the lower classes their overt repertoires seem to be
little influenced by such viewing or listening.
Finally, it should be recognized that urbanization may also foster
certain kinds of differentiation. Whereas the number of different ethnic
groups (and, therefore, the number of mutually exclusive language
groups) may decline, new social differentiations and new occupational
and interest groups normally follow in the wake of industrialization.
These latter commonly develop sociolects and specialized usages of
their own, thus expanding the repertoires of many speakers. Even the
rise of languages of wider communication frequently results in differen-
tiation rather than in uniformation. The spread of English as a second

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
286 Joshua A. Fishman
language in the past 50 years has resulted in there being more varieties
of English today (including Indian English, East African English, Fran-
glais, Spanglish and others) rather than less. It is, of course, true that
certain languages, now as in the past, are in danger of dying out.
Nevertheless, others frequently regarded as 'mere varieties' rather than
as full-fledged languages, are constantly being 'born' in terms of diffe-
rentiating themselves within the linguistic repertoires of certain inter-
action networks, and, at times, of entire speech communities. Moderni-
zation is a complex phenomenon. While it depresses the status and
decreases the number of speakers of certain varieties (e.g., in recent
years: Frisian, Romansch, Landsmal, Yiddish) it raises the status and
increases the speakers of others (Macedonian, Neo-Melanesian, Indo-
nesian, Swahili, etc.).
Our own American environment is an atypical example. It reveals
the uniformation that results from the rapid urbanization and industria-
lization of dislocated populations. We must not confuse the American
experience with that of the rest of the world (Greenberg 1965). In addi-
tion, we must come to recognize that American uniformation, whether
in speech or in diet, is at times a surface phenomenon. It is an added
variety to the repertoires that are still there and that are still substantial
if we will but scratch a little deeper (Fishman 1967a).

6.0 SOCIETAL BILINGUALISM: STABLE AND TRANSITIONAL

Societal bilingualism has been referred to so many times in the previous

pages that it is time that we paused to consider it in its own right
rather than as a means of illustrating more general sociolinguistic
phenomena. The psychological literature on bilingualism is so much
more extensive than its sociological counterpart that workers in the
former field have often failed to establish contact with those in the
latter. It is the purpose of this section to relate these two research
traditions to each other by tracing the interaction between their two
major constructs: bilingualism (on the part of psychologists and psycho-
linguists) and diglossia (on the part of sociologists and sociolinguists).

6.1 Diglossia
In the few years that have elapsed since Ferguson (1959a) first ad-
vanced it, the term diglossia has not only become widely accepted by
sociolinguists and sociologists of language, but it has been further ex-
tended and refined. Initially it was used in connection with a society
that recognized two (or more) languages or varieties for intrasocietal

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 287
communication. The use within a single society of several separate codes
(and their stable maintenance rather than the displacement of one by
the other over time) was found to be dependent on each code's serving
functions distinct from those considered appropriate for the other code.
Whereas one set of behaviors, attitudes and values supported, and was
expressed in, one language, another set of behaviors, attitudes and
values supported and was expressed in the other. Both sets of behaviors,
attitudes and values were fully accepted as culturally legitimate and
complementary (i.e., non-conflictual) and indeed, little if any conflict
between them was possible in view of the functional separation between
them. This separation was most often along the lines of a H(igh)
language, on the one hand, utilized in conjunction with religion, edu-
cation and other aspects of High Culture, and an L(ow) language, on
the other hand, utilized in conjunction with everyday pursuits of hearth,
home and lower work sphere. Ferguson spoke of H as "superposed"
because it is normally learned later and in a more formal setting than
L and is, thereby, superposed upon it.
To this original edifice others have added several significant conside-
rations. Gumperz (1961, 1962, 1964a, 1964b, 1966) is primarily res-
ponsible for our greater awareness that diglossia exists not only in
multilingual societies which officially recognize several 'languages', and
not only in societies that utilize vernacular and classical varieties but,
also, in societies which employ separate dialects, registers, or functio-
nally differentiated language varieties of whatever kind. He has also
done the lion's share of the work in providing the conceptual apparatus
by means of which investigators of multilingual speech communities
seek to discern the societal patterns that govern the use of one variety
rather than another, particularly at the leven of small group interaction.
Fishman (1964, 1965a, 1965c, 1965d, 1965e, 1966a, 1968c), on the
other hand, has attempted to trace the maintenance of diglossia as well
as its disruption at the national or societal level. In addition he has
attempted to relate diglossia to psychologically pertinent considerations
such as compound and coordinate bilingualism (1965b). Finally, Kaye
(1970) has indicated that diglossia is often a far more flexible,
changeable and even ill-defined status, particularly in its linguistic
aspects, than has often been presumed. The present section represents
an extension and integration of these several previous attempts.
For purposes of simplicity it seems best to represent the possible
relationships between bilingualism and diglossia by means of a four-
fold table such as shown in Figure 13.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
288 Joshua A. Fishman
6.2 Speech Communities Characterized by both Diglossia and Bilin-
gualism

The first quadrant of Figure 13 refers to those speech communities in

which both diglossia and bilingualism are widespread. At times such

DIGLOSSIA
+ —
BILINGUALISM

1. Both diglossia 2. Bilingualism

+
and bilingualism without diglossia

3. Diglossia without 4. Neither diglossia

bilingualism nor bilingualism

Figure 13. The Relationships between Bilingualism and Diglossia

communities comprise an entire nation, but of course this requires ex-

tremely widespread (if not all-pervasive) bilingualism and, as a result,
there are really few nations that are fully bilingual and diglossic. An
approximation to such a nation is Paraguay, where more than half of
the population sneaks both Spanish and Guarani (Rubin 1962, 1968).
A substantial proportion of the formerly monolingual rural population
has added Spanish to its linguistic repertoire in connection with matters
of education, religion, government, and High Culture (although in the
rural areas social distance or status stressing more generally may still be
expressed in Guarani). On the other hand, the vast majority of city
dwellers (being relatively new from the country) maintain Guarani for
matters of intimacy and primary group solidarity, even in the midst
of their more newly acquired Spanish urbanity (See Figure 14). Note
that Guarani is not an 'official' language (i.e., recognized and utilized
for purposes of government, formal education, the courts, etc.) in
Paraguay, although it was finally recognized as a "national language"
at the 1967 constitutional convention. It is not uncommon for the H
variety alone to be recognized as 'official' in diglossic settings without
this fact threatening the acceptance or the stability of the L variety
within the speech community. However, the existence of a particular
'official' or 'main' language should not divert the investigator from
recognizing the fact of widespread and stable multilingualism at the
levels of societal and interpersonal functioning (see Table 7).
Below the level of nationwide functioning there are many more
examples of stable diglossia co-occurring with widespread bilingualism.
The Swiss-German cantons may be mentioned since their entire popu-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
 The Sociology of Language 289

(Joan Rubin 1968)

Location

Rural-Guarani Non-Rural

I
Formality-Informality

Formal-Spanish Non-Formal

Non-Intimate Intimate
Spanish

t
Seriousness of Discourse

Non-Serious Serious
Guarani

f
First Language Learned
Predicted Language Proficiency
Sex
Figure 14. National Bilingualism in Paraguay:
Ordered Dimensions in the Choice of Language in a Diglossic Society
(Joan Rubin, 1968)
lation of school age and older alternates between High German (H)
and Swiss German (L), each with its own firmly established and highly
valued functions (Ferguson 1959a; Weinreich, U. 1951,1953a). Hughes
(1970) has demonstrated how English (H) and French (L) diglossia-
and-bilingualism are peripheral and external in many Montreal agencies

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
290 Joshua A. Fishman

TABLE 7
Linguistic Unity and Diversity, by World Region

No. of Countries by Percent of Population Speaking Main Language

x uiai
160-69 1j 50-59| 40-49 | 30-39
90-100 180-89| 70-79| | 20-29
| 110-14
| | 10-100%
Europe 17 4 2 2 2 — — — — 27
East and South
Asia 5 3 4 3 1 4 — 1 — 21
Oceania» 2 2
Middle East and
Northern Africa 8 6 2 3 1 2 — — — 22
Tropical and
Southern Africa 3 — — 2 5 8 7 5 3 33
The Americas 15 6 — — 2 2 1 — — 26
World Total 50 19 8 10 11 16 8 6 3 131

Source: Table 1 (D. Rustow, 1967)

a Not including New Guinea, for which no breakdown by individual languages
was available.
and businesses in which clients (or customers) and management (or
owners) must interact although coming from different origins. On the
other hand, in plants (where no customers/clients are present) the
communication between workers and management reveals bilingualism-
and-diglossia of a hierarchical and internal nature. Traditional (pre-
World War I ) Eastern European Jewish males communicated with each
other in Hebrew ( H ) and Yiddish ( L ) . In more recent days many of
their descendants have continued to do so in various countries of
resettlement, even while adding to their repertoire a Western language
(notably English) in certain domains of intragroup communication as
well as for broader intergroup contacts (Fishman 1965a, 1965e; Wein-
reich, U. 1953a; Weinreich, M . 1953). This development differs signifi-
cantly from the traditional Eastern European Jewish pattern in which
males whose occupational activities brought them into regular contact
with various strata of the non-Jewish coterritorial population utilized
one or more coterritorial languages (which involved H and L varieties
of their own, such as Russian, German or Polish on the one hand, and
Ukrainian, Byelorussian or 'Baltic' varieties, on the other), but did so
primarily for intergroup purposes. A similar example is that of upper
and upper middle class males throughout the Arabic world who use
classical (Koranic) Arabic for traditional Islamic studies, vernacular
(Egyptian, Syrian, Lebanese, Iraqui, etc.) Arabic for informal conver-
sation, and, not infrequently, also a Western language (French or
English, most usually) for purposes of intragroup scientific or techno-
logical communication (Blanc 1964; Ferguson 1959a; Nader 1962).
A l l of the foregoing examples have in common the existence of a

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
 The Sociology of Language 291

*n >n u-i in

f-' ^ N o «^
3 r f t rt rt N N
01
*e3

r - oo m »o r -

§ tL,
I

£
e O O "t I-;
o
s Tf 00 «
rn rn ci

S c
I
H
£ £
wi
c
1 0
Go 1
£ 01
•a
I
IP S c w-j vo r
o
s ^ pi -h r^ oo c i
N in \o
IP

C t-mcSTt.-ioovomr- ^ a\
3
•O s

•a
2
£3
O
2

« « ^ « ^ t >t t w
«? T "V +
<60 2 £ to
n
<o «o «ft o
cn ^f «o vo t -

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
292 Joshua A. Fishman
fairly large and complex speech community such that its members have
available to them both a range of compartmentalized roles as well as
ready access to these roles. If the role repertoires of these speech com-
munities were of lesser range, then their linguistic repertoires would
also be(come) more restricted in range, with the result that one or more
separate languages or varieties would be(come) superfluous. In addition,
were the roles not compartmentalized, i.e., were they not kept separate
by dint of association with quite separate (though complementary)
values, domains of activity and everyday situations, one language (or
variety) would displace the other as role and value distinctions merged
and became blurred. Finally, were widespread access not available to
the range of compartmentalized roles (and compartmentalized languages
or varieties) then the bilingual population would be a small, privileged
caste or class (as it is or was throughout most of traditional India or
China) rather than a broadly based population segment.
These observations must lead us to the conclusion that many modern
speech communities that are normally thought of as monolingual are,
rather, marked by both diglossia and bilingualism, if their several
registers are viewed as separate varieties or languages in the same sense
as the examples listed above. Wherever speech communities exist whose
speakers engage in a considerable range of roles (and this is coming
to be the case for all but the extremely upper and lower levels of
complex societies), wherever the access to several roles is encouraged or
facilitated by powerful social institutions and processes, and finally,
wherever the roles are clearly differentiated (in terms of when, where
and with whom they are felt to be appropriate), both diglossia and
bilingualism may be said to exist. The benefit of this approach to the
topic at hand is that it provides a single theoretical framework for
viewing bilingual speech communities and speech communities whose
linguistic diversity is realized through varieties not (yet) recognized as
constituting separate 'languages'. Thus, rather than becoming fewer in
modern times, the number of speech communities characterized by
diglossia and the widespread command of diversified linguistic reper-
toires has increased greatly as a consequence of modernization and
growing social complexity (Fishman 1966b). In such communities each
generation begins anew on a monolingual or restricted repertoire base
of hearth and home and must be rendered bilingual or provided with
a fuller repertoire by the formal institutions of education, religion,
government or work sphere. In diglossic-bilingual speech communities
children do not attain their full repertoires at home or in their neigh-
borhood playgroups. Indeed, those who most commonly remain at
home or in the home neighborhood (the pre-school young and the
post-work old) are most likely to be functionally monolingual, as

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 293
Lieberson's tables on French-English bilingualism in Montreal amply
reveal (see Table 8). Once established, and in the absence of rapid and
extensive social change, bilingualism under circumstances of diglossia
becomes an ingredient in the situational and metaphorical switching
patterns available for the purposes of intra-communal communicative
appropriateness. Many conversations and utterances demonstrably
'mean something else', depending on the language in which they are
expressed (Table 9) even when all other factors are kept constant
(Kimple et al. 1969).

6.3 Diglossia without Bilingualism

Departing from the co-occurrence of bilingualism and diglossia we
come first to polities in which diglossia obtains whereas bilingualism is
generally absent (quadrant 3). Here we find two or more speech com-
munities united politically, religiously and/or economically into a single
functioning unit notwithstanding the socio-cultural cleavages that sepa-
rate them. At the level of this larger (but not always voluntary) unity,
two or more languages or varieties must be recognized as obtaining.
However one (or both) of the speech communities involved is (are)
marked by relatively impermeable group boundaries such that for
'outsiders' (and this may well mean those not born into the speech
community, i.e., an emphasis on ascribed rather than on achieved status)
role access and linguistic access are severely restricted. At the same
time linguistic repertoires in one or both groups are limited due to role
specialization.
Examples of such situations are not hard to find (see, e.g., the many
instances listed by Kloss 1966a). Pre-World War I European elites often
stood in this relationship with their countrymen, the elites speaking
French or some other fashionable H tongue for their intra-group pur-
poses (at various times and in various places: Danish, Salish, Provencal,
Russian, etc.) and the masses speaking another, not necessarily linguis-
tically related, language for their intra-group purposes. Since the maj-
ority of elites and the majority of the masses never interacted with one
another they did not form a single speech community (i.e., their linguis-
tic repertoires were discontinuous) and their inter-communications were
via translators or interpreters (a certain sign of intra-group monolin-
gualism). Since the majority of the elites and the majority of the masses
led lives characterized by extremely narrow role repertoires their lin-
guistic repertoires too were too narrow to permit widespread societal
bilingualism to develop. Nevertheless, the body politic in all of its eco-
nomic and national manifestations tied these two groups together into
a 'unity' that revealed an upper and a lower class, each with a language

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
294 Joshua A. Fishman
TABLES 9 a n d 9 b
The Interpretation of Language Switching (English-Spanish) given
both Bilingualism and Diglossia (Kimple et al., 1969)
Analysis of variance for items requiring subjective judgement:
Conversation 1 : boy calls girl for date

Item no. Source df MS F

10 (length of family's Treatments 3 4.13 3.50»

residence in N.Y.) Within 45 1.18
11 (length of boy's Treatments 3 11.70 6.69b
residence in N.Y.) Within 45 1.75
12 (kind of job held Treatments 3 2.46 3.97a
by girl's father) Within 45 .62
13 Treatments 3 .17 .23
Within 45 .73
14 Treatments 3 .18 .51
Within 45 .35
15 Treatments 3 .32 .76
Within 45 .42
16 (naturalness of conversation Treatments 3 2.32 7.03 b
between boy and girl) Within 45 .33
17 (naturalness of conversation Treatments 3 2.19 3.91a
between mother and girl) Within 45 .56
18 Treatments 3 .67 1.45
Within 45 .46
a = p < 05 b = p < 01
Analysis of variance for items requiring subjective judgement:
Conversation 2: Invitation to stay for dinner

Item no. Source df MS F

9 Treatments 3 .37 .73

Within 45 .51
10 Treatments 3 3.33 1.29
Within 45 2.58
11 Treatments 3 1.42 1.89
Within 45 .75
12 Treatments 3 .06 .65
Within 45 .93
13 (naturalness of conversation Treatments 3 2.97 5.82b
between mother and guest) Within 45 .51
14 (naturalness of conversation Treatments 3 3.96 8.25b
between boy and guest) Within 45 .48

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 295
appropriate to its own restricted concerns. Some have suggested that the
modicum of direct interaction that does occur between servants and
masters who differ in mother tongue contributes to bringing into being
the marginal languages (pidgins) for which such settings are known
(Grimshaw, in press).
Thus, the existence of national diglossia does not imply widespread
bilingualism amongst rural or recently urbanized African groups (as
distinguished from somewhat more Westernized populations in those
settings); nor amongst most lower caste Hindus, as distinguished from
their more fortunate compatriots the Brahmins, nor amongst most lower
class French-Canadians, as distinguished from their upper and upper
middle class city cousins, etc. In general, this pattern is characteristic
of polities that are economically underdeveloped and unmobilized,
combining groups that are locked into opposite extremes of the social
spectrum and, therefore, groups that operate within extremely restricted
and discontinuous linguistic repertoires (Friederich 1962, Fishman
1969a, Pool 1969). Obviously such polities are bound to experience
language problems as their social patterns alter as a result of industrial-
ization, widespread literacy and education, democratization, and mod-
ernization more generally. Since few polities that exhibit diglossia
without bilingualism developed out of prior socio-cultural consensus or
unity, rapid educational, political or economic development experienced
by their disadvantaged groups or classes is very likely to lead to de-
mands for secessionism or for equality for their submerged language(s).
The linguistic states of Eastern Europe and India, and the language
problems of Wales and Belgium stem from origins such as these. This
is the pattern of development that may yet convulse many African and
Asian nations if their de-ethnicized and Westernized elites continue to
fail to foster widespread and stable bilingual speech communities that
incorporate the masses and that recognize both the official language(s)
of wider communication and the local languages of hearth and home
(Figure 15).

6.4 Bilingualism without Diglossia

We turn next to those situations in which bilingualism obtains whereas
diglossia is generally absent (quadrant 2). Here we see more clearly
than before that bilingualism is essentially a charcterization of individual
linguistic versatility whereas diglossia is a characterization of the social
allocation of functions to different languages or varieties. Under what
circumstances do bilinguals function without the benefit of a well un-
derstood and widely accepted social consensus as to which languages is
to be used between which interlocutors, for communication concerning

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
Brought to you by | The University of Texas at Austin
Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 297
what topics or for what purposes? Under what circumstances do the
varieties or languages involved lack well defined or protected separate
functions? Briefly put, these are circumstances of rapid social change,
of great social unrest, of widespread abandonment of prior norms before
the consolidation of new ones. Children typically become bilingual at a
very early age, when they are still largely confined to home and neigh-
borhood, since their elders (both adult and school age) carry into the
domains of intimacy a language learned outside its confines. Formal
institutions tend to render individuals increasingly monolingual in a
language other than that of hearth and home. Ultimately, the language
of school and government replaces the language of home and neighbor-
hood, precisely because it comes to provide status in the latter domaine
as well as in the former, due to the extensive social change to which
home and neighborhood have been exposed (See section 7, below).
Many studies of bilingualism and intelligence or of bilingualism and
school achievement have been conducted within the context of bilin-
gualism without diglossia (for a review see Macnamara 1966), often
without sufficient understanding on the part of the investigators that this
was but one of several possible contexts for the study of bilingualism
(Corpas 1969, Metraux 1965). As a result many of the purported
'disadvantages' of bilingualism have been falsely generalized to the
phenomenon at large rather than related to the absence or presence of
social patterns that reach substantially beyond bilingualism (Fishman
1965b, 1966a).
The history of industrialization in the Western world (as well as in
those parts of Africa and Asia which have experienced industrialization
under Western 'auspices') is such that the means (capital, plant, organi-
zation) of production have often been controlled by one speech com-
munity while the productive manpower was drawn from another
(Deutsch 1966). Initially, both speech communities may have main-
tained their separate diglossia-with-bilingualism patterns or, alternatively,
that of an overarching diglossia without bilingualism. In either case,
the needs as well as the consequences of rapid and massive industrial-
ization and urbanization were frequently such that members of the
speech community providing productive manpower rapidly abandoned
their traditional socio-cultural patterns and learned (or were taught)
the language associated with the means of production much earlier than
their absorption into the socio-cultural patterns and privileges to which
that language pertained. In response to this imbalance some react(ed)
by further stressing the advantages of the newly gained language of
education and industry while others react(ed) by seeking to replace
the latter by an elaborated version of their own largely pre-industrial,
pre-urban, pre-mobilization tongue (Fishman 1968c).

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
298 Joshua A. Fishman
Under circumstances such as these no well established, socially re-
cognized and protected functional differentiation of language obtains in
many speech communities of the lower and lower middle classes. Dislo-
cated immigrants and their children (for whom a separate 'political
solution' is seldom possible) are particularly inclined to use their mother
tongue and other tongue for intra-group communication in seemingly
random fashion (Fishman, Cooper and Ma 1968; Nahirny and Fishman
1965; Herman 1961). Since the formerly separate roles of the home
domain, the school domain and the work domain are all disturbed by
the massive dislocation of values and norms that result from simulta-
neous immigration and industrialization, the language of work (and of
the school) comes to be used at home. As role compartmentalization
and value complementarity decrease under the impact of foreign models
and massive change the linguistic repertoire also becomes less compart-
mentalized. Languages and varieties formerly kept apart come to in-
fluence each other phonetically, lexically, semantically and even gram-
matically much more than before. Instead of two (or more) carefully
separated languages each under the eye of caretaker groups of teachers,
preachers and writers, several intervening varieties may obtain differing
in degree of interpénétration. Under these circumstances the languages
of immigrants may come to be ridiculed as 'debased' and 'broken'
while at the same time their standard varieties are given no language
maintenance support.
Thus, bilingualism without diglossia tends to be transitional both in
terms of the linguistic repertoires of speech communities as well as in
terms of the speech varieties involved per se. 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). Furthermore, pidginization (the
crystallization of new fusion languages or varieties) is likely to set in
when members of the 'work force' are so dislocated as not to be able
to maintain or develop significantly compartmentalized, limited access
roles (in which they might be able to safeguard a stable mother tongue
variety), on the one hand, and when social change stops short of per-
mitting them to interact sufficiently with those members of the 'power
class' who might serve as standard other-tongue models, on the other
hand.

6.5 Neither Diglossia nor Bilingualism

Only very small, isolated and undifferentiated speech communities may
be said to reveal neither diglossia nor bilingualism (Gumperz 1962;

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 299
Fishman 1965c). Given little role differentiation or compartmentali-
zation and frequent face-to-face interaction between all members of
the speech community, no fully differentiated registers or varieties may
establish themselves. Given self-sufficiency, no regular or significant
contact with other speech communities may be maintained. Neverthe-
less, such groups - be they bands or class - are easier to hypothesize
than to find (Owens 1965; Sorensen 1967). All speech communities
seem to . have certain ceremonies or pursuits to which access is limited,
if only on an age basis. Thus, all linguistic repertoires contain certain
terms that are unknown to certain members of the speech community,
and certain terms that are used differently by different sub-sets of
speakers. In addition, metaphorical switching for purposes of emphasis,
humor, satire or criticism must be available in some form even in
relatively undifferentiated communities. Finally, such factors as exoga-
my, warfare, expansion of population, economic growth and contact
with others all lead to internal diversification and, consequently, to re-
pertoire diversification. Such diversification is the beginning of bilin-
gualism. Its societal normification is the hallmark of diglossia. Quadrant
four tends to be self-liquidating.
Many efforts are now underway to bring to pass a rapprochement
between psychological, linguistic and sociological work on bilingualism
(Fishman and Terry 1969). The student of bilingualism, most particu-
larly the student of bilingualism in the context of social issues and social
change, should benefit from an awareness of the various possible re-
lationships between individual bilingualism and societal diglossia il-
lustrated in this section. One of the fruits of such awareness will be
that problems of transition and dislocation will not be mistaken for
the entire gamut of societal bilingualism.

7.0 LANGUAGE MAINTENANCE AND LANGUAGE SHIFT

Modern history reveals at least five major instances of language shift,
i.e., where huge populations adopted a new language or variety into
their repertoires, whether or not at the same time they also gave up
a language or variety that they had previously used. The instances
referred to are (a) the vernacularization of European governmental,
technical, educational, cultural activity, (b) the Anglification/Hispani-
zation of the populations of North/South America respectively (Table
10), (c) the adoption of English and French as languages of elite for wider
communication throughout much of the world, but particularly so in
Africa and Asia, (d) the Russification of Soviet-controlled populations,
and most recently (e) the growing displacement of imported languages
of wider communication and the parallel vernacularization of govern-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
300 Joshua A. Fishman
table 10
1940-1960 Totals for 23 Non-English Mother Tongues in the USA
(Fishman 1966c)
Tot^l (Change
Language 1940 Total 1960 Total
n %
Norwegian 658,220 321,774 —336,446 —51.1%
Swedish 830,900 415,597 —415,303 —50.0%
Danish 226,740 147,619 —79,121 —65.1%
Dutch/Flemish 289,580 321,613 + 32,033 + 11.1%
French 1,412,060 1,043,220 —368,840 —26.1%
German 4,949,780 3,145,772 —1,804,008 —36.4%
Polish 2,416,320 2,184,936 —231,384 —9.6%
Czech 520,440 217,771 —302,669 —58.2%
Slovak 484,360 260,000 —224,360 —46.3%
Hungarian 453,000 404,114 —48,886 —10.8%
Serbo-Croatian 153,080 184,094 + 31,014 +20.3%
Slovenian 178,640 67,108 —111,532 —62.4%
Russian 585,080 460,834 —124,246 —21.2%
Ukrainian 83,600 252,974 + 169,374 + 202.6%
Lithuanian 272,680 206,043 —66,637 —24.4%
Finnish 230,420 110,168 —120,252 —52.2%
Rumanian 65,520 58,019 —7,501 —11.4%
Yiddish 1,751,100 964,605 —786,495 —44.9%
Greek 273,520 292,031 + 18,511 +6.8%
Italian 3,766,820 3,673,141 —93,679 —2.5%
Spanish 1,861,400 3,335,961 + 1,474,561 +79.2%
Portuguese 215,660 181,109 —34,551 —16.0%
Arabic 107,420 103,908 —3,512 —3.3%
Total 21,786,540 18,352,351 —3,434,189 —15.8%

In 1940 the numerically strongest mother tongues in the United States

were German, Italian, Polish, Spanish, Yiddish, and French, in that order. Each
of these languages was claimed by approximately a million and a half or more
individuals. In 1960 these same languages remaind the 'big six' although their
order had changed to Italian, Spanish, German, Polish, French, and Yiddish.
Among them, only Spanish registered gains (and substantial gains at that) in this
20-year interval. The losses among the 'big six' varied from a low of 2.5% for
Italian to a high of 44.9% for Yiddish. The only other languages to gain in
overall number of claimants during this period (disregarding the generational
distribution of such gains) were Ukrainian, Serbo-Croatian, 'Dutch'/Flemish, and
Greek. The greatest gain of all was that of Ukrainian (202.6%!). Most mother
tongues, including five of the 'big six', suffered substantial losses during this
period, the sharpest being that of Danish (65.1%). All in all, the 23 non-English
mother tongues for which a 1940-1960 comparison is possible lost approximately
one-sixth of their claimants during this interval. Yet the total number of claim-
ants of non-English mother tongues in the United States is still quite substantial,
encompassing nearly 11% of the total 1960 population (and an appreciably
higher proportion of the white population).»

a The 1940 and 160 totals shown in Table 10 must not be taken as the totals
for all non-English mother tongue claimants in those years. Figures for Armenian

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 301
mental, technical, educational and cultural efforts in many parts of
Africa and Asia. Having previously noted (section 5) that divergence
and differentiation of the verbal repertoire are reflections of societal
distance and segmentation, we must now point out that the socio-
cultural changes that carry with them changes in verbal repertoires are
themselves differentially associated with the various speech communi-
ties and speech networks of any polity. As a result, not only are the
verbal repertoires of communities and networks that experience the
greatest socio-cultural change the most likely to be altered, but the re-
pertoires of those who gain most in economic, political or other socio-
cultural status are the most likely to be adopted or copied by others who
see opportunities for desirable changes in their own status by so doing.
The study of language maintenance and language shift focuses upon
cell 2 of Figure 13 above and is basically concerned with the relation-
ship between degree of change (or degree of stability) in language usage
patterns, on the one hand, and ongoing psychological, cultural or social
processes, on the other hand, in populations that utilize more than one
speech variety for intra-group or for inter-group purposes. That lan-
guages (or language varieties) sometimes displace each other, among
some speakers, particularly in certain interpersonal or system-wide
interactions, has long aroused curiosity and comment. However, it is
only in quite recent years that this topic has been recognized as a field
of systematic inquiry among professional students of language behavior.
It is suggested here that the three major topical subdivisions of this
field are: (a) habitual language use at more than one point in time or
space; (b) antecedent, concurrent or consequent psychological, social
and cultural processes and their relationship to stability or change in
habitual language use; and (c) behavior toward language, including
directed maintenance or shift efforts. It is the purpose of this section
to discuss each of these three topical subdivision briefly, to indicate
their current stage of development, and to offer suggestions for their
further development.

7.1 Habitual Language Use at More Than One Point in Time

The basic datum of the study of language maintenance and language
shift is that some demonstrable change has occurred in the pattern of

were reported in 1940 but not in 1960. Figures for Chinese and Japanese were
reported in 1960 but not in 1940. Total figures for "All other" languages were
reported in both years. None of these inconsistent or non-specific listings are in-
cluded in Table 10. Adding in these figures, as well as the neccessary genera-
tional estimates based upon them, the two totals would become 1940: 22,036,240;
1960: 19,381,786.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
302 Joshua A. Fishman
habitual language use. The consequences that are of primary concern
to the student of language maintenance and language shift are not
interference phenomena per se but, rather, degrees of maintenance or
displacement in conjunction with several sources and domains of
variance in language behavior. Thus, the very first requirement of in-
quiry in this field is a conceptualization of variance in language beha-
vior whereby language maintenance and language displacement can be
accurately and appropriately ascertained. In the course of their labors
linguists, psychologists, anthropologists and other specialists have deve-
loped a large number of quantitative and qualitative characterizations of
variance in language behavior. By choosing from among them and
adding to them judiciously, it may be possible to arrive at provocative
insights into more sociolinguistic concerns as well. Whether those as-
pects of variance in language behavior that have, in the past, been con-
ceived of as qualitative, can be rendered ultimately commensurable with
those that have more frequently been considered quantitative is a topic
to which we will return, after first considering the two aspects separa-
tely.

7.11 Degree of Bilingualism

For the student of language maintenance and language shift the quan-
tification of habitual language use is related to the much older question
of ascertaining degree of bilingualism. This question, in turn, has been
tackled by a great number of investigators from different disciplines,
each being concerned with a somewhat different nuance. Linguists have
been most concerned with the analysis of bilingualism from the point
of view of switching or interference. The measures that they have pro-
posed from their disciplinary point of departure distinguish between
phonetic, lexical and grammatical proficiency and intactness (Mackey
1962). At the other extreme stand educators who are concerned with
bilingualism in terms of total performance contrasts in very complex
contexts such as the school or even the society (Manuel 1963). Psycho-
logists have usually studied degrees of bilingualism in terms of speed,
automaticity, or habit strength (Macnamara 1966). Sociologists have
relied upon relative frequencies of use in different settings (Hayden
1964, Hofman 1966a, 1966b, Nahirny and Fishman 1966). Thus, since
a great number of different kinds of bilingualism scores or quotients
are already available, the sociolinguistically oriented student of lan-
guage maintenance and language shift must decide which, if any, are
appropriate to his own concerns. Since the study of this topic cannot be
reduced to or equated with the concerns of any particular discipline it
seems highly likely that a combination or organization of approaches

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 303
to the measurement and description of bilingualism will uniquely cha-
racterize the study of language maintenance and language shift.

7.12 The Need for a Combination of Interrelated Measures

It would seem that the linguist's interest in itemizing examples of inter-

ference and switching introduces an outside criterion into the study of
language maintenance and language shift which may not at all corres-
pond to that utilized by speech communities or speech networks under
study. The linguist's distinction between what is English and what is
French and the distinction made by English-French bilinguals may
differ so widely that the linguist's conclusions about the drift of shift,
based upon interference and switch data, may be seriously in error.
However, even where a linguist is obviously interested only in a
carefully delimited question about the relative frequency of a particular
instance or class of interferences or shifts, it is clear that it may be far
easier to answer this question in some cases than in others (e.g., it may
be easier to answer in connection with encoding than in connection with
inner speech; it may be easier to answer in connection with writing than
in connection with speaking; it may be easier to answer in connection
with formal and technical communication than in connection with inti-
mate communication), for the 'density', stability and clarity of inter-
ference and switching varies for the same individual from occasion to
occasion and from situation to situation. Although interference and
switching are lawful behaviors, there are advanced cases of language
shift in which even linguists will be hard pressed to determine the
answer to "which language is being used?", particularly if a single
supra-level answer is required.
Similarly, concern with relative proficiency, relative ease and automa-
ticity, and relative frequency of language use in a contact setting are
also not necessarily indicative of overall language maintenance or shift.
Conclusions based on such measures may be particularly far off the
mark in bilingualism-plus-diglossia settings in which most speakers use
both languages equally well (correctly), effortlessly and frequently but
differ primarily in connection with the topics, persons, and places (or,
more generally, the situations and situation types or domains) in which
these language are used. Thus, in conclusion, the contribution that the
student of language maintenance and language shift can make to the
measurement of bilingualism, is precisely his awareness (a) that various
measures are needed if the social realities of multilingual settings are
to be reflected and (b) that the measures can be organized in terms of
relatively general variance considerations. Of the many approaches to
variance in language use that have been suggested the following is both

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
304 Joshua A. Fishman
simple enough for easy presentation as well as sufficiently involved to
imply that even greater complexity exists not too far below the surface.

7.13 Media Variance: Written, Read and Spoken Language

Degree of maintenance and shift may be quite different in these very
different media. Where literacy has been attained prior to interaction
with an 'other tongue', reading and writing in the mother tongue may
resist shift longer than speaking. Where literacy is attained subsequent
to (or as a result of) such interaction the reverse may hold true (Fish-
man 1965e). More generally, the linguist's disinclination to be con-
cerned with the written language is a luxury that cannot be afforded in
the study of language maintenance and language shift, where the con-
trasts involved are so frequently between languages that vary greatly in
the extent to which they have literacy or other 'higher' functions for
the speech networks under study.

7.14 Overtness Variance

Degree of maintenance and shift may be quite different in connection
with inner speech (in which ego is both source and target), comprehen-
sion (decoding, in which ego is target), and production (encoding, in
which ego is the source). Where language shift is unconscious or
resisted, inner speech may be most resistant to interference, switching
and disuse of the mother tongue. Where language shift is conscious and
desired, this may less frequently be the case (Fishman 1965f).
Location of Bilingualism: The Domains of Language Behavior. The
qualitative aspects of bilingualism are most easily illustrated in connec-
tion with the location of language maintenance and language shift in
terms of domains of language behavior. What is of concern to us here
is the most parsimonious and fruitful designation of the societally or
institutionally clusterable occasions in which one language (variant,
dialect, style, etc.) is habitually employed and normatively expected
rather than (or in addition to) another.
7.141 The Domains of Language Behavior and the Compound-
Coordinate Distinction. If the concept of domains of language behavior
proves to be as fruitful and as manageable a one as seems to be likely
on the basis of recent empirical evidence it may also yield beneficial
results in connection with other areas of research on bilingualism, e.g.,
in connection with the distinction between coordinate and compound
bilingualism (Ervin and Osgood 1954, p. 140). The latter distinction
arose out of an awareness (mentioned by several investigators over the
years) that there are "at least two major types of bilingual functioning",

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 305
one (the compound type) being "characteristic of bilingualism acquired
by a child who grows up in a home where two languages are spoken
more or less interchangeably by the same people and in the same
situations" and the other (the coordinate) being "typical of the 'true'
bilingual, who has learned to speak one language with his parents, for
example, and the other language in school and at work. The total
situations, both external and emotional, and the total behaviors
occurring when one language is being used will differ from those
occurring with the other." From our previous discussion of domains of
language behavior it is clear these two types of bilingual functioning
(more accurately put, two extremes of a continuum of psycho-neurolo-
gical organization) have been distinguished on the bases of some
awareness, however rudimentary, that bilinguals vary with respect to
the number and overlap of domains in which they habitually employ
each of their languages. However, this is true not only initially, in the
acquisition of bilingualism (with which the compound-coordinate
distinction is primarily concerned) but also subsequently, throughout
life. Initially coordinate bilinguals may become exposed to widespread
bilingualism in which both languages are used rather freely over a larger
set of overlapping domains. Similarly, compound bilinguals may
become exposed to a more restrictive or dichotomized environment in
which each language is assigned to very specific and non-overlapping
domains.
Going one step further it appears that the domain concept may facili-
tate a number of worthwhile contributions to the understanding of the
compound-coordinate distinction in conjunction with language main-
tenance and language shift per se. Thus, domain analysis may help orga-
nize and clarify the previously unstructured awareness that language
maintenance and language shift proceed quite unevenly across the several
sources and domains of variance in habitual language use. Certain do-
mains may well appear to be more maintenance-prone than others (e.g.,
the family domain in comparison to the occupational domain) across all
multilingual settings charecterized by urbanization and economic develop-
ment, regardless of whether immigrant-host or co-indigeneous populations
are involved. Under the impact of these same socio-cultural processes
other domains (e.g., religion) may be found to be strongly maintenance
oriented during the early stages of interaction and strongly shift oriented
once an authoritative decision is reached that their organizational base
can be better secured via shift. Certain interactions between domains
and other sources of variance may remain protective of contextually
'disadvantaged' languages (e.g., family domain: internal speech, hus-
band-wife role relations), even when language shift has advanced so far
that a given domain as such has been engulfed. On the other hand,

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
306 Joshua A. Fishman
if a strict domain separation becoms institutionalized such that each
language is associated with a number of important but distinct domains,
bilingualism may well become both universal and stabilized even though
an entire population consists of bilinguals interacting with other bilin-
guals. Finally, in conjunction with language maintenance and language
shift among American immigrant groups, the interaction between do-
main analysis and the compound-coordinate distinction may prove to
be particularly edifying.
As suggested by Figure 16, most late 19th and early 20th century
immigrants to America from Eastern and Southern Europe began as
compound bilinguals, with English assigned to quite specific and restrict-
ed domains. With the passage of time (involving increased interaction

BILINGUAL DOMAIN OVERLAP TYPE

FUNCTIONING
TYPE Overlapping Domains Non-Overlapping Domains

Compound 2. Second Stage: More 1. Initial Stage: The

('Interdependent' immigrants know more immigrant learns English
or fused) English and therefore can via his mother tongue.
speak to each other either English is used only in
in mother tongue or in Eng- those few domains (work
lish (still mediated by the sphere, governmental
mother tongue) in several sphere) in which mother
domains of behavior. In- tongue cannot be used.
creased interference. Minimal interference.
Only a few immigrants
know a little English.

Coordinate 3. Third Stage: The 4. Fourth Stage: English

('Independent') languages function inde- has displaced the mother
pendently of each other. tongue from all but the
The number of bilinguals most private or restric-
is at its maximum. ted domains. Inter-
Domain overlap it at its ference declines. In
maximum. The second most cases both languages
generation during child- function independently;
hood. Stabilized inter- in others the mother
ference. tongue is mediated by
English (reverse direc-
tion of Stage 1, but
same type).

Figure 16. Type of Billingual Functioning and Domain Overlap

During Successive Stages of Immigrant Acculturation

with English-speaking Americans, social mobility, and acculturation

with respect to other-than-language behaviors as well) their bilingualism
became characterized, first, by far greater domain overlap (and by far

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 307
greater interference) and then by progressively greater coordinate func-
tioning. Finally, language displacement advanced so far that the mother
tongue remained only in a few restricted and non-overlapping domains.
Indeed, in some cases, compound bilingualism once more became the
rule, except that the ethnic mother tongue came to be utilized via
English (rather than vice-versa, as was the case in early immigrant
days). Thus the domain concept may help place the compound-coor-
dinate distinction in socio-cultural perspective, in much the same way
as it may well serve the entire area of language maintenance and lan-
guage shift.
7.142 The Dominance Configuration. Section 7.141 above clearly
indicates the need for basic tools of a complex and sophisticated sort.
Precise measurement of degree of maintenance or displacement will be
possible only when more diversified measures of degree of bilingualism
(including attention to media and overtness variance) are at hand.
Precise measurement of domains of maintenance or displacement will
be possible only after concerted attention is given to the construction
of instruments that are based upon a careful consideration of the
various domains of language behavior (and the role-relations, topics and
locales - these being the three components of situational variation)
mentioned in a scattered international literature. The availability of
such instruments will also facilitate work in several related fields of
study, such as the success of intensive second-language learning pro-
grams, accurate current language facility censuses, applied 'language
reinforcement' efforts, etc. Given such instruments, the inter-corre-
lations between the several components of variance in degree of bilin-
gualism will become amenable to study, as will the variation of such
inter-correlations with age or with varying degrees of language ability,
opportunity and motivation. The relationship between maintenance or
displacement in the various domains of language will also become sub-
ject to scrutiny. Speculation concerning the relationship between shifts
in degree and direction of bilingualism and shifts in the domains of
bilingualism will finally become subject to investigation. Finally, out of
all the foregoing, it will become possible to speak much more meaning-
fully about the dominance configurations of bilinguals and of changes
in these configurations in language maintenance - language shift con-
texts.
7.143 Some Preliminary Suggestions. Figures 17 and 18 are pri-
marily intended to serve as possible presentation formats for dominance
configurations based upon several domains and sources of variance in
language behavior mentioned earlier in this discussion. The types of
language use data favored by linguists, psychologists and educators
have been set aside temporarily in favor of grosser 'frequency use' data.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
308 Joshua A. Fishman

Source of Variance

Family Neighb. Work Jew. Rei/Cult

role-rels. role-rels. role-rels. role-rels.
Media Overtness 1 2 3 1 2 1 2 3 1 2

Speaking Production
Comprehension
Inner

Reading Production
Comprehension

Writing Production
Comprehension

Figure 17. Intragroup Yiddish-English Maintenance and Shift in the United States:
1940-1970 Summary Comparisons for Immigrant Generation 'Secularists' Arriving
Prior to World War I ('Dummy Table' for Dominance Configuration).

Media Overtness Domains Role-Relations Summary-Ratings

1940 1970

Speaking Production Family Husband-Wife Y Y

Parent-Child
Grandparent- Y E
Grandchild
Other: same E
generation
Other: younger Y Y
generation
E E
Neigh- Friends Y E
borhood Acquaintances Y E

Work Employer-Employer E E
Employer-Employee E E
Employee-Employee E E

Jewish Supporter-Writer,
Rel./Cult Teacher, etc. Y Y
Supporter-Supporter Y Y

Figure 18. Part of 'Dummy Table' in Greater Detail

However, of primary interest at this time are the suggested parameters

rather than the rough data presented. An inspection of these figures
reveals several general characteristics of the dominance configuration:

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 309
(a) the dominance configuration summarizes multilingual language use
data for a particular population studied at two points in time and space;
(b) a complete cross-tabulation of all theoretically possible sources
and domains of variance in language behavior do not actually obtain.
In some instances, logical difficulties arise. In others, occurrences are
logically possible but either necessarily rare or rare for the particular
populations under study; (c) each cell in the dominance configuration
summarizes detailed process data pertaining to the particular role-
relations (parent-child, teacher-pupil, etc.) pertinent to it and the
situations, network types (open and closed) and/or transaction types
(interactional and personal) encountered; (d) some of the domains
utilized do not correspond to those listed in section 7.2 above, nor are
all of the domains previously listed utilized here. This should sensitize
us further to the probability that no invariant set of domains can prove
to be maximally revealing, notwithstanding the efforts expended in
pursuit of such a set (Dohrenwend and Smith 1962, Jones and Lambert
1959, Mackey 1962, Schermerhorn 1964); (e) an exhaustive analysis of
the data of dominance configurations may well require sophisticated
pattern analysis or other mathematical techniques which do not neces-
sarily assume equal weight and simple additivity for each entry in each
cell; (f) a much more refined presentation of language maintenance or
language shift becomes possible than that which is provided by means
of mother tongue census statistics (Kloss 1929, Nelson 1947). Word
naming scores, self-ratings of frequency of usage, observed occurrences
of various phonological, lexical, or gramatical realizations, all of these
and many other types of scores or indices can be utilized for domi-
nance configuration analysis of speech communities or networks. The
need to summarize and group language usage data necessarily leads to
some loss of refinement when proceeding from specific instances of
actual speech in face to face interaction to grouped or categorized data.
However, such summarization or simplification is an inevitable aspect
of the scientific process of discovering meaning in continuous multivar-
iate data by attending to differential relationships, central tendencies,
relative variabilities and other similar characterizations. Moreover, the
ultimate 'summary' nature of the dominance configuration and the
further possibilities of collapsing domains according to higher order
psychological or sociological similarities (e.g., 'public' vs. 'private' lan-
guage use) obviates the proliferation of atomized findings.

All in all, the dominance configuration represents a great and diffi-

cult challenge to students of bilingualism and of language maintenance
or language shift. It is possible that once this challenge is recognized,
serious problems of configurational analysis will also arise, as they have
in other substantive areas requiring attention to patterns of quantitative

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
310 Joshua A. Fishman
or qualitative measures. However, it is unnecessary to prejudge this
matter. It does seem fitting to conclude that the dominance configura-
tion - if it is to have maximal analytic value - might best be limited to
those aspects of degree of bilingualism and of location of bilingualism
which further inquiry may reveal to be of greatest relative importance
and independence. Focused attention on the study of spoken production
(as initially suggested by Table 11) has amply demonstrated the rich
yield that a self-imposed limitation of this kind can produce in appro-
priately selected speech communities (Fishman, Cooper, Ma et al.
1968).

7.2 Psychological, Social and Cultural Processes Related to Stability

or Change in Habitual Language Use
The second major topical subdivision of the study of language mainte-
nance and language shift deals with the psychological, social and cul-
tural processes associated with habitual language use. Under certain
conditions of interaction the relative incidence and configuration of
bilingualism stabilizes and remains fairly constant over time within
various bilingual-diglossic speech communities. However, under other
circumstances one variety or another may continue to gain speakers to
the end that bilingualism initially increases and then decreases as the
variety in question becomes the predominant language of the old and
the mother tongue of the young. The second subdivision of the study
of language maintenance and language shift seeks to determine the
processes that distinguish between such obviously different conditions
of interaction as well as processes whereby the one condition is trans-
formed into the other. The processes pertaining to this topical subdi-
vision may be conceived of either as antecedent, concurrent (contex-
tual), or consequent variables, depending on the design of particular
studies. Their major common characteristic is that they are primarily
outside of language per se.

7.21 The Paucity of Cross-Cultural and Diachronic Regularities.

Just as an understanding of social-behavior-through-language must de-
pend upon a general theory of society so the understanding of language
maintenance or language shift must depend on a theory of socio-culture
contact and socio-cultural change. Furthermore, it would seem that
since we are concerned with the possibility of stability or change in lan-
guage behavior on the one hand, we must be equally concerned with
all of the forces contributing to stability or to change in societal beha-
vior more generally, on the other. Thus the selection of psychological,
social and cultural variables for the study of language maintenance and
language shift may well be guided not only by impressions of what seem

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
 The Sociology of Language 311
t- o o •<t m •o
a
rn 1 vc VO ri C9

3
E £
„ 1
m
o _
O
k, ^ o o W)
. c 1 (S
^
o >/-! V
mO r^
n
r^
m 2«3Ji*
rft,
1- 3
O"z 1 vc 00 O d H-S
a CM . nO
<A
<u ^

^ o _ o cn Tf O 3 e*
SO >
-c 00 o
oo\ o
x-i (N m or> •5 .H
'2 (S
<S is
CT\
<z \o Tl" »-H o t •O a)
r-i •<t (S C k"00
o
<50 o fn o o a ^
o. s ^ v-i >ri
c VO 00 vi o .2 °
3 iE u> SO oo 00 eU «
O vo (S »o 00 r- b
< Z z u9
C 23
cs r^ oo •o o
K ^ VO rn 9
vi e C-4 ts (S 1 1 °02'S 3
i- D
3 ^
tJ
«Ifl V
<•3 u ft. u- w-i o 1 1
-<3 vo
^ V
Os •g*
fn oo o o
O ed^ 00 o\ >r<i o
c
« m ^
-a w-i <S vo r- ••I- Pi a
S
<c-s: 4> o,
„
^ **
s <n t-i
o B
o
s £ O r- OS t-;
o
>ri o <ri 00 r- 0\
u a
on G o>
— CT\

J «—i 00 •Sjs
tt •8 < Z £ m JS ©
m m fS 1-1 1 §
< > c 3
M 1
H s 3 O r- >3
o iS ^
O £s i a« o m m oo ri
(N
Tf 2 ••§
o
1-. 3
**

•S tj Uh cr C\ f^ o\ r» o n _ « 5O
<3 cs t-H q,
«3
O o OO 2
'O
3 a
"s o i-H as 1 1 I 1 C>
ft. O
V
3 J)
<H> << VD w-1 v-i 1 1 1
c til 1
C
« ft
o
tx 00 ra o t^ 00
6s- I-* r4 so m 19 >,
s u> W| 00 sc 00
•a i-H i-H O vi rl-
0 < Z > rj r)
fS n
§ o\ m SO t-
vP «
o ci 00° c-4 >r, vo
i c ts ts ^

ft- & VO o TJ- <N i

s S 73
<u
3 4-t C
s >>
O
VO 00<
„
VD oo
o
o
^
vi
fc, O ca
6£ Z
fS 1—
vc r- >0 n
^

fi n
»
™ ii
s
3
a
u •a a .5
c
0
&
Os
^
O ••j-
CS &
-
•a a
1s 2
a
£
ft,so
u t. •o
~OT3 c
C .2 u
S «>
•e-gg
¡g
o
c
if
s l
o G m O *

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
312 Joshua A. Fishman
to be the most relevant processes in a particular contact situation but
also by more general theories of personal, social, and cultural change.
This is not to imply that all forces leading to change in other-than-
language behaviors necessarily also lead to language shift. Indeed,
whether or not this is the case (or, put more precisely, a determination
of the circumstances under which language and non-language behaviors
change concurrently, consecutively or independently) constitutes one
of the major intellectual challenges currently facing this field of inquiry.
If this challenge is to be met, it will be necessary for the study of
language maintenance and language shift to be conducted within the
context of studies of intergroup contacts that attend to important other-
than-language processes as well: urbanization (ruralization), industriali-
zation (or its abandonment), nationalism (or de-ethnization), nativism
(or cosmopolitanization), religious revitalization (or secularization), etc.
Our current state of generalizeable knowledge in the area of language
maintenance and language shift is insufficient for the positing of rela-
tionships of cross-cultural or diachronic validity. Indeed, many of the
most popularly cited factors purportedly influencing maintenance and
shift have actually been found to 'cut both ways' in different contexts
or to have no general significance when viewed in broader perspective.
Thus, Kloss illustrates that no uniform consequences for language main-
tenance or language shift are derivable from (a) absence or presence
of higher education in the mother tongue, (b) larger or smaller num-
bers of speakers, (c) greater or lesser between-group similarity, and (d)
positive or hostile attitudes of the majority toward the minority (Kloss
1966b, pp. 9-13). The presence of so many ambivalent factors is a
clear indication that complex interactions between partially contribu-
tory factors (rather than a single overpowering factor) must frequently
be involved and that a typology of contact situations (as well as a theory
of socio-cultural change) may be required before greater regularity
among such factors can be recognized.
Although debunking represents a rather primitive level of scientific
development it may be a necessary stage on the path to greater matu-
rity. Although we cannot currently formulate universally applicable
regularities in our area of inquiry we can indicate that several attempts
along these lines fall somewhat short of their mark:
7.211 A Few Questionable Generalizations.
7.2111 Language Maintenance is a Function of Intactness of Group
Membership or Group Loyalty, Particularly of Such Ideologized Ex-
pressions of Group Loyalty as Nationalism. Among the evidence
pointing to the need for refining or justifying this view is that which
reveals that the Guayqueries of Venezuela preserved their groupness by
preserving their property relations while giving up their language and

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 313
religion (Hohenthal and McCorkle 1955), that lower caste groups in
India pursue Sanskritization (emulation) rather than solidarity as a
means of group mobility, that "the Raetoromans, like the Italian
Swiss, cultivate the fullest possible loyalty to their language with-
out aspiring to such nationalistic goals as political independence"
(Weinreich 1953a, p. 100), that the 'Yiddishist' movement in Eastern
Europe before and after World War I similarly concentrated on a lan-
guage program rather than on political organization (Weinreich 1953,
p. 100), that second and third generation Americans frequently main-
tain "cultural (refinement) bilingualism" after ethnic group loyalty dis-
appears at any functional level and, vice versa, that vestiges of behavio-
ral ethnicity often remain generations after language facility has been
lost (Fishman and Nahirny 1964); that many Auslandsdeutsche main-
tained their self identification as Germans in the midst of Polish or
Ukrainian majorities, long after completely giving up their German
mother tongue (Kuhn 1930, 1934); that Language loyalty is low in
many newly developing and highly nationalistic African states (Bros-
nahan 1963b, Spencer 1963), etc. Thus, it would seem, on the one hand,
that language maintenance has continued under various and highly dif-
ferent forms of group membership, some of which have involved signifi-
cant changes in traditional social relationships and in pre-established
role-relations. On the other hand, it appears that group loyalty can
be similarly (if not more) ubiquitous, continuing both with and without
language maintenance. The American readiness to use language as an
index of acculturation may, in itself, be quite culture bound (Samora
and Dean 1956). Hymes' observation that "some languages do not en-
joy the status of a symbol crucial to group identity" (Hymes 1962, p.
30) and Weinreich's observation that "the connection (between lan-
guage maintenance and group maintenance) is thus at least flexible and
cannot be taken entirely for granted" (Weinreich 1953, p. 100) really
represent important intellectual challenges for the study of language
maintenance and language shift. We very much need a more refined
understanding of the circumstances under which behaviors toward lan-
guage and behaviors toward the group are related to each other in par-
ticular ways. We can recognize today that the pre-World War II views
of many German students of language maintenance and language shift
(as to whether language and language consciousness create — or are
derived from - race, peoplehood and consciousness of kind) were too
simplified and too colored by then current political considerations.
However, the fact remains that the relationship between language-
saliency and group-saliency is almost as speculative today as it was at
that time, although it seems clear that a language undergoing massive
displacement may be retained most fully by increasingly atypical and

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
314 Joshua A. Fishman
self-consciously mobilized populations as displacement progresses. Ne-
vertheless, it is also clear that ideologies normally mobilize only a
relatively younger, more active and, perhaps, more alienated or dislo-
cated segment of any large population. Language maintenance may
depend most on nationalist ideologies in populations whose lives have
otherwise been greatly dislocated and it may also depend least on such
ideologies in those populations that have best preserved their total social
context against the winds of change (Fishman, 1969d).
The nationalism of several African and Asian countries seems to
be much more characterized by nationism than by the nationalistic
elaboration of ethnicity per se. It is much more concerned with the
instrumental political and economic conditions of nationhood than with
the socio-cultural content of peoplehood. The political and administra-
tive limits of new nations are now usually defined in advance of their
formation rather than in the process of their formation. The new nations
are less frequently formed as the result of the 'painful but glorious'
unification of hitherto particularistics who have groped to define the
language, the history, the customs, and the missions that unite them and
set them apart from others. They are formed along supra-ethnic lines
that normally follow colonial demarcations which depended on the
fortunes of conquest and the skills of treaty-making. Political and
economic self-determination are much more prominent considerations
in the new nations than is cultural self-determination of the European
pre- and post-World War I variety. Political leadership is much more
evident than cultural leadership. The Western experience has typically
been that industrialization preceded urbanization and (particularly in
Eastern Europe) that nationalism preceded nationism and that the
first set of phenomena preceded the second. In the new nations, the
reverse sequences seem to be more common, and these may be among
the major socio-cultural determinants de-emphasizing language issues in
connection with local or regional languages, on the one hand, and which
favor continued use of supra-regional and colonial languages on the
other. Indeed, it may be that language concerns are most noticeable
today where we find socio-cultural distinctions remaining (even after the
attainment of considerably more politico-operational integration than
has currently been attained in most new nations), particularly when
hitherto backward, exploited or disadvantaged groups begin to expe-
rience great and rapid economic and cultural development in their own
areas of primary population concentration (as, e.g., the French-Cana-
dians, Flemings, Jura-regionists, etc.). The displacement of Western
languages of wider communication in Africa and Asia is coming - parti-
cularly in connection with mass education and governmental operations
and services - and it is coming on socio-cultural integrative grounds,

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 315
but it is still just coming, rather than having arrived together with inde-
pendence.
7.2112 Urban Dwellers are More Inclined to Shift; Rural Dwellers
(More Conversative and More Isolated) are Less Inclined to Shift. This
is one of the most reasonable and best documented generalizations in
the study of language maintenance and language shift. Nevertheless, it
runs counter to the first mentioned generalization, above, in that
consciousness of ethnicity and the espousal of nationalism have been
primarily urban phenomena. Language revival movements, language
loyalty movements, and organized language maintenance efforts have
commonly originated and had their greatest impact in the cities. Intelli-
gentsia and middle class elements, both of which are almost exclusively
urban, have frequently been the prime movers of language maintenance
in those societies which possess both rural and urban populations.
Indeed, urban groups have been 'prime movers', organizers or mobi-
lizers more generally, that is in connection with other than language
matters as well as in connection with language behavior and behavior
toward language. Thus, whereas small rural groups may have been
more successful in establishing relatively selfcontained traditional inter-
action patterns and social structures, urban groups, exposed to inter-
action in more fragmented and specialized networks, may reveal more
conscious, organized and novel attempts to preserve or revive or change
their traditional language. The urban environment does facilitate
change. However, the direction of such change has not always favored
language shift at the expense of language maintenance. When it has
favored the one and when the other (and when urban-inspired lan-
guage shift has actually signified a return to a languishing ancestral
language), represents a further challenge to this field of study.
Discussions of rurality-urbanness in relation to language maintenance
have often unwittingly combined two related but importantly separate
factors: separation and concentration. Thus, rurality is often not so
much significant for language maintenance because of a higher relative
concentration of own-mother-tongue population as because rural popu-
lations can isolate themselves consciously — or are more isolated even
without particularly wanting to be - from differently speaking popu-
lations. Data from several countries illustrate this aspect of rurality. In
the United States in 1940 the 'second generation foreign white stock'
(that is native born individuals of foreign born parents) was regularly
more retentive of its ethnic mother tongues - regardless of whether this
stock was derived from less retentive old-immigrant (Scandinavian and
German) or from more retentive new-immigrant (Southern and Eastern
European) groups - if living in rural than if living in urban areas
(Haugen 1953, Table 12). Seemingly, at that time, it was more possible

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
316 Joshua A. Fishman

O O ri o o O O vo O O t-
o o vo oo
t-_ ° O I mm « o
»
»
¡Cí
3 VO 00
c f ->"

•o o o o o o oo O o oo
c O 00
o) vo vo
. « VO
"3 V-, «-1 vo o\
E 00 W-1 m ra en

O O I O O TT o o vo o o n O O 1/1
•o O oo , o <s • oo f s (X 00 o , O «í
c ve ¡ 00 OS VO 00
i r-_ «H S
cd oo" vo O L
"o o - 8 vo v i
cl, VO

00
vo
O
o
o
o
8
N
o
o
VO © rt VI O oC o r-" o •t" O o
Ov vo VO N OC l/l v i vg r-. CS >/"1 00 t-
s « N N °° ©
< 'T. Ö «^W
r- d >o vo vo VO
o\ i-H o oo
o ^ O o o
s
vo O VO 00
cd i- o VO
o t-
S
UH m
O
O cí

a\
8 o*-' >o
£
S í Ü 00 M OO o
«I °
o
00
o
°
£
O
^ ^O VO
^

m oo oo oo O vo
•t -t TT -H r» m

O O T- O O t- o o m O O (- o o -d- O O <n
m o k; OO Tt • n o - ;
X <=> cñ
Zm
M
s . a o j r-^ 5 rt JJ" °° 0 ° OV.Ç
TJ- VO T- c i «O
VO TT VO « vo e s

O O TT © O t O O o o t- o o O O OV
e
o O vo ' v o oo - j 00 v o -i c^ N rA VO Tf - ¡
•o OO 5 <*i 1 1 vo^es Ü <=> 2
u oo e s oo m 0\ 00 vo <s vo
£
CO
<r\ cí -q- vo VO oo
(M

>> o o O O O O «r> O O vo
cd » 2 6
VO 00 ^ VO O
~ « o . K O
^V O VO s s »
£
t-H o r^ Cí m •«i
O d m m <N
Z O «

• a. ° • Ä 0 •. oo
tìv O
C • a. ° • a, a • s. «
. 50 o
. oo o
: «
. oo o
: cd -¡3 . to ^ • £f
' s
o
2
• c8 ©
'S ' & S ^ t i
ä S ä S
o cm » S %a S o c 2 ' c í " e S
o 3 » © cd <o X
O ecd Ä« » 5 o cd » 2 « »
4-* 55 J ei
Vi J B¿ CO • J « S j « t/5 h4 P5 i i J Pá
13 « g c
cd u- C UH
cd -
•Q ? cd 3 x> 3 .cd
Pi
« è « à
O O

CO c ò ^ CO
D Ù

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 317

o o •>» S O N © O o o oo
52 n N 1- N oo o V N
CO ( S t- S 1i vo
00 <S1 <S •<t cs m
>/i v£>
o \ so

•O O S « O O 00 o o <s O O C- o o o* o O ,
C 00 f N ' oo
03 ° io o O vo
s a p oo vo I
"3 <s o \ VO Tt
£

O O f> o o o o oo O O CI O O OS
•O 00 ^ r-i <s o <^1 oo 2 ^ oo"
c 1 r- 3 <"¡.55
ov c> VO t-T VO m
"o VO o \
0-

0 O 0 0 O O
00 cs VO VO 00 00
(S 00 W-J^ ON (M 00,
•c o" O V—I O cr> 0 O 0" 0 m 0
r-; 00 t - VO O 00 T—< 00 •<t 0 »ri
00 v i m O n r - <N ^ ts
00 <0 ci r-
m
o\
VO
r- 0" fn
fS
cn Ti
0 0 O 0 r) 0 0
•>* r) O 00 tM C-)
00_ cn
ri ri
Wl VO
ri
rry <
O
u
0\ o\ VO CM

< <s
oN
0 0 0 O O 0 0 0 O 0 f- 0 0 O O O 0
\VO 3 <0 8 o\
<s"%5
VO 00 rrt 00
r » VO Ov O •<t VO

z
D ri cif S rf cn
d
a
i«
«D
O
o
,C
O O t- O O o o O O 00
0O 00 - J M vo o s s
^ o\ vo ~ a \ CM
CO •o rj r f r-i
T3
•a
e<S
CM 0 0 * 0 O O TT o o o\ o o « o o CM
oo N N ^
SS S 00
°S ^ CM
•a 00,0 ® °° 3 £
u o\ r j CM —
i 00 Tf CM H
1/5

>>
o o O O 00 O O CM o o oo o o o o
cs o o t ,VO - i >e u i o cm 00 t o , 00 VO
CT\ •VO
"„vo <N VO VO rt
S
u. <S <S Q m CM ON <s t - ov >
O HI CM VO Wl CM CO
Z r-- m •c
•a
§
• a. c . <u c & c . o e • Si, <3 • a, s
: ff-s : 8 P j
: 9 . 3
•. re
2P o
•o
« 1
S> 8 j: m B
! « a i i l iQ s4-» •s I I 1 1 1
3 CO V CO
a I _) Pi
2
M J
S «
Pi J Pi 2 CS u
CO - J P i ° a <0
2
CO h J
OS
Pi
« C/3

a a S h i Pi
o co
o X) c
!3
3
§ g
Pi
3

>-: § h
•8
H Z zz z

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
318 Joshua A. Fishman
to hand on more traditional ways of life, including the traditional
mother tongue, in rural areas, particularly in those that were populated
largely by others of the same language background. Such separation no
longer made much difference in the United States in 1960 (Fishman
1966c).
Similarly, non-rurality in India (as well as a more advanced level of
education which accompanies non-rurality) is positively related to
claiming English as a subsidiary language in contemporary India (Table
13), but it is negatively related to the claiming of Hindi as a subsidiary
TABLE 1 3
The Best Predictors of District Variation in English Claiming
(N = 129 districts)
(Das Gupta and Fishman, 1971)
Cumulative
Predictor r CumR CumR2 AR2 Far2
% Male Pri + Jr -,336b .336 .113 — —

% Male Matric + ,176a .497 .247 .134 22.3c

Rural pop/Total pop -.054 .649 .421 .174 37.8c
% Immigrants .038 .659 .434 .013 2.8
% Female Matric .057 .670 .448 .014 3.1
Crude Literacy -.067 .672 .452 .004 < 1
% Female Pri + Jr -.146 .678 .459 .007 1.6
Agricult/% Rural -.122 .679 .461 .002 < 1
Workers in retail .039 .679 .462 .001 < 1
Persons/sq. mile .056 .680 .463 .001 < 1
Workers in manuf. -.005 .681 .463 .000 0.0
Scheduled caste .021 .681 .464 .001 < 1

a = significant at .05 level

t> = significant at .01 level
c = significant at .001 level

language (Table 14). Seemingly, the acquisition of English depends on

institutions, higher schools, government bureaus, organizations and
media (newspapers, motion pictures) not readily available in the rural
areas. However, the acquisition of Hindi (in non-Hindi mother tongue
areas) depends more on lower schools, on radio broadcasts and on
federal governmental agricultural demonstration and assistance pro-
grams and these are available in rural areas. Thus, rurality in India
means well nigh full separation from English acquisition opportunities
and, therefore, a relative intensification of Hindi acquisition opportuni-
ties. Language shift is occurring in both settings, but in different direc-
tions as a result of the differential separations that rurality represents

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 319
TABLE 1 4
The Best Predictors of District Variation in Hindi Claiming
(N — 75 districts) (Das Gupta and Fishman, 1971)
Cumulative
Predictor r CumR CumR2 AR2 FAR2
% Male Pri + Jr. .425a .425 .181 — —

Crude Literacy -.167 .619 .384 .203 23.6c

% Female Matric + -.019 .635 .403 .019 2.2
% Male Matric + -.163 .680 .462 .059 7.7b
AgricuIt/% Rural ,303 b .719 .518 .056 8.0b
% Immigrants -.086 .736 .542 .024 3.6
% Female Pri + Jr. .055 .744 .553 .011 1.6
Rural Pop/Total Pop .030 .746 .556 .003 < 1
Persons/Sq.Mile -.120 .747 .558 .002 < 1
Scheduled caste .046 .748 .559 .001 < 1
Workers in Manuf. -.051 .752 .565 .006 < 1
Workers in Retail -.142 .753 .566 .001 < 1

a = significant at .05 level

b = significant at .01 level

« = significant at .001 level

for English and for Hindi (Das Gupta and Fishman, in press). Of
course, separation need not depend on rurality and can occur - although
less readily - in urban areas as well. Lieberson (in press) has shown
that "separating occupations" can serve language maintenance quite as
well as does the separation factor in rurality (Table 15).
The impact of population concentration, i.e., the proportion that
speakers of language X are of the total co-territorial population of a
particular administrative unit, is quite another matter from rurality per
se. Of course, rurality is related to population concentration in general
but as we have used it here, concentration is a proportional matter
rather than merely an absolute one. Once again, there is much evidence
that population concentration is important in language maintenance, but
this is true in urban rather than in rural settings. Thus, Lieberson (in
press) has shown that in cities in which the proportions of non-English-
speaking immigrants were higher in 1900 the proportion of second
generation Americans unable to speak English was also higher (Table
16). Sixty years later, those non-English mother tongues that were
numerically in the strongest position in the United States were exactly
those that constituted the highest relative proportions of the total popu-
lations of the states in which their claimants were concentrated (Table
17). Seemingly, a relatively large community of speakers is necessary,

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
320 Joshua A. Fishman
TABLE 1 5
Foreign Born White Males Unable to Speak English,
By Occupation, 1890 (Lieberson, 1971)
Occupation Per Cent Unable to

All 23

Agricultural Laborers 28
Miners (coal) 55
Stock Raisers, Herders 52
Professional Service 8
Dentists 4

Lawyers 2
Bartenders 6
Launderers 30
Auctioneers 4
Clerks and Copyists 6

Salesmen 5
Artificial Flower Makers 30
Brick and Tile Makers 46
Harness and Saddle Makers 10
Iron and Steel Workers 33

Printers, Lithographers 8
Tailors 29
Tobacco and Cigar Factory Operatives 44

Note: Persons born in England, Ireland, Scotland, and Canada (English) are
excluded since it is assumed that virtually all could speak English prior to
migration.

in many immigrant settings at least, in order for language maintenance

to be most useful as well as most likely in the increasingly urban context
with which it is faced. Under circumstances of high relative concen-
tration non-English schools, publications, broadcasting, organization
activity, and, above all, non-English family patterns can more readily be
maintained in interactional American urban environments. Thus, not
only is an inter-group diglossia fostered in urban centers with a high
relative concentration on non-English speakers, but in addition, intra-
group diglossia, in terms of the separate societal allocation of functions,
becomes more of a possibility. Soviet developments during the past few
decades also seem to reveal similar processes with respect to the co-
existence of Russian and the languages of at least the major Soviet
minorities (Table 18).

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 321
TABLE 16
Proportion Unable to Speak English in Cities,
Second Generation Cross-Tabulated by Foreign Born, 1900
(Lieberson, 1971)

Cities Classified by Mean Proportion Unable to Speak English Among

Proportion of Foreign
Born Unable to
Speak English Foreign Born Second Generation

.10+ .1957 .0065

.05 to .09 .0682 .0005
.04 or less .0267 .0003

Data based on 20% sample of cities with 25,000 or more population. 'Foreign
born' refers to Foreign Born Whites; 'Second Generation' refers to Native
Whites of Foreign Parentage.

7.2113 The More Prestigeful Language Displaces the Less Prestige-

ful Language. Our earlier discussions of sources of variance and do-
mains of language behavior may have prepared us for the realization
that language prestige is not a unit trait or tag that can be associated
with a given language under all circumstances. Indeed, our earlier dis-
cussions were necessary precisely because the prestige of languages can
vary noticeably from one context to another for the same interlocutors,
as well as from one speech network to another within the same speech
community. It is for this very reason that Weinreich recommends that
"as a technical term . . . 'prestige' had better be restricted to a language's
value in social advance" (Weinreich 1953a, p. 79). However, even this
limitation does not make the concept 'prestige' any more useful for
research purposes since social advance itself is relative to various
reference groups. Advance in family and neighborhood standing may
require a different language than advance in occupational or govern-
mental standing. The fact that an overall hierarchy of reference groups
may exist does not mean that the top-most reference group will be do-
minant in each face-to-face situation.
It may be precisely because 'prestige' obscures so many different
considerations and has been used with so many different connotations
that the relationship between prestige data and language maintenance or
language shift data has been more uneven than might otherwise be ex-
pected. Thus, whereas Hall claims that "it is hard to think of any
modern instance in which an entire speech community is under pressure
to learn a sub-standard variety of a second language" (Hall 1952, p.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
322 Joshua A. Fishman

5 »
o <u

'Ü 2
¿3 CO
tí « H
O O
e 5
o o
° Z
<J

•a
* —< T-t es r í ' - ' ^ - ' — ' ' — — rs es
id
Pi
c
cd
.o o o r - r ) s o « n r - - o e s © t - » ' ^ - v i s o — v s o o s ^ - e n r - - < t o \ o s Os
Lh ^ n m m O M O o o w N ^ i û O m N c o f t w O r n o n « os
o o o o o s o o o N t — o s O \ o s o s o s r - - o o o o o o o o o o o o a o o s t - - r ' - s o vo

M
C »-H SO m oo TT <-» i o «s r - «O Os oo e s Os e n o
cd f—1 i-H e s es es
Pi

'S g B NO «S Os t - 00 O CS o es o\ o r - oo e n Os v o
ri m en >o e n NO r- SO o e n t'- 00 00 e n SC
o 'S" r- m r- »—« C^Ì es o o
g§§
E 8 - S o
<S o o o o o o O o o o o s o
S § 2 O o o o o o o O o o o o o o O o
â o -

ri
5 ««
• 2 **
"5 «5
CO m
- h <s M
00 e n " O
O r i Tf
M r i oo
N N M <W W M l '—• * i
_ _
I
• o ÛC t^ O T
a » 0
00 O ^ CO O -«I- ^ " t " 1 " I ^ t . ^ ° NO
04*0 « r o 00 e*í
o S o \ S_ . o vo oo O M V U o e H N i ^ r - o - i M n ï i Q O h m n
Ph O e S ON •—I 1—< O e n ^ t"; os ^ N oo » N
— <u v C en Tf o vp-vg r t ^f m t^" oo t o m » h ( n vo" m O e s v T v i " oo
S "3 es «n m Ti- • ï •ta
s
o « or
H .S u

o a S
t-< . s
O j j ï r» r- en o 00 00 r - Os e s CJ es en r- VO O V I SO e s
e s w-i r » •<*rs r- o r- T—1 e s o s O e s Os r - O v> l'- m T-H Tf
o 55 ® 00 »-« r ^ en 00 < SO Os T Í T f ' o s . SO es es en so TT Os e s
fi S " vo
-t
r-* e s en O es o •/-T o o o C VO os so o vo "/"T vo t— es" e s
•Si S U 2 NO o s o s r - r- < o s •n W-l oo f » o 00 1 -s- SO >n es en en
a> - g - 3 TT s o v p en «—1 e s
t-i ed o
O r t »
fcUö
M

§ o 9
§
'3 JS
3 so cd
9 a 5 ** 3 o .o -g
o « « i l
U d i ¿ Pi
O ^ i n ^ h o o o i O H N

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 323
TABLE 1 8
Proportions of Russians in populations of Union Republics in 1926
and 1959, and percentage of migrant and non-migrant populations
using Russian as native language
(Lewis, 1971)
% of Russians in
% usine Russian 1959
population
Republic
1959
1926
Total Urban non-migrant migrant

Russia 78 83.0 87.2

Ukraine 9 16.9 29.9 12.0 23.0
Belorussia 8 8.2 19.4 15.0 28.0
Uzbekistan 6 13.5 33.4 0.3 12.6
Kazakhistan 20 42.7 57.6 1.2 4.3
Azerbaidjhan 10 13.6 24.9 1.2 9.3
Armenia 2 3.2 4.5 8.0 15.0
Georgia 4 10.1 18.8 0.4 8.0
Lithuania - 8.5 17.0 0.1 3.5
Moldavia 9 10.2 30.8 3.0 15.0
Latvia - 26.6 34.5 1.4 25.0
Tadzhikstan 5 13.3 35.3 0.5 18.0
Turkmenia 8 17.5 35.4 0.6 6.7
Estonia — 20.1 30.8 0.5 25.0
Kirgisia 12 30.1 51.8 0.3 16.0

Sources — (a) Figures for 1926 and 1959 are drawn from the respective Census
returns, (b) Volova, N.G. Voprosy Dvuyazychaya na Severnom Kaukaza, So-
vetskaya Etnografiya 1967, N o . 1, 27-40.

19), it is really not very hard to do so: A Low German dialect displaced
Lithuanian in East Prussia before World War I, although many Li-
thuanians there were highly conversant with Standard German (Gerul-
lis 1932). Unstandardized Schwyzertütsch is replacing Romansh, al-
though several generations of Raetoromans have known Standard Ger-
man as well (Weinreich 1951, pp. 284-286). Standard German complete-
ly displaced Danish in a trilingual area of Schleswig, but it was itself
then increasingly displaced by the local Low German dialect (Selk
1937). Obviously, Schwyzertütsch maintains itself quite successfully in
competition with Standard German; Landsmaal achieved considerable
success (into the 1930's, at the very least) in competition with Dano-
Norwegian; Yiddish won speakers and adherents among Russified, Po-
lonized and Germanized Jewish elites in Eastern Europe before and
after World War I; Castillian speaking workers settling in more in-
dustrialized Catalonia tend to shift to Catalan, etc. Indeed, the entire
process whereby a few classical languages were displaced by 'lowly'

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
324 Joshua A. Fishman
vernaculars and whereby some of the latter, in turn, were later displaced
by still other.and even 'less prestigeful' vernaculars (Deutsch 1942; the
latter varieties are still referred to as 'dialects' in many popular [as well
as in all too many socio-linguistically insensitive though scholarly] pu-
blications, e.g., Yiddish, Ukrainian, Byelo-Russian, Flemish, Afrikaans,
Macedonian, to mention only European derivatives) indicates that the
prestige notion is easily discredited unless serious qualifications and
contextual redefinitions are attempted. This too may be an appropriate
task for the study of language maintenance and language shift.
Quite clearly it is not some mystically invariant prestige of a language
or variety that need concern us, but, rather the highly variant fates and
fortunes of its speakers. The triumphs of English, Spanish (and Portu-
guese) in the New World is a triumph of physical might, of economic
control and of ideological power. None of these are language factors per
se, but languages that happen to be associated with such powerful forces
and developments can open up advantages to their speakers far beyond
those available to non-speakers of these languages. Under circumstan-
ces in which desired socio-cultural change follows from verbal reper-
toire change schools and media and organizations and programs
have no difficulty facilitating shift (as e.g. in Israel, See Figure 19).
Without such circumstances - and they are usually differentially avai-
lable to various population segments - neither better pedagogic
approaches nor more intense exhortation can have major impact on
language shift.

7.22 Toward More General Theory and a More Inclusive Comparative

Approach.
7.221 When bilingual speech networks are in touch with each
other on the one hand, as well as with monolingual speech net-
works on the other, they are differentially involved in the crucial socio-
cultural processes that influence or regulate their interaction. These
processes serve to increase or decrease interaction between populations
or sub-populations in question, to either detach them from or to
confirm them in their accustomed sources of authority, to either lead
them to influence others or to be particularly receptive to influence from
others, to either emphasize or minimize their own groupness and its
various manifestations, to either rise or fall in relative power or control
over their own and each other's welfare, to either view with positiveness
or negativeness the drift of the interaction between them and to react
toward this drift on the basis of such views. We must look to these
engulfing socio-cultural processes and, particularly, to indices of indi-
vidual and group involvement in them, in our efforts to explain the
direction or rate of language maintenance and language shift.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 325
Index of Hebrew Speaking in the Jewish Population (Aged 2 and over), by age at
Immigration and Year of Immigration.

Age at immigration

61 60 59 58 57 56 1951-55 1936-40
Year of immigration

Figure 19. Population and Housing, Census, 1961; Government of Israel,

Jerusalem.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
326 Joshua A. Fishman
7.222 However, after having appropriately selected and specified
one or more variables from among the endless subtleties that make up
the 'process' of socio-cultural change, it may still be found that their
cross-cultural and diachronic study reveals inconsistent results. The
'same' process (e.g., 'urbanization', as measured by constant indices
such as those selected and cross-culturally applied by Reissman 1964)
may result in language shift away from hitherto traditional languages in
some cases, in language shift back to traditional languages in other
cases, while revealing significantly unaltered maintenance of the status
quo in still others. Under such circumstances a typology of contact
situations might serve to control or regularize a number of group or
contextual characteristics, in the manner of moderator variables, and,
by so doing, reveal greater order in the data.
We all have an intuitive impression that the "American immigrant
case" is different from the "Brazilian immigrant case" (Willems 1943);
that the "Spanish conquest case" (Bright 1960; Dozier 1951) is different
from the "Anglo-American conquest case" (Cook 1943; Gulick 1958);
that the "immigrant case", in general, is different from the "conquest
case" in general; that the "Yiddish speaking immigrant to America
case" (Fishman 1965f) is different from "German speaking immigrant
to America case" (Kloss 1966b), etc. The question remains how best to
systematize these intuitive impressions, i.e., what variables or attributes
to utilize in order that contact situations might be classified in accord
with the differences between them that we sense to exist. In the terms
of R.A. Schermerhorn's recently formulated typology (1964) the "Ame-
rican immigrant case" immediately prior to World War I would be
characterized as revealing (i) sharply unequal power configurations
between non-English speaking immigrants and English-speaking "old-
Americans"; (ii) incorporation (rather than extrusion or colonization)
as the type of control exercised by American core society over the im-
migrants; (iii) marked plurality and recent immigration (rather than
duality, intermediate plurality without recent immigration, or any other
of a continuum of patterns) as the plurality pattern-, (iv) intermediate
stratification and substantial mobility within the stratification pattern;
(v) widespread mutual legitimization of acculturation and de-ethnization
as the interpretation of contact in philosophical or group image terms;
and (vi) growing industrialization, mass culture and social participation
as major social forces.
Given the above typological framework, it has proved possible to
summarize the current status of language maintenance and language
shift among pre-World War I immigrants in terms of a very few pre-
contact factors, host factors, and product factors. Unfortunately, Scher-
merhorn's typology for intergroup contacts is so recent that it has not

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 327
yet been widely tested on either practical or theoretical grounds, whether
in conjunction with language maintenance-language shift or in con-
junction with other topics in the area of intergroup relations. While it
may be expected that any typology based upon six parameters, each
with several subdivisions, is likely to be somewhat unwieldy and require
simplification, it is clear that Schermerhorn's system has at least heuris-
tic value for the sociology of language from Verdoodt's efforts to put
it to use in such fashion (in press).
At the opposite extreme of complexity from Schermerhorn's typo-
logy is one which is derivable from an intensive review of the extensive
literature on Auslandsdeutschtum (Kuhn 1934). One of the major dif-
ferentiations among the German settlers seems to have been the original
legitimization and concentration of their settlements. A three-way break
is recognizable here: Stammsiedlungen (settlements founded as a result
of official invitation and assistance from non-German governments),
Tochtersiedlungen (settlements founded by those who left the earlier
Stammsiedlungen and who settled elsewhere as groups, but without
governmental invitation or assistance), and Einsiedlungen (the in-
migration of German individuals or of small occupationally homoge-
neous groups into non-German communities). Another related distinc-
tion is that between the relative 'cultural development' of the settlers
and their hosts. During the decade before the Second World War the
two most frequently recognized co-occurrences were (a) Einsiedlungen
of 'culturally more mature' Germans living in the midst of a 'culturally
less developed' population, as opposed to (b) Stamm- and Tochtersied-
lungen of 'culturally younger' Germans surrounded by a 'more mature,
nation-oriented' population. Thus, although only two diagonal cells of
a theoretically complete two-by-two typology are extensively discussed
it is possible to find examples of the remaining cells as well. Even when
limited to the two co-occurrences mentioned above very interesting and
consistent differences appear both in rate and in stages of language
shift and acculturation. The implications of this rough typology and of
the regularities that it has suggested deserve consideration in connection
with quite different intergroup contact settings.
7.223 Although the study of language maintenance or language shift
need not be completely limited to the comparison of separate cases it
is nevertheless undeniably true that the comparative method is quite
central to inquiry within this topic area. Certainly the comparative me-
thod is indispensible in our pursuit of cross-cultural and diachronic re-
gularities. Assuming that a relatively uniform set of appropriate socio-
cultural process-measures could be selected and applied and assuming
that a recognizably superior typology of contact situations were
available it would then become possible to study:

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
328 Joshua A. Fishman

00

Total
1960
h x O T+ rt on VI CICT\r*-i ao«m«iOH(fi
a i n n 't f « m vo —• « N v; ON
VO

»a vooow-i
>o N
oo n o>-< rf ^ a\ « o
cs ^ «
^ t- o n w «
m

214 31
English

H
1960

rt «t o c-4 in Tf in o m »-h o O »n »-i r--

a VO 00 CS »—I

va oor^-oo ^ m« r, m
^ cs rt cs h—< oe ^H ^ N oo r^
m"

107 15
tj-
Mixed
1960

co m o\ •^•rtfn o rt vi ^ o "n r j -n -H M r -
a -h fs «

e-*
vo «
m THo\t>
<n r -
J-TJ-TJ-
rtoo-^-oo
r^ m oc
^ o c oooo m« in IN
r^ oc

377 54
Mother
Tongue
1960

ri«"-! v>
rt es
cn ^ rt cn voOrti^-oMov
-rt <N ^ ^

vi vo vi eh h inMrtVì M » n io o>
crt*< oo cs & a\ • < es «
—
1930
Total

a m c^ ^ co
CS »-H

vS> n f- N m o o MNOO O m N O r t - M
181 17
English

^ CO T-T TH R J rrt
1930

_ -H IN oo
S r-
csoo es es cs o o -H
o\ »n o «-» m I-H
'

VA r^cso\ to^t
WRTRT
oCO es o\' es »-ir^csvoocscs
RIRTHCS CS RT
Mixed
1930

_ vo o\ es RO^rt OO cs ^T rt
O es -H to —1m«1o h o w ON
•<T
1

VO O — CT\ I H o oo Vi oo o\ n vo o\ t^ vo
737 69

oo co vi cs o\ oo \oo\c~o\ r»*ovio\oovooo
Mother
Tongue
1930

aortvo O v> vo Vi vi Tt
0 c s v o ^< esci oo oo « ^ m » H .H -i t
Other Germanic
Other Romance
Ethnic Groups

Near Eastern
Scandinavian
Other Slavic

Far Eastern
Hungarian

All Others
Ukrainian
German

Total
Spanish
French

Jewish

Italian
Polish
Czech
Greek

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
 The Sociology of Language 329

oo oo so - 1 « O3 OS \o m vi »
so ¿3 so O f - 00 n N viH a
Os o © <N r- —i ts t O) HN
—H C4

ao rn N O N CI O *n <s o o >o r-
O t« 00 -H CS so os n
SO « oo TC O
r- so so r- m so O so C
r- cJ •<f —i ts ^H © ^H *~H (N
w m n

o
o
VO 4 t io
Os m »-<
o-g
¡e * vi TJ" SO m ^ t- •<t o\ >o t -
«I Tf so rj SO rt rt •<t -t <-> ->t cs
<M
e
43
os »n so ic n ^ vo <s so os t^ to
ON so *o Os so OS •>t cs oo so vs oo
VO OO OO —. os oo
Os © © \c O sf m f - h n
-H SO 00 t» ai t r- o\ so r- -h (N N n « n
»-I ' 3 2 H N N ts so csi N <1 <N
T3
C
<3

£ —• -a- r- m r-- m h fi © m oo © oo oo m
- i . §2 13
^ ® ©
m 2 t-
O ts
<n t m o s H
f- © wi
os oo m m m Os ©
os so < mm

8
S J 3
•c
3
© M £ o
m (S
o o
2 I OS m
1 °
S i 00 Tf o o SO o O m
ft. Os SO ( OS cs 1 ° I I I
fS s

3
£
m O n N
2 I i-H
© m vs Tf Os h« <S VD O N h
-H SO m vo oo ^H
C

o Ko Tf oo r- (S vo o m no i— < Tt >n oo o O h co
a- t-1 oo «o o O oo ON
o O o3 ^ ON ON SO r- o\ © 00 o\ t""-
43 So
° § oo m ^t fO O O N O 00 vp NO
r-» m H o\ r> H m 00 CS <*•> m ^ ON
N© Ox »n r- vn cn cs

o-
o
t- 5 a
o j S s
°b -a B la s? a S
™ 3co ^O» J3
O4) o
'e o ata Ph g "43 H
•B ta o b n
(L, fiO SfcOU
M

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
330 Joshua A. Fishman
(i) The same language group in two separate interaction contexts
that are judged to be highly similar (with respect to primary socio-cul-
tural process[es] and contact type), e.g., two separate German Stamm-
siedlungen in rural Poland.
(ii) The same language group in two separate interaction contexts
judged to be quite dissimilar (with respect to major socio-cultural pro-
cessives] and contact type), e.g., one German-Swiss community in contact
with Swiss Raetoromans and another German-Swiss community in Cin-
cinnati, Ohio.
(iii) Different language groups in two separate interaction contexts
judged to be highly similar (with respect to major socio-cultural pro-
cesses] and contact type), e.g., a Polish speaking and a Slovak speaking
community, both of rural origin, in Cincinnati, Ohio.
(iv) Different language groups in two separate interaction contexts
judged to be quite dissimilar (with respect to major socio-cultural
processfes] and contact type), e.g., a German Stammsiedlung in rural
Poland and a Slovak community in Cincinnati, Ohio.
Thus, by judiciously contrasting groups, socio-cultural processes and
types of contact situations (not necessarily taken two at a time, if higher
level interaction designs prove to be feasible) it should become possible
to more meaningfully apportion the variance in language maintenance
or language shift outcomes. Furthermore, the greater our insight with
respect to socio-cultural processes and the more appropriate our typo-
logy of intergroup contact situations, the more possible it becomes to
meaningfully assemble and analyze language maintenance and lan-
guage shift files. Such files would permit both cross-cultural and
diachronic analysis, of primary as well as of secondary data, based upon
comparable data, collected and organized in accord with uniform sets
of socio-cultural processes and contact categories. This state of affairs is
still far off but it is the goal toward which we might attempt to move
within this second topical subdivision of the study of language mainte-
nance and language shift, once more basic methodological and concep-
tual questions reaching a somewhat more advanced level of clarification.

7.3 Behavior Toward Language

The third (and final) major topical subdivision of the study of lan-
guage maintenance and language shift is concerned with behavior
toward language (rather than with language behavior or behavior
through language), particularly, with more focused and conscious beha-
viors on behalf on either maintenance or shift per se. Strictly speaking,
this subdivision may be properly considered a subtopic under 7.2,
above. However, it is of such central significance to this entire field of

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 331
inquiry that it may appropriately receive separate recognition. Three
major categories of behaviors toward language are discernible within
this topical subdivision:

7.31 Attitudinal-Affective Behaviors.

We know all too little about language oriented attitudes and emotions
(running the gamut from language loyalty — of which language
nationalism is only one expression - to language antipathy - of which
conscious language abandonment is only one expression) as distin-
guished from attitudes and emotions toward the 'typical' speakers of
particular language variants. The features of language that are con-
sidered attractive or unattractive, proper or improper, distinctive or
commonplace, have largely remained unstudied. However, in multilin-
gual settings, particularly in those in which a variety of 'social types' are
associated with each language that is in fairly widespread use, language
per se (rather than merely the customs, values and cultural contributions
of their model speakers) are reacted to as 'beautiful' or 'ugly', 'musical'
or 'harsh', 'rich' or 'poor', etc. Generally speaking, these are language
stereotypes (Fishman 1956). However, the absence or presence of a
'kernel of truth' (or of verifiability itself) is entirely unrelated to the
mobilizing power of such views.
The manifold possible relationships between language attitudes and
language use also remain largely unstudied at the present time. Al-
though Lambert reports a positive relationship between success in
school-based second language learning and favorable attitudes toward
the second language and its speakers (Lambert et al 1963), this finding
need not be paralleled in all natural multilingual contact settings. Thus,
Ruth Johnston reports a very low correlation between subjective and
objective (external) assimilation in the language area (1963b). Many
older Polish immigrants in Australia identified strongly with English,
although they hardly spoke or understood it several years after their
resettlement. On the other hand, many young immigrants spoke English
faultlessly and yet identified strongly with Polish, although they spoke
it very poorly (1963a). Similarly, in summarizing his findings concerning
current language maintenance among pre-World War I arrivals in the
United States coming from rural Eastern and Southern European back-
grounds, Fishman reported a long-term distinction between attitudes
and use, namely, an increased esteem for non-English mother tongues
concomitant with the increased relegation of these languages to fewer
and narrower domains of language use (Fishman 1965f). In the latter
case, the particular non-English mother tongues in question were now
found to be viewed positively and nostalgically by older first and second
generation individuals who had formerly characterized these tongues as

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
332 Joshua A. Fishman
ugly, corrupted and grammarless in pre-World War II days. Younger
second and third generation individuals were found to view these mother
tongues (almost always via translations) with less emotion but with even
more positive valence. Instead of a "third generation return" (Hansen
1940) there seemed to be an "attitudinal halo-ization" within large
segments of all generations, albeit unaccompanied by increased usage.
This development (a negative relationship over time between use rates
and attitudinal positiveness) was not predictable from most earlier stu-
dies of language maintenance or language shift in immigrant or non-
immigrant settings. We are far from knowing whether its explanation in
American contextual terms (i.e., in terms of the greater acceptability of
marginal rather than either primordial or ideologized ethnicity) would
also apply to other settings in which similar circumstances might obtain.
Recent methodological clarification of the language-attitude area (Fish-
man and Agheyisi 1970) should now make it possible for workers to
move ahead in this area along a broad front of little explored topics and
approaches.

7.32 Overt Behavioral Implementation of Attitudes, Feelings and

Beliefs.
Both language reinforcement ('language movements') and language
planning may be subsumed under this heading. Language rein-
forcement may proceed along voluntary as well as along official routes
and encompasses organizational protection, statutory protection, agita-
tion and creative production. As for language planning, it has not always
been recognized that much (if not most) of its activity (codification,
regularization, simplification, purification, elaboration, and the imple-
mentation and evaluation of all of the foregoing) occurs in the context
of language maintenance or language shift (Fishman 1966c, Ch. 21).
The possible relationships between language reinforcement (or lan-
guage planning), on the one hand, and the waxing or waning of actual
language use (or of other socio-cultural processes) are largely unknown
at this time. Data from the American immigrant case imply that a num-
ber of unexpected relationships may obtain in that novel reinforcements
may be introduced as actual language use diminishes. Thus, as even
some of the more 'exotic' mother tongues (i.e., mother tongues not
usually considered to be among the major carriers of European civi-
lization and, therefore, hitherto usually associated only with foreign
ethnicity in the minds of 'average Americans' (Hayden 1966)) have
ceased to be primarily associated with immigrant disadvantages or with
full-blown religio-ethnic distinctiveness among their own sometime-and-
erstwhile-speakers, they have been increasingly introduced as languages
of study at the university, college and public high school levels (Haugen

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 333
1953, Kloss 1966b). At the same time, massive displacement seems to
have had greater inhibitory impact on language planning efforts in the
American immigrant case than it has had on language reinforcement
efforts. The latter are essentially conservative and seem to require less
in the way of highly specialized leadership. The former are frequently
innovative and dependent upon expert personnel working in concert
with compliance producing or persuasive authority. To what extent this
differential impact also holds true in other types of language shift set-
tings is currently unknown but worthy of study.
Advocates of languages that are undergoing displacement are often
much more exposed to (and identified with) the values and methods of
their linguistic competitors than were their less exposed (and less threa-
tened) predecessors. As a result, they are more likely to adopt organized
protective and publicity measures from more 'advantaged' co-territorial
(other-tongue) models to serve language maintenance purposes (Fish-
man 1969a). The introduction of a few ethnically infused languages into
the curricula of American high schools, colleges and universities repre-
sents just such a recent innovation on behalf of mother tongue main-
tenance - and an even more de-ethnicized one (Nahirny and Fishman
1965) one than was the innovative establishment of ethnic group news-
papers, schools, cultural organizations and camps prior to World War
I. In contrast, the normal processes of controlled language change and
the more aroused processes of conscious language planning may require
more than 'last ditch' ingenuity. However, to what extent reinforcement
and planning are differently balanced given varying degrees of displace-
ment or augmentation is currently unknown but worthy of study. In
addition to its importance in its own right, the overall study of the
relationship between language attitudes and language behaviors (Fish-
man 1969c) will also gain greatly from attention to topics such as this.

7.33 Cognitive Aspects of Language Response

Constantly flitting between the above two categories and overlapping
partially with the one, with the other, or with both are such matters as:
consciousness of mother tongue (or 'other tongue') as an entity separate
from folkways more generally; knowledge of synchronic variants, lan-
guage history and literature; and perceptions of language as a com-
ponent of 'groupness'. We have little systematic information concerning
the circumstances under which language consciousness, language know-
ledge and language-related groupness-perceptions do or do not enter
into reference group behavior in contact situations. As a result, it is
difficult to say at this time whether or when language maintenance and
language shift are ideologically mediated as distinguished from their
more obvious situational and instrumental determinants discussed thus

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
334 Joshua A. Fishman
far. We recognize very gross long-term contrasts in this connection,
namely, that there were periods and regions when language "was in no
way regarded as a political or cultural factor, still less as an object of
political or cultural struggle" (Kohn 1945, p. 6); that there were
other periods and regions marked by a sharp increase in such regard,
so that language became a principle "in the name of which people . . .
(rallied) themselves and their fellow speakers consciously and explicitly
to resist changes in either the functions of their language (as a result of
language shift) or in the structure or vocabulary (as a consequence of
interference)" (Weinreich 1953a, p. 99), and that there currently seems
to be less of this than previously, particularly if we compare African
with European nationbuilding. However, gross differentiations such as
these are patently insufficient to enable us to clarify the conditions
under which language becomes a prominent component in perceptions
of 'own-groupness' and 'other-groupness'. This topic (language-related
groupness-perception) is, of course, closely related to one previously
mentioned, namely, the role of language in group membership and in
group functioning (see section 7.2111, above). In the American immi-
grant case we have seen a growing dissocation between self-perceived
ethnic identification and language maintenance. Far from being viewed
as necessary components of groupness (whether in the sense of resul-
tants or contributors) non-English mother tongues appear to be viewed
increasingly in terms of non-ethnic cultural and non-ethnic practical

TABLES 2 0 a a n d 2 0 b
Attitudes and beliefs with respect to Spanish among Ordinary
Puerto Ricans (OPR) and Intellectuals, Leaders and Artists (ILA) in
the Greater New York Metropolitan Area (Fishman 1969e)

OPR ILA
Response (« = 32) (n - 20)
No 20 (62%) 2(10%)
Yes 12(38%) 18 (90%)

TABLE 2 0 b
Are there many "Nuyorquinos" who do not speak or understand
Spanish?
OPR ILA
Response (n = 29) (n = 20)

Yes (many do not understand) 2 ( 7%) 1 ( 5%)

Most understand little and speak poorly 3 (10%) 4 (20%)
Most understand well but speak poorly 3 (10%) 14(70%)
Most speak and understand without real difficulty 21 (73%) 1 ( 5%)

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 335
considerations. At the same time, some form of ethnic self-identification
is frequently still reported by many of those who no longer claim any
facility at all in their ethnic mother tongues, implying that in several
American immigrant-derived groups some kind of ethnicity usually
appears to be a much more stable phenomenon than language mainte-
nance. Indeed, some groups are able to maintain newspapers, schools and
organization long after they have lost their non-ethnic mother tongues
(Table 19a and Table 19b). Most immigrants became bilingual much
before they embarked on de-ethnization or seriously contemplated the
possibility of bi-culturism. However, there were obviously exceptions to
this process, both in the United States and in other contact settings. We
certainly do not seem to be in a position to indicate the underlying
regularities in this subtle area of inquiry at the present time, except to
point out that the segments of the population among which language
consciousness, language interest, and language-related groupness-per-
ceptions are likely to be in evidence are normally quite small and elitist
in nature (Tables 20a and 20b).
We know very little about the interaction among the three compo-
nents of behavior toward language or about the interaction between
any of these components and the larger psychological, social and cultu-
ral processes discussed earlier. Rather than being a 'natural', omni-
present condition, either in monolingual or in multilingual settings,
heightened and integrated behaviors toward language may be related to
somewhat rare and advanced symbolic and ideological extensions of
primordial ethnicity. Such extensions may well require a particular
level of socio-cultural development and a particular group of custodians
for their preservation and further elaboration. They almost certainly
require a relatively advanced level of elitist concentration on intra-
elitist concerns, often in advance of elitist concerns for communication
within the masses. Nevertheless, none of these desiderata need have
invariable consequences for behavior toward language. Even where
heightened and integrated behaviors toward language are culturally
present they will not be equally operative in all situations or among all
population subgroups. Furthermore, even where they are culturally
present they need not be uniformally related to other symbolically ela-
borated forms of behavior. Thus, this area remains the most unsyste-
matized topical sub-division of the study of language maintenance and
language shift. Perhaps it can be clarified in the future as a result of
concomitant clarification and constant interrelation in connection with
the two other major sub-divisions within this field of inquiry.

7.34 Interference and Switching.

Within the topical subdivision of behavior toward language we once

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
336 Joshua A. Fishman
again meet the topic of interference and switching, first introduced in
section 7.1, above. The absence or presence of interference and
switching can have cognitive, affective and overt implementational
implications for language maintenance and language shift. Certainly,
both interference and switching are related to the domains and variance
sources of bilingualism, on the one hand, and to socio-cultural processes
and type of interaction, on the other hand. Moreover, within this
topical subdivision it is appropriate to stress that where attitudes and
awareness concerning purism obtain, interference is sometimes viewed
as an imperfection - not in the speaker or in his productions but in the
language itself. At the opposite pole, there are multilingual contact
situations in which conscious, purposive interference obtains. In these
instances speakers attempt to incorporate into their language usage
as many elements or features as possible from another language
including (in very advanced cases) interference in stress patterns,
intonation, and Denkformen. In either case (i.e., when interference
occurs although it is considered undesirable, or when interference
occurs and is considered desirable) interference is not always
considered to be all of one piece. Certain occurrences are considered
to be more acceptable, excusable, permissible, necessary than others. In
either case it can become a factor in hastening language shift, particu-
larly since bilinguals tend to interpret interference in each of the
languages known to them quite differently. Finally, at a point when
language shift is appreciably advanced, certain sounds and forms of the
language undergoing displacement may become so difficult for the
average speaker (while errors in connection with them may become so
stigmatized among purists) that this in itself may accelerate further
shift. All in all, recognition of interference, attitudes toward interfer-
ence, and the behavioral consequences of interference represent in-
teresting and important topics within the field of language maintenance
and language shift.

7.4 A Glance Back and a Glance Ahead

Various language maintenance and language shift phenomena have

long been of interest to scholars and to laymen. Several sub-topics
within this area have indisputed relevance to the daily concerns and
joys of millions. Others, of more theoretical interest, are closely related
to topics of recognized concern to linguists, anthropologists, sociologists,
psychologists, political scientists, educators, etc. Culture contact and
language contact will always be with us, and out of these contacts will
come modifications in habitual behavior as well as attempts to restrain

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 337
or channel such modifications. Whether (or when) language habits
change more or less quickly than others, whether or when language
loyalties are more or less powerful than others, indeed, whether (or
when) men can live in a supraethnic tomorrow without strong links
(linguistic or non-linguistic) to their ethnic yesterday and today -
these are questions to which there are currently no definitive answers.
However, interest in social-psychological aspects of language behavior
is currently growing (whether under that name or under the name of
sociolinguistics, anthropological linguistics, ethno-linguistics, the ethno-
graphy of speaking, the ethnography of communication, the sociology
of language, or some other designation). In most instances, there is some
recognition of behavior toward language as a crucial topic within the
field of social behavior through language. This growing interest will
undoubtedly contribute answers to many of the currently unanswerable
questions within the field of language maintenance and language shift.
Three major subdivisions of the study of language maintenance and
language shift have been suggested. The first deals with the precise
establishment of habitual language use in a contact situation. This re-
quires instruments just beginning to become available for the measure-
ment of degree of bilingualism and of location of bilingualism along
sociologically relevant dimensions. Degree of bilingualism, hitherto re-
cognizable in terms of automaticity, proficiency, and code-intactness at
the phonetic, lexical and grammatical levels, must also be investigated
with respect to media variance and overtness variance. Location of
bilingualism requires investigation with respect to functional diversifi-
cation in appropriately designated domains of language, each domain
being abstracted from patterned role-relations, topics, locales and/or
other lower order phenomena. The complex relationships between the
several components of degree of bilingualism and location of bilingua-
lism may be represented by a dominance configuration which, in turn,
may or may not be reducible to a single index of direction of bilingua-
lism. The drift of language maintenance or language shift may be
established by diachronic measures pertaining to some or all af the
above factors.
The second major topical subdivision of the study of language main-
tenance and language shift deals with psychological, social and cultural
processes that are associated with ascertained changes in habitual lan-
guage use. No conceptual systematization of these processes is currently
available although several preliminary typologies of "contact situations"
exist and require further refinement in cross-cultural perspective. The
greatest encouragement in this topical subdivision comes from the
accelerating interdisciplinary work on socio-cultural and politico-ope-
rational change (including work on development and modernization).

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
 338 Joshua A. Fishman
To the extent that the study of language maintenance and language
shift will become increasingly linked to ongoing theoretical and em-
pirical refinements in the study of psycho-socio-cultural stability and
change more generally the more rapidly will mutually rewarding pro-
gress occur.
The third (and final) major subdivision of the study of language
maintenance and language shift pertains to behavior toward language,
including (but not limited to) more focused and conscious behaviors
on behalf of maintenance or shift. Three major sub-topics within this
topic are recognizable: Attitudinal-affective behaviors (loyalty, anti-
pathy, etc.), overt behavioral implementation (control or regulation of
habitual language use via reinforcement, planning, prohibition, etc.),
and (overlapping partially with each of the two foregoing sub-topics)
cognitive behaviors (language consciousness, language knowledge, lan-
guage-related group-perceptions, etc.).
Two socio-linguistic patterns, that of the urban American immigrant
and that of the urban French-Canadian nationalist, have been repeated
many times in the past century. The increasing use of Russian alone by
Soviet minorities - particularly the smaller ones - whether they be im-
migrants to large urban centers in other regions or outnumbered by
Russians and various other immigrants into their own regions, has
followed the same path as the increasing use of English alone by immi-
grants to the United States, the increasing use of Spanish alone by
indigenous Indian populations moving to urban centers throughout
Latin America, or the increasing use of Wolof alone by the diverse Sene-
galese populations that began to move to Dakar more than a generation
ago. Similarly, the increasing use of the mother tongue in the domains
of education, industry and government (which had previously 'belonged',
so to speak, to English), that has increasingly typified French-Canada,
is not at all unlike the growing displacement of English or another
Western language of wider communication in Puerto Rico, Tanzania,
Kenya, India, Pakistan, Malaysia and the Philippines. The one group
of cases illustrates the general inability of dislocated populations to
maintain domain separation and, therefore, a sufficiently distinctive
functional allocation of codes in their verbal repertoires, such as to
render their mother tongues necessary for membership and status even
within the home, neighborhood and other intra-group domains. The
other group of cases illustrates the generally far greater ability of
sedentary populations to withstand the onslaught of foreign-inspired
political, educational, social and economic domination. If domain
separation is maintained, at least between the L domains of home and
neighborhood and the H domains of government, education and
religion, a subsequent mobilization of the indigenous population around

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 339
a new, nationalist proto-elite may yet lead to the introduction (or re-
introduction) of the vernacular into those domains from which it has
been barred or displaced.
In the urban American immigrant case - as in all instances in which
severely dislocated populations have been presented with tangible op-
portunities to share in new role-relationships and in vastly improved
power- and status-networks - a new language initially entered the
verbal repertoire of the speech community for marginal metaphorical
purposes only. Situational and metaphorical switching both were pos-
sible only with respect to several varieties of the ethnic mother tongue
or its H + L matrix. However, with the passage of time intra-group
power, status and even membership per se, all come to be granted
on the basis of mastery of the new language. As a result, the ethnic
mother tongue became increasingly relegated to metaphorical purposes
(humor, contrast, tenderness) and, therefore, to oblivion as a third
generation arose that had itself directly experienced none of the
situations upon which the metaphorical functions of the ethnic mother
tongue rested in the usage of 'old timers' and the second generation.
In the case of less dislocated populations - where the absence of
widespread social mobility or of physical extirpation from established
roles and networks helped preserve the distinction between intra-group
and extra-group domains - the new language normally gained metapho-
rical recognition only insofar as the majority of intra-group networks
and role relations were concerned. As a result, it served primarily as
an intergroup H for the few well-placed individuals with inter-group
roles. Little wonder then that among the rank and file of such less
dislocated populations - including the Alsatians discussed by Tabouret-
Keller (1968) and by Verdoodt (1971) and the Swabians discussed
by Fishman and Lueders (1971) - H varieties do not displace L
varieties and indeed, are themselves easily displaced by yet newer
H varieties resulting from the temporary intrusions of new political
authorities.
The above sketch is still more suggested than demonstrated. It de-
pends more on theoretical parsimony than on empirical data. The ex-
haustive study of language maintenance and language shift ultimately
requires not merely theory but also theory tested and revised in the
light of hard data. Since the basic instruments and theory required for
the establishment of degree and direction of language maintenance or
language shift are now beginning to be available (certainly this is true
relative to the situation five years ago) it would now seem to be most
crucial to devote increasing amounts of theoretical and empirical atten-
tion to comparative (cross-network, cross-speech community, cross-
polity and cross-cultural) study of the psycho-socio-cultural antecedents

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
340 Joshua A. Fishman
and concomitants of language maintenance and language shift. The next
few years will doubtlessly see the greatest progress precisely along these
lines, i.e., along lines for which the social anthropologist, social psycho-
logist and sociologist - rather than the linguist - must take primary
responsibility.

8.0 SOCIO-CULTURAL ORGANIZATION:

LANGUAGE CONSTRAINTS AND LANGUAGE REFLECTIONS

One of the major lines of social and behavioral science interest in

language during the past century has been that which has claimed that
the radically differing structures of the languages of the world constrain
the cognitive functioning of their speakers in different ways. It is only
in relatively recent years - and partially as a result of the contributions
of psycholinguists and sociolinguists - that this view (which we shall
refer to as the linguistic relativity view) has come to be replaced by
others: (a) that languages primarily reflect rather than create socio-
cultural regularities in values and orientations and (b) that languages
throughout the world share a far larger number of structural universals
than has heretofore been recognized. While we cannot here examine
the work related to language universals (Greenberg 1966; Osgood
1960), since it is both highly technical and hardly sociolinguistic in
nature, we can pause to consider the linguistic relativity view itself as
well as the linguistic reflection view which is increasingly coming to
replace it in the interests and in the convictions of social scientists. It
is quite clear why so much interest has been aroused by the question of
language as restraint and language as reflection of socio-cultural orga-
nizations. Both of these views are undirectional. One posits that lan-
guage structure and language usage are fundamental and 'given' and
that all behavior is influenced thereby. The other claims that
social organization and behavior are prior and language merely reflects
these. A position on one side or another of this argument must be
taken by those who are interested in changing or influencing the 'real
world' of behavior.

8.1 Grammatical Structure Constrains Cognition

The strongest claim of the adherents of linguistic relativity - whether

by Whorf (1940,1941), Hoijer (1951,1954), Trager (1959), Kluckhohn
(1961), or by others - is that cognitive organization is directly con-
strained by linguistic structure. Some languages recognize far more tenses
than do others. Some languages recognize gender of nouns (and, there-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 341
fore, also require markers of gender in the verb and adjective systems)
whereas others do not. Some languages build into the verb system
recognition of certainty or uncertainty of past, present, or future action.
Other languages build into the verb system a recognition of the size
shape and color of nouns referred to. There are languages that signify
affirmation and negation by different sets of pronouns just as there are
languages that utilize different sets of pronouns in order to indicate
tense and absence or presence of emphasis. Some languages utilize tone
and vowel length in their phonological systems whereas English and
most other modern European languages utilize neither. There are lan-
guages that utilize only twelve phonemes while others require more than
fifty. A list of such striking structural differences between languages
could go on and on - without in any way denying that each language
is a perfectly adequate instrument (probably the most adequate instru-
ment) for expressing the needs and interests of its speakers. That the
societies using these very different languages differ one from the other
in many ways is obvious to all. Is it not possible, therefore, that these
socio-cultural differences - including ways of reasoning, perceiving,
learning, distinguishing, remembering, etc. - are directly relatable to the
structured differences between the languages themselves? The Whorfian
hypothesis claims that this is indeed the case (Fishman 1960).
Intriguing though this claim may be it is necessary to admit that
many years of intensive research have not succeeded in demonstrating
is to be tenable. Although many have tried to do so no one has
successfully predicted and demonstrated a cognitive difference between
two populations on the basis of the grammatical or other structural
differences between their languages alone. Speakers of tone languages
and of vowel length languages and of many-voweled languages do not
seem to hear better than do speakers of languages that lack all of these
features. Speakers of languages that code for color, shape and size
in the very verb form itself do not tend to categorize or classify a
random set of items much differently than do speakers of languages
whose verbs merely encode tense, person and number (Carroll and
Casagrande 1958). Whorf's claims (namely, that " . . . the background
linguistic system [in other words, the grammar] of each language is
not merely a reproducing instrument for voicing ideas, but rather is
itself the shaper of ideas, the program and guide for the individual's
mental activity, for his analysis of impressions, for his synthesis of his
mental stock in trade. Formulation of ideas is not an independent
process, strictly rational in the old sense, but it is part of a particular
grammar and differs, from slightly to greatly, between grammars"
1940) seem to be overstated and no one-to-one correspondence between
grammatical structure and either cognitive or socio-cultural structure

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
342 Joshua A. Fishman
measured independently of language has ever been obtained. Several
of the basic principles of sociolinguistic theory may help explain why
this is so, although the psychological maxim that most men think about
what they are talking about (i.e., that language structure is always
being struggled with via cognitive processes) should also be kept in
mind.
In contrast with the older anthropological-linguistic approach of
Whorf, Sapir, Kluckhohn, Korzybski and others who pursued this
problem during the first half of the twentieth century, sociolinguistics is
less likely to think of entire languages or entire societies as categorizable
or typable in an overall way. The very concepts of linguistic repertoire,
role repertoire, repertoire range and repertoire compartmentalization
argue against any such neat classification once functional realities are
brought into consideration. Any reasonably complex speech community
contains various speech networks that vary with respect to the nature
and ranges of their speech repertoires. Structural features that may be
present in the speech of certain interaction networks may be lacking
(or marginally represented) in the speech of others. Structural features
that may be present in certain varieties within the verbal repertoire of
a particular interaction network may be absent (or marginally repre-
sented) in other varieties within that very same repertoire. Mother-
tongue speakers of language X may be other-tongue speakers of lan-
guage Y. These two languages may co-exist in a stable diglossic pattern
throughout the speech community and yet be as structurally different
as any two languages chosen at random.
Certainly, all that has been said above about the difficulty in
setting up 'whole-language' typologies is equally true when we turn
to the question of 'whole-society' typologies. Role repertoires vary from
one interaction network to the next and roles themselves vary from one
situation to the next within the same role-repertoire. Distinctions that
are appropriately made in one setting are inappropriate in another and
behaviors that occur within certain interaction networks do not occur
in still others within the same culture. The existence of structured bicul-
turism is as real as the existence of structured bilingualism and both of
these phenomena tend to counteract any neat and simple linguistic
relativity of the kind that Whorf had in mind.
Nevertheless, there are at least two large areas in which a limited
degree of linguistic relativity may be said to obtain: (a) the structuring
of verbal interaction and (b) the structuring of lexical components.
The first area of concern points to the fact that the role of language
(when to speak, to whom to speak, the importance of speaking per se
relative to inactive silence or relative to other appropriate action)
varies greatly from society to society (Hymes 1966). However, this type

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 343

Data of
(Cognitive) Behavior
Data of Language data Non-linguistic
Language Characteristics ('cultural themes') data
Lexical or 'semantic' Level 1 Level 2
characteristics
Grammatical characteristics Level 3 Level 4

Figure 20. Schematic Systematization of the Whorfian Hypothesis (Fishman 1960)

Level 1 of the Whorfian ('linguistic relatively') hypothesis predicts that speakers
of languages that make certain lexical distinctions are enabled thereby to talk
about certain matters (for example, different kinds of snow among speakers of
Eskimo and different kinds of horses among speakers of Arabic) that cannot as
easily be discussed by speakers of languages that do not make these lexical
distinctions. Similarly, Level 3 of the Whorfian hypothesis predicts that speakers
of languages that possess particular grammatical features (absence of tense in
the. verb system, as in Hopi, or whether adjectives normally precede or follow
the noun, as in English vs. French) predispose these speakers to certain cultural
styles or emphases (timelessness; inductiveness vs. deductiveness). These two
levels of the Whorfian hypothesis have often been criticized for their anecdotal
nature as well as for their circularity in that they utilized verbal evidence for
both their independent (causal) and dependent (consequential) variables. Level 2
of the Whorfian hypothesis predicts that the availability of certain lexical items or
distinctions enables the speakers of these languages to remember, perceive, or
learn non-linguistic tasks more rapidly or completely than can the speakers of
languages that lack these particular lexical items or distinctions. This level of
the Whorfian hypothesis has been demonstrated several times—most recently
and forcefully in connection with the differing color terminologies of English and
Zuni—but it is difficult to argue that the absence of lexical items or distinctions
in a particular language is more a cause of behavioral differences than a reflec-
tion of the differing socio-cultural concerns or norms of its speakers. As soon
as speakers of Zuni become interested in orange (color) they devise a term for
it. Language relatively should be more stable and less manipulable than that!
Level 4 of the Whorfian hypothesis is the most demanding of all. It predicts
that grammatical characteristics of languages facilitate or render more difficult
various non-linguistic behaviors on the part of their speakers. This level has yet
to be successfully demonstrated via experimental studies of cognitive behavior.

of relativity has nothing to do with the structure of language per se

in which Whorf was so interested. T h e second area of concern deals
with lexical taxonomies and with their consequences in cognition and
behavior. However, these border o n being linguistic reflections of socio-
cultural structure rather than being clearly and solely linguistic con-
strains that inevitably and interminably must bring about the particular
behaviors to which they are supposedly related. It is to a consideration
of these lexical taxonomies that w e n o w turn.

8.2 Lexical Structure Constrains Cognition

For many years it was believed that the only tightly structured levels of

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
344 Joshua A. Fishman
language were the grammatical (morphological and syntactic), on the
one hand, and the phonological, on the other. These two levels cer-
tainly received the brunt of linguistic attention and constituted the
levels of analysis of which linguists were most proud in their inter-
actions with other social and behavioral scientists^ By contrast, the
lexical level was considered to be unstructured and exposed to infinite
expansion (as words were added to any language) and infinite inter-
ference (as words were borrowed from other languages). A small but
hardy group of lexicographers (dictionary makers) and etymologists
(students of word origins) continued to be enamoured of words per se
but the majority of linguists acted as though the lexicon was the black
sheep, rather than a bona fide member in good standing, of the linguistic
family. The discover of structured parsimony in parts of the lexicon has
done much to revive linguistic interest in the lexical level of analysis.
The discovery as such is one in which psychologists, anthropologists and
sociologists were every bit as active as were linguists themselves (if not
more so). This may also explain why the interrelationship between
lexical organization and behavioral organization has been so prominent
in conjunction with the investigation of lexical structure.
The psychological contributions to this area of analysis take us back
to one level of the Whorfian hypothesis (see level 2 in Figure 20).
Psychologists had long before demonstrated that the availability of ver-
bal labels was an asset in learning, perception and memory tasks (see,
e.g., Carmichael et al. 1932; Lehmann 1889; Maier 1930). A new
generation of psychologists has recently set out to determine whether
this could be demonstrated both interlinguistically (i.e., by comparing
different languages) as well as intralinguistically (i.e., within a given
language) on a structured set of behaviors that corresponded to a
structured portion of lexicon.
They chose the color spectrum to work with because it is a real
continuum that tends to be environmentally present in all cultures.
Nevertheless, the investigators hypothesized that language labels for the
color spectrum are culturally idiosyncratic. These labels not only chop
up the color continuum into purely conventional segments in every lan-
guage community, but they probably do so differently in different
language communities. By a series of ingenious experiments, Brown
and Lenneberg (1954), Lenneberg (1953, 1957), Lantz and Stefflre
(1964), and others have demonstrated that this was indeed true. They
have demonstrated that those colors for which a language has readily
available labels are more unhesitatingly named than are colors for
which no such handy labels are available. They have shown that the
colors for which a language has readily available labels (i.e., highly
codable colors) are more readily recognized or remembered when they

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 345
must be selected from among many colors after a delay subsequent to
their initial presentation. They have demonstrated that somewhat diffe-
rent segments of the color spectrum are highly codable in different
language communities. Finally, they have shown that the learning of
nonsense-syllable associations for colors is predictably easier for highly
codable colors than for less codable colors that require a phrase - often
an individually formulated phrase - in order to be named.
All in all, this series of experiments has forcefully shown that the
availability of a structured set of terms has both intralinguistic as well
as interlinguistic consequences. However, in addition, it has underscored
the equally important fact that every speech community has exactly such
terms for those phenomena that are of concern to it. Certainly, artists,
painters, and fashion-buyers have a structured color terminology that
goes far beyond that available to ordinary speakers of English. The
relative absence or presence of particular color terms in the lexicon
of a given speech network is thus not a reflection of the state of that
network's code per se as much as it is a reflection of the color interests,
sensitivities and conventions of that network at a particular time in
its history.
A color terminology is merely one kind of folk-taxonomy, i.e., it is
an example of the many emic semantic grids that are contained in the
lexicons of all speech communities. Other such examples are the kinship
terminologies of speech communities, their disease or illness termino-
logies, their plant terminologies, their terms of address, etc. (Basso
1967; Conklin 1962; Frake 1961,1962; Pospisil 1965; Friederich 1966;
Metzger and Williams 1966; Price 1967; Wittermans 1967; etc.). In
each of these instances the particular lexicons involved constitute "un
systeme on tout se tient".
Each such system is considered by its users to be both literally ex-
haustive and objectively correct. Nevertheless, each system is socially
particularistic, i.e., for all of its self-evident objectivity ("what other
kind of kinship system could there possibly be?" — we can imagine the
average member of each of the scores of such systems asking himself),
it is a reflection of locally accepted conventions rather than a necessary
reflection either of nature or of language per se. This last is particularly
well demonstrated in the work of Friederich (on Russian kinship terms),
Wittermans (on Javanese terms of address), and Basso (on Western
Apache anatomical terms and their extension to auto parts; see Figure
21).
The Russian revolution brought with it such fargoing social change
that the kinship terms in use in Czarist days had to be changed to some
degree. In contrast with the refined stratificational distinctions that
existed in Czarist days - distinctions that recognized gradations of

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
346 Joshua A. Fishman

Figure 21. Lexical structure and social change (Basso, 1967)

ndg bi tsi ("man's body").

ni
("face") ebiyi* ("entrails")
M
O
gon ("hand and arm"
do ("chin and jaw")

3
XI
wos ("shoulder")

T3

pit ("stomach")
ta ("forehead")
C

òi ("intestine")

jisole ("lung")
ze' ("mouth")

ysn ("back")
CS

inda ("eye")

tsçs ("vein")
zik ("liver")
ke' ("foot")
JS

ji ("heart")
Kko ("fat")

cj ("nose")
60

'3

Note: Black bars indicate position of additional (unextended) anatomical terms.

Figure 21a. Taxonomic Structure of Anatomical Set

Western Apache

nalbil bi tsi ("automobile's body")

ni a ebiy<'
("machinery
under hood")

"a.
ze' ("gas pipe opening")

ts?s ("electrical wiring")

ta ("front of cab," "tc
wos ("front fender")
do ("front bumper")

/an ("bed of truck")

gon ("front wheel")

ci ("radiator hose")
kai ("rear fender")

jisole ("radiator")
inda ("headlight")
ke' ("rear wheel")

ji ("distributor")
pit ("gas tank")
z/k ("battery")
liko ("grease")

cj ("hood')

a "Area extending from top of windshield to bumper"

Figure 21b. Taxonomic Structure of Extended Set

power, wealth and proximity within the universe of kin, n o t unlike

t h o s e that w e r e recognized i n t h e larger universe of social and e c o n o m i c
relationships - Soviet society stressed far f e w e r and broader distinctions.
A s a result, various kinship terms w e r e a b a n d o n e d entirely, others were
m e r g e d a n d other w e r e expanded. A very similar d e v e l o p m e n t trans-
pired in Javanese with respect to its highly stratified system of terms

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 347

Superiors

Equal and Equal and Not

Solidary Solidary
T V

Inferiors

(a)

Superior and f Superior and Not

Solidary T V Solidary

Equal and Equal and Not

Solidary Solidary

T V

Inferior and 1 Inferior and Not

Solidary T V Solidary

(b)

Figure 22. The two-dimensional semantic (a) in equilibrium and (b) under tension
(Brown and Gilman, 1960)

Solidarity comes into the European pronouns as a means of differentiating address

among power equals. It introduces a second dimension into the semantic system
on the level of power equivalents. So long as solidarity was confined to this level,
the two-dimensional system was in equilibrium (see Figure 22a), and it seems to
have remained here for a considerable time in all our languages. It is from the
long reign of the two-dimensional semantic that T derives its common definition
as the pronoun of either condescension or intimacy and V its definition as the
pronoun of reverence or formality. These definitions are still current but usage
has, in fact, gone somewhat beyond them.
The dimension of solidarity is potentially applicable to all persons addressed.
Power superiors may be solidary (parents, elder siblings) or not solidary (officials
whom one seldom sees). Power inferiors, similarly, may be as solidary as the old
family retainer and as remote as the waiter in a strange restaurant. Extension of
the solidarity dimension along the dotted lines of Figure 22b creates six categories
of persons defined by their relations to a speaker. Rules of address are in con-
flict for persons in the upper left and lower right categories. For the upper left,
power indicates V and solidarity T. For the lower right, power indicates T and
solidarity V.
Well into the nineteenth century the power semantic prevailed and waiters, com-
mon soldiers, and employees were called T while parents, masters, and elder
brothers were called V. However, all our evidence consistently indicates that in
the past century the solidarity semantic has gained supremacy. The abstract result
is a simple one-dimensional system with the reciprocal T for the solidary and
the reciprocal V for the nonsolidary.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
348 Joshua A. Fishman
of address. The impact of post-war independence, industrialization, ur-
banization and the resulting modification or abandonment of traditional
role-relationships led to the discontinuation of certain terms of address
and the broadening of others, particularly of those that implied rela-
tively egalitarian status between interlocutors. Howell's review of changes
in the pronouns of address in Japan (1967) also makes the same point,
as did his earlier study of status markers in Korean (1965). Not only
does he indicate how individuals change the pronouns that they use in
referring to themselves and to each other, as their attitudes and roles
vis-a-vis each other change, but he implies that widespread and cumu-
lative changes of this kind have occurred in Japan since the war, to the
end that certain pronouns have been practically replaced by others. Cer-
tainly the best known study of this kind is Brown and Gilman's review
of widespread Western European social change with respect to the use
of informal (T) vs. formal (V) pronouns and verb forms for the third
person singular (1960). Feudalism, renaissance, reformation, the French
Revolution, 19th century liberalism and 20th century democratization
each had recognizable and cumulative impact. As a result, both T and
Y forms were retained in interclass communication (except in the
case of English) but the;- differential use came to indicate differences
primarily in solidarity or differences in solidarity and in power rather
than differences in power alone as had been the case in the early
middle ages (See Figure 22).
Note that the complexities of the pre-revolutionary kinship taxono-
mies in Russia did not keep Russians from thinking about or from
engaging in revolution. Note also that the revolution did not entirely
scrap the pre-existing kinship taxonomy. Similarly, the Apache anato-
mical taxonomy did not preclude (but rather assisted) taxonomic orga-
nization of automobile parts. Thus, while we are clearly indicating the
untenability of any strong linguistic relativity position when we show
that semantic taxonomies are subject to change, expansion and contrac-
tion as the socio-cultural realities of their users change, we are also
demonstrating that their linguistic reflection of social reality is also
likely to be both slow and partial. Nevertheless, as between the two, the
taxonomic reflection of socio-cultural reality is more likely to have
widespread heuristic utility at any given time, however much the exist-
ence of such taxonomies is likely to be constraining in the momentary
cognitive behavior of individual members of socio-cultural systems.
The emic distinctions which underlie these taxonomies are differen-
tially constraining for various interaction networks within any speech
community. Some networks (e.g., the networks of quantitative scien-
tists) can repeatedly rise above the cognitive constraints of the taxono-
mies current in their speech communities. These networks are likely to
be the ones that are most actively engaged in social change and in

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 349
taxonomic change as well. Other networks are unable to break out of
the socio-cultural taxonomies that surround them. In such cases, as,
e.g., in connection with Kantrowitz' race relations taxonomy among
White and Negro prison inmates (1967; See Figure 23), or Price's
botanical taxonomies among the Huichols (1967), these taxonomies may
be taken not only as useful reflections of the cognitive world of the
speech community from which they are derived but also as forceful
constraints on the cognitive behavior of most, if not all, of the indi-
vidual members of these networks.
CONCEPTS OR names USED names USED IN C O M M O N CONCEPTS OR names
EXCLUSIVELY BY NEGROES BY BOTH NEGROES AND USED EXCLUSIVELY BY
WHITES WHITES

A white man who A white man who

does not discriminate associates with
against Negroes Negroes

(
1
free thinker
1
nigger lover

A Negro who believes

white are superior, A Negro who is not
and acts subservient aggressive or does
to them not insist on his
equal rights with
whites
sander, smoke-
jejfjeffer,jejf- blower, easy going-
davis, jeff artist, black slave
Charlie mccarthy,
chalk eyes, renegade,
shuffler, some timer,
uncle torn, devil lover,
stays in uncle iom's-
cabin, hoosier lover

A Negro who hates

A Negro who constantly whites, and expresses
tells both Negroes and it vehemently and
whites that Negroes freely among Negroes
must be accorded the
same status and rights
as whites civil rights nigger,
freedom rider, little•
civil rights man, rocker, lumumba, a-
aggressive man, free- race man, mau mau,1 martin luther king,
speaker, man of equal rights man mau mau preacher, •
reasoning muslim, pale hater,
Figure 23. Examples of Vocabulary Used Differently by muslim, pale hater,
torn torn guy
White and Negro Prison Inmates (Kantrowitz, ADS, 1967).

8.3 Lexical Structure Reflects Social Organization

There are, however, more pervasive (and, therefore, seemingly less sys-
tematic) ways in which lexicons in particular and languages as a whole

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
350 Joshua A. Fishman
are reflective of the speech communities that employ them. In a very
real sense a language variety is an inventory of the concerns and
interests of those who employ it at any given time. If any portion of
this inventory reveals features not present in other portions this may
be indicative of particular stresses or influences in certain interaction
networks within the speech community as a whole or in certain role-
relationships within the community's total role-repertoire. Thus, Ep-
stein's study of linguistic innovation on the Copperbelt of Northern
Rhodesia (1959) revealed that the English and other Western influences
on the local languages were largely limited to matters dealing with
urban, industrial and generally non-traditional pursuits and relation-
ships. Similarly, M. Weinreich's meticulous inquiry into the non-Ger-
manic elements in Yiddish (1953) sheds much light on the dynamics of
German-Jewish relations in the 11th Century Rhineland.
Like all other immigrants to differently-speaking milieus, Jews, learn-
ing a variety of medieval German in the 11th Century, brought to this
language learning task sociolinguistic norms which incorporated their
prior verbal repertoire. In this case the repertoire consisted of a ver-
nacular (Loez, a variety of Romance) and a set of sacred languages
(Hebrew-Aramaic). However, the pre-existing sociolinguistic norms did
not impinge upon the newly acquired Germanic code in either a random
fashion or on an equal-sampling basis. Quite the contrary. Both the
Romance and the Hebraic-Aramaic elements in Yiddish were over-
whelmingly retained to deal with a specific domain: traditional religious
pursuits and concerns. The Christological overtones of many common
German words, for example lesen (to read) and segnen (to bless), were
strong enough to lead to the retention of more neutral words of
Romance origin (leyenen and bentshn) in their stead. Similarly, Hebrew
and Aramaic terms were retained not only for all traditional and sancti-
fied objects and ceremonies but also in doublets with certain Germanic
elements in order to provide contrastive emphases: bukh (book) vs.
seyfer (religious book, scholarly book); lerer (teacher) vs. melamed or
rebi (teacher of religious subjects), etc. Thus, Yiddish is a wonderful
example of how all languages in contact borrow from each other selec-
tively and of how this very selectivity is indicative of the primary
interests and emphases of the borrowers and the donors alike (for
examples pertaining to early Christianity see Knott 1956, Mohrman
1947, 1957). Indeed, M. Weinreich has conclusively demonstrated
(1953, 1967, etc.) that a language not only reflects the society of its
speakers but, conversely, that societal data per se is crucial if lan-
guage usage and change are to be understood.
Findling's work too (1969) is interpretable in this fashion, demonstra-
ting as it does that Spanish and English among Puerto Rican youngsters

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 351
and adults in the Greater New York Metropolitan area reflect different
psycho-social needs and conflicts. In word-association tasks Findling
found his subjects mentioning humans more frequently in English than
in Spanish and more frequently in the work and education domains
than in the home and neighborhood domains (Table 21a and 21b).
According to various previous studies in the area of personality theory,
the prevalence of human terms in such unstructured tasks is indicative
of 'need affiliation', that is, the need to be accepted into positive
relationships with others. Findling therefore maintains that the language
of Puerto Ricans in New York reveals this need to be stronger (because
less gratified) in English interactions and in Anglo-controlled domains
than in Spanish interactions and Puerto Rican controlled domains.
Knowing, or suspecting, as we do from other sources, that Puerto
Ricans in New York are struggling for acceptance in an Anglo-dom-
inated world, Findling's interpretations seem reasonable and intriguing
indeed.

TABLE 2 1 a
Analysis of Variance of Human Ratio
(Need Affilation) Scores (Findling 1969)

Source
of Variance Sum of Squares df Mean Square F F95 F99

Between subjects 19,573.09 31

Occupation (C) 110.73 1 110.73 .17 4.17 7.56
Error (b) 19,463.08 30 648.77

Within subjects 65,904.10 288

Language (A) 701.69 1 701.69 3.78» 4.17 7.56
Domain (B) 12,043.27 4 3,010.82 12.10b 2.44 3.47
AB ' 239.49 4 59.87 .48 2.44 3.47
AC 181.84 1 181.84 .98 4.17 7.56
BC 1,855.50 4 463.87 1.86 2.44 3.47
ABC 446.16 4 111.54 .89 2.44 3.47
Error (w) 50,436.15 270
Errorj (w) 5,571.17 30 185.71
Error2 (w) 29,851.83 120 248.77
Errors (w) 15,013.15 120 125.11

Total 85,477.19 319

a = p > .07
b = p > .01

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
352 Joshua A. Fishman
TABLE 2 1 b
Mean Need Affilatiort Ratio Scores by Language and Domain

Domain

Language Work Educa- Re- Neigh- Home Total

tion ligion borhood
English 33 24 20 17 14 22
Spanish 28 23 17 13 14 19
Total 30 23 18 15 14 20

8.4 Language Behavior and Societal Behavior; a Circular Process of

Mutual Creations

The difference between the language constraint view and the language
reflection view is related to the difference between being interested in
language as langue and language as parole. It is also related to the
difference between being interested in inter-cultural variation and being
interested in 'intra-societal variation. Obviously, the sociology of lan-
guage is more fully at home with the latter level of analysis, in both
cases, than with the former. However, the latter level too can be over-
stated, particularly if it is claimed that not only is language behavior a
complete index to social behavior, but, also, that it is nothing more
than an index of such behavior. While indices are merely passive,
language behavior is an active force as well as a reflective one. Language
behavior feeds back upon the social reality that it reflects and helps to
reinforce it (or to change it) in accord with the values and goals of
particular interlocutors.
When Weinreich relates that Yiddish (then Judeo-German) came to
be the vernacular of Rhineland Jewry because Jews and non-Jews on
the eastern shore of the Rhine shared open networks and because higher
status in these Jewish-Gentile networks also came to provide Jews with
higher status in their own closed networks, he is saying much more than
that language usage reflects social interaction. Of course, Judeo-German
was a reflection of the fact that Jews and Gentiles participated in
common open networks. However, Judeo-German also helped imple-
ment and reinforce these networks, and, thus, became a co-participant
in creating or preserving the social reality that it reflected. Similarly,
when Weinreich tells us that Judeo-German became increasingly more
indigeneously normed (and therefore increasingly more Yiddish and
less Judeo-German) he is referring to much more than a linguistic

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 353
reflection of the primacy of its closed networks for this Jewish com-
munity. He is also telling us that the uniquely Jewish aspects of Yiddish
(in phonology, lexicon and grammar) also helped foster the primacy
of Jewish closed networks for its speakers. As a result, Yiddish not only
reflected (as it does today) the cohesiveness and separateness of its
speakers, but it helped to preserve and to augment these characteristics
as well.
Thus, both unidirectional views are outgrowths of an artificial search
for independent variables and original causes. The original cause of
any societal behavior may well be of some interest but it is a historical
interest rather than a dynamic one with respect to life as it continues
round about us. If we can put aside the issue of 'what first caused what'
we are left with the fascinating process of ongoing and intertwined
conversation and interaction. In these processes language and societal
behavior are equal partners rather than one or the other of them being
'boss' and 'giving orders' to the other.

9.0 APPLIED SOCIOLOGY OF LANGUAGE

On of the wisest maxims that Kurt Lewin bequeathed to social psycho-
logy is that which claims that "nothing is as practical as a good theory".
In addition, social science theory is undoubtedly enriched by attempting
to cope with the real problems of the workaday world. Thus, if social
science theory is really any good (really powerful, really correct), it
should have relevance for practitioners whose work brings them into
contact with larger or smaller groups of human beings. Applied socio-
logy of Language attempts both to enrich the sociology of language and
to assist in the solution of societal language problems. The applied
sociology of language is of particular interest whenever: (a) language
varieties must be 'developed' in order to function in the vastly new
settings, role-relationships or purposes in which certain important net-
works of their speakers come to be involved, or (b) whenever important
networks of a speech community must be taught varieties (or varieties
in particular media or uses) that they do not know well (or at all) so
that these networks may function in the vastly new settings, role-
relationships or purposes that might then become open (or more open)
to them. In many instances (a) and (b) co-occur, that is, language varie-
ties must be both developed and taught in order that important networks
within a speech community may be fruitfully involved in the new
settings, role-relationships and purposes that have become available to
them. This is but another way of saying that planned language change
and planned social change are highly interrelated activities and that the
sociology of language is pertinent to their interaction.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
354 Joshua A. Fishman
Comments on the uses of sociology of language must keep in mind
four separate categories of actual and potential users, namely, linguists
and sociologists on the one hand, and the users of linguistics as well as
the users of sociology on the other hand. The sociology of language as
a hybrid or bridge-building specialization, is useful not only as it per-
tains to the front line of contact between science and society but also
as it enables those in theoretical heartlands to understand their basic
fields afresh and in refreshing ways. Application and applicability are
themselves an endless array of concentric circles that surround all
immediate problems in an ever-widening and interlocking flow. It is
never wise to rigidly declare some knowledge 'useful' and other 'use-
less', for neither knowledge nor usefulness (nor even the very problems
to which both are referred) hold still long enough for such judgments to
be more than myopic indicators of how near or far we stand with
respect to a particular and often fleeting goal. All knowledge is useful,
and if at any point in time we nevertheless grope toward a consideration
of the 'uses of X', it is merely because for some particular purposes at
some particular time some knowledge may seem more useful than
others.
This section proceeds by reviewing a few recognized topics within
applied linguistics in order to illustrate and document a point of view
with respect to the usefulness of the sociology of language. Its point
of departure is Charles A. Ferguson's well known attempt to divide
applied linguistics into its six most common American branches: the
creation and revision of writing systems, literacy efforts, translation
work, language teaching efforts, and language policy efforts (Ferguson
1959; for a German and a Soviet view see Kandler 1955 and Andreev
and Zinder 1959). Although it will be impossible to give equally de-
tailed attention to all five of these branches of applied linguistics here,
it would seem that essentially similar questions must be addressed to
each of them; namely, what has been accomplished without formal
sociolinguistic awareness and sophistication? what has the sociology
of language contributed to more recent applied linguistic efforts in
these topical areas? finally, what more could the sociology of lan-
guage contribute to these (and even to other) applied linguistic concerns
if its practitioners were to really take both parts of this hybrid field
with equal seriousness and with the deep technical and theoretical profi-
ciency that they both require.

9.1 Creation of Writing Systems

The sophistication of phonological theory, both that of the early part
of this century as well as that of very recent years, and the recent

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 355
linguistic interest in theories of writing systems and in the relations
between such systems and spoken language are, and have long been,
powerful linguistic contributions to the worldwide efforts to create
writing systems for pre-literate peoples. However, the very sophisti-
cation of the linguist's professional skills in code description and code
creation merely intensified the separation trauma when it became ob-
vious that it was necessary to go outside the code and to confront the
real world if writing systems were not only to be devised (this being the
only apparent concern of Pike 1947 or Ray 1963) but also employed.
The first steps in this direction were moderate indeed. These consisted
of Vachek's (1945/49 and 1948) and Bolinger's (1946) protests (among
others) that the writing system must be viewed separately from the
spoken code, i.e., that it could not properly be viewed as merely the
phonetic transcription of the spoken code, and that it was basically a
"visual system" (being not unlike the language interaction of the deaf
in this respect) with regularities all its own.
The reverberations of these early protests are still with us. As Berry
has pointed out (1958), new alphabets have clearly become less purely
phonemic and more inclined to the "use of reason and expedience"
(rather than to rely on phonemicization alone) in their pursuit of
acceptance. Indeed, the latter concern, that of acceptance, has tended to
replace the former, that of "reduction to writing", and, as a result,
arguments pertaining to intra-(written) code phenomena have tended
to recede evermore into the background. While "phonetic ambiguity"
is still considered a "bad" thing and while it is generally agreed that
"words pronounced differently should be kept graphically apart" (Brad-
ley 1913/14) it is considered to be an even "worse thing" if alphabets
of exquisite perfection remain unused or unaccepted. More and more
work on the creation of writing systems has shown awareness of the fact
that such nonacceptance is only to a relatively minor degree governed
by intra-code ambiguities, inconsistencies or irrelevancies (all of these
being rampant characteristics of the most widely used writing systems
today and throughout history). Time and again in recent years the
greater importance of extra-code phenomena has been hinted at (Gelb
1952, Bowers 1968), pointed to (Sjoberg 1964, 1966, Walker 1969)
and, finally, even listed and catalogued (Nida 1954, Smalley 1964).

9.2 Desired Similarity and Dissimilarity

Perhaps because their attention is basically directed toward intra-code
factors, linguists and applied linguists were quickest to notice those
extra-code factors in the adoption or rejection of writing systems which
indicated societal preferences or antipathies for writing conventions

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
356 Joshua A. Fishman
associated with some other language or languages. Thus, among the
"practical limitations to a phonemic orthography" Nida (1954) dis-
cussed the fact that both the Otomi and the Qeuchua "suffer from
cultural insecurity" and want their writing systems not only to "look
like Spanish" but to operate with the same graphematic alternances as
does Spanish, whether these are needed or not in terms of their own
phonemic system. In a related but crucially different vein Hans Wolff
recommended (1954) that Nigerian orthographies be created not only
in terms of tried and true technically linguistic criteria (such as "accu-
racy, economy and consistency") but that "similarity to the ortho-
graphies of related languages" also be used as a guide. Of course,
Wolff was merely following in the footsteps of the Westermann Script
of the late 20's, which, in its fuller, more generally applicable form,
became the All-Africa Script of the International African Institute
(Anon. 1930). However, he was also following in the tradition that
placed the linguist or other outside expert in the position of judging not
only which languages were sufficiently related in order to deserve a
common writing system, but that placed them in the position of deciding
whether such similarity in writing systems was or was not a 'good thing'
and whether it was or was not desired by the speech communities in-
volved.
However, once having stepped outside of the charmingly closed
circle of intra-code considerations, Pandora's box had been opened
never again to be shut. In very recent days, to mention only such
examples, Serdyuchenko has assured us that the Cyrillic alphabet is
used as the model in "the creation of new written languages in the
USSR" only because of the widespread and still growing interest in
subsequently more easily learning Russian, just as Sjoberg (1966)
mentions Tlingit insistence that their orthography "follow the rather
chaotic orthographic patterns of English wherever possible in order to
conform to the demands of the broader society (p. 217)", and the
Institut Français d'Afrique Noire concludes that speakers of African
vernaculars in Francophone countries want their orthographies to look
as French as possible (Smalley 1964). Walker (1969), like Serdyuchenko
before him (1962), is quite willing to champion such modeling at the
explicit expense of maximal phonemic efficiency. Recently the Bamako
Meeting on the Use of the Mother Tongue for Literacy (February 28
- March 5, 1966, UNESCO sponsored) went a step further. It not only
recommended that new writing systems be similar to those of unrelated
but important languages for the learners (Bowers 1968), but it also
warned of "possible repercussions of a technical and economic nature"
following upon the adoption of Non-European diacritics and special
letters in the standard transcriptions of West African languages (Ferru

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 357
1966). Such letters and diacritics, it is pointed out, increase the cost
of printing and typing, as well as the cost of manufacturing printing and
typing equipment, and do so at the time when the per capita cost
of printed or typed material is already likely to be troublesomely high
in view of the limited number of consumers available for them in
newly literate societies. On these same grounds the Institut d'Afrique
Noire insisted as far back as 1959 that "when symbols have to be made
up they should be typable on a standard French typewriter" (Smalley
1959).
The obverse case has been less fully documented, namely, that in
which newly literate communities have desired a more distinctive
writing system, one that they could call their own or one that would
more effectively differentiate their language from others with respect to
which they sought not similarity but rather dissimilarity. Dickens'
(1953) discussion of the Ashante rejection of the Akuapem-based
writing system for standard Twi (in the late 30's and early 40's) is one
such case. Another is Ferguson's brief reference to the fact that St.
Stefan of Perm (14th century) purposely created a separate alphabet
for the Komi (giving "some of the letters an appearance suggestive of
the Tamga signs in use among the Komi as property markers and deco-
rations" 1967, p. 259) "so that the Komi could regard the writing
system as distinctively theirs and not an alphabet used for another
language." There must be many examples of this kind, e.g., St. Mesrop's
creation of the Armenian alphabet in the fifth century, utilizing in part
characters like those of far-distant brother-Menophysite Christians in
Ethiopia with whom contact had probably been made (according to
Olderogge) as a result of the presence of both Armenian and Ethiopian
churches in Jerusalem. Another such example is Sequoyah's syllabary
which was "not associated with aliens but developed within the Chero-
kee language community itself" (Walker 1969, p. 149; also see White
1962). Finally, to the above cases there must be added the few prelim-
inary studies of indigenous African and Asian scripts of relatively small
communities that weathered competitive pressures precisely because of
their real or assumed local origins (e.g., Dalby 1967, 1968, Hair 1963,
Stern 1968, Stewart 1967). Perhaps the relative reluctance to docu-
ment such cases is not unrelated to the more general reluctance of
those who practice applied linguistics upon others to recognize the
frequent desires of non-literate peoples to be themselves (albeit "in a
modern way"), rather than merely to be imitative copies of ourselves
(whether we be"" Chinese, Russian, Arab, French, British, American,
Spanish, or Portuguese).

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
358 Joshua A. Fishman

9.3 'A Little More Complicated Than That'

If economics answers all questions with 'supply and demand', psycho-

logy with 'stimulus and response' and education with 'it all depends',
then the first contribution of the sociology of language to applied lin-
guistics is doubtlessly to stress the fact that the relations and inter-
penetrations between language and society are 'a little more complicated
than that', whatever that may be. Indeed, although it is nearly half a
century since Radin first implied that the adoption (actually, the bor-
rowing) of an alphabet by an aboriginal people was a fascinatingly
complex and internally differentiated chain of social processes, we
have not to this very day seriously followed up this seeming complexity,
let alone tried to reduce it to some underlying set of basic dimensions.
Our technical expertise and theoretical sophistication lead us more
readily to agree with Burns' (1953) early conclusion, based on sad
experience with the failure of "linguistics without sociology" in Haiti,
that the choice of an orthography has widespread social and political
implications. They also lead us to continually admire Garvin's accounts
(1954, also see 1959) of his attempts to achieve consumer consensus
and participation in the creation of a standard orthography for Ponape,
and to share his disappointment that even this was not enough to assure
the use of that orthography. Beyond such agreement and admiration,
however, we can only suggest that the process of gaining acceptance
for technically sound writing systems is even 'a little more complicated
than that'. In spelling out this complexity applied sociology of lan-
guage uniquely stresses that it is crucial to systematically look outside
of the linguistic system itself if one is to locate the reasons for the
differential acceptance or rejection of programs of linguistic change.
Modern sociology of language can contribute most by linking this
particular topic of applied linguistics with the body of theory and
practice that has grown up in connection with the acceptance of other
systematic innovations, the planning of social change more generally,
and the amelioration of the inevitable dislocations that follow upon
the introduction of innumerable innovations and changes of which
new writing systems are merely symptomatic.
The creation of writing systems is itself necessarily an outgrowth
of culture contact, if not of political and economic domination from
outside. Thus, the creation of a writing system is singularly unlikely to
be viewed dispassionately and its propagation and acceptance by indi-
genous networks are necessarily viewed as having implications for group
loyalty and group identity. Latinization, Arabization, Cyrillization or
Sinoization are not merely fargoing indications of desired (and fre-
quently of subsidized or directed) social change and cognitive-emotional

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 359
reorganization, but they have immediate consequences for the relevance
of traditional elitist skills and implications for the distribution of new
skills and statuses related to literacy and to the philosophy or ideology
which is the carrier of literacy.
The creation of writing systems is significant only insofar as it leads
to the acceptance and implementation of writing systems. The latter
are broadly revolutionary rather than narrowly technical acts. They
succeed or fail far less on the basis of the adequacy of their intra-code
phonological systems or on the basis of their fidelity to model systems,
than on the basis of the success of the larger revolutions with which
they are associated; revolutions in the production and consumption of
economic goods (leading to new rural-urban population distributions,
new jobs, new training programs, new avocations, new pastimes, and
new purposive social groups) and revolutions in the distributions of
power and influence. All of these both lead to and depend upon an
increasing number of new texts and new written records. Thus, when
sociolinguistic attention is finally directed to the creation of writing
systems it will be focused upon the organization, functioning and
disorganization of an increasingly literate society. This is potentially
a very useful addition to the linguist's disciplinary focus because even
more than writing changes speech (via 'spelling pronunciations') literacy
changes speakers and societies. It is this perspective on the creation of
writing systems - as always, a perspective which is outside of the
linguistic system alone - that is part of the programmatic promise of
the sociology of writing systems.
How will such attention improve or alter the creation of writing
systems? Precisely by relating the problem of creation to the problem
of acceptance, of impact, of possible dislocation, of possible manipu-
lation, of possible exploitation, of possible redistribution of power and,
in general, of the dependency of the very best writing system on
revolutionary processes at their most pragmatic as well as at their most
symbolic.

9.4 Orthographic Reform

To some extent such liberation and immersion are more advanced with
respect to the study and planning of orthographic reforms, perhaps
because the truly vast amount of technical linguistic effort invested in
these reforms has yielded such meager results. Even though ortho-
graphic reform may be so sweeping as to involve the complete replace-
ment of one writing system by another (and, in that sense, it may be
viewed as a subcategory of the topic just reviewed), it deals with already

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
360 Joshua A. Fishman
literate networks and as a result, more clearly reveals the societal
ramifications and reverberations of seemingly technical linguistic ad-
justments.
If the introduction of a newly created writing system easily threatens
to change established lines of relative advantage and disadvantage,
practical and symbolic, the revision of traditional orthographies most
often obviously attempts to do so. Orthographic change represents
def '.ture from an established written tradition and, as such, it must
cope with the gatekeepers of that written tradition, the poets, priests,
principals and professors, and the institutions and symbols that they
create and serve, or be destined to oblivion. Indeed, the greater and
grander the tradition of literacy, literature and liturgy in an ortho-
graphic community, the less likely that even minor systematic ortho-
graphic change will be freely accepted and the less likely that any
orthographic change will be considered minor.
In this connection we have a larger number of rather detailed and,
to some extent, sociolinguistically oriented descriptions, than is the
case for the creation of writing systems, but, as yet we have no socio-
logical analyses or hypotheses per se. The soci-culturally contextualized
descriptions of orthographic reforms in the USSR (Kolarz 1967, Oren-
stein 1959b, Quelquejay and Bennigsen 1961, Serdyuchenko 1965,
Weinreich 1953, Winner 1952), Turkey (Rossi 1927, 1929, 1935, 1942,
1953, Heyd 1954, Ozmen 1967, Gallagher 1967 and 1969), Norway
(Haugen 1966a, which contains an exhaustive bibliography of other
studies) and Vietnam (Haudricourt 1943, Nguyen dinh Hoa 1960,
Thompson 1965) again point to the literally revolutionary nature of
the societal processes that have often accompanied system-wide ortho-
graphic change. On the other hand, the available descriptions of "ar
less successful attempts to bring about orthographic change under less
dramatic circumstances, e.g., in Japan (DeFrancis 1947, Holton 1947,
Meyenburg 1934, Scharshmidt 1924, Toshio 1967), Haiti (Valdman
1968, Burns 1953) and Israel (Rabin 1969), or to bring about the
orthographic unification of closely related languages in the absence of
accompanying societal unification, e.g., in India (Anon 1963, Jones
1942, Ray 1963), Africa (Dickens 1953, Ward 1945) and Indonesia-
Malaysia (Alisjahbana 1969 and in press), all indicate the difficulties
encountered and the failures experienced thus far.
However, there is no justification for interpreting the above cited
investigations as implying 'revolutionary success and non-revolutionary
failure' as the proper summation of experience with orthographic
reform. In earlier centuries a great deal of orthographic reform seems
to have been accomplished both quietly and successfully without the
involvement of mobilized populations or, indeed, of any other popula-

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 361
tion segments than 'the authorities' whose business it was to make wise
decisions for the community. The initial orthographic distinctions
between Serbian and Croatian, or between Ruthenian (Ukrainian) and
Polish, were decided upon by representatives of God and/or Caesar
who sought to cultivate differences between speech communities that
were otherwise 'in danger' of religious, political and linguistic unifica-
tion. Indeed, the Ausbau languages (in Kloss' sense, 1952) are all
instances of the success of applied linguistics and should be carefully
studied as such. The restoration of written Czech (and Slovak) in
Latin script was engineered by Count Sedlnitzsky, the administrative
director of the Austro-Hungarian police and one of the most influential
officials under the Emperor Francis (early 19th century), by subsidizing
the publication of the Orthodox prayer book in Latin letters as "an
important device to fight the political danger of the Pro-Russian Pan-
Slav movement (Fischel 1919, p. 57). The Roumanian shift from
Cyrillic to Latin script in 1863 was accomplished by a painless edict
which sought to further that nation's self-defined Latinizing and Chris-
tianizing role in the heathen 'Slavo-Moslem' Balkans (Kolarz 1946). In
more recent days Irish orthography has been changed without arousing
unusual interest or opposition (Macnamara 1969), as was the type font
(from an 'Irish looking' font to an ordinary roman font). Indeed, the
relative ease with which these changes were made may be a reflection
of the lack of widespread Irish interest or concern for the Language
Revival.
Not only has there been much successful orthographic reform without
revolutionary change (particularly where mass mobilization along lan-
guage-related lines was absent for one reason or another) but there
has also been a good bit of unsuccessful orthographic reform even when
these have been accompanied by revolutionary social changes. Thus,
the Soviet 'rationalization' of Yiddish orthography (Szajkowski 1966)
initially aimed at both the phonetization of words of Hebrew-Aramaic
origin, as well as at the discontinuation of the socalled final letters of
the traditional Hebrew alphabet. However, twenty five, years after the
October revolution, the names of the grandfathers of Modern Yiddish
literature were neither spelled p i p S " >0"HXDD n ^ m a S ^ S H E ? »

-
a s ^ x Ti p n x x s n y s wrr1?
(as they had been throughout the 20's and 30's), nor were they spelled
d d ^ s ? m ^ w n x p s cnn^ 1 ? p n r , n n s o i D i n ^ s n r a a
(as they had been before the Revolution and continued to be everywhere
outside of the Soviet Union), but, rather, in an attempt to reach a com-
promise that would maximize the propaganda value of the few permit-
ted Yiddish publications primarily distributed to and published for

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
362 Joshua A. Fishman
readers outside of the USSR: p i O ^ " ,D"HXDO "laD^TD 17*7S7"r2S7»
' T T

a s n ^ x DS^X® p x p n a s tm"1?
However, even in its heyday the Soviet revolution in Yiddish ortho-
graphy could not overcome the visual traditions of the orthographic
community. The initial silent aleph at the beginning of words that would
otherwise begin with the vowels and 1 was never dropped, regardless
of its phonemic uselessness, perhaps because the initial silent aleph in
such cases was considered to be too strong a visual convention to be
tampered with (Hebrew writing itself - i.e., the visual precursor to
written Yiddish - never beginning words with vocalic ^ or 1).
A far more widely renowned revolutionary attempt at orthographic
reform which has failed (certainly thus far) is the once promised phone-
tization of (Northern Mandarin) Chinese. While the basic sources
available to us in English (DeFrancis 1950 and 1968, Mills 1956 and
Hsia 1956) all agree that the Latinized New Writing was abandoned
sometime late in the fifties, the reason for this abandonment can
still only be surmised.
By 1956 it had become necessary to defend the "Han (Chinese)
language phonetization draft plan" as being concerned with an alphabet
(Latin) which was truly progressive and international rather than neces-
sarily related to any anti-proletarian class (Chinese Writers Language
Reform Committee 1956, Wu Yu-Chang 1956). By 1959 Chou En-
lai had officially demoted phonetization from its original goal of imme-
diate "liberation and development of the whole Chinese language from
the shackles of the monosyllabic Chinese characters" (Ni Hai-shu 1949,
cited by DeFrancis 1968) to third place and the indefinite future, after
both simplification of the traditional characters and adoption of a
spoken standard for "Common Speech" had been attained (Chou En-
lai 1965). While work on the first two tasks is constantly going on
in a very direct fashion (see, e.g., Wu Yu-Chang 1965) work on the
latter is primarily nominal (that is, phonetization is kept alive as a
distant goal but is not substantively advanced) and indirect (i.e.,
phonetization is utilized for subsidiary purposes, such as annotating
novel or complex Chinese characters in technical texts, furthering
instruction in the Common Speech among speakers of other regional
languages, or creating "initial alphabetic scripts" for illiterate non-
Chinese speaking minorities). Indeed, while phonetization has recently
been reported to be superior for such special purposes as telegraphic
communication (Wu Yu-Chang 1964) and minority group initial literacy
(Li Hui 1960) the traditional characters have again been proclaimed
as superior in connection with general education for the bulk of the
population among whom these characters are viewed as symbolic of

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 363
education and the standard pronunciation (Serruys 1962)! The goal of
phonetization is, seemingly, still a long way off and may or may not
be reached any more rapidly than the withering away of the state (See
several references to this effect in Kwan-wai Chiu 1970).
From the foregoing examples it is clear that if we but dichotomize
both 'success' (acceptance) and 'revolutionary social change' we have
examples of all four possible types of co-occurrences: successful ortho-
graphic revision with and without revolutionary social change and revo-
lutionary social change with and without successful follow-through of
planned orthographic revision. The discussions of revolutionary social
change thus far encountered in studies of either the creation of writing
systems or the revision of orthographies is still far too crude to be
considered as more than rough labeling. As sociolinguistic description it
is regrettably out of touch with the sizable modernization literature in
qioq S5JDBJ ii ASojodojqjuB pue A3o[opos 'aouops jeoijijod 'soiuiouooa
the concepts and the technical data collection methods and data ana-
lysis skills needed to inquire into the intensity, extensity or continuity
of the change forces and processes or the counter-change forces
and processes that underly the gross labels so frequently encountered.
It is also unfortunate that there are so few localized case studies of
variation in sub-group reactions to new writing systems or to revised
orthographies and, conversely, proportionally so many commentaries,
studies, evaluations, and recommendations that deal with entire coun-
tries, continents, and even the world at large. The result is an imbalance
with respect to the usual mutual stimulation between micro-analysis
and its emphases on process and function concerns, on the one hand,
and macro-analysis and its emphases on structure, quantification, com-
positing and weighing of parameters on the other hand. Either type
of study, when pursued too long without correction from the other,
becomes myopic and, therefore dangerous for theory as well as (or
even more so) for application. However, whereas both macro and
micro-studies are equally necessary for the growth of general sociolin-
guistic theory, the future of applied sociology of language is particularly
tied to within-context studies (within nation X, within region Y, within
district Z) and, therefore, to ever more detailed studies of differential
acceptance processes, rather than to studies of large-scale between-
context variation (the latter not having as immediate applied significance
for any particular within-context problem).

9.5 Language Planning

Perhaps the area of applied linguistics which most clearly illustrates the
full complexity of societal phenomena which the sociology of language

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
364 Joshua A. Fishman

Les grandes enquêtes de l'Office

du Vocabulaire français de Paris
Depuis quelques années, l'opinion publique est alertée sur les dangers que
court la langue française, et divers organismes, tant officiels que privés, ont
été créés pour assurer la défense d'un idiome parlé par la communauté sans
cesse grandissante des francophones: 90 millions aujourd'hui, ils seront
200 millions avant l'an 2000. Mais comment obtenir que la langue fran-
çaise, ciment de cette francophonie dont on parle tant, ne s'altère pas, ne
dégénère pas, ne devienne pas un jargon à la syntaxe désordonnée et au
vocabulaire truffé de néologismes mal assimilés?
Les règles du bon langage, on peut les connaître. Les atteintes au bon
langage, tout le monde les connaît également, ou peut les connaître à la
lecture de maints ouvrages ou articles de presse. Ce que l'on connaît mal,
c'est le moyen d'obtenir que les règles soient mieux respectées et l'invasion
des néologismes contenue. L'opinion sait que l'usage est à corriger selon
les principes d'une norme raisonnable. Mais elle ignore comment la norme,
une fois établie, peut descendre jusqu'à l'usage, c'est-à-dire comment l'usage
peut être orienté, sinon dirigé par la norme de manière efficace.
Précisément, la troisième Biennale de la langue française a inscrit à l'ordre
du jour de ses prochains travaux une formule significative d'enquête: De
la norme à l'usage.
L'Office du Vocabulaire français, qui, on le sait, est à l'origine des Bien-
nales, entend participer à cette recherche et fournir aux orateurs de la
Biennale qui désireront s'informer sur l'opinion du grand public éclairé
une riche documentatior Celle-ci sera puisée à la meilleure source qui soit:
le groupe que constituent les membres consultants de l'Office du Voca-
bulaire français.
C'est pourquoi nous vous demandons de bien vouloir répondre au question-
naire que voici:
1. Pensez-vous, d'une manière très générale, qu'il soit possible d'agir sur
les habitudes de langage d'une grande communauté humaine?
N.B. — A cette première question, les membres de l'Office du Voca-
bulaire français répondront, de toute évidence, par un « oui ». Mais
toute personne qui répondra « non » devra développer ses raisons, qui
pourront être constructives.
2. Pensez-vous que l'Enseignement soit le seul dispensateur de la norme
et que nulle action sur l'usage ne soit concevable hors de l'école?
3. Croyez-vous que les adultes puissent recevoir un enseignement prolongé
de la langue française, de la même manière qu'ils reçoivent, par les
publications spécialisées, par les revues de vulgarisation scientifique ou
technique, un enseignement prolongé en histoire, en physique, en histoire
naturelle, en géographie . . . et même en àstronautique?
Nous vous serions reconnaissants de bien vouloir répondre à ces questions,
en portant en tête de votre lettre vos nom et prénom, profession, adresse.
Vous voudrez bien ajouter, également en tête de votre réponse, selon votre
choix:
J'accepte que mon nom figure dans un compte rendu de synthèse (signa-
ture);
Je désire garder l'anonymat (id.).
Les réponses devront être adressées au secrétariat de

l'Office du Vocabulaire français, 17,

rue de Montparnasse, Paris-Vie, France
Brought to you by | The University of Texas at Austin
Figure 24. From Le Travailleur (Worchester, Mass.), June Authenticated
7, 1969
Download Date | 2/16/20 10:45 AM
The Sociology of Language 365
may someday enable us to understand is that which is concerned with
language planning. Just as sociolinguistic inquiry into the creation of
writing systems and into the revision of orthographies permits us to first
recognize and to then refine our appreciation of the magnitudes of
social change and social planning (if not social dislocation) with which
such activities are commonly associated, so the systematic sociolinguistic
study of language planning as a whole (incorporating the creation of
writing systems and the revision of orthographies, but going beyond
them to conscious governmental efforts to manipulate both the structure
and the functional allocation of codes within a polity) enables us to
appreciate the societal complexity impinging on the determination,
implementation and evaluation of language policy as a whole. The
study of language planning is the study of organized efforts to find solu-
tions to societal language problems (Jernudd and Das Gupta 1969).
As such, it is necessarily most dependent - of all the fields of applied
language concerns - on the sociology of language and on the social
sciences as a whole in order to move from theory to informed practice
(Figure 24).
Of the language planning studies recently completed or currently
underway a few have dealt with the cost-benefit analysis of alternative
or hypothetically alternative decisions between which governmental or
other bodies must choose (Jernudd 1969, Thurburn 1969). Others have
discussed the pressure functions focused upon decision making/decision
implementing bodies in the language field (whether the latter be legis-
lative-executive within government or political-religious-literary-acade-
mic outside of government) from a variety of special interest groups
running the gamut from professional associations of educators, to
manufacturers of typewriters and publishers of textbooks, to spokesmen
for literary, journalistic and ideological groupings, etc. (Das Gupta
1969). There are now several theoretical models (happily commen-
surable) of the interaction of sentimental and instrumental integrative
and disintegrative forces in the language planning process (Kelmann
1969, Fishman 1971). There are recent critiques and integrations of
the literature on the evaluation of planned change in education, industry,
agriculture and other areas of conscious societal planning, in an effort
to suggest evaluative methods that might be most fruitfully adopted
for the evaluation of success or failure in language planning (Macna-
mara 1971, Rubin 1971). A four-country study has recently gotten
underway (involving linguists, anthropologists, political scientists, socio-
logists, psychologists and educationists) in order to obtain roughly com-
parable data concerning the processes of language planning per se in
each of the above context (decision making, pressure functions, national
integration, implementation and evaluation). Obviously, the study of

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
366 Joshua A. Fishman
language planning is rapidly moving away from intra-code efficiency
considerations alone (the latter being the primary emphasis of Tauli
1968) and moving steadily into ever-richer contextual concerns. Hope-
fully, as language planning and social planning agencies become more
aware of the possible contributions of applied sociology of language
they may become more inclined to involve sociolinguists and other
language specialists in guiding the decision making process itself rather
than merely in implementing decisions already reached. Several signs
already point in this direction. Thus, the several nations of East Africa
are interested in the current "Survey of Language Use and Language
Teaching" (Prator 1967) in order to adopt (or revise) language opera-
tions in schools, mass media, public services, etc., on the basis of more
precise information as to the age, number, location and interactions of
the speakers of various local languages. Similarly, the Philippine govern-
ment has long followed a policy of evaluating language policy in the
area of education via research projects dealing with such matters as
the advisability of initiating education in the local mother tongues
and introducing the national language (Filipino) only in some optimal
subsequent year (Ramos et al. 1967). The Irish government has spon-
sored "motivation research" and opinion polls in order to determine
how its citizens view the Irish language and how they react to the
government's efforts to "restore" it to wider functions (Anon. 1968).
One of the most widely cited guides to governmental language policies
and their educational implications is an applied sociolinguistic report
issued by Unesco and dealing with "The Use of the Vernacular in
Education" (Anon. 1953). Once a policy is adopted it is then necessary
to implement it. Such implementation not only takes the obvious route
of requiring and/or encouraging the functional re-allocation of varieties
but also their phonological, lexical and grammatical realization along
prescribed lines. Language agencies, institutes, academies or boards
are commonly authorized to develop or plan the variety selected by
policymakers. Such agencies are increasingly likely to seek feedback
concerning the effectiveness or the acceptability of the 'products'
(orthographies, dictionaries, grammars, spellers, textbooks, translation
series, subsidized literary works, etc.) that they have produced. Socio-
logists of language have already produced many studies which language
agencies are likely to find extremely useful in terms of their implications
for the work that such agencies conduct.
The difficulties encountered and the lessons learned in planned
lexical expansion to cope with the terminology of modern technology,
education, government and daily life are recounted by Alisjahbana
(1962, 1965, 1971), Bacon (1966), Morag (1959), Passin (1963) and
Tietze (1962) in their accounts of language planning in Indonesia,

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 367
Central Asia, Japan, Israel and Turkey, respectively. The problems of
planned language standardization have been illuminated by Ferguson
(1968), Garvin (1959), Guxman (1960), Ray (1963), U. Weinreich
(1953b), Havranek (1964), Valdman (1968) and Twaddell (1959) in suf-
ficiently general terms to be of interest in any speech community where
this process needs to be set in motion.
Even the very process of government involvement in language issues
has begun to be documented. In this connection one must mention the
reports of the Irish government on its efforts to restore the Irish lan-
guage (Anon. 1965); Goodman's review of Soviet efforts to provide
- as well as deny - indigenous standard languages to the peoples under
their control (Goodman 1960); Haugen's many insightful reports of the
Norwegian government's attempts to cope with language conflict by
both protecting and limiting the linguistic divergence of its citizenry
(Haugen 1961, 1966a, 1966b); Heyd's account of language reform in
modern Turkey (Heyd 1954); Lunt's account of the studied efforts in
Titoist Yugoslavia to separate Macedonian from Serbian and from
Bulgarian (Lunt 1959) and Mayner's comments on the attempts to fuse
Serbian and Croatian in that country (Mayner 1967); the contrasts
between different parts of Africa noted by Mazrui (1967), Armstrong
(1968), Polome (1968) and Whiteley (1968); Mills' report of how Com-
munist China advanced and retreated in connection with the writing
reform it so desperately needs (Mills 1956); Wurm's descriptions of the
very beginnings of language policy in reference to Pidgin English ("Neo-
melanesian") in New Guinea (Wurm and Laycock 1961/62), and
several others (e.g., Brosnahan 1963b; LePage 1964, Fishman 1968c)
of more general or conceptual relevance.
One of the most necessary areas of applied sociology of language is
that which deals with educational problems, related to language policy
formulation or evaluation. In this connection there have been studies of
the organization and operation of bilingual schools (Gaarder 1967); of
the academic consequences of compulsory education via the weaker
language for most learners (Macnamara 1966, 1967); of different
approaches to teaching hitherto untaught mother tongues (Davis 1967);
of varying South American and West Indian approaches of teaching
both local and 'wider' languages (Burns 1968, LePage 1968, Rubin
1968); of difficulties in teaching English (as the compulsory school
language for non-English speakers) encountered by teachers who are
themselves non-native speakers of English (Lanham 1965); and, more
specifically, of the problem of teaching standard English to speakers of
very discrepant, non-standard varieties of that language (Stewart 1964,
1965). A more generalized interest in applied sociology of language is
that shown by the recent Canadian Royal Commission on Bilingualism

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
368 Joshua A. Fishman
and Biculturism (Royal Commission 1965, 1967, 1968). It authorized
studies not only on bilingual schooling but also on bilingualism in
broadcasting, in industrial operations, in military operations and in the
operation of various other societal enterprises.
Notwithstanding the obvious recent strengthening of applied socio-
logy of language, several nations throughout the world are currently en-
gaged in language planning without anything like the information avail-
able to them in other areas of planning. Sociolinguistic research on lan-
guage planning must aim, first, to locate, then, to apportion the varia-
tion in behavior-toward-language which is to be observed in language
planning contexts. It must seek detailed knowledge of how ortho-
graphic decisions (or script decisions, or national language decisions,
or nomenclature decisions, etc.) are arrived at, how they are differen-
tially reacted to or followed up by agencies inside and outside govern-
ment, how they are differentially accepted or resisted by various popu-
lation segments, how they are differentially evaluated and how sub-
sequent policies and plans are differentially modified as a result of
feedback from prior policy and planning. The sociology of language is
just now beginning to describe the variation that constantly obtains in
all of these connections. After this has been done sufficiently well and
in sufficiently many contexts it should begin to successfully account for
this variation and, at that point, be able to offer suggestions that are
useful from the point of view of those seeking to influence, implement
or evaluate language planning in the future.

9.6 Some Straws in the Wind

However, even in the absence of the amount of detail and sophistication
that is needed before practical information becomes available, 'sociolin-
guistically motivated' changes in applied linguistics are clearly on the
increase. Not only are such topics as the creation of writing systems,
the reform of orthographies and language planning more generally
marked (as we have seen) by a constantly increasing awareness of
societal interpénétration and of the need for truly professional com-
petence (which is more than simply being either critical or admiring)
if one is to understand, let alone influence, the societal forces at work,
but such awareness is growing in most other fields of applied linguistics
as well.
The planning, implementation and evaluation of literacy campaigns
increasingly ceased being merely applied linguistics plus education
(pedagogy) plus ethnography as the period of immediate post World
War II exuberance was left behind (Smith 1956). What is currently
being developed in this field goes beyond advice on how to establish

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 369
proper local contacts and obtain official cooperation (Young 1944,
Russell 1948), important though such advice undoubtedly is. It goes
beyond care to adapt programs to local needs (Jeffries 1958), to utilize
a variety of methods on a variety of fronts (Ivanova and Voskresensky
1959), or to evaluate outcomes broadly enough to include health, eco-
nomic and other pertinent indices (UNESCO 1951). Current efforts
to advance literacy are increasingly based upon efforts to more fully
understand the meaning and impact of literacy via small pilot studies
which seek to recognize and weigh alternatives (Correa and Tinbergen
1962, Lewis 1961, McClusker 1963) and clarify the societal dimensions
of literacy enterprises in different context (Goody 1963, Hayes 1965,
Nida 1967, Schofield 1968, Wurm 1966).
A similar systematic intrusion of societal considerations has become
noticeable in the field of translation. It is here, in particular, that socio-
linguistic differentiation of language into varieties and of speech com-
munities into situations is beginning to be felt, perhaps more so than
in any other field of applied linguistics. One cannot read Catford's
Linguistic Theory of Translation (1965) without being delighted by the
fact it is far broader than 'immaculate linguistics' alone, and one
cannot read Wonderly's Bible Translations for Popular Use (1968)
without wishing that its sensitivity to social varieties and social occa-
sions were part of the professional orientation of translation for far
more worldly purposes as well. Certainly the deep concern with re-
cognizing the significance of functional variation in language variety
use, the sensitivity shown with respect to the situational analysis of
repertoires of social and linguistic behavior - viewing Bible reading and
listening as kinds of situations that may require particular kinds of
language - and the repeated attention given to the contextual-functional
differences between written and spoken language (and the multiple va-
rieties of each) must sooner or later feed back into religious work on
the creation of writing systems and on literacy more generally. This is,
indeed, the beginning of technical sociolinguistic utility for an applied
field. Having once embarked along the path or recognizing that all of
the factors influencing communicative appropriateness in a particular
speech community also influence the acceptability and the impact of
translations in that community, the probability of mutual enrichment
between application and theory for both fields of endeavor (translation
and sociology of language) is indeed very great.
The same may yet be the case for the huge field of language
teaching, where the contacts with the sociology of language are still
far more tenous, if only because the contacts between an elephant and
a sparrow must always be rather incomplete. Nevertheless, although
the problems and prospects of language teaching could easily swallow

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
370 Joshua A. Fishman
up or trample underfoot not only all of the sociology of language but
also all of sociology, psychology and linguistics per se, first linguistics
and, more recently, the sociology of language have had some impact on
the beast. A valuable introduction to sociology of language has been
presented to language teachers generally by Halliday, Mcintosh and
Strevens (1964). In this introduction teachers are urged to recognize
the different uses (and, therefore, the different varieties) of language
that coexist within speech communities rather than, as has usually
been the case thus far, to persist in the erroneous and deadening fiction
that there is always only one (and always the same one) correct variety.
More recently we have witnessed a deluge of 'sociolinguistically
oriented' interest in the language of disadvantaged speakers of non-
standard English with Bernstein's work (e.g. 1964) being best known
in England and Labov's (1965) or McDavid's (1958) in the U.S.A.
Most of the products of this interest seek to contrastively highlight the
basic structure of the speech of such communities so that teachers may
be able to more successfully recognize and overcome the difficulties that
learners will encounter when confronted with the phonological and
grammatical structures of standard (school) English (e.g., Labov 1966d,
1968b, Wulfram 1969, Baratz and Shuy 1969). Nevertheless much (if
not all) of what is currently offered to teachers in this connection is
merely "sociolinguistically oriented" (in that it recognizes that minority
group members often utilize varieties of English unfamiliar to others)
rather than sociology of language proper (Fishman 1969b, Fishman and
Lueders, in press). "Sociolinguistically oriented" advice is now also
being directed toward teachers of bilinguals (Anderson 1969, Boyd
1968, Gaarder 1965). Such teachers are admonished that learners
should be encouraged to maintain or acquire repertoires (incorporating
several varieties) in each of their languages - rather than to displace all
non-standard varieties in favor of one artificial standard version of
each. Teachers of bilinguals are being urged to enable their students
to select from each repertoire in accord with the norms for communi-
cative appropriateness of the particular networks with which they (the
pupils) seek mutually accepting interaction (Fishman and Lovas 1970,
Mackey 1970). Nevertheless, the teachers of bilinguals (particularly
in countries of mass immigration) have just begun to be shown how
to influence the bilingual settings in which they and their students
live and in which one or another of their languages may be roundly
ignored, if not attacked, as soon as school is over (Andersson and
Boyer 1970, John and Horner 1971).
Of course, the distance is still considerable between "sociolinguisti-
cally oriented" advice or sensitivity training for teachers and any more
complete interrelationship between teaching methodology and sociology

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 371

of language. Thus the education of bilinguals is still viewed primarily

within the context of disadvantaged and dislocated minorities (whose
lot in life will be far easier if only they learn English, French, Russian,
etc.) rather than within the broader context of worldwide experience
with bilingual education — whether in conjunction with elitist bilingua-
lism, traditional bilingualism, or, more generally, widespread and stable
(i.e., non-dislocated) bilingualism. As a result, the education of speakers
of non-standard English is being pondered without awareness, for
example, of the fact that most students entering German, French,
Italian, and other schools during the past century have also been
speakers of non-standard varieties of their respective languages. A
true meeting of education and the sociology of language will enable
both to discover why proportionally so many dialect speakers do and
did seem to become readers and speakers of the standard language
(and even of classical languages) in other parts of the world whereas so
few seem to accomplish this in the U.S.A. today (Fishman and Lueders,
in press). As with many other social science fields, a severe test of the
power of the sociology of language will be its ability to be useful in
the world of affairs. The education of non-literates, of bilinguals and
of non-standard speakers are all fields about which the sociology of
language must have more to say if it is really a discipline worth
listening to at all.

APPENDIX

Linguistics: The Science of Code Description . . . and More

If one part of sociolinguistics comprises the 'study of the characteristics

of language varieties' then we must turn to that science that has special-
ized in the systematic description of language: linguistics. To attempt
to describe and analyze language data, in this day and age, without a
knowledge of linguistic concepts and methods is to be as primitive as
to try to describe and analyze human behavior more generally (or the
functions of language varieties and the characteristics of their speakers)
without knowledge of psychological and sociological concepts and
methods.
It is no more possible to provide an adequate introduction to linguis-
tics 'in one easy lesson' than to provide such for sociology or psychol-
ogy. Nevertheless, it may be possible to briefly sketch some of the major
concerns and methods of linguistics that bear upon sociolinguistics.
The purpose of the next few pages, therefore, is to bring about 'linguis-
tics appreciation', and of a very selective sort at that, rather than to

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
372 Joshua A. Fishman
present a full-fledged introduction to what is a very technical and
complicated science which intersects the humanities, the social sciences
and the natural sciences in its various subdivisions. The specialist
knows full well that 'music appreciation' is not the same as music
mastery. Similarly, 'linguistics appreciation' is not the same as lin-
guistics mastery. Nevertheless, it is a beginning.
As a formal discipline, particularly insofar as the American academic
scene is concerned, linguistics is a very recent field of specialization.
The Linguistic Society of America was founded only in 1924. (The
oldest linguistic society in the world, that of Paris, was founded in
1864.) Even today, when the number of linguists and linguistics pro-
grams in American universities is greater than ever before, there are
only some two score graduate linguistics departments in the United
States. Nevertheless, this discipline has not only come to be of prime
interest to a growing band of dedicated scholars and practitioners
within linguistics per se, but it has also in very recent years forcefully
come to the attention of all other disciplines that recognize the centrality
of verbal interaction in human affairs. Interdisciplinary contacts between
linguistics and anthropology have been well established since the very
appearance of linguistics in American universities. The anthropological
linguist is a well recognized and highly regarded specialist among lin-
guists and anthropologists alike. Indeed, linguistics is recognized as a
'branch' of anthropology in many textbooks and training programs. Of
more recent vintage is psycholinguistics. Most recent of all is socio-
linguistics, an interdisciplinary field which is just now beginning to
train specialists that can bridge linguistics and sociology-social psychol-
ogy in such a manner as to expand the horizons of both.
Descriptive Linguistics. The basic field in which most (if not all) lin-
guists have been trained is that which is known as descriptive or syn-
chronic linguistics. As its names imply, this field focuses upon the syste-
matic description of a given language in a given time and place. It is
not historical; it is not comparative; it is not prescriptive. Its emphasis
is definitely on spoken language, the assumption being that written
language is both derivative and different from natural language or
speech.
It is common for the uninitiated to think of a language as being well
represented by an unabridged dictionary. This view implies that the
way to describe a language is to consider its components to be words.
Any careful or consistent and exhaustive presentation and definition of
the words of a language (which may be exactly what dictionaries
attempt to do) would, therefore, from this point of view, be considered
a description of that language. For most linguists, however, there are
two other kinds of systematic presentations which are considered even

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 373

more basic to their goal of describing language: the sound system of

a language and the grammatical system of that language.
The branch of linguistics that is concerned with the systematic des-
cription of the sounds (phones) of a language is phonology. Some of
the more general subspecialties within phonology are articulatory pho-
netics (how tongue, lips, teeth, vocal chords, velum, nasal passage and
other speech organs produce the sounds of language) and acoustic
phonetics (the physical properties of the sound waves or signals emitted
by the speaker). Linguistics have devised for purposes of phonetic
notation the International Phonetic Alphabet which is roughly adequate
for the transcription of speech in all languages, although minor adjust-
ments or additions to it are required in most individual cases.
On the foundation of these more general branches of phonology
linguistics has been able to establish the study of phonemics, i.e., the
study of those sounds that enter into meaningful contrasts or combina-
tions in a given language, as compared to all of the physically differ-
entiable sounds of a language (which are of interest in phonetics).
A skilled phonetician differentiates far more fine shades of language
sounds than do the native speakers of any particular language. Phonetic
analysis is now sufficiently refined to demonstrate that no two speakers
of a given language pronounce their words in exactly the same way.
Indeed, the degree of refinement available to phonetic analysis has gone
so far that it is possible to show that even an idiolect (the way of
speaking that characterizes an individual) is not entirely consistent. The
same individual does not pronounce the same word in the same way on
all occasions of the same type. Into this endless series of successively
refined analysis of language sound differences phonemics seeks to
introduce the parsimony that derives from a knowledge of those sound-
differences that are meaningfully distinctive (i.e., that serve to distin-
guish between linguistic signs and their meanings) for the native speakers
of a particular language. The following brief example may illustrate the
phonemic approach to demonstrable phonetic differences.
Let us consider the " b " sound in English, Arabic and Bengali. That
each of these languages has some sound that the American man-in-
the-street would unhesitatingly represent by the letter b is, for linguis-
tics, a nonstructural comment and, therefore, one of no particular
interest. It is of interest, however, to point out that in English aba and
apa are differentiated, the voiced bilabial stop ("b) in the first being
considered clearly different from the unvoiced bilabial stop ( " p " ) in the
second because the difference between b and p is crucial to recognizing
the difference in meaning between "bit" and "pit", "bet" and "pet", and
hundreds of other meaningful contrasts. In Arabic, on the other hand,
no such meaningful substitutions of b and p are made. T h e native

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
374 Joshua A. Fishman
speaker of Arabic says only aba and uses a p sound only under
special conditions, such as before s or t. More generally put, whatever
sound differences exist in the p-b range in Arabic are not distinctive,
i.e., they do not regularly signal meaning differences.
Thus, it is not enough to say that both English and Arabic have a
b sound, for the sound functions far differently in the two languages.
In English b and p function as phonemically different sounds (and,
therefore, are notated / p / and / b / ) ; in Arabic they do not.
The absence or presence of a meaningful contrast between b and p
takes on even greater linguistic significance if Bengali is examined. Not
only are / b / and / p / differentiated by the ordinary native speaker of
Bengali, but, in addition, an unaspirated p (as in the English spin) is
differentiated from an aspirated p (as in the English pin). Similarly, an
unaspirated b is regularly differentiated from an aspirated b.
Note, that while English recognizes a phonemic (meaning-related)
difference between two sounds (one voiced and one unvoiced) that
represent only a meaningless difference in Arabic, Bengali recognizes a
further phonemic difference between two pairs of sounds (each with an
aspirated and an unaspirated component) that represent only meaning-
less phonetic differences in English. Furthermore, as the English and
Bengali languages change over time, changes in their "b" sounds will
presumably be correlated with changes in their "p" sounds, precisely
because these sounds are systematically related to each other.
It is in this last respect — i.e., in terms of systematic interrelation-
ships - that descriptive linguistics is interested in the sounds of a
language. This is also why descriptive linguistics is sometimes referred
to as structural linguistics. It is not merely the sounds of a language
that are of interest to linguistics, nor even the meaningfully different
sounds, but, above all, the systematic links that exist between the
meaningfully different sounds of a language. The phonemes of a
language, like all other features of a language at a given point in time,
are part of a system (a 'structure') that operates as a whole. Changes
in one part of the structure affect the other parts; indeed, in true Gestalt
fashion, any phonemic part can be truly appreciated only in terms of
the phonological whole. A famous linguist of the first part of this
century was the first to emphasize that language is a system in which
every part has its (interlinked) place ("un systeme ou tout se tient"; de
Saussure 1916) and this structural dictum has since then come to
characterize not only descriptive linguistics but other branches of lin-
guistics as well.
So basic is descriptive linguistics to linguistic science as a whole
that another example of its concerns, this time at the level of gramma-
tical analysis, is in order. Such an example is particularly desirable

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 375
because the grammatical structure of language is so completely inter-
woven with its sound structure, so much so that some linguists
claim that phonological analysis depends on and must be part of an
exhaustive grammatical analysis (although most linguists consider phon-
ology, grammar, lexicon and semantics as quite separate levels of
analysis).
Just as there is a minimal unit of meaningful sound (actually, of
substitutionally meaningful sound, since the sounds in question are
not meaningful per se), the phoneme, so is there a minimal unit of
meaningful grammatical (i.e. of ordered or environmental) form, the
morpheme. As a result, one branch of grammatical study is known as
morphology. It studies the ordered relationships between small mean-
ingful segments such as occur within words. (Syntax, on the other
hand, studies the ordered relationships between units such as words
in a phrase or utterance.) Thus, many English verbs form the past
tense by adding a morpheme, which may be represented as to
the present tense of the verb: I open - I opened, / d j means past tense
in English. Similarly, many English nouns form their plural by adding
a morpheme, which may be represented as | z | , to their singular: car
- cars. In both of these instances, however, the morphemes in question
occur in several different forms that also differ somewhat as to their
sound. Functionally equivalent alternatives of the same morpheme are
referred to as allomorphs, precisely because there is no functional
difference between them, however much they may differ in sound, just
as sounds that revealed no functional difference were referred to
earlier as allophones. The allomorphs of for the common, pro-
ductive English verbs may sound like a d (as in opened), like a t
(as in laughed), or like ed (as in mended). However, these allomorphs
are not used at random. How would linguistics provide a rule to
indicate when the native speaker of English employs which? What
would such a rule be like?
To begin with, linguists would list as many verbs as possible that
utilize each variant of the { d j morpheme. Such a list might initially
look like that shown in Table 1. After inspecting the array of final
sounds in each of the columns of that table a linguist is able to do
that which no ordinary native speaker of English can do: formulate
a very few rules which summarize the systematic variation in the
three allomorphs of { d j . Such rules might proceed as follows:

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
376 Joshua A. Fishman
TABLE 1
Allomorphs of {d} in the past tense of some common, productive
English verbs
ed t d
mend bank open
lift cook use
boot drop save
raid help bomb
kid walk mail
tend laugh try
sift shop play
hoot stamp radio
shade rank hinge
hand staff rig

1. If the verb stem ends in /t/ or /d/ the past tense ends in ed (with
the exception of a small number of verbs that retain the same form
in past and present: cut, hit, put),
2. If the verb stem ends in a voiceless stop (other than /t/) or in a
voiceless spirant, the past tense ends in t,
3. Otherwise, the past tense ends in d.
The above three brief rules pertain to the phonological conditioning
of allomorphs. The allomorphs of |d| are realized according to their
phonological environment. Thus, variations in grammatical form and
variations in phonological form may and frequently do coincide. In
general, linguistics has traditionally pursued two kinds of structured
variation: variation relatable to change in meaning (such as the subst'
tutional meaning that underlies phonemic analysis) and variation re-
latable to change in environment (such as the positional meaning that
underlies morphemic analysis). Further synchronic variation in lan-
guage, i.e., variation that cannot be identified either with change in
meaning (i.e. change in referent) or with change in linguistic environ-
ment, when geographic area is held constant, has traditionally been
thought of as 'free variation', i.e. as variation (not to say 'irregularity')
due to factors outside of langue (the latent structure underlying speech)
and, therefore, outside of the descriptive rules pertaining to langue.
It is in some of the kinds of free variation - in variations which may
co-occur with differences in a speaker's alertness or emotional state,
with differences in topic, role relationship, communicational setting or
interpersonal purpose - that sociolinguistics (and other interdisciplinary
studies of language usage) attempts to discover additional regularity.
Linguistics has long been aware that 'free variation' might have a

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 377
structure of its own. However, that structure (when and if it obtains)
has usually been considered as being part of the structure of the
speech event rather than part of the structure of the speech code per
se. Although descriptive linguistics has emphasized the spoken lan-
guage, the speech act itself was long considered to be outside of the
domain of linguistics, for the speech act, just as the message content of
speech, was considered to be part of 'communication' (long considered
by linguists to be an outer or surface phenomenon) rather than part and
parcel of langue per se (the heart of the matter). Many famous linguists
have warned against confusing the two.
Thus, if it appeared that certain phonemic, morphological, syntactic
or lexical regularities were not always as regular as one would hope
(time and place remaining constant) this was attributed to the irre-
gularity of parole (speaking, behavioral realization) as distinct from the
systematic and abstract purity of langue (language, underlying structure)
with which linguists should really be concerned. Parole is subject to
many factors that produce variation (among those not previously
mentioned: fatigue, anger, limitation in memory span, interruptions,
etc.). These are all factors of 'degree', of 'more or less', of 'sometimes'.
It was thought that the goal of linguistics was to cut through these
psychological and sociological sources of 'static' and to concern itself
with matters that were clearcut enough to be viewed as all or none
phenomena: the basic code which, at any given time and place, might
be considered to be one and the same for all who employed it. Thus,
not only were linguists warned to distinguish sharply between parole
and langue (de Saussure 1916), but they were also admonished to keep
their distance from psychological or sociological data and theories
which were viewed as inherently more concerned with the highly
variable and seemingly irregular processes of verbal interaction and
communication (and, therefore, with the messy data of parole) than
with the pure code underlying these processes (Bloomfield 1933). It is
only in more recent days, when the traditionally rigid distinction
between langue and parole has come to be reexamined and when the
varying interaction between them has come to be pursued that larger
groups of linguists and of social scientists have found things to say to
each other.
Other Branches of Linguistics. Other branches of linguistics — some
of them older than descriptive linguistics (even though the latter has
come to be so central to all linguistic pursuits) - have long been on
friendlier terms with the social sciences. Historical (diachronic) lin-
guistics, for example, in studying the changes that occur in a given code
over time (sound changes, grammatical changes and word changes) has,
of necessity, been interested in human migrations, intergroup contacts

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
378 Joshua A. Fishman
(conquest, trade), and whatever other diversification within a speech
community that leads some of its members to interact with each other
differentially (rather than equally or randomly), or that leads some of
its members to interact with outsiders much more than do the rest.
Historical linguistics (also known as comparative linguistics) focuses
on tracing how one, earlier, parent ('proto') code subsequently divided
into several related but separate ('sister' or 'daughter') codes or, alter-
natively, how several codes were derived from one pre-existing code.
Although time is the crucial dimension in the development of families
of languages between which genetic relationship can be shown to
exist, as it is in the reconstruction of all common ancestries, never-
theless, historical linguists realize full well that the language changes
that occurred were due to differential interaction and contact processes
that transpired as time passed, rather than to the mere passing of time
per se. As a result, historical linguistics has interacted fruitfully with
history, archeology, anthropology and with other disciplines that can
provide information concerning coterritorial influences between popu-
lations. In recent years, the fluctuating interaction between langue and
parole (e.g., how one of the alternative systems of speaking available
to a speech community spreads through the entire speech community
and, increasingly displacing other alternatives, becomes an unvarying
part of its basic code) has been studied by linguists working with social
science concepts and methods of data collection and data analysis on
what would once have been considered a 'purely' comparative problem
(Labov 1963; Haugen 1961). The ties between comparative linguistics
and the social sciences become stronger as the dynamics of language
change come under increasing linguistic scrutiny, as distinct from the
static, stepwise contrasts between the written records of one century
and those of another that formerly dominated this field of study.
Another branch of linguistics that has frequently maintained close ties
to the social sciences is dialectology (also known as linguistic or dialect
geography). In comparison to historical linguistics this branch is con-
cerned with variation in language on some dimension other than time.
The achronic dimension with which dialectologists have most common-
ly been concerned has been geographic space or distance. Language that
are employed over considerable expanses are often spoken somewhat
differently (or even quite differently) in different parts of their speech
areas. These differences may be phonological, such as President Ken-
nedy's "Cuber" (for Cuba) and "vigah" (for vigor), where a Philadel-
phian would have said "Cubah" and "vigor" while many a Southerner
would have said "Cubah" and "vigah". Dialect differences may
also apply to the lexicon (milk shake vs. frappe; soda vs. pop) and even
to parts of the grammatical system. Dialectologists have traditionally

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 379
prepared linguistic atlases to show the geographic distribution of the
linguistic features that have been of interest to them. Such atlases consist
of maps on which are indicated the geographical limits within which
certain usages are current. These limits are known as isoglosses (Wein-
reich, U. 1962; Herzog 1965).
However, dialectologists are well aware that the variations that are
of interest to them are not due to geographical distance per se, but
rather to the interactional consequences of geographic and other kinds
of 'distance'. Phonological, lexical and grammatical uniformity may
obtain over large geographic expanses when settlement is simultaneous
and when verbal interaction as well as common identification are
frequent. On the other hand, major language differences (sometimes
referred to as 'social dialects' or 'sociolects') may arise and be main-
tained within relatively tiny geographic areas (e.g. in many cities) where
the above conditions do not obtain. Considerations such as these have
led many dialectologists, particularly those who have been interested
in urban language situations, to be concerned with educational, occu-
pational, religious, ethnic and other social groups and societal processes
(although all or most of these groups may, in part, be traceable to
originally diverse geographic origins) rather than with geographic dis-
tance per se. As a result, the ties between dialectologists and social
scientists (not to mention sociolinguists) have been many and strong,
particularly in recent years (Blanc 1964; Ferguson and Gumperz 1960),
when the entire speech act - rather than merely the code rules abstrac-
ted from the speech act - has come to be of interest to an increasing
number of dialectologists (Hymes 1962).
Of late, many linguists have taken to examining the structure of
language - rather than the structure of particular languages - and to
doing so in order to discover the nature of those fundamental human
capacities which make for the competence of native speakers. Native
speakers possess a rare gift which they themselves usually overlook: the
ability to generate sentences that are recognized as structurally accept-
able in their speech communities, and, what is more, to generate only
such sentences. Many linguists now believe that a linguistic theory than
can specify an adequate grammar (i.e. the rules that native speakers
implicitly grasp and that constitute their native speaker competence)
will also specify the language acquiring and language using nature of
man. These linguists say that only an adequate theory of human capa-
city to acquire and use language will yield an adequate theory of what
language itself is (Chomsky 1957, 1965).
Sociolinguistics may ultimately serve similarly basic purposes in the
ongoing quest of the social sciences to understand communicative com-
petence as a fundamental aspect of the social nature of man. The

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
380 Joshua A. Fishman
sociolinguistic theory that can specify adequate communicative com-
petence (i.e. the rules that native members of speech communities
implicitly grasp and that constitute their native member sociolinguistic
behavior) will also specify the nature of social man as an acquirer and
utilizer of a repertoire of verbal and behavioral skills. Man does not
acquire or use his communicative competence in a single-code or single-
norm community. Indeed, pervasively homogeneous communities with
respect to communicative and other social behaviors do not exist except
in the simplified worlds of some theorists and experimentors. Ulti-
mately, sociolinguistics hopes to go beyond comfortably simple theory
concerning the nature of communicative competence in the conviction
that only an adequate theory of human capacity to acquire and to use
a repertoire of interlocking language varieties and their related beha-
viors will yield an adequate theory of what communicative competence
in social man really is.
Just as there are branches of linguistics that seek to study langue and
language alone (indeed, to study language at its 'deepest', most abstract,
and, therefore, at its socially most uninvolved) so are there branches of
linguistics that have departed from a strict separation between langue
and parole (since parole has its very definite structure, since parole
constantly influences langue, and since the individual's meaningful
differentiation must be referred to, even though these are outside of
langue per se, in order to establish a description of phonemic and other
distinctions). Similarly, some branches of social psychology (and other
social sciences as well) have moved closer to linguistics. Many sociolo-
gists and social psychologists now realize (whereas few did so a
decade ago) that the norms that apply to and that may be thought of
as generating human verbal interaction pertain not only to the com-
municative content and context of that interaction but to its linguistic
form as well. As linguistics is developing outward - in the hopes of
some: to become an all-encompassing science of language behavior-
sociology and social psychology are developing toward increasing tech-
nical competence in connection with language description and analysis.
The sociology of language is one of the byproducts of these two com-
plementary developments and, as such, it must refer not only to the
work of linguists but attend as well to those topics that are essentially
sociological and social psychological and to which few linguists have,
as yet, paid much attention.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 381
SUGGESTED READINGS

I. A Few Popular Introductions and Overviews

1. Hall, Robert A., Jr., Linguistics and Your Language (New York: Dou-
bleday, 1960).
2. Orenstein, Jacob, and Wm. W. Gage, The ABC's of Language and
Linguistics (Phila. and New York: Chilton, 1964).
3. Bolinger, Dwight, Aspects of Language (New York: Harcourt, Brace
and World, 1967).
II. Some Traditional American Texts
1. Bloomfield, Leonard, Language (New York: Holt, Rinehart and Wins-
ton, 1933).
2. Gleason, H. A., Jr., An Introduction to Descriptive Linguistics (revised)
(New York: Holt, Rinehart and Winston, 1961).
3. Hockett, Charles F., A Course in Modern Linguistics (revised) (New
York: Macmillan, 1963).
4. Sapir, Edward, Language (New York: Harcourt, Brace, 1921). (Paper-
back: New York: Harvest Books, 1955).
III. One Classic and Three Recent Texts by European Authors
1. de Saussure, Ferdinand, Course in General Linguistics (translation of
1916 French Original by Wade Baskin) (New York: Philosophical
Library, 1959).
2. Martinet, Andre, Elements of General Linguistics (translation of 1960
French Original by Elisabeth Palmer) (Chicago: University of Chicago
Press, 1964).
3. Halliday, M. A. K., Angus Mcintosh, and Peter Strevens, The Linguis-
tic Sciences and Language Teaching (London: Longmans, Green, 1964).
4. Robins, Robert H., General Linguistics: An Introductory Survey (Bloom-
ington: Indiana University Press, 1966).
IV. Examples of the Newer 'Transformationalist Approach
1. Bach, Emmon W., An Introduction to Transformational Grammars
(New York: Holt, Rinehart, Winston, 1964).
2. Chomsky, Noam, Aspects of the Theory of Syntax (Cambridge: MIT
Press, 1965).
3. Langacker, Ronald W., Language and its Structure (New York: Har-
court, Brace and World, 1968).
V. Journals to Glance At
1. Language.
2. Lingua.
3. Linguistics.
4. Linguistic Reporter.
5. International Journal of American Linguistics.
6. Language in Society.

VI. References
1. Rutherford, Phillip R., A Bibliography of American Doctoral Disserta-
tions in Linguistics, 1900-1964 (Washington: Center for Applied Lin-
guistics, 1968).
2. Linguisticinformation (Washington: Center for Applied Linguistics,
1965).

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
382 Joshua A. Fishman
3. Cartter, Allan M., "Doctoral programs in linguistics", in his An Assess-
ment of Quality in Graduate Education (Washington: American Coun-
cil on Education, 1966).
4. Various articles on linguistic topics in the new International Encyclo-
pedia of the Social Sciences (New York: Macmillan, 1968). (See the
review of these topics in Language, 1969, 45, 458-463.)

10.0 BIBLIOGRAPHY

Alisjahbana, S. Takdir,
1962 "The modernization of the Indonesian language in practice", in his
Indonesian Language and Literature: Two Essays (New Haven, Yale
University, Southeast Asia Studies), 1-22.
Alisjahbana, S. Takdir,
1965 "New national languages: a problem modern linguistics has failed to
solve", Lingua, 15, 515-530.
Alisjahbana, S. Takdir,
1969 "Some planning processes in the development of the Indonesian/Malay
language", Consultative Meeting on Language Planning Processes.
(Honolulu, IAP); subsequently in (Rubin, Joan and Jernudd, Bjorn,
(eds.), Can Language be Planned? (Honolulu, East-West Center Press)
1971.
Alisjahbana, S. Takdir (ed.),
1971 The Modernization of the Languages of Asia (Kuala Lumpur, Uni-
versity of Malaysia Press).
Andersson, Theodore,
1969 "Bilingual schooling: oasis or mirage?" Hispania, 52, 69-74.
Andersson, Theodore and Mildred Boyer (eds.),
1970 Bilingual Schooling in the United States, 2 vols., Washington, D.C.,
USGPO.
Andreev, N. D. and L. R. Zinder,
1959 Osnovnye problemy prikladnoj lingvistiki, Voprosy Jazykoznaiya, no. 4,
1-9.
Anon.,
1930 Practical Orthography of African Languages, (International Institute of
African Languages and Cultures, Memorandum 1) (Oxford, Oxford
University Press) revised edition.
Anon.,
1953 The Use of Vernacular Languages in Education (Paris, UNESCO).
Anon.,
1963 A Common Script for Indian Languages (Delhi, Ministry of Scientific
Research and Cultural Affairs).
Anon.,
1965 The Restoration of the Irish Language (Dublin, The Stationery Office),
Also note Progress Report for the Period Ended 31 March, 1966 and
Progress Report for the Period Ended 31 March 1968. (Dublin, The
Stationery Office) 1966 and 1968.
Anon.,
1968 A Motivational Research Study for the Greater Use of the Irish
Language, 2 vols. (Croton-on-Hudson (N.Y.), Ernest Dichter Inter-
national Institute for Motivational Research).

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 383
Armstrong, Robert,
1968 "Language policies and language practices in West Africa", in (Fishman,
J. A., C. A. Ferguson and J. Das Gupta (eds.), Language Problems of
Developing Nations. (New York, Wiley), 227-236.
Bacon, Elizabeth E.,
1966 "Russian influence on Central Asian languages", in her Central Asians
Under Russian Rule (Ithaca, Cornell University Press).
Baratz, Joan and Roger W. Shuy,
1969 Teaching Black Children to Read (Washington, Center for Applied
Linguistics).
Barker, George C.,
1947 "Social Functions of Language in A Mexican-American Community",
Acta Americana, 5, 185-202.
Basso, Keith H.,
1967 "Semantic aspects of linguistic acculturation", American Anthropologist
69, 471-477.
Barker, George C.,
1947 "Social Functions of Language in A Mexican-American Community",
Acta Americana 5, 185-202.
Basso, Keith H.,
1967 "Semantic aspects of linguistic acculturation", American Anthropologist
69, 471-477.
Bell, Daniel,
1961 The End of Ideology (New York: Collier).
Bernstein, Basil,
1964 "Elaborated and restricted codes: Their social origins and some con-
sequences", American Anthropologist 66, No. 6, Part 2, 55-69.
1966 "Elaborated and restricted Codes: An Outline", Sociological Inquiry 36,
254-261.
Berry, Jack,
1958 "The making of alphabets", Proceedings of the International Congress
of Linguistics (Oslo: Oslo University Press), 752-764; also reprinted in
J. A. Fishman (ed.), Readings in the Sociology of Language (The
Hague: Mouton, 1968), 737-753.
Blanc, Chaim,
1964 Communal Dialects in Baghdad (Cambridge: Harvard University Press).
Blom, Jan Peter, and John J. Gumperz,
In press "Some social determinants of verbal behavior", in John J. Gumperz
and Dell Hymes (eds.), The Ethnography of Communication: Direct-
ions in Sociolinguistics (New York: Holt, Rinehart and Winston).
Bloomfield, Leonard,
1933 Language (New York: Holt).
Bock, Philip K.,
1964 "Social Structure and Language Structure", Southwestern Journal of
Anthropology 20, 393-403; also in J. A. Fishman (ed.), Readings in
the Sociology of Language (The Hague: Mouton, 1968), 212-222.
Bolinger, D. L.,
1946 "Visual morphemes", Language 22, 333-340.
Bonjean, Charles M.,
1966 "Mass, class and the industrial community: A comparative analysis of
managers, businessmen and workers", American Journal of Sociology
72, 149-162.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
384 Joshua A. Fishman

Bottenberg, R. A., and K. H. Ward, Jr.,

1963 Applied Multiple Linear Regression (Lackland, Tex.: Lackland AF
Base PRL-TDR-63-6).
Boulding, Kenneth,
1963 "The death of the city: A frightened look at postcivilization", in
Oscar Handlin and John Burchard (eds.), The Historian and the City
(Cambridge: MIT Press and Harvard University Press), 145.
Bowers, John,
1968 "Language problems and literacy", in J. A. Fishman, C. A. Ferguson
and J. Das Gupta (eds.), Language Problems of Developing Nations
(New York: Wiley), 381-401.
Boyd, Dorothy L.,
1968 "Bilingualism as an educational objective", The Educational Forum 32,
309-313.
Bradley, Henry,
1913/14 "On the relation between spoken and written language", Proceed-
ings of the British Academy 6, 212-232.
Bright, William,
1960 "Animals of acculturation in the California Indian languages", Uni-
versity of California Publications in Linguistics 4, No. 4, 215-246.
Broom, Leonard, and D. Norval Glenn,
1966 "Negro-White differences in reported attitudes and behavior", Sociology
and Social Research 50, 187-200.
Brosnahan, L. F.,
1963a "Some historical cases of language imposition", in Robert Spencer (ed.),
Language in Africa (Cambridge, Eng.: Cambridge University Press),
7-24.
1963b "Some aspects of the linguistic situation in tropical Africa", Lingua 12,
54-65.
Brown, Roger W., and Albert Gilman,
1960 "The pronouns of power and solidarity", in Thomas A. Sebeok (ed.),
Style in Language (Cambridge and New York: Technology Press of
MIT and Wiley), 253-276; also in J. A. Fishman (ed.), Readings in the
Sociology of Language (The Hague: Mouton, 1968), 252-275.
Brown, Roger W. and Eric H. Lenneberg,
1954 "A study in language and cognition", Journal of Abnormal and Social
Psychology 49, 454-462.
Burns, Donald,
1953 "Social and political implications in the choice of an orthography",
Fundamental and Adult Education 5(2), 80-85.
1968 "Bilingual education in the Andes of Peru", in J. A. Fishman, C. A.
Ferguson and J. Das Gupta (eds.), Language Problems of the Develop-
ing Nations (New York: Wiley), 403-414.
Carmichael, L., H. P. Hogan, and A. A. Walter,
1932 "An experimental study of the effect of language on the perception
of visually perceived form", Journal of Experimental Psychology 15,
73-86.
Carroll, John B., and J. B. Casagrande,
1958 "The function of language classifications in behavior", in E. Maccoby,
T. Newcomb and E. Hartley (eds.), Readings in Social Psychology
(New York: Holt), 18-31.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 385
Catford, J. C.,
1965 A Linguistic Theory of Translation (London: Oxford University Press).
Chinese Written Language Reform Committee,
1956 "Several points concerning the Han language phoneticization plan
(draft) explained", Current Background, No. 380 (March 15), 4-13.
Chomsky, Noam,
1957 Syntactic Structures (The Hague: Mouton).
1965 Aspects of the Theory of Syntax (Cambridge: MIT Press).
Chou En-lai,
1965 "Current tasks of reforming the written language", in Reform of the
Chinese Written Language (Peking: Foreign Language Press), 7-29.
Cohen, Jack,
1965 "Some statistical issues in psychological research", in B. B. Wolmand
(ed.), Handbook of Clinical Psychology (New York: McGraw-Hill),
95-121.
1968a "Prognostic factors in functional psychosis: A study in multivariate
methodology", invited address at the New York Academy of Sciences,
March 18, mimeographed.
1968b "Multiple regression as a general data-analytic system", Psychological
Bulletin 70, 426-443.
Conklin, Harold C.,
1962 "Lexicographic treatment of folk taxonomies", in Fred W. House-
holder and Sol Saporta (eds.), Problems in Lexicography (Bloomington:
Indiana University Research Center in Anthropology, Folklore and
Linguistics), Publication 21, 119-141.
Cook, S. F.,
1943 'The conflict between the California Indian and white civilization",
Ibero-Americana 21, 1-194, 1-55; 23, 1-115; 24, 1-29.
Corpas, Jorge Pineros,
1969 "Inconvenientes de la enseñanza bilingüe a la luz de la fisiología
cerebral", Noticias Culturales (Bogotá), No. 99, 1-4.
Correa, Hector, and Jan Tinbergen,
1962 "Quantitative adaptation of education to accelerated growth", Kyklos
15, 776-785.
Dalby, David,
1967 "A survey of the indigenous scripts of Liberia and Sierra Leone: Vai,
Mende, Loma, Kpelle and Bassa", African Languages Studies 8, 1-51.
1968 "The indigenous scripts of West Africa and Surinam: Their inspiration
and design", African Languages Studies 9, 156-197.
Das Gupta, Jyotirindra,
1969 "Religious loyalty, language conflict and political mobilization", Con-
sultative Meeting on Language Planning Processes (Honolulu: EWC-
IAC); subsequently in Joan Rubin and Bjorn Jernudd (eds.), Can
Language be Planned? (Honolulu: East-West Center Press, 1971).
Das Gupta, Jyotirindra, and Joshua A. Fishman,
1971 "Interstate migration and subsidiary language-claiming; an analysis of
selected Indian census data", International Migration Review, 5, no. 2
Davis, Frederick B.,
1967 Philippine Language-'Teaching Experiments (Phillippine Center for
Language Study, No. 5) (Quezon City: Alemar-Phoenix).
DeFrancis, John,
1947 "Japanese language reform: Politics and phonetics", Far Eastern Survey

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
386 Joshua A. Fishman
16, No. 19, 217-220.
1950 Nationalism and Language Reform in China (Princeton: Princeton Uni-
versity Press).
1968 "Language and script reform (in China)", Current Trends in Linguistics
(The Hague: Mouton), 130-150, also in J. A. Fishman (ed.), Advances
in the Sociology of Language (The Hague: Mouton, 1971).
de Saussure, Ferdinand,
1959 Course in General Linguistics (translated by Wade Baskin) (New York:
Philosophical Library). Original (French) publication: 1916.
Deutsch, Karl W.,
1942 "The trend of European nationalism - the language aspect", American
Political Science Review 36, 533-541.
1966 Nationalism and Social Communication (Cambridge: MIT Press) (2nd
edition).
Dickens, K. J.,
1953 "Unification: The Akan dialects of the Gold Coast", in The Use of
Vernacular Languages in Education (Paris, UNESCO), 115-123.
Dohrenwend, Bruce P., and J. Robert Smith,
1962 "Toward a theory of acculturation", Southwest Journal of Anthropology
18, 30-39.
Dozier, Edward P.,
1951 'Resistance to acculturation and assimilation in an Indian pueblo",
American Anthropologist 53, 56-66.
Edelman, Martin, Robert L. Cooper, and Joshua A. Fishman,
1968 "The contextualization of school children's bilingualism", Irish Journal
of Education 2, 106-111.
Ellis, Dean S.,
1967 "Speech and social status in America", Social Forces 45, 431-437.
Epstein, A. L.,
1959 "Linguistic innovation and culture on the Copperbelt, Northern Rho-
desia", Southwestern Journal of Anthropology 15, 235-253; also in
J. A. Fishman (ed.), Readings in the Sociology of language (The
Hague: Mouton, 1968), 320-339.
Ervin, Susan M., and Charles E. Osgood,
1954 "Second language learning and bilingualism", Journal of Abnormal
and Social Psychology 49, Supplement, 139-146.
Ervin-Tripp, Susan M.,
1964 "An analysis of the interaction of language, topic and listener", Ameri-
can Anthropologist 66, Part 2, 86-102; also in J. A. Fishman (ed.),
Readings in the Sociology of Language (The Hague: Mouton, 1968),
192-211.
1969 "Sociolinguistics", in L. Berkowitz (ed.), Advances in Experimental
Social Psychology, Vol. 4 (New York: Academic Press), 91-165; also in
J. A. Fishman (ed.), Advances in the Sociology of Language (The
Hague: Mouton, 1971).
Ferguson, Charles A.,
1959a "Diglossia", Word 15, 325-340.
1959b "Myths about Arabic" ( = Monograph Series on Languages and Lin-
guistics) (Georgetown University) 12, 75-82; also in J. A. Fishman (ed.),
Readings in the Sociology of Language (The Hague: Mouton, 1968),
375-381.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 387
1965 "Directions in sociolinguistics; Report on an interdisciplinary seminar",
SSRC Items 19, No. 1, 1-4.
1967 "St. Stefan of Perm and applied linguistics", in To Honor Roman
Jakobson (The Hague: Mouton); also in J. A. Fishman, C. A. Fergu-
son and J. Das Gupta (eds.), Language Problems of Developing Na-
tions (New York: Wiley, 1968), 253-266.
1968 "Language development", in J. A. Fishman, C. A. Ferguson and J. Das
Gupta (eds.), Language Problems of Developing Nations (New York:
Wiley).
Ferguson, Charles A., and J. John Gumperz (eds.),
1960 "Linguistic diversity in South Asia: Studies in regional, social and
functional variation", International Journal of American Linguistics 4,
No. 1 (entire issue).
Ferguson, Charles A., and Raleigh Morgan, Jr.,
1959 "Selected Readings in applied linguistics", Linguistic Reporter, Supple-
plement 2, 4 pp.
Ferru, Jean Louis,
1966 "Possible repercussions of a technical and economic nature of the
particular letters for the standard transcription of West African lan-
guages", Damako (Mali) Meeting on the Standardization of African
Alphabets, Feb. 28-March 5, 1966. UNESCO (CLT) Baling.
Findling, Joav,
1969 "Bilingual need affiliation and future orientation in extragroup and
intragroup domains", Modern Language Journal 53, 227-231; also in
J. A. Fishman (ed.), Advances in the Sociology of Language 2 (The
Hague: Mouton, 1971).
Fischel, A.,
1919 Der Panslawismus bis zum Weltkrieg (Stuttgart/Berlin: Cotta).
Fischer, John L.,
1958 "Social influences in the choice of a linguistic variant", Word 14, 47-56.
Fishman, Joshua A.,
1956 "The process and function of social stereotyping", Journal of Social
Psychology 43, 27-64.
1960 "A systematization of the Whorfian Hypothesis", Behavioral Science 8,
323-339.
1964 "Language maintenance and language shift as a field of inquiry".
Linguistics 9, 32-70.
1965a Yiddish in America (Bloomington, Ind.: Indiana University Research
Center in Anthropology, Folklore and Linguistics), Publication 36.
1965b "Bilingualism, intelligence and language learning", Modern Language
Journal 49, 227-237.
1965c "Varieties of ethnicity and language consciousness", Georgetown
University Monograph Series on Languages and Linguistics 18, 69-79.
1965d "Who speaks what language to whom and when?", Linguistique, No. 2,
67-88.
1965e "Language maintenance and language shift; The American immigrant
case within a general theoretical perspective", Sociologus 16, 19-38.
1965f "Language maintenance and language shift in certain urban immigrant
environments: The case .of Yiddish in the United States", Europa
Ethnica 22, 146-158.
1966a "Bilingual sequence at the societal level", On Teaching English to-
Speakers of Other Languages 2, 139-144.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
388 Joshua A. Fishman
1966b "Some contrasts between linguistically homogeneous and linguistically
heterogeneous polities", Social Inquiry 36, 146-158. Revised and ex-
panded in J. A. Fishman, C. A. Ferguson and J. Das Gupta (eds.),
Language Problems of Developing Nations (New York: Wiley, 1968),
53-68.
1966c Language Loyalty in the United States (The Hague: Mouton).
1966d "Planned reinforcement of language maintenance in the United States;
Suggestions for the conservation of a neglected national resource", in
J. A. Fishman, Language Loyalty in the United States (The Hague:
Mouton), Chapter 21.
1967a "The breadth and depth of English in the United States", University
Quarterly, March, 133-140.
1967b "A Sociology of Language (Review)", Language 43, 586-604.
1968a Readings in the Sociology of Language (The Hague: Mouton).
1968b "Sociolinguistic Perspective on the Study of Bilingualism", Linguistics
39, 21-50.
1968c "Sociolinguistics and the language problems of developing nations",
International Social Science Journal 20, 211-225.
1968d "A sociolinguistic census of a bilingual neighborhood", in Bilingualism
in the Barrio, J. A. Fishman, R. L. Cooper and Roxana Ma, et al.
(New York: Yeshiva University). Final Report to Department of
Health, Education, and Welfare under Contract No. OEC-1-7-062817-
0297; also American Journal of Sociology, 1969d, 75, 323-339.
1969a "National language and languages of wider communication in the
developing nations", Anthropological Linguistics 11, 111-135.
1969b "Literacy and the language barrier", Science 165, 1108-1109.
1969c "Bilingual attitudes and behaviors", Language Sciences, No. 5, 5-11;
also in J. A. Fishman, R. L. Cooper and Roxana Ma et al., Bilingual'
ism in the Barrio (New York: Yeshiva University, 1968). Final Report
to D H E W under contract No. OEC-1-7-062817-0297, and Blooming-
ton, Ind., Language Sciences Series, 1971.
1969d ' T h e impact of nationalism on language planning: some comparisons
between early 20th century Europe and subsequent developments in
South and South-East Asia", Consultative Meeting on Language Plan-
ning Processes (Honolulu: East-West Center, Institute of Advanced
Projects); also in Can Languages be Planned?, Joan Rubin and Bjorn
Jernudd (eds.) (Honolulu: East-West Center Press), 1971.
1969e "Puerto Rican intellectuals in New York; Some intragroup and inter-
group contrasts", Canadian Journal of Behavioral Sciences 1, 215-226.
In press "The links between micro- and macro-sociolinguistics in the study of
who speaks what language to whom and when", in John J. Gumperz
and Dell Hymes (eds.), The Ethnography of Communication: Direc-
tions in Sociolinguistics (New York: Holt, Rinehart and Winston); also
in J. A. Fishman, R. L. Cooper and Roxana Ma et al., Bilingualism in
the Barrio (Bloomington, Ind., Language Science Series, 1971).
Fishman, Joshua A., and Rebecca Agheyisi,
1970 "Language attitude studies", Anthropological Linguistics, 12, 137-157.
Fishman, Joshua A., Robert C. Cooper, and Roxana Ma et al.,
1968 Bilingualism in the Barrio. Final Report on Contract-OEC-1-7-062817-
0297 to D H E W (New York: Yeshiva University); also Bloomington,
Ind., Language Sciences Series, 1971.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 389
Fishman, Joshua A., and Lawrence Greenfield,
1970 "Situational measures of normative language views in relation to
person, place and topic among Puerto Rican bilinguals", Anthropos 65,
602-618; also in Advances in the Sociology of Language, Vol. 2, J. A.
Fishman (ed.) (The Hague: Mouton, 1972).
Fishman, Joshua A., and Eleanor Herasimchuk,
1969 "The multiple prediction of phonological variables in a bilingual speech
community", American Anthropologist 71, 648-657; also in J. A. Fish-
man (ed.), Advances in the Sociology of Language, Vol. 2 (The Hague:
Mouton, 1972).
Fishman, Joshua A., and John C. Lovas,
1970 "Bilingual education in sociolinguistic perspective", TESOL Quar-
terly, 4, 215-222.
Fishman, Joshua A., and Erika Lueders,
1970 "What has the sociology of language to say to the teacher?" (On
teaching the standard variety to speakers of dialectal or sociolectal
varieties), in C. Cazden, V. John and D. Hymes (eds.), The Functions
of Language (New York: Teachers College Press).
Fishman, Joshua A., and Vladimir C. Nahirny,
1966 "The ethnic group school in the United States", in J. A. Fishman et al.,
Language Loyalty in the United States (The Hague: Mouton), Chap-
ter 6; also Sociology of Education (1964), 37, 306-317.
Fishman, Joshua A., and Charles Terry,
1969 "The validity of sensus data on bilingualism in a Puerto Rican neigh-
borhood", American Sociological Review 34, 636-650. •
Frake, Charles O.,
1961 "The diagnosis of disease among the Subanun of Mindanao", American
Anthropologist 63, 113-132.
1962 "The ethnographic study of cognitive systems", in T. Gladwin and
William C. Sturtevant (eds.), Anthropology and Behavior (Washington,
D. C.: Anthropological Society of Washington), 77-85; also in J. A.
Fishman (ed.), Readings in the Sociology of Language (The Hague:
Mouton, 1968), 434-446.
Friederich, Paul,
1966 "The linguistic reflex of social change: From Tsarist to Soviet Russian
kinship", Sociological Inquiry 36, 159-185.
1962 "Language and politics in India", Daedalus, Summer, 543-559.
Gaarder, A. Bruce,
1965 "Teaching the bilingual child: Research, development and policy", The
Modern Language Journal 49, 165-175.
1967 "Organization of the bilingual school", Journal of Social Issues 23,
110-120.
Gallagher, Charles F.,
1967 "Language rationalization and scientific progress", Paper prepared for
Conference on Science and Social Change, California Institute of Tech-
nology, October 18-20.
1969 "Language reform and social modernization in Turkey", Consultative
Meeting on Language Planning Processes (Honolulu, EWS-IAP); sub-
sequently in Joan Rubin and Bjorn Jernudd (eds.), Can Language be
Planned? (Honolulu: East-West Center Press. 1971).
Garfinkel, Harold,
1967 Studies in Ethnomethodology (New York: Prentice Hall).

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
390 Joshua A. Fishman
Garfinkel, Harold and H. Sacks (eds.),
in press Contribution in Ethnomethodology (Bloomington: Indiana Uni-
versity Press).
Garvin, Paul L.,
1954 "Literacy as problem in language and culture", Georgetown Uni-
versity Monograph Series on Languages and Linguistics 7, 117-129.
1959 "The standard language problem: Concepts and methods", Anthropolo-
gical Linguistics 1, No. 2, 28-31.
Geertz, Clifford,
1960 "Linguistic etiquette", in his Religion of Java (Glencoe: Free Press);
also in J. A. Fishman (ed.), Readings in the Sociology of Language
(The Hague: Mouton, 1968), 282-295.
Gelb, I. F.,
1952 A Study of Writing (Chicago: University of Chicago Press).
Gerullis, Georg,
1932 "Muttersprache und Zweisprachigkeit in einem preussisch-litauischen
Dorf', Studi Baltici 2, 59-67.
Glenn, Norval D.,
1966 "The trend in differences in attitudes and behavior by educational
level", Sociology of Education 39, 255-275.
1967 "Differentiation and massification: Some trend data from national
surveys", Social Forces 46, 172-179.
Glenn, Norval D., and J. L. Simmons,
1967 "Are regional cultural differences diminishing?", Public Opinion Quar-
terly 31, 176-193.
Goodenough, Ward H.,
1965 "Rethinking 'status' and 'role': Toward a general model of the cultural
organization of social relationships", in M. Banton (ed.), The Relevance
of Models for Social Anthropology (New York: Praeger), 1-24.
Goodman, Elliot R.,
1960 "World state and world language", in his The Soviet Design for a
World State (New York: Columbia University Press), 264-284; also in
J. A. Fishman (ed.), Readings in the Sociology of Language (The Hague:
Mouton, 1968), 717-736.
Goody, Jack,
1963 Literacy in Traditional Societies (London: Cambridge University Press),
Greenberg, Joseph R.,
1965 "Urbanism, migration and language", in Hilda Kuper (ed.), Urbaniza-
tion and Migration in West Africa (Los Angeles and Berkeley: Uni-
versity of California Press), 50-59.
1966 Universals of Language, 2nd edition (Cambridge, Mass.: The MIT
Press).
Greenfield, Lawrence,
1968 "Situational measures of language use in relation to person, place and
topic among Puerto Rican bilinguals", Bilingualism in the Barrio. Final
Report to DHEW re Contract No. DEC-1-7-062817-0297 (New York:
Yeshiva University).
Grimshaw, Allen D.,
1969 "Sociolinguistics and the sociologist", The American Sociologist 4,
312-321.
in press "Some social sources and some social functions of pidgin and
creole languages", in D. Hymes (ed.), Proceedings of the Social Science

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 391
Research Council Conference on Creolization and Pidginization (Cam-
bridge: Cambridge University Press).
Gulick, John,
1958 "Language and passive resistance among the eastern Cherokees", Ethno-'
history 5, 60-81.
Gumperz, John J.,
1958 "Dialect differences and social stratification in a North Indian village",
American Anthropoligst 60, 668-682.
1961 "Speech variation and the study of Indian civilization", American An-
thropologist 63, 976-988.
1962 "Types of linguistic communities", Anthropological Linguistics 4, No. 1,
28-40; also in J. A. Fishman (ed.), Readings in the Sociology of
Language (The Hague: Mouton, 1968), 460-476.
1964a "Linguistic and social interaction in two communities", American An-
thropologist 66, No. 2, 37-53.
1964b "Hindi-Punjabi code switching in Delhi", in Morris Halle (ed.), Pro-
ceedings of the International Congress of Linguistics (The Hague:
Mouton).
1966 "On the ethnology of linguistic change", in William Bright (ed.), Socio-
linguistics (The Hague: Mouton), 27-38.
1967 "The linguistic markers of bilingualism", Journal of Social Issues 23,
No. 2, 48-57.
Guxman, M. M.,
1960 "Some general regularities in the formation and development of na-
tional languages", in M. M. Guxman (ed.), Voprosy Formirovanija
Nacional'nyx Jazykov (Moscow), 295-307; also in J. A. Fishman (ed.),
Readings in the Sociology of Language (The Hague: Mouton, 1968),
766-779.
Hair, P. E. H„
1963 "Notes on the discovery of the Vai script, with a bibliography", Sierra
Leone Language Review 2, 36-49.
Hall, Robert A., Jr.,
1952 "Bilingualism and applied linguistics", Zeitschrift für Phonetik und
allgemeine Sprachwissenschaft 6, 13-30.
Halliday, M. A. K., Angus Macintosh, and Peter Strevens,
1964a The Linguistic Sciences and Language Teaching (London: Longmans).
1964b "The users and uses of language", in H. A. K. Halliday, A. Macintosh
and P. Strevens ^eds.), The Linguistic Sciences and Language Teaching
(London: Longmans, Green); also in J. A. Fishman (ed.), Readings in
the Sociology of Language (The Hague: Mouton, 1968), 139-169.
Hamilton, Richard F.,
1965 "Affluence and the worker: The West German case", American Jour-
nal of Sociology 71, 144-152.
Hansen, Marcus L.,
1940 The Immigrant in American History, Arthur M. Schlesinger (ed.) (Cam-
brige: Harvard University Press).
Haudricourt, A. G.,
1943 "De l'origine des particularités de l'alphabet Vietnamien", Dan Viet-
nam, No. 3, 61-68.
Haugen, Einar,
1953 The Norwegian Language in America; A Study in Bilingual Behavior
(Philadelphia: University of Pennsylvania Press), 2 volumes; 2nd ed.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
392 Joshua A. Fishman

(Bloomington: Indiana University Press, 1969).

1961 "Language planning in modern Norway", Scandinavian Studies 33, 68-
81; also in J. A. Fishman (ed.), Readings in the Sociology of Language
(The Hague: Mouton, 1968), 673-687.
1966a Language Planning and Language Conflict; The Case of Modern Nor-
wegian (Cambridge: Harvard University Press).
1966b "Linguistics and language planning", in Wm. Bright (ed.), Sociolinguis-
tics (The Hague: Mouton), 50-66.
1966c "Dialect, Language, Nation", American Anthropologist 68, 922-935.
Havranek, Bohuslav,
1964 "The functional differentiation of the standard language", in Paul L.
Garvin (ed.), A Prague School Reader on Esthetics, Literary Structure
and Style (Washington, D. C.: Georgetown University Press), 1-18.
Hayden, Robert G., and Joshua A. Fishman,
1966 ' T h e impact of exposure to ethnic mother tongues on foreign language
teachers in American high schools and colleges", in J. A. Fishman et
al., Language Loyalty in the United States (The Hague: Mouton), Chap-
ter 13; also, Modern Language Journal (1964), 48, 262-274.
Hayes, Alfred S.,
1965 Recommendations of the Work Conference on Literacy (Washington:
Center for Applied Linguistics).
Herman, Simon N.,
1961 "Explorations in the social psychology of language choice", Human
Relations 14, 149-164; also in J. A. Fishman (eds.), Readings in the
Sociology of Language (The Hague: Mouton, 1968), 492-511.
Hertzler, Joyce O.,
1965 The Sociology of Language (New York: Random House).
Herzog, Marvin I.,
1965 The Yiddish language in Northern Poland: Its Geography and History
(Bloomington, Ind.: Indiana University Research Center in Anthro-
pology, Folklore and Linguistics), Publication 37.
Heyd, Uriel,
1954 Language Reform in Modern Turkey (Jerusalem: Israel Oriental So-
ciety).
Hodges, Harold M.,
1964 Social Stratification: Class in America (Cambridge: Schenkman).
Hofman, John E.,
1966a "Mother tongue retentiveness in ethnic parishes", in Joshua A. Fish-
man et al., Language Loyalty in the United States (The Hague: Mouton).
1966b "The language transition in some Lutheran denominations", in J. A.
Fishman et al., Language Loyalty in the United States (The Hague:
Mouton), Chapter 10; also in J. A. Fishman (ed.), Readings in the
Sociology of Language (The Hague: Mouton, 1966), 620-638.
Hofman, John E. and Haya Fisherman,
1971 "Language Shift and Maintenance in Israel", International Migration
Review, 5, no. 2, 204-226; also in J. A. Fishman (ed.), Advances in the
Sociology of Language, Vol. 2 (The Hague: Mouton, 1972).
Hohenthal, W. D., and T. McCorkle,
1955 "The problem of aboriginal persistence", Southwestern Journal of
Anthropology 11, 288-300.
Hoijer, H.,
1951 "Cultural implications of the Navaho linguistic categories", Language

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 393
27, 111-120.
1954 ' T h e Sapir-Whorf hypothesis", in H. Hoijer (ed.), Language in Culture
(= American Anthropological Association, Memoir No. 79) (Chicago:
University of Chicago Press), 92-104.
Holton, Daniel C.,
1947 "Ideographs and Ideas", Far Eastern Survey 16, No. 19, 220-223.
Howell, Richard W.,
1965 "Linguistic status markers in Korean", Kroeber Anthropological So-
ciety Papers 55, 91-97.
1967 "Terms of address as indices of social change", paper presented at
American Sociological Association Meeting, San Francisco, Sept. 1967.
Huf fine, Carol L.,
1966 "Inter-Socio-Economic Clan Language Differences: A Research Re-
port", Sociology and Social Research 50, 3, 351-355.
Hughes, Everett C.,
1970 "The linguistic division of labor in industrial and urban societies",
Georgetown University Monograph Series on Languages and Linguis-
tics, in press; also in J. A. Fishman (ed.), Advances in the Sociology
of Language, Vol. 2 (The Hague: Mouton, 1972).
Hsia, Tao-tai,
1956 China's Language Reforms (New Haven: Yale University Press).
Hymes, Dell H.,
1962 ' T h e ethnography of speaking", in T. Gladwin and W. C. Sturtevant
(eds.), Anthropology and Human Behavior (Washington, D. C.: Anthro-
pology Society of Washington), 13-53; also in J. A. Fishman (ed.),
Readings in the Sociology of Language (The Hague: Mouton, 1968),
99-138.
1966 "Two types of linguistic relativity", in Wm. Bright (ed.), Sociolinguis-
tics (The Hague: Mouton), 114-157.
1967a "Why Linguistics Needs the Sociologist", Social Research 34, 7, 632-
647.
1967b "Models of interaction of language and social setting", Journal of
Social Issues 23, No. 2, 8-28.
Ivanova, A. M., and V. D. Voskresensky,
1959 "Abolition of adult illiteracy in USSR, 1917-1940", Fundamental and
Adult Education 11, 3, 131-186.
Jeffries, W. F.,
1958 "The literacy campaign in Northern Nigeria", Funding and Adult
Education 10, 1, 2-6.
Jernudd, Bjorn,
1969 "Notes on economic analysis and language planning", Consultative
Meeting on Language Planning Processes (Honolulu: EWC-IAP); sub-
sequently, in Joan Rubin and Bjorn Jernudd (eds.), Can Language be
Planned? (Honolulu: East West Center Press, 1971).
Jernudd, Bjorn, and Jyotirindra Das Gupta,
1969 "Towards a theory of language planning", Consultative Meeting on
Language Planning Processes (Honolulu: EWC-IAP); subsequently, in
Joan Rubin and Bjorn Jernudd (eds.), Can Language be Planned?
(Honolulu: East-West Center Press, 1971).
Jernudd, Bjorn H., and Tommy Willingsson,
1968 "A sociolectal study of the Stockholm region", Svenska Landsmal och
Svenskt Folkliv 289, 140-147.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
394 Joshua A. Fishman
John, Vera, and Vivian Horner,
1971 Early Childhood Bilingual Education (New York: Modern Language
Association).
Johnston, Ruth,
1963a "Factors in the Assimilation of Selected Groups of Polish Post-War
Immigrants in Western Australia", Unpublished Ph.D. Dissertation,
University of Western Australia (Perth); subsequently: Immigrant As••
similation; A Study of Polish People in Western Australia (Perth: Pater-
son Brokensha, 1965).
1.963b "A new approach to the meaning of assimilation", Human Relations
16, 295-298.
Jones, D.,
1942 Problems of a National Script for India (Hartford: Hartford Seminary
Foundation).
Jones, Frank E., and Wallace E. Lambert,
1959 "Attitudes toward immigrants in a Canadian community", Public
Opinion Quarterly 23, 538-546.
Joos, Martin,
1950 "Description of language design", Journal of the Acoustical Society
of America 22, 701-708. Reprinted in his Readings in Linguistics
(Washington, D. C.: American Council of Learned Societies, 1958),
349-356.
1959 "The isolation of styles", Georgetown University Monograph Series
on Languages and Linguistics 12, 107-113; also in J. A. Fishman (ed.),
Readings in the Sociology of Language (The Hague: Mouton, 1968),
185-191.
Kandier, C.,
1955 "Zum Aufbau der angewandten Sprachwissenschaft", Sprachforum 1,
3-9.
Kandori, Takehiko,
1968 "Study of dialects in Japan", Orbis 17, 47-56.
Kantrowitz, Nathan,
1967 'The vocabulary of race relations in a prison", Paper presented at
American Dialect Society Meeting, Chicago, December, 1967.
Kaye, Alan S.,
1970 "Modern standard Arabic and the colloquials", Lingua 24, 374-391.
Kelman, Herbert C.,
1969 "Language as aid and barrier to involvement in the national system",
Consultative Meeting on Language Planning Processes (Honolulu: EWC-
IAP); subsequently, in Joan Rubin and Bjorn Jernudd (eds.), Can Lan-
guage be Planned? (Honolulu: East-West Center Press, 1971); also in
J. A. Fishman (ed.), Advances in the Sociology of Language 2 (The
Hague: Mouton, 1972).
Kimple, James, Jr., Robert L. Cooper, and Joshua A. Fishman,
1969 "Language switching in the interpretation of conversations", Lingua 23,
127-134.
Kloss, Heinz,
1929 "Sprachtabellen als Grundlage für Sprachstatistik, Sprachenkarten und
für eine allgemaine Soziologie der Sprachgemeinschaften", Vierteljahrs-
schrift für Politik und Geschichte 1(7), 103-117.
1952 Die Entwicklung Neuer Germanischer Kultursprachen (Munich, Pohl).
1966a "Types of multilingual communities: A discussion of ten variables",

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 395
Sociological Inquiry 36, 135-145.
1966b "German-American language maintenance efforts", in J. A. Fishman
et al., Language Loyalty in the United States (The Hague: Mouton),
Chapter 15.
1967 " 'Abstand' languages and 'Ausbau' languages", Anthropological Lin-
guistics 9, No. 7, 29-41.
Kluckhohn, Clyde,
1961 "Notes on some anthropological aspects of communication", American
Anthropologist 63, 895-910.
Knott, Betty I.,
1956 "The Christian 'special language' in the inscriptions", Vigiliae Chris-
tianae 10, 65-79.
Kohn, Hans,
1945 The Idea of Nationalism: A Study of its Origin and Background (New
York: Macmillan).
Kolarz, Walter,
1946 Myths and Realities in Eastern Europe (London: Lindsay Drummond).
1967 Russia and her Colonies (Hamden, Conn.: Archon Books) (originally
published in 1952).
1969 The Peoples of the Soviet Far East (Hamden, Conn.: Archon Books)
(originally: 1954).
Kuhn, Walter,
1930 Die jungen deutschen Sprachinseln in Galizien: ein Beitrag zur Methode
der Sprachinselforschung (Münster: Aschendorffsche Verlagsbuchhand-
lung).
1934 Deutsche Sprachinselforschung (Plauen: Gunther Wolff).
Kwan-Wai Chiu, Rosaline,
1970 Language Contact and Language Planning in China (1900-1967); A
Selected Bibliography (Quebec: Les Presses de l'Université Laval).
Labov, William,
1963 "The social motivation of a sound change", Word 19, 273-309.
1964 "Phonological correlates of social stratification", American Anthro-
pologist 66, No. 2, 164-176.
1965 "On the mechanism of linguistic change", Georgetown University
Monograph Series in Language and Linguistics 18, 91-114.
1966a "The effect of social mobility on linguistic behavior", Sociological
Inquiry 36, 186-203.
1966b "Hypercorrection by the lower middle class as a factor in linguistic
change", in Wm. Bright (ed.), Sociolinguistics (The Hague: Mouton),
84-101.
1966c The Social Stratification of English in New York City (Washington:
Center for Applied Linguistics).
1966d "Stages in the acquisition of standard English", in Roger W. Shuy (ed.),
Social Dialects and Language Learning (Champaign: NCTE).
1968 "The reflection of social processes in linguistic structures", in J . A.
Fishman (ed.), Readings in the Sociology of Language (The Hague:
Mouton), 240-251.
Labov, William, Paul Cohen, Clarence Robins, and John Lewis,
1968 A Stuây of the Non-Standard English of Negro and Puerto Rican
Speakers in New York City. Final Report, Cooperative Research Proj-
ect No. 3288 (New York: Columbia University), 2 vols.
Lambert, Wallace E., R. C. Gardner, H. C. Barick, and K. Tunstall,

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
396 Joshua A. Fishman
1963 "Attitudinal and cognitive aspects of intense study of a second lan-
guage", Journal of Abnormal and Social Psychology 66, 358-368.
Lanham, L. W.,
1965 'Teaching English to Africans: A crisis in education", Optima 15,
December, 197-204.
Lantz, D e lee, and Volney Stefflre,
1964 "Language and cognition revisited", Journal of Abnormal and Social
Psychology 49, 454-462.
Lehmann, A.,
1889 "Uber Wiedererkennen", Philos. Stud. 5, 96-156.
Lenneberg, Eric H.,
1953 "Cognition in ethnolinguistics", Language 29, 463-471.
1957 "A probabilistic approach to language learning", Behavioral Science 2,
1-12.
LePage, Robert,
1964 The National Language Question (London: Oxford University Press).
1968 "Problems to be faced in the use of English as the medium of educa-
tion in four West Indian territories", in J. A. Fishman, C. A. Ferguson
and J. Das Gupta (eds.), Language Problems of Developing Nations
(New York: Wiley), 431-441.
Levine, William L., and H. J. Crockett,
1966 "Speech variation in a Piedmont community: Postvocalic r", Socio-
logical Inquiry 36, 204-226.
Lewis, E. Glynn,
1971 "Migration and language in the USSR", International Migration Re-
view, 5, no. 2, 147-179; also in J. A. Fishman (ed.), Advances in the
Sociology of Language, Vol. 2 (The Hague: Mouton, 1972).
Lewis, W. Arthur,
1961 "Education and economic development", Social and Economic Studies
10, No. 2, 113-127.
Li Hui,
I960 "The phonetic alphabet - short cut to literacy", Peking Review 13,
No. 28 (July 12).
Lieberson, Stanley,
1965 "Bilingualism in Montreal: A demographic analysis", American Jour-
nal of Sociology 71, 10-25; also in J. A. Fishman (ed.), Advances in
the Sociology of Language, Vol. 2 (The Hague: Mouton, 1972).
1971 "Language shift in the United States: some demographic clues", Inter-
national Migration Review 5, no. 2.
Lind, Andrew W.,
1969 "Race relations in New Guinea", Current Affairs Bulletin (Sydney:
Australia) 44, No. 3, 34-48.
Lindenfeld, Jacqueline,
1969 'The social conditioning of syntactic variation in French", American
Anthropologist 71, 890-898; also in J. A. Fishman (ed.), Advances in
the Sociology of Language, Vol. 2 (The Hague: Mouton, 1972).
Lunt, Horace G.,
1959 "The creation of Standard Macedonian: Some facts and attitudes",
Anthropological Linguistics 1, No. 5, 19-26.
Ma, Roxana, and Eleanor Herasimchuk,
1968 "The linguistic dimensions of a bilingual neighborhood", in J. A. Fish-
man, R. L. Cooper and Roxana Ma et al., Bilingualism in the Barrio

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 397
(New York, Yeshiva University). Final Report to Department of Health,
Education and Welfare under Contract No. OEC-1-7-062817-0297; also
in J. A. Fishman (ed.), Advances in the Sociology of Language, Vol. 2
(The Hague: Mouton, 1972).
Mackey, William F.,
1962 "The description of bilingualism", Canadian Journal of Linguistics 7,
51-85.
1970 "A typology of bilingual education", The Foreign Language Annals 3,
596-608; also in J. A. Fishman (ed.), Advances in the Sociology of
Language, Vol. 2 (The Hague: Mouton, 1972).
Macnamara, John,
1966 Bilingualism in Primary Education (Edinburgh, Edinburgh University
Press).
1967 "The effects of instruction in a weaker language", Journal of Social
Issues 23, 121-135.
1969 "Successes and failures in the movement for the restoration of Irish",
Consultative Meeting on Language Planning Processes (Honolulu: EWC-
LAP), subsequently in Joan Rubin and Bjorn Jernudd (eds.), Can Lan-
guage be Planned? (Honolulu: East-West Center Press, 1971).
Maier, Norman R. F.,
1930 "Reasoning in humans, I. On direction", Journal of Comparative Psy-
chology 10, 115-143.
Manuel, Herschel T.,
1963 The Preparation and Evaluation of Interlanguage Testing Materials
(Austin: University of Texas). Mimeographed report, Cooperative Re-
search Project Number 681.
Mayner, Thomas F.,
1967 "Language and nationalism in Yugoslavia", Canadian Slavic Studies 1,
333-347.
Mazrui, Ali A.,
1967 "The national language question in East Africa", East Africa Journal,
No. 3, 12-19.
McClusker, Henry F., Jr.,
1963 An Approach for Educational Planning in the Developing Countries
(Menlo Park: Stanford Research Institute).
McCormack, William,
1960 "Social dialects in Dharwar Kannada", In C. A. Ferguson and J. J.
Gumperz (eds.), "Linguistic Diversity in South Asia", UAL 4, No. 1,
79-91.
1968 "Occupation and residence in relation to Dharwar dialects", in M.
Singer and B. S. Cohn (eds.), Social Structure and Social Change in
India (New York: Viking Fund), 475-486.
McDavid, Raven I.,
1958 "The dialects of American English", in W. N . Francis (ed.), The Struc-
ture of American English (New York: Ronald).
Metraux, Ruth W.,
1965 "A study of bilingualism among children of U.S.-French parents",
French Review 38, 650-655.
Metzger, Duane, and Gerald E. Williams,
1966 "Some procedures and results in the study of native categories: Tzeltal
'firewood' ", American Anthropologist 68, 389-407.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
398 Joshua A. Fishman
Meyenburg, Erwin,
1934 "Der heutige Stand der Romazi-Bewegung in Japan", Forschungen und
Fortschritte 10, No. 23-24.
Mills, II. C.,
1956 "Language reform in China", Par Eastern Quarterly 15, 517-540.
Mohrmann, Christine,
1947 "Le latin commun et le latin des Chretiens", Vigiliae Christiannae 1,
1-12.
1957 "Linguistic problems in the Early Christian Church", Vigiliae Chris-
tiannae 11, 11-36.
Morag, Shelomo,
1959 "Planned and unplanned development in modern Hebrew", Lingua 87,
247-263.
Nader, Laura,
1962 "A note on attitudes and the use of language", Anthropological Lin-
guistics 4, No. 6, 24-29; also in J. A. Fishman (ed.), Readings in the
Sociology of Language (The Hague: Mouton, 1968), 276-281.
Nahirny, Vladimir C., and Joshua A. Fishman,
1965 "American immigrant groups: ethnic identification and the problem of
generations", Sociological Review 13, 311-326.
Nelson, Lowry,
1947 "Speaking of tongues", American Journal of Sociology 54, 202-210.
Nguyen dinh Hoa,
1960 The Vietnamese I Mguage (Saigon: Department of National Education)
(= Vietnam Culture Series, No. 2).
Nida, Eugene A.,
1954 "Practical limitations to a phonemic orthography", Bible Translator 5,
35-39 and 58-62.
1967 "Sociological dimensions of literacy and literature"; Chapter 11 of
Floyd Shacklock et al. (eds.), World Literacy Manual (New York: Com-
mittee on Literacy and Christian Literature).
Orenstein, Jacob,
1959 "Soviet language policy: Theory and practice", Slavic and East Euro-
pean Journal 17, 1-24.
Osgood, Charles E.,
1960 "The cross-cultural generality of visual-verbal-synesthetic tendencies",
Behavioral Science 5, 146-149.
Owens, Roger C.,
1965 'The patrilocal band: A linguistically and culturally hybrid social unit",
American Anthropologist 67, 675-690.
özmen, Yücel,
"A Sociolinguistic Analysis of Language Reform in Turkey 1932-1967,
with Special Reference to the Activities of the Turk Dil Kurumu", MS
Thesis, Georgetown University, 1967 (unpublished).
Passin, Herbert,
1963 "Writer and journalist in the transitional society", in Lucian W. Pye
(ed.), Communication and Political Development (Princeton: Princeton
University Press), 82-123; also in J. A. Fishman, C. A. Ferguson and
J. Das Gupta (eds.), Language Problems of Developing Nations (New
York: Wiley, 1968), 442-458.
Pike, Kenneth L.,
1947 Phonemics: A Technique for Reducing Languages to Writing (Ann

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 399
Arbor: University of Michigan Press).
Polome, Edgar,
1968 ' T h e choice of official languages in the Democratic Republic of the
Congo", in J. A. Fishman, C. A. Ferguson and J. Das Gupta (eds.),
Language Problems of Developing Countries (New York: Wiley), 295-
312.
Pool, Jonathan,
1969 "National development and language diversity", La Monda Lingvo-
Problemo 1, 129-192; also in J. A. Fishman (ed.), Advances in the
Sociology of Language, Vol. 2 (The Hague: Mouton, 1972).
Pospisil, Leopold,
1965 "A formal semantic analysis of substantive law: Kapauka Papuan laws
of land tenure", American Anthropologist 67, Part 2, 186-214.
Prator, Clifford H„
1967 "The Survey of Language Use and Language Teaching in Eastern
Africa", Linguistic Reporter 9, No. 8.
Price, P. David,
1967 "Two types of taxonomy: A Huichol ethnobotanical example", Anthro-
pological Linguistics 9, No. 7, 1-28.
Quelquejay, C., and A. Bennigsen,
1961 The Evolution of the Muslim Nationalities of the USSR and Their
Linguistic Problems (London: Central Asian Research Center).
Rabin, Chaim,
1969 "Spelling reform: Israel, 1968", Consultative Meeting on Language
Planning Processes (Honolulu: EWC-IAP); subsequently in Joan Rubin,
Bjora Jernudd (eds.), Can Language be Planned? (East-West Center
Press, 1971).
Radin, Paul,
1924 "The adoptation of an alphabet by an aboriginal people", Cambridge
University Reporter (Proceedings of the Cambridge Philological So-
ciety), Nov. 25, 27-34.
Ramos, Maximo, Jose V. Aguiler and Bonifacio P. Sibayan, The Determination
and Implementation of Language Policy (= Philippine Center for Lan-
guage Study, Monograph 2) (Quezon City: Alemar-Phoenix).
Ray, Punya Sloka,
1963 Language Standardization (Chapter 9: Comparative description and
evaluation of writing systems, 106-120) (The Hague: Mouton).
Read, Allen Walker,
1967 "The splitting and coalescing of widespread languages", Proceedings
of the Ninth International Congress of Linguistics, 1129-1134.
Reissman, Leonard,
1964 The Urban Process: Cities in Industrial Societies (New York: Free
Press).
Ross, Allan S. C.,
1956 "U and non-U; An essay in sociological linguistics", in N. Mitford (ed.),
Noblesse Oblige (London, Hamish Hamilton), 11-38.
Rossi, Ettore,
1927 "La questione dell' alfabeto per le lingue turche", Oriente Moderno 7,
295-310.
1929 "II nuovo alfabeto latino introdotto in Turchia", Oriente Moderno 9,
32-48.
1935 "La riforma linguistica in Turchia", Oriente Moderno 15, 45-57.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
400 Joshua A. Fishman
1942 "Un decennio di rifonna linguistica in Turchia", Oriente Moderno 22,
466-477.
1953 "Venticinque anni di rivoluzione dell'alfabeto e venti di riforma lin-
guistica in Turchia", Oriente Moderno 33, 378-384.
Royal Commission on Bilingualism and Biculturism,
1965 A Preliminary Report (Ottawa, Queen's Printer).
1967 Book 1: General Introduction, The Official Languages (Ottawa: Queen's
Printer).
1968 Book II: Education (Ottawa: Queen's Printer).
Rubin, Joan,
1962 "Bilingualism in Paraguay", Anthropological Linguistics 4, No. 1, 52-
58.
1968 "Language and education ino Paraguay", in J. A. Fishman, C. A.
Ferguson and J. Das Gupta (eds.), Language Problems in Developing
Nations (New York: Wiley), 477-488.
1969 "Education and language planning", Consultative Meeting on Language
Planning Processes (Honolulu: EWC-IAP), subsequently in Joan Rubin
and Bjorn Jernudd (eds.), Can Language be Planned? (Honolulu: East-
West Center Press, 1971); also in J. A. Fishman (ed.), Advances in the
Sociology of Language, Vol. 2 (The Hague: Mouton, 1972).
Russell, J. K., "Starting a literacy campaign", Books for Africa 18, No. 2, 17-20.
Rustow, Dankwart A.,
1967 A World of Nations: Problems of Political Modernization (Washington:
Brookings Institution); also, adapted as "Language, modernization and
nationhood - an attempt at typology", in J. A. Fishman, C. A. Fergu-
son and J. Das Gupta (eds.), Language Problems of Developing Nations
(New York: Wiley, 1968), 87-106.
Samora, Julian, and Wm. N . Deane,
1956 "Language usage as a possible index of acculturation", Sociology and
Social Research 40, 307-311.
Scharshmidt, Clemens,
1924 "Schriftreform in Japan: Ein Kulturproblem", Mitteilungen des Semi-
nars für Orientalische Sprachen 26/27, No. 1, 183-186.
Schegloff, Emanuel A.,
1968 "Sequencing in conversational openings", American Anthropologist 70,
1075-1095; also in J. A. Fishman (ed.), Advances in the Sociology of
Language (The Hague: Mouton, 1971).
Schermerhorn, Richard A.,
1964 "Toward a general theory of minority groups", Phyton 25, 238-246.
Schofield, R. S.,
1968 "The Measurement of Literacy in Pre-Industrial England", in Jock
Goody (ed.), Literacy in Traditional Societies (London: Cambridge
University Press).
Shnore, Leo,
1966 ' T h e rural-urban variable: A n urbanite's perspective", Rural Sociology
21, 137.
Selk, Paul,
1937 Die sprachlichen Verhaltnisse im deutsch-dänischen Sprachgebiet süd-
lich der Grenze (Flensburg: Verlag Heimat und Erbe) (Ergänzungsband,
1940).
Serdyuchenko, G. P.,
1962 "The eradication of illiteracy and the creation of new written languages

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 401
in the USSR", International Journal of Adult and Youth Education 14,
1, 23-29.
1965 Elimination of illiteracy among the people who had no Alphabets
(Moscow, USSR: Commission for UNESCO, Ministry of Education,
RSFSR), 16 pp.
Serruys, Paul L. M.,
1962 Survey of the Chinese Language Reform and the Anti-Illiteracy Move-
ment in Communist China (= Studies in Communist Chinese Termi-
nology, No. 8) (Berkeley: Center for Chinese Studies, UC-B).
Shuy, Roger W. (ed.),
1966 Social Dialects and Language Learning (Champaign: NCTE).
Sjoberg, Andree F.,
1964 "Writing, speech and society: Some changing interrelationships", in
Proceedings of the Ninth International Congress of Linguists (The
Hague: Mouton), 892-897.
1966 "Socio-cultural and linguistic factors in the development of writing
systems for preliterate peoples", in Wm. Bright (ed.), Sociolinguistics
(The Hague: Mouton), 260-276.
Smalley, William A.,
1964 Orthography Studies: Articles on New Writing Systems (London: United
Bible Societies) (Help for Translators, Vol. 6).
Smith, Alfred G.,
"Literacy Promotion in an Underdeveloped Area", Madison, University
of Wisconsin, Unpublished Ph.D. thesis, 1956.
Sorensen, Arthur P., Jr.,
1967 "Multilingualism in the Northwest Amazon", American Anthropologist
69, 670-684.
Spencer, John (ed.),
1963 Language in Africa (Cambridge: Cambridge University Press).
Stern, Theodore,
1968 "Three Pwo Karen scripts: A study of alphabet formation", Anthro-
pological Linguistics 10, No. 1, 1-39.
Stewart, Gail,
1967 "Present-day usage of the Vai script in Liberia", African Language
Review 7, 71-74.
Stewart, Wm. A.,
1964 Non-Standard Speech and the Teaching of English (Washington: Center
for Applied Linguistics).
1965 "Sociolinguistic factors affecting English teaching", in Roger W. Shuy
(ed.), Social Dialects and Language Learning (Champaign: NCTE), 10-
18.
1968 "A sociolinguistics typology for describing national multilingualism",
in J. A. Fishman (ed.), Readings in the Sociology of Language (The
Hague: Mouton), 531-545.
Szajkowski, Zosa,
1966 Catalogue of the exhibition on the history of Yiddish orthography
from the spelling rules of the early sixteenth century to the standard-
ized orthography of 1936 (New York: Yivo Institute for Jewish Re-
search).
Tabouret-Keller, Andree,
1968 "Sociological factors of language maintenance and language shift: A
methodological approach based on European and African examples",

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
402 Joshua A. Fishman
in J. A. fishman, C. A. Ferguson and J. Das Gupta (eds.), Language
Problems of Developing Nations (New York: Wiley), 107-118.
Tanner, Nancy,
1967 "Speech and society among the Indonesian elite: A case study of a
multilingual community", Anthropological Linguistics 9, No. 3, 15-40.
Tauli, Valter,
1968 Introduction to a Theory of Language Planning (Uppsala: Acta Uni-
versitatis Upsaliensis ( = Studia Philologiae Scandinavicae Upsaliensia).
Thompson, Laurence C.,
1965 A Vietnamese Grammar (Seattle: University of Washington Press).
Thorburn, Thomas,
1969 "Cost-benefit analysis in language planning", Consultative Meeting on
Language Planning Processes (Honolulu: EWC-IAP); subsequently in
Joan Rubin and Bjorn Jernudd (eds.), Can Language be Planned?
(Honolulu: East West Center Press, 1971); also in J. A. Fishman (ed.),
Advances in the Sociology of Language (The Hague: Mouton, 1972).
Tietze, Andreas,
1962 "Problems of Turkish lexicography", 1JAL 28, no. 2, part 4, 263-272
( = Publication no. 21 of Indiana Univ. Center of Anthropology, Folklore
and Linguistics).
Toshio, Yamada,
1967 "The writing system: Historical research and modern development",
Current Trends in Linguistics 2, 693-731.
Trager, George L.,
1959 "The systematization of the Whorf hypothesis", Anthropological Lin-
guistics 1, No. 1, 31-35.
Twadell, W. I.,
1959 "Standard German", Anthropological Linguistics 1, No. 3, 1-7.
UNESCO,
1951 The Haiti Pilot Project (Paris: UNESCO).
Vachek, Joseph,
1945-1949 "Some remarks on writing and phonetic transcription", Acta Lin-'
guistica (Copenhagen), 5, 86-93.
1948 "Written language and printed language", Recueil Linguistique de
Bratislava 1, 67-75. Reprinted in J. Vachek (ed.), A Prague School
Reader in Linguistics (Bloomington: Indiana University Press, 1964),
453-560.
Valdman, Albert,
1968 "language standardization in a diglossia situation: Haiti", in J. A.
Fishman, C. A. Ferguson and J. Das Gupta (eds.), Language Problems
of Developing Nations (New York: Wiley), 313-326.
Verdoodt, Albert,
1971 "The differential impact of immigrant French speakers on indigenous
German speakers: A case study in the light of two theories", Interna-
tional Migration Review,• 5, no. 2, 138-146; also in J. A. Tishman (ed.),
Advances in the Sociology of Language, Vol. 2 (The Hague: Mouton, 1972).
Walker, Willard,
1969 "Notes on native writing systems and the design of native literacy
programs", Anthropological Linguistics 11, No. 5, 148-166.
Ward, Ida C.,
1945 Report of an Investigation of some Gold Coast Language Problems
(London: Crown Agents for the Colonies).

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
The Sociology of Language 403
Weinreich, Max,
1953 "Yidishkayt and Yiddish: On the impact of religion on language in
Ashkenazic Jewry", in Mordecai M. Kaplan Jubilee Volume (New
York: Jewish Theological Seminary of America); also in J. A. Fishman
(ed.), Readings in the Sociology of Language (The Hague: Mouton,
1968), 382-413.
1967 "The reality of Jewishness versus the ghetto myth: The sociolinguistic
roots of Yiddish", in To Honor Roman Jacobson (The Hague: Mouton),
2199-2211.
Weinreich, Uriel,
1951 "Research problems in bilingualism, with special reference to Switzer-
land". Unpublished Ph.D. dissertation. Columbia University.
1953a Languages in Contact (New York: Linguistic Circle of New York).
1953b "The Russification of Soviet Minority Languages", Problems of Com-
munism 2(6), 46-57.
1962 "Multilingual Dialectology and the New Yiddish Atlas", Anthropologi-
cal Linguistics 4, No. 1, 6-22.
White, John K„
1962 "On the revival of printing in the Cherokee Janguage", Current An-'
thropology 3, 511-514.
Whiteley, W. H.,
1968 "Ideal and reality in national language policy: A case study from
Tanzania", in J. A. Fishman, C. A. Ferguson and J. Das Gupta (eds.),
Language Problems of Developing Nations (New York: Wiley), 327-344.
Whorf, Benjamin L.,
1940 "Science and linguistics", Technology Review 44, 229-231, 247-248.
1941 'The relation of habitual thought to behavior and to language", in
L. Speier (ed.), Language, Culture and Personality (Menasha, Wise.:
Sapir Memorial Publication Fund), 75-93.
Willems, Emilio,
1943 "Linguistic changes in German-Brazilian communities", Acta Americana
1, 448-463.
Winner, T. G.,
1952 "Problems of Alphabetic Reform among the Turkic Peoples of Soviet
Central Asia", Slavonic and East European Review, 132-147.
Wittermans, Elizabeth P.,
1967 "Indonesian terms of address in a situation of rapid social change",
Social Forces 46, 48-52.
Wolff, Hans,
1954 Nigerian Orthography (Zaria: Gaskiya Corp.).
Wolfram, Walter A.,
1969 A Sociolinguistic Description of Detroit Negro Speech (Washington:
Center for Applied Linguistics).
Wonderly, William L.,
1968 Bible Translation for Popular Use (London: United Bible Societies).
Wu Yu-Chang,
1956 "Concerning the (draft) Han Language Phonetization Plan", Current
Background, No. 380 (March 15), 4-20.
1964 "Widening the use of the phonetic script", China Reconstructs 13,
No. 6, 29-31.
1965 "Report of the current tasks of reforming the written language and the

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM
404 Joshua A. Fishman
draft scheme for a Chinese phonetic alphabet", in Reform of the
Chinese Written Language (Peking: Foreign Language Press).
Wurm, S. A.,
1966 "Language and literacy", in E. K. Fisk (ed.), New Guinea on the
Threshold (Canberra: Australian National University), 135-148.
Wurm, S. A., and D. C. Laycock,
1961/62 "The question of language and dialect in New Guinea", Oceania
32, 128-143.
Young, R. R.,
1944 "An adult literacy campaign in Sierra Leone", Oversea Education 15,
No. 3, 97-100.

Brought to you by | The University of Texas at Austin

Authenticated
Download Date | 2/16/20 10:45 AM