In Whither Music?, his brilliant
survey of the history and theory of music, Leonard Bernstein develops an
analogy between parsing sentences and parsing music. For both sentences
and for music, Bernstein finds that parsing is pretty much useless, indeed
"unbearable" and largely "meaningless" -- unless we have a theoretical
context that gives the parsing meaning. Surprisingly, Bernstein builds
his theoretical context for a "musical linguistics" by using basic concepts
of Chomsly's transformational-generative grammar, primarily deletion and
embedding. Using these concepts, Bernstein returns to parsing and finds
it essential to his explanation for laymen of musical history and
theory. I have emphasized "for laymen" because the grammar we teach in
our schools should be a grammar for laymen, not for specialists in linguistics.
The two transformational linguistic
concepts that Bernstein adopts are fundamental to KISS grammar, and they
can easily be explained in terms of traditional terminology. Deletion
is, in part, comparable to the traditional concept of "ellipsis," the omission
of words that are assumed to be understood. But it is much more powerful
than is traditional ellipsis. According to transformational theory, all
language is composed from very simple, deep-structure sentences, such as
The jewel was in the drawer and The jewel was red. These
two sentences can be combined into The red jewel was in the drawer
by deleting The jewel was from the second sentence and embedding
the adjective red into the first. Another way of looking at this
is to say that the second sentence was reduced (to the adjective
red) and then embedded into the first. As I will try to show in
several different contexts later, deletion (reduction) and embedding enable
us not only to simplify traditional grammar, but also to explain many aspects
of natural syntactic development. Although they are crucial to putting
grammar "in motion," these transformational concepts in themselves do not
provide a good pedagogical context. A major boost toward developing such
a context came in 1956, with George Miller's publication of "The Magic
Number Seven, Plus or Minus Two."
Miller's Theory of Short-Term Memory
Any pregnant new idea is
subsequently developed, extended, and altered such that the meanings of
terms become confused. "Short-term memory," for example, is now used by
some psychologists and educators to refer to what we can remember next
week from instruction that we receive today. This was not what Miller had
in mind. Miller was concerned with what the mind can handle from one second
to the next. (Because of the confusion in terms, Miller's original
concept is now often referred to as "working memory.")
In his primary experiments,
Miller put an assistant and a human subject in a quiet room. The assistant
had a list of sequences of random numbers:
The list, of course, was longer than my example. The assistant would read a sequence, and the human subject was asked to repeat it. The assistant then recorded the response as either correct or incorrect. The experiment, naturally, was repeated over and over again, with different assistants and different subjects. The results produced a very nice bell curve. Most people were able to remember all the sequences that were seven digits in length. Some people had trouble with six-digit sequences; a few had problems with five digits. At the other end, some people managed the eight digit sequences very well, and a few managed those that were nine digits long. This gave Miller his "magic number seven, plus or minus two." Miller concluded that humans basically have a seven-slot working memory.1, 6, 9, 8, 3, 6, 8
6, 8, 4, 3, 2
7, 5, 4, 3, 8, 6, 5, 3, 1, 7, 8, 9, 0, 7
The Model as Validation of the Rules We Teach
Given the model, students can understand
why some rules are more important than others. Using the model, of course,
students see one word chunking to another word, or to a group of other
words. This "chunking" is the dynamic equivalent of the traditional concept
"modification." Note, for example, that the model explains why dangling
or misplaced modifiers are an "error." They are not an error because they
will violate somebody's sense of etiquette (which is what I used to think);
they are an error because the reader will have problems determining what
they chunk to.
Because the model is dynamic,
it also explains why comma-splices, run-ons, and fragments are usually
considered as errors. One of the things psycholinguists have done (and
students can reproduce the experiment on their own) is to give "subjects"
one sentence followed by a second sentence. Asked to repeat the first
sentence, most people will reproduce the basic ideas, but they will use
different syntax and vocabulary. This has led linguists to conclude that
a main clause is generally processed in STM and then dumped to "long-term
memory." The periods (or other ending punctuation marks) at the end of
main clauses thus signal to readers that this sentence should be dumped
to "long-term memory," thereby clearing STM for the next main clause. Comma-splices
and run-ons can be errors because they do not signal the "dump." Readers,
therefore, try to chunk words from the next main clause into the preceding
one. That doesn't work. Processing crashes. Fragments cause crashes for
a different reason. Readers become accustomed to the basic main-clause
pattern -- every word in every sentence (with the exception of interjections)
chunks to another word or construction until everything chunks back to
the subject and finite verb pattern of a main clause. Fragments can confuse
readers because a "dump" is signaled before a complete pattern has been
established.
In addition to punctuation errors,
the model suggests the importance of some spelling errors. Learning to
read is not just a matter of learning vocabulary words. It is a matter
of learning how the syntax of English works, and then, using what one has
learned to make educated guesses. As noted above, after "want," a reader
expects an answer to "want what?" Similarly, "to" always evokes the question
"to what?"; "too" never does. Using the wrong spelling leads readers either
to expect a "what" when they shouldn't (because "too" was meant), or to
get a "what" when they didn't expect it. ("She wanted to go too the store.")
Similar problems occur with the misspelling of "its," "their," etc.
Although an occasional error (of
any of the types discussed above) may not cause a reader a problem, numerous
errors cause very serious problems. To illustrate this to my students,
I note that we can run on loose, rocky ground, trip, and usually keep running.
But if we trip a few times in quick succession, sooner or later, we will
end up on our behinds. And there are other sentences that are more comparable
to running on ice -- one good slip and down we go. CRASH!
What happens in a reader's mind
during a serious crash? I processed our model sentence using only four
of the seven slots of STM. I am not an expert in psycholinguistics, and
I am sure that the situation is more complex than I present it, but I suggest
to students that we use those other "free" slots to keep track of other,
more global aspects of what we are reading -- the topic sentence of the
paragraph, the thesis of the paper, etc. But when a crash occurs, the reader
has to focus more slots on the crash site. Something has to get thrown
out -- it may be the topic sentence of the paragraph, it may be the writer's
thesis. The more errors there are, the more opportunities there are for
the reader to crash.
Although a lot more could be said
about the psycholinguistic model and the question of errors, perhaps the
most important point is the radical shift it creates within the pedagogical
situation. The teacher is no longer imposing rules without reason. In fact,
the teacher doesn't even have to impose the rules at all -- the model does
it. The teacher, instead of being viewed as the policeman, becomes the
"coach."
The Psycholinguistic Model as an Introduction to Normative Stylistics
I don't remember reading much about
syntax and the "texture" of sentences, but it can be an interesting and
helpful concept to present to students. The question of "style" is itself
interesting because many people consider "style" to be something positive
-- "She did it with style." In a broader sense, however, "style" simply
means a repeated pattern. And, since sentences all have to be based on
an S/V/C pattern, every piece of writing has style. Part of this style
results from syntactic texture, and readers sense that texture unconsciously,
just as we sense any other texture. A KISS Approach, however, can make
many aspects of syntactic texture very easy to see and discuss.
One of the most basic aspects
of style involves main-clause length. The theory suggests that the brain
processes a main clause in STM and, when given an appropriate signal, dumps
it to long-term memory. This dumping creates a rhythm that is clearly
perceptible to all readers, and it is a rhythm which students can investigate
scientifically by counting the number of words in a passage, counting the
number of main clauses, and dividing the total number of words by the total
number of main clauses. What they will have calculated, of course,
is precisely Hunt's words per T-unit. If each member of a class analyzes
a sample of his or her own writing in this way, and the results are averaged,
the results will probably be close to the figures arrived at by Hunt, O'Donnell,
Loban, etc. When my college Freshmen do this, as a group they almost
always average within a half word of fifteen words per main clause.
What is important about
this is that it sets a basic stylistic norm. I introduce students to the
statistical results of the researchers, results which show the increasing
length of main clauses with age. I also point out that their readers do
not count words, but they definitely do sense the rhythm as a result of
their "dumping" the contents of STM to LTM and clearing STM for the next
sentence. I demonstrate this by simply using sequences of numbers::
clearly creates a different rhythm than does1, 2, 3, 4, 5, 6, 7, 8, dump
1, 2, 3, 4, 5, 6, 7, 8, 9, dump
1, 2, 3, 4, 5, 6, 7, dump
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, dumpIn analyzing their own writing, some of the students found themselves at the low end; some, at the high; and most, in the middle.
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, dump
1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, dump.
The Psycholinguistic Model and the KISS Theory of Natural Syntactic Development
As we look at the psycholinguistic
model and the KISS theory of natural syntactic development, we need to
keep in mind both O'Donnell's allusion to "formulas" and the question of
general cognitive development. The latter includes the level of abstraction
of the writing. The student who writes "Mother said we can go if it doesn't
rain" uses a conditional "if … then" construction, but the elements being
discussed are all simple and concrete. This "if … then" should not be considered
the same as the "if … then" in "If train A starts from Washington at 40
mph . . ." We can't simply count words or constructions. That having
been said, I would suggest that our psycholinguistic model helps explain
the research results of Hunt, O'Donnell, and Loban.
The KISS web site includes a rather
lengthy essay on "Language as a Stream of Meaning." It suggests that the
stream originates from the river of ideas that is in the writer's head.
Dostoevsky is reported to have said that only 10% of what was in his head
ever got down on paper, but the same is true even of the weakest of our
student writers. Lots of ideas flow into and out of our brains, and as
we try to write, we have no choice but to select some of those ideas and
attempt to embody them in words and sentences. The basic sentence pattern,
of course, is the main clause -- a subject, finite verb, and possible complement.
This is, of course, also the basic sentence pattern of fourth graders.
Let's assume, for a few moments,
that subordinate clauses, appositives, and gerundives are beyond the ability
of the fourth grader -- that any appearances of such constructions are
the result of formulas. In Vygotsky's terms, we are assuming that they
are beyond these students' "zone of proximal development." Vygotsky
visualized the zone as that area between two concentric circles. The inner
circle represents those ideas and concepts which the child has mastered.
The area between the two circles, the zone of proximal development, represents
those concepts which the student currently has the ability to learn. The
area outside the outer circle represents those ideas which are currently
beyond the range of the students' ability to comprehend. In math, for example,
the concept of multiplication is beyond the zone of any student who cannot
yet add and subtract. Once the student understands addition and subtraction,
the circles grow and multiplication comes into the zone of proximal development.
Even if subordinate clauses, appositives,
and gerundives are beyond the fourth graders zone, as Hunt suggested, there
is still a lot within that zone. In addition to compounding subjects, finite
verbs, and complements, Hunt suggested, for example, the embedding and
deletion that result in 1) a predicate adjective in one clause being embedded
in another -- "The red jewel was in the drawer." 2) a prepositional phrase
from one main clause being embedded in another -- "The jewel in the drawer
was red" and 3) a clause with "have" being reduced to a possessive -- "I
like to ride my new bicycle." And Hunt's list was meant to be suggestive,
not exhaustive. Since the research shows that subordinate clauses do not
blossom until seventh grade, we can probably conclude that between fourth
and the end of sixth, most students are mastering a variety of other constructions
such as those described by Hunt. The point here is that even though we
do not yet know enough about what students are learning, learning is certainly
going on. Forcing these students to "learn" what we want them to (clauses,
gerundives, appositives) probably interferes with their natural learning.
The development of the subordinate
clause appears to us to be a major expansion of the zone, or in Piaget's
terms, an ascent to a higher plateau, primarily because it is easier for
us to see it. When it does occur, of course, it is a major expansion because
it requires the student to process at least two S/V/C patterns, as S/V/C
patterns (and not as formulas), within the student's short-term memory.
If we compare the processing in STM with juggling, the number of things
juggled at one time at least doubles. It should be obvious that it will
take some time to master this new skill. The research suggests seventh
and eighth grades, but as I noted previously, a few of my college Freshmen
have still not mastered it. It is not, as Hunt suggested, "the language"
which "sets a practical limit on the number of subordinate clauses" (1965,
p. 69). For young students, it is their linguistic, and psycholinguistic
experience. (Note that, as suggested above, at the upper end of development,
the limit is imposed by the psycholinguistic processing abilities of intended
readers.)
Once subordinate clauses have
been mastered (in Piaget's sense of the term), natural development, as
suggested in the previous chapter, turns to the appositive and gerundive,
Hunt's "late" bloomers. Why are they late bloomers? Hunt perceived most
of the late blooming constructions as deriving from reductions of clauses,
and he was probably basically right.
They live in Endicott, which is a city in New York.can be reduced to the appositive
They live in Endicott, a city in New York.Likewise
When they were going to the park, they saw an accident.can be reduced to the gerundive
Going to the park, they saw an accident.If, forgetting formulas for a moment, we assume that all of these constructions derive from the reductions of clauses, then they obviously must develop after the student masters the subordinate clause.
1. As students grow older (and have more to say), there is a natural tendency to combine more ideas into one main clause.
2. Much of that combining involves embedding of ideas from one clause into another, and, at about seventh grade, the development extends to embedding one entire clause (as a subordinate clause) into another main clause.
3. Throughout the process, the combining and embedding involve reductions. Before the development of the subordinate clause, most of the embedding and reduction results in the addition of simple adjectives, adverbs, and prepositional phrases; after the development of the subordinate clause, this process extends to appositives and gerundives.
There is, of course, still a lot that we need to learn about natural syntactic development, particularly in the area of formulas. The process is complex, and we need to be careful in the way that we teach. The first rule of doctors, however, is to do no harm. We need to be sure that, in intervening in this process of natural development, we too do no harm.
As I suggested earlier in
this chapter, the model presents a dynamic view of grammar that we can
present to students. The model justifies the "rules," and it can be used
to give students some basic insights into style. The model also gives the
study of grammar a clearly defined, relatively easily attainable goal --
the students' objective is to understand and be able to discuss how any
word in any sentence is "processed" by readers. The grammatical concepts
that students need to learn in order to understand and use the model are
the subject of Part Four. Taken all at one time, they may appear to be
complicated, but if they are studied in a KISS approach, they are not as
difficult as they may appear.