Social Science Information

http://ssi.sagepub.com/

Unpackaging literacy
Sylvia Scribner and Michael Cole
Social Science Information 1978 17: 19
DOI: 10.1177/053901847801700102

The online version of this article can be found at:

http://ssi.sagepub.com/content/17/1/19

Published by:

http://www.sagepublications.com

On behalf of:

Maison des Sciences de l'Homme

Additional services and information for Social Science Information can be found at:

Email Alerts: http://ssi.sagepub.com/cgi/alerts

Subscriptions: http://ssi.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://ssi.sagepub.com/content/17/1/19.refs.html

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
Sylvia Scribner
Michael Cole

Unpackaging literacy

One of the important services anthropology has traditionally

provided the other social sciences is to challenge generalizations
about human nature and the social order that are derived from
studies of a single society. The comparative perspective is especially
valuable when the topic of inquiry concerns psychological
&dquo;consequences&dquo; of particular social practices, such as for example,
different methods of child-rearing (permissive vs. restrictive) or
schooling (formal vs. nonformal) or mass communication (oral vs.
literate). It is surely a hazardous enterprise to attempt to establish
causal relationships among selected aspects of social and individual
functioning without taking into account the totality of social
practice of which they are a part. How are we to determine whether
effects on psychological functioning are attributable to the
particular practices selected for study, or to other practices with
which they co-vary, or to the unique patterning of practices in the
given society? When we study seemingly &dquo;same&dquo; practices in
different societal contexts, we can better tease apart the distinctive
impact of such practices from other features of social life.

This paper was prepared for a Conference on Writing, sponsored by the National
Institute of Education, June 1977. Its preparation and the research on which it was
based were supported by the Ford Foundation, Grant No. 740-0255.

19

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
20

impact of such practices from other features of social life.

Here we apply one such comparative approach to questions
about reading and writing practices and their intellectual impact.
Our approach combines anthropological field work with
experimental psychological methods in a study of &dquo;literacy without
schooling&dquo; in a West African traditional society. We hope our
findings will suggest a new perspective from which to examine
propositions about the intellectual and social significance of
literacy whose uncertain status contributes to our educational
dilemmas. ,

These dilemmas have been repeatedly stated. They revolve

around implications for educational and social policy of reports
that students’ writing skills are deficient, and that there is a
&dquo;writing crisis&dquo;. Is this the case if so, is it really a matter for
concern? Does it call for infusion of massive funds in new
research studies and methods of instruction? Or is it merely a signal
that we should adjust our educational goals to new &dquo;technologies
of communication&dquo; which reduce the need for high levels of
literacy skill? (See for example Macdonald, 1973)
These questions call for judgments on the social importance of
writing and thus raise an even more fundamental issue: on what
grounds are such judgments to be made? Some advocate that
pragmatic considerations should prevail and that instructional
programs should concentrate on teaching only those specific
writing skills that are required for the civic and occupational
activities student groups may be expected to pursue. Many
educators respond that such a position is too narrow and that it
overlooks the most important function of writing, the impetus that
writing gives to intellectual development. The argument for the
general intellectual importance of writing is sometimes expressed as
accepted wisdom and sometimes as knowledge revealed through
psychological research. At one end of the spectrum there is the
simple adage that &dquo;An individual who writes clearly thinks
clearly,&dquo; and at the other, conclusions purporting to rest on
scientific analysis, such as the recent statement that &dquo;the cognitive
restructuring caused by reading and writing develop the higher
reasoning processes involved in extended abstract thinking&dquo;
(Farrell, 1977, p. 451).
This is essentially a psychological proposition and one which is
increasingly moving to the forefront
of discussion on the &dquo;writing
problem.&dquo; Our research speaks to several serious limitations in
developing this proposition as a ground for educational and social

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
 21

policy decisions. One of these is the frailty of the evidence for

generalizations about the dependency of certain cognitive skills on
writing, and the other is the restricted model of the writing process
from which hypotheses about cognitive consequences tend to be
generated. Before presenting our findings on Vai literacy, we shall
briefly consider each of these in turn.

Speculations about cognitive consequences of literacy

What are the sources of support for statements about intellectual
consequences of literacy? In recent decades, scholars in such
disciplines as philology, comparative literature and anthropology
have advanced the thesis that over the course of history, literacy
has produced a &dquo;great divide&dquo; in human modes of thinking.
Havelock (1963) brilliantly speculated that the advent of alphabetic
writing systems and the spread of literacy in post-Homeric Greece
changed the basic forms of human memory. Goody and Watt
(1963) maintained that these same historic events laid the basis for
the development of new categories of understanding and new
logical operations, and in subsequent studies Goody (1977) has
concluded that potentialities for graphic representation promote
unique classificatory skills. Ong’s (1958) historical analyses of
prose literary genres in the fifteenth century led him to conclude
that the invention of the printing press gave rise to a new form of
intellectual inquiry uniquely related to the printed text.
Intriguing as these speculations are, their significance for a
theory of psychological consequences for individuals in our society
is problematic on two counts. These scholars derive evidence for
cognitive effects of literacy from historical studies of cultural and
social changes associated with the advent of widespread literacy.
Inferences about cognitive changes in individuals are shaky if they
rest only on the analysis of cultural phenomena. The incon-
clusiveness of the great debate between Levy-Bruhl and Franz Boas
(see Cole and Scribner, 1974) on the &dquo;logicality of primitive
thought&dquo; reminds us of the limitations of reliance on cultural data
as sole testimony to psychological processes. Secondly we need to

distinguish between historical and contemporaneous causation (see

Lewin, 1936). The development of writing systems and the
production of particular kinds of text may, indeed, have laid the
basis historically for the emergence of new modes of intellectual

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
22

operation, but these over time, may have lost their connection with
the written word. There is no necessary connection between the
modality in which new operations come into being and the
modality in which they are perpetuated and transmitted in later
historical epochs. Forms of discourse initially confined to written
text may subsequently come to be transmitted orally through
teacher-pupil dialogue, for example, or through particular kinds of
&dquo;talk&dquo; produced on television shows. One cannot leap to the
conclusion that what was necessary historically is necessary in
contemporaneous society. There is no basis for assuming, without
further evidence, that the individual child, born into a society in
which uses of literacy have been highly elaborated, must personally
engage in writing operations in order to develop &dquo;literate modes of
thought&dquo;. That may be the case, but it requires proof, not simply
extrapolation from cultural-historical studies.
While most psychologists have been interested in the psycho-
linguistic aspects of reading, some have concerned themselves with
these theoretical conjectures on the cognitive consequences of
writing. Vygotsky (1962) considered that writing involved a
different set of psychological functions from oral speech.
Greenfield (1968) has suggested that written language in the schools
is the basis for the development of &dquo;context-independent abstract
thought&dquo; - the distinguishing feature of school-related intellectual
skills. Scribner (1968) speculated that mastery of a written language
system might underlie formal &dquo;scientific&dquo; operations of the type
Piaget has investigated. Olson (1975) argues that experience with
written text may lead to a mode of thinking which derives
generalizations about reality from purely linguistic, as contrasted
to empirical, operations. In his view, schooling achieves
importance precisely because it is an &dquo;instrument of literacy&dquo;.
&dquo;There is a form of human competence,&dquo; he states, &dquo;uniquely
associated with development of a high degree of literacy that takes
years of schooling to develop&dquo; (p. 148).
These views, too, lack clear-cut empirical tests. Greenfield was
extrapolating effects of written language from comparisons of
schooled and unschooled child populations, but it is clear that such
populations vary in many other ways besides knowledge of a
written language system. Olson, to our knowledge, has developed
his case from a theoretical analysis of the kind of inferential
operations that the processing of written statements &dquo;necessarily&dquo;
entails. Scribner employed the same method of procedure.

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
 23

These are perfectly satisfactory starting points for a theory of the

intellectual consequences of reading and writing but they do not
warrant the status of conclusions. At a minimum, we would want
evidence that the consequences claimed for literacy can be found in
comparisons of literate and nonliterate adults living in the same
social milieu whose material and social conditions of life do not
differ in any systematic way.
We not only lack evidence for theoretical speculations about the
relationship between writing and thinking, but in our opinion, the
model of writing which underlies most psychological theorizing is
too restricted to serve as a guide for the necessary research.

Some dominant conceptions of writing

Although all disciplines that concern themselves with writing

acknowledge that it has different &dquo;functions&dquo;, these are often
conceived as external to the writing act itself that is the functions
-

being served by writing are not seen as intrinsic to an analysis of

component skills. In theory and in practice, writing is considered a
unitary (although admittedly complex) phenomenon representing
some given and fixed set of processes. These processes, it is

assumed, can be ferretted out and analysed by the psychologist,

linguist and educator without regard to their contexts of use.
Writing, together with reading, are described as &dquo;abilities&dquo; which it
is the task of education to enhance.
&dquo;The writing process&dquo; is typically identified with the production
of written discourse or text. Non-textual uses of writing, such as the
notational systems employed in mathematics and the sciences
which also require complex symbol manipulation, are excluded
from the domain of writing, along with other types of graphic
representation which use non-linguistic elements (diagrams, codes,
maps, for example).
In practice, a prototypical form of text underlies most anaiyses
of the writing process.’ This is the expository text or what Britton
and his colleagues (Britton et al., 1975) characterize as
transactional writing. Transactional writing is described as writing
in which it is taken for granted that the writer means what he says
and can be challenged for its truthfulness and its logicality: &dquo;... it
is the typical language of science and intellectual inquiry of ...

planning, reporting, instructing, informing, advising, persuading,

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
24

arguing and theorising&dquo; (Martin et al, 1976, pp. 24, 25).

Models of the cognitive skills involved in writing are intimately
tied up with this type of text. Thus, in making the claim that certain
analytic and inferential operations are only possible on the basis of
written text, Olson (1975) selects the analytic essay to represent the
&dquo;congealed mental labor&dquo; represented in writing. Nonliterate and
literate modes of thought are basically distinguished by their
differential experience with the production and consumption of
essayist text.
The development of writing skills is commonly pictured as a
course of progression toward the production of expository text.
Bereiter’s (1977) suggested model, for example, rests on the
assumption that there is a lawful sequence in the growth of writing
competence and that this sequence progresses toward the production
of a well-crafted story or a logically coherent discussion of a
proposition. At the apex of progressively more complex structures of
writing skills is epistemic writing writing that carries the function
-

of intellectual inquiry. (Similar views are expressed by Moffett,

1968.)
What is apparent from this somewhat caricature-like sketch is that
most of our notions of what writing is about, the skills it entails and
generates, are almost wholly tied up with school-based writing.
Centrality of the expository text and well-crafted story in models of
the writing process accurately reflects the emphasis in most school
curricula. A recently completed study of secondary schools in
England (Martin et al, 1976) found that writing classed as
transactional (see definition above) constituted the bulk of written
school work, increasing from 54 percent of childrens’ writing in the
first year to 84 percent in the last. Since such writing skills are both
the aim of pedagogy and the enabling tools which sustain many of
the educational tasks of the school, their preeminence in current
research does not seem inappropriate. But we believe that near-
exclusive preoccupation with school-based writing practices has
some unfortunate consequences. The assumption that logicality is in
the text and the text is in school can lead to a serious under-
estimation of the cognitive skills involved in non-school, non-essay
writing, and, reciprocally, to an overestimation of the intellectual skills
that the essayist test &dquo;necessarily&dquo; entails. This approach binds the
intellectual and social significance of writing too closely to the image
of the &dquo;academic&dquo; and the professional member of society, writ large.
It tends to promote the notion that writing outside of the school is of

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
 25

little importance and has no significant consequences for the

individual. The writing crisis presents itself as purely a pedagogical
problem - a problem located in the schools to be solved in the
schools through the application of research and instructional
techniques. What is missing in this picture is any detailed
knowledge of the role and functions of writing outside of school,
the aspirations and values which sustain it, and the intellectual
skills it demands and fosters. As our study of literacy among the
Vai indicates, these facts are central to an evaluation of the
intellectual and social significance of writing.

Three literacies among the Vai

The Vai, a Mande-speaking people of northwestern Liberia, like

their neighbours, practice slash-and-burn rice farming using simple
iron tools, but they have attained a special place in world history as
one of the few cultures to have independently invented a phonetic

writing system (Dalby, 1967; Gelb, 1952; Koelle, 1854). Remark-

ably, this script, a syllabary of 200 characters with a common core
of 20-40, has remained in active use for a century and a half within
the context of traditional rural life and in coexistence with two
universalistic and institutionally powerful scripts the Arabic and -

Roman alphabets. Widely available to all members of the society

(though in practice confined to men), Vai script is transmitted
outside of any institutional setting and without the formation of a
professional teacher group.
The fact that literacy is acquired in this society without formal
schooling and that literates and non-literates share common
material and social conditions allows for a more direct test of the
relationship between literacy and thinking than is possible in our
own society. Among the Vai we could make direct comparisons of
the performance on cognitive tasks of reasonably well-matched
groups of literate and non-literate adults. To do so, however,
required us from the outset to engage in an ethnographic enterprise
not often undertaken with respect to literacy the study of literacy
-

as acquired and practised in the society at large. Our effort to

specify exactly what it is about reading and writing that might have
intellectual consequences and to characterize these consequences in
observable and measurable ways forced us away from reliance on
vague generalizations. We found ourselves seeking more detailed

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
26

and more concrete answers to questions about how Vai people

acquire literacy skills, what these skills are, and what they do with
them. Increasingly we found ourselves turning to the information
we had obtained about actual literacy practices to generate

hypotheses about cognitive consequences.

From this work has emerged a complex picture of the wide range
of activities glossed by the term &dquo;writing&dquo;, the varieties of skills
these activities entail and the specificity of their cognitive
consequences.

What writing &dquo;is&dquo; among the Vai

Our information about Vai literacy practices comes from a number

of sources: interviews with some 700 adult men and women, in
which anyone literate in one of the scripts was questioned
extensively on how he had learned the script and what uses he made
of it; ethnographic studies of literacy in two rural towns;’
observations and records of Vai script teaching sessions and
Qur’anic schools; analyses of Vai script and Arabic documents as
they relate to Vai social institutions (see Goody, Cole and Scribner,
1977).
We estimate that 30 percent of the adult male population is literate
in one of the three scripts, the majority of these in the indigenous Vai

script, the next largest group in Arabic and the smallest in English.
There is a substantial number (20%) of literate men who read and
write both Vai and Arabic and a small number of triliterates. Since
each script involves a different orthography, completion of a
different course of instruction and, in the case of Arabic and
English, use of a foreign language, multiliteracy is a significant
accomplishment.33
As in other multiliterate societies, functions of literacy tend to be
distributed in regularly patterned ways across the scripts, bringing
more clearly into prominence their distinctive forms of social
organization, and transmission and function. In a gross way, we can
characterize the major divisions among the scripts in Vai life as
follows: English is the official script of political and economic
institutions operating on a national scale; Arabic is the script of
religious practice and learning; Vai script serves the bulk of personal
and public needs in the villages for information preservation and
communication between individuals living in different locales.

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
 27

In daily practice these distinctions are often blurred, raising a

host of interesting questions about the personal and situational
factors which may influence the allocation of literacy work to one
or another script.

English script has least visibility and least impact in the

countryside. It is learned exclusively in Western-type government
and mission schools, located for the most part outside of Vai
country. Students leave home to pursue their education and to win
their place in the &dquo;modern sector&dquo;. Little is seen of English texts in
the villages, but paramount chiefs and some clan chiefs retain
clerks to record court matters in English, and to maintain official
correspondence with administrative and political functionaries.
Arabic writing, on the other hand, is an organic part of village
life. Almost every town of any size has a Qur’anic school
conducteg by a learned Muslim (often the chief or other leading
citizen). These are usually &dquo;schools without walls&dquo; groups of -

boys ranging in age from approximately 4 years to 24, who meet

around the fire twice a day for several hours of recitation and
memofization of Qur’anic verses which are written on boards that
each child holds. (Qur’anic teaching in West Africa is described in
Wilks, 1968). In Islamic tradition, committing the Qur’an to
memory (internalizing it in literal form) is a holy act, and the
student’s progress through the text is marked at fixed intervals by
religious observances and feasting. Initially, learning can only
proceed by &dquo;rote memorization&dquo; since the students can neither
decode the written passages nor understand the sounds they produce.
But students who persevere learn to read (that is, sing out) the text
and to write out passages -
still with no understanding of the
language. Some few who complete the Qur’an go on to advanced
study under tutorship arrangements, learning Arabic as a language
and studying Islamic religious, legal and other texts. In Vai country,
there are a handful of outstanding scholars. with extensive Arabic
libraries who teach, study and’ engage in textual commentary,
exegesis and disputation. Thus Arabic literacy can relate individuals
to text on both the &dquo;lowest&dquo; (repetition without comprehension) and
&dquo;highest&dquo; (analysis of textual meaning) levels. Arabic script is used
in a variety of &dquo;magico-religious&dquo; practices, its secular uses include
correspondence, personal journal notes and occasionally trade
records. The overwhelming majority of individuals with Qur’anic
training, however, do not achieve understanding of the language and
their literacy activities are restricted to reading or writing

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
28

out known passages of the or frequently used prayers, a

Qur’an
service performed for others well as for oneself.
as

Approximately 90 percent of Vai are Muslim and, accordingly,

Qur’anic knowledge qualifies an individual for varied roles in the
community. Becoming literate in the Arabic language means
becoming integrated into a close-knit but territorially extended
social network, which fuses religious ideals, fraternal self-help,
trade and economic relationships with opportunities for continuing
education (see Wilks, 1968).
Knowledge of Vai script might be characterized as &dquo;literacy
without education&dquo;. It is typically learned within a two week to two
month period with the help of a friend, relative or citizen who
agrees to act as teacher. Learning consists of committing the
characters to memory and practice in reading, first lists of names,
later personal letters written in the Vai script. Demonstration of the
ability to write a letter without errors is a common terminating
point for instruction. With rare exceptions, there are no teaching
materials except such letters or other written material as the teacher
may have in his personal possession. &dquo;Completion of lessons&dquo; is
not the endpoint of learning: there are frequent consultations
between ex-student and teacher. For the practised scribe as well as
the novice, literacy activities often take a cooperative form (e.g. A
goes to B to ask about characters he cannot make out) and
sometimes a contentious one (e.g. A and B dispute whether a given
character is correct or in error).
Vai script uses are overwhelmingly secular. It serves the two
classical functions of writing: memory (preserving information
over time) and communication (transmitting it over space) in both

personal and public affairs, with a heavy emphasis on the

personal.44
From an analytic point of view, focusing on component skills, it
is useful to classify script functions according to whether or not
writing involves the production of text or non-text materials. Non
-textual uses range from very simple activities to complex record
-keeping. Among the simple activities are the uses of individual
written characters as labels or marking devices (e.g., marking
chairs lent for a public meeting with the names of owners,
identifying one’s house, clarifying information displayed in
technical plans and diagrams).’ Record-keeping, most typically a
list-making activity, fulfills both social cohesion and economic
functions. Lists of dowry items and death feast contributions,

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
 29

family albums of births, deaths, marriages all help to regulate

-

the kinship system of reciprocal rights and obligations. Lists

enlarge the scope and planful aspects of commercial transactions:
these include records of yield and income from cash-crop farming,
proceeds netted in marketing, artisan records of customer orders
and payments received.
A mere &dquo;listing of lists&dquo;, however, fails to convey the great
variation in levels of systematicity, organization and completeness
displayed in records. Some are barely decipherable series of names;
others orderly columns and rows of several classes of information.
Some genealogies consist of single-item entries scattered
throughout copy books, others of sequential statements which
shade off into narrative-like texts.
The more expert Vai literates keep public records from time to
time when asked to do so. These are less likely to be continuing
series than single list assignments: house tax payments for the
current year, work contributions to an ongoing public project such
as road- or bridge-building, a population headcount and the like.
Personal correspondence is the principal textual use of the script.
Letter-writing is a ubiquitous activity which has evolved certain
distinctive stylistic devices, such as conventional forms of
salutation and signature. It is not uncommon to see letters passed
from hand to hand in one small town, and many people who are
not personally literate participate in this form of exchange through
the services of scribes. Since Vai society like other traditional
cultures developed and still maintains an effective system of oral
contact and communication by message and &dquo;grapevine&dquo;, reasons
for the popularity of letter-writing are not self-evident, especially
since all letters must be personally sent and hand-delivered.
Protection of secrets and guarantee of delivery are among the
advantages most frequently advanced in favour of letters rather
than word-of-mouth communication.
For all its popularity, letter-writing is circumscribed in ways
which simplify its cognitive demands: a majority of Vai literates
correspond only with persons already known to them (78 percent of
literates interviewed in our sample study reported they had never
written to nor received a letter from a stranger). Many factors
undoubtedly contribute to this phenomenon, among which the
non-standardized and often idiosyncratic versions of script
characters must figure prominently, but it is significant for
hypotheses about intellectual skills that written communication

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
30

among the Vai draws heavily upon shared background information

against which the &dquo;figural&dquo; news is exchanged.
What about other texts? The first thing to note is that all textual
material is held in private; texts are rarely circulated to be read,
though on occasion and under special circumstances they might be
made available for copying. Thus the relationship of Vai script
literates to text is primarily as producer or writer, seldom as reader
on non auto-authored work. This social arrangement has several

important consequences. One is that reading is not an activity

involving assimilation of novel knowledge or material; another is
that existing texts reflect what people choose to write about,
depending on their own interests and concepts of what writing is
&dquo;for&dquo;. Many texts are of a cumulative nature that is, they are
-

not set pieces, but rather comprise &dquo;journals&dquo; or &dquo;notebooks&dquo;.

Each such &dquo;book&dquo; might contain a variety of entries, some auto-
biographic (personal events, dreams), other impersonal and factual
(facts of town history, for example). While not read as continuous
texts, such materials are often used as important source books or
data records and depending on their scope and age, may serve as
archives.6
Some texts fit recognizable, in terms of Western literacy, genres.
There are histories, for example, fables, books of maxims,
parables, and advice. In at least one instance, we have been able to
obtain a set of documents of a Muslim self-help organization which
included a Vai-script written constitution and bylaws (see Goody,
Cole and Scribner, 1977). As in the case of lists, the range of skills
reflected in texts is broad: &dquo;histories&dquo; may be a collection of what
were originally notes on scattered sheets of paper, assembled under
one cover with no apparent chronological or other ordering; at the
other extreme they might be well-organized and fluent narrations
of a clan history or ambitious accounts of the origin and migration
of the Vai people as a whole. While we do not know the relation-
ship between written and oral history and narrative, and thus
cannot determine whether written works are continuous or dis-
continuous with respect to the oral tradition, there clearly are
individual texts which bear the stamp of creative literary and
intellectual work. But it must be added that texts of this nature are
the exception; most histories are brief, often fragmentary and
written stories rare discoveries.
There are two types of text rarely found so far; Britton’s (1975)
two polar types the poetic, concerned with exploring personal
-

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
 31

experiences and feelings, and the transactional or expository, basically

concerned with examining ideas or presenting a persuasive argument.
Vai script literates are known in the community and admired for
their knowledge of books. Motivations sustaining script use are not
restricted to pragmatic ones; individuals will cite its utilitarian
value for correspondence, records and &dquo;secrets&dquo; but will as often
speak about the importance of the &dquo;book&dquo; for self-education and
knowledge and for preserving the history and reputation of the Vai
people. To be looked upon with respect and to be remembered in
history are important incentives to many Vaijournal-writers.
It is apparent from this quick review that Vai people have
developed highly diversified uses for writing and that personal
values, pride of culture, hopes of gain a host of pragmatic,
-

ideological and intellectual factors sustain popular literacy. The

-

level of literacy that obtains among the Vai must, however, on

balance be considered severely restricted. Except for the few Arabic
scholars or secondary school English students, literacy does not
lead to learning of new knowledge nor involve individuals in new
methods of inquiry. Basic processes of production, trade and
education are little affected by the written word.

Effects of literacy

Should conclude that these restrictions disqualify indigenous Vai

we

literacy &dquo;real literacy?&dquo; It clearly has social consequences for its

as

practitioners and (we hypothesized) might have identifiable cognitive

consequences as well. It seemed unlikely, however, that it would
have the very general intellectual consequences which are presumed
to be the result of high levels of school based literacy.
Nonetheless, this possibility was explored as part of our major sur-
vey of Vai adults at the outset of the project. In fact, we found no evi-
dence of marked differences in performance on logical and classifi-
catory tasks between non-schooled literates and non-literates. Conse-
quently, we adopted a strategy of making a functional analysis of
literacy. We examined activities engaged in by those knowing each of
the indigenous scripts to determine some of the component skills in-
volved. On the basis of these analyses, we designed tasks with different
content but with hypothetically similar skills to determine if prior
practice in learning and use of the script enhanced performance.

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
32

Communication skills

Since letter-writing is the most common use to which Vai script is

put, it is reasonable to look here for specific intellectual
consequences. In the psychological literature, written
communication is considered to impose cognitive demands not
encountered in face-to-face oral communication. In writing,
meaning is carried entirely by the text. An effective written
communication requires sensitivity to the informational needs of
the reader and skill in use of elaborative linguistic techniques. We
believed it reasonable to suppose that Vai literates’ experience in
writing and receiving letters should contribute to the development
of these communicational skills. To test this proposition, we
adapted a communication task used in developmental research
(Flavell, 1968). With little verbal explanation, subjects were taught
to play a simple board game and then were asked to explain the
game without the board present to someone unfamiliar with it.
We compared a full range of literate and non-literate groups,
including junior high and high school students, under several
conditions of play. Results were quite orderly. On several indices of
amount of information provided in an explanation, groups
consistently ranked as follows: high school students, Vai literates,
Arabic literates, and non-literates. Vai literates, more often than
other non-student groups, provided a general characterization of
the game before launching into a detailed account of rules of play.
If there is anything to the notion that what is acquired in a
particular literacy is closely related to practice of that literacy, the
differential between Vai and Arabic literates is exactly what we
would expect to find: on the average, Vai literates engage in letter-
writing more frequently than Arabic literates. It is interesting, too,
that both Vai and Arabic letter-writing groups were superior to all
non-literate groups.

Memory

We were also able to show specific consequences of Qur’anic

learning. Regardless of what level of literacy they attain, all Arabic

literates begin by learning to recite passages of the Qur’an by heart,
and some spend many years in the process. Learning by
memorization might promote efficient techniques for learning to

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
 33

memorize. To test this possibility, we employed a verbal learning

task (Mandler, 1969) involving processes that our observations
indicated matched those in Qur’anic memorization. In this task, a
single item is presented on the first trial and a new item is added on
each succeeding trial for a total of 16 trials and 16 items. The
subject is required to recall the words in the order presented. Our
comparison groups were the same as those used in the
communication experiment. English students again ranked first,
but in this task, Arabic literates were superior to Vai literates as
well as to non-literates in both amount recalled and in preservation
of serial order. If this superiority were simply the manifestation of
&dquo;better general memory abilities&dquo; on the part of Qur’anic scholars,
we would expect Arabic literates to do better in all memory tasks,
but this was not the case. When the requirement was to remember
and repeat a story, Qur’anic students did no better, and no worse,
than other groups. When the requirement was to remember a list of
words under free recall conditions, there were no significant
performance differentials. Superiority of Arabic literates was
specific to the memory paradigm which shadowed the learning
requirements of Qur’anic school.

Language analysis
In a third domain, we were again able to demonstrate the
superiority of Vai literates. Vai script is written without word
division, so that reading a text requires as a first step the analysis of
separate characters followed by their integration into meaningful
linguistic units. Our observations of Vai literates &dquo;decoding&dquo;
letters suggested that this process of constructing meaning was
carried out by a reiterative routine of sounding out characters until
they &dquo;clicked&dquo; into meaningful units. We supposed that this
experience would foster skills in auditory perception of
semantically meaningful but deformed (i.e. slowed down)
utterances. Materials consisted of tape recordings in which a native
speaker of Vai read meaningful Vai sentences syllable by syllable at
a two-second rate. The task was to listen and to repeat the sentence
as well as to answer a comprehension question about it. Vai
literates were better at comprehending and repeating the sentence
than Arabic literates and non-literates; and Vai literates with
advanced skills performed at higher levels than Vai literates with

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
34

elementary skills. Comparisons of performance on repetition of

sentences in which words, not syllables, were the units showed no
differences among literate groups but a sizeable one between all
literate and non-literate populations. The comparison of the two
tasks isolates skill in syllable integration as a specific Vai script
related skill.
Taken as a group, these three sets of studies provide the strongest
experimental evidence to date that activities involved in reading and
writing may in fact promote specific language-processing and
cognitive skills.

Implications

Our research among the Vai indicates that, even in a society whose
primary productive and cultural activities continue to be based on
oral communication, writing serves a wide variety of social
functions. Some of the pragmatic functions we have described are
by no means trivial, either in indigenous terms or in terms of the
concerns in economically developed countries for the promotion of
&dquo;functional literacy&dquo; skills. Vai literates routinely carry out a
variety of tasksusing their script which are carried out no better
(and perhaps vorse) by their English-educated peers who have
completed a costly twelve year course of school study. The record
keeping activities which we described briefly in earlier sections of
this paper provide the communities within which the liferates live
with an means of local administration. The fact that court
effective
cases were once recorded in the script and that religious texts are
often translated into Vai as a means of religious indoctrination
suggest that uses of writing for institutional purposes are fully
within the grasp of uneducated, but literate, Vai people.
While the bulk of activities with the Vai script may be
characterized in these pragmatic terms, evidence of scholarly and
literary uses, even rudimentary ones, suggest that non-schooled
literates are concerned with more than the &dquo;immediate personal
gain&dquo; aspects of literacy. We could not understand in such
narrowly pragmatic terms the effort of some Vai literates to write
clan histories and record famous tales nor the ideological
motivations and values sustaining long years of Qur’anic learning.
Of course we cannot extrapolate from Vai society to our own,
but it is reasonable to suppose that there is at least as wide a range

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
 35

of individual aspirations and social practices capable of sustaining

a variety of writing activities in our own society as among the Vai.
Since our social order is so organized that access to better-paying
jobs and leadership positions commonly requires writing skills,
there are even more powerful economic and political incentives at
work to encourage interest. It seems premature to conclude that
only schools and teachers are concerned with writing and that
writing would perish in this era of television if not artificially kept
alive in academic settings.
An alternative possibility is that institutionalized learning
programs have thus far failed to tap the wide range of
&dquo;indigenous&dquo; interests and practices which confer significance on
writing. Ethnographic studies of writing in different communities
and social contexts -
in religious, political and fraternal groups -

might help broaden existing perspectives.

Our research also highlights the fact that the kind of writing that
goes on in school has a very special status. It generates products
that meet teacher demands and academic requirements but may not
fulfill any other immediate instrumental ends. Is this an un-
avoidable feature of writing instruction?
When we look upon school-based writing within the context of
popular uses of writing found among the Vai, we are also
impressed by what appears to be the unique features of the
expository or essay type text. In what non-school settings are such
texts required and produced in our own society? Although
developmental models of writing place such texts at the &dquo;highest
stage&dquo; of writing ability, we find it difficult to order different types
of texts and writing functions to stages of development. Our
evidence indicates that social organization creates the conditions
for a variety of literacy activities, and that different types of text
reflect different social practices. With respect to adult literacy, a
functional approach appears more appropriate than a
developmental one. The loose generalization of developmental
models developed for work with children to instructional programs
with adolescents and adults is certainly questionable.
With respect to intellectual consequences, we have been able to
demonstrate that literacy-without-schooling is associated with
improved performance on certain cognitive tasks. This is certainly
important evidence that literacy does &dquo;count&dquo; in intellectual terms,
and it is especially important in suggesting how it counts. The con-
sequences of literacy that we identified are all highly specific and

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
36

closely tied to actual practices with particular scripts; learning the

Qur’an improved skills on a specific type of memory task, writing
Vai script letters improved skills in a particular communication
task. Vai literates and Arabic literates showed different patterns of
skills, and neither duplicated the performance of those who had
obtained literacy through attendance at Western-type English
schools.
The consequences we were able to identify are constrained by the
type of practices common in Vai society. We did not find, for
example, that performance on classification tasks and logic
problems was affected by non-school literacy. This outcome
suggests that speculations that such skills are the &dquo;inevitable
outcome&dquo; of learning to use alphabetic scripts or write any kind of
text are overstated. Our evidence leaves open the question of
whether conceptual or logical skills are promoted by experience
with expository text; in fact if our argument that specific uses
promote specific skills is valid, we should expect to find certain
skills related to practice in written exposition. The challenging
question is how to identify these without reintroducing the
confounding influence of schooling.
Perhaps the most challenging question of all, is how to balance
appreciation for the special skills involved in writing with an
appreciation of the fact that there is no evidence that writing
promotes &dquo;general mental abilities&dquo;. We did not find superior
&dquo;memory in general&dquo; among Qur’ anic students nor better language
integration skills &dquo;in general&dquo; among Vai literates. Moreover,
improvements in performance that appear to be associated with
literacy were thus far only observed in contrived experimental
settings. Their applicability to other domains is uncertain. We do
not know on the basis of any controlled observation whether more
effective handling of an experimental communication task, for
example, signifies greater communication skills in non-
experimental situations. Are Vai literates better at communicating
anything to anybody under any circumstances than Arabic literates
or non-literates? We doubt that to be the case, just as we doubt that
Qur’anic learning leads to superior memory of all kinds in all kinds
of situations. There is nothing in our findings that would lead us to
speak of cognitive consequences of literacy with the notion in mind
that such consequences affect intellectual performance in all tasks
to which the human mind is put. Nothing in our data would
support the statement quoted earlier that reading and writing entail

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
 37

fundamental &dquo;cognitive restructurings&dquo; that control intellectual

performance in all domains. Quite the contrary: the very specificity
of the effects suggests that they may be closely tied to performance
parameters of a limited set of tasks, although as of now we have no
theoretical scheme for specifying such parameters. This outcome
suggests that the metaphor of a &dquo;great divide&dquo; may not be
appropriate for specifying differences among literates and non-
literates under contemporary conditions.
The monolithic model of what writing is and what it leads to,
described at the beginning of this paper, appears in the light of
comparative data to fail to give full justice to the multiplicity of
values, uses and consequences which characterize writing as social
practice.

Sylvia Scribner is Senior Research Associate at the Rockefeller University. She is

engaged in studies of cognitive effects of schooling and literacy, of verbal
memory and reasoning among children, and is Co-Director. (with Michael Cole)
of &dquo;Literacy research among the Vai, West Africa&dquo;.
Michael Cole is Professor of Ethnographic Psychology and Experimental
Anthropology at the Rockefeller University and Director of the Laboratory of
Comparative Human Cognition.
They are co-authors of several articles and two books: Culture and thought: A
psychological introduction (1974) and Literacy: Social and psychological
implications of writing in a traditional society (forthcoming). Authors’ address:
The Rockefeller University: 1230 York Avenue, New York, NY 10021, USA.

Notes
1. The narrative text is also a common prototype, but we are leaving aside for
the time being approaches to creative writing which have largely been initiated and
developed outside the public school system.
2. These were carried out by Michael R. Smith, an anthropologist from
Cambridge University.
3. Because this phenomenon is rarely encountered in our own culture, we tend to
peg our "basic skills models" of writing very closely to the particular characteristics
and structure of a single orthographic system and assumptions of pre-writing fluency
in the language represented. As Fishman (1975) suggests this was the case with
bilingualism, studies of multiscript-using communities might well enlarge the
framework in which basic research on literacy is conducted. For accounts of other
non-industrial societies with a number of simultaneously active scripts, see Gough
(1968); Tambiah (1968); Wilder (1972). Schofield (1968) reminds us that between the
16th and 19th centuries in England, early instruction in reading and writing was
conducted with texts in English while higher education was conducted in classical
Latin.

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
38

4. Public functions of Vai script appear to be declining as English becomes

mandatory for administrative and judicial matters.
5. Gelb (1952) presents an interesting argument that social
origins of non-pictorial writing systems are to be found m the use of individualized
symbols as brands of ownership.
6. It is reported (Scribner, field notes) that an entire Vai community in
Monrovia was able to retain its right to disputed land because an elderly kinsman
had recorded in his book the names of the original deed-holders.

References
Bereiter, C.
1977 Integration of skill systems in the development of textual writing
competence. (Mimeo.)

Britton, J.; Burgess, T.; Martin, N.; McLeod, A.; Rosen, H.

1975 The development of writing abilities. London, Macmillan Edinburgh.

Cole, M.; Scribner, S.

1974 Culture and thought. New York, Wiley.

Dalby, D.
1967 "Further indigenous scripts of West Africa: Mandingo, Wolof and Fula
alphabets and Yoruba ’holy’ writing", African Language Studies 8: 1-51.

Farrell, T.J.
1977 "Literacy, the basics, and all that jazz", College English (January):
443-459.

Fishman, J.A.
1975 "The description of societal bilingualism", pp. 605-611 in: J.A. Fishman;
R.L. Cooper; R. Ma (eds.). Bilingualism in the Barrio. Bloomington,
Ind., Indiana University Publications.

Flavell, J.H.; Botkm, P.J.; Fry, C.L.; Wright, J.W.; Jarvis, P.E.
1968 The development of role-taking and communication skills in children.
New York, Wiley.

Gelb, I.J.
1952 A study of writing. Chicago, III., The University of Chicago Press.

Goody, J.
1977 "Literacy and classification:On turning the tables", in: R.K. Jam (ed.).
Text and context: The social anthropology of tradition. Philadelphia,
Penn., Institute for the Study of Human Issues.

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
 39

Goody, J.; Cole, M.; Scribner, S.

1977 "Writing and formal operations: A case study among the Vai", Africa 47
(3).

Goody, J,; Watt, I.

1963 "The consequences of literacy", Comparative Studies in Society and
History 5: 304-345.

Gough, K.
1968 "Implications of literacy in traditional China and India", pp. 69-84 in: J.
Goody (ed.). Literacy in traditional societies. Cambridge, Mass.,
Cambridge University Press.

Greenfield, P.
1968 Oral or written The consequences for cognitive development in
language:
Africa and the United States. Paper presented at the Symposium on
Cross-Cultural Cognitive Studies, Chicago, III.

Havelock, E.
1963 Preface to Plato. Cambridge, Mass., Harvard University Press.

Koelle, S.W.
1854 Outlines of a grammar of the Vai language. London, Church Missionary
House.

Lewin, K.
1936 A dynamic theory of personality. New York, McGraw-Hill.

Macdonald, J.B.
1973 "Reading in an electronic media age", pp. 23-29 in: J.B. Macdonald
(ed.). Social perspectives on reading. Delaware, Ohio, International
Reading Association.

Mandler, G.
1969 "Input variables and output strategies in free recall of categorized lists",
The American Journal of Psychology 82 (4): 531-539.

Martin, N.; D’Arcy, P.; Newton, B.; Parker, R.

1976 Writing and learning across the curriculum 11-16. London, Ward Lock
Educational.

Moffett, J.
1968 Teaching the universe of discourse. Boston, Mass., Houghton-Mifflin.

Olson, D.R.
1975 "Review of Toward a literate society", pp. 109-178 in: J.B. Carroll; J.
Chall (eds.). Proceedings of the National Academy of Education, vol. 2.
1978 "From utterance to text: The bias of language in speech and writing",
Harvard Educational Review.

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010
40

Ong, W.
1958 Ramus, method, and the decay of dialogue. Cambridge, Mass., Harvard
University Press. (Reprinted by Octagon Press, 1974).
1971 Rhetoric, romance and technology: Studies on the interaction of
expression and culture. Ithaca, NY, Cornell University Press.

Schofield, R.S.
1968 "The measurement of literacy in pre-industrial England", pp. 311-325 in:
J. Goody (ed.). Literacy in traditional societies. Cambridge, Mass.,
Cambridge University Press.
Scribner, S.
1968 Cognitive consequences of literacy. New York, Albert Einstein College of
Medicine. (Mimeo).

Tambiah, S.J.
1968 "Literacy in a Buddhist village in north-east Thailand", pp. 85-131 in: J.
Goody (ed.). Literacy in a traditional society. Cambridge, Mass.,
Cambridge University Press.

Vygotsky, L.S.
1962 Thought and language. Cambridge, Mass., MIT Press.

Wilder, B.
1972 An examination of the phenomenon of the literacy skills of unschooled
males in Laos. Michigan State University. (Ph.D. dissertation).

Wilks, I.
1968 "The transmission of Islamic learning in the Western Sudan", pp.
161-197 in: J. Goody (ed.). Literacy in traditional societies. Cambridge,
Mass., Cambridge University Press.

Downloaded from ssi.sagepub.com at UNIV CALIFORNIA SAN DIEGO on November 15, 2010