I found this week’s reading very interesting because they, like many parts of this class thus far, made me analyze aspects of artistic and everyday visual perception in a way that I was not accustomed to, and allowed me to see and conceive of many things I hadn’t considered. For this reason I particularly like the two Solso chapters, and I consistently found myself amazed at the distinctions he brought up and the way they were illustrated.
In chapter 7, he outlines the ways that we analyze the space between objects through monocular depth cues. He describes how our assessment of relative size is affected by both size of the image that is seen (bottom-up processing) and knowledge of the object (top-down processing), and how occluded objects tend to appear as complete but layered shapes. I found his examples of shadow particularly riveting; not only does the addition of darkness and highlights create the illusion of depth in a two-dimensional image, but that concavity and convexity can be reversed with the simple reversal of those shadows (by turning the picture upside down, for example). He includes orientation, elevation within the frame, linear perspective, and texture gradients, and makes very interesting claims about how color influences depth perception in relation to atmospheric depth (where colors and shapes become more blurry, pale, and blue as they recede) and the fact that warm colors simply advance to the foreground over cool colors. The church in Milan was a particularly fascinating example of these tools at work (I had to stare at the photograph for several minutes before I could convince myself that it was not in fact three-dimensional).
I ran into trouble with these depth cues, however, in Arnheim, when he claims that the United States map provides unreal depth effects that must be avoided: “...we see that a corner of Wyoming lies on top of a corner of Utah, and that a corner of Colorado lies on top of Nebraska. No knowledge that this is not so prevents us from seeing what we see” (p. 248). As I read this section, I was perplexed, as I didn’t think I had ever experienced the illusion before; of course when I looked at a map I immediately saw it this way, having just read the chapter (for the life of me I cannot now remember if I ever saw the states as overlapping now). What interested me beyond this point, however, as I was looking at the map more thoroughly, was how states like Pennsylvania, New Mexico, Missouri, Mississippi, Montana, Florida, etc. did not provide this illusion to the same extent, unless I tried very hard to make it happen. Firstly, I hypothesized that our general schema of maps (see next paragraph) presupposes that states/countries border each other rather than overlapping, and that they are not perfect, simple shapes. When the states do remind us of more simple shapes (e.g. the rectangle-shaped states of the West), the overlapping illusion is easier to sustain, or for Arnheim, impossible to escape.
Part of this is of course related to chapter 8 in Solso, where he explicates prototypes and how they influence how we think things should appear as opposed to how they do appear. Solso explains that certain schemata, derived from our experiences with the world, are activated when we participate in certain activities, telling us what to expect. In this way, Solso argues, and I am quite inclined to agree, that though all viewers perceive a visual stimulus in the same way (i.e. light is reflected off of it, which falls on the retinas, and signals are sent to the brain), this stimulus will activate different schemata that will result in the stimulus being experienced in a different way. In other words, the semantic value or “meaning” of the stimulus will change.
These schemata become abstracted into prototypes, Solso says, which in turn can influence our conceptualization of art and perception of individual art objects. Particularly, art often serves to violate these prototypes resulting in visual dissonance and psychological tension. In some ways, Solso seems to claim that this is in fact the purpose of much of art, “to demand active participation [from the viewer] in the construction of ‘reality’” (p. 237).
Finally, Solso also revisits the question of artists and scientists, brought up somewhat facetiously in the Wade article (I think?) from a few weeks ago. Solso, however, puts artists and scientists in the same camp, seeking out aspects of perception that excite viewers in some way. He says artists “do not invent art” in the same way that scientists do not invent science: “As scientists discover the laws of the universe that are congruent with mind, artists discover visual images of the world that are harmonious with mind” (p. 257). Of course, “harmonious” can have many different meanings, and I don’t think Solso intends to exclude visual dissonance. In fact, I think he means that dissonant with expectation or not, art (or perhaps “successful” art, just like successful theories) accords with the mind in such a way that it merges with the mind of the viewer and provokes thought.
I think that Tessa's choice of map was really interesting, in that it really showed the depth cues in more way than one. I had also not seen the states as overlapping before reading the Arnheim chapters, and I completely agree with the 'prewiring' that Tessa sites as a cause for this. However, the map she included made it so much easier for me to see the states as occluding one another because of their colors. As Solso says, "warm colors seem to advance while cold colors recede" (213). At first I had trouble visualizing this phenomenon- my thought was that, of course if the warm color is on top of the cold color, it will look as though it is in front. I was picturing a completely blue background with an orange shape in the middle. However, this map, especially where Texas and Oklahoma meet, illustrates the point perfectly- I see Oklahoma as occluded by Texas, as if the pan handle is just created by the occlusion, and more of Oklahoma exists underneath Texas. This is the case for me when I view most of the orange-colored states bordered by blue-colored ones.
ReplyDeleteIn keeping with the Solso reading, but fairly unrelated to Tessa's comment: Another big moment for me in Solso was his mention of the "motion parallax". As Solso says, "motion provides the essential information on which depth is based" (204). I keep a Tumblog which is basically a normal blog but for a few added features, such as a 'dashboard' which displays all the posts of the people you 'follow' in chronological order. Over the past year I have been noticing a real love affair among the people I follow and .gif animations. While many of them convey motion in a way that still seems relatively without depth, I have noticed a definite increase in the number of .gifs that convey depth to a surprising and often very visually appealing degree. Here is one that came up on my dashboard recently that I think really takes advantage of the ability of the .gif format to employ the motion parallax: http://27.media.tumblr.com/tumblr_kwl78ma6aG1qzyas5o1_500.gif The animation doesn't necessarily convey the actual physical motion of the subjects. Rather, the image is of people in basically a stagnant position- but it has such depth that in a lot of ways, we see it as three dimensional. For me, Solso totally explained this internet phenomenon, and it was really satisfying to understand just why people find these images appealing enough to post them on their blogs all the time- they add the motion parallax!
This is another link to a tumblr called "wonder-tonic": http://wondertonic.tumblr.com/post/320915609/medieval-illuminated-manuscript-disasters
ReplyDelete- this post is a really funny take on the issues we've been discussing with regards to perspective and realism in older/ancient art. Not really an enlightening post, but I saw it & thought of the class..