After my talk at Comic-Con I was delighted to be able to chat with Scott McCloud for awhile. He actually came to my talk, and I think it was one of the first times he's ever seen me present my research live, despite our having known each other for about 15 years now.
One of the things we talked about was the use of scientific method to study the visual language of comics. My approach of course is based on a scientific view of this research.
McCloud's work often gets flack from various people within academia for not being theoretically rigorous enough, being ignorant of "the literature" that came before it (mostly outside of English), and for essentially not being "scholarly" enough.
However, I think that McCloud's ideas—as basic as some of them are—are perhaps the most scientific approach to comics up until that time. In this way, this work distinctly separates itself from other theories, like Groensteen's System of Comics, which are distinctly un-scientific.
What I mean is that McCloud's theory is based out of a desire to understand how certain aspects of comics were different in the world. Unless I'm mistaken, McCloud's theories greatly grew out of a desire to explain how the things he was seeing in Japanese manga were different from those in American comics. This was the observation that provided a problem, the theory provided the solution.
But, here's the key thing: he then used his theory as a tool to actively quantitatively analyze the properties of actual comics. This is especially salient in his cross-cultural comparisons of panel transitions. He actively used his categories in a useful way to 1) confirm the hypotheses of his original observations, and 2) to illuminate aspects of the medium in interesting ways.
This was what was so novel and interesting about McCloud's book when it came out. Regardless of whether he got the actual theorizing right, or if his work paid lip service and connected to a scholarly tradition, the most important insight was in its methodology. Very few approaches to theories about the visual language of comics have done this prior to McCloud, and very few other than mine have done it since.
On this point, I should mention that my work has largely grown out of this same trend. Besides the empirical experiments and data collection, my theory has grown by looking at actual examples and then applying the theory—seeing where it goes wrong—and then revising the theory.
In fact, this is how I first noticed that I needed to expand the number of transitions beyond McCloud's six, and eventually how I realized that transitions did not work and that an alternative approach was necessary: I was trying to apply McCloud's transitions to actual comics using the same type of quantitative analysis that I found in Understanding Comics.