Re: [xmca] New Valsiner SEmiots paper on MCA website at lchc

From: Kellogg (
Date: Mon Nov 20 2006 - 00:29:11 PST

Thanks, Steve, and other painting fans out there. If there were enough pixels on the site to show it, you would realize that my grasp of technique is, alas, very far from stunning. There is a good reason why I do landscapes and not portraits: a tree stump never complains when you can't get the likeness right.

It seems to me that there is also a good reason why Jaan Valsiner deals very briefly with realist Dutch paintings and then discusses monuments and architecture. There are three key "tricks" to Dutch realism which have nothing to do with pleromatization and everything to do with why the kids in my data want to say what they know how to say rather than know how to say what they want to say. In particular, I think the Dutch realist "tricks" explain a lot about why the kids tend to underspecify their verbs and overspecify their nouns (or, if you profer, undergeneralize nouns and overgeneralize verbs).

The Dutch found out about using multiple translucent layers of oil paint before almost anybody else, and they perfected it. Now the received dogma is that doing this allows you to simulate the layers of human skin, and for a while I really believed this (and wasted many days painting bones and muscles, then cover them with a blue and red network of veins, then white fat, and finally skin). But what it really does is to superimpose many approximations, like those Galton photographs that Vygotsky rejected as the basis of conceptual thought.

In the same way, the kids go for fluent simplicity, superimposing multiple approximations of verbs rather than trying to get exactly the right verb. With nouns they focus on objects, and generalize; but with verbs, they do the opposite, starting with a general relationship and then trying to specify it.

The second trick is that when the Dutch masters do landscapes, they give you a thin strip of actual landscape (maybe the bottom quarter or third of the canvast) and the rest is all sky and clouds. Clouds are even more uncomplaining than tree stumps--because they are fractal in nature, big and small, far and near, they look the same, so nobody complains when you don't get the perspective right (for the same reason you can never tell how far away clouds are from a plane window). Two thirds of the canvas, then, is invariant under error.

In the same way, the kids need a language system that is invariant under error, and the way they do this is to try to repeat something they've heard, because even if they repeat it wrongly they know it will be understood by someone who heard the original model. They go for concrete objects in their environment, and since verbs are not concrete objects, they repeat the ones they hear most frequently.

But the third Dutch trick is the CONTEXT--every picture has a frame. Within that picture frame, what counts is not detailed representation at all--that is an illusion. It is the relationships that matter. Claude Lorrain proved this by painting a realistic picture of the sun. This should be impossible, because you cannot look at the sun, whereas a picture is made to be stared, nay, gaped at. But Claude's sun is realistic, because realism in painting involves the relationships between shades and colors and shapes and sizes and not the actual representations.

In the same way, the kids see verbs not as verbs in the way we think of verbs, but rather a way of establishing realistic relationships between the nouns in their noun-filled world (and this explains their overwhelming preference for the "X is Y" formula). Again, this is a form of schematization, an impoverishment....

...unless we accept that the world really is ineffable and unrepresentable (I think this is what the word "overwhelming" really means), and that it really is out there. To accept this is to accept that any attempt to depict it realistically in painting (and in grammar) is not a pleromatic embellishment, but rather an iterative form of schematization.

But this does not seem true for monuments, which have to be out there in the world, and which would necessarily suffer by comparison if they tried to be schematic representations of their surroundings. Not one of the Dutch painting tricks (or my kids' versions of them) will work when you are designing a cathedral entrance or a gargoyle-lined parapet.

David Kellogg
Seoul National University of Education

xmca mailing list

This archive was generated by hypermail 2b29 : Wed Jan 03 2007 - 07:14:21 PST