"Building Blocks of Graphics come in Syntactic Categories"

Graphics -- such as diagrams, maps, charts, and information visualizations -- have often been connected with the concept of "graphic language" or "visual language". Assuming that something like "grammar" is a useful notion in the graphic domain, are such "visual grammars" in any way similar to linguistic grammars, in that they are based on some set of distinct syntactic categories of building blocks, and on rules for the combination of these syntactic categories?

Building partly on the work of Clive Richards and Jock Mackinlay, I propose such a set of distinct syntactic categories of building blocks of graphics, each category with its own combination rules. I claim, and will try to show, that this set of syntactic categories applies to every visual representation of information. Reminiscent of syntactic categories of linguistic constituents, the proposed syntactic categories of graphic constituents can help us to analyze how we interpret and how we create visual representations.