Semiology of graphics

by Jacques Bertin

Originally published in French in 1967, Semiology of Graphics holds a significant place in the theory of information design. Founded on Jacques Bertin’s practical experience as a cartographer, Part One of this work is an unprecedented attempt to synthesize principles of graphic communication with the logic of standard rules applied to writing and topography. Part Two brings Bertin’s theory to life, presenting a close study of graphic techniques including shape, orientation, color, texture, volume, and size in an array of more than 1,000 maps and diagrams.

Genres: Design, Nonfiction, Cartography, Art Design, Art, Reference, Philosophy

415 pages, Hardcover

First published January 1, 1967

Reviews

[Andy Chu · Rating 4 out of 5]

I managed to find this rare book in the Palo Alto Library a few years ago. It's like a precursor to Tufte's books, but it's far more detailed and dense.

It's been awhile, but I believe most of it is essentially covering theme of accurately representing data with graphics.

[Michael Scott · Rating 5 out of 5]

TODO full review:
+++ One of the classics of information visualization cited by the classics of information visualization, Jacques Bertin's Semiology of Graphics (Sémiologie graphique: les diagrammes, les réseaux, les cartes) was written in 1965 and published in 1967. Although this book was translated in English only in 1983, from a second-edition print (1973), Edward R. Tufte includes references to it in his own classic, The Visual Display of Quantitative Information.
+++ It is very interesting to understand how a scientist would take on providing a unified theoretical structure for the field of information visualization; in short, it's experiment after experiment, exploring the vast design space with determination and creativity. The author asks many of the important questions: what can we show with existing plot types? How to construct a good variant of a known plot type? What kinds of questions can the viewer ask from each type of plot? What factors affect perception, per type of plot and per type of question? Can we change color, etc.? Can we combine factors? Can we combine types of graphs? How does information need to change, depending on the level of detailed matching the question? How to communicate the information visually, so that it becomes memorable (as opposed to a recording that replaces the memory)? Etc.
++/- The Semiology of Graphics proposes a grand theory of visualizing information through diagrams, networks, and maps. Although parts of this theory have since been invalidated (for example, Tufte is sternly against hatching in filling forms and against radial graphs of all shapes and kinds, including the spider graphs on p. 195/200-1, and Scott Berinato reminds us of how the 2010s have changed assumptions about visual perception and especially color-related perception), it is very interesting to see what has stayed with us (notions of how value, color, form, orientation, scale, and even grain/hatching affect the eye).
++/-- The author also proposes a new set of symbols, some of which have loosely mathematical meaning, meant to describe the various ways graphs are created and their components composed. This results in a rather cryptical structure, and also in tabular summaries that are often difficult to relate to the plots they describe, especially maps. (But I recall reading as a kid many maps encoded using this kind of approach, and liking them to bits.) The structure is used to map some of the classical visualizations, such as Charles de Fourcroy's 1782 plot Tableau Poleometrique (p.202-3) and William Playfair's 1805 plot Inquiry into the Permanent Causes of the Decline and the Fall of the Powerful and Wealthy Nations (p.222).
+++ Plot diversity: some plots are rare today (e.g., the triangular construction on p.232-3, curve superposition into a multiscale plot on p.235, the multi-graph combined with multiple other variables on p.281), but we also see great ideas about some of the more modern designs (e.g., the table in the right-hand column on p.236 seems to be a kind of sparkline, the ordinal comparisons on p.248 are now modern rank-change plots, map profiles are still drawn this way, the multidimensional problems on p.254-68 are proto-big data treatments, the graphs on p.277 are on par with the latest radial displays of genetic knowledge, the vectors used to depict winds in different spots on the map are still used in meteorological maps).
+ At the end, there is the lament about the sore state of education, which, according to the author, gives much too much time to learning the national language and the art of drawing, but allocates nothing to learning information visualization. The situation remains largely the same today, in France as in other countries. (This is no indictment of the arts, just a pragmatic concern about the future citizens and voters in these countries.)
- Perhaps the only negative point: this is not a book for the beginner in information visualization. Start with Stephen Few and Nathan Yau, work your way up to Edward R. Tufte and John W. Tukey, then breathe deeply and dive in.
(+++ Reading this in French doubled the fun.)

[Gabe · Rating 2 out of 5]

I only skimmed this book. It turned out not to be quite what I was hoping for. I was interested in the book because, as a student in scientific research, I have to prepare graphical representations of my data, and I wanted a guide to understanding the principles behind constructing graphs and graphical representations. Apparently this book is one of the classics of the subject, but I found it surprisingly confusing. It's a translation from French, and there were plenty of sentences that just seemed overly complicated. The fact that the title uses an uncommon, academic sounding word should be a hint that the whole book is kind of like that. I was especially interested in the typography of graphics, and the book doesn't seem to give this topic much attention. Some of the pictures are neat, and it does make you think about things like which dimension of graphical differentiation to use to display data, but I don't think I'll be using this book much for guidance.

[Kenny · Rating 5 out of 5]

Probably one of the best books to start with if you're interested in visualizing data. It covers the fundamental principles of visualization very detailed and thoroughly. It's a bit outdated, since it's from 1967, so don't expect stuff about interactivity, e.g. zooming, filters, highlights, etc. But other than that, it's a brilliant book that I recommend to everyone interested in graphics and visualization.

[Synaps · Rating 5 out of 5]

This is a manual combined with a style guide. It unpacks every aspect of visual communication on the basis of what is proven to work best for the human eye. The sheer amount of guidance makes both for a difficult read and a perfect reference book to go back to when in doubt.

[Georges Hattab · Rating 4 out of 5]

A comprehensive book and to the point. Pretty quick read if you're familiar with the field. However, more depth would have been appreciated in many instances.

[Francis Harvey · Rating 3 out of 5]

A book of its times; a book for knowledge archaeology today. The graphical variables, for Bertin, mediate the mind/body divide

[Emily · Rating 5 out of 5]

An essential.