By Gordon Rugg
Making a design interesting can be a significant challenge for designers, particularly when working in a well-established field.
Two simple but effective ways of making a design interesting are:
- making the design novel, in terms of deep structure and/or surface structure
- making the design difficult or impossible to parse.
Both these approaches have a long history in applied design, though usually without much explicit reference to the underlying principles.
This article discusses the underlying principles, and then looks at practical implications for making a design novel. I’ll look at the issue of parsing a design in a separate article, for reasons of space.
Novelty has long been recognised as an important issue in design. However, it’s traditionally been viewed as something subjective and vague, in terms of practical definitions and guidelines.
That situation has changed in recent years, with the introduction of concepts from information theory and knowledge representation. The underlying theory in both those fields is powerful and elegant. I’ll start by looking at this way of measuring novelty, and then go on to the practical implications.
Background theory: Measuring novelty
One key concept is that you can measure novelty by counting how often a particular thing has already occurred. The less often it has occurred, the more novel it is. This simple principle is widely used in online search engines, in the form of inverse document frequency weighting. Here’s how it works.
If I search for the word “cat” on Google, I get about 456,000,000 results. If I search for the word “civet” on Google, I get about 788,000 results. What the search engine now does is to transform those numbers by inverting them, for technical reasons, into 1/456,000,000 and into 1/788,000 respectively. What this means is that:
- at a practical level, we have a simple, objective method of measuring novelty
- the numbers from this method show the word “civet” to have a novelty value very much higher than the novelty value for the word “cat”.
So far, so good. This approach has been invaluable in online search, and we’ll return to it in later articles. However, when you apply it to design, you hit the issue of deep structure versus surface structure as complicating variables. These concepts are viewed as very familiar in some design-related fields, but not in others, so I’ll now unpack what they are, and what their implications are for design.
Background theory: Deep structure and surface structure
These concepts crop up in various disciplines under different names, such as classes and instances. As is often the case, the ancient Greeks spotted this issue early on, and, as is also often the case, the ancient Greeks then proceeded to invent plausible but profoundly misleading explanations and categorisations that caused needless confusion for centuries afterwards. Later philosophers such as Jung helpfully added to the confusion by introducing concepts such as archetypes. I’m not planning to go into detail about topics such as Plato’s concept of essences, but any readers who have an interest in classical philosophy might like to try comparing the classical approaches with the approaches that are described below.
I’ll start with a picture of a cat. It’s a grainy low-res picture, but it will be immediately familiar to most readers.
Grumpy Cat is a specific, individual cat. In technical terms, Grumpy Cat is an instance of the class of cat. The class of cat is in turn a member of the class of pet.
There are different ways of slicing up the same concept for different purposes. For instance, we might want to treat the instance of a particular cat as belonging to the class of tabby cat, which then in turn belongs to the class of cat. Another option would be to treat the instance of a particular cat as belonging to the class of cat, which then in turn belongs to the class of carnivores.
If you’d like to know more, there are systematic, formal ways of handling the concept of using different classifications for different purposes; facet theory in particular is highly relevant, and is described in one of our other posts.
That’s a very brief overview of the theoretical foundations; the next section is about the practical implications.
Practical implications: Surface structure and deep structure
If you want to make a design novel, then you have a choice. You can make it novel at the level of surface structure, or at the level of deep structure, or both.
Surface structure is about instances, as opposed to classes, which is why we had to cover that topic before tackling this one. Deep structure, conversely, is about the underlying structure – usually classes, but often actions. Here’s an example that should make the distinction clearer. I’ve shown the surface structure in italics.
Han Solo has a crew member who is a Wookie.
One deep structure for this sentence is as follows (deep structure in bold):
The progagonist has an associate who is part-human and part-animal.
So what? So, if we now look for other examples of the same deep structure, we get these, among numerous others:
Gilgamesh has a companion who is a beast -man.
Lyra has a familiar who is a daemon.
So, you can produce novelty at the level of surface structure (Gilgamesh and Han Solo and Lyra are very different from each other at the surface level) while keeping the same deep structure. This is very well known territory in literature; the classic example of this work, Propp’s Morphology of the Folktale, dates from the early part of the twentieth century. It’s also well known in media research, sometimes under the name of TV tropes.
It’s less well known, though, in other fields, so in this article I’ll describe the underlying principles, and look at how they can be applied to a broader range of design fields.
Mapping out the permutations
The table below shows the permutations of familiar and novel, and of deep surface and surface structure.
Each of these can be used for different purposes.
Familiar surface structure, familiar deep structure
One everyday example of this combination is a TV soap opera, with the same cast across numerous episodes, and with a familiar set of plots. Another example is a faithful remake of an old film or TV show. There is a substantial market for this category of production, usually because it’s comfortably familiar in an often unpredictable world. This is well recognised in the media industries, where it’s often described in terms of finding a best-selling formula or recipe. Barbara Cartland’s oeuvre is one widely-cited example of this principle carried to a logical conclusion.
Novel surface structure, familiar deep structure
Novices tend to be poor at spotting deep structure, and tend to focus on the surface structure.
In practical terms, this means that you can often create a new product by re-using a well-established deep structure, such as a layout for a house, or a design for a car, or a plot for a story, and changing the surface structure. This is usually cheap and simple, since the key structural elements are unchanged, and since surface structures are usually easy to change. Examples include as a new type of surface for the exterior walls of the house, or a new shape for the lights on the car, or characters with new quirks within the well-established plot.
This has the advantages of being easy to do, and of having a deep structure that is comfortingly familiar to the target audience. There is, however, the risk that the audience will eventually start to spot the deep structure regularities, and become bored with them. This is a well known phenomenon in the world of TV tropes.
For some niches this isn’t a problem. A classic example is material for children, such as children’s novels or children’s TV, where the same deep structures can be re-used year after year, because the children will probably move on to other interests before they spot the deep structure regularities. For longer-term audiences though, such as audiences for crime series or thrillers or horror, it’s more likely to be an issue.
Novel deep structure, familiar surface structure
For experienced, jaded audiences, one way of introducing novelty is to use the same surface elements, but to change the deep structure. This can produce a new, and often unsettling, way of seeing familiar scenes.
The first Terminator movie did a brilliant job of this by taking familiar elements (e.g. the heroine is menaced by an assailant) and then giving a twist to what happens next. In most horror movies, for instance, the deep structure convention is that when the heroine is menaced by an assailant, there’s some plot contrivance that prevents her from going to the cops for protection. In Terminator, however, the heroine is menaced by the Terminator (familiar plot element) and then goes to the cops. The step of going to the cops is familiar from other genres, so what’s novel is the sequence of familiar elements.
Similarly, in those genres where going to the cops is a familiar story element, it’s usually followed by the element of the cops tell the heroine to go home because there’s no evidence for her story, whereas in Terminator it’s followed instead by the element of shoot-out with the cops – again, a familiar element, but in an unfamiliar juxtaposition.
When new technologies are marketed, as in the early days of cars and aircraft, a common way of making them more acceptable to the target market is to obscure much of the new deep structure by using a familiar, reassuring surface structure to lessen the perceived transition. In the case of cars, airships, and passenger aircraft, the initial market was rich early adopters, so the products included interior design features that were familiar and luxurious, regardless of whether they were functionally optimal.
This made it possible to keep the perceived novelty within comfortable bounds, and also made it easy to vary the perceived novelty at the level of surface structure, to conform with feedback from the users.
Novel deep structure, novel surface structure
It’s possible to introduce novelty in both surface and deep structure at the same time, but the risk is that this will put too much novelty into the mix, and will overwhelm the audience. An example of this was the initial negative reaction of many critics to Stravinsky’s The Rites of Spring, which had a lot of novelty both in deep structure and surface structure.
If you get this combination right, though, you may manage to establish an entire new genre or market, with all the associated opportunities and rewards. An example is the mobile phone, which has been massively successful and popular as a technology.
Usually, novices in a given area focus on the surface structure, whereas experts tend to focus on the deep structure. This is well known to script writers and to writers of student examinations; in both cases, you can recycle the same deep structure year after year by just changing the surface structure above it. However, once an audience has seen enough cases to become familiar with the deep structure, then you need to use other methods to make the product interesting.
One is to change the deep structure, as described above.
Another method is to make use of another aspect of how people process information, namely parsing. That is the topic of the companion article to this one.
There’s a whole world of TV tropes, which is fascinating, but can be addictive. A good place to start is on the TV tropes site.
In case you’ve always wondered about Barbara Cartland but been afraid to ask, here’s the opening sentence of one of her books, as a small taster.
The Lioness and the Lily.
Chapter 1: 1841.
“As the Earl of Rockbrook drove down the drive of the enormous Georgian mansion which had been in his family since the days of Charles II, he felt no pride of possession.”
Pingback: One hundred Hyde & Rugg articles, and the Verifier framework | hyde and rugg
Pingback: 150 posts and counting | hyde and rugg
Pingback: Why Hollywood gets it wrong: Conflicting conventions | hyde and rugg