-
Language & Communication, Vol. 9, No. 2/3, pp. 99-106, 1989.
0271-5309189 $3.00 + 0.00 Printed in Great Britain. Pergamon Press
plc
HOW DOES WRITING RESTRUCTURE THOUGHT?
ROY HARRIS
This paper challenges the romantic view of the evolution of
writing and the role of the Greeks in the development of alphabetic
literacy. It introduces the concept of autoglottic space and argues
that this affords a better explanation of how writing restructures
thought. At the same time it emphasizes that such questions cannot
be decontextualized from the social and political circumstances
attendant upon the introduction of writing into particular
cultures, nor from the diverse purposes which writing may
serve.
Writing is a concept which has undergone a remarkable
transformation in the Western intellectual world over the past 30
years. In the wake of influential work by Derrida (1967a, b), Goody
(1968), Havelock (1963, 1976, 1982), McLuhan (1962, 1964), Parry
(1971) and others, writing is no longer regarded as a mere
substitute for speech, or as a useful way of preserving and
transmitting knowledge, but as an active and powerful cultural
agency in its own right. What distinguishes the literate from the
preliterate society is no longer seen as being the possession of a
superior communications technology which has overcome the intrinsic
limitations of the spoken word and makes it possible to accumulate
records and accounts ad infinitum. These advantages, long
recognized within the traditional view of writing, now tend to be
regarded as merely incidental and external. According to the modern
view, the essential innovation which writing brings is not a new
mode of exchanging and storing information but a new mentality.
Propositions such as Writing restructures thought and Writing
restructures consciousness (Ong, 1982) have become almost cliches
in current discussions of literacy. So much so that it may now be
timely to pause and consider exactly what truth, if truth it is,
these cliches purport to capture. Whatever truth it may be, there
can be no doubt that it is often accompanied, and perhaps obscured,
by an accretion of misconceptions and fantasies which might be
called, to avoid any harsher term, romantic. These include a
romantic view of the origin of writing, a romantic view of the role
played by the alphabet in the history of writing, and a romantic
view of the contribution made by the Greeks to alphabetic writing.
To disentangle the thesis that writing restructures thought from
these and similar encumbrances is an essential preliminary
step.
It should be-but probably is not-superfluous to insist at the
outset that there are forms of writing which have nothing to do
with language at all, let alone with the respresentation of speech.
Those who claim that writing restructures thought, however, usually
have in mind verbal as opposed to non-verbal forms of writing, and
this restriction will be taken for granted in what follows.
The romantic view of the origin of writing is a very
ethnocentric view, and specifically a Eurocentric view. It is a
view summed up with admirable concision in the eighteenth century
by Charles Davy, who described as the communis opinio of the
day:
Writing, in the earliest ages of the world, was a delineation of
the outlines of those things men wanted to remember, rudely graven
either upon shells or stones, or marked upon the leaves or bark of
trees (Davy. 1772).
Correspondence relating to this paper should be addressed to
Professor R. Harris, Language & Communication, c/o Pergamon
Press, Headington Hill Hall, Oxford OX3 OBW, U.K.
99
-
100 ROY HARRIS
In short, writing originated as an evolutionary adaptation of
pictorial representation. Thus writing systems which retain
vestiges of pictorial figures are more primitive, and those which
have none more advanced. This thesis, in one version or another,
survived as received wisdom among twentieth-century authorities on
writing (Cohen, 1958; Diringer, 1962; Fevrier, 1948; Gelb, 1963)
and was taught as such to students of linguistics (Bloomfield,
1935), for it fitted in admirably with the orthodox linguistic
doctrine of the primacy of speech. The history of writing was seen
accordingly in terms of a universal progression from pictograms to
logograms to phonograms, culminating eventually in the alphabet.
Given this scenario, writing systems such as Chinese, which had
failed to progress to the alphabetic stage, were seen as inferior
or retarded, an opinion still widely held to this day. As Havelock
reminds us, when Boswell objected to describing the Chinese as
barbarians, Johnsons reply was Sir, they have not an alphabet. (One
might perhaps have expected him to add: And they have no grammar,
Sir, either.)
The romantic view of the origin of writing has not gone
unchallenged (Harris, 1986). But inasmuch as challenging it leads
automatically to calling in question the position of the alphabet
as the crowning achievement in human attempts to devise writing
systems, the challenge is likely to be met with astonished
incomprehension. For those who regard the alphabet as the
constitutional framework of our culture (Levin, 1987, p. 114)
anyone who appears to undervalue alphabetic writing is likely to be
seen as someone who neither appreciates nor even understands the
intellectual achievement that went into it (Levin, 1987, p.114).
According to Levin, an acquaintance with Latin, Greek or Hebrew
alphabetic writing opens up for the sophisticated reader endless
vistas of unique audio-visual aesthetic pleasure, of being able to
drink in and relish, hour after hour, page after page, how neatly,
how beautifully each of these did justice to the sounds . . .
(Levin, 1987, p. 116). Small wonder that those accustomed to
indulge in such delights have little patience with the unromantic
question of whether the letters of the alphabet represent sounds at
all.
The same scenario which bills Chinese writing as a system
devised by barbarians predictably reserves the noblest role for the
Greeks. The romantic view of the Greek contribution to writing may
be summarized as follows. The Greeks did not just add letters- for
vowels to a borrowed Phoenician inventory of letters for
consonants. They did something far more profound. By a typical
stroke of Greek analytic genius, they rethought the whole question
of how to represent spoken language by means of visible marks and
came up with what modern phonologists later rediscovered as the
phoneme principle. In other words, the Greeks converted the
alphabet into a phonemic notation.
Among champions of this alleged Greek achievement the most
eloquent is undoubtedly Havelock, for whom the Greeks were the
first people to devise a true alphabet.
Atomism and the alphabet alike were theoretical constructs,
manifestations of a capacity for abstract analysis, an ability to
translate objects of perception into mental entities, which seems
to have been one of the hallmarks of the way the Greek mind worked
(Havelock, 1982, p. 82).
Furthermore, the Greeks were not the sole beneficiaries of this
revolutionary insight, for the new system could identify the
phonemes of any language with accuracy. The cultural consequences
of this were far-reaching in Havelocks view.
Thus the possibility arose of placing two or several languages
within the same type of script and so greatly accelerating the
process of cross-translation between them. This is the
technological secret which made possible the construction of a
Roman literature upon Greek models-the first such enterprise in the
history of mankind (Havelock, 1982, p. 85)
The views briefly delineated above are open to ejection on
various counts. Not only do they romanticize the history of writing
in conformity with the cultural biases of Western
-
HOW DOES WRITING RESTRUCTURE THOUGHT? 101
education, and in so doing carry conjecture far beyond the
bounds of available evidence, but they make it all but impossible
to come to grips with the question of how writing restructures
thought. In order to correct this distortion of perspective, it is
necessary to bear three points in mind.
First, it no more follows from the fact that various writing
systems include what appear to be recognizably iconic characters
that writing is derived from pictorial representation than it
follows from the fact that various writing systems include
characters recognizable as simplified geometric shapes that writing
is derived from geometry. To conclude otherwise would be to confuse
form with function. The question of the origin of writing must not
be conflated with the question of the ancestry of particular
written characters (Harris, 1986).
Second, since we cannot know exactly how the inventors of the
alphabet (whoever they were) pronounced their native language,
there is no basis for discussing any original correspondence
between the characters and the sound system. Furthermore, any
argument which depends on taking primitive forms of the alphabet as
evidence for a putative sound system would be patently
circular.
Third, attributing the discovery of the phonemic principle to
the Greeks on the evidence afforded by Greek alphabetic writing is
like crediting the inventor of the kettle with discovering the
principle of the internal combustion engine. If the Greeks were
unaware of the phonemic principle they could hardly have made it
the basis of the Greek alphabet. Mysticism at this point comes to
the rescue of the romantic hypothesis, and the Greeks are said at
least to have achieved an unconscious phonemic analysis (Robins,
1979, p. 13). But what kind of cognitive feat is an unconscious
phonemic analysis? It hardly helps the romantic case to say that
the Greeks hit upon the phonemic principle without realizing it.
That will not do for at least two reasons. First, it contradicts
the romantic story that the alphabet is a work of analytic genius.
Second, the phonemic principle is not the kind of thing it is
possible to hit upon without realizing it, any more than it is
possible to grasp the Newtonian law of gravitation without knowing
you have done so. That is precisely why we attribute a grasp of the
law of gravitation to Newton, and not to his many predecessors in
human history who had been hit on the head by falling apples.
If we look soberly at the historical evidence we are forced to
conclude that if at any time in Graeco-Roman antiquity some unsung
genius did work out a version of the phonemic principle, then
either that insight died with its author, or else those to whom it
was imparted as a useful principle for devising writing systems
must have entirely failed to understand the message. The facts
which force us to this conclusion are the following.
First, phonetics was manifestly the weakest branch of Greek
linguistics (Robins, 1979, p. 24). People who were no better
informed about the mechanisms of articulation than the Greeks were
not likely to have understood enough about speech production to
arrive at a clear understanding of the phonemic principle.
Furthermore, if the Greeks had understood this principle, or even
groped towards it, it is quite mysterious why they never discuss
it. For example, it is certainly relevant to the topics debated by
Socrates in Platos Cratylus: but it is quite clear from that
discussion that in Platos day the sharpest minds in Greece, far
from grasping the phonemic principle, were still floundering with
the crudest notions of sound symbolism as an explanation for the
structure of human speech. This is a devastating piece of testimony
against the romantic hypothesis. It admits only two
interpretations. Either the institution of the Greek alphabet had
nothing to do with the phonemic principle; or else it had, but by
the time of Socrates the connexion had already been forgotten.
Either way, it is clear that from Plato onwards the phonemic
principle
-
102 ROY HARRIS
simply is not grasped in Greek discussions of language. This
must cast serious doubts on any thesis to the effect that the Greek
system of letters embodies the worlds first example of phonemic
analysis.
No less devastating is the internal evidence afforded by
comparison of the two major applications of alphabetic writing in
Graeco-Roman antiquity: namely, to record the classical Greek and
Latin texts that survived as the basis for Western education.
Inspection of the Greek case in isolation might possibly lead one
to view it as an example of phonemic notation, if one were prepared
to overlook certain inconsistencies, including the muddle it makes
of aspirate consonants. What gives the game away is the Latin case,
which fails to cope with the phonemic structure of the Latin vowel
system altogether. Latin spelling treats the long and short vowels
as allophonic variants, which they are not. Now if the Greeks had
understood the phonemic principle, it is inconceivable that the
Romans did not, especially those Romans most thoroughly versed in
Greek culture, and who wrote specifically on linguistic topics (for
example, Varro and Quintilian). Quintilian, who knew everything the
Greek grammarians had written, actually raises the question of the
adequacy of the alphabet; but it is quite clear that he had no
conception of interpreting the alphabet in terms of the phonemic
principle.
The only way to salvage the romantic hypothesis would be to
retroject the Greek discovery of the phonemic principle back into a
remote past, and postulate an ancestral dialect with exactly the
inventory of phonemes to match the postulated inventory of original
letters. But this postulation is entirely gratuitous, for a more
plausible account of events is readily available. This unromantic
account has been given many times; and no doubt it is rather
humdrum compared with the attractive story about Greek analytic
genius. This alternative is simply that the Greeks found that
unless they added some letters for vowels, the Phoenician alphabet
of consonants could not be used for Greek in any satisfactory way.
The reason for this does not take much genius to discover. It is a
reflection of the fact that Indo-European languages have a
typically different morphological structure from Semitic languages.
One must stress here the term morphological: the crucial factor has
nothing to do with phonology at all, much less with the alleged
Greek capacity for unconscious phonemic analysis.
Debunking romantic fantasies about writing would hardly be worth
anyones time did they not threaten to conceal from us a more
realistic picture of how writing restructures thought. The threat
arises in a number of ways, of which the following are two. First,
there is the inculcation of anachronistic views about the
relationship between language and art, and about the social
functions of language and art, which are comprehensible only after
the advent of writing. Second, there is the perpetuation, in an
extremely subtle form, of the cultural chauvinism (Ong, 1982, p.
18) of the nineteenth century, which assumes that the zenith of
human intelligence and cultural achievement is represented by and
reflected in the kind of education available to a member of the
upper classes of white, Western industrialized countries. In other
words, we are dealing with an attitude which assimilates the plight
of the poor illiterate Westerner to the plight of the poor
preliterate savage, even while romanticizing it as primitive,
innocent, ancestral, close to Nature, etc.
This comes out in the form of insisting that the mental
difference between literacy and non-literacy has to do with memory.
Allegedly, oral cultures are obsessed with the problem of handing
down in oral formulaic capsules pearls of communal wisdom which
would otherwise be forgotten. By contrast, literate cultures do not
have to impose this mnemonic burden on themselves, because they can
just write it all down. So they are free to use their
-
HOW DOES WRITING RESTRUCTURE THOUGHT? 103
collective minds in other ways. The restructuring of thought
which writing brings is based on this supposed liberation of
psychological space from the duties of storage.
The clearest statement of this theory comes from Havelock, who
applies it specifically to the literate revolution in Greece; but
his argument, if valid, would clearly apply far more generally. He
writes:
The alphabet, making available a visualized record which was
complete, in place of an acoustic one, abolished the need for
memorization and hence for rhythm. Rhythm had hitherto placed
severe limitations upon the verbal arrangement of what might be
said, or thought. More than that, the need to remember had used up
a degree of brain-power-of psychic energy-which now was no longer
needed . . . The mental energies thus released, by this economy of
memory, have probably been extensive, contributing to an immense
expansion of knowledge available to the human mind (Havelock, 1982,
p. 87).
This thesis about the limitations of memory has significant
affinities with common nineteenth-century views of linguistic
impoverishment: for instance, the view that primitive languages
have impoverished vocabularies, the view that the uneducated farm
labourer has an impoverished mastery of the vocabulary of English
by comparison with the squires son, and the view that dialects are
impoverished versions of a hypothesized standard language. The
common factor uniting all these views is that literacy broadens the
mind and widens the intellectual horizons; because the human memory
is finite, and those of us who cannot relieve the memory by
recourse to writing or by having access to written records are
condemned to trying to remember what would otherwise be forgotten.
Writing, in brief, was the technology by which the human mind was
at last freed from the cultural burden of constant oral
repetition.
Whatever plausibility this theory has begins to sag as soon as
the contention that literacy broadens the mind is called into
question. Like all new technologies, writing was a mixed blessing
in human history. In many respects, literacy can narrow the mind
just as easily as broaden it. Furthermore, as Socrates was aware,
it can be argued that the effect of reliance on writing is not to
liberate psychological space but merely to weaken the memory. In
addition, writing itself imposes storage requirements on the mind.
Learning ones letters is not a simple matter of familiarizing
oneself with a couple of dozen arbitrary shapes. (This is the
simplistic notion which seems to underlie the common contention
that an alphabet is superior to a syllabary because it reduces the
number of symbols. On the contrary, substitution of an alphabet for
a syllabary may well increase memory load. If reduction of the
basic number of symbols were a criterion of efficiency, then
clearly the binary system would be the optimum system of numerals.
Significantly, only computers find this to be the case. Human
beings dont.)
Nor does the theory that literacy makes possible a release of
psychic energy provide any specific answer to the question of
exactly how thought is restructured by writing. If it is true that
writing restructures thought, that is merely a particular case of
the more general truth that all new intellectual tools restructure
thought. The abacus restructured human thought much more profoundly
than the alphabet. The camera restructured human thought much more
profoundly than the calculus. The question with every new
intellectual tool is always: how does this innovation make possible
or foster forms of thought which were previously difficult or
impossible?
The key to understanding how writing restructures thought lies
in seeing how writing facilitates a variety of forms of autoglottic
inquiry. The actual development of these facilities then depends on
other factors in the cultural equation as it obtains in a
particular historical context. Nevertheless, this does not make it
impossible to formulate certain generalizations about the process
itself.
-
104 ROY HARRIS
Inevitably, the introduction of writing reshapes the whole
framework of communicational concepts available to a community. It
does so simply by destroying the equation of language with speech.
But to describe the revolution in these terms is already to
describe it retrospectively from the vantage point of literacy.
Less misleadingly, we might say that it was the invention of
writing that made speech speech and language language (Harris,
1983, p. 15), although any endorsement of that formula would be
vacuous unless it took account of the various ways in which the
distinction between language and speech may be drawn in literate
societies. [For a discussion of this distinction see Harris
(1983).]
The same point may be put in slightly different although no less
retrospective terms as follows. The restructuring of thought which
writing introduces depends on prising open a conceptual gap between
sentence and utterance. This is the locus for the creation of
autoglottic space; and it is into this autoglottic space that the
syllogism is inserted in the Western tradition. Without that space,
no formalization of the kind we now call logic would be possible.
(Exactly how this and cognate formalizations develop depends, as
noted above, on other factors in the cultural equation. One can see
this by comparing different cultures. In Europe the advent of
writing led to the formalizat.ion of logic before the formalization
of grammar. In India it was the other way around: first the
formalization of grammar and only then the formalization of logic.
This is an example of the kind of intercultural difference which
the romantic view of writing obscures.)
Writing is crucial here because autoglottic inquiry presupposes
the validity of unsponsored language. Utterances are automatically
sponsored by those who utter them, even if they merely repeat what
has been said before. Sentences, by contrast, have no sponsors:
they are autoglottic abstractions. The Artistotelian syllogism,
like the Buddhist Panchakdruni, presupposes writing.
It is sometimes argued (Ong, 1982, pp. 78ff.) that preliterate
cultures have available an alternative model of unsponsored
language in the form of proverbs, songs, etc. (where these are
traditional, i.e. essentially authorless). But this is to overlook
a number of important differences. For instance, proverbs acquire
vicarious sponsorship by being cited appropriately by parents,
elders and others noted for wise speech. Songs likewise are the
responsibility of those who sing them. In short, there is no
pragmatic divorce in these cases between a linguistic object and an
episode of discourse. More importantly, however, proverbs, songs,
etc. offer no general model of unsponsored language; precisely
because it is not the case than any utterance fulfils the
requirements of a proverb, song, etc. Whereas writing affords
exactly that dimension of generalization: in principle, any
utterance will have a written counterpart in a culture which has an
adequately developed script. The creation of autoglottic space
depends not merely on the fact that written words have a physical
existence which is independent of their authors existence, but
also, and more crucially, on the fact that writing offers a form of
unsponsored language which is not limited to particular categories
of speech act or verbal practice.
Thus in a literate culture it is relatively less difficult than
in a primary oral culture to distinguish consistently what is said
and what is meant from the person who said it and the occasion on
which it was said. In a primary oral culture there are no genuinely
autological forms of verbal knowledge because there is no
technology by means of which words and their relationships can be
decontextualized at will. Writing constitutes such a technology: it
thereby introduces a level of verbal conceptualization which
detaches words from their human sponsors. It is the availability of
this level of conceptualization which makes it possible for
Socrates to ask questions like What is justice?.
-
HOW DOES WRITING RESTRUCTURE THOUGHT? 105
If this is correct, then the role of literacy in the evolution
of Greek thought is quite different from the role assigned to it
within the romantic perspective. The pivotal movement in the
intellectual development of ancient Greece is not accounted for
simply by pointing to the introduction of writing, or to its
interiorization (whatever that may mean), and even less by
anachronistic construals of the Greek alphabet as a phonemic
notation. The contrast between Homeric patterns of thought and
Socratic strategies of inquiry is striking but inexplicable, until
and unless we introduce a different factor into the picture. This
factor is political. It has to do on the one hand with the advent
of democracy, and on the other hand with the emergence of the
Sophists. The pivotal balance is swayed by the assimilation of all
forms of intellectual inquiry to political debate. The paradox of
Classical Greek culture lies precisely in that assimilation. At the
very moment when, according to the romantic view, a new and
powerful restructuring of thought became available, Greece
renounced it and apparently reverted to an essentially oral
strategy for acquiring knowledge. Socrates, according to the
adulatory account which Plato has left us of his activities, is the
supreme apostle of what in the Indian tradition would be called the
Sabdapramana. Plato, although in many respects Socrates faithful
disciple, ultimately reached a different view.
As Julius Tomin demonstrates in a recent paper (Tomin, 1988), it
can be argued that the conflicting attitudes towards writing to be
found in Platos works are explicable chronologically by reference
to the failure of the philosophers political experiment in Sicily.
Having begun his career as a Socratic sceptic about writing (as
witnessed in the Phaedrus), Plato eventually ends up in the Laws as
a convert to the new technology. Laws must be set down in
writing:
the lawgiver must not only write down the laws, but in addition
to the laws, and combined with them, he must write down his
decisions as to what things are good and what bad; and the perfect
citizen must abide by these decisions no less than by the rules
enforced by legal penalties (Laws, 823A).
One might take the argument a stage further by pointing out that
although Plato treats laws as requiring sponsors in practice (i.e.
legislators, kings, etc.), the criteria by which laws are to be
justified in effect treat laws as examples of unsponsored language:
that is to say, laws are precepts having virtues or demerits which
are to be evaluated independently of their sponsors or their
sponsors intentions. Without the autoglottic space afforded by
writing in a society which can survey, as Plato says, not only the
writings and written speeches of many other people, but also the
writings and speeches of the lawgiver (Laws, SSSC), it would be
difficult indeed to evaluate what the law says as distinct from
what the law-giver says. That surveyability, which places tyrants
and poets on a par by substituting texts for both, is the
contribution which writing makes to the structuring of conceptual
space in a literate culture.
REFERENCES
BLOOMFIELD, L. 1935 Language, London.
COHEN, M. 1958 La grande invention de IPcriture et son
kvolution, Paris.
DAVY, C. 1772 Conjectural Observations on the Origin and
Progress of Alphabetic Writing, London
DERRIDA, J. 1967a LEcriture et la diffkrence, Paris.
DERRIDA, J. 1967b De la grammatologie, Paris.
DIRINGER, D. 1962 Writing, London.
-
106 ROY HARRIS
FEVRIER, .I. G. 1948 Histoire de Ikriture, Paris.
GELB, I. J. 1963 A Study of Writing, 2nd edn. Chicago.
GOODY, J. (Ed.) 1968 Literacy in Traditional Societies,
Cambridge.
HARRIS, R. 1983 Language and Speech. In Harris, R. (Ed.)
Approaches to Language, Oxford.
HARRIS, R. 1986 The Origin of Writing, London.
HAVELOCK, E. A. 1963 Preface to Plato, Cambridge, MA.
HAVELOCK, E. A. 1976 Origins of Western Literacy, Toronto.
HAVELOCK, E. A. 1982 The Literate Revolution in Greece and its
Cultural Consequences, Princeton
LEVIN, S. 1987 Review of Harris 1986, General Linguistics 27,
113-117.
McLUHAN, M. 1962 The Gutenburg Galaxy, Toronto.
McLUHAN, M. 1964 Understanding Media, New York.
ONG, W. J. 1982 Orality and Literacy, London/New York.
PARRY, M. 1971 The Making of Homeric Verse, Oxford.
ROBINS, R. H. 1979 A Short History of Linguistics, 2nd edn.
London.
TOMIN, J. 1988 With the Phaedrus on his mind. Unpublished
paper.