∆ • Perceptions of the Heider-Simmel Film • ∆

Here's a follow-up to the Karl Heider event of May 1st, and my digression about his father Fritz Heider (1896-1989), a prominent psychologist who once created a little film of a big idea. As Laura Kissel reported in her comment (below), a 16mm print of the untitled work now sits on her desk in Columbia. (I guess the print is being retired, as it is now 65 years old.) And there's a good backstory, as well as a telling tale of the unexpected ways in which orphan films can serve noble ends.

In 1944, professors of psychology Fritz Heider and Marianne Simmel published an article: "An Experimental Study of Apparent Behavior," American Journal of Psychology 57.2 (April 1944): 243-59. There they describe how the piece was made:

The film, one frame of which is shown in Fig. 1, was produced by a trick-film method. The geometrical figures were cut from cardboard and placed on a horizontal translucent-glass plate illuminated from above. A mirror below the plate threw the image into the camera. After each exposure the figures which were to be shown in motion were moved a short distance, then another exposure was made, and so on. (p. 245)

The study is now widely cited as a classic experiment in cognitive psychology. It's even made its mark on media studies. One scholar uses it when addressing how viewers cognitively make sense of edited film sequences. Are these abilities cultivated as we are "thoroughly schooled in the grammar of cinema"? Or "are they endemic to our way of perceiving the world?" asks J. Yellowlees Douglas. In "Gaps, Maps and Perception" she describes how Heider and his student Marianne Simmel "projected an animated film featuring a small moving triangle, a small moving circle, a large moving square, and an empty rectangle to viewing subjects who unanimously described their movements in animate, causal, or intentional terms." ["Gaps, Maps and Perception: What Hypertext Readers (Don't) Do," Perforations 3, 2.3 (199?), citing Jerome Bruner, Actual Minds, Possible Worlds (Harvard U Press, 1985), 18.]

More specifically, in 1943, Heider and Simmel projected the film (a silent, 16mm, b/w animation) for 114 undergraduate women at Smith College. The students were divided into three groups, each of which saw the film twice.

The first group was told simply "Write down what happened in the picture."

The second "was instructed to interpret the movements of the figures as actions of persons." After the screening, the viewers wrote answers to ten questions. Nine followed the example of the first: "What kind of a person is the big triangle?" (Answers included aggressive, warlike, irritable, dumb, stupid, ugly, shy, sly and quick to take offense.) The final item on the questionnaire was quite cinematic: "Tell the story of the movie in a few sentences[!]." Said movie lasted only 2 1/2 minutes, by the way. Tough to write coverage for that.

"In the third experiment," H&S reported, "the same picture was shown in reverse." So 44 of the women saw the film projected backwards! Presumably without being told that the projector was running in reverse. More on this in a moment.

While Laura Kissel and Karl Heider have located a 16mm print, she points out that digitized versions of it appear on the web now. Yale visual anthropologist Karen Nakamura showed a digitized version during her May 1st tribute to Karl's work. The animation still does its work, and on viewers other than college sophomores. Or so the reactions to Karen's screening seem to indicate.

On the oracular YouTube, it's tagged "Social Illusion" by 18-year-old uploader and college student TressaJT. In 2007, she added a soundtrack of her roommate narrating the action while watching the movie for the first time. Smart. Tressa's summary of the experiment reads:

This film by researchers Heider & Simmel (1944) simply portrays geometric objects moving around a 2D plane, but people will usually describe the film in anthropomorphic terms, i.e., as as [sic] the social interactions of three human-like characters who posses [sic] personalities, emotions, and intentions.

A more "advanced" application of the film appears on
Anthropomorphism.org [!]. Educators at Carnegie Mellon University created this as a user-friendly site to teach concepts in psychology. It's both surprising and gratifying that the project "re-purposing" this orphan film is funded by National Science Foundation grants for "Cognitive and Social Design of Assistive Robots" and other research in human-robot interaction.

Sara Kiesler, a professor of computer science and the site's author, constructs a lesson asking "Under what conditions do we attribute humanlikeness to nonhumans?" Part of her answer is "Social Context Cues." Here is Kiesler's description of the movie.

Fritz Heider and Marianne Simmel made a silent animation on motion-picture film (1944) in which two triangles and a circle move against and around each other and a diagram of a house. Virtually all people (except for autistic kids) make up a social plot in which the big triangle is seen as an aggressor. Speed or slow the animation, or run it backward, and the illusion of humanlikeness disappears.

Curiously, her last sentence misstates the findings of the 1944 experiment. Heider and Simmel reported that "all but two" viewers of the reversed projection "interpreted the movements as human actions." Kiesler's references to things not in the original study -- autistic kids, speeding or slowing the film -- suggest that she might be conflating later studies with the one for which our dear film was made. This glitch aside, Kiesler's lessons in anthropomorphism are quite useful and very clearly laid out. The elemental movie works well in this webspace.

One mystery remains about the film vs. the streaming video. On both of the above websites, the video's duration is 75 seconds. That indicates 45 feet of 16mm film transferred at 24 frames per second. If slowed to 18fps (common on classroom projectors) it would still last only 100 seconds, according to Kodak's handy film calculator. But the 1944 publication says the film ran "about 2 1/2 min."

Either (a) Heider and Simmel meant that their movie lasted a total of 2 1/2 minutes when run twice, or (b) some original footage is missing from what we see on the web. The former seems more likely. Perhaps we will learn more when Documentary Educational Resources (which distributes Karl Heider's ethnographic films) transfers the 16mm print arriving from Columbia.

For those who can't wait, there are "remakes" (in color) available on the web. Here's a Flash animation replicating the form of the original film. The UCLA Communication Studies department created this for Cogweb, an online research tool for "the study of human cognition" in "communication and the arts."

For further information about this long-lived film and its creator/s one can consult Fritz Heider, The Life of a Psychologist: An Autobiography (U of Kansas Press, 1983) and Fritz Heider: The Notebooks 6 vols., ed. Marijana Benesh-Weiner (Spring-Verlag, 1989).

----------------------------
Other versions:

Der Film zu einem Experiment von Heider

∆ Zeichentrickfilm (2min, 21 sec.) [I thought this might be the uncut original at 2 1/2 minutes, but the 2:21 label is not accurate. This version runs only 47 seconds.]

∆ Another derivative remake is this QuickTime by Emre Yilmaz, posted on Prof. Michael J. Black's Brown University site.

∆ Same as above, but with Swedish commentary.

∆ In 1993, one study showed children (3 to 5 years old) four videos. One was the original Heider and Simmel film, the other three were reworkings of it, "altered with a digital special-effects generator." Diane S. Berry and Ken Springer, "Structure, Motion, and Preschoolers' Perceptions of Social Causality," Ecological Psychology 5.4 (1993): 273-78.

∆ Elisabetta Zibetti, Elizabeth Hamilton and Charles Tijus, "Contextual Categorization in Interpreting Perceived Events as Actions," Cognitive Science (under revision). Their abstract reads in part:

The goal of the first experiment was to collect the shortest segments, called 'primitives,' from participants viewing Heider and Simmel's 1944 animated film. Participants were asked to segment the film [David Bordwell's cognitive film studies influence? Emphasis added.] and to describe the action segments. In the second, we modified the physical characteristics of the geometric figures in the film and collected verbalizations for the primitive segments that were either scrambled, in the original-order, or embedded in the full-length film. The results show that . . . participants construct very different stories about objects when these are made to vary only in shape and size.

∆ Here Mr./Ms. Square, Circle and Triangle show up in a 1996 text book.


Vicki Bruce, Patrick R. Green and Mark A. Georgeson, Visual Perception: Physiology, Psychology and Ecology 3rd ed. (Psychology Press, 1996), 318.


∆ Ashley Buia, Narreh Ghazarian, Annie Kimbell, Amanda Vinsant and David Yoselevsky," Attribution & Emotional Intelligence: Intrepretations of the Heider-Simmel Film" (2006).
Below: Their poster summarizing an experiment with a short version the Heider film, conducted by 5 students at Clark University (under Prof. James D. Laird). They showed a one-minute version of the video (at both 'fast' and 'slow' speeds) to 39 undergraduates.
www.clarku.edu/activelearning/present/asd/2006posters