The Organization of Knowledge
Networks and Schemas
by Gerald Grow, Ph.D.
Professor of Journalism
Florida A&M University
copyright © 1994, 1996
The Network Model
It is still a mystery how knowledge is stored, but cognitive researchers
offer two main models for knowledge--the network model and the schema model.
In the network model, knowledge is stored in "a network of interrelated
propositions." (See Figure 2.) Networks are simple "node-link"
structures which can be related in complex ways. You might think of a network
as a collection of contents ("propositions") which are interconnected
in very specific ways that reveal important relationships among. A similar
analysis of knowledge structure is widely used in research on artificial
intelligence. Note 19 In the
network model, connections, meaning, and learning are intertwined concepts:
"When no meaning (no connections) can be created, nothing is learned."
Note 20 We will return to the
importance of connections when considering the implications of this model
for writers.
Figure 2. Memory represented as a propositional network composed
of node-link structures. Not shown here are the many other links that radiate
from each element to other elements in the memory network. The connections
among nodes give information meaning and make it accessible. (Adapted from
E. Gagne, 1985, p. 79)
Schema Theory
Schema theory provides a different view of how knowledge is stored, though
network theory and schema theory are compatible and often used together.
Network theory tends to present a somewhat mechanistic view of mind, modeled
after the interconnections of computer memory. Schema theory presents a
more creative, goal-oriented view of mental activity.
A schema is a generalized mental model which is used to organize memory,
to focus attention, to interpret experience, and to codify actions. A schema
is similar to a prototype or template, except that schemas are active, self-activating,
self-revising processes. Note 21
As Anderson and Pearson explain it,
- Schemas are abstract (they contain summary information and prototypical
categories, rather than details about a specific case) and they are
- structured (they represent relationships among component parts).
- Schemas are dynamic--they change, develop, interact.
- They arise in response to situations (e.g., the schema for starting
your car comes to mind when you insert the key).
- They shape perceptions (in the dark it is easy to mistake a
bush for a bear).
- They provide context and vocabulary for interpreting
what we read. (A student moving from newswriting class to typography class
shifts the interpretation of the word "lead" from "opening
paragraph" to "spacing between lines of type.")
- They organize experience and modify themselves to accommodate
new experiences.
Schemas are thought to express relationships among parts called "slots"
(also known as nodes or variables). The schema for "going to a restaurant,"
for example, has a slot for "ordering an appetizer," and another
for "paying the bill." (See Figure 3 .) The schema for "face"
has as slot for "nose." The schema for "country" has
slots for "location on the globe, size, type of government, geographical
features," etc.
Figure 3. The Restaurant Schema. According to schema theory, we understand
the world in terms of prototypical patterns (scripts, schemas, narratives)
in which are embedded a vast array of relationships, concepts, and vocabulary
words.
When a schema is brought to mind and used to interpret some event, the schema's
slots are then "instantiated" with the particular details of the
moment. Into the schema of "face," you insert the particular details
of a new face. The schema relieves you of the burden of counting the number
of noses, eyes, ears, etc., on the face--you see the details of the new
face on the template for "face."
"Placing" something by recognizing its role in a familiar schema
("instantiating" it) is a good deal like placing an incident in
a familiar story. If someone read this passage aloud,
"Now these three children had a faithful
nurse called Nana; but Mr. Darling was angry with her and chained her up
in the yard; and so all the children flew away,"
you would immediately understand every detail of it--if you had recently
been reading about Peter Pan, who flew away with the Darling children while
Nana, the dog who was their nurse, strained against her chain. Note 23
If you imagine how confusing this passage would be to someone not familiar
with that story, you can understand how easily readers become confused when
they are not familiar with the schemas underlying what has been written.
Surely the practice of writing in a multicultural world requires us to gain
a deeper appreciation of the schemas that writers do and do not share with
various groups of readers.
Schemas play key roles in many cognitive processes.
They help us pay attention, comprehend, interpret, remember, make inferences,
set expectations, reason, solve problems, understand language structures,
read, write, explain what we know, and have a sense of humor. Note
24
Schemas (and the overall worldview of which they are a part) tend to be
self-maintaining. Indeed, some psychological theories treat schemas as the
active means by which we continuously affirm our sense of self as well as
our knowledge. No one is certain how schemas are acquired, but they do not
always require the slow, repetitive reinforcement envisioned in older theories.
Rather, people have the ability to pick up prototypical situations quickly,
even with one exposure. It is also unclear how people keep from cluttering
their minds with proliferating, conflicting, or superstitious schemas. (How
do we weed our mental gardens and forget what is wrong or useless?)
People seem to have a built-in impulse to organize experience in the form
of schemas. For the most part, schemas appear to be learned gradually; many
are learned indirectly. Recent educational theory, however, has begun to
consider the possibility of teaching schemas directly.
The Effects of New Knowledge
The structures of the mind constantly receive new information and respond
to it. New information can have several effects on a reader's existing knowledge
structures. Three effects identified by Rumelhart and Norman include: Note 25
- Accretion: The new information may fit into a slot in an existing
schema, and thus be quickly comprehended (e.g., the score for the second
game of the World Series).
- Restructuring: A reader may use new information to create a
new schema. (Eating certain things can help prevent cancer; oat bran is
one of those.)
- Tuning: A reader may use new information to "tune"
an existing schema so it is more accurate, complete, or useful. (Oat bran
does not help prevent cancer after all, but other foods seem to.)
These activities are essentially the same as Piaget's categories of assimilation,
accommodation, and equilibrium. New information can also have a few other
effects on current knowledge, including two that are important for this
paper:
- Dissonance: A reader may use new
information to create and maintain cognitive dissonance. When new information
does not fit with what is known, that information can be placed in long-term
suspension as an anomaly--something apparently real but presently incomprehensible.
The psycholinguistic explanation of reading (e.g., Frank Smith) implies
that people activate their schemas to define what is "known"
and then actively seek anomalies in order to test and update the known.
It is like walking into a familiar room and sensing that some one object
has been moved. Cognitive psychology's emphasis on the integrative power
of the mind may have led us to neglect an innate human appetite for the
spice of uncertainty. Within manageable proportions, everybody loves a
mystery.
- Confirmation: As Boulding points out, a reader may also use
new information to clarify or increase the degree of certainty of a schema
(what Boulding calls an "image."). Note
26
One of the most interesting questions in current research asks under
what conditions a schema is revised, supplemented, replaced, or abandoned.
Such changes are near the heart of learning. When important schemas closely
related to one's sense of self ("core schemas") are superceded
or shattered, extreme confusion, suffering, and meaninglessness can result.
Though sometimes as indurate as a deep-seated prejudice, core schemas may
also be shattered by something as insubstantial as Desdemona's handkerchief.
Knowledge, education, or even just learning to read, can be dangerous. Notes 27 - 28
Comprehension
Schema theory explains comprehension this way: "A reader comprehends
a message when he is able to bring to mind a schema that gives a good account
of the objects and events described in the message." Note
29
In some cases, this use of "schema" is what we commonly refer
to as "context." For example, the statement "The notes were
sour because the seam split" makes no sense until you are informed
that it is a statement about a bagpipe (Anderson's example).
Classic research by Bransford and Johnson showed that it is possible to
write paragraphs that are utterly incomprehensible until the reader receives
an explanatory schema (such as a drawing or title) by which to interpret
them--at which time they suddenly become quite clear. In other experiments,
a simple title could induce readers to interpret an ambiguous passage as
being about activities as different "playing cards" or "wrestling."
A poor reader, a newcomer to a culture, or a reader ignorant of geography,
government, etc., may lack the schemas that others fluently use to make
meaning out of what they are reading. Or such readers may apply an inappropriate
schema--as children often do--and misunderstand. Notes
30 - 31
To understand even simple words (charge, fast, strike, roll), you must first
know which schema to instantiate them into (i.e., what content to interpret
them in). For example, contrast the meaning of "check" in the
schemas for
- "eating out,"
- "paying bills,"
- "making a list," and
- "playing chess."
Only the most abstract and technical terms, which are locked into single
schemas (hilum, hispid, heterostylous), have single meanings. Words do not
"have" multiple definitions; rather, the same word ("check")
appears as the label for a different slot in several different schemas.
The meanings are in the schemas (the contexts), not in the word. Vocabulary,
then, comes from a knowledge of how the world works, not from what words
mean.
As Frank Smith and others have emphasized, the
overall meaning of a passage is not accreted by adding up the meanings of
individual words; the meanings of the words themselves are determined by
the larger context of meaning which the reader brings to the passage. Note 22
Indeed, due to the inherent ambiguity of individual words, constant interpretation
is essential; Note 33 many words
gain specific meaning only when placed in an appropriate schema. "Some
comprehension of the whole is required before one can say how individual
words should sound, or deduce their meaning in particular utterances, and
even assert their grammatical function," Frank Smith wrote. "There
is only one way in which print can be understood, and that is by having
meaning brought to it." Notes
34 - 35
Any moment of meaning depends on a finely-balanced dance between the familiar
patterns in the mind and the unique characteristics of the present moment.
Schema theory proposes that, by automatically processing the vast bulk of
familiar experiences, we can more fully attend to what is new and different.
By making the familiar into background, schemas make the strange come forward.
However, to someone without appropriate schemas, everything is strange.
All readers bring with them information that
is not given in the text, and it is to a large degree the activation of
this prior knowledge that makes reading possible. Note
32
Meaning is not assembled slowly from incoming chunks of verbal propositions
(as earlier models suggest). Rather, onto the linear, analytic sequence
of unfolding syntax, experienced readers quickly superimpose a holistic
scenario of meaning, and then read selectively to adjust the fit. They guess
the meaning of the passage, then correct their guesses as they proceed.
In this way, readers treat text not as information to be decoded but as
evidence to be interpreted in solving another mystery of meaning.
Forward
Contents
Gerald Grow's Home Page