Level 6.5 Statistical Stylistics
and Advanced Analytical Questions
A green background in the right column indicates that the exercise is
in the printable version.
Although some readers may already be familiar
with it, it may be best to start this book with the essay on this topic
in the Background Essays:
An Essay on Statistical Exercises
and KISS Grammar
Victor Shklovsky, a famous
Russian literary critic, claimed that the purpose of literature is to make
life strange. We see the world around us so often that we fail to notice
what we see. Similarly, KISS statistical studies make sentence structure
strange. Statistical analysis slows us down -- it forces us to look at
various aspects of sentences in ways that we otherwise would not. How many
words are in the average written sentence? When is a sentence too long?
Too short? What makes adults' sentences longer and more sophisticated than
those of fourth graders? If students do several statistical analyses of
their own writing and compare their results with those of their classmates,
they can see for themselves that their sentences are short -- or long --
or, as Goldilocks preferred, "just right."
Can third and fourth graders, for example,
do a statistical analysis? This is an essential question. Within the KISS
framework, there are so many possibilities that the specific answer will
depend on the interests and skill of the teachers or parents. Suggestions
are offered below in "Some Suggestions about Graphing
and Statistics." The other primary practical question is, How
are statistical exercises used in KISS Grammar?
Note that the process of statistical
analysis may be more important than the numerical results. This is especially
true if students work in small groups to check their analyses of their
own writing. During such small group work I have occasionally had students
come up to me and say "I can't find -- my group can't find -- any subordinate
clauses in my writing." There were none. Most of the students in the class,
of course, did have lots of subordinate clauses in their writing. This
student, therefore, clearly discovered on his own that his writing was
somewhat lacking. He was much more motivated to learn how to analyze sentences
-- and to do some sentence-combining exercises on subordinate clauses.
At the other end, I have been asked, "Can
a subordinate clause be in a subordinate clause that is in a subordinate
clause that is in a subordinate clause?" The answer to that question is
"Yes," but professional writers rarely embed clauses that deeply because
such deep embedding makes sentences more difficult to understand.
In this case, the student became interested in the de-combining exercises.
The process may be more important than the
results, but my students are interested in the norms. To give students
a context for their results, you can use the data in KISS's "Cobweb
Corner." As some people are aware, however, statistical studies can
be very suspect. Sample size, sample selection, what is being counted,
how, and how it is being reported are all complex questions. In "Cobweb
Corner," you will find several studies done within the KISS framework,
but you will also find references to, and data from, several major studies
done in the 1960's and 70's on the natural syntactic development of school
children. (The dates suggest why "Cobweb Corner" is so named. In those
two decades, statistical analysis was widely discussed; since then, it
has been largely ignored.) When one steps back and looks at the overall
picture, these studies show definite trends in such things as words per
main clause, the frequency with which subordinate clauses are used, the
use of appositives, etc.
For more on the history and purpose of statistical
stylistics, see the "Essay on Statistical
Exercises and KISS Grammar." Most importantly, don't become obsessive
about the results. Enjoy watching the students learn for themselves that
their sentences are too short, too long, or "just right." Ultimately, after
all, the length (or shortness) of their sentences should be their decision.
Our job is to give them tools that they can use to see their writing in
the context of the world in which they live.
The possibilities for statistical analysis
of texts are endless, so the projects described below cover only the basics.
My Own Writing
See the Teachers' Notes (TN) for explanations and suggestions.
The first project
(for KISS Levels 1 and 2) focuses on words per sentence, words
in prepositional phrases, and words that students cannot yet analyze.
Directions for all
Level 1 and 2 Assignments
The objectives of this project are:
1.) to have students focus on sentences as opposed to fragments.
You will probably have to help young students work their way, step by step,
through the directions. For the sake of simplicity, the directions do not
include an analysis of compounds, but you can add that by, for example,
having the students count the number of finite verb phrases and the number
of those phrases that are second, third, etc. elements in the same S/V/C
pattern. For example, in
Sally saw her brother, called him, and then ran after him.
there are three finite verbs, but only the last two ("called" and "ran")
would be counted as compounds. To get the number that corresponds to that
in the "Summary Table of Statistics,"
the number of compounds is divided by the total number of finite verbs.
2.) to let students compare the average length of their own sentences
to those of their peers.
3.) to enable students to see how prepositional phrases add to the
length of sentences (especially in comparison to the writing of their peers).
4.) to show students how much of their own writing they can already
explain (by counting the number of words that they cannot explain as a
percentage of the total words in the passage).
The second project
is for KISS Levels 3 and up. It calculates words per main clause
and subordinate clauses per main clause. These are the primary "yardsticks"
that the researchers in the 1970's used to measure "syntactic maturity."
Remember that the primary objective is to enable students to see how their
own writing compares to that of their peers. Remind the students that "long"
is not the main objective. Main clauses that are far above average cause
problems for readers; main clauses that are well below average sound childish.
Exercises for Analyzing Students' Writing
| Every "Practice/Application" exercises includes
space for two samples of the writing of students at a specific grade level.
Just before these exercises, students are given one of the following "Warm-up"
exercises to go over the method of analysis.
of Students' Writing for Statistical Analysis
| Every "Practice/Application" section has spaces
for a set of five exercises for the analysis of the writing of students
at specific grade levels. Two of these exercises are devoted to editing
the writing of their peers. (These are indexed in KISS
Level 6.1) In two of the exercises (indexed here), students are asked
to do a statistical analysis of the writing of their peers. The fifth exercise
asks students to write their own responses to the prompt and then to analyze
their writing in the same way that they analyzed that of their peers.
Note that the statistical summaries listed
below are from the averaging of all the samples in a set, not just those
used as for exercises within the "Complete" books.
of Third Graders
See also Grade
3, Statistical Stylistics
|Average words per main clause = 6.0
Average subordinate clauses per main clause = .13
As of now, this study includes only the five samples from the 2001
Student Guide for Arizona's Instrument to Measure Standards.
The details for these statistics are in "Cobweb
The prompt for these samples is:
Most people have at least one thing that means
a lot to them. Think of something you have that you would like to keep
forever. Tell about it so that your readers can picture it in their minds
and understand why it is special.
A Graph Format
of Fourth Graders
See also Grade 4, Statistical
|Average words per main clause = 9.9
Average subordinate clauses per main clause = .50
From what I have seen, these numbers seem high. For more on this,
see the Summary of Statistics in "Cobweb
Suggestions about Graphing and Statistics:
The "How" and the "What"
How? Graphs or Stats?
In 1965, Kellogg Hunt published
Structures Written at Three Grade Levels. The study is foundational
in that it demonstrated that words per main clause can be used as a fundamental
yardstick of syntactic maturity. Roughly speaking, the older someone is,
the more words they will use, on average, in written main clauses (distinct
from "sentence"). His results, like the results of most professional research,
were reported with occasional graphs, but mainly with statistics.
As early as third grade students can use the graph
to illustrate words per sentence. Use one column (or row) for each sentence,
and fill in one block in that column for each word in the sentence. The
graph will represent the length (and the variety in the length) of the
sentences in the analyzed passage.
Even some college students have trouble calculating
statistical averages, but third graders can make and understand graphs.
As examples, KISS currently has two graph formats:
At a much more complex
level, students could color code the graphs for the types or functions
of the words in the sentences:
light blue = adverbs
The color codes emphasize how much of the text students can already explain.
You may find, however, that they also help students note the differences
in the grammatical structures that students use—or do not use.
blue = finite verbs
light green = adjectives
green = subjects
brown = complements
pink = words in prepositional phrases
red = conjunctions
purple = Nouns Used as Adverbs, Interjections, or Direct Address
gray = words that the students are not expected to be able to explain
diagonal green/blue cells = contractions, such as "I'll"
You should, of course, modify the directions
and format of graphs to suit the needs of your students. As I have stated
elsewhere in KISS materials, "The study of grammar is a science." And analyzing
real data is the essence of science. Thus these graphs can introduce students
to the scientific method. But "The teaching of grammar is an art." Teachers
need to adjust the graphing method to the abilities and time-constraints
of their students. For example, you might want to have students color-code
just the prepositional phrases. Such a graph will show them how much of
a sentence is composed of such phrases, and where they are located. This
may seem like an extremely simplistic graph, but when you look closely
at the writing of some students, you will find very few prepositional phrases.
A graph of a single passage may be interesting,
but it becomes much more meaningful when placed in context. The KISS graphs
of analyzed passages may give some context, but here again the best context
is graphs of the writing of their peers. Another approach would be to have
each student graph the length and prepositional phrases in the first ten
sentences of different texts (or of chapters of the same text). They could
then share their graphs and discuss the differences in sentence length
and in the frequency and placement of prepositional phrases.
Unfortunately, we educators often miss opportunities
to develop skills over multiple years of education. For example, almost
all primary school students learn how to make graphs. Thereafter, graphing
is basically forgotten. KISS graphs, however, have a purpose and can be
used as frequently as one wishes. Suppose that third graders make three
graphs a year -- two on the writing for the KISS statistical stylistics
passages, and one of their own writing. In fourth grade (and every grade
thereafter), they would already understand the basic methods and purposes
of the graphs. Doing them would be easier and faster, and, if they kept
the graphs on their own writing from previous years, they could compare
the graph of this year's work to those. (Mother Nature promises that, if
nothing else, the sentences will generally become longer.)
As noted above, the most meaningful statistics
are probably words per main clause. Although this can be demonstrated on
graphs, the numbers are what are most important here. KISS, of course,
does not introduce clauses until Level Three. Before that time, graphs
(and any statistical calculations) need to be based on words per sentence.
Once students can identify clauses, graphs may become more time-consuming
than they are worth. The following ambitious worksheet was developed for
a project on analyzing the opening sentences of the nine chapters of Henrietta
Marshall's Stories of Robin Hood Told to Children:
This brings us to the question of what to look at and count.
Some Notes on What to Look At
Once one gets into statistical analysis of
sentence structure, one can find an almost infinite number of things to
count. But the more constructions you add to a statistical study, the more
time it takes to do, and the more cumbersome it becomes to some students.
The following are suggested because they involve either stylistic questions
or questions of syntactic growth.
For KISS Levels One and Two
In addition to what was suggested above, you
might want to have students make graphs or produce statistics on the following.
Sentence Openers --
Prepositional Phrases and/or "But." Some students begin almost
all of their sentences with the grammatical subject (perhaps modified by
adjectives). Some teachers try to get students to vary their sentence openers.
At KISS levels one and two, students will not be able to identify subordinate
clauses or verbal as sentence openers, but they should be able to recognize
prepositional phrases. Using the same format, students can explore that
"rule" about not beginning a sentence with "But."
Format: Have the students look at the
first ten sentences in ten texts. The texts should be from different things
that they are reading, but one text should be something that they themselves
wrote. Count the number of sentences that begin with a prepositional phrase,
or with "But." The results can be reported as statistics (3 of 10, or 30%
begin with "But") or on graphs. (Ten columns represent the ten texts, ten
rows, the sentences. Red blocks represent prepositional phrases; green
Compounding is introduced in KISS Level 1.4. Once you start looking at
sentences from the KISS perspective (once you can identify the grammatical
constructions from which the sentences are constructed and you start using
that ability to analyze style), you will probably agree that compounding
is a very important aspect of style. KISS Level One teaches students to
identify subjects, verbs, and complements, and the compounding of these
three constructions is itself stylistically significant.
If students are analyzing randomly selected
texts, they may have some trouble with this. They will be confused by verbals
(verbs that function as nouns, adjectives, or adverbs). Distinguishing
finite verbs from verbals is the focus of KISS Level 2.1.6, but even after
they have studied that, students may still have trouble mastering it. Thus
they may need some help with the analysis. In most sentences, however,
the subjects, finite verbs, and complements are relatively easy to identify.
Format: You can basically use the same
format as that suggested for sentence-openers, but you may want to have
students make three graphs based on the same texts, one for compound subjects,
one for compound (finite) verbs, and one for compound complements. For
subjects, have them count the first ten subjects (not sentences) and note
the number of them that are in compounds. The same is true for finite verbs.
Complements present a complication.
For KISS Levels Three (Clauses) and Up
Once students get to KISS Level Three, the
most important statistics are words per main clause and subordinate clauses
per main clause. The following project should be relatively easy and fruitful.
Words per Main Clause.
This can be calculated by counting the number of words in a passage and
counting the number of vertical lines in the analyzed passage. (In KISS
analyses, vertical lines indicate the end of a main clause.) As noted previously,
words per main clause is a primary measure of syntactic maturity.
Format: The format you want to use
here really depends on what you want students to see. Simple statistical
averages are helpful in suggesting norms. My college Freshmen, for example,
average between 14.5 and 15.5 words per main clause. When these students
analyze their own writing, the numbers give them a frame of reference against
which they can make some tentative judgments about their own writing --
too short, too long, or "just right." Averages, however, hide the variety
in main clause length.
You may, therefore, want to have students
make graphs of the number of words in the first ten main clauses in two
or more selections, one of which should be their own writing. The problem
with graphs here is that writers don't always write sentences that fit
within the graphs. Suppose, for example, you make a graph on which ten
columns represent the first ten main clauses. And, to be "safe," your graph
has thirty rows to represent words per each main clause. Sooner or later
you will find a passage that has some main clauses that are thirty (or
fifty) words long. This is not a serious problem. Students can simply write
the number of words in that main clause at the top of the column. (I'm
just sharing my experience, and letting you know what to expect.
Words per Main Clause
and Compound Main Clauses. As previously noted, words per main
clause is a fundamental yardstick of syntactic maturity. Stylistically,
however, you will probably be surprised by the frequency with which professional
writers compound main clauses.
Format: You can have students count
compounds as a percentage of total main clauses. Or you can have them use
graphs (as in the preceding suggestion) and color code for compounds. For
example, they can color the blocks for words in the first main clause in
a sentence blue. If the next clause is a compound, color the blocks that
represent it yellow (or whatever color they prefer).
per Main Clause. These projects may be interesting for students
who are working at KISS Levels 3.1 and 3.2. The research of Hunt, Loban,
and O'Donnell suggests that subordinate clauses "blossom" around seventh
grade. Of course, younger writes do use them, but O'Donnell in particular
would probably argue that their clauses are "formulas" -- set strings of
words that are learned as a whole. For example, "When we get home, ask
mother what you should do." Children, according to O'Donnell, pick up the
strings, learn how to replace words in them, and produce more advanced
syntactic structures without real mastery of the underlying constructions.
Formats: Younger students who cannot
calculate statistical averages can represent subordinate clauses per main
clause graphically simply by having every column in the graph represent
a main clause. The rows then represent subordinate clauses. Having analyzed
a text, they can go, main clause by main clause, filling in a block in
a column for each subordinate clause in its main clauses. (Even most professional
sentences rarely have more than five subordinate clauses in a main clause,
so this type of graph does not run into the problem of having sentences
overflow the graph format, which will happen if every block represents
a word in a sentence.) Although these graphs are extremely simply, they
may become very informative for students if, for example, every student
in a class (or small groups of students) analyze the first ten (or twenty)
sentences in texts by different writers (professional as well as student).
They can then compare the graphs and see one aspect of how different writers
use subordinate clauses differently.
For students who can identify subordinate
clauses and do the math, statistical averages are fairly easy. In KISS,
we place brackets [ ] around subordinate clauses and vertical lines at
the end of main clauses. To calculate subordinate clauses per main clause,
students simply need to count the opening brackets in an analyzed passage
and divide that number by the number of vertical lines.
These statistics (and even more so the graphs)
become interesting when used to compare different writers or groups of
writers. Imagine, for example, a graph in which the columns represent ten
(or more) different writers, and the columns represent the average number
of subordinate clauses per main clause used by each writer. My "educated"
guess is that some professionals would use very few, whereas others would
use quite a few. The graph would clearly show students that the use of
subordinate clauses can vary -- it is not a matter of right or wrong, but
rather a matter of style.
Clauses per Main Clause. I noted in the introduction to
this section that a student asked, "Can a subordinate clause be in a subordinate
clause that is in a subordinate clause that is in a subordinate clause?"
Students who study Level 3.1 of the KISS Approach will automatically learn
the answer to that question, but the answer raises two more -- how often
should this occur, and how deeply should clauses be embedded? One of the
major problems, particularly of my weaker adult writers, is that they write
extremely long sentences, with very heavy embedding of clauses, and they
lose control. "Cobweb Corner" has a summary page on the "Type and Embedded
Level of Subordinate Clauses." Again, let me note that those statistics
are very tentative in that many more samples need to be studied, but your
students can do such studies themselves.
Format: A very simple study for a single
text is to have the columns represent main clauses, and the rows represent
subordinate clauses. Color code the embedding levels. For example, Level
One = blue; Level Two = yellow; Level Three = green; Level Four = red,
etc. Then simply go through the text, main clause by main clause, and fill
in the blocks. Graphs of texts by different writers would probably be more
interesting in terms of stylistic differences.
of Stylistic Interest. Any grammatical construction can be studied
stylistically, but for purposes of natural syntactic development (the natural
increase in sentence length and complexity) and for purposes of style,
consider gerundives, appositives, and noun absolutes that
function as adverbs. Each of these can be fairly easily counted (and
graphed) "per main clause" in ways noted above. The professional studies
that I have seen simply count these as individual constructions, but you
may find it more interesting to also calculate the average number of words
per construction. Young writers, for example, may use these constructions,
but the constructions are very short:
Walking home, we saw an accident.
Older writers are more likely to elaborate the construction:
Walking home because it was such a beautiful day,
we saw an accident.
The gerundive phrase in the first example consists of two words -- an "average"
of two words per gerundive. In the second example, that same phrase has
been modified by a seven-word subordinate clause, thereby creating a nine-word
gerundive phrase. Averaged together, the two examples are 5.5 words per
gerundive phrase. Researchers who simply count the appearance of these
construction miss, I would suggest, an important aspect of stylistic and
The "Chicken" and "Aluminum" passages were used
by Kellogg Hunt in an interesting study of "late blooming" constructions."
See the "Notes" for more information.
"Chicken" and the "Aluminum" Studies
the Bull Pen
| The items listed here are statistical projects that were
done in the past. They may be developed in more detail later.
Opening Sentences from the Nine Chapters of
Stories of Robin Hood Told to the Children,
by Henrietta Elizabeth Marshall
| The data sheet for this set of exercises may
be too ambitious. It includes information about the number of words in
prepositional phrases, the total number of words that students can explain,
the opening words of sentences, the compounding of finite verbs, and the
relative number of the types of complements. All of this can be time-consuming,
so you may want to do all of just one exercise, or you may want to limit
the analysis to just the number of words in prepositional phrases, or the
number of sentences that open with something other than simple subject
"The Ethical and Pedagogical Importance of
the Principle of Habit"
| This is an excellent essay that all high school
students should read and discuss. (I first read it as a Freshman in college.)
The seven selections, a total of 1023 main clauses, average 24.4 words
per main clause.
Analytical Questions -- or "What Counts as What?
One Main Clause? Or Compound Complements?
From: "Niagara Falls," by Rupert Brooke in Modern
Essays Selected by Christopher Morley. New York: Harcourt, Brace, 1921.
New York: Bartleby.com.
The following is a 106-word sentence:
There are touts
insinuating, and touts
raucous, greasy touts,
brazen touts, and upper-class,
refined, gentlemanly, take-you-by-the-arm-touts;
who intimidate and touts
who wheedle; professionals,
amateurs, and dilettanti, male and female;
who would photograph you with your arm round a young lady against a faked
background of the sublimest cataract,
who would bully you into cars, char-?-bancs, elevators, or tunnels, or
deceive you into a carriage and pair,
who would sell you picture post-cards, moccasins, sham Indian beadwork,
blankets, tee-pees, and crockery, and touts,
finally, who have no apparent object in the world, but just purely, simply,
merely, incessantly, indefatigably, and ineffugibly to tout.
Adjectival, Adverbial, or Interjection?
From: "Samuel Butler: Diogenes of
the Victorians," by Stuart P. Sherman in Modern Essays Selected by Christopher
New York: Harcourt, Brace, 1921. New York: Bartleby.com.
UNTIL I met the Butlerians
I used to think that the religious spirit in our times was very precious,
there was so little of it.
Is "there was so little of it" a
run-on, a subordinate clause that functions as an intejection, or an adverbial
subordinate clause? (I counted it as an adverbial clause even though there
is no conjunction because it indicates why Sherman thought it was so precious.)
He was not even—till
his posthumous disciples made him so—a person of any particular importance.
Should the "till" clause
be counted as adverbial, as an interjection, or both? (I counted it as
an interjection. Counting it as both messes up the total number of clauses.)
Prepositional Phrase or
From a student's paper:
I have to wear a dress shirt and tie to work.
Is "work" the place or the act?
Can he take it off while working? (This is a 50-50 call based on subjective
From students' papers:
"The women also have the right to dress as
they think is appropriate."
"The women also have the right to dress [Adv. as [Inj
they think] *it* is appropriate]."
Women demand more rights, more freedom, and above
all they demand to be treated as equal with men.
The "as equal" is a reduced subordinate
clause -- "as if they are equal with men."
What if the person they are trying
to make a deal with was a woman?
This "What if" is a reduction of
"What happens if . . . ."
[W]e are after a system that scarcely knows it is a system
. . . . (from "The Almost Perfect State," by Don Marquis.
Essays: Selected by Christopher Morley NEW YORK: HARCOURT, BRACE, 1921;
NEW YORK: BARTLEBY.COM, 2000).
"Are after" = "want"
More than One Question in
Compound Subject or Appositive?
Adjectival Clause or Interjection?
From: "Beer and Cider," by George Saintsbury in Modern
Essays Selected by Christopher Morley. New York: Harcourt, Brace, 1921.
New York: Bartleby.com.
The curious “white ale,” or lober agol—which,
within the memory of man, used to exist in Devonshire and Cornwall, but
which, even half a century ago, I have vainly sought there—was, I believe,
drunk quite new; . . . .
Since "lober agol" is simply a different name for "white
ale," should it be counted as an appositive instead of a compound subject?
(I counted it as a subject.)
Since the "which" clause is set off by dashes, should
it be counted as adjectival or as an interjection? (It's both, but I count
it as an interjection.)
From students' papers:
Syntactically Awkward Sentences
If a problem can be resolved among each other that would be
the best solution.
The function of the "that" clause is unclear.
The student initially stated the solution as an adverbial clause. Having
done so, he (unconscioiusly) realized that that cluase cannot be a subject,
and so restated it with the "that," and then tacked on the verb. What is
probably meant is "Resolving a problem within the group would be the best
I think the women are getting a little carried away with filing sexual
harassment charges and believe that an agreement can be made between the
supervisor and his female employees.
What is the subject of "believe" - "I" or "women"?
Drop the "I think"? -- The women are getting a little carried away with
filing sexual harassment charges. An agreement can be made between the
supervisor and his female employees.???