Psychology is the
science that concerns about the mind, the behaviours exhibited and how they
could be explained. Several theories have been developed over the years,
complementing or competing with each other. Beside structural, functional,
behavioural, psychoanalytic and cognitive approaches, just to name a few, the
relative young ecological theory has evolved. It provides an interesting view
on and explanation of human and animal action. We will use ecological
psychology to understand ‘natural behaviour’ and thus ‘natural interaction’.
Ecological
psychology is “the branch of psychology that deals with the relations between
individuals and their environment, geographically and social” (Webster’s New
Dictionary 1994). Specifically it “emphasizes the animal’s activities in
encountering its environment” (Reed 1996). An important concept in this theory
is affordances, suggested and consequently developed by J.J.Gibson (Gibson 1979).
The affordances of
an individual’s environment are what that environment offers (provides,
permits, makes available), either for good or for ill[1]. They encompass what can be done to, in, on,
with or about an object/substance/event. Potentially affordances are the
behavioural possibilities of an environmental layout taken with reference to an
individual’s action capabilities. Effectively they are a property of the
coupling of a user to a given environmental layout. We ordinarily detect,
discriminate and recognize affordances. We behave so as to achieve desirable
and beneficial affordances or to avoid undesirable and injurious affordances.
Each individual can
directly perceive affordances (values, meanings, significance) as they are
external to him. Their meaning is defined in terms of behavioural consequences
(e.g. avoiding drowning is a consequence of behaving relative to water).
Available information about appropriate affordances may not be perceived at all
(as evidenced by walking into a glass door), or may be misperceived or misclassified
(as evidenced by taken a bite out of a wax apple).
The sets of
capabilities for action that define the behavioural repertoire of the
individual, are called effectivities. They are defined in relationship to the
affordances of the environment. Realising an intended effect, i.e. effectuating
an affordance, means taking advantage of what is offered, bringing an effect to
pass, accomplishing, or making operative. Perception guides the selection of
effectivities and can be mediated through devices (e.g. a Geiger counter for
detection of radioactivity). Available information sets the occasion for
perception of affordances but cannot cause it. And affordances set the occasion
for choices among alternative actions. They are “opportunities for action, not
causes or stimuli; they can be used and they can motivate an organism to act,
but they do not and cannot cause even the behaviour that utilized them” (Reed
1996). Motivation sets the occasion for intentions but does not select one.
Intentions are constrained by opportunities, by personal and social values, and
it’s the individual who chooses. In the end there is no causal relationship
whatsoever between available information and resulting behaviour. It’s more of
regulation nature: “The ability to use information allows the individual to
know where and when to behave so as to take advantage of available
affordances”, thus “behaviour (…) is not caused” (Reed 1996) but facilitated.
The individual always has a choice – including the decision not to act, not to
achieve or not to avoid.
Under the eye of
this theory the initial problem of determining what ‘natural behaviour’ is
means to take a closer look on humans, their social affordances, i.e. they
afford interacting with, and the relation to our field of interest, i.e. agents
in augmented reality.
Today affordances of
technology mediated interfaces are drastically different from affordances of
humans and our environment to which we have accustomed during evolution. The
interaction with machines are likely to require different actions from users in
order to accomplish the same goal and they are rather rudimentary and
unnatural. Communication with e.g. computers is still very much narrowed,
mostly to keyboard, mouse and display, first two only for input, latter only for
output. In contrast humans are used to and capable of handling multi-modal,
multi-threaded activity and conversation threads. Humans can distinguish
between them by using different communication mechanisms, directing
conversation to different participants, or just relying on context. In a normal
human-human interaction the communication space includes the task space (e.g.
two architects discuss over a physical model of a house that exists between
them), whereas in computer-supported interaction the computer is always an
external entity (e.g. a spreadsheet figure on a projector wall in front of CEOs
in a business meeting). Communication space and task space separation results
from the fact that simple projector or display based systems do not effort
another kind of interaction. Their effectivities for sending and receiving
communications along the channels to which we are naturally accustomed do not
match human affordances. As the human being is highly flexible it adopts to the
situation, spending extra effort on changing his interaction mode and giving
away opportunities to interact more expressively and meaningfully.
Summarised
affordance theory applied to interface design means that classical interfaces
simply didn’t afford human like interactions. These environments and systems
were not designed to support or encourage humans in a human like way. User
frustration is common place e.g. with the Office Assistant in Microsoft’s
Office Package (Doyle 1999) and may discourage, or even prevent
interaction. To improve UI functional design could help to point out what
affordance properties an interface has – it should look like what it efforts
(e.g. a humanoid agents efforts human like behaviour and interaction), not hide
what it efforts. As we could learn this does not guarantee the perception or
the effectuation of affordances. Nothing can force the user to employ the
offered means and communication channels. He/she must become or be made aware
of the extended opportunities. But at least there are the chance and the
promising advantages of easier, human-like interactions over the broad
bandwidth of human expressions.
Like with cognitive
ergonomics the result is a reduction in workload that can free the person for
higher-order cognitive tasks such as planning, decision-making, supervising,
and co-ordinating with other agents – both human and artificial.
[1]
Following summary composed from handouts in course
PSY450, University of Canterbury, Dean Owen, 2003