The following is a brief overview of the model I am developing to organize a description of complex layout in book design. Many models for graphic description have been proposed in the past; I have tried to integrate existing models into my own framework where appropriate, in part as an effort to help consolidate terminology in this field. Figure 1 is a simplified illustration of the framework; one of the models is it based on is a flexible linguistic model proposed by Crystal and Davy (1969), which allows for “ranges of appropriateness” within categories (useful when dealing with graphic features) and links grammatical features with situational variables, generating a context-relevant analysis. It has been highly modified for application to graphic description, in response to other models and to my experience in examining the books themselves.
The micro-level of analysis deals with the basic component parts of a complex layout. This includes design characteristics such as typographic information (type hierarchy, line length, justification, etc) and elements such as keylines or boxed areas. It also includes descriptions of illustrations, photographs and diagrams. Some relevant models in the area of pictorial/diagrammatic representation include those of Ashwin (1979), Bertin (1968), Twyman (1985), and Richards (2000).
Pictorial characteristics in Figures 2 and 3 can be readily compared at the micro-level: Figure 2 makes use of detailed, close-cropped photography, Figure 3 of heavily outlined, cartoony illustrations and flat colour diagrams. Actual descriptions of these features can be lengthy; one of the goals of this project is to establish which categories contain meaningful information, such that they will have predictive power in identifying “book genres” (after Crystal & Davy, 1969: 41). The final framework would confine itself to these categories rather than providing a total description of a given book.
Micro-level components come together to form ever larger discursive structures on the book page, over a 2-page spread, and over the whole book. The macro-level involves characteristics of groups of graphic elements, up to the boundary point of the 2-page spread. The idea here is that groups of signs are also signs. Groups can be described according to Gestalt qualities (Bruce & Green 1990: 110-114), in terms of their semantic coherence as “rhetorical clusters” (Schriver 1997: 343-4), or in terms of picture/text relationships. Descriptions could also employ Twyman's Schema for the study of graphic language (1979). The Schema is useful in describing internal characteristics of macro-features rather than those of the entire book, where illustrated science texts tend to fall under the same categories.
The diagram in Figure 4, for example, combines fairly naturalistic drawing with flat areas of colour and labels; it contains figurative, non-figurative, literal and non-literal features (see Richards 2000: 97). It is self-contained in comparison with the leaf-breakdown diagram of Figure 3, which is surrounded by text and spans two pages.
I would propose that another quality of design at this level is that coherent design groupings can be nested: groups combine with other groups to form larger meaningful groupings on the book page. Nested and recursive structures have been discussed in analyses of diagrams and maps (Engelhardt 2002: 13-18), which suggests another avenue for analysis: perhaps the book page could also be analyed as a diagram?
Meta-level features have to do with the underlying principles or structures governing the design of the book. They include characteristics such as the complexity of the underlying grid, the degree to which the layout adheres to the grid, and design standards such as the density and layout of graphic elements. This category also includes structural qualities of the book as a whole: organising principles, size, binding, etc. For example, the constellation of images in Figure 2 is only loosely organized around a grid, in contrast with those in Figures 3 and 4, in which the elements fit into the grid like a puzzle. The graphic elements in Figure 4, a book the size of a paperback novel, are fewer and larger in relation to overall page size. These are meta- rather than macro-level features because they refer to universal attributes of the book.
Another way of looking at groupings has been proposed by Waller, who refers to typography as the “macro-punctuation” of a text at the level of discourse (1987: 34). This is because segments of text can be integrated through typographic devices (marginal notes, boxed items, etc.) to modify a written argument. Waller's definition is more writing-centred than the one proposed here: with reference to science texts, illustrations and diagrams can be seen as inherent parts of the content, such that meaning is centred in the whole layout rather than in the purely written text.
There are many references to the semiotic concepts of syntax, semantics and pragmatics in discussions of communication design (Goldsmith 1983, Waller 1988, Sless 1986b, Richards 2000, etc.) In earlier stages of this research, I drew distinctions between the observable forms of graphic language (signs and syntax), the meanings they represent (semantics), and the conditions of reading (pragmatics). A simpler distinction can be drawn, however: that of formal features versus context-relevant features. I will provisionally call the former rule-based and the latter context-based.
Rule-based features are those that relate to systematic or codified aspects of graphic language. In his discussion of graphic analysis, Waller cites a linguistics-based approach that seeks to uncover “logical rules” in design (Waller 1987: 36). He also suggests the term “logic of assembly” to refer to the visual structure of a page imparted by a modular grid (Waller 1988: 238). The internal logic of a book will reflect the constraints of printing technology as well as the rules of language and typography. Context-based features are structures or meanings contingent on social and cultural context, the reader's previous knowledge, and the actual conditions of reading. To clarify the distinction, a parallel can be drawn with Saussure's definitions of langue and parole in semiology: rules are analogous to langue, the social and cultural system of language rules. Context is analogous to parole, the individual instances of actual speech (Saussure 1922/1983:172-173). Many more parallels can be drawn: rule-based structures deal with the logical and systematic, and draw on our internal schemata of pictorial forms (Gombrich 1968). Contextual features deal with the particular, and draw on shared social or historical codes for rhetorical significance. They may relate to underlying features such as the “voice of the author” or the intended role of the book, often covert characteristics. These categories do not exist in an either/or relationship: it is possible for one characteristic to fit under both, for different reasons, because they simply define different ways in which graphic features are meaningful.
The systematic use of a modular grid is a rule-bound feature at the meta-level of analysis. The underlying grid of Figure 5 is simpler than those of Figures 6 and 7, as it contains fewer grid fields. The text in Figure 5 makes less use of typographic cueing than the text in Figures 6 and 7, which are more typographically differentiated through the course of the book: another meta-level feature. In terms of context, the group of detailed pinecone illustrations of Figure 6 alludes to specimens displayed in a sober manner in a display cabinet. The photographs in Figure 7 are also detailed, but represent particular examples rather than the universal depicted in scientific illustration: their dense and random positioning around the type suggests a naturalist's scrapbook. Also, many of the ruling lines in Figure 7 do not meet, suggesting the traditional “look” of letterpress; a contextual feature at the micro-level. This is how the categories of rule and context cut across the three levels of analysis to organise a description of complex layout in books.
The framework described above is meant to be a tool to facilitate an analysis of book design as a form of communication. It seeks to emphasize how graphic communication contributes in a substantial way to meaning, and may also allow the 'graphic language' of books to be described and possibly traced through history.