By what bastard grandiloquence have we fooled ourselves into the theory of predictable meaning for iconographic writing? Design favors icons because they are fast to read, fast to load on the computermachine, fit on the tiny flatspaces where we today look, and are somewhat prettier than words. However, after centuries of icons representing shops and components of systems, design today finds icons faltering at their primary job: meaning.
How we killed icons
Why are icons leaky carriers of signification?
Chiefly, because they are used so much and in such degraded circumstances. The places where we put icons, such as systems, tools, and documents, are now:
- being produced and replaced much faster than usual
- being used to communicate with larger audiences than usual
- relied upon to communicate with strangers, foreigners, and people from other (sub)cultures
- relied upon to communicate with absolute beginners who want it easy and now
We are asking a lot of icons and are trying to communicate with people who it is basically harder to talk to.
Second, the built environment and tools of contemporary society are becoming harder to represent with icons. The actual functions of a contemporary streaming media player are far more complex than that of a wax cylinder player or cassette or even audio CD. We add to play and stop: pause, favorite, add to playlist, skip for next song, repeat, switch to shuffle, and surely soon more clever options too.
Third, the design imperative for “clean” looks has simplified visual representations and minimum viable mental models to the point that there is precious little to draw an icon of. For example, one of these icons represents the app menu. The other represents the phone’s dialer.
They look the same because, in each case, you enter a menu with several large, equally sized buttons. Isn’t this the endgame of all technologies? A whimsy board of several large clean buttons that do your bidding faster, cheaper, sooner, more pleasantly?
Looking at the longer historical trend, industry has succeeded in turning big, heavy machines into pretty, virtual machines that run on small, sleek electronic machines. And while these new-fangled tools are easier to use, they are harder to draw. Consider the phone.
It was easier to draw a hundred years ago and may soon become impossible to draw at all, much like the tiny dot on your phone that is your camera or microphone. When the “telephone” functions are assimilated into some other device like the camera was into the smartphone (or the mitochondria into the cell), a “phone” won’t be something you can draw either. It would be like drawing teleconferencing today: someone on a laptop with a face on it.
What was hard about older technologies was also what made them iconic. Consider the visually rich features of a phonograph: the needle, the record, the album cover, the turntable, the speakers (or before that, the horn of the Victrola).
Icons Reflect Some Quality of the Thing You’re Representing
C.S. Peirce explained in the 19th century that icons are distinct from symbols because icons reflect some quality of that which they signify. In contrast, symbols are entirely conventional. The word “tree” is a pure symbol, but an icon of a tree at least reflects some visual qualities of a tree. It is arbitrary.
But that it is arbitrary does not mean we can arbitrate it. No, you can’t draw a triangle and call it a teleconferencing icon anymore than you can invent a new unified spoken language for humanity and expect people to learn it.
In a user interface, an icon has single literal meaning: what it does. When your icons vary from the expected meaning, clever users soon learn what that icon actually means now in this case.
The arbitrary meaning of a symbol, such as the hamburger menu or the word “tree,” can find widespread acceptance only slowly. Usually, it needs a big break: a single product, document, place or institution that uses it consistently and happens to be wildly successful. Once a billion people have ordered a Big Mac, it becomes a phrase most readily reused to describe things other than McDonald’s sad excuse for a large burger.
There is a little workaround here that takes advantage of living history. While, for most people, a camera is just a software function of the plastic rectangular solid called a “smartphone,” almost everyone recognizes the lens and body of the cameras of 50 years back.
So long as we have some cultural memory of the older technology being replaced, icons can draw on our antiquated sensibilities to communicate a helpful older meaning. The smartphone icon for a phone is a receiver, which is increasingly rare to actually use. But the word used is “dial,” referring to rotary phones, or “call,” referring to calling cards and actual verbal projection. Yet none of this is a problem.
In some cases, especially when replacing technologies with no lasting charm, this trick works less well. Save and Voicemail are nasty examples. Floppy disks have disappeared completely from cultural memory and voicemail is widely considered unpleasant. A save icon in a cloud app showing a floppy disk might do better to find some more welcome referent for saving things than a disk. (Or, better, a cultural revival of the floppy disk as an object of nostalgia or admiration could make it a nice icon.)
While antiquarianism and creative imagism can hold the failure of meaning at bay for a little while, the real transition we need is to a set of icons that are no longer iconic. Icons must become symbolic to have reliable meanings. They must be matters of convention through and through, without relying on any resemblance to qualify themselves. The “x” in the upper right, the magnifying glass for search, the pencil for edit. These are close.
But even more importantly, the digital industries will have to reach the terrible end of their prodigious growth and lock in a mature market where dominant demographics already know what symbol to expect for voicemail, how to add a layer to an image, and what a message icon looks like.
This will not happen for several decades. When it does, symbols will become hieroglyphics, with increasingly abstract meanings that describe software functions quite distant from the analog technologies we are still proud to have replaced.
Until then, icons will be tempting but unreliable.