Review article of Kress and Van Leeuwen's "Reading Images" (Routledge 1996)

What is it about?

"Reading Images: The Grammar of Visual Design" is one of the pioneering books in visual studies. Kress and Van Leeuwen adapt Hallidayan Systemic Functional Linguistics to present and adapt concepts that can be used as tools to analyse visuals. Pertinent concepts include "offer" pictures & photographs (depicted persons do not look into the "camera") versus "demand" pictures & photographs (they look into the "camera"); the significance of lower versus upper regions of pictures as well as that of central versus peripheral regions and left-right orientations; realistic versus abstract styles of depiction; "vectors" (= imagined "lines" that connect depicted persons with each other or with objects by their direction of looking or their actions). This review article offers a summary of the book, and discusses both what I think is good in it and what I think is bad in it.

Featured Image

Photo by John Jennings on Unsplash

Photo by John Jennings on Unsplash

Why is it important?

The book, whose revised edition was published in 2006, is marketed as a textbook usable at BA level, which suggests that the views expressed present a generally shared "state-of-the-art" in visual studies. Particularly young students may therefore believe the book's ideas to be tried and tested, and thus self-evident, and consequently adopt them uncritically as the basis for their own analyses. Unfortunately the book contains a number of ideas that are deeply problematic. By and large I consider the concepts presented as potentially useful; however, they are too indiscriminately applied, and any "pattern" presented is based on far too few examples to be convincing. Other issues: (1) the authors show little awareness of the degree to which the genre of an image steers potential interpretations; (2) the authors overextend the analogy between language and visuals, which leads to strained and implausible conclusions. Put differently: visuals have a structure, but no "grammar"; (3) the authors underestimate the degree to which interpretations are triggered -- not by visual information alone but by visual information accompanied by written text. That is, they analyse information that is not of a purely visual, but of a multimodal nature; (4) The authors are overly eager to show how their concepts can expose ideological assumptions underlying the visuals they analyse. This bias often conflicts with, and thus comes at the price of sacrificing, methodological precision.

Perspectives