The Connection between Embodied Cognition and Learning: 3 Examples from Physics Education

When I started this blog 8 years ago, it was described as ‘eclectic’.  Part of that is because, like most blogs, it is a slow form of stream of conscious, blogging about stuff that interests me.  But also as a researcher you look for theoretical connections between things that on the surface may look very different.  One such connection that has been a focus of some of my research is the application of embodied cognition research and theory to explain various anomalies in educational research and new techniques for instruction and educational technology, as described in a recent post about an upcoming AERA symposium on embodied cognition and education I am organizing.

For example, researchers have found that attending to student gestures or using gestures while explaining concepts or procedures (for example in a math class) helps student understanding, and also having students interact with and physically manipulate models (such as acting out a story or physically manipulating a simulation) helps student reading comprehension or physics understanding.

But the first time I across this connection between embodied cognition and learning was 15 years ago when working on an undergraduate thesis about physics misconceptions (“intuitive physics”).  I wrote a review of research on the area.  This is actually a very broad area, so it ended up being a massive task for me to review it.  Well, massive for an undergraduate.

The specific anomaly that I came across involved a test question about dropping an object from a plane.  In this problem you see a diagram or animation of an airplane traveling from left to right.  The airplane drops an object, say a heavy ball or box or bomb or whatever.  The task is to draw or identify the “correct” path the object takes as it falls to the ground.  The student is usually supposed to ignore air resistance, but that doesn’t really make much difference to student answers.  See the crude diagram below:

Airplane drop problem

The misconception (identified by Michael McCloskey and other researchers in the 1980s) is that a significant percentage of students think the object drops straight down.  In fact it follows a path like the solid line in the diagram.  It didn’t matter the form of the question, be it a diagram or animation, the misconception is still seen.  The theory, or explanation, for this misconception was that this is a visual perception error or illusion based on our past visual experience.  From the perspective of the airplane (imagine you are a bomber in the plane), neglecting air resistance, the object would appear to fall straight down from your perspective.

What is the anomaly in this research?  In the animated form of this problem, students watch the object fall along one of these paths (both correct and incorrect) and then are asked to draw the path that the object took.  In all the drawings of students with the misconception, they drew the object as falling “behind” the actual path it took.  So for example, after the “correct” animation, students drew the object as falling straight down, and after the ‘straight down’ animation, some students even drew the object as falling backwards (moving to the left).  The anomaly, however, is that when students were shown an animation of the object falling ahead of the plane (the red dotted path above), students drew the path correctly, with no misconception.  Suddenly, they “saw” the path correctly.

You could still explain this with the visual perception theory by adding a new constraint.  Perhaps objects that move ahead of another object are visually segregated from the other object and no longer perceived from the perspective of the other object.  Another theory though is that in this case the object appears to “shoot out” from or be “thrown” from the plane, and not just passively dropped.  This is similar to how we drop wadded up paper balls into trash cans, for example.  We don’t walk by the trash can and passively drop it, calculating the relative velocity and height to get it in the trash can.  We throw it in the can.  We actively control it.

This research is from the 1980s, before embodied cognition became known, but there were (and are) perceptual-motor theories of perception that can perhaps better explain this and related phenomenon, such as for example the motor theory of speech perception.  These theories revolve around the idea that what and how we visually (and aurally) perceive are connected to our actions and embodied capabilities and embodied experience, not just purely visual experience.  For example the McGurk effect: you watch a video of a person visually speaking one thing, but the audio is of something similar but different.  We tend to hear what the person is visually pronouncing with their mouth, or at least our auditory perception by what we see the person physically saying.

A second physics education example is perhaps a little more clear.  This is the curved tube problem, shown below:

Curved Tube Problem

A ball travels through a curved tube and exits out the other end.  The misconception is that some students believe the ball will continue to travel in a curvilinear path after it exits the tube.  From Newtonian physics, we know the ball show travel in a straight line absent external forces.  This misconception can’t be explained as a purely visual perception error.  We have no visual perceptual experience or perspective that would explain this misconception.  But from motor experience, we have experience controlling the motion of objects with our actions.  We sometimes believe we can continue to control or influence them even after contact.  Such as leaning your head or body when playing a videogame or playing pool, or there is the classic video/photo of the hitter waving his arms to make a baseball stay fair:

Carlton Fisk waving arms after homerun

And indeed when participants were given another version of the curved tube in which they manually controlled the ball (or in this case, a puck), there was evidence for this.  A curved path was drawn on top of a table.  Participants were to push or “shoot” a puck through this path without it touching the lines.  You could do this by pushing it diagonally through the curved path, but many “wound up” the puck by moving their arm and hand in a curved path before releasing the puck, with hopes it would continue curving through the path.

A third example of the role of embodiment in physics conceptions is from what was called microcomputer-based labs (MBL).  This is when sensors (such as optical distance sensors) are combined with computers to allow things like pushing a car back and forth along a track, and the computer graphs its motion in real-time (position vs. time, or velocity or acceleration versus time).  It turns out this is an extremely effective and fast way to help students understand how to interpret graphs of motion.  Before, many students have a “graph as picture” misconception.  If students are shown a graph like the one below and asked to describe what the car whose motion it describes is doing, many might say that it is a graph of a car going up and over a hill.  Actually it is a graph of a car moving faster and then slowing back down to the original speed.


Graph of velocity over time


Research has shown that if students can physically manipulate the motion of the car and see the graph change in real-time, they learn the concepts very fast (less than 20 minutes in some studies).  If however, the feedback is delayed (by as little as a few seconds) or if students watch a video of the car moving instead of actively controlling it, it becomes much less effective.

In more generally-relevant recent research on animations/videos and diagrams, people are finding that animations and videos depicting dynamic processes aren’t inherently more effective than static diagrams for learning purposes.  In fact on average, diagrams have a slight edge.  Part of this is because with diagrams you can take your time, explore and revisit different parts of the diagram, “mentally animate” what’s going on, and so forth.  When watching a video or animation, it may go too fast for you (or too slow), you might miss or not understand part of it.  However, what has been shown to be even more effective than either option (video/animation or static diagrams) are user-controllable diagrams (or animated, controllable simulations).  If you let students control the movement of an object, for example, or the changing of a variable, their scores and other measures of understanding are much higher than from passive animations or static diagrams alone.

There are plenty of other examples of the role of embodiment in physics education (like understanding pulley systems), reading education, music education, math education, and other areas of science education (like biology), etc.

Posted in embodiment, learning sciences, research
3 comments on “The Connection between Embodied Cognition and Learning: 3 Examples from Physics Education
  1. Hey Doug, thanks for the article. I was brought up in the cognitivist tradition about the same time as you were. At times, I was lost in blind alleys pushing abstract representations around to create meaning. Neuroscience and the theories embodied cognition are liberating much of this, and as a sim designer, are a great help to me.

    I recently had the pleasure of operating a high fidelity flight simulator – talk about embodied cognition! Move this and see and feel the consequences, fantastic. In the meantime, we’ll keep trying to make these embodied experiences availlable to the mass market of learners.

    I’d love to follow up on a few of the references you mention for my own presentations – can you point me in the direction of these studies you talk about:

    “also having students interact with and physically manipulate models (such as acting out a story or physically manipulating a simulation) helps student reading comprehension or physics understanding”

    “However, what has been shown to be even more effective than either option (video/animation or static diagrams) are user-controllable diagrams (or animated, controllable simulations)”


  2. edtechdev says:

    Yeah I was too lazy to include the citations 🙂

    For reading, see Arthur Glenberg’s work:

    For user-controllable diagrams and animations vs. diagrams, see Richard Lowe’s work:

    and publications by Chan & Black:

    and this previous post I forgot about:

Comments are closed.


Enter your email address to follow this blog and receive notifications of new posts by email.

Join 4,349 other followers

%d bloggers like this: