Tversky, B. (2001). Chapter 4: Spatial schemas in depictions. In M. Gattis (Ed.), Spatial
Schemas and Abstract Thought (pp. 79-112). Cambridge: MIT Press.
JUSTIFICATION:
Discusses and empirically documents how graphics provide support for educational functions such
as maintaining attention, supporting memory, providing mental models, and facilitating discovery.
A heavy chapter from one of the important scholar/researchers in the field.
Bar graphs, pie charts -> for representing quantitative data
Flow charts, trees, networks -> for representing qualitative data
Icons, pictorial representations -> concepts and information difficult to visualize
HOW GRAPHICS RELATE MEANING
Tversky presents an analysis of how pictorial representations, based on their
sub-components (she calls elements) and their relationship in space, express
meaning. Two graphic inventions (3D and animation) are also explored in terms
of the functions they serve and how their design should be best implemented.
The “elements” are the basic components of a graphic representation, that which
represent individual elements in the real world. Their spatial relationship to each
other represent the relationship between elements. This parallel can be seen in
conventional maps as to the “what” vs. “where” information supported in visual
cognition: information about critical components themselves and information in
how the components relate to one another. These graphic representations are
external, functional to reduce cognitive load for memory and processing while still
being available for accessing, integration, and operation for conceptual
organization. Externalized graphical representations are also inherently
collaborative, allowing multiple people to simultaneously cognate information
facilitating group information.
metonymy – a figure depiction where an associated object represents the
concept (ex. a trash can represents an unwanted file of words or a file folder
represents a group of words)
synecdoche – a figure depiction where a part is used to represent a whole, or a
whole for a part (ex. a foot for “walking” or a gas pump for “a gas station is
nearby”)
nominal scales – clustering by category
Graphic displays for nominal scales often use grouping to convey meaning,
capitalizing on Gestalt principles for mental representation. Simply, element in
close spatial proximity are more “related” than those further apart.
Devices for categorical/nominal relations:
 spaces between words, rows, columns
 use of parenthetical devices, boxes, circles, Venn diagrams
 visual devices such as color, shading, texturizing
Devices for simple subordination:
 indentation
 size (shape, font)
 superposition
 highlighting
 punctuation
Devices for complex subordination:
 order
 trees (hierarchy)
Tables both group and juxtapose elements simultaneously by using spatially
arrayed rows and columns. Empty space is a technique used for grouping.
Some visual signs can also be used for grouping, falling into either the Gestalt
principles of enclosure or similarity. Other features of graphical representations
can convey categorical relations such as colors and lines (political boundaries or
geographical features [blue for ocean, beige for desert]). Colors can convey
qualitative information as well (deeper blue for deeper water) and can be used
metaphorically (degree of excitation in an MRI).
In graphs, bars seem to be container like and lines seem to make connections.
(Zacks and Tversky 1999). Therefore, categorical or discrete concepts are
mapped into depictions that “contain” and ordinal (sequential) concepts are
mapped into depictions that “connect.”
3D
“Perceiving and comprehending the 3-D details, however, can be problematic
from flattened images. Such perceptual difficulties abound, especially for odd
views or scenes with multiple objects. The retina itself is essentially flat, and
beyond small distances, perception in depth depends on a set of cues such as
occlusion, foreshortening, and relative size rather than on stereoscopic vision.
Those clues are only clues however; they are fallible. Their limitation yield error
in depth perception and in object recognition alike (e.g., Loomis, DaSilva, Fujita
and Fukusima, 1992). 3D interfaces suffer from exactly the same limitation.
Even within the bounds of stereoscopy, only one view of an object a scene is
present at once. This means that mental representations of three dimensions
must be constructed from separate views. It also means that some objects may
be occluded by others, as frequently happens in 3-D depth arrays of bars.
Frontal views of 3-D bars often become reversible figures, hard to stabilize in
order to inspect. (pp. 102-103)”
Research has shown that professionals in architecture have difficulty
conceptualizing 3 dimensions on a 2 dimensional display, which is why they
prefer to design in 2D representations (Arnheim 1977; Suwa and Tversky 1996).
Interpretation in this manner is especially difficult for novices (e.g., Cooper,
Schacter, Ballesteros and Moore, 1992; Gobert 1999). Constructing 3D
conceptual representations from 2D graphic representation is a difficult task
(Shah and Carpenter 1995), as shown by students who when presented 3D
graphical representations preferred to describe them by focusing on 2D verbal
descriptions (?not sure this is a valid study).
Learners express their preference for 3D data representations when
 showing data to others rather than self
 remembering data rather than examining it
ANIMATION
A natural use for a series of changing cells is to convey actions or processes or
time. This holds true for concrete actions (people and mechanical motions) and
abstract actions (weather, causal relations of algorithms). Some dynamic events
take place to quickly for the learner to perceive, which may produce temporal
misconceptions. Even events that are slow may be more effectively conceived
as discrete steps, unless the timing is of vital importance to the concept being
conveyed. This problem of accuracy is one disadvantage of animation. Static
series of depictions can be easily inspected, with the advantage of re-inspection,
with the opportunity to draw relationships to other elements for understanding.
Animation has been shown to be effective in teaching RxT=D problems, where
motion is conveyed as a moving point (Baek and Lane 1988). Another
instructional success has come with using animated diagrams for changing the
focus of the learners’ attention. Tversky suggests utilizing animations in those
situations which have a combination relevant to cognitive naturalness.
Specifically, ones which are slow enough in which perception can follow
important information within an event and simple enough to focus attention on
the desired instructional material. She also suggests that animations should
reflect people’s natural temporal processes, that is, discrete if the situation
should be discrete and continuous when it should be continuous. They should
definitely be used to portray changes over time or when sequence is of great
importance.
research by analysis -> gives clues to how graphics are used and produced by
designers
research by observation of how they are used -> gives clues in how they should
be designed
Bertin (1981) – function of graphics
 record information
 communicate information
 process information
 good ones provide a simplification of the above functions
Tversky – function of graphics
 attract attention and interest
 record information
 facilitate memory
 facilitate communication
 provide models of actual and theoretical worlds – there is a continuing debate
about how much information should be conveyed in a ‘successful diagram’, a
debate exists between creating the maximum amount of detail and realism vs.
minimalism to give maximum amount of focus to a particular aspect
 convey meaning, facilitate discovery and inference
“A major purpose of graphic displays is to represent visually concepts and
relations that are not inherently visual. Graphic displays use representations of
elements and the spatial relations among them to do so.” (p. 111)
“This review suggests a perhaps deceptively simple maxim: use spatial elements
and relations naturally.” (p. 111)