If visual images had a grammar or, at least, any structural system for human communication as Kress and Leeuwen said, we would strat to learn its rules in elementary schools, I guess. I do not mean that K&L are in wrong thoughts. Rather, this is a very fresh trial which grammatically structuralizes visual images as languages.
Only they didn't attempt such work; Christian Metz, a film theorist, also mentioned film has a language earlier. In Film Language: A Semiotics of Cinema, his book, Metz focuses on narrative structure, proposing the "Grand Syntagmatique", a system for categorizing scenes (known as "syntagms") in films (from wiki, http://en.wikipedia.org/wiki/Christian_Metz_(critic)). He argued there are the unified and identified patterns in films that are kind of visual arts. My point is that he did the work in a film which is a limited visual field. K&L tend to generalize visual images in a broad extent.
LANGUAGE:VERB::VISUAL:VECTOR
LANGUAGE:NOUN::VISUAL:VOLUME
This analogies, K&L suggests, may be applied to visual images for mainly communicative purposes. For example, it is hard to just say that theoretical models such as Shannon and Weaver's model (already used as an example by K&L) are good represetatives to explain visual structures. This is because these are designed to account for their logical thinking well. In other words, the models are to translated thoughts built by languages into visuals. I think all graphics for communication are initially designed from linguistic thoughts.
I think it fundamentally differs in the role of languages and visuals. While lanaguages are communication, visuals are representation. It seems to be similar but it's so different in their purposes. Communication is based on reciprocal behaviors, so that it is important what is conveyed to interact each other. But, representation is literally expression itself for one who expresses something, so that it is very individualistic and subjective. Maybe it is a tool to represent unconsciousness excluded from logical thoughts. Then, structural systems between languages and visuals cannot be camparable in any way at all..... :)
