In Praise of Prose
I've just read the excellent The Back of the Napkin, by the visualization consultant Dan Roam. Now, say the words 'visualization consultant' to people and they are likely to recoil slightly, imagining some person or organization churning out fantastically complicated systems diagrams in dense Powerpoint decks. But Dan's book is about representing information simply, and he demonstrates pretty persuasively that most strategic problems can be drawn up pretty simply. (There's a line of thought among strategic analysts that says that the ones that can't arise from faulty strategies.) His cartooning style also makes the book a delight to thumb through.
One of the endorsements on the back of the book troubled me a bit, though. It's from the designer Roger Black, and it says:
Visual information is much more interesting than verbal information. So if you want to make a point, do it with images, pictures, or graphics.
As a linguist you might not expect me to agree. And I don't, but not for purely territorial reasons. You see, while The Back of the Napkin demonstrates that simple visualisation is an effective way of showing the whats, whos, wheres and how-manys of situations and problems, I'm not convinced that it's the best way of addressing the whens and whys. I'm about to undermine my own point by using a bit of visualisation by means of bold text, but syntax - the structural features of language that most of us call 'grammar' - is an extremely good way of encoding processes and argument.
Sometimes, graphics are great. I could read you off a list of the amount of money spent by every UK government department in the 2007-8 financial year, and by the end you'd be mad and you'd only be able to remember a fraction of it. If I gave it to you written down, it would be a long list, and it would take you time to compare spending between the Nuclear Decommissioning Authority and the Rural Payments Agency. That's because writing is not a helpful or intuitive way of representing quantities. So instead I would just show you the Public Spending Atlas drawn up by the Guardian and the Institute for Fiscal Studies, and all would become clear.
Visualisation has its limits, though. A good diagram can tell you how a system is or how it should be, but will struggle to tell you why it isn't how it should be. Suppose you run a widget-making business with a faulty supply chain. I can show you a diagram of that chain and a diagram of a new improved chain, and both diagrams could be fairly simple and use no words other than labels for parts of the system. But how do I explain why your existing supply chain is wrong? What I could do is elaborate on the original diagram, adding boxes for the users of your widgets, fat and thin arrows for supply and demand levels, dotted lines to represent delivery time, etc., and hope that the diagram is complete enough that you can work out the problem and recognize that the new system is better. But I wouldn't. I'd use a single sentence to connect the old diagram to the new: 'Your supply chain is faulty because it doesn't get widgets to widget-users quickly enough.'
Within the syntax of that easy sentence are the basics of what the problem is, and why - in other words, what you need to know. I can then elaborate on that. 'This is because you send everything from your warehouse by second-class post. Send it first-class and the time lag will disappear. You can fund the extra postage costs by making your executives fly economy class to meetings instead of business class.' No diagrams required, and if you want you can rearrange the syntax to make the process sound linear rather than relational: 'Make your executives fly economy class; save money; spend money posting everything first class; then your customers will get their widgets quicker. Then they'll recommend you to their friends; then their friends will buy widgets from you; then you'll be rich and beloved of widget-users.' One of the beauties of syntax is that you can reconfigure it fast.
The insistence on diagramming everything seems to be one of the major bugbears people have with strategy consultants. The flow chart is the stock joke of consultant-haters because it seems to be an uneasy attempt to turn grammar into graphics. Powerpoint's ubiquitous process flow arrows come in for even more ridicule, because they describe linear processes with no decision points. They are, in other words, just sentences with arrows around them.
Consultant-haters also loathe the proliferation of bullet points, and not without reason. Just as flow charts try to make graphics do the work of grammar, bullet points do the opposite. Bullet points are fairly useful for simple lists of items, but in that case you might as well use a graphic as it will be easier on the eye. For any kind of argument, they impose an artificial syntax which makes everything look like a linear process. This can wreak havoc with systems, as Edward Tufte shows in his book The Cognitive Style of Powerpoint, as it can lead us to assume that the relationship between concepts is paratactic, when often it isn't.
So I disagree that graphics are always best. Continuous prose, despite its fall from fashion in the world of business, remains an elegant and powerful way to argue persuasively. Whereas graphics allow us to see whole systems at a glance, syntax is a much better way of reasoning about change and its consequences, and therefore of explaining the need for new and better systems. Michael Halliday's systemic functional grammar model argues that every sentence tells a story about the relationships between things, people or concepts in the world. It would be a shame if systems thinkers were to forget that systems are dynamic, or to abandon their best means of telling a good story.
# Alex Steer (14/06/2009)