Although business communication relies heavily on the visual, current
approaches to graphics and text design are prescriptive and unsystematic.
A 12-cell schema of visual coding modes and levels provides a model for
describing and evaluating business documents as flexible systems of visual
lar^uage. Emphasizing clarity and objectivity, the "information design"
movement has generated guidelines for creating functional visual displays.
However, visual language in business communication is seldom rhetorically
"neutral" and requires adc^tation to the contextual variables of each
document, a goal the writer can achieve by combining visual and verbal
planning in the same holistic process.
A Systematic Approach to Visual
Language in Business Communication
Charles Kostelnick
Iowa State University
WRITING DOES NOT COMMUNICATE unta it is seen, until
it becomes an artifact for visual inspection (Gelb, 1980). From
the primitive making of lists, business commimicators have
created permanent records by transcribing symbols onto two-
dimensional surfaces. With the proliferation of print technology
formerly reserved for specialists, contemporary business com-
municators have unprecedented control over visual design.
However, participation in a visually intensive "print culture"
(Ong, 1982) does not guarantee visual sensibility. While the
"technologizing of the word" opens new avenues for visual
expression—everything from page design to bar charts to
scanned images—ihe thing-like quality of texts and the per-
ceptual reliance on the eye for information processing are
largely unacknowledged and imexamined. In theory and
application, visual language in business communication re-
mains rudimentary and prescriptive, and confined to a super-
ficial or decorative rather than a functional or rhetorical role
in the communication process.
To compensate, we need a systematic approach for analyzing
how visual elements affect the readability and the rhetoric of
business documents, and even for describing the visual elements
themselves. Terms such as "layout," "format," and "graphics"
lack precision and connote low level skills rather than serious
visual thinking. The visual language of business communication
is a flexible system of symbols, marks, and spatial variations
29
30 THE JOURNAL OF BUSINESS COMMUNICATION ' 25:3:SUMMER 1988
that operates on several levels within and outside the text. Used
effectively, visual langus^e enables readers to process informa-
tion—in letters, brochures, reports, newsletters—clearly and effi-
ciently. The "infonnation design" movement has generated
guidelines for creating functional, user-oriented documents. How-
ever, far from being generic or rhetorically "neutral" (Kinross,
1985), visual choices are bound to a perceptual and rhetorical
context. Each document embodies an autonomous system re-
quiring adaptation to a particular audience and purpose. Visual
planning and invention, therefore, must be part of the global
strategy of the document and thoroughly integrated into the
process.
THE SPECTRUM OF VISUAL INTENSITY
Writing and reading are intensely visual acts in which writer
and reader negotiate meaning on a two-dimensional field. Since
this field can be varied and manipulated into infinite configura-
tions, all documents are not visual in the same way or to the
same d^ree (Bemhardt, 1986). Thus we could devise a spectrum
of visual intensity ranging from very low to very
LOWVISUAL ^ ^ HIGH VISUAL
INTENSITY INTENSITY
What, then, are the limits of such a continuum? What character-
istics determine whether a document has a high or low level of
visual intensity?
At the low end we might place a conventional essay or report,
which submei^es explanations and arguments in continuous text,
or a novel with its symbols, metaphors, and complex narrative
strands that readers must extract from seamless blocks of text
(Bemhardt, 1986; Herrstrom, 1984). StUl, even a novel relies on
spatial segmentation—books, chapters, paragraphs—to signal tran-
sitions in place, time, and voice (Sterne even includes his own
diagrams). The low end of the continuum calls for something far
less visual: ideally, a string of unembellished signs flowing in a
sii^e direction, unrestricted by margins or pages. Rendering the
text entirely in upi>er case and eliminating punctuation would
further reduce visual dependency. The reader would then have no
sense of spatial direction other than the syntactic relations of
linguistic signals. Like a message moving across a video screen in
VISUAL LANGUAGE • KOSTELNICK 31
an airport, the text would approximate the perceptual imperman-
ence of an utterance.
At the high end of the visual continuum, des ^i i ^ and seeing
assume a far greater role in shaping and receivir^ the messi^e. As
text is segmented into chimks, chimks into lists, and lists into
matrices, charts, and dis^rams, words become "thing-like" (Ong,
1982) entities transfixed spatially and graphically (through line
work, arrows, geometric forms) on the page. Even more thing-like
and reliant on visual inspection are representative images—for
example, a picture in a proposal of key employees seated aroimd a
conference table. Still, visual intensity does not necessarily depend
on realism because messages can be purely visual but abstract. For
example, to show the enrollment patterns of a college, I can use
pictures of each class or even of each student (achievii^ optimum
realism), but a bar chart or a line graph would display the trends
more clearly and concisely and with greater visual immediacy.
Thus on one end of the spectrum we might place a piurely (or as
pure as possible, given that it must be seen) linguistic messt^e, on
the other end a purely visual message which defies simple transla-
tion into words. However, the conditions for the extremes rarely
obtain: most business documents are visual hybrids. When trans-
ferred to the page, the linear messs^e flashing across the screen
undergoes a visual metamorphosis: words are shaped by a type
style, xmderlined, thickened, s^mented into lines, distributed
vertically and horizontally across the page, and circumscribed by
graphic marks. The picture of the employees at the conference
table, on tiie other hand, can be reduced to a line drawing, a plan
view of the office, a corporate logo, or a series of abstract marks,
each of which (including a color photo) may depend on linguistic
coding to make any sense. And so inside the extremes of the con-
tinuum occurs a vast interminglii^ of visual signs, and of the visual
and the verbal. Like a chemical reaction, the signs are interdepen-
dent: different combinations alter the visual effects of individual
signs, which collectively alter the meaning of the message. Because
these combinations are bound perceptually and rhetorically to
context, each document contains an idiosyncratic system of visual
language. The visual intensity of each system depends on the inter-
play of several levels and coding modes.
A SCHEMA OF VISUAL LANGUAGE
What, then, are the constituent elements of visual language?
First, we need to define the coding modes that comprise the raw
32 THE JOURNAL OF BUSINESS COMMUNICATION • 25:3:SUMMER 1988
materials of visual language; and second, the levels at which the
modes operate within actued documents.
Visual language can be encoded in three modes:
• Alphanumeric/symbolic: textual particles, including letters
that comprise linguistic units; numbers; and symbols such as
parentheses, dollar signs, and ampersands.
• Graphic: marks, lines, tones, and colors which encode geo-
metric shapes (squares, circles, bars), realistic im£^es (a face, a
building), or abstract forms (a corporate logo).
• Spatial: the distribution of alphanumeric/symbolic and
graphic signs across a plane, creating visual syntax among textual
or graphic particles.
These coding modes are integrated on four distinct levels: intra-,
inter-, extra-, and supra-textual. Together these four levels and
three modes comprise a 12-ceIl schema of visual language. '
The Intra-textual level (cells 1-3) is coded primarily in the
alphanumeric/symbolic mode and controls the local form, size,
posture, and embellishment of textual elements. Consider the fol-
lowing sample of text:
During the past month, our clientele in each region has increased as fol-
lows: E^t Region, 31 clients; South Region, 40 clients; and West Region,
25 clients.
In the alphanumeric/symbolic mode (cell 1), I can select a type-
face with or without serifs, change the thickness (bold, l^ht) or
posture (e^. , italics) of the typeface, choose upper or lower case,
and use symbols (&, $, #) in place of words; in the spatial mode
(cell 2), enlarge or shrink the type size, choose the linear spacing
between characters and words (12 pitch; 10 CPI; ti^ht or loose
spacing, kerning); and in the graphic mode (cell 3) use marks to
regulate the linear flow of text (commas, periods, colons) or to
emphasize key ideas (underlining), or transform letters into ex-
pressive icons. Intra-textual choices, therefore, r^ul at e thousands
of points on a plane: given only the common variations in each
coding mode cited above, the possible permutations of our sample
easily number in the thousands! Intra-textual choices can be local
(a word in boldface) or global (a type size for an entire text), but
are essentially a one-dimensional, point-by-point regulation of
text.
The Inter-textual level (cells 4-6) is coded primarily in the spa-
tial mode, generating visual cues that enable readers to search for
and retrieve information. Distributed horizontally and vertically
on an invisible two-dimensional grid, a seamless text can be
transformed into a highly variegated system "surfacing" (Herr-
VISUAL LANGUAGE • KOSTELNICK
33
Intra
Inter
Extra
Supra
A 12-CelI Schema of Visual Communication:
Four Levels and Three Coding Modes
Alphanumeric/
Symbolic
1
micro-level textual
form: style, size,
weight, and posture
of letters, numbers,
and symbols
4
serial and segment-
ing devices:
headings, letters,
numbers; typestyle
variations showing
textual structure
7
decoding devices:
legends, captions,
labels, numerical
description of data
10
macro-level serial
and segmenting
devices: section
titles, numbers;
page headers,
pagination
Spatial
2
local spacing
between characters
and textual particles:
CPl, picas, kerning
5
vertical/horizontal
arrangement of text:
line endings,
indentation; lists,
matrices, trees,
flow diagrams
8
configuration of
schematic and picto-
rial sign systems;
plotting of data on
X-Y axes, circles,
other forms; depth
of pictorial images
11
cohesion of entire
document over
several planes:
page breaks, size;
location of extra-
textuals within text
Graphic
3
marks: punctuation,
underscoring;
iconicity of letters
and words
6
cueing devices:
bullets, icons;
syntactic devices:
line work and
arrows on tables,
charts, diagrams
9
schematic: lines,
tones, colors, and
textures on data
displays; pictorial:
resolution of details
on images, abstract
to realistic
12
coding marks
unifying pages or
sections of text:
line work, color,
icons, logos, tabs
34 THE JOURNAL OF BUSINESS COMMUNICATION ' 25:3:SUMMEE 1988
Strom, 1984, p. 229) textual structure.^ In our sample, of course,
some inter-textual decisions had already been made: the move-
ment of the text linearly to the r^ht, the line break and continua-
tion below and to the left, the distance between lines. We can
introduce numerous other inter-textual elements to reshape the
text:
5. During the past month, our clientele in each region has increased as follows:
• East Region, 31 clients
• South Region, 40 clients
• West Region, 25 clients
In the alphanumeric/symbolic mode (cell 4), " 5. " cues us that the
statement is part of a series; headings, roman numerals, initial let-
ters, and symbols are also serial devices. Spatially (cell 5), the
second half of the text is segmented into units and aligned verti-
cally as a list. Graphically (cell 6), bullets identify each item in the
secondary series. In more dynamic inter-textual configurations-
tables, matrices, flow charts, decision trees—spatial coding (the
positioning of words and figures) creates a visual syntax, and
graphic cues (arrows, lines, geometric shapes) form a system of
"macro-punctuation" (Waller, 1980), among textual particles.
The division between the inter- and extra-textual levels is sub-
tle but perceptually and semantically critical. The two examples
below can illustrate the distinction:
PLAN
Except for the spatial positioning of "PLAN," both messages use
similar visual language. But their meanings and visual codes differ
radically. On the left, the textual particle "PLAN" (cell 5) is
circumscribed graphically (cell 6), indicating one of perhaps several
functions in a decision-making process. Framed and embellished,
the word is the thing-like center of attention, like the word STOP
on a road sign. The same applies to textual particles in tables,
matrices, flow charts, decision trees, and so on: all are inter-textual
arrangements, simply at a higher level of intensity than a seamless
essay. In the example on the r^ht, however, the textual particle is
merely a tag (cell 7) identifying an image (a plan of an office, a
desk, a post card) coded spatially (cell 8: top view, proportional to
VISUAL LANGUAGE • KOSTELNICK
35
the real thing) and graphically (cell 9: outlining the perimeter).
Alternatively, the image may be nonrealistic, representing a quan-
tity, say the dollar amount or time in hours of generating a plan
(as opposed to other costs). In either case, textual elements assist
the decoding of the message (legends, labels, captions), but infor-
mation is processed primarily through seeing rather than reading.
In the right-hand example, "PLAN" is useful for decoding the
message but secondary to the visual image, which stands outside
the text.
The Extra-textual level (cells 7-9), therefore, may rely on the
alphanumeric/symbolic mode for interpretation, but coding is
primarily spatial and graphic. Sign systems range in level of ab-
straction from pictorial (low level of abstraction—photographs,
renderings) to schematic (high level of abstraction—logos, graphs,
diagrams; see Twyman 1979). Below, the data in our sample mes-
sage are displayed schematically:
During tiiepast month, our clientele in each region has increased as foUows:
East South West
Alphanumeric signs (cell 7) encode the X axis with geographic
locations, the Y axis with figures; spatially (cell 8), plotting the
data on the grid determines the vertical distance between figures
and the horizontal distance between regions; graphically (cell 9),
shaded bars encode the data, lines locate the axes, and tick marks
gauge the scale. All of the coding modes are open to variation.
"East" can be simplified to "E," and figures on the Y axis re-
placed by exact values above each bar. Spatially, the distances
between values and between regions can be increased or decreased,
or the axis shifted 90 degrees so the bars run horizontally. Graphi-
cally, the bars can be cross-hatched, colored, or changed to simple
lines, and the tick marks replaced by a light grid.
If the extra-textual were pictorial rather than schematic, the
same variables would come into play: spatially, the image of the
employees at a conference table can be close or distant or rotated
36 THE JOURNAL OF BUSINESS COMMUNICATION • 25.3:SUMMER 1988
at different angles from the real thing; graphically, it can be coded
at a h ^ level of realism (with plenty of surface details, down to
the wood grain on the table) or reduced to an abstract stick draw-
ing or a stylized logo. The coding choices for either schematics or
pictorials are interdependent: the graphic coding of a logo may
flatten an image spatially; the spatial coding of data on a bar chart
may restrict the range of graphic choices. Like the intra- and inter-
textual levels, visual language at the extra-textual level depends on
the contextual variables of the system.
The Supra-textual level (cells 10-12), like the inter-textual,
operates chiefly in the spatial mode, though on a more global
scale, arranging extra-textuals (graphs, pie charts, pictures) within
the text and regulating the flow from one page or section (para-
graph breaks at the bottom of the page; breaks between parts of a
report) to the next to create a coherent document. For example,
in the last version of our sample, the spatial mode (ceU 11) locates
the extra-textual relative to the text ("During the past month . . ."):
centered directly beneath the text, or to the left or to the right; or
isolated on a separate page in close proximity or in an appendix.
In the alphanumeric mode (cell 10), supra-textual elements include
segmenting and serial devices (page headers, section titles, pagina-
tion); for example, I could label the bar graph in the sample "Fig-
ure 6," indicating a supra-textual relation with other extra-textuals
in the document. In the graphic mode cohesive devices—line work,
page color, logos—unify sections of a document, providing visual
markers for readers. In our exsimple, a line border around the bar
chart would further demarcate the transition from text to extra-
text. By mapping out the terrain, the supra-textual level controls
the three-dimensional configuration of the document, enabling
readers to comprehend its logic and hierarchy and to retrieve
essential information.
Each document blends the coding modes and levels into an idio-
syncratic visual system incorporating most of the cells on the
matrix. Consider the visual choices for a simple analytical memo
report. At the intra-textual level, typefaces, point sizes (cell 1),
spacing (cell 2), and graphic cues (cell 3), selected from a menu
of computer-generated options, create interest and emphasis and
regulate the intra-linear flow of signs. At the inter-textual level,
headings (cell 4), indentation, and vertical spacing (cell 5) struc-
ture the text, and graphic coding around textual particles (cell 6)
forms tables and matrices. At the extra-textual level, spatial coding
distributes data (monthly sales f^res, production costs) on a grid
VISUAL LANGUAGE • KOSTELNICK 37
(cell 8), graphic coding records data with shaded bars (cell 9), and
alphanumeric coding (cell 7) explains each variable. At the supra-
textual level, pages are numbered and labeled with subject, date,
author, and number (cell 10), composed spatially to locate the bar
charts within the text and to ensure continuous flow from page to
page (cell 11), and coded with icons, colors, tabs, or separate pages
(blank or titled) to differentiate sections of the report (cell 12).
Individually, none of these choices may be novel, but when com-
bined in the same document they compose a system of visual lan-
guage dependent, as all visual images are, on context for their per-
ceptual qualities (Amheim, 1969). Hence, the levels and coding
modes are interdependent. The cell 1 selection of a type size, pos-
ture, and style expands or contracts the document and therefore
has supra-textual consequences. Conversely, a cell 11 decision to
write a "one-page resume" impinges on intra- and inter-textued
choices. The cell 4 decision to deploy a variety of headings may re-
quire cell 1 assistance to establish a hierarchy. The spatial and
graphic coding of a bar chart (cells 8, 9) influence supra-textual
page design. Choices for each coding mode and level, therefore,
are contingent upon other choices.
The 12-cell matrix opens up these choices to public purview,
enabling us to describe how visual elements coedesce to form co-
herent systems. Just as each writing situation engenders a different
combination of linguistic choices, so too with visual language:
each document requires conscious planning so that the modes and
levels work harmoniously to achieve the same goals. To the extent
that the cells on the matrix compose an integrated system that
contributes to the purpose of the message, the document obtains
a certain level of visual sophistication or intensity.
INFORMATION DESIGN: TOWARD A
FUNCTIONAL VISUAL LANGUAGE
Of course not all visual systems succeed in accommodating the
reader. What makes visual language reader-oriented? In a fast-paced
business environment, effective visual display simplifies the reader's
infonnation processing. As Hartley (1985) puts it, "The printed
page should provide a reliable frame of reference from within
which the readers can move about, leave and return without con-
fusion" (p. 18). Above all else, visual language should be function-
al. And on an idiomatic and conventional level, it usually is: words
38 THE JOURNAL OF BUSINESS COMMUNICATION ' 25:3:SUMMER 1988
ordinarily flow horizontally across the page; letters have signature
blocks; bars on graphs stand for measured units. These generic
conventions enable the reader to place visual codes into the slots
of menteil schemata (see Schumacher, 1981). A one-page document
that follows the conventions of a letter and arrives in an envelope
with a booklet "fits" nicely into the slot of "cover letter," a per-
suasive message, widely distributed, that entices the reader to
examine an accompanying document. On the other hand, a letter
set in boustrophedon (as the ox plows, back and forth), or hori-
zontally across the length of the page, poses an obvious "misfit"
to most readers. By activating schemata, generic conventions foster
reader expectations about the whole message.
Visual conventions, however, define only general patterns,
change over time and from one discourse community or culture to
another, and may not provide the optimum visual patterns for
information processing. To compensate for these shortcomings,
the "information design" movement has tried to establish more
stable, culturally neutral, and empirically verifiable guidelines.
Combining theory, practice, and research, information design
draws upon several disciplines—graphic design, perception, educa-
tional and cognitive psychology, and semiotics—to discover reader-
oriented visual patterns for many of the levels and coding modes
on the matrix.
From the outset, the information design movement soi^ht to
create a modem, rational, ideologically free system of signs acces-
sible to a mass audience. Its intellectual roots can be traced to the
functional modernism of the Bauhaus in the 2O's (Kinross, 1985)
and indirectly to the visual dynamism of Futurist and Constructi-
vist text displays. Each of these movements experimented with
radically new modes of visual design, emphasizing clean, nontradi-
tional forms and typefaces compatible with Machine Age technol-
ogy (see Crouwel, 1979). Together these fresh approaches to
textual display and image-making provided an aesthetic and func-
tional springboard for early information designers. For example,
Neurath's "isotype" (international system of picture education)
symbols epitomized the quest for a rational, cross-cultural language
of extra-textual s ^ s for displaying quantitative data. After World
War II the search for simple, universal, and technically precise
forms continued in the invention of typeface systems, for example
the serifless Univers style (Kinross, pp. 23 & 27). On a supra-
textual level, the appeal to the universal was reinforced by the
laws of gestalt psychology (similarity, good figure, equilibrium).
V I S U A L L A N G U A G E • K O S T E L N I C K . , _ 3 9
which transcended style and culture by codifying perceptual ex-
periences detached from language or ideology.
The information des^n movement, therefore, has evolved into a
modem philosophy of visual display emphasizing objectivity, clari-
ty, and the integration of text with simple forms. The movement
has had a wide impact on advertising, govemment documents and
forms, educational texts, and scientific and technical publications.
Today theorists and practitioners continue to develop and test
methods for functional designs. On the inter-textual level. Hartley
(1985), Wright (1977), and several others have devised methods of
employing headings, vertical spacing, and graphic cues to surface
discourse visually, while Hom (1982) has codified these elements
into a comprehensive system of "structured writing." On the extra-
textual level, Tufte (1983) and Bertin (1981) apply functionalist
principles to the visual display of quantitative data. Both advocate
economy through density, arguing that graphs and charts should
display the maximum information in the smallest space with the
least ink. Bertin invokes semiotic theory to define visual clarity:
graphic sign systems should be "monosemic" (pp. 177-179), the
relations among s ^ s conveying a single, unambiguous meaning.
Consistent with its functionalist tradition, contemporary informa-
tion design theory values economy, objectivity, and semantic and
structural transparency.
Further supporting the functionalist approach are a wealth of
empirical studies aimed at identifying optimum visual displays. At
the intra-textual level. Tinker (1963) found that type sizes be-
tween 9 and 12 point maximize readii^ efficiency, lower case
letters are more readable than upper case, and excessive italic print
reduces legibility. On the inter-textual level. Hartley (1984) has
studied, with mixed results, the effects of horizontal and vertical
structuring of text on readability, while Hartley and Trueman
(1985) have shown, through extensive experiments, that headings
improve reader performance. Wright (1968) experimented with the
spatial configurations of tables, finding that users of currency
conversion tables perform better with redundant lists than with
matrices. On the extra-textual level, Macdonald-Ross (1977)
summarized over five decades of empirical research on the graphic
display of quantitative data, concluding for example that horizon-
tal bars are superior to unkeyed circles. Psychologists have studied
the effects of spatial depth and degree of graphic detail in the
processing of representative images (Perkins, 1980). They found
that readers need a modest amount of acculturation, should be
40 THE JOURNAL OF BUSINESS COMMUNICATION ' 25.3.SUMMER 1988
given only selected information pertinent to the task, and perceive
depth on a two-dimensional surface differently from the "real
thing." At the supra-textual level, guidelines can be derived from
hemispheric brain research (see, for example, Welford, 1984).
Because readers process visual information more effectively in
the right brain (the left field of vision), pictures and schematics
are generally better placed to the left, explanatory text to the
right.^
Many of the empirical findings jind theoretical positions cited
above presume a universal reader performing generic, noncon-
textual tasks. Wright (1977) sums up the empiricist's dilemma:
"Researchers often consider themselves free to ask if format A is
better than format B, without needing to specify better for what?"
(p. 112). Some of the findings—in particular, those of Hartley and
Wright—are user-centered and tied to specific audiences and con-
texts. Wright's experiments with tabular displays are bound to the
readers' experience with visual arrangements of figures, the situa-
tion in which the display is used, and the readers' familiarity with
the subject. M£iny case studies of public documents are con-
textually sensitive to the special needs posed by audience and
situation. Such studies entail pretesting and interviewing during
the design process to assure quality control and to convert visual
language to a "plain English" style intell^ible to the nonspecialist
(Redish, Felker, and Rose, 1981; Wright, 1979).
The functional guidelines of the information design movement,
therefore, specify which configurations afford the most efficient
processing of information, enabling us to engineer clear visual
conduits. The emphasis on information privileges function over
decoration, just as the Bauhaus did. A well-designed text repli-
cates on the page the sleek steel and glass of a modem h^h-rise:
functional, aesthetically austere, and culturally neutral. By meld-
ing text and form, the information design movement revolution-
ized the visual language of texts and made design a bona fide part
of the creative process, from planning to production.
THE MYTH OF RHETORICAL NEUTRALITY
Ideally, information des^n deploys visual language to ensure
an unadulterated channel between sender and receiver. However,
visual elements themselves—especially in business messages—obtain
meaning beyond the simple transfer of information. "' Pure' inf or-
VISUAL LANGUAGE " KOSTELNICK 41
mation," asserts Bonsiepe (1965), "exists for the designer only in
arid abstraction. As soon as he b^ins to give it concrete shape, to
bring it within the range of experience, the process of rhetorical
infiltration b^ins" (p. 30). Simply put, a "rhetoric of neutrality"
is a "pipe-dream," for even the mimdane visual language of a train
timetable has a certain rhetorical force, intended or not, that af-
fects the reader's judgment (Kinross).
This rhetorical "infiltration" stems partly from taste and cul-
tural values. Futurist and Constructivist text displays, the fore-
runners of infonnation design, were considered iconoclastic when
they first appeared; today they seem modem but pass^. Then as
now, they relinquish any claims to neutrality. The Machine Age
modemism of the information design movement also constitutes a
style—streamlined, simple, and low-key—but also subject to the
vicissitudes of intellectual and aesthetic favor, and hence anything
but "objective."* The invention and evolution of styles continually
transform taste and perception, in art and in business communica-
tion. An early corporate logo, letterhead, or trademark juxtaposed
to its contemporary version demonstrates the visual distance be-
tween past and present, and presages future recreation.
Technology also influences the rhetoric of visual langus^e. At
the outset of the century most texts were handwritten in one of
four styles: business, civil service, text, or legal. The appearance of
the typewriter transformed "micro-level" (Waller, 1980) visual
choices, forcing characters into uniform, equally spaced units on a
grid and initiating a whole new set of intra- and inter-textual con-
ventions (Walker, 1984). Ninety years ago a letter done in business
script would have signaled the nature of the message but otherwise
attracted little attention. Today it would totally alter the rhetoric
of the communication: the writer would suffer (an entry-level
application letter) or benefit (a letter from a CEO to a consumer)
from the unorthodox visual coding.
Cultural and technological changes are long-term, evolutionary
tendencies that gradually modify the shared perception of large
discourse communities. The choices business communicators make
everyday are highly contextual and vary from one audience and
visual system to another. Information design can weed out percep-
tually poor choices—illegible type styles, line spacing, and data dis-
plays—but cannot necessarily prescribe rhetorically appropriate
choices goveming the arrangement, style, or tone of a particular
document.
For example. Tinker found that many common typefaces were
42 THE JOURNAL OF BUSINESS COMMUNICATION ' 25:3:SUMMER 1988
equally legible but that readers preferred some faces over others.
The "atmosphere value" or "semantic quality" of typefaces (Bar-
tram, 1982) opens the door to subjectivity and rhetorical judg-
ment.* What "pleases" readers in one message may not be pleasing
or rhetorically consistent in another. Visual choices demand adap-
tation to the particular audience, purpose, and circumstances of
the document. A simple intra-textual (cell 1) choice of a typeface
can have a major impact on the rhetoric of a messf^e. For an in-
formal message to a co-worker, a handwritten note or a dot-matrix
printout reinforces the familiar tone and transience of the docu-
ment. However, a manager (one that values job security) submitting
an annual report to a board of directors or a proposal to a govem-
ment s^ency will specify a typeface that projects credibility and
permanence.
Although technical capability helps, visual rhetoric is not neces-
sarily constrained by available technology. Direct mail sales let-
ters use typewritten serif styles even though virtually any typeface
could be used, given the bulk printing. Indeed, to make messages
appear more spontaneous, senders frequently underline key points
with insouciant strokes of the pen (in red ink) to personalize the
visual style. Urgent mess£^es (6.g., collections) are customarily
typed entirely in upper case, sacrificing legibility for a more
official, authoritarian tone. These choices may seem trivial and
subjective, but they are made for rhetorically sound reasons. Add
to these variations the ability to highlight text with boldface and
multiple fonts and point sizes now commonly available in desk-top
printers, the business communicator will experience far greater
freedom to adapt intra-textual tone and style to the rhetorical task.
The same holds for information design on any level of the matrix.
Inter-textual structuring—headings, vertical spacing, graphic cues-
has rhetorical consequences, surfacing certain aspects of the mes-
sage while embedding others. Opening up a text with spatial and
graphic coding is equivalent to persuading visually, impelling read-
ers to value selected pieces of information and to acquiesce to logi-
cal and hierarchical connections that make them cohere visually.
In the second version of our sample text, the list with bullets estab-
lishes visual parallelism among the three regions. Perhaps the West
R^ion is new and the writer wants to emphasize its status as a
full and equal sector. Perhaps isolating these elements ai ^es that
the vice-president for sales is doing satisfactorily in all regions, or
that by comparison the West Region is faltering, or that more
employees should be hired in the South. Depending on the con-
VISUAL LANGUAGE • KOSTELNICK 43
text, selective textual surfacing advances the writer's ai l ment .
In an extremely sensitive message (a perfonnance appraisal), the
writer may even eschew any sort of inter-textual variation to
embed key information, sacrificing lability to maintain rhetori-
cal integrity.
The third version of our sample detaches the data from the
text, reshapes them graphically, and tells the reader in a resonant
voice: "Here are the facts, bare and hrefutable!" The visual style,
bordering on hyperbole, can of course be modified. Spatially, the
bars can be extended or contracted on the grid, the graphic codii^
ornamented or streamlined, the alphanimieric codii^ expanded—
depending on the designer's argument and rhetorical stance. The
differences among regions can be accentuated or diminished, the
graph made to appear scientific or aesthetically pleasing, clumsy or
self-serving. Any of these effects influences the reader's impression
of the data as well as of the author's integrity and is far from neu-
tral. Even acquiescing to someone else's configuration of the data
is still to choose, and the reader to receive, the visual rhetoric of
the message.
Together these choices coalesce to drive the visual rhetoric of
the document. Consider the range of visual decisions that go into
an annual report to stockholders: the company logo, charts and
graphs to display sales and production figures; apian and line draw-
ing of a new plant; photographs introducing board members; and
textual displays of expenses, revenues, and earnings. Aloi^ with
these are numerous supra-textual decisions regarding the location
of extra-textuals, the size and number of pages, and graphic cues
such as color and line work intended to coordinate the whole
document, add interest to complex data, instill confidence in in-
vestors, and accurately project the company's image (innovative,
consumer oriented, h^h-tech). The visual system is both informa-
tional and persuasive, enabling investors to locate pertinent facts
and, more importantly, enabling the company to market its per-
formance. Although annual reports are more complex than every-
day messages, maint{iining full rhetorical control over any business
document requires careful visual planning. In other words, the
writer must attend to the visual system during the composing
process as rhetorical strategies take shape.
44 THE JOURNAL OF BUSINESS COMMUNICATION • 25:3:SUMMER 1988
COORDINATING VISUAL AND VERBAL RHETORIC:
A HOLISTIC PROCESS
Planning an entire document is a "holistic process" (Barton &
Barton, 1985) in which the visual and the verbal modify, comple-
ment, and compensate for one another to accomplish the rhetori-
cal goals of the communication. Linguistic invention entails in-
numerable lexicEil choices; visual invention materializes from the
wealth of elements defined by the matrix. Both are driven by the
same purpose and by many of the same principles and concepts:
style, tone, level of technicality, interest, benefits, arrangement,
readability, and so on. During the planning stage of a document -
say a proposal persuading a skeptical reader to act on a recom-
mendation-Hhe visual rhetorician considers several options:
• Stimulate reader interest through a logo, picture, or data
display (cells 8, 9) and through highly variegated inter-textual
stmcture (cells 4, 5, 6).
• Surface selected arguments and benefits through intra- and
inter-textual variations (cell 1: boldfacing and all capitals; cell 3:
underlining; cell 6: bullets to highlight items in a list).
• Adapt the style of data displays (cells 8-9) to a nonspecialist
(cell 7 alphanumeric coding), and project a serious tone necessary
to engage this reader (cell 1 choice of a typeface; cells 6, 9, 12
graphic coding).
• Map out the document with inter- and supra-textual cues
(cells 4-6, 10-12) that guide the busy reader through the informa-
tion, increasing readability.
Because these and many other strategies apply to visual as well
as verbal problem-solving, early planning synchronizes the two
from the start.
Early planning also exploits the special rhetorical capabilities of
the visual and the verbeil. For example, a bar graph can often argue
with statistics more clearly and emphatically than inter-textual dis-
plays of figures; a logo can establish the identity of an oi^Jiniza-
tion or a project more succinctly and forcefully than an explana-
tory paragraph. On the other hand, in a sales letter a poetic
description of the scenery at a resort may be more provocative
than a high resolution photograph, and in a n^ative recommenda-
tion letter a seamless text more tactful than one that is visually
transparent.
Because the visual and verbal are interdependent, holistic plan-
ning enables the writer-designer to monitor the interaction between
VISUAL LANGUAGE • KOSTELNICK 45
the two. A highly vari^ated inter-textual system chaises sentence
structure and syntax, increasing the frequency of parallelism and
fragments and eliminating the need for verbal transitions (Bem-
hardt, 1986). Converting text-bound figures to a data display (as
in the third version of our example) contracts and reshapes tiie
text. On the other hand, a complex line graph may require textual
expansion to hi ghl i t and explain key data. Style and tone are
also interactive: a set of informally written instructions on how
to pot plants calls for simple line drawings, instead of photo-
graphs, to illustrate the task; the linguistic tone of a sales letter
theme or slogan sets the visual tone for images on the envelope
and enclosures.
Globally, supra-textual decisions can regulate the linguistic evo-
lution of the text. Thing-like, the page is a tablet, the writer's unit
of workspace, like an artist's canvas, an architect's story, or a
sculptor's marble slab. The tablet as visual workspace both compli-
cates and simplifies the writer's task. Some undergraduates strug-
gle to find enough material to fill up a "one-page" r4sum^, whUe
others try desperately to conflate several pages into one. As items
are arranged on the tablet, the writer strikes a balance between the
functional exigencies of information design (the user may give the
document only a cursory glance) and an appropriate rhetorical
stance—ihe need to appear trustworthy, ambitious, successful, and
employable. To these ends, elements are moved about, shaped, cut,
and recast linguistically (in phrases and fragments). Through the
symbiotic growth of the visual and the verbal, the document
evolves within the supra-textual boundaries of the tablet.
Visual design, like writing, is a process of discovery. This dis-
covery process occurs within and outside the text. Articulating
a text with intra- and inter-textual cues-H;ypographical variations,
paragraph breaks, headings, lists, tables—unfolds and structures
ideas, enabling the writer to resee the logic of the mess^e. Creat-
ing visual clarity demands clear thinking, and rethinking, about
content (Hartley, 1984, p. 502). Extra-textuals also act as dis-
covery tools (Bertin, 1981). The process of translating data into a
visual sign system surfaces patterns and relationships submerged in
a table of figures. This happens rudimentarily in the third version
of our example: the designer must experiment with graphic aad
spatied coding to find a rhetorically suitable configuration. If we
had additional figures—about client increases in previous periods,
the clients themselves (individuals, businesses, govemment agen-
cies), or dollar volumes per client or region—permuting these data
46 THE JOURNAL OF BUSINESS COMMUNICATION • 25:3:SUMMER 1988
on the X-Y axes would enable us to find an insightful pattem.
Supple and unpredictable, data displays are powerful discovery
tools, much like freewriting, zero drafting, and outlining.
At the outset of the holistic process, therefore, visual thinking
may be provisional: designing a document is an evolutionary
process entailing invention, revision, and editing. Visual language
is rich and flexible, allowing infinite syntactic variation of the
constituent elements outlined on the matrix. However, if these
elements eire to be transformed into inventive patterns, they must
be treated as bona fide rhetorical tools guided by the same strate-
gies that govern the linguistic text. Asserting complete rhetoricsd
control over a document, therefore, demands visual thinking from
planning to production.
CONCLUSION
Business communication is intensely visual: as practitioners,
researchers, and instructors we are inundated with an array of visi-
ble forms. Although we use, teach, and presumably value effective
visual language, we have no system for describii^, analyzing, or
evaluating it. This is partially because d e s ^ is viewed as a ready-
made set of conventions or as a task performed by specialists—
in either case extrinsic to the writing process. However, in the
making of business documents, writing and designing, reading and
seeing, are interdependent, complementing and impinging on each
other to satisfy the goals of the communication. Combined in any
business document, the three coding modes and four levels of
visual language comprise an integrated system enablii^ the reader
to process information efficiently from the rhetorical perspective
the writer intends. Adapting the visual system to the message is
essential in business communication because audience and purpose
vary from one message to another and because writers have the
autonomy to design their own documents.
Successful adaptation requires attention to design throughout
the writing process. We have only rudimentary knowledge about
how visual and verbal thinking interact durii^ this process or how
the contextual variables of visual language affect readers. What we
do know, instinctively though perhaps unconsciously as members
of a "print culture," is that visual thinking is intrinsic to the dis-
covery and communication of meaning. To use visual language
fluently, or to understand its stmcture and nuances, we must ex-
VISUAL LANGUAGE • KOSTELNICK 47
plore it explicitly and systematically, with the same rigorous scru-
tiny given the linguistic text.
NOTES
' Twyman (1979) has proposed a graphic schema containing four modes
of symbolization and seven methods of configuration. His matrix contains
twenty-eight "cells" and classifies primarily spatial variations at the inter-
and extra-textual levels.
^ Some of Twyman's methods of configuration are particularly rele-
vant here: Pure linear (a continuous flowing text; e.g., in circular form).
Linear interrupted (conventional text with line breaks). List, Matrix (tables),
and Linear branching (a tree, an organizational chart).
^ Design guidelines derived from hemispheric research are contingent
upon at least two variables: the kind of information processing required by
the task (part-by-part vs. holistic analysis) and the amount of verbal thinking
the viewer is doing at the time.
^ See Kinross for a detailed account of the aesthetic and ideological
evolution of the information design movement and its claims to rhetorical
"neutrality."
^ The "atmosphere value" of typefiaces has long been the subject of
study and speculation. For example, Ovink experimented with several type-
&ces used in literature and advertising, grouping their subjective effects under
categories such as "luxury/refinement," "economy/precision," and "strength."
But the moods and cultural values associated with intra-textual styles can
change dramatically: serifless letter forms, which are typically allied with
modernism, were initially revived in the eighteenth century because they
evoked "rugged antiquity" (Spencer, 1969, pp. 29-30).
REFERENCES
Arnheim, Rudolf (1969). Visual thinking. Berkeley, CA: tJniversity of Cali-
fornia Press.
Barton, Ben F. & Barton, Marthalee S. (1985). Toward a rhetoric of visuals
for the computer era. The Technical Writing Teacher, 12,126-145.
Bartram, David (1982). The perception of semantic quality in type: IHffer-
ences between designers and nondesigners. Information Design Joumal,
3, 38-50.
Bemhardt, Stephen A. (1986). Seeing the text. College Composition and
Communication, 37,66-78.
Bertin, Jacques (1981). Graphics and graphic information-processing (William
J. Berg and Paul Scott, Trans.). New York: De Gruyter.
Bonsiepe, Gui (1965). Visual/verbal rhetoric. Vim, 14-16, 23-40.
Crouwel, Wim (1979). Typography: A technique of making a text 'legible.'
In Paul Kolers, Merald E. Wrolstad, and Herman Bouma (Eds.), Proces-
sing of Visible Language, {1, pp. 151-164). New York: Plenum.
Gelb, I. J. (1980). Principles of writing systems within the frame of visual
communication. In Kolers, Wrolstad, and Bouma {2: pp. 7-24).
48 THE JOURNAL OF BUSINESS COMMUNICATION ' 25:3:SUMMER 1988
Hartley, James (1985). Designing instructional text (2nd ed.). New York:
Nichols.
Hartley, James (1984). Space and structure in instructional text. In Ronald
Easterby and Harm Zwaga (Eds.), Infonnation design (pp. 497-515). New
York: Wiley.
Hartley, James & Trueman, Mark (1985). A research strategy for text design-
ers: The role of headings. Instructional Science, 14,99-155.
Herrstrom, David Sten (1984). Technical writing as mapping description onto
diagram: The graphic paradigms of explanation. Journal of Technical
Writing and Communication, 14,223-240.
Horn, Robert E. (1982). Structured writing and text design. In David H.
Jonassen (Ed.), The technology of text (pp. 341-365). Englewood Qiffs,
NJ: Educational Technology Publications.
Kinross, Robin (1985). The rhetoric of neutrality. Design Issues, 2(2), 18-30.
Macdonald-Ross, Michael (1977). How numbers are shown: A review of re-
search on the presentation of quantitative data in texts. AV Communica-
tion Review, 25, 359-409.
Ong, Walter (1982). Orality and literacy: The technologizing of the word.
New York: Methuen.
Perkins, D. N. (1980). Pictures and the real thing. In Kolers, Wrolstad, and
Bouma (2, pp. 259-278).
Redish, Janice C, Felker, Daniel B., & Rose, Andrew M. (1981). Evaluating
the effects of document design principles. Information Design Journal, 2,
236-243.
Schumacher, Gary M. (1981). Schemata in text processing and design. Infor-
mation Design Journal, 2,17-27.
Spencer, Herbert (1969). The visible word (2nd ed.). New York: Hastings.
Tinker, Miles (1963). Legibility of print. Ames, IA: Iowa State University
Press.
Tufte, Edward (1983). The visual display of quantitative information. Che-
shire, CT: Graphics Press.
Twyman, Michael (1979). A schema for the study of graphic language. In
Kolers, Wrolstad, and Bouma (i , pp. 117-150).
Walker, Sue (1984). How typewriters changed correspondence: An analysis
of prescription and practice. Visible Language, 18,102-117.
Waller, Robert H. W. (1980). Graphic aspects of complex texts: Typography
as macro-punctuation. In Kolers, Wrolstad, and Bouma (2, pp. 241-253).
Welford, A. T. (1984). Theory and application in visual displays. In Easterby
and Zwaga (pp. 3-18).
Wright, Patricia (1979). The quality control of document design. Information
Design Journal, 1, 33-42.
Wright, Patricia (1977). Presenting technical information: A survey of re-
search findings. Instructional Science, 6, 93-134.
Wright, Patricia (1968). Using tabulated information. £;:?o«omfcs, 11, 331-343.
doc_944160055.pdf
approaches to graphics and text design are prescriptive and unsystematic.
A 12-cell schema of visual coding modes and levels provides a model for
describing and evaluating business documents as flexible systems of visual
lar^uage. Emphasizing clarity and objectivity, the "information design"
movement has generated guidelines for creating functional visual displays.
However, visual language in business communication is seldom rhetorically
"neutral" and requires adc^tation to the contextual variables of each
document, a goal the writer can achieve by combining visual and verbal
planning in the same holistic process.
A Systematic Approach to Visual
Language in Business Communication
Charles Kostelnick
Iowa State University
WRITING DOES NOT COMMUNICATE unta it is seen, until
it becomes an artifact for visual inspection (Gelb, 1980). From
the primitive making of lists, business commimicators have
created permanent records by transcribing symbols onto two-
dimensional surfaces. With the proliferation of print technology
formerly reserved for specialists, contemporary business com-
municators have unprecedented control over visual design.
However, participation in a visually intensive "print culture"
(Ong, 1982) does not guarantee visual sensibility. While the
"technologizing of the word" opens new avenues for visual
expression—everything from page design to bar charts to
scanned images—ihe thing-like quality of texts and the per-
ceptual reliance on the eye for information processing are
largely unacknowledged and imexamined. In theory and
application, visual language in business communication re-
mains rudimentary and prescriptive, and confined to a super-
ficial or decorative rather than a functional or rhetorical role
in the communication process.
To compensate, we need a systematic approach for analyzing
how visual elements affect the readability and the rhetoric of
business documents, and even for describing the visual elements
themselves. Terms such as "layout," "format," and "graphics"
lack precision and connote low level skills rather than serious
visual thinking. The visual language of business communication
is a flexible system of symbols, marks, and spatial variations
29
30 THE JOURNAL OF BUSINESS COMMUNICATION ' 25:3:SUMMER 1988
that operates on several levels within and outside the text. Used
effectively, visual langus^e enables readers to process informa-
tion—in letters, brochures, reports, newsletters—clearly and effi-
ciently. The "infonnation design" movement has generated
guidelines for creating functional, user-oriented documents. How-
ever, far from being generic or rhetorically "neutral" (Kinross,
1985), visual choices are bound to a perceptual and rhetorical
context. Each document embodies an autonomous system re-
quiring adaptation to a particular audience and purpose. Visual
planning and invention, therefore, must be part of the global
strategy of the document and thoroughly integrated into the
process.
THE SPECTRUM OF VISUAL INTENSITY
Writing and reading are intensely visual acts in which writer
and reader negotiate meaning on a two-dimensional field. Since
this field can be varied and manipulated into infinite configura-
tions, all documents are not visual in the same way or to the
same d^ree (Bemhardt, 1986). Thus we could devise a spectrum
of visual intensity ranging from very low to very
LOWVISUAL ^ ^ HIGH VISUAL
INTENSITY INTENSITY
What, then, are the limits of such a continuum? What character-
istics determine whether a document has a high or low level of
visual intensity?
At the low end we might place a conventional essay or report,
which submei^es explanations and arguments in continuous text,
or a novel with its symbols, metaphors, and complex narrative
strands that readers must extract from seamless blocks of text
(Bemhardt, 1986; Herrstrom, 1984). StUl, even a novel relies on
spatial segmentation—books, chapters, paragraphs—to signal tran-
sitions in place, time, and voice (Sterne even includes his own
diagrams). The low end of the continuum calls for something far
less visual: ideally, a string of unembellished signs flowing in a
sii^e direction, unrestricted by margins or pages. Rendering the
text entirely in upi>er case and eliminating punctuation would
further reduce visual dependency. The reader would then have no
sense of spatial direction other than the syntactic relations of
linguistic signals. Like a message moving across a video screen in
VISUAL LANGUAGE • KOSTELNICK 31
an airport, the text would approximate the perceptual imperman-
ence of an utterance.
At the high end of the visual continuum, des ^i i ^ and seeing
assume a far greater role in shaping and receivir^ the messi^e. As
text is segmented into chimks, chimks into lists, and lists into
matrices, charts, and dis^rams, words become "thing-like" (Ong,
1982) entities transfixed spatially and graphically (through line
work, arrows, geometric forms) on the page. Even more thing-like
and reliant on visual inspection are representative images—for
example, a picture in a proposal of key employees seated aroimd a
conference table. Still, visual intensity does not necessarily depend
on realism because messages can be purely visual but abstract. For
example, to show the enrollment patterns of a college, I can use
pictures of each class or even of each student (achievii^ optimum
realism), but a bar chart or a line graph would display the trends
more clearly and concisely and with greater visual immediacy.
Thus on one end of the spectrum we might place a piurely (or as
pure as possible, given that it must be seen) linguistic messt^e, on
the other end a purely visual message which defies simple transla-
tion into words. However, the conditions for the extremes rarely
obtain: most business documents are visual hybrids. When trans-
ferred to the page, the linear messs^e flashing across the screen
undergoes a visual metamorphosis: words are shaped by a type
style, xmderlined, thickened, s^mented into lines, distributed
vertically and horizontally across the page, and circumscribed by
graphic marks. The picture of the employees at the conference
table, on tiie other hand, can be reduced to a line drawing, a plan
view of the office, a corporate logo, or a series of abstract marks,
each of which (including a color photo) may depend on linguistic
coding to make any sense. And so inside the extremes of the con-
tinuum occurs a vast interminglii^ of visual signs, and of the visual
and the verbal. Like a chemical reaction, the signs are interdepen-
dent: different combinations alter the visual effects of individual
signs, which collectively alter the meaning of the message. Because
these combinations are bound perceptually and rhetorically to
context, each document contains an idiosyncratic system of visual
language. The visual intensity of each system depends on the inter-
play of several levels and coding modes.
A SCHEMA OF VISUAL LANGUAGE
What, then, are the constituent elements of visual language?
First, we need to define the coding modes that comprise the raw
32 THE JOURNAL OF BUSINESS COMMUNICATION • 25:3:SUMMER 1988
materials of visual language; and second, the levels at which the
modes operate within actued documents.
Visual language can be encoded in three modes:
• Alphanumeric/symbolic: textual particles, including letters
that comprise linguistic units; numbers; and symbols such as
parentheses, dollar signs, and ampersands.
• Graphic: marks, lines, tones, and colors which encode geo-
metric shapes (squares, circles, bars), realistic im£^es (a face, a
building), or abstract forms (a corporate logo).
• Spatial: the distribution of alphanumeric/symbolic and
graphic signs across a plane, creating visual syntax among textual
or graphic particles.
These coding modes are integrated on four distinct levels: intra-,
inter-, extra-, and supra-textual. Together these four levels and
three modes comprise a 12-ceIl schema of visual language. '
The Intra-textual level (cells 1-3) is coded primarily in the
alphanumeric/symbolic mode and controls the local form, size,
posture, and embellishment of textual elements. Consider the fol-
lowing sample of text:
During the past month, our clientele in each region has increased as fol-
lows: E^t Region, 31 clients; South Region, 40 clients; and West Region,
25 clients.
In the alphanumeric/symbolic mode (cell 1), I can select a type-
face with or without serifs, change the thickness (bold, l^ht) or
posture (e^. , italics) of the typeface, choose upper or lower case,
and use symbols (&, $, #) in place of words; in the spatial mode
(cell 2), enlarge or shrink the type size, choose the linear spacing
between characters and words (12 pitch; 10 CPI; ti^ht or loose
spacing, kerning); and in the graphic mode (cell 3) use marks to
regulate the linear flow of text (commas, periods, colons) or to
emphasize key ideas (underlining), or transform letters into ex-
pressive icons. Intra-textual choices, therefore, r^ul at e thousands
of points on a plane: given only the common variations in each
coding mode cited above, the possible permutations of our sample
easily number in the thousands! Intra-textual choices can be local
(a word in boldface) or global (a type size for an entire text), but
are essentially a one-dimensional, point-by-point regulation of
text.
The Inter-textual level (cells 4-6) is coded primarily in the spa-
tial mode, generating visual cues that enable readers to search for
and retrieve information. Distributed horizontally and vertically
on an invisible two-dimensional grid, a seamless text can be
transformed into a highly variegated system "surfacing" (Herr-
VISUAL LANGUAGE • KOSTELNICK
33
Intra
Inter
Extra
Supra
A 12-CelI Schema of Visual Communication:
Four Levels and Three Coding Modes
Alphanumeric/
Symbolic
1
micro-level textual
form: style, size,
weight, and posture
of letters, numbers,
and symbols
4
serial and segment-
ing devices:
headings, letters,
numbers; typestyle
variations showing
textual structure
7
decoding devices:
legends, captions,
labels, numerical
description of data
10
macro-level serial
and segmenting
devices: section
titles, numbers;
page headers,
pagination
Spatial
2
local spacing
between characters
and textual particles:
CPl, picas, kerning
5
vertical/horizontal
arrangement of text:
line endings,
indentation; lists,
matrices, trees,
flow diagrams
8
configuration of
schematic and picto-
rial sign systems;
plotting of data on
X-Y axes, circles,
other forms; depth
of pictorial images
11
cohesion of entire
document over
several planes:
page breaks, size;
location of extra-
textuals within text
Graphic
3
marks: punctuation,
underscoring;
iconicity of letters
and words
6
cueing devices:
bullets, icons;
syntactic devices:
line work and
arrows on tables,
charts, diagrams
9
schematic: lines,
tones, colors, and
textures on data
displays; pictorial:
resolution of details
on images, abstract
to realistic
12
coding marks
unifying pages or
sections of text:
line work, color,
icons, logos, tabs
34 THE JOURNAL OF BUSINESS COMMUNICATION ' 25:3:SUMMEE 1988
Strom, 1984, p. 229) textual structure.^ In our sample, of course,
some inter-textual decisions had already been made: the move-
ment of the text linearly to the r^ht, the line break and continua-
tion below and to the left, the distance between lines. We can
introduce numerous other inter-textual elements to reshape the
text:
5. During the past month, our clientele in each region has increased as follows:
• East Region, 31 clients
• South Region, 40 clients
• West Region, 25 clients
In the alphanumeric/symbolic mode (cell 4), " 5. " cues us that the
statement is part of a series; headings, roman numerals, initial let-
ters, and symbols are also serial devices. Spatially (cell 5), the
second half of the text is segmented into units and aligned verti-
cally as a list. Graphically (cell 6), bullets identify each item in the
secondary series. In more dynamic inter-textual configurations-
tables, matrices, flow charts, decision trees—spatial coding (the
positioning of words and figures) creates a visual syntax, and
graphic cues (arrows, lines, geometric shapes) form a system of
"macro-punctuation" (Waller, 1980), among textual particles.
The division between the inter- and extra-textual levels is sub-
tle but perceptually and semantically critical. The two examples
below can illustrate the distinction:
PLAN
Except for the spatial positioning of "PLAN," both messages use
similar visual language. But their meanings and visual codes differ
radically. On the left, the textual particle "PLAN" (cell 5) is
circumscribed graphically (cell 6), indicating one of perhaps several
functions in a decision-making process. Framed and embellished,
the word is the thing-like center of attention, like the word STOP
on a road sign. The same applies to textual particles in tables,
matrices, flow charts, decision trees, and so on: all are inter-textual
arrangements, simply at a higher level of intensity than a seamless
essay. In the example on the r^ht, however, the textual particle is
merely a tag (cell 7) identifying an image (a plan of an office, a
desk, a post card) coded spatially (cell 8: top view, proportional to
VISUAL LANGUAGE • KOSTELNICK
35
the real thing) and graphically (cell 9: outlining the perimeter).
Alternatively, the image may be nonrealistic, representing a quan-
tity, say the dollar amount or time in hours of generating a plan
(as opposed to other costs). In either case, textual elements assist
the decoding of the message (legends, labels, captions), but infor-
mation is processed primarily through seeing rather than reading.
In the right-hand example, "PLAN" is useful for decoding the
message but secondary to the visual image, which stands outside
the text.
The Extra-textual level (cells 7-9), therefore, may rely on the
alphanumeric/symbolic mode for interpretation, but coding is
primarily spatial and graphic. Sign systems range in level of ab-
straction from pictorial (low level of abstraction—photographs,
renderings) to schematic (high level of abstraction—logos, graphs,
diagrams; see Twyman 1979). Below, the data in our sample mes-
sage are displayed schematically:
During tiiepast month, our clientele in each region has increased as foUows:
East South West
Alphanumeric signs (cell 7) encode the X axis with geographic
locations, the Y axis with figures; spatially (cell 8), plotting the
data on the grid determines the vertical distance between figures
and the horizontal distance between regions; graphically (cell 9),
shaded bars encode the data, lines locate the axes, and tick marks
gauge the scale. All of the coding modes are open to variation.
"East" can be simplified to "E," and figures on the Y axis re-
placed by exact values above each bar. Spatially, the distances
between values and between regions can be increased or decreased,
or the axis shifted 90 degrees so the bars run horizontally. Graphi-
cally, the bars can be cross-hatched, colored, or changed to simple
lines, and the tick marks replaced by a light grid.
If the extra-textual were pictorial rather than schematic, the
same variables would come into play: spatially, the image of the
employees at a conference table can be close or distant or rotated
36 THE JOURNAL OF BUSINESS COMMUNICATION • 25.3:SUMMER 1988
at different angles from the real thing; graphically, it can be coded
at a h ^ level of realism (with plenty of surface details, down to
the wood grain on the table) or reduced to an abstract stick draw-
ing or a stylized logo. The coding choices for either schematics or
pictorials are interdependent: the graphic coding of a logo may
flatten an image spatially; the spatial coding of data on a bar chart
may restrict the range of graphic choices. Like the intra- and inter-
textual levels, visual language at the extra-textual level depends on
the contextual variables of the system.
The Supra-textual level (cells 10-12), like the inter-textual,
operates chiefly in the spatial mode, though on a more global
scale, arranging extra-textuals (graphs, pie charts, pictures) within
the text and regulating the flow from one page or section (para-
graph breaks at the bottom of the page; breaks between parts of a
report) to the next to create a coherent document. For example,
in the last version of our sample, the spatial mode (ceU 11) locates
the extra-textual relative to the text ("During the past month . . ."):
centered directly beneath the text, or to the left or to the right; or
isolated on a separate page in close proximity or in an appendix.
In the alphanumeric mode (cell 10), supra-textual elements include
segmenting and serial devices (page headers, section titles, pagina-
tion); for example, I could label the bar graph in the sample "Fig-
ure 6," indicating a supra-textual relation with other extra-textuals
in the document. In the graphic mode cohesive devices—line work,
page color, logos—unify sections of a document, providing visual
markers for readers. In our exsimple, a line border around the bar
chart would further demarcate the transition from text to extra-
text. By mapping out the terrain, the supra-textual level controls
the three-dimensional configuration of the document, enabling
readers to comprehend its logic and hierarchy and to retrieve
essential information.
Each document blends the coding modes and levels into an idio-
syncratic visual system incorporating most of the cells on the
matrix. Consider the visual choices for a simple analytical memo
report. At the intra-textual level, typefaces, point sizes (cell 1),
spacing (cell 2), and graphic cues (cell 3), selected from a menu
of computer-generated options, create interest and emphasis and
regulate the intra-linear flow of signs. At the inter-textual level,
headings (cell 4), indentation, and vertical spacing (cell 5) struc-
ture the text, and graphic coding around textual particles (cell 6)
forms tables and matrices. At the extra-textual level, spatial coding
distributes data (monthly sales f^res, production costs) on a grid
VISUAL LANGUAGE • KOSTELNICK 37
(cell 8), graphic coding records data with shaded bars (cell 9), and
alphanumeric coding (cell 7) explains each variable. At the supra-
textual level, pages are numbered and labeled with subject, date,
author, and number (cell 10), composed spatially to locate the bar
charts within the text and to ensure continuous flow from page to
page (cell 11), and coded with icons, colors, tabs, or separate pages
(blank or titled) to differentiate sections of the report (cell 12).
Individually, none of these choices may be novel, but when com-
bined in the same document they compose a system of visual lan-
guage dependent, as all visual images are, on context for their per-
ceptual qualities (Amheim, 1969). Hence, the levels and coding
modes are interdependent. The cell 1 selection of a type size, pos-
ture, and style expands or contracts the document and therefore
has supra-textual consequences. Conversely, a cell 11 decision to
write a "one-page resume" impinges on intra- and inter-textued
choices. The cell 4 decision to deploy a variety of headings may re-
quire cell 1 assistance to establish a hierarchy. The spatial and
graphic coding of a bar chart (cells 8, 9) influence supra-textual
page design. Choices for each coding mode and level, therefore,
are contingent upon other choices.
The 12-cell matrix opens up these choices to public purview,
enabling us to describe how visual elements coedesce to form co-
herent systems. Just as each writing situation engenders a different
combination of linguistic choices, so too with visual language:
each document requires conscious planning so that the modes and
levels work harmoniously to achieve the same goals. To the extent
that the cells on the matrix compose an integrated system that
contributes to the purpose of the message, the document obtains
a certain level of visual sophistication or intensity.
INFORMATION DESIGN: TOWARD A
FUNCTIONAL VISUAL LANGUAGE
Of course not all visual systems succeed in accommodating the
reader. What makes visual language reader-oriented? In a fast-paced
business environment, effective visual display simplifies the reader's
infonnation processing. As Hartley (1985) puts it, "The printed
page should provide a reliable frame of reference from within
which the readers can move about, leave and return without con-
fusion" (p. 18). Above all else, visual language should be function-
al. And on an idiomatic and conventional level, it usually is: words
38 THE JOURNAL OF BUSINESS COMMUNICATION ' 25:3:SUMMER 1988
ordinarily flow horizontally across the page; letters have signature
blocks; bars on graphs stand for measured units. These generic
conventions enable the reader to place visual codes into the slots
of menteil schemata (see Schumacher, 1981). A one-page document
that follows the conventions of a letter and arrives in an envelope
with a booklet "fits" nicely into the slot of "cover letter," a per-
suasive message, widely distributed, that entices the reader to
examine an accompanying document. On the other hand, a letter
set in boustrophedon (as the ox plows, back and forth), or hori-
zontally across the length of the page, poses an obvious "misfit"
to most readers. By activating schemata, generic conventions foster
reader expectations about the whole message.
Visual conventions, however, define only general patterns,
change over time and from one discourse community or culture to
another, and may not provide the optimum visual patterns for
information processing. To compensate for these shortcomings,
the "information design" movement has tried to establish more
stable, culturally neutral, and empirically verifiable guidelines.
Combining theory, practice, and research, information design
draws upon several disciplines—graphic design, perception, educa-
tional and cognitive psychology, and semiotics—to discover reader-
oriented visual patterns for many of the levels and coding modes
on the matrix.
From the outset, the information design movement soi^ht to
create a modem, rational, ideologically free system of signs acces-
sible to a mass audience. Its intellectual roots can be traced to the
functional modernism of the Bauhaus in the 2O's (Kinross, 1985)
and indirectly to the visual dynamism of Futurist and Constructi-
vist text displays. Each of these movements experimented with
radically new modes of visual design, emphasizing clean, nontradi-
tional forms and typefaces compatible with Machine Age technol-
ogy (see Crouwel, 1979). Together these fresh approaches to
textual display and image-making provided an aesthetic and func-
tional springboard for early information designers. For example,
Neurath's "isotype" (international system of picture education)
symbols epitomized the quest for a rational, cross-cultural language
of extra-textual s ^ s for displaying quantitative data. After World
War II the search for simple, universal, and technically precise
forms continued in the invention of typeface systems, for example
the serifless Univers style (Kinross, pp. 23 & 27). On a supra-
textual level, the appeal to the universal was reinforced by the
laws of gestalt psychology (similarity, good figure, equilibrium).
V I S U A L L A N G U A G E • K O S T E L N I C K . , _ 3 9
which transcended style and culture by codifying perceptual ex-
periences detached from language or ideology.
The information des^n movement, therefore, has evolved into a
modem philosophy of visual display emphasizing objectivity, clari-
ty, and the integration of text with simple forms. The movement
has had a wide impact on advertising, govemment documents and
forms, educational texts, and scientific and technical publications.
Today theorists and practitioners continue to develop and test
methods for functional designs. On the inter-textual level. Hartley
(1985), Wright (1977), and several others have devised methods of
employing headings, vertical spacing, and graphic cues to surface
discourse visually, while Hom (1982) has codified these elements
into a comprehensive system of "structured writing." On the extra-
textual level, Tufte (1983) and Bertin (1981) apply functionalist
principles to the visual display of quantitative data. Both advocate
economy through density, arguing that graphs and charts should
display the maximum information in the smallest space with the
least ink. Bertin invokes semiotic theory to define visual clarity:
graphic sign systems should be "monosemic" (pp. 177-179), the
relations among s ^ s conveying a single, unambiguous meaning.
Consistent with its functionalist tradition, contemporary informa-
tion design theory values economy, objectivity, and semantic and
structural transparency.
Further supporting the functionalist approach are a wealth of
empirical studies aimed at identifying optimum visual displays. At
the intra-textual level. Tinker (1963) found that type sizes be-
tween 9 and 12 point maximize readii^ efficiency, lower case
letters are more readable than upper case, and excessive italic print
reduces legibility. On the inter-textual level. Hartley (1984) has
studied, with mixed results, the effects of horizontal and vertical
structuring of text on readability, while Hartley and Trueman
(1985) have shown, through extensive experiments, that headings
improve reader performance. Wright (1968) experimented with the
spatial configurations of tables, finding that users of currency
conversion tables perform better with redundant lists than with
matrices. On the extra-textual level, Macdonald-Ross (1977)
summarized over five decades of empirical research on the graphic
display of quantitative data, concluding for example that horizon-
tal bars are superior to unkeyed circles. Psychologists have studied
the effects of spatial depth and degree of graphic detail in the
processing of representative images (Perkins, 1980). They found
that readers need a modest amount of acculturation, should be
40 THE JOURNAL OF BUSINESS COMMUNICATION ' 25.3.SUMMER 1988
given only selected information pertinent to the task, and perceive
depth on a two-dimensional surface differently from the "real
thing." At the supra-textual level, guidelines can be derived from
hemispheric brain research (see, for example, Welford, 1984).
Because readers process visual information more effectively in
the right brain (the left field of vision), pictures and schematics
are generally better placed to the left, explanatory text to the
right.^
Many of the empirical findings jind theoretical positions cited
above presume a universal reader performing generic, noncon-
textual tasks. Wright (1977) sums up the empiricist's dilemma:
"Researchers often consider themselves free to ask if format A is
better than format B, without needing to specify better for what?"
(p. 112). Some of the findings—in particular, those of Hartley and
Wright—are user-centered and tied to specific audiences and con-
texts. Wright's experiments with tabular displays are bound to the
readers' experience with visual arrangements of figures, the situa-
tion in which the display is used, and the readers' familiarity with
the subject. M£iny case studies of public documents are con-
textually sensitive to the special needs posed by audience and
situation. Such studies entail pretesting and interviewing during
the design process to assure quality control and to convert visual
language to a "plain English" style intell^ible to the nonspecialist
(Redish, Felker, and Rose, 1981; Wright, 1979).
The functional guidelines of the information design movement,
therefore, specify which configurations afford the most efficient
processing of information, enabling us to engineer clear visual
conduits. The emphasis on information privileges function over
decoration, just as the Bauhaus did. A well-designed text repli-
cates on the page the sleek steel and glass of a modem h^h-rise:
functional, aesthetically austere, and culturally neutral. By meld-
ing text and form, the information design movement revolution-
ized the visual language of texts and made design a bona fide part
of the creative process, from planning to production.
THE MYTH OF RHETORICAL NEUTRALITY
Ideally, information des^n deploys visual language to ensure
an unadulterated channel between sender and receiver. However,
visual elements themselves—especially in business messages—obtain
meaning beyond the simple transfer of information. "' Pure' inf or-
VISUAL LANGUAGE " KOSTELNICK 41
mation," asserts Bonsiepe (1965), "exists for the designer only in
arid abstraction. As soon as he b^ins to give it concrete shape, to
bring it within the range of experience, the process of rhetorical
infiltration b^ins" (p. 30). Simply put, a "rhetoric of neutrality"
is a "pipe-dream," for even the mimdane visual language of a train
timetable has a certain rhetorical force, intended or not, that af-
fects the reader's judgment (Kinross).
This rhetorical "infiltration" stems partly from taste and cul-
tural values. Futurist and Constructivist text displays, the fore-
runners of infonnation design, were considered iconoclastic when
they first appeared; today they seem modem but pass^. Then as
now, they relinquish any claims to neutrality. The Machine Age
modemism of the information design movement also constitutes a
style—streamlined, simple, and low-key—but also subject to the
vicissitudes of intellectual and aesthetic favor, and hence anything
but "objective."* The invention and evolution of styles continually
transform taste and perception, in art and in business communica-
tion. An early corporate logo, letterhead, or trademark juxtaposed
to its contemporary version demonstrates the visual distance be-
tween past and present, and presages future recreation.
Technology also influences the rhetoric of visual langus^e. At
the outset of the century most texts were handwritten in one of
four styles: business, civil service, text, or legal. The appearance of
the typewriter transformed "micro-level" (Waller, 1980) visual
choices, forcing characters into uniform, equally spaced units on a
grid and initiating a whole new set of intra- and inter-textual con-
ventions (Walker, 1984). Ninety years ago a letter done in business
script would have signaled the nature of the message but otherwise
attracted little attention. Today it would totally alter the rhetoric
of the communication: the writer would suffer (an entry-level
application letter) or benefit (a letter from a CEO to a consumer)
from the unorthodox visual coding.
Cultural and technological changes are long-term, evolutionary
tendencies that gradually modify the shared perception of large
discourse communities. The choices business communicators make
everyday are highly contextual and vary from one audience and
visual system to another. Information design can weed out percep-
tually poor choices—illegible type styles, line spacing, and data dis-
plays—but cannot necessarily prescribe rhetorically appropriate
choices goveming the arrangement, style, or tone of a particular
document.
For example. Tinker found that many common typefaces were
42 THE JOURNAL OF BUSINESS COMMUNICATION ' 25:3:SUMMER 1988
equally legible but that readers preferred some faces over others.
The "atmosphere value" or "semantic quality" of typefaces (Bar-
tram, 1982) opens the door to subjectivity and rhetorical judg-
ment.* What "pleases" readers in one message may not be pleasing
or rhetorically consistent in another. Visual choices demand adap-
tation to the particular audience, purpose, and circumstances of
the document. A simple intra-textual (cell 1) choice of a typeface
can have a major impact on the rhetoric of a messf^e. For an in-
formal message to a co-worker, a handwritten note or a dot-matrix
printout reinforces the familiar tone and transience of the docu-
ment. However, a manager (one that values job security) submitting
an annual report to a board of directors or a proposal to a govem-
ment s^ency will specify a typeface that projects credibility and
permanence.
Although technical capability helps, visual rhetoric is not neces-
sarily constrained by available technology. Direct mail sales let-
ters use typewritten serif styles even though virtually any typeface
could be used, given the bulk printing. Indeed, to make messages
appear more spontaneous, senders frequently underline key points
with insouciant strokes of the pen (in red ink) to personalize the
visual style. Urgent mess£^es (6.g., collections) are customarily
typed entirely in upper case, sacrificing legibility for a more
official, authoritarian tone. These choices may seem trivial and
subjective, but they are made for rhetorically sound reasons. Add
to these variations the ability to highlight text with boldface and
multiple fonts and point sizes now commonly available in desk-top
printers, the business communicator will experience far greater
freedom to adapt intra-textual tone and style to the rhetorical task.
The same holds for information design on any level of the matrix.
Inter-textual structuring—headings, vertical spacing, graphic cues-
has rhetorical consequences, surfacing certain aspects of the mes-
sage while embedding others. Opening up a text with spatial and
graphic coding is equivalent to persuading visually, impelling read-
ers to value selected pieces of information and to acquiesce to logi-
cal and hierarchical connections that make them cohere visually.
In the second version of our sample text, the list with bullets estab-
lishes visual parallelism among the three regions. Perhaps the West
R^ion is new and the writer wants to emphasize its status as a
full and equal sector. Perhaps isolating these elements ai ^es that
the vice-president for sales is doing satisfactorily in all regions, or
that by comparison the West Region is faltering, or that more
employees should be hired in the South. Depending on the con-
VISUAL LANGUAGE • KOSTELNICK 43
text, selective textual surfacing advances the writer's ai l ment .
In an extremely sensitive message (a perfonnance appraisal), the
writer may even eschew any sort of inter-textual variation to
embed key information, sacrificing lability to maintain rhetori-
cal integrity.
The third version of our sample detaches the data from the
text, reshapes them graphically, and tells the reader in a resonant
voice: "Here are the facts, bare and hrefutable!" The visual style,
bordering on hyperbole, can of course be modified. Spatially, the
bars can be extended or contracted on the grid, the graphic codii^
ornamented or streamlined, the alphanimieric codii^ expanded—
depending on the designer's argument and rhetorical stance. The
differences among regions can be accentuated or diminished, the
graph made to appear scientific or aesthetically pleasing, clumsy or
self-serving. Any of these effects influences the reader's impression
of the data as well as of the author's integrity and is far from neu-
tral. Even acquiescing to someone else's configuration of the data
is still to choose, and the reader to receive, the visual rhetoric of
the message.
Together these choices coalesce to drive the visual rhetoric of
the document. Consider the range of visual decisions that go into
an annual report to stockholders: the company logo, charts and
graphs to display sales and production figures; apian and line draw-
ing of a new plant; photographs introducing board members; and
textual displays of expenses, revenues, and earnings. Aloi^ with
these are numerous supra-textual decisions regarding the location
of extra-textuals, the size and number of pages, and graphic cues
such as color and line work intended to coordinate the whole
document, add interest to complex data, instill confidence in in-
vestors, and accurately project the company's image (innovative,
consumer oriented, h^h-tech). The visual system is both informa-
tional and persuasive, enabling investors to locate pertinent facts
and, more importantly, enabling the company to market its per-
formance. Although annual reports are more complex than every-
day messages, maint{iining full rhetorical control over any business
document requires careful visual planning. In other words, the
writer must attend to the visual system during the composing
process as rhetorical strategies take shape.
44 THE JOURNAL OF BUSINESS COMMUNICATION • 25:3:SUMMER 1988
COORDINATING VISUAL AND VERBAL RHETORIC:
A HOLISTIC PROCESS
Planning an entire document is a "holistic process" (Barton &
Barton, 1985) in which the visual and the verbal modify, comple-
ment, and compensate for one another to accomplish the rhetori-
cal goals of the communication. Linguistic invention entails in-
numerable lexicEil choices; visual invention materializes from the
wealth of elements defined by the matrix. Both are driven by the
same purpose and by many of the same principles and concepts:
style, tone, level of technicality, interest, benefits, arrangement,
readability, and so on. During the planning stage of a document -
say a proposal persuading a skeptical reader to act on a recom-
mendation-Hhe visual rhetorician considers several options:
• Stimulate reader interest through a logo, picture, or data
display (cells 8, 9) and through highly variegated inter-textual
stmcture (cells 4, 5, 6).
• Surface selected arguments and benefits through intra- and
inter-textual variations (cell 1: boldfacing and all capitals; cell 3:
underlining; cell 6: bullets to highlight items in a list).
• Adapt the style of data displays (cells 8-9) to a nonspecialist
(cell 7 alphanumeric coding), and project a serious tone necessary
to engage this reader (cell 1 choice of a typeface; cells 6, 9, 12
graphic coding).
• Map out the document with inter- and supra-textual cues
(cells 4-6, 10-12) that guide the busy reader through the informa-
tion, increasing readability.
Because these and many other strategies apply to visual as well
as verbal problem-solving, early planning synchronizes the two
from the start.
Early planning also exploits the special rhetorical capabilities of
the visual and the verbeil. For example, a bar graph can often argue
with statistics more clearly and emphatically than inter-textual dis-
plays of figures; a logo can establish the identity of an oi^Jiniza-
tion or a project more succinctly and forcefully than an explana-
tory paragraph. On the other hand, in a sales letter a poetic
description of the scenery at a resort may be more provocative
than a high resolution photograph, and in a n^ative recommenda-
tion letter a seamless text more tactful than one that is visually
transparent.
Because the visual and verbal are interdependent, holistic plan-
ning enables the writer-designer to monitor the interaction between
VISUAL LANGUAGE • KOSTELNICK 45
the two. A highly vari^ated inter-textual system chaises sentence
structure and syntax, increasing the frequency of parallelism and
fragments and eliminating the need for verbal transitions (Bem-
hardt, 1986). Converting text-bound figures to a data display (as
in the third version of our example) contracts and reshapes tiie
text. On the other hand, a complex line graph may require textual
expansion to hi ghl i t and explain key data. Style and tone are
also interactive: a set of informally written instructions on how
to pot plants calls for simple line drawings, instead of photo-
graphs, to illustrate the task; the linguistic tone of a sales letter
theme or slogan sets the visual tone for images on the envelope
and enclosures.
Globally, supra-textual decisions can regulate the linguistic evo-
lution of the text. Thing-like, the page is a tablet, the writer's unit
of workspace, like an artist's canvas, an architect's story, or a
sculptor's marble slab. The tablet as visual workspace both compli-
cates and simplifies the writer's task. Some undergraduates strug-
gle to find enough material to fill up a "one-page" r4sum^, whUe
others try desperately to conflate several pages into one. As items
are arranged on the tablet, the writer strikes a balance between the
functional exigencies of information design (the user may give the
document only a cursory glance) and an appropriate rhetorical
stance—ihe need to appear trustworthy, ambitious, successful, and
employable. To these ends, elements are moved about, shaped, cut,
and recast linguistically (in phrases and fragments). Through the
symbiotic growth of the visual and the verbal, the document
evolves within the supra-textual boundaries of the tablet.
Visual design, like writing, is a process of discovery. This dis-
covery process occurs within and outside the text. Articulating
a text with intra- and inter-textual cues-H;ypographical variations,
paragraph breaks, headings, lists, tables—unfolds and structures
ideas, enabling the writer to resee the logic of the mess^e. Creat-
ing visual clarity demands clear thinking, and rethinking, about
content (Hartley, 1984, p. 502). Extra-textuals also act as dis-
covery tools (Bertin, 1981). The process of translating data into a
visual sign system surfaces patterns and relationships submerged in
a table of figures. This happens rudimentarily in the third version
of our example: the designer must experiment with graphic aad
spatied coding to find a rhetorically suitable configuration. If we
had additional figures—about client increases in previous periods,
the clients themselves (individuals, businesses, govemment agen-
cies), or dollar volumes per client or region—permuting these data
46 THE JOURNAL OF BUSINESS COMMUNICATION • 25:3:SUMMER 1988
on the X-Y axes would enable us to find an insightful pattem.
Supple and unpredictable, data displays are powerful discovery
tools, much like freewriting, zero drafting, and outlining.
At the outset of the holistic process, therefore, visual thinking
may be provisional: designing a document is an evolutionary
process entailing invention, revision, and editing. Visual language
is rich and flexible, allowing infinite syntactic variation of the
constituent elements outlined on the matrix. However, if these
elements eire to be transformed into inventive patterns, they must
be treated as bona fide rhetorical tools guided by the same strate-
gies that govern the linguistic text. Asserting complete rhetoricsd
control over a document, therefore, demands visual thinking from
planning to production.
CONCLUSION
Business communication is intensely visual: as practitioners,
researchers, and instructors we are inundated with an array of visi-
ble forms. Although we use, teach, and presumably value effective
visual language, we have no system for describii^, analyzing, or
evaluating it. This is partially because d e s ^ is viewed as a ready-
made set of conventions or as a task performed by specialists—
in either case extrinsic to the writing process. However, in the
making of business documents, writing and designing, reading and
seeing, are interdependent, complementing and impinging on each
other to satisfy the goals of the communication. Combined in any
business document, the three coding modes and four levels of
visual language comprise an integrated system enablii^ the reader
to process information efficiently from the rhetorical perspective
the writer intends. Adapting the visual system to the message is
essential in business communication because audience and purpose
vary from one message to another and because writers have the
autonomy to design their own documents.
Successful adaptation requires attention to design throughout
the writing process. We have only rudimentary knowledge about
how visual and verbal thinking interact durii^ this process or how
the contextual variables of visual language affect readers. What we
do know, instinctively though perhaps unconsciously as members
of a "print culture," is that visual thinking is intrinsic to the dis-
covery and communication of meaning. To use visual language
fluently, or to understand its stmcture and nuances, we must ex-
VISUAL LANGUAGE • KOSTELNICK 47
plore it explicitly and systematically, with the same rigorous scru-
tiny given the linguistic text.
NOTES
' Twyman (1979) has proposed a graphic schema containing four modes
of symbolization and seven methods of configuration. His matrix contains
twenty-eight "cells" and classifies primarily spatial variations at the inter-
and extra-textual levels.
^ Some of Twyman's methods of configuration are particularly rele-
vant here: Pure linear (a continuous flowing text; e.g., in circular form).
Linear interrupted (conventional text with line breaks). List, Matrix (tables),
and Linear branching (a tree, an organizational chart).
^ Design guidelines derived from hemispheric research are contingent
upon at least two variables: the kind of information processing required by
the task (part-by-part vs. holistic analysis) and the amount of verbal thinking
the viewer is doing at the time.
^ See Kinross for a detailed account of the aesthetic and ideological
evolution of the information design movement and its claims to rhetorical
"neutrality."
^ The "atmosphere value" of typefiaces has long been the subject of
study and speculation. For example, Ovink experimented with several type-
&ces used in literature and advertising, grouping their subjective effects under
categories such as "luxury/refinement," "economy/precision," and "strength."
But the moods and cultural values associated with intra-textual styles can
change dramatically: serifless letter forms, which are typically allied with
modernism, were initially revived in the eighteenth century because they
evoked "rugged antiquity" (Spencer, 1969, pp. 29-30).
REFERENCES
Arnheim, Rudolf (1969). Visual thinking. Berkeley, CA: tJniversity of Cali-
fornia Press.
Barton, Ben F. & Barton, Marthalee S. (1985). Toward a rhetoric of visuals
for the computer era. The Technical Writing Teacher, 12,126-145.
Bartram, David (1982). The perception of semantic quality in type: IHffer-
ences between designers and nondesigners. Information Design Joumal,
3, 38-50.
Bemhardt, Stephen A. (1986). Seeing the text. College Composition and
Communication, 37,66-78.
Bertin, Jacques (1981). Graphics and graphic information-processing (William
J. Berg and Paul Scott, Trans.). New York: De Gruyter.
Bonsiepe, Gui (1965). Visual/verbal rhetoric. Vim, 14-16, 23-40.
Crouwel, Wim (1979). Typography: A technique of making a text 'legible.'
In Paul Kolers, Merald E. Wrolstad, and Herman Bouma (Eds.), Proces-
sing of Visible Language, {1, pp. 151-164). New York: Plenum.
Gelb, I. J. (1980). Principles of writing systems within the frame of visual
communication. In Kolers, Wrolstad, and Bouma {2: pp. 7-24).
48 THE JOURNAL OF BUSINESS COMMUNICATION ' 25:3:SUMMER 1988
Hartley, James (1985). Designing instructional text (2nd ed.). New York:
Nichols.
Hartley, James (1984). Space and structure in instructional text. In Ronald
Easterby and Harm Zwaga (Eds.), Infonnation design (pp. 497-515). New
York: Wiley.
Hartley, James & Trueman, Mark (1985). A research strategy for text design-
ers: The role of headings. Instructional Science, 14,99-155.
Herrstrom, David Sten (1984). Technical writing as mapping description onto
diagram: The graphic paradigms of explanation. Journal of Technical
Writing and Communication, 14,223-240.
Horn, Robert E. (1982). Structured writing and text design. In David H.
Jonassen (Ed.), The technology of text (pp. 341-365). Englewood Qiffs,
NJ: Educational Technology Publications.
Kinross, Robin (1985). The rhetoric of neutrality. Design Issues, 2(2), 18-30.
Macdonald-Ross, Michael (1977). How numbers are shown: A review of re-
search on the presentation of quantitative data in texts. AV Communica-
tion Review, 25, 359-409.
Ong, Walter (1982). Orality and literacy: The technologizing of the word.
New York: Methuen.
Perkins, D. N. (1980). Pictures and the real thing. In Kolers, Wrolstad, and
Bouma (2, pp. 259-278).
Redish, Janice C, Felker, Daniel B., & Rose, Andrew M. (1981). Evaluating
the effects of document design principles. Information Design Journal, 2,
236-243.
Schumacher, Gary M. (1981). Schemata in text processing and design. Infor-
mation Design Journal, 2,17-27.
Spencer, Herbert (1969). The visible word (2nd ed.). New York: Hastings.
Tinker, Miles (1963). Legibility of print. Ames, IA: Iowa State University
Press.
Tufte, Edward (1983). The visual display of quantitative information. Che-
shire, CT: Graphics Press.
Twyman, Michael (1979). A schema for the study of graphic language. In
Kolers, Wrolstad, and Bouma (i , pp. 117-150).
Walker, Sue (1984). How typewriters changed correspondence: An analysis
of prescription and practice. Visible Language, 18,102-117.
Waller, Robert H. W. (1980). Graphic aspects of complex texts: Typography
as macro-punctuation. In Kolers, Wrolstad, and Bouma (2, pp. 241-253).
Welford, A. T. (1984). Theory and application in visual displays. In Easterby
and Zwaga (pp. 3-18).
Wright, Patricia (1979). The quality control of document design. Information
Design Journal, 1, 33-42.
Wright, Patricia (1977). Presenting technical information: A survey of re-
search findings. Instructional Science, 6, 93-134.
Wright, Patricia (1968). Using tabulated information. £;:?o«omfcs, 11, 331-343.
doc_944160055.pdf