Hidden in the vast majority of typefaces is one of the most well-known optical illusions in the world. Can you spot the illusion, and can you identify the famous illustration that demonstrates its power over human perception?
I’ll give you a hint on the first part: Look at the last three letters in the word “OPTICAL” as it appears above, written in all-caps, in one of the most ubiquitous typefaces of the modern era, Times New Roman. Look closely:
Something about the letters is “off.” Do you see it? How about here:
Notice how the peak of the A extends above the dotted line resting flush atop the L? Now look at the C; like the peak of the A, its upper curve extends beyond the horizontal line that marks the height of the L. But then, its lower extent dips below the bottom line, too.
These encroachments, which typographers call “overshoots,” add to a letter’s absolute height. If you brought the bottom of the C in line with the bottoms of A and L, you would observe that C is taller than A is taller than L, despite the fact that they are all of “equal” font-size. This feature is not unique to C, A, and L; letters with curved tops and bottoms (like Os and Ss and Gs) and pointy extremities (like Ns and Ms and Vs) tend to overshoot the upper and lower bounds (the cap-height and baseline, respectively, in typographer lingo) of letters with flatter or boxier extremities, sometimes by as much as 5% of the letter’s cap-height.
But why is this? Why would a type designer make some letters intentionally bigger than others?
The simple answer is that it looks better. In fact, if typographers didn’t use overshoot, words would look strange. Letters of equal metric height would appear, to our eyes and our brains, disproportional. In any given string of letters, there exists a tension between what is mathematically correct and what is visually correct—and in typography, the latter almost always wins out. “Part of typeface design is managing this eternal friction between logic and optics,” explains acclaimed type designer Tobias Frere-Jones, in a comprehensive post on the mechanics of overshoot published earlier this year. “It’s always there, no matter the style. It is what placates the stubborn oddities of human perception.”
The Müller-Lyer Illusion
But the question remains: Where does this “stubborn oddity” originate? In his blog post, Frere-Jones links our perception of a letter to the proportion of its shape that is fully and squarely aligned with the upper and lower edges of its form. Consider a letter like H, the outline of which is basically a box. Its upper and lower bounds, its four corners, match uniformly with its baseline and cap-height. In contrast, “only a narrow sliver of an O is the full height, and the rest of the shape falls away,” says Frere-Jones. “The parts that are too short greatly out number the parts that are big enough, so we conclude—wrongly, but very reliably—that the round shape is too small.”
The phenomenon behind what Frere-Jones is describing is something most people are familiar with, even if they’ve never thought of it in relation to typography. Does the phrase “Müller-Lyer Illusion” ring a bell? If it doesn’t, the following illustrations almost certainly will.
In the 1880s, German psychiatrist Franz Müller-Lyer devised an experiment that demonstrated simply and compellingly the fallibility of human perception. The experiment showed how the addition of inward- or outward-pointing chevrons to the ends of two line segments of equal length could cause one segment to be perceived as shorter or longer than the other. Today, it’s one of the most famous illusions in the world, and it’s not hard to see why. The illustration is dead simple to reproduce, and its mind-bending effect is powerful. It’s the classic optical illusion, as common in books as it is in listicles all over the internet. It’s also popular in scientific circles; to this day, it remains one of the most widely investigated—and widely debated—illusions in vision research. And if you turn the illusion on its side, you’ll see why it’s relevant to type design, as well:
This screenshot comes from a Seattle Nerd Night presentation delivered last year by Kevin Larson, a researcher on Microsoft’s Advanced Reading Technologies team and the author of the outstanding typography essay “The Science of Word Recognition or: How I Learned To Stop Worrying and Love the Bouma.” According to Larson, who works with designers and psychologists to create reader-friendly typefaces, the Müller Lyer Illusion is the “psychological reason” for overshoot. Letters like I, Larson explains, bear a strong visual resemblance to a line segment with terminal chevrons pointing inward. This makes them appear taller than letters like O, the shape of which is more akin to that of a line segment with outward-pointing chevrons.
“Typographers will always design Os [with overshoots],” says Larson, “because they believe it is going to be, visually, the same size as that I.” On the page, an O that is the same height as an I will look a little on the small side, “even though mathematically they’re the exact same size.”
Tobias Frere-Jones’ essay on overshoot contains a number of examples of where this design choice is appropriate or inappropriate, and features several illustrations and animations that help illustrate his points. It is the first installment in a series of posts where Frere-Jones will explore what he calls “typeface mechanics,” or “the behind-the-scenes work that makes typefaces visually functional.” You can catch up on the full series here.
The portion of Kevin Larson’s Nerd Nite talk that deals with overshoot is only a couple minutes long, but the full talk—which is a little over half an hour long—is outstanding. Watch the entire thing here: