Friedman, Sarah L., and Marguerite B. Stevenson. 1975. Developmental Changes in the Understanding of Implied Motion in Two-dimensional Pictures. Child Development 46:773-778.
This paper reports on an experiment testing how action lines are understood by children. They compared figures with postural cues showing motion (i.e. they look like they're running) with those with motion lines, and those that have polymorphic features (i.e. repeating legs over and over to show motion). They tested preeschoolers, first graders, sixth graders, and college students.
Massive increase in the understanding of “cartoon” conventions (i.e. motion lines) occured between the ages of first grade and sixth grade. (possibly due to developmental reasons... or to increased exposure to comics?). Reliance on postural cues decreases from first grade through college. Understanding of polymorphic representation increases greatly between pre-school and first grade, then levels out. The insinuation is that conventional cues are relied on more and more as people age.
Gross, Dana, Nelson Soken, Karl S. Rosengren, Anne D. Pick, Bradford H. Pillow, and Patricia Melendez. 1991. Children’s Understanding of Action Lines and the Static Representation of Speed of Locomotion. Child Development 62:1124-1141.
A similar study also tested children’s knowledge of action lines to determine whether body posture or action lines contributed more to understanding with seven and nine year olds, and adults. The stimuli showed a running figure (one silhouetted, one photo-like), without lines, with lines trailing it, or with lines behind it. They also had a task where subjects drew their own figures running.
They found that the relationship of action lines to the meaning of locomotion is non-arbitrary, though exposure to drawings using it is necessary to understand its convention. Again, children attenuated more to postural cues than adults did. The younger children did not distinguish between action lines and background lines to the same degree as older children and adults.
Interestingly, it also asked subjects what the lines represented, showing that children contributed meaning to lines where adults did not. Many children gave visible and invisible explanations for the lines – such as “air moving” or “wind.” Adults simply accepted the lines as symbolic representations. In cartoon drawings, adults treated lines as “path-of-movement metaphors,” but for photos were treated as non-arbitrary cues for perceptual movement.
(Updated 8/25): These conclusions are interesting for a few reasons... one, it shows that not all graphic things are universally or transparently understood by everyone. It takes some degree of development to reach that understanding for things that we take for granted as obvious. This is different than some of the findings on speech balloons or thought bubbles.
The fact that the development happened most in pre-puberty aligns with many other developmental changes, like language acquisition. Whether that development has to do with exposure or just age, it's hard to tell.
Also, its interesting that there was a change in how they considered what the lines were — from an iconic meaning (claiming the lines are "wind") to a purely symbolic meaning. This too is consistent with developmental changes in other domains.