1.
THE EDUCATIONAL ENTERPRISE
1.1
The complaints voiced by educators, students, and society at large suggest grave
disillusionment with the efficacy of public education in contemporary America:
what is being taught is not retained, not useful, and not relevant to the tasks
learners will face in later life. In this environment, learner motivation is
generally low. The ominous failure quota dooms a substantial portion of American
youth to an existence permanently devoid of professional and social
opportunities. The bizarre Darwinian notion of the “grade curve”—a
statistical construct that demands a balance of superior grades against failures
or near-failures, with a broad mass of learners in the middle—suggests that
failure is not viewed as a personal tragedy, but as a normal condition of
schooling. Educators who undertake to encourage superior performance for all
learners can become the target of committees against “grade inflation “—a
purported “overbalance” of favorable grading.
1.2
When public schooling became generally available in the eighteenth and
nineteenth centuries, the traditional curriculum was uphel—a curriculum which
by and large goes back to the schools of Ancient Greece. The question of whether
such a curriculum is relevant to the needs of modern children has seldom been
officially raised. Educators “assumed a cultural trust, a vast body of
unspoken but shared routines that freed them from the need to explain what they
were up to” (Shaughnessy 1976: 153). A methodology that is not openly
expounded and discussed can hardly be revised.
1.3
My investigations of American schooling have left me with the impression that
the main orientation is behavioristic. despite the dramatic downfall of
behaviorism as an account of human knowledge and abilities. Altogether too much
energy is expended on rote acquisition and recitation of isolated facts. Testing
and evaluation inhabit a disquieting world where everything is either
“correct” or “incorrect.” The consequences of forcing humanistic
domains into such a methodology are disastrous. Language skills must then be
measured by “grammar quizzes,” and writing exercises graded via mechanical
tabulation of surface errors or deviations from prestigious usage. Literature is
demeaned and impoverished into a contest of identifying quotations or authors’
names and biographies. In short, CREATIVITY is routinely discouraged because it
eludes the distinction of “correct and incorrect responses.” Getzels and
Jackson (1962) found that standard measurements of intelligence (“IQ tests”)
scarcely measure creativity at all. The result is doubtless a loss of motivation
among learners who justly feel that their individual talents and skills are
welcome and approved only for a restricted range of pre-decided tasks of
questionable value and relevance. Typically, the failure to learn in such an
environment is blamed on the individual child (cf. criticism in Dittmar 1976:
95).
1.4
In cognitive terms, education is on the wrong track as long as it stresses
EPISODIC KNOWLEDGE over CONCEPTUAL-RELATIONAL (cf. Kintsch 1977a: 284; Groeben
1978: 15). The learner’s mind is cluttered with an array of incidental facts
which elude integration into. a coherent functioning system of world-knowledge.
Each batch of poorly digested facts is promptly forgotten after testing, because
the systemic organization needed for application is not provided. I submit that
this dreary situation can be vastly improved if we shift the whole emphasis of
schooling away from the memorization of facts toward the development of a
powerful and flexible set of strategies for acquiring, organizing, and
applying knowledge irrespective of the specific content in a task or
textbook (cf. 1.5.6; IV.3.18). These strategies should become the openly
proclaimed theme of all curricular subjects from the first to the last
year of schooling. A fully co-ordinated corpus of learning materials reflecting
this priority should be provided for each and every classroom. The judgment of
performance should be done not for the sake of discrimination and retribution
against the individual child, but for the sake of diagnosis and arranging of
training priorities. The strategies I envision include the following (cf. the
list of processing strategies in IV.3.17ff.):
1.4.1
strategies for general problem-solving in the sense of 1.6.7ff.;
1.4.2
strategies for flexibility in task management;
1.4.3
strategies for decomposing large, difficult tasks into small, simple ones;
1.4.4
strategies for focusing and distributing attention wisely;
1.4.5
strategies for judging the efficiency, effectiveness, and appropriateness of
available options (cf. 1.4.14);
1.4.6
strategies for building, implementing, and revising goal-directed plans;
1.4.7
strategies for weighing competing goals and deciding among them;
1.4.8
strategies for analyzing and learning from failure;
1.4.9
strategies for reasoning by causality (cause, reason, enablement, purpose);
1.4.10
strategies for reasoning by generalization from examples;
1.4.11
strategies for hierarchical reasoning (e.g. superclass inclusion, cf. III.3.19);
1.4.12
strategies for analogical reasoning (e.g. metaclass inclusion, cf. III.3.20);
1.4.13
strategies for differential reasoning (e.g. class definition and determination);
1.4.14
strategies for reasoning from incomplete knowledge;
1.4.15
strategies for computing relative probabilities and success chances;
1.4.16
strategies for adapting to improbable or non-expected occurrences;
1.4.17
strategies for modifying available systems or resources in response to a
reasonable motivation;
1.4.18
strategies for organizing, integrating, and storing content;
1.4.19
strategies for judging and maintaining interestingness and informativity;
1.4.20
strategies for arguing in support of views and beliefs;
1.4.21
strategies for self-reliance in intellectual undertakings of all kinds.
1.5
Though this list of abilities is not necessarily novel (cf. Newell & Simon
1972; Collins 1977, 1978; Resnick 1977), its implications for the development of
the human intellect are far from having been fully appreciated. The raising of
human INTELLIGENCE can be achieved by training learners to decouple these
strategies from the performance of individual tasks in the classroom (cf.
1.5.6;
IV.3.18).
Only then can learners operate on the high plane of types of knowledge and
operations, rather than on the low plane of daily “facts” and
“assignments.” Rote memorization and mechanical activities would sink down
to reasonable proportions. The teacher would act as a diagnostician or
specialized consultant, and not as a relentless inspector of exact details. We
would no longer equate rapid, accurate storage and recall with intelligence at
large—a misconception which educators seem to share with many psychologists.
We should rather undertake to explore and manage the rate at which individual
children can discover and apply cognitive strategies to the broadest possible
range of tasks and knowledge. Every school activity must be supportive of this
general objective. Children must realize that the different subjects and topics
in the various classes are all intended to cultivate a common repertory of
intellectual abilities. Learning materials must not be viewed as
incontrovertible artefacts to be swallowed and disgorged under stressful
conditions, but only as sample tools for rehearsing the cognitive skills that
learners will need for successful thought and action in later life.
1.6
I view activities of utilizing texts as the center of such an educational
enterprise.
The production and reception of efficient, effective, and appropriate texts
demands all of the twenty-one cognitive abilities I have cited. Hence, the
language-based disciplines offer a unique pivotal domain from which cognitive
development across the whole curriculum can be coordinated. Although my plans
for the re-organization of schooling along these lines are still in the primary
stage of development, I shall suggest in this chapter at least some aspects
worthy of consideration.
1.7 The shift of emphasis and methodology will consume considerable time and
resources at first, but in the long run, will lead to an enormous increase in
the power and successfulness of schooling will be attainable. A major source of
support for the already overworked teacher will be a new approach to computer-assisted
instruction. The notion of GENERAL PROBLEMSOLVING (mentioned in 1.6.7)
evolved from the realization that general strategies of reasoning and
procedure can and should be recognized as independent of the demands of any
particular task domain (cf. Ernst & Newell 1969; Newell & Simon
1972:414).1 [1. See footnote 20 to Chapter I.] Powerful
computer programs are already being developed for training precisely such skills
as I have listed above over a wide range of curricular subjects; geography,
chemistry, mathematics, medicine, and machine assembly (cf. Carbonell Sr. 1970;
Papert 1973; Collins, Warnock, & Passafiume 1974; Collins & Grignetti
1975; Collins, Warnock, Aiello, & Miller 1975; Brown & Burton 1975,
1977; Davis, Buchanan, & Shortliffe 1977; Collins 1977, 1978; Brown,
Collins, & Harris 1978). These programs differ fundamentally from the
conventional computer-assisted instruction in which there is a rigid sequence of
specific, material-dependent question-answer routines. In these new programs,
knowledge is stored not as rote facts, but as NETWORKS that allow the automatic
tutor to use facts in many ways, to ask flexible, mutually relevant questions,
and to deal perspicaciously with a variety of student replies and inquiries.
Hence, the tutor does not simply drill facts; it encourages the student
to acquire and apply general reasoning strategies such as Socratic logic,
generalization from examples, analogical reasoning, and even reasoning in
absence of knowledge (cf. Collins 1977, 1978). Errors are taken as opportunities
to show how the correlation of strategies can be improved, and not as ones to
administer humiliating punishments.
1.8
There is a further interdependency between education and a science of texts. The
acquisition of both humanistic and scientific knowledge cannot occur without
well-organized discourse. Indeed, a good portion of the work in mastering a
discipline lies in mastering its established mode of discourse (cf. for
instance, Bross 1973). The science of texts must provide clear, workable
standards for the production of all texts employed in learning. Norbert Groeben
(1978: 83) asserts that the success of learning strategies depends heavily on
the READABILITY of instructional texts—a factor all too often passed over
hitherto.
1.9
No one would deny that the educational enterprise is a vast and complex field
whose reorganization will be an enormous undertaking. Yet I hold that complexity
will diminish dramatically via the emphasis upon a common base of cognitive
training throughout the curriculum. Just as “mainstream” linguistics has
been unduly concerned with isolated aspects of language, so that a clear picture
of communication overall never emerged; in the same way, preoccupation with
scattered bits and pieces of knowledge in education renders learning
unnecessarily laborious, and solutions to individual problems arbitrarily
intricate. A science of texts with open borders toward all language-related
disciplines (cf. 1.1.2) and with emphasis upon the deeper processes of human
cognition can become the paradigm of a new outlook in both science and the
humanities, so that educational concerns of whatever kind can be integrated
within a forceful continuity and relevance.
2.
TRADITIONAL GRAMMAR VERSUS APPLlED LlNGUISTICS
2.1
For thousands of years, the socialization and education of children has involved
some kind of language training. That training seldom became institutiona1ized at
the outset of language acquisition: actual schooling usually began at an age
where extensive prior ability and experience could be safely presupposed.
Schooling focused less on the language as a complete intersystem, than on
certain PROBLEMATIC aspects deemed to require MANAGEMENT. Unfortunately,
management is difficult to achieve without a workable description or explanation
of language (I.1.9).
2.2
Traditional grammar drew on a diversity of sources that may well be hard to
reconcile with each other. Logic, philosophy, rhetoric, literature, public
attitudes, personal views of grammarians, and even grammars of entirely
different languages (notably Latin) have served as inspiration. The outcome was
a wide diversity of principles that were applied sporadically or at
cross-purposes. The “rules” of English usage are a striking demonstration.
The “rule” forbidding double negation came from logic, where two negatives
yield an affirmative; in normal speaking, they often yield an emphatic negative.2
[2. See. for instance. (6) and (7) in the ‘Tom
Tit Tot’ story (VIII.2.20). Perhaps the extra negatives are used to direct
focus to precisely those elements in which expectations are being overturned (d.
IV.l.2S), i.e., where a non-negative state of affairs is desirable or
normal.] The “rules” that
proscribe split infinitives and sentence-final prepositions were derived from
the structure of Latin. And many “rules” were introduced to discredit the
speech habits of non-prestigious social groups. Not surprisingly, a system of
“rules” like these is not very helpful to the learner, because: (1) the
principles at work are vague and inconsistent; (2) the emphasis is mostly on
what not to say or write; and (3) there is a divergence between the
official version of the language and the usage people can find all around the
culture.
2.3
Weaknesses like these were seized upon by modern linguistics as grounds for
rejecting the whole undertaking of traditional grammar. Paul Roberts (1958)
announced “the grammarian’s funeral” and exto11ed the precision and
objectivity of linguistic methods. The linguist vowed to describe languages as
they are, not as they should be in the opinion of small,
prestigious groups. Each language was investigated in its own terms, not in
terms of Latin. Despite these important improvements, early optimism about the
application of linguistics to language training soon waned. Linguistics was
addressing essentially different issues from traditional grammar, and was not
suitable for language management in any direct way. The analysis of utterances
into structures of minimal units offers no obvious means for discovering how
language options should be selected and utilized in communication. Linguistics presupposes
communication just as traditional grammar often had done.
2.4
In retrospect, the limitations upon “applied linguistics” for language
training seem obvious. Educators failed to appreciate the high cost of the
division of language into “langue vs. parole,” or “competence vs.
performance.” If taken in an extremely narrow sense—as in the recent polemic
of Dresher and Hornstein quoted in I.1.17.3—the idealized version of language
is far removed from the practical concerns of the teacher. Similarly, the
mysterious “language acquisition device” postulated as an innate mechanism
for building transformational grammars in the human brain-the rules are too
numerous and intricate to be learned in any other way-has not materialized so
far.
2.5
I believe the notion of “applied linguistics” will make a valuable
contribution only for a linguistics of actualization.3
[3.
Scc for example the contributions in Kohonen & Enkvist (eds.), 1978.]
The study of virtual systems of phonemes, morphemes, lexical
items, and sentence patterns can offer only an incomplete insight into the
operations of language in use. While there are many options in such
virtual repertories, only some options are EFFICIENT to use, EFFECTIVE in
getting things achieved, and APPROPRIATE to the demands of textuality in a
particular setting (cf. I.4.14). By discarding altogether the evaluative component
of traditional grammar, linguistics cannot advance much beyond the task of description;
even explanation will eventually have to account for the value
judgments that underlie language users’ motives for selecting language
options. The errors of old-style evaluation evolved from the belief that one can
state once and for all what language options are “correct” or
“incorrect” under all circumstances.
2.6
The monumental Grammar of Contemporary English (Quirk, Greenbaum, Leech,
& Svartvik 1972) demonstrates that a traditional grammar can both reflect
the realities of current usage and provide helpful criteria for selecting
options in common situations.3a [3a. Today (2004) I would
recommend the Longman Grammar of English over all others.] The grammar is
based not on the opinions of a few grammarians, but on an exhaustive survey of
English usage directed by Randolph Quirk. Socially and regionally marked options
are described as such rather than proscribed as “incorrect.” Current methods
of linguistics are deployed without dogmatic allegiance to a single approach.
The scope of the survey approximates completeness as far as any educational
purpose is likely to demand. Thus, the work vindicates the undertaking of
traditional grammar and opens the way to a reasonable integration of grammar
into developmental education.
2.7
One can continue to debate the question of whether linguistics has a commitment
to language training or not. I hold the task of language management to be among
the most pressing issues in the whole educational enterprise; what right have we
to talk about theoretical “competence” if our language theories cannot be
used to develop practical competence? To insist any longer on carefully screened
abstractions as the only “scientific” object for the study of languages is
to incur the onus of evaded responsibility toward the intellectual development
and self-expression of our children.
3.
THE TEACHING OF READING
3.1 In the absence of a useable account of language actualization, traditional reading instruction has laid undue emphasis upon the recognition of single words and sentences. Occurrences in which the reader’s “response” did not fit the “stimulus” on the page were dismissed as “errors” to be eradicated at all costs. Ernst Z. Rothkopf (1976: 109) surveys reading education and concludes: “human theories during the last three decades treated learning as if it were the passive consequence of bombardment by environmental particles.” I have devoted considerable space in this volume to supporting the view that, on the contrary, knowledge acquisition requires circumspect prediction and matching operations that integrate new material into prior systems. Changes of the surface presentation are normal and natural for experienced readers. “Miscues” in reading aloud (cf. VII.3.14) happen routinely among all reader groups and provide important evidence of the cognitive strategies at work (cf. Goodman & Burke 1973). Indeed, I would view miscues as a signal of a fluent reader who is not chained slavishly to the printed page.
3.2
The study of READABILITY has been mainly understood as the proportionality
between processing effort and obtainable knowledge during the activity of
reading( cf. VII.2.24). Klare (1963) cites some thirty readability
“formulas,” many of them derivative to some extent from the work of Rudolf
Flesch (1949). In most cases, reliance is placed on the superficial standards of
length and complexity of words or sentences. Rothkopf (1976: 108) criticizes
such standards:
The
lexical characteristic chiefly tapped is familiarity. Vividness and concreteness
are neglected. Exposition and organization are disregarded completely. Content
factors are ignored.
3.3
As far as 1 can judge, the usual measurements of readability reveal some severe
misunderstandings of the nature of texts and textuality. For one thing, I cannot
see how the options of virtual systems of language can be assigned a
degree of difficulty for all possible contexts of occurrence. One can, for
example, conduct experiments to find out what sentence patterns are easy to read
(cf. Groeben 1978: 18ff.); but it is surely an error to demand that the easiest
patterns must always be used in order to make the text optimally readable. On
the contrary: the constant use of easy patterns was shown in VII.2.18ff. to
yield an egregiously unreadable story. I consider the principle of the
"least effort" whlly misconceived as a standard of human activities at
large and of the reading of texts in particular. Readers will gladly expend more
effort, provided that the text awakens interest and rewards the effort with
informative insights. The "least effort" principle has bequeathed
us an armory of inane, boring readers for children (once called
"Dick-and-Jane" books in America) because of the view that we should
pursue only processing ease and ignore processing depth (cf.
I.4.14; IV.1.6).
3.4
A sensible measurement of readability cannot be content with looking at the
surface text or at virtual systems alone. We need to consider all of the
operations that map the surface text onto an underlying representation and
utilize that representation as relevant to a task. Cunningham (1978) found
children who performed quite well at recognizing and defining words (virtual
system: lexicon), but quite poorly at understanding and remembering the content
of whole passages (actual system: text). He attained some improvement by having
the children make up titles or endings for stories—tasks in which large
blocks of knowledge must be organized and tested for relevance. The restoration
of scrambled stories (cf. Kintsch, MandeI, & Kozminsky 1977) encourages
children to use schemas actively.
3.5
It seems clear that we can hardly measure the effort demanded for reading if we
have no model of the processes being used by readers. My own model leads me to
suppose that readability depends on the kind of problems which a text
presents—e.g. discontinuities in the surface or in the text-world model-and
the goals to be achieved by solving those problems. Literary and poetic texts
are frequently problematic because of their reorganization of the world and of
discourse about the world (cf. VII.1.8.4f.); but their readability is such that
they often outlive all other texts of their time and continue to be read widely
even now. 1 suspect that the use of creative texts wll do much to engage
children in learning to read.
3.6
The role of a science of texts in this area is readily evident. We must set up
integrated models of the reading processes and test the variables that could
genuinely affect operations, e.g.: (1) extent of discontinuities within
texts or textual worlds; (2) extent of discrepancies between texts or
textual worlds and prior knowledge or expectations; (3) extent of redundancy among
textual levels (cf. VII.2.18); (4) extent of individual reading experience; (5)
the outlook of reader groups on orders of informativity; (6) distribution
of attention; (7) extent of recall. All of these variables must be
considered before we can decide how readable a given text should be for a given
audience: we will be able to measure readability as relative, not absolute
(cf. Hirsch 1977), and in terms of cognitive procedures rather than
of surface texts.
4.
THE TEACHING OF WRITING
4.1
In the wake of an alarming decline in writing skills, the demand for a
full-scale theory of the writing processes has become acute. Following the
priorities of conventional linguistics, research has dwelt chiefly on structural
analysis, as Richard L. Larson (1976: 71) remarks:
We have, in studies of form, largely a record of
search for formulas and patterns in discourse, and a record of advice on the
properties that well-ordered discourse ought, in the a priori judgement of
theorists, to exhibit. But the reasons for the effectiveness of different
patterns, the ways in which their parts interact, the most useful techniques of
deciding upon particular sequences of steps in composing [. . .] have been dealt
with slightly, hesitantly, or not at all.
4.2
Perhaps we should inquire not what applied linguistics can offer, but rather
what applied linguistics should become if it is to have something worth
offering.4 [4.
One direct offshoot of linguistics for composition is the method of
"sentence combining" introduced by Bateman and Zidonis (1964) and
popularized by Mellon (1969) and O'Hare (1971). I have argued, however, that
sentence combining is of little value until we have a workable account of the motives
for combining sentences in a given context (Beaugrande 1979h) (cf.
1X.4.Sf.). Here again, the distinction between virtual and actual is crucial.] The following contributions appear desirable (cf. Beaugrande 1978e):
4.2.1 a realistic account of the cognitive activities
in writing;
4.2.2 a statement and classification of the options
and categories of the WRITTEN mode as opposed to the SPOKEN;
4.2.3 an account of the relative efficiency,
effectiveness, and appropriateness of written options in recognizable contexts;
4.24 a statement of the procedures for applying the
options of writing that are relevant to a plan toward a goal;
4.2.5
a model of orderly decision and selection based on all of the above;
4.2.6
a corresponding methodology for presenting and training the skills involved.
4.3 I expect that such an approach to writing will make it possible to diagnose writing problems which hitherto have simply been viewed as a chaotic array of surface "errors." The strangely disordered texts many untrained writers produce arise, I think, from competition between two divergent systems: face-to-face communication vs. public written discourse (cf. VII.2.6).
4.4 Consider the well-known error of the "comma splice," in which a writer fuses two otherwise independent sentences together with only a comma. The traditional treatment in composition classes was to advocate simply replacing the comma with a period or a semicolon. However, we might inquire how these comma splices evolve. In my view, this usage is intended to signal a dose relatedness of content that makes the writer hesitate to use a period. It would thus be sensible to connect the sentences into a single sentence with a subordinating junctive (V. 7.6f). My pilot projects show that beginning writers can use this sort of approach to attain considerable improvement.
4.5
The focus of a theoretically well-founded methodology of writing should be upon
MOTIVATION and DECISION. Learners who acquire workable standards for evaluating
their own prose as a protocol of decision-making need not rely
constantly on the teacher's feedback. Instead, they can compare their text to
their current motivations and goals, and revise inadequate decisions
accordingly. In this fashion, untrained writers can distribute their attention
selectively during several phases rather than trying to manage all writing
operations successfully in the first run. Presumably, this latter proceeding
simply overloads processing resources. We must therefore break the writing task
down into sufficiently small subtasks which any learners can manage,
irrespective of their prior experience and social background. As training
progresses, the ability to co-ordinate more and more subtasks at once should
rise in the same fashion that active storage can hold more material when larger,
better integrated "chunks" are formed (cf. III.3.11.6).
4.6
A research group I directed at the University of Florida, including Richard
Hersh, Patsy Lynn, Genevieve Miller, Nathan Robinson, and Patty Street,
initiated pilot projects along the lines I have sketched so far. A brief
demonstration can be found in these four alternative text fragments:
(260.la) Many people are thrifty these days. (260.13)
My husband is thrifty. (260.3a) He saves used toothpicks for firewood.
(260.lb) Many people are thrifty these days. (260.2b)
My thrifty husband saves used toothpicks for firewood. (260.lc) Many people are
extravagant these days. (260.2c) My husband is thrifty. (260.3c) He saves used
toothpicks for firewood.
The
decision in focus here is whether to place the modifier ‘thrifty’ of the
follow-up sentences (260.2a-d) in the “attributive” position (before the
head ‘husband’) or the “predicate” position (after a linking verb). In a
textual world where thriftiness is usual and hence expected (samples a and b),
the attributive positioning (260.2b) is a better choice than the predicate
(260.2a). In a textual world where extravagance is the norm, the focus-creating
predicate positioning (260.2c) is superior to the attributive (260.2d). We can
notice these differences just by reading the samples aloud and listening for the
stressed items. Of course, the samples are artificial in several respects: (1)
the normal state of things is not usually expressed in an immediately preceding
sentence; (2) a contrajunctive expression like ‘however’ would probably be
used in cases such as (260.2c); and (3) the opposition of ‘thrifty’ vs.
‘extravagant’ is more pronounced than would often be found in spontaneous
texts. Nonetheless, I hold such samples to be useful precisely because they make
their point so obviously and insistently.
4.7
In systemic terms, the point of such an exercise is to temporarily (and
artificially) STABILIZE the greater part of acualization operations in order to
direct focus toward a single variable factor. This channeling of focus might be
compared to the intensified apperception of a moving “figure” against a
stationary “ground” (cf. IV.2.5). Further decisions we have rehearsed with
this tactic include; main clause vs. relative clause or verbal participle for
expressing events and actions; active vs. passive; and two sentences vs. one
subordinative sentence (cf. Beaugrande, in preparation, for more details). In
all cases, the crucial factor is the degree of expectedness vs. informativity
within the underlying content (cf. VII.2.20ff.).
4.8
Notice that such training is not at all centered upon ERRORS, i.e. upon negative
instructions about what not to write. All samples like (260a) through
(260d) are well-formed according to the grammar of English, however described.
What I am trying to instill is an awareness of differences in efficiency,
effectiveness, and appropriateness of options which are all in themselves
allowable. Only in this way do I see some hopes for a positive approach
that lends an understanding of what good writing is rather than only what
it is not . Learning to write well is learning to navigate between the
extreme poles of the known and the unknown, the expected and the unexpected,
between conventionality and uniqueness, between economy and expenditure, and
between processing ease and processing depth (cf. VII.2.26).
4.9 Future research on composition depends in
particular upon contributions such as these:
4.9.1
Perspicacious fact-finding techniques are needed to probe the prior
discourse systems of beginning writers. As I indicated in IX.4.3f., surface
errors are probably symptomatic of competition. The exclusive reliance on
face-to-face communication can leave a learner quite unfamiliar with the
different organization of the written mode (cf. IX.4.3; Rubin 1978a). In
general, the access to the text receiver is subject to greater mediation in
writing, so that a higher degree of structural designing is required to
compensate (cf. Iser 1976: 114).
4.9.2
Empirical studies can discover how representative learner groups react in a
writing situation with stabilized content, such as narrating the course of
events in a silent film. Research on READABILITY as outlined in IX.3.5 should
provide a firm basis for our claims that certain textual organizations deserve a
higher rating than others.
4.9.3
lntegrated theoretical models of the writing processes must be developed in
co-ordination with the considerations of fact-finding techniques and empirical
studies. Each model must have HUMAN ACTIVITIES as the goal of its cognitive
interests (cf. 1.1.5).
4.9.4
Operational training materials must be developed that mediate between the
insights of the theoretical models and the exigencies of language management.
This mediation is especially important when aspects of a theory are too complex
to be capable of direct implementation as exercises.
4.10
My inclination is to look toward PROBLEM-SOLVING as the most general and
flexible model for writing activities of all kinds. I have suggested throughout
this volume that problem-solving as search for connectivity of points in a space
can serve to model all major aspects of textuality. John Richard Hayes and Linda
Flower (1978) have observed intriguing analogies between writing activities and
the solving of arithmetic and algebra problems. I take this evidence as
supportive of my claim that the utilization of texts taps all central cognitive
operations that are performed in intellectual acts of every kind (cf.
IX.1.4ff.). If so, a theory of writing must be an integral part of a theory of
cognition at large (cf. Bruce, Collins, Rubin, & Gentner 1978; Collins &
Gentner 1978).
5.
THE TEACHING OF FOREIGN LANGUAGES
5.I
For many people in universities, the term “applied linguistics” is taken to
denote only linguistically oriented approaches to instruction of foreign
languages. In America, the rise of this application came during and after World
War II when there arose an acute need for extreme fluency in strategic languages
such as German, Japanese, and Russian. The aid of such linguists as Leonard
Bloomfield and Charles Carpenter Fries was enlisted to develop special intensive
programs on the basis of current linguistic methods. For example, speaking and
hearing were declared pre-eminent over writing and reading. Exact descriptions
of sound systems were prepared. Care was expended to describe languages as used
by the average native speaker.
5.2
These improvements have, in the long run, failed to advance the state of the
discipline materially. In no other educational domain, as far as I know, is the
disproportion between expenditure of time or effort and successful learning so
dismally manifest as in foreign language instruction. Behaviorism has become
firmly entrenched in a stimulus-response methodology of endless repetition and
imitation—the so-called “audio-lingual” method (ear and tongue with no
brain in between, perhaps). Any materials that appealed to the cognitive
abilities of the learner, such as the presentation of language systems and
regularities, were taboo. The use of the native language was banned from the
classroom under the illusion that it would equally disappear from the
learners’ awareness (no stimulus, hence no response). Tedious “pattern
drills” in which inane, prefabricated sentences were varied by word
replacements bore little resemblance to natural communication. An “acoustic
grammar” was even proposed in which learners were trained to respond to
foreign language stimuli without knowing at all what they were saying (R. Morton
1966).
5.3
Under conditions like these, the failure rate is hardly astonishing. The
audio-lingual method is degrading if not openly wasteful of human resources in
its denial of any more sophisticated learning strategies than a “classical
conditioning” in the tradition of Ivan Pavlov (1927), aided by desirable
“grades” as a kind of “reinforcement” (cf. Thorndike 1931). The method
is horribly inefficient, covering as it does such a tiny and unrealistic range
of situations, and leaving the learner helpless in actual communication. Only
the surface aspects of sounds, morphemes, and syntax are treated with any
thoroughness, and only for pointless sentences devoid of sensible context. As a
result, the strategies of DISCOURSE PLANNING are left untouched, along with most
factors of situationality.
5.4
To be sure, the complexities of teaching a foreign language in a classroom
setting are alarmingly vast. Perhaps a reorganization of our methods, however,
would bring substantial improvements. First, it is clear that the teaching
situation can under no conditions suffice to impart the entire language.
Hence, I would propose the design of an artificially restricted intersystem
functioning with maximally powerful rules and options, the POWER of rules
and options being their ability to handle a wide range of cases with simple
operations. In grammar, power would be a trait of rules which generate the
highest number of paradigm members or syntactic patterns from the least numerous
steps (cf. Beaugrande 1979c). In vocabulary, power would obtain for items used
to express, explain, or circumscribe the largest amount of concepts (cf. the
“basic English” of Ogden 1932). These criteria would discourage the
inclusion of rare vocabulary or roles that belong to an unrealistically elevated
style of speaking (e.g. many subjunctives and anterior tenses in French).
5.5
Another crucial consideration is that we have been teaching VIRTUAL systems
without an account of ACTUALIZATION strategies. If actualization must guide
virtual systems and can even override them, as I claim (cf. 1.3.4.3), then not
even basic language instruction can afford to dispense with communicative
CONTEXTS (cf. Wienold 1973). Casual conversations, for instance, often fail to
exhibit the formation rules upon which so much effort is expended in the
classroom. I propose that textbooks be replaced, at least in part, with
subtitled film media capable of showing integrated communicative situations
(including the vital gestures and facial expressions used in the foreign
culture). The “drilling” of empty forms should be removed as much as
possible from the class to the terminals for computer-assisted instruction.
5.6
An important obstacle to the acquisition of a foreign language can be called
INTERFERENCE: the influence of the native language on the use of the foreign
one. Interference studies have mostly been concerned with the divergencies in
organization of the grammar, (e.g. different tenses) or the lexicon (e.g.
different distribution of expressions for a conceptual domain). My own pilot
projects suggest that discourse planning and actualization strategies are
also major sources of interference. I accordingly developed adjunct strategies
which the learners can superpose upon their native language strategies to
combat interference. This tactic enabled several pilot classes I taught to
attain a marked improvement in a period of about five to seven weeks.
5.7
The question of strategies indicates that an operational account of native
language in use is antecedent to a really workable methodology for
acquiring a foreign language. The hopes that transformational grammar would fill
this requirement (e.g. in Achtenhagen 1969) naturally led to disillusionment. I
would not view the poor results of applied linguistics in the past, however, as
proof that linguistics is inherently inapplicable. Considerably more may become
possible if we can agree to establish a linguistics which assumes its proper
role in a comprehensive account of meaningful human activities.
6.
TRANSLATION STUDIES
6.1
Text linguistics can offer a substantial contribution to translation studies
(see especially Dressler 1970b, 1972b, 1974b). “Mainstream” linguistics of
virtual systems had comparatively little to offer, because translating is always
a matter of actualization (cf. Beaugrande 1978a: ch. I). The celebrated disaster
of machine translation shows that a processor with only a grammar and a
dictionary can be constantly misled or entangled in alternative readings. The
processor simply couldn’t perform the problem-solving which discovers
or imposes connectivities upon language occurrences (cf. 1.6.8).
Considerable knowledge about how events and situations in the world are
preferentially organized and combined is needed here (cf. Wilks 1972). The Yale
Artificial Intelligence project directed by Roger Schank applies programmed
knowledge of events and situations to the understanding of reports coming
directly from the wires of news services; it then uses language-independent
representations to paraphrase, summarize, or answer questions about the reports
in Chinese, Russian, Dutch, and Spanish (cf. Schank& Abelson 1977;
Cullingford 1978 has many Spanish examples).
6.2
A central domain of translation studies is CONTRASTIVE LINGUISTICS (cf. surveys
in Nickel [ed.] 1971, 1972). Early trends using the descriptive approach
contrasted only virtual systems: phonemes, morphemes, grammatical
paradigms, and syntax in particular (Catford 1964; Ellis 1966). This outlook
emphasizes formal differences of the surface so strongly that translating seems
incapable of having a sound theoretical basis. Georges Mounin (1963) devotes an
elaborate treatise to the refutation of this conclusion, noting along the way
that linguistic models of the time were too myopic. Later appeals to
transformational grammar (e.g. Nida 1964) brought no great headway, because here
also, we are dealing with a purely virtual system.
6.3
Translating is possible only because human beings share an experiential world
and perhaps also at least some universal processing strategies (cf. IV.3.17; IX.
1.4). Those factors stand in ASYMMETRICAL relationships to the surface options
of individual languages (cf. I.6.12). These relationships have such complexity
that they are unlikely to be inferred by working from the surface alone. At
most, we could search for surface tendencies that reflect processing universals,
e.g. the placement of known or expected material before unknown or unexpected
(cf. IV.3), or the structural differentiation of coreferring pro-forms (cf.
V.4). Yet similar principles of sequencing and differentiation are no guarantee
that the resulting forms will be similar in various languages.
6.4
For a linguistics of actualization processes, the EQUlVALENCE between a text and
its translation can be neither in form nor lexical meanings, but only in the
experience of text receivers (Beaugrande 1978a, 19791). Translating is then
an issue of INTERTEXTUALITY (1.4.11.6) in which mediation works across different
language intersystems. The danger is that the translator will interpose his or
her own receiver experience as the only possible one for the text. For example,
the translator might bridge or fill in all discontinuities and discrepancies
such that the goal language receiver finds the text wholly devoid of
informativity and interest. This tactic is applied with lamentable frequency in
translating LITERARY and POETIC texts, and the multiplicity of functions and
meanings required for aesthetic experience is often destroyed (cf. Beaugrande.
1978a). Instead of arguing over “free” versus “literal” translating, we
might find the true opposition in “receiver-based” and
“translator-based” translating: only the former can claim communicative
equivalence. The question of how and whether forms or meanings are preserved can
only be settled in such a framework.
7.LITERARY
STUDIES
7.1
For many years, literary studies was the main discipline for the investigation
of whole texts. In absence of the necessary groundwork, discussions had to be
carried on without consistent or explicit theoretical models of texts and text
processing. The nature of literary and poetic texts as vehicles for alternative
organizations of the world and of discourse about the world (VII.1.8.4f.)
renders the construction of such models particularly intricate. While it was
often recognized that these texts represent deviations from some norm (ef.
Riffaterre 1959, 1960; Mukařovský 1964; Thorne 1969; Enkvist 1973), no
integrated language theories were available to describe what norms should be
consulted to begin with.
7.2
These matters were demonstrated by the attempt to postulate a “generative
grammar of literary texts” (cf. Bierwisch 1965b; van Dijk 1972b; Ihwe 1972).
It was reasoned that a set of subsidiary transformational roles could be added
onto the standard grammar of the language, such that literary and poetic texts
could be generated. Two objections seem readily evident. First, a grammar
expanded in this fashion would gain such ominous power that every conceivable
structure would be producible, so that the grammar would in the end have
explained nothing at all. Second, the effectiveness of literary and poetic texts
arises from the MODIFICATIONS performed on language systems for that
particular occasion. If the modifications were rule-governed, they would
lose quite a bit of their informativity and interestingness.
7.3
Perhaps we could envision literariness and poeticalness as arising from
CREATIVITY, modelled as the motivated modification of systems (Beaugrande
1979e). The strategies at work might be formulated like this: (1) insert a
highly non-systemic pathway into some actualized text system (e.g.
bizarre combinations of concepts or expressions); (2) test the interestingness
of the resulting configuration; and (3) test the relevance of the
configuration for the task of mediating insights into alternative
organizations of the world and of discourse about the world. Whether a given
text is effective and acceptable or not depends on the proportionality between
the effort expended on systems modification and the enriching insights thus
mediated. Many highly deviant texts (e.g. Baroque, Gongorism) seem to us
disproportionate along this dimension.
7.4
Possibly, the notion of STYLE as a literary issue might also be approached along
the lines I propose. I suggested in 1.2.10 that style arises from the
characteristic mapping procedures among the various levels of participating
language systems. Some of these mappings affect the distribution of phonemes and
morphemes (cf. Jakobson & Jones 1970) and the formatting of sentences (cf.
Ohmann 1964). Yet to depict style fully, we must consult all the mapping from
the phases of PLANNING and IDEATION all the way to the LINEARIZATION of the
surface text (cf. VII.2.8ff.). We must note the particular modifications performed
in or between these phases as an additional non-obligatory contributor to style.
Finally, we must confront the observed processes with the EXPECTATIONS of real
or potential text receivers about the respective domains (cf. IV.1.23ff.).
7.5
It would also be worth investigating how literary and poetic modifications can
be integrated into existing systems and thus create an impetus for further
modifications (cf. Mukarovsky 1964). This evolution can even occur in miniature
within a single text where modifications in one part of a text can be overthrown
by others in a later part (cf. Riffaterre 1959, 1960). Of course, the epochs of
the past and their communicative systems are removed from direct empirical
investigation. However, we can ascertain many things via the self-contextualizing
nature of texts (I.4.7). If system users must orient themselves toward the
functions of the systems themselves (whether or not modifications are
performed), then we can reconstruct a substantial portion of the systems via the
procedural analysis of sample texts. Eventually, the key role of literary
discourse within the total universe of discourse in historically evolving
societies may become clear.
7.6
I have conducted pilot classes in which the teaching of literature was
carried on along the lines I advocate. After a short introductory demonstration,
I turn the class largely over to the students. They each present a poem or song
text which, for whatever reason, they find appealing. Then they ask and answer
these questions: (I) what is the overall topic; (2) what steps and goals does
the writer’s plan contain; (3) what elements seem unusual, out of place, or
surprising; and (4) what motivations can be found for the use of those elements.
I have observed a dramatic increase in the abilities of understanding and
appreciating poetry. Indeed, many insights which my (largely first-year) college
students articulated are worthy of dissemination in scholarly journals, and I
eventually hope to present the method and its results. [“Eventually” is
right: see Beaugrande 1989, 1991, 1992.]
8.
A FINAL WORD
8.1
Following this brief overview of interdisciplinary applications, I must rest my
case in favor of a science of texts-a fitting conclusion for a method whose very
substance is so highly interdisciplinary. I have undertaken to marshal a broad
spectrum of support for my thesis that linguistics can and should explore texts
and textuality from the standpoint of human activities in actual utilization.
Such an approach can be useful both for traditional linguistic issues such as
phrase structures or grammar and for new issues emerging from the investigation
of cognitive processes.
8.2
I am not at all convinced that a broader scope must lead to an alarming increase
in the complexity of our theories and models. The complexity problem in
linguistics is of course important, and was handled in the past via severe
reductions of the domain of study (Weinrich 1976: 74ff.). But I believe that the
development of INTERACTIVE theories and models in which operations are closely
interconnected from system to system will show that the separation of language
systems in fact increases rather than reduces complexity. A case in point is the
explosion of arbitrary rules for an autonomous syntax.
8.3
I would draw an analogy to the famous “Waltz effect”: Although there are
staggering numbers of abstractly possible ways to label vertices in
analyzing visual objects (e.g. cubes shown only as outlines), David Waltz (1975)
found that the interaction of physical constraints reduces the number of actually
occurring labelling down from billions to a few thousand. I foresee the same
trend in a “linguistic Waltz effect”: the infinity of sentences, the vast
multitudes of natural language expressions and their senses, the variety of
communicative situations, the huge expanse of human knowledge and
experience—all these do not simply complicate the picture of communication;
they also impose mutual control, determinacy, and relevance upon each other. The
principle of continuity as the regulative stability of all participating systems
drastically limits the decisions and selections that are probable among
the total range of the abstractly possible.
8.4
I am therefore confident that the kind of science of texts which I have essaysed
to present will advance much more rapidly than the perspectives of
“mainstream” linguistics could lead one to believe. The new perspectives
will encompass wider, more diverse, and more vital issues, and, at the same
time, move closer to an understanding of the powerful processes of human
cognition and expression.
The
following is a program written by Robert F. Simmons in UT LISP 1.5 for the
University of Texas computer. The first portion represents the ‘rocket’ text
in terms of Hom clauses in successor arithmetic (Simmons & Correira 1978).
The bracketing shows hierarchical depths, as usual in LISP implementation. The
symbol “<“ is for antecedent-consequent theorems, meaning: “is true,
if…”. The letters “TF” followed by a number designate
“transformations” for ordering the clauses. “R” of course stands for
‘rocket’.1 [1.
A detailed rationale for the treatment of this particular text in terms of
clausal logic is now available in Simmons and Chester (1979). There, an
extensive rule formalism is provided for automatic extraction of the network
from the text and generation from the network back to the text.]
(GENPRINT RULES)
((V2 ROCKET) < (SETTING
R) (EPI R) (TF (2 3) (A 1 WAS FLOWN)))
((SETTING R) < (TOPIC R)
(STATES R) (TF (2 3) (THERE WAS A R)))
((TOPIC
R) < (R ISA ROCKET) (STOOD IN A NEW MEXICO DESERT) (SPEC R) (TF (4 3) (A R
3)))
((SPEC R) < (A GREAT)
(BLACK AND YELLOW) (V2) (FORTY-SIX FEET LONG) (TF (2 3 4 R 5)))
((STATES R) < (EMPTY IT
WEIGHED 5 TONS) (STATE 1 R) (TF (2 3) (2))) ((STATE1 R) < (IT CARRIED 8 TONS
OF FUEL) (KIND R) (TF (2 3) (2)))
((KIND R) < (—ALCOHOL
AND LIQUID OXYGEN) (TF (2)))
((EPI
R) < (EVERYTHING WAS READY) (ACT R) (TF (2 3) (2 AND IT FLEW OFF)))
((ACT R) < (PREACT R) (FLlGHT
R) (TF (2 3) (AFTER TAKEOFF IT FLEW))) ((PREACT R) < (PREACT1 R) (PREACT2 R)
(TF (2 3) (SIGNALS WERE GIVEN)))
((PREACT1 R) <
(SCIENTISTS AND GENERALS WITHDREW) (PURPOSE1) (TF (2 3) (2)))
((PURPOSE1)
< (SOME DISTANCE TO CROUCH BEHIND EARTH MOUNDS) (TF (2)))
((PREACT2 R) < (TWO RED
FLARES ROSE) (PURPOSE2) (TF (2 3) (2))) ((PURPOSE2) < (AS A SIGNAL TO FIRE
THE ROCKET) (TF (2)))
((FLIGHT
R) < (ASCEND R) (CRUISE R) (DESCEND R) (TF (2 34) (THE R ASCENDED AND LATER
PLUNGED TO EARTH)))
((ASCEND R) < (THE GIANT
R ROSE) (HOWROSE R) (TF (23) (2))) ((HOWROSE R) < (WITH A GREAT ROAR AND A
BURST OF FLAME) (HOW1) (TF
((HOW1)
< (THE ROCKET ROSE SLOWLY THEN FASTER AND FASTER (DISPLAY1) (TF (2 3) (2)))
((DISPLAY1)
< (BEHIND THE ROCKET TRAILED 60 FEET OF YELLOW FLAME)
(SOON THE FLAME LOOKED LlKE A YELLOW
STAR (TF (2 3)))
((CRUISE R) < (IN A FEW
SECONDS THE R WAS TOO HIGH TO BE SEEN)
(CRUISE2) (TF (2 3) (2)))
((CRUISE2)
< (BUT RADAR TRACKED IT AS THE ROCKET SPED UP TO 3000 MPH) (TF (2)))
((DESCEND R) < (TIME1)
(IT PLUNGED INTO EARTH 40 MILES FROM THE
((TIME1) < (A FEW
MINUTES LATER) (OBSERVER1) (TF (23) (2») (OBSERVER1) < (THE PILOT OF A
WATCHING PLANE SAW THAT) (TF (2)))
((X ISA X))
Below
are some summaries generated by truncating the hierarchy at various depths of
nesting. Each “say proof” is more detailed than its predecessor.
*VALUE:
T*
(SAY PROOF 1)
*VALUE: (A (V2 ROCKET) WAS
FLOWN)
(SAY PROOF 2)
*VALUE: (THERE WAS A ROCKET
(EVERYTHING WAS READY (AFTER TAKEOFF IT FLEW)))
(SAY PROOF 3)
*VALUE: (THERE WAS A ROCKET
(EVERYTHING WAS READY (SIGNALS WERE GIVEN (THE ROCKET ASCENDED AND LATER PLUNGED
TO EARTH))))
(SAY PROOF 4)
*VALUE: (A ROCKET (STOOD IN
A NEW MEXICO DESERT) (EMPTY IT WEIGHED 5 TONS) (EVERYTHING WAS READY (SIGNALS
WERE GIVEN (THE ROCKET ASCENDED AND LATER PLUNGED TO EARTH))))
(SAY PROOF 5)
*VALUE: (A GREAT (BLACK AND
YELLOW) V2 ROCKET (FORTY-SIX FEET LONG) (STOOD IN A NEW MEXICO DESERT) (EMPTY IT
WEIGHED 5 TONS (IT CARRIED 8 TONS OF FUEL)) (EVERYTHING WAS READY (SCIENTISTS
AND GENERALS WITHDREW (TWO RED FLARES ROSE) (THE GIANT ROCKET ROSE (IN A FEW
SECONDS THE ROCKET WAS TOO HIGH TO BE SEEN) (IT PLUNGED INTO EARTH 40 MILES FROM
THE STARTING POINT)))))
(SAY PROOF 6)
*VALUE:
(A GREAT (BLACK AND YELLOW) V2 ROCKET (FORTY-SIX FEET LONG) (STOOD IN A NEW
MEXICO DESERT) (EMPTY IT WEIGHED 5 TONS (IT CARRIED 8 TONS OF FUEL (—ALCOHOL
AND LIQUID OXYGEN))) (EVERYTHING WAS READY (SCIENTISTS AND GENERALS WITHDREW
(SOME DISTANCE TO CROUCH BEHIND EARTH MOUNDS) (TWO RED FLARES ROSE (AS A SIGNAL
TO FIRE THE ROCKET)) (THE GIANT ROCKET ROSE (WITH A GREAT ROAR AND A BURST OF
FLAME) (IN A FEW SECONDS THE ROCKET WAS TOO HIGH TO BE SEEN (BUT
RADAR TRACKED IT AS THE
ROCKET SPED UP TO 3000 MPH)) (A FEW MINUTES LATER (IT PLUNGED INTO EARTH 40
MILES FROM THE STARTING POINT)))))))
(SAY PROOF 7)
*VALUE: (A GREAT (BLACK AND YELLOW) V2 ROCKET (FORTY-SIX FEET LONG) (STOOD IN A NEW MEXICO DESERT) (EMPTY IT WEIGHED 5 TONS (IT CARRIED 8 TONS OF FUEL (—ALCOHOL AND LIQUID OXYGEN))) (EVERYTHING WAS READY (SCIENTISTS AND GENERALS WITHDREW (SOME DISTANCE TO CROUCH BEHIND EARTH MOUNDS) (TWO RED FLARES ROSE (AS A SIGNAL TO FIRE THE ROCKET)) (THE GIANT ROCKET ROSE (WITH A GREAT ROAR AND A BURST OF FLAME (THE ROCKET ROSE SLOWLY THEN FASTER AND FASTER)) (IN A FEW SECONDS THE ROCKET WAS TOO HIGH TO BE SEEN (BUT RADAR TRACKED IT AS THE ROCKET SPED UP TO 3000 MPH)) (A FEW MINUTES LATER (THE PILOT OF A WATCHING PLANE SAW THAT) (IT PLUNGED INTO EARTH 40 MILES FROM THE STARTING POINT)))))))