VIRTUAL WORLDS AND THEIR ROLE IN INVESTIGATING CHANGE IN COGNITIVE…

VIRTUAL WORLDS AND THEIR ROLE IN INVESTIGATING
    CHANGE IN COGNITIVE MODELS OF MOTION

                               Denise Whitelock
                          IET, The Open University
                 Walton Hall, Milton Keynes MK7 6AA, England
                      E-Mail: d.m.whitelock@open.ac.uk

                                Simon Holland
                            Computing Department
      The Open University, Walton Hall, Milton Keynes MK7 6AA, England
                        E-Mail: s,holland@open.ac.uk




                                    Abstract
   There is considerable preliminary evidence that virtual environments are
   highly effective for allowing participants to internalise data and acquire new
   understanding. However little principled empirical work appears to have
   been carried out on the effectiveness of virtual reality in environments for
   such purposes. A systematic approach is explored to the study of the
   effectiveness of virtual environments for studying change in students' mental
   models of motion.




                          This paper appeared as
   Whitelock, D. and Holland, S. (1992) Virtual Worlds and their Role in
   Investigating Change in Cognitive Models of Motion. In Colloquium Digest
   of the Institute of Electrical Engineers' IEE Colloquium on Using Virtual
   Worlds. Pages 2/1-2/5. IEE Electronics Division, Savoy Place, London.




                                    1
VIRTUAL WORLDS AND THEIR ROLE IN INVESTIGATING CHANGE IN
COGNITIVE MODELS OF MOTION

D. Whitelock and S. Holland

1 INTRODUCTION

Previous research into children's ideas about physical phenomena has shown that these ideas
are very different from those of the scientist. There is a large body of research data, mainly
descriptive accounts of children's reasoning, which is not always easy to interpret. (For
extensive overviews see Driver and Erickson (1983) and Gilbert and Watts (1983)). There
is also a difference of opinion about how these conceptions should be viewed: it is not clear
whether children's ideas in dynamics should be described as systematic mental structures or
as ad hoc temporary constructions.

Following the work of McCloskey (1983) and Whitelock et al (1989) it is our belief that
commonsense ideas about motion can be represented as causal models which are different
from the Newtonian view of motion. Students' failure to answer dynamics problems
correctly (as demonstrated by Caramazza et al 1981; Clement 1982; diSessa 1985; Larkin et
al 1980; McDermott 1984; McCloskey 1983; Viennot 1979; White 1983 and many others)
reveals a lack in their knowledge of the causal principles that underly the formulas they have
been taught. We believe that experiences that enable students to acquire accurate causal
models will make them less likely to develop misconceptions, and better able to understand
the principles underlying Newtonian mechanics.

We hypothesise that a sequence of pre-designed activities carried out in a virtual environment
with a head-tracking head-mounted display and a dataglove will lead students to build up an
accurate qualitative model of the laws of motion more effectively than conventional computer
simulations or instruction using physical apparatus such as found in schools. We propose to
build a collection of virtual worlds designed to aid particular stages of acquiting accurate
causal models of motion. In order to understand the prima facie evidence for our hypothesis,
we need to consider theories of how commonsense conceptions of motion arise.



2 WHY DO PEOPLE HAVE TROUBLE LEARNING DYNAMICS?

People have difficulty understanding dynamcis, even undergraduates and post-graduates
(Viennot 1979; McCloskey 1983; diSessa 1985) and their misconception are diffiuclt to
change. Evidence suggests that people use a non-Newtonian framework to understand
motion.

McCloskey's work (1983) led him to suspect that intuitive beliefs about motion play a role,
not only in people's thinking about hypothetically moving objects, but also in their
interaction with moving objects. These intuitive beliefs "appear to be grounded in a
systematic intuitive theory of motion that is inconsistent with fundamental principles of
Newtonian mechanics". In order to investigate systematically the relation between beliefs
and actions, McCloskey devised two types of tasks to see if intuitive ideas about motion
might influence how high school and college students performed the tasks. One of the task
was to investigate, through actions, intuitive beliefs about the motion of dropped objects.
The other task was to investigate subjects' ideas about circular motion. Subjects were asked
to push a small object across a table so that it would pass through a 90 degree segment of a
large circular ring-shaped area marked out on a table.

Simon Holland, Department Of Computing Science, Kings College, University of
Aberdeen, Aberdeen AB9 2UB. Email, simon@uk.ac.abdn.csd.
Denise Whitelock, Institute of Educational Technology, Open University, Milton
Keynes, MK7 6AA Email, dm_whitelock@uk.ac.open.acs.vax.



                                              2
In the first task, undergraduate students were asked to walk across a room and while they
were walking to drop a golf ball so that it would hit a target marked on the floor. On 45% of
the students acted as if they knew the ball would travel forward as it fell. 49% acted as if the
ball would fall straight down and land immediately below its point of release, while 6% acted
as if the ball would move backwards as it fell.

In the second task, 25% of the subjects moved the puck in a curved path before releasing it,
appearing to believe that it would continue to curve after release and so would follow the arc
of the ring segment. These subjects were surprised when the puck failed to curve.

McCloskey suggests that the undergraduates viewed circular motion as not being
fundamentally different from motion in a straight line. Both forms of motion are generated
by imparting the appropriate impetus to an object. However a pupil working with a
Newtonian framework would realise that for an object to move in a circle, it must be acted on
constantly by an outside force that tends to deflect the motion from linearity.

McCloskey's methodology is an important and useful contribution to our understanding of
subjects' intuitive ideas in dynamics. He believes there is some structure to these intuitive
notions. His studies suggest that students' errors are systematic, and he remarks that their
intuitive theory bears a striking resemblance to the pre-Newtonian theory of impetus (an idea
previously proposed by Viennot). His technique has removed problems of mathematical
manipulation, interpretation of graphical data and diagrams from the understanding of the
problem. There are, however, difficulties in adopting an historical model to describe
spontaneous reasoning in dynamics, in that the model does not account for all the different
responses found in the literature.

A formal model of commonsense understanding of motion based on work by Whitelock,
Ogborn and Bliss (Bliss et al, 1989) and further developed by Whitelock (1990) gives a
causal structure to this framework, and predicts that commonsense ideas about motion will
be highly resistant to change by abstract argument or presentation. However, the model
suggests that in environments where the effects of gravity, air resistance and friction, etc.
can be varied, directly experienced action upon objects could lead to changes in deeply held
commonsense understandings of motion. The model seeks to look for the primitives of
reasoning about motion and suggests that an understanding of motion is achieved by
constructing a class of stereotypical motions from a set of primitive actions. These include
SEE/HOLD, LET GO and MOVE YOURSELF. Piaget proposed that during the
sensorimotor stage the child uses primitive actions to construct the motion of object, time,
space and causality, i.e. to construct a commonsense theory of motion. Therefore if such
actions can be performed on objects in environments where gravity, air resistance, friction,
etc. can be varied, this could lead to changes in deeply-rooted intuitive conceptions of
motion. A Virtual Reality Environment provides the only practical means to perform such an
experiment.

3 PILOT RESEARCH AT THE OPEN UNIVERSITY.

O'Shea and others (Twigger et al, 1991) have developed a two-dimensional software system
known as DM3 (Direct Manipulation of Mechanical Worlds) in order to help children reflect
critically upon and modify their own conceptions of the laws of motion. O'Shea and others
(Taylor et al, 1991) have also been involved in the educational evaluation of ARK (the
Alternative Reality Kit), a two-dimensional computer-based environment in which the laws
of physics can be modelled and altered. While these environments have shown some success
in helping children articulate their own notions of the laws of motion, it is not yet clear the
extent to which they help children to from stable, accurate, causal models of motion. White
and Horovitz (1988) carrying out similar work in American schools report some success, but
it has proved difficult to produce similar results to date in the UK.

There is evidence to believe that an environment that engendered a stronger sense of physical
involvement and presence might do considerably better. Virtual reality differs from two
dimensional computer simulations in at least three important ways; its use of three

                                               3
dimensions: the very high degree of sensorimotor interactivity: and the compelling sensory
cues that engender a strong subjective sense of "physical presence" and "direct experience"
(Zeltzer, 1990). The importance of this subjective feeling of sensorimotor immersion can be
seen by reference to theories of how commonsense conceptions of motion arise.

We have already noted Whitelock's (1990) formal causal model of commonsense
conceptions of motion, based on the psychological primitives of "support" and "effort".
This theory suggests that commonsense conceptions of notion are acquired through early
sensorimotor activity, and that while they are highly resistant to reasoned argument,
exposure to new sensorimotor experiences, in which objects can be thrown and played with
in the apparent absence of gravity and friction, may give rise to changes in these
conceptions. However, it is not known the stages by which such conceptions might change,
or exactly what their new nature would be. Investigations are needed to discover the extent
that commonsense notions can be effected by new experiences not normally possible in the
real world, and the extent to which such comceptions will be transferable to the real world.



4 RELATION TO OTHER VR WORK
Reviews of work on virtual reality environments can be found in Brooks et al (1988,1990)
and Chung et al (1989). Existing virtual environments which allow professionals to
internalise large amounts of data rapidly include work on: headmounted displays and gloves
for information management (Fisher et al, 1986); the GRIP and GROPE systems for
interactive molecular studies (Chung et al, 1989); the Walkthrough Project (Brooks, 1986)
for virtual building exploration; and work by Iwata (1990) , Weimer and Ganapathy (1989)
and Waldern and Edwards (1985) with applications in industrial design. Kreuger (1990)
and others have written speculative papers on the use of virtual reality in education.

One of the few virtual environments (using a different technology from our proposal, with
force-feedback) that has been systematically used and evaluated for education and research is
GROPE, due to Brooks et al (1990), for work in protein docking and electrical force field
investigation. Users reported a radically improved situation awareness and a new
understanding of receptor sites, force fields and drug docking. Unexpected perceptual
phenomena were observed, all of which worked in favour of the effectiveness of the tool.
There is considerable preliminary evidence that virtual environments are highly effective for
allowing participants to internalise data and acquire new understanding. However, little
principled empirical work appears to have been carried out on the effectiveness of virtual
reality environments for such purposes. The proposed work would be one of the first
systematic studies of the effectiveness of a virtual environment for teaching.

5 PROPOSED VIRTUAL WORLD TUITION

We propose that students should undertake activities in four related microworlds. (Their
physics understanding as a result of their experiences will be measured with performance
scores on a pre and post test). The group using the virtual environment will be invited to
undertake activities in four related virtual microworlds. They will intitially be given time to
acclimatise to the equipment and to each new virtual world. The domains covered are (a)
projectiles in zero gravity with no air resistance or friction, (b) experience of relative motion
(c) motion of projectiles under gravity (d) motion of projectiles with speed-dependent air
resistance.

The first microworld is designed to help develop an intuitive familiarity with Newton's first
law. This environment will simulate the effects of zero gravity without air resistance or
friction. For their first task in this world, students will be set the (impossible) task of
throwing projectiles in such a way that they follow curved paths. This microworld also
permits comparison between motion of projectiles of different masses with impulse rockets
attached. The size or number of impulses can be contolled. As in all of the microworlds,
projectiles may be set to flash at regular intervals, which causes them to leave 'tracer' trails in
space. There will be a facility for traces to leave projections on reference x, y, z, planes.

                                                4
Students may examine the traces from different viewpoints, and may uniformly copy, shrink
and keep traces for subsequent comparison, so that speed-time graphs may be produced.

The second microworld is concerned with relative motion. Students are initially invited to
'drive' in a vehicle toward a given distant target, acoss a platform moving at right angles at
uniform speed. They are invited to predict how their trails will appear from the point of view
of an observer on the moving platform, and from the point of view of an observer at rest
with respect to the stationary target. Other possible activities in this microworld might
involve a platform that moves with uniform acceleration, and space ships in relative motion
in three dimensions. The tasks in this microworld are designed to develop an intuitive sense
of how relative motions combine when viewed from different frames of reference. It
appears to be difficult to foster a good subjective sense of relative motion using conventional
two dimensional computer simulations (Whitelock et al 1991), although we suspect such a
sense may contribute greatly to the understanding the motion of projectiles under gravity.

The third microworld is similar to the first, but introduces gravity, which may be varied in
strength. The motion of projectiles of different masses may be compared when thrown or
fired, using rockets. The fourth microworld is an extension of the third, allowing 'gases' to
be introduced to provide different air resistances. The motion of projectiles of different
masses may be compared when thrown or fired using rockets against air resistance.
Comparisons can be made in the final microworld with gravity on or off. The exact form of
these microworlds may be adjusted or refined before the formal experiments, subject to
experiences in the informal pilot stage during development.

We are aware that the Virtual Worlds Simulation will be far from perfect. The different mass
and weight of projectiles will not be directly felt. Whole-body acceleration forces will not
be simulated. Students will feel the effect of gravity of their own bodies. However our own
preliminary experiences in virtual worlds, suggest that the combination of perceptual effects
that are well-simulated by the apparatus will be strong enough to give rise to the effects that
we hypothesise.

6 CONCLUSION

There is considerable preliminary evidence that virtual environments are highly effective for
allowing participants to internalise data and acquire new understanding. However little
principled empirical work appears to have been carried out on the effectiveness of virtual
reality in environments for such purposes. The proposed work would offer a systematic
study of the effectiveness of a virtual environment for studying change in students' mental
models of motion.


REFERENCES

Bliss J., Ogborn, J. and Whitelock, D. (1989) Secondary pupils' commonsense theories of motion.
 International Journal of Science Education Vol. 11, No 3.
Brooks F.P. (1988) Grasping Reality through illusion: interactive graphics serving Science.
 Keynote address published in CHI 1988 Proceedings Addison-Wesley Reading Mass
Brooks, F.P., Ouh-Young, M., Batter, J.J. & Kilpatrick, P.J. (1990) "Project GROPE - Haptic
 displays for scientific visualisation." ACM Computer Graphics, vol.24, no.4, Aug 1990 pp 177-
 185.
Brooks, F.B. (1986) Walkthrough - A Dynamic Graphics System for Simulating Virtual Buildings.
 Proc 1986 Workshop on Interactive 3d Graphics 9-21.
Caramazza A., McCloskey M. and Green B.(1981). Naive beliefs in 'sophisticated 'subjects:
 Misconceptions about trajectories of objects. Cognition,9, 117-23.
Clement J. (1982). Students' perceptions in elementary mechanics. American Journal of Physics,
 50, 66-71.
Chung, J.C., Harris, M.R., Brooks, F.P., Fuchs, H., Kelley, M.T., Hughes, J.,Ouh-Young, M.,
 Cheung, C., Holloway, R.L. & Pique, M. (1989) "Exploring Virtual Worlds with Head-mounted
 Displays"; SPIE vol.1083 Three Dimensional Visualization and Display Technologies, pp 42-52.

                                               5
diSessa, A. (1982). Unlearning Aristotelian physics: a study of knowledge-based learning,
  Cognitive Science, 6(1), 37-75
Fisher, S.S. (1990) "Virtual Interface Environments" in Laurel, B. (Ed) "The Art of Human-
  Computer Interface Design"; Addison-Wesley, pp 423- 438.
Fisher, S.S., McGreevy, M., Humphries, J. & Robinett, W. (1986) "Virtual Environment Display
  System"; Proc. 1986 ACM Workshop on Interactive Graphics, (October 23-24, 1986) Chapel
  Hill, NC.
Hayes P. (1979). The naive physics manifesto in Michie D.(ed). Expert systems in the microelectric
  age. Edinburgh University Press.
Iwata, H. (1990) "Artificial Reality with Force-feedback: Development of Desktop Virtual Space
  with Compact Master Manipulator"; Computer Graphics, vol.24, no.4, August 1990, pp 165-170.
Kruegar, Myron. (1982) Artificial Reality. Addison Wesley, New York.
Lanier, J. (1989) Personal communication at Nato conference on multi-media interface design,
  Lucca, Italy, Sept 1989.
Larkin, J.H., McDermott J.,Simon D.P. and Simon H.A., (1980). Expert and novice performance
  in solving physics problems. Science, 208, 1335-42.
McCloskey ,M. (1983) Intuitive Physics, Scientific American 284 (4).
McDermott, L.C. (1984). Research on conceptual understanding in mechanics. Physics Today, 37,
  24-32.
Osborne, R.J. (1985). Building on children's intuitive ideas. In R.J. Osborne and P. Freyberg
  (Eds), Learning in Science: The implications of children's science. Heinemann, Auckland.
Pausch, R. (1991) Virtual Reality on $5 a day. CHI 1991 Proceedings, pages 265-270, ACM
  Press, New York.
Rowley,T. (1991) Personal communication at IEE Colloqium on Virtual Reality, Leicester, Italy,
  Sept 1991.
Taylor, J. et al (1990) Discourse and Harmony: Preliminary Findings in a Case-Study of Multi-
  Media Collaborative Problem Solving. Technical Report, IET, Open University.
Twigger, D. et al (1991) The 'Conceptual change in Science' Project. In Journal of Computer
  Assisted Learning 7, 144-145.
Viennot, L. (1979). Spontaneous reasoning in elementary dynamics, European Journal of Science
  Education,1 (2), 205-221.
Waldern, J.D. and Edwards E.A. (1985) A three Dimensional computer graphics workstation.
  Technical report.
Weimer, D. & Ganapathy, S.K. (1989) "A Synthetic Visual Environment with Hand Gesturing and
  Voice Input"; Proceedings of ACM SIGCHI, Human Factors in Computing Systems Conference
  1989, pp 235-240.
White B. (1983). Sources of difficulty in understanding Newtonian dynamics. Cognitive Science, 7
  (1), 41-32.
White B. & Horowitz P.,(1990). Computer Microworlds and Conceptual Change: A new approach
  to Science Education in Boyd-Barrett O.& Scanlon E. (Eds). Computers and Learning. Addison -
  Wesley.
Whitelock, D. (1987) What does the student Know? Investigating a commonsense theory of
  motion in R.Lewis and E.D. Tagg (Eds). Trends in Computer Assisted Education.(Blackwell
  Scientific Publications).
Whitelock, D.M. (1990) How well does a causal model of commonsense reasoning represent the
  structure of children's understanding about motion? Proceedings of the British Educational
  Research Association Conference, Roehampton.
Whitelock, D.M. (1991) Investigating a model of commonsense thinking about causes of motion
  with 7-16 year old pupils. International Journal of Science Education vol. 13, 3 pp 321-340.
Whitelock, D., Taylor J., O'Shea T., Scanlon E., Sellman R., Clark P. and O,Malley C.
  (1991).Investigating Computer Supported Collaborative Learning about Collisions with a Change
  of Reference.Technical Report No139 Institute of Educational Technology, Open University.
Zeltzer, D. (1990) "Virtual Environments:Where Are We Going?"; Proc. 12th International IDATE
  Conference, November 14-16, 1990, Montpellier, France.




                                            6