Content-Type: text/html
SPECIFICITY AND IMAGERY IN WRITING:
TESTING THE EFFECTS OF "SHOW, DON'T TELL"
James W. Tankard, Jr.
Laura J. Hendrickson
Department of Journalism
University of Texas at Austin
Austin, Texas 78712
(512) 471-1997
James W. Tankard, Jr., is a professor and Laura Hendrickson is a doctoral
candidate in the Department of Journalism at the University of Texas at Austin.
They are grateful to the following instructors who allowed time for their
students to participate in this study: Jane Ballinger, Bill Israel, and Randy
Sumpter.
ABSTRACT _ One of the most common pieces of advice on writing is "Show, Don't
Tell." This study clarifies that advice by explicating the meaning of
"showing," specifying some possible effects of "showing," and conducting an
experiment to investigate those effects. The study found "show" sentences were
usually perceived as more interesting, engaging, and informative than their
"tell" counterparts, but the evidence was less clear that "show" sentences are
seen as more objective, clearer, or more believable. _ END OF ABSTRACT
One often-heard piece of conventional wisdom about good writing, whatever its
genre, is "show, don't tell." The advice often is given in writing books and by
writing teachers. Mencher (1991) in his book News Reporting and Writing, says:
"We might start with Tolstoy who, in describing the strength of his masterwork
War And Peace, said, 'I don't tell; I don't explain. I show; I let my
characters talk for me'" (p. 138).
Tolstoy was a novelist, of course, but this advice has not been restricted to
novelists, as is clear from its mention in a journalism textbook. Journalists
also hear of this traditional ground rule of good writing. Mencher goes on to
say:
One of the reporter's first writing rules might be: Show, don't tell. Telling
makes readers passive. Showing engages readers by making them draw the
conclusions, see the significance of the facts the writer presents. Good
writers let the words and actions of the participants do the work (p. 139).
So the journalist is advised to create an experience for the reader _ an image,
a scent, a sound, a touch _ and to make the reader draw the conclusions. Noble
(1991), in his advice to fiction writers, says: "Why show and not tell? By and
large, readers pick up a book to be entertained, and there's little
entertainment value in being told what they are reading" (pp. x - xi).
So what exactly does this "show, don't tell" folk wisdom mean? And what part
does it play in the writings of a journalist? Some elements of the rule may
seem more immediately applicable to journalistic writing, such as the call to be
specific, which can also make writing more visual and create what Noble calls
"word-pictures." He says:
There are a lot of adjectives to describe writing which touches the reader and
creates a word-picture: vigorous, vivid, sensuous, sentimental, challenging.
Writing like this brings a story to life, and there is a singular reason for it.
The writer has injected drama, he or she has made things happen and has tickled
the imagination . . .We must never be far from the drama trough because this is
where we get the food for "showing" instead of "telling" (p. x).
The specificity and imagery called for by "showing" often may be missing in
journalistic writing, perhaps needlessly. As Carol A. Turkington (1993), has
said:
It took me years to undo the damage of my newspaper beginnings. But with the
help of some gifted editors, I slowly and painfully learned that creating an
article with some color didn't mean I'd sold my journalistic soul. It meant I'd
found it. I learned that my job as a writer is not just to string together a
bunch of facts and phrases, but to weave a sensual tapestry of sights and sounds
that help readers not simply to know, but to understand.
Successful fiction and nonfiction writer Tom Wolfe (1973), in his discussion of
the New Journalism, says:
If you follow the progress of the New Journalism closely through the 1960's, you
seen an interesting thing happening. You see journalists learning the
techniques of realism _ particularly of the sort found in Fielding, Smollett,
Balzac, Dickens and Gogol _ from scratch. By trial and error, by "instinct"
rather than theory, journalists began to discover the devices that gave the
realistic novel its unique power, variously known as its "immediacy," its
"concrete reality," its "emotional involvement," its "gripping" or "absorbing"
quality (p. 31).
We should point out that the emphasis on "showing" in writing has caused some
writers to speak up in defense of the art of telling. Marrazzo (1993) reminds
us that abandoning telling for the sake of exclusively showing can cause a
reader to feel "distanced" (p. 32). She calls for a combination of showing and
telling to draw the reader in. For the magazine journalist, this may be the
most practical and useful way to approach writing, although it seems clear that
the "show, don't tell" advice has endured.
Showing rather than telling can be accomplished in a variety of ways. A number
of aspects to the "show, don't tell" advice have evolved:
1. Be concrete. Newsom and Wollert (1988), in advising writers to be
concrete, recommend using "steak and beer" instead of "sustenance" (p. 57). The
authors note that problems often arise not from a word's unfamiliarity, but from
its level of abstraction. They say:
. . . all words, to some degree, are abstract. But the closer the word is to
something visual, something real, the easier the word will be to understand.
Good writers prefer words that give the reader something to see. . . . Good
writing is characterized by concrete nouns and action words (p. 57).
Strunk and White, in The Elements of Style, tell us that good writing is
concrete and specific. They suggest that this calls up pictures for the reader.
And calling up pictures is one way to show, rather than tell. Strunk and White
say:
If those who have studied the art of writing are in accord on any one point, it
is on this: the surest way to arouse and hold the attention of the reader is by
being specific, definite, and concrete. The greatest writers _ Homer, Dante,
Shakespeare _ are effective largely because they deal in particulars and report
the details that matter. Their words call up pictures. (p. 21).
The authors offer the example that the sentence "A period of unfavorable weather
set in" is not as effective as "It rained every day for a week."
Concrete language is at the lower levels of the "abstraction ladder" described
by Hayakawa (1964). He gives the example "Mrs. Levin makes good potato
pancakes," and notes that it is at a lower level of abstraction than "the
culinary art has reached a high state in America."
2. Be specific. Newsom and Wollert say that when general statements are used,
they should be backed up by specific examples to make the communication clear.
For instance, they recommend "The audience stood and applauded for five minutes"
rather than "He was a big hit with the audience" (p. 58). They also offer the
following example:
Consider this sentence: "People in the village eat a lot of fruit." The
sentence is clear enough, but it is general. "Every villager eats two pounds of
bananas, peaches and apples daily" is specific (p. 58).
Other authors have written about the need for specificity in writing. Noble
says that specificity is essential in appealing to the senses. He also says,
"Keep things specific, and what do we have? Imagery which becomes drama which
becomes showing" (p. 137).
Conrad (1993), in an article about "the four deadly sins of description," sums
up the fourth deadly sin this way: Don't generalize _ Be specific. He calls
this Chekhov's Golden Rule and asks the writer not to say the sky was "sort of
bluish," but to say that it was "like the sky in a child's painting, the same
deep blue from top to bottom." Is this too much space _ even for a magazine
feature? Specificity need not always take more space. For example, Conrad also
says:
. . . remember when you sit down to write it is: not a drink but a martini;
not a dog but a poodle; not a flower but a rose; not a sleigh but a Rosebud;
not a hat but a borsolino; not a cat but an Abyssinian; not a gun but a .44
colt on a frontier frame; not a painting but Manet's "Olympia."
3. Use language that appeals to the senses. Showing implies that we appeal to
the senses. When we show people something, we give them a visual image. We
appeal to their sense of sight and create what Noble (1991) calls "a picture
forged from our words." Noble (p. x), says "Writer's need help in learning how
to develop word-pictures for readers."
Roberts has said, "It took me years to appreciate it, but there is no better
admonition to the writer than 'make me see'" (1994, p. 7).
But we can appeal to other senses as well. McKenzie (1994), in an article
about writing fiction for children, said, "Recall sights and sounds, tastes and
smells. Be brief. Choose words that are graphic."
4. Use quotes and/or dialogue. Noble, who calls it dialogue because he's
talking about fiction, says, "Good dialogue . . . serves a useful purpose
because it carries dramatic impact, it shows rather than tells, and the reader
can get immediately involved" (p. 54). Wolfe says dialogue is the single most
effective device for establishing character (p. 31). Spikol (1993), in a
discussion of nonfiction writing, says "it is very difficult to instill life
into characters without using what they themselves use to communicate: their own
words" (p. 72).
5. Use figures of speech. Similes, metaphors and other literary devices can
bring vividness to writing. Some writers choose to "show" by using metaphors to
create a visual image, as in this line from a National Geographic story about
Cyprus: "Time stands still along narrow Hermes Street, which slashes across the
center of the Old City of Nicosia like a deep wound that has never healed"
(Szulc, p. 109). Noble (p. 44), while cautioning that metaphors and similes
not be overused, calls for the writer to use them skillfully and says,
"Comparing and contrasting images is what similes and metaphors do, and the
purpose is to build up that word-picture and clarify it. To show it, not tell
it."
6. Write in terms of action or narrative. Conrad (1993) explains his first
deadly sin of description this way: "Don't let your description, no matter how
beautifully written, bring your narrative to a halt." But not just any
narrative will do. Noble says that "narrative is the heart of storytelling" and
is "what sweeps us along" but adds that it must contain drama. He says:
Good narrative has drama . . . Where the drama slips away, the narrative becomes
lecture instead of storytelling, and the reader loses that all-important
word-picture. But narrative with drama . . . ah, that's showing not telling
(p. 104).
7. Use strong verbs. Strong verbs can lead to greater specificity, and thus
to showing. In Gibson's (1989) book The Writer's Friend, the author concedes
that the verb "is" does the job "properly on occasion" but that:
We have no weaker verb than is. Any professional writer can find something
stronger than that. Your vocabulary surely contains a livelier verb that tells
us something occurs, that something acts, that something happens (p. 36).
With his examples, Gibson also shows that stronger verbs can lead to greater
specificity throughout the sentence. For example, the sentence "Farmers in this
area are poor" becomes "Farmers in this area don't have enough money to buy this
year's seed" (p. 37). The second sentence shows, rather than tells.
One way to address the usefulness of the "show, don't tell" advice is to gather
empirical evidence for the assumption that showing has a different effect on
readers than does telling. The goal of this study was to examine the
differences in the way readers respond to sentences that "show" rather than
"tell."
For the purpose of this study, we have concentrated on the related aspects of
concreteness, appealing to the senses, and specificity of language.
It seems to us to be useful to think of most of the other factors mentioned
above as separate and distinct variables that could (and probably should) be
examined in other studies. For instance, it would be easy to look in another
study at the effects of direct quotations vs. indirect quotations.
Hypotheses
1. "Show" sentences will be seen by readers as more interesting than "tell"
sentences.
This hypothesis is expected to be supported because one of the main arguments
for "showing" rather than "telling" in writing is that showing makes writing
more engaging, entertaining, attention-arousing, or gripping than telling
(Mencher, 1991, p. 139; Noble, 1991, pp. x-xi; Strunk and White, 1979, p. 21;
Wolfe, 1973, p. 31).
2. "Show" sentences will be seen as more believable than "tell" sentences.
This hypothesis is expected to be supported because "telling" often presents
the writer's conclusions without supporting detail. For instance, writers who
are "telling" might say that a concert was "terrific," that a view was
"awesome," or that an accident was "tragic." Readers should find these
conclusions drawn by the writer to be less objective, and therefore less
believable, than writing that "shows" by reporting specific, concrete details.
An article on nonfiction writer John McPhee provides some testimony about the
link between concrete details and believability. The piece states that "McPhee
has a passion for details, for they convince readers that he deals in
actualities" (Howarth, 1976, p. 10).
Research Questions
1. Which is perceived as more clear, a "show" sentence or "tell" sentence?
We did not hypothesize about the effects of "showing" versus "telling" on
sentence clarity because arguments have been made on both sides. Newsom and
Wollert and others suggest that specific, concrete language makes writing
clearer by giving the reader something to visualize, suggesting that "show"
sentences should be clearer. But it also is true that "show" sentences often
will be longer, and research on readability (Klare, 1963) suggests that longer
sentences typically are more difficult to read.
2. Which is perceived as more informative, a "show" sentence or a "tell"
sentence?
We also did not hypothesize about the effects of "showing" versus "telling" on
sentence informativeness because arguments exist on both sides. An argument can
be made that "show" sentences are more informative because they provide more
detailed information, but, again, if a sentence gets too long because of added
detail, it could become less readable, and thus less informative. At least one
author suggests that "showing" isn't always enough if a writer wants to convey a
character's thoughts or feelings and that showing-only writing can feel
"distanced" to the reader (Marrazzo, 1993, p. 12).
Method
The hypotheses and research questions were investigated by means of an
experiment.
Overview
The experiment was designed to present the same sentences to some participants
in a "show" version and to other participants in a "tell" version. Ten
sentences were presented. Two experimental packets were prepared so that a
sentence that appeared in a "show" version in one packet would appear in a
"tell" version in the other packet. In each packet, "show" and "tell" sentences
alternated.
Each sentence was followed by six semantic differential scales selected to
measure the dependent variables specified in the hypotheses and research
questions. The packet also contained some questions about the subjects'
frequency of reading daily newspapers, magazines and books, and their gender.
Packets of the two types were arranged in a random order and passed out to
students in selected journalism classes. Participants were asked to give their
"honest reaction" to the sentences. They were shown how to use the rating
scales by means of an example on the blackboard. They were informed that the
people sitting next to them might not have the same questionnaire, and that they
should concentrate on their own packets.
Experimental Materials
Recent issues of several popular magazines (Reader's Digest, The New Yorker,
National Geographic, National Wildlife, and Guideposts) were searched for
examples of "show" sentences, or sentences containing specific, concrete
language that appealed to the senses. These "show" sentences were rewritten so
there was a "tell" sentence for each "show" sentence. The "tell" sentences
were produced by stripping out specific, concrete language and replacing it with
more general terms. Ten pairs of sentences were used in the experiment.
Six semantic differential scales were selected to measure the dependent
variables specified in the hypotheses and research questions. Two scales were
selected to measure the dependent variable specified in each of the two
hypotheses and one scale was selected to measure the dependent variable in each
of the two research questions.
The quality of "interestingness" mentioned in Hypothesis 1 was measured with
the scales "interesting-dull" and "engaging-unengaging." The quality of
"believability" mentioned in Hypothesis 2 was measured with the scales
"objective-subjective" and "believable-unbelievable." (There were two steps in
our logic here: that "show" sentences should be seen as more objective than
"tell" sentences, and that objective sentences should be seen as more
believable.) The quality of "clarity" mentioned in Research Question 1 was
measured with the scale "clear-unclear." The quality of "informativeness"
mentioned in Research Question 2 was measured with the scale
"informative-uninformative."
Participants
Research participants were 80 undergraduate students in four journalism classes
_ three sections of "Writing for the Mass Media," the first journalism class
taken by most students, and one section of "Copy Editing." The experimental
materials were presented in the second or third week of the semester, before the
topic of "Show, Don't Tell" was discussed in class.
Participants were randomly assigned to packets by putting the packets in an
order determined by a random numbers table before taking the packets to class.
Then the packets were passed out systematically row by row.
Analysis
Differences between the "show" and "tell" versions in the ratings of the
sentences were examined by means of t-tests. (The sample size was actually 79
for most of the t-tests because one subject only responded on one of the six
scales for each sentence.)
Results
Results are presented for each of the ten sentences in Tables 1-10 (copies of
the tables can be obtained by writing the authors).
Hypothesis 1, suggesting that "show" sentences will be seen by readers as more
interesting that "tell" sentences, was generally supported. For seven of the 10
sentences, the "show" version was rated as significantly more interesting than
the "tell" version. For nine of the 10 sentences, the "show" version was rated
as significantly more engaging than the "tell" version.
Hypothesis 2, suggesting that "show" sentences will be seen as more believable
than "tell" sentences, did not receive strong support. For four of the ten
sentences, the "show" version was seen as significantly more objective than the
"tell" version, but for one sentence the "tell" version was seen as
significantly more objective than the "show" version. For only one of the ten
sentences was the "show" version seen as more believable than the "tell"
version.
The first research question asked whether a "show" sentence or a "tell"
sentence would be seen as more clear. The answer was mixed, with the "show"
version rated as significantly more clear for four sentences and the "tell"
version rated as significantly more clear for two sentences. In the four
sentences in which the "show" version was rated as more clear, this happened
even though the "show" sentences were longer than the corresponding "tell"
sentences.
The second research question asked whether a "show" or a "tell" sentence would
be seen as more informative. The answer was more definite here, with the "show"
version rated as significantly more informative for eight out of ten sentences.
Conclusions
Numerous authorities have suggested that writers should "show, don't tell" if
they want their writing to be more effective. This study both adds to and asks
questions about that conventional wisdom by taking a closer look at what
"showing" might mean and by specifying some of the effects of "showing" rather
than "telling."
"Showing" in writing could refer to a number of things, including being
concrete, being specific, using language that appeals to senses, using quotes
and/or dialogue, using figures of speech, writing in terms of action or
narrative, and using strong verbs. In this study, we defined "showing" as being
specific, being concrete, and appealing to the senses.
Our experiment found strong evidence that, as many experts predicted, "show"
sentences are seen as more interesting and engaging than "tell" sentences. The
experiment also provided evidence that "show" sentences are seen as more
informative than "tell" sentences. While "showing" rather than "telling"
apparently can make writing more interesting, engaging, and informative to
readers, this study, at least, did not provide conclusive support for the idea
that it always makes writing more believable and clear. But the hypothesis and
the research question addressing believability and clarity produced such mixed
results that the role of other variables becomes worth considering.
The hypothesis that "show" sentences would be seen as more believable than
"tell" sentences was tested with the scales of believable-unbelievable and
objective-subjective. We found some evidence, but less strong than in the first
hypothesis, that "show" sentences are seen as more objective than "tell"
sentences, but this did not carry over to make the "show" sentences more
believable. While four of the "show" sentences were seen as more objective than
their "tell" counterparts, one of the "tell" sentences was seen as significantly
more objective than the "show" version. In only one of ten tests was the
"show" sentence more believable than the "tell" sentence.
That one sentence is worth examining for some clues about sentence
believability, however. The "show" version ("At the local playground, weeds
poke through cracked concrete and climb over collapsed, rusted swing sets.") was
seen on the average as nearly one point more believable on the
believable-unbelievable scale than the "tell" version ("The local playground is
in disrepair."). Like a good "show" sentence should, the first version allows
the reader to draw his or her own conclusion about the state of the playground,
while the second sentence _ the "tell" version _ draws a conclusion without
providing evidence or imagery.
It is difficult to tell why this distinction did not hold up to produce
statistically significant differences for the other nine sentences, but other
variables simply may have played a stronger role. For example, two of the
sentence pairs were in first person in both versions, which might tend to make
both of them believable, and subjective language may occasionally have slipped
into the "show" version of some sentences.
Let's look at the "show" sentence that was seen as less objective than the
"tell" version: "A drunk walking unsteadily home from a bar on November 2 saw a
red Dodge sedan snap a U-turn, whip into a side street and cruise slowly by an
alley." The word "drunk" could have interfered here in two ways. First, it
might have been seen as a subjective and derogatory term in itself, damaging the
perception of the sentence's objectivity. Second, the sentence is reporting
what was seen by a man who has been described as "drunk," perhaps calling into
question the objectivity of the man's own idea of what he saw.
When we had subjects rate sentences for clarity, this also produced mixed
results, although twice as many "show " sentences (four) as "tell" sentences
(two) were identified as clearer. The four "show" sentences that were rated as
more clear than corresponding "tell" sentences were longer than the "tell"
sentences. This is contrary to the most likely prediction from readability
research, which is that shorter sentences are more understandable. The fact
that these four sentences were seen as more clear than their shorter, more
simply constructed "tell" versions provides more support for the basic power of
"showing."
In general, the study supports the recommendation to writers to "show" rather
than "tell," particularly if they want their writing to be more interesting,
more engaging, more informative, more objective, and clearer. For these five
dependent variables, particularly for the first three, the power of "showing"
rather than "telling" seems strong. The collection of 10 sentences chosen for
the subjects to read varied in length, structure, and subject matter, and the
show vs. tell distinction seems to have overridden all of those factors to
produce statistically significant differences in favor of the "show" sentences
most of the time. The "show" sentences were rated as more interesting,
engaging, and informative regardless of whether they were long or short, and
regardless of whether they were about virtual reality, a local playground, or a
view from a mountaintop. And the "show" sentences were rated higher than the
"tell" sentences, not by contrast, but standing alone. In other words, because
subjects saw only one version of each sentence pair, their ratings were not
based on comparisons between the two versions, but on the merits of each
sentence on its own.
This is not to say that there is not a place for "tell" writing, however. We
noted earlier that Marrazzo warns that too much "showing" can make the reader
feel distanced. In fact, when we were gathering sentences to use in our
experiment we found examples in Reader's Digest and other publications of
sentences that seemed to be deliberately combining showing and telling.
This study has some limitations. The experiment was conducted with journalism
students at a large state university as the research participants. Before we
can assume that the results apply widely, further testing should be done with
other kinds of people.
Another limitation is that the study was done with the sentence as the unit
that people were reading. The subjects did not see these sentences in the
context of a larger story. The drawbacks or advantages of a show-only approach
might be different in the more realistic context of a full article, since
sentences often gain their meaning, interestingness, informativeness,
believability, and clarity by their association with the sentences around them.
This question should be investigated further in studies in which "show" and
"tell" are manipulated in articles rather than sentences. This kind of research
might allow the effects of "showing" to build up over a number of sentences, and
some of the possible effects of "showing" _ on believability, for instance _
might become even more apparent. This approach also opens up the possibility of
studying different dependent variables and more complex independent variables _
such as what combination of showing and telling is more likely to make an
article memorable or an opinion piece persuasive.
References
Conrad, Barnaby. (1993, May). "The Four Deadly Sins of Description," The Writer,
pp. 9-12.
Gibson, Martin L. (1989). The Writer's Friend. Ames: Iowa State University
Press.
Hayakawa, S. I. (1964). Language in Thought and Action, 2d. ed. New York:
Harcourt, Brace & World.
Klare, George. (1963). The Measurement of Readability . Ames, Iowa: Iowa State
University Press.
Marrazzo, Carol-Lynn. (1993, October). "Show and Tell," Writer's Digest, pp.
32-35.
McKenzie, Ellen Kindt. (1994). "Appealing to the Senses When Writing for
Children," The Writer, pp. 21-23.
Mencher, Melvin. (1991). News Reporting and Writing, 5th ed. Dubuque, Iowa: Wm.
C. Brown.
Newsom, Doug, and James A. Wollert. (1988). Media Writing: Preparing
Information for the Mass Media, 2d ed. Belmont, Calif.: Wadsworth.
Noble, William. (1991). Show, Don't Tell: A Writer's Guide. Middlebury,
Vermont: Paul S. Eriksson.
Roberts, Eugene L., Jr. (1994, May). "Writing for the Reader." The Red Smith
Lecture in Journalism. The Department of American Studies, University of Notre
Dame, Notre Dame, Indiana.
Spikol, Art. (1993, June). "How Will You Tell It?" Writer's Digest, pp. 72-73.
Strunk, William, Jr., and E. B. White. (1979). The Elements of Style. New
York: Macmillan Publishing Co., Inc.
Turkington, Carol A. (1993, October). "Color Your Articles Sold," Writer's
Digest, pp. 29-31.
Wolfe, Tom. (1973). The New Journalism. New York: Harper & Row.