[See Mohan Matthen, “Why do movie effects get dated?” at NewAPPS.]
Consider two famous stop-motion sequences by Ray Harryhausen: the Cyclops sequence from The 7th Voyage of Sinbad and the skeletons sequence from Jason and the Argonauts.
In the fiction of Sinbad, there is nothing to indicate that the Cyclops moves otherwise than like a large humanoid animal. Its motions should appear smooth like those of humans. But they don’t. The difference is apparent especially when human actors and the Cyclops are presented in the same frame. Harryhausen may have intended that the motions should be realistically depicted, but his technique falls short, especially now that computer animation sets the standard.
The skeletons’ motion as depicted has the slight jerkiness and excessive clarity of stop-motion animation. But skeletons, after all, if they could move, might very well move in an uncanny manner. One can see the sequence as depicting realistically the uncanny motions of skeletons brought to life.
Or one can see it as not-quite-realistically depicting ordinary, smooth, physically correct motions.
Which option predominates will depend on conventions of depiction. The overarching convention that governed the reception of works like Harryhausen’s, at the time of their making was a version of realism (see this 1941 article in Popular Mechanics). In particular, the depicted motions, other than magical, of fictional creatures were to appear as much like those of human or animal actors as possible. Realism, together with an emphasis on technical ingenuity and painstaking labor, has been part of the ideology of special effects from the start.
Also among the cultural norms surrounding special effects is that the spectator is invited to view them with a critical eye. The tagline of Superman, “You will believe a man can fly”, succinctly sums up the challenge taken up by special effects. Conviction will depend not only on the depiction but on the range of instances with which the viewer is familiar. What the critical eye looks for will therefore change as techniques change.
What the critical eye looks for will also change according to the imputed target of depiction. Rotoscoping, a technique in which the figures in a cel animation are traced from a filmed human figure, frame by frame, enables animators to depict human movements much more accurately than traditional freehand animation. Even now the results look more realistic (Snow White, 1937; Gulliver’s Travels, 1939).
But compared with what 3D computer animation can now accomplish, it falls short — if you take the target to be the realistic depiction of the human figure in motion. Shading, for example, is absent or minimal. Spectators in 1939 did not expect detailed shading, and still less a physically realistic treatment of reflection, and the absence of those features did not, therefore, tell against the realism of Rotoscoped animation.
When we watch older movies, we tend, I think, to revise our estimate of the target, rather than judge it to fall short by current standards. Viewers of Pinocchio were said to marvel at the “realism of the shadows, the highlights on jewels and on shiny objects, the mists, the dusts, and the water” (Popular Mechanics, January 1940, 21–22). Judge for yourself: look at the liquids, the smoke, the shadows… Pinocchio may have looked more realistic than its predecessors. It may still look more realistic. But does it approach the realism of today’s animation?
Standards change, targets change. Which standards do I apply? What governs my imputation of targets to animations? They aren’t arbitrary, of course. On the other hand, they aren’t determined by features of the medium or of the human perceptual system. Viewers learn which standards to apply and which targets are likely to be aimed at from critics, from peers — the avenues, in other words, by which one picks up collective norms generally. In that respect the realism of animations is a matter of convention.