Let’s take a look at the atoms of visual design: letterforms.

This is the letter R, typeset in Gill Sans.

Gill Sans was designed by Eric Gill and released in 1928.

There’s a couple of things that I’d like to point out about this letter right here:

One of them is that it has the perceptual characteristic of consistent weight all throughout;

however, the actual thickness of the letter varies considerably over different parts.

So, if you look at some points of the letter, like the bottom,

you can see that they are much thicker than at other points in the letter.

And what you see in these atoms is exactly what you’ll see throughout the entirety of graphic design,

which is that we’re going for something that feels perceptually balanced —

or may have perceptual uniformity — when the actuality may be quite different.

This letter was also my very first assignment as a graphic design student at the Rhode Island School of Design:

We had to draw this letter, one foot high, in black paint, on a white background.

It may seem like a silly exercise,

but it’s a great way to understand intimately the curves and strokes of the letterform,

and it’s also really good at training both your eye and training your hand;

Maybe it’s also a hazing ritual that I’ve just post-hoc rationalized.

Anyhow, moving along, let’s look at the different parts of a typeface.

One term you may be familiar with from using on your computer is the point size or the font size.

And what’s notable about this is

it seems like a number that ought to be set in stone and consistent across all typefaces,

but it’s really not.

Most of our terminology from typefaces comes from when they were set in lead,

and the point size of a typeface really just means the height of the lead block that the type was set in,

which has some relationship to the actual letter size, but is not completely consistent.

For example, you can see, for the same point size,

here’s a letter S in Gill Sans, and a letter S in Calibri,

and you can that the Calibri S is lower than the Gill Sans S.

Our second term is leading.

If we have solid blocks of lead type that form the letters,

the spacing that we get between the lines was achieved by putting thin strips of lead in between the lines,

and that was called the leading.

It’s customary to have 20% of your font size as your leading;

and so, if in your word processor you’ve set 10-point type,

it will usually automatically add an extra 2 points of spacing to put 12 points in between each line.

The height of the lowercase letters is called the x-height, from the “x”.

Some typefaces have a higher x-height, like Lucida Bright or Georgia;

other typefaces have a lower x-height, like Baskerville.

And this is important for design — especially online —

because typefaces that have a higher x-height will be easier to read at smaller point sizes

and on low resolution devices like most screens.

Conversely, typefaces with a low x-height, like Baskerville,

are used when you’re trying to give elegance, or a “Belle Époque” feel to something.

And sprouting up and dropping down from the x-height

are the ascenders — like the top stem of the h and the dot of the i —,

and your descenders — like the y, p, g.

In general, a typeface that has a low x-height will tend to have long ascenders and descenders,

and a typeface that has a tall x-height will tend to have more squashed ascenders and descenders.

Typefaces can also vary in their weight.

So here have an example of light on the left, regular or roman in the middle, and bold on the right.

Sometimes you’ll see weights in between like demibold or semibold,

and, even to the further to the right of bold, you’ll see black.

Number six on the list are serifs.

Serifs are the doohickeys on the end of letters that give them that chiselled look.

[There are] a couple more that are bonus:

Some faces (typefaces) will also include small caps — which sure are exactly what they sound like,

smaller versions of capital[s and] numbers.

These can be used as headers, or they can also be used

to put acronyms and other capital texts inside a block of body text.

And one thing that you may not know is —

just like letters come in uppercase and lowercase, numbers do too.

These numbers, set in Neutraface,

provide an example of something that has just small ascenders and descenders;

in some cases these can be more dramatic.

Lowercase numbers are useful

when you want numbers to fit in with a larger block of text and not stand out.

One of the most common questions I get asked about graphic design is:

“Which typeface should I use?”

And a lot of people have heard a story

where serif typefaces are generally recommended for body text

and sans serif typefaces are recommended for headers.

And that’s, on the whole, not bad advice.

There is an often-repeated claim that the reason that this is a good heuristic

is because serif typefaces, with those doohickeys at the end,

give your eyes something to latch onto, and consequently they’re easier to read.

It’s an interesting hypothesis.

So I looked into this a little bit for all of you,

and there’s some challenges in trying to figure this out.

For starters, individual differences in readers dwarf any manipulation effects of the typeface —

like, some people read a whole lot faster than others,

and so it’s difficult to measure differences in the readability of typefaces.

Also, reading requires familiarity:

It’s not like we’re born knowing what a letter R looks like; we need to practice.

And so if you learn with one style of the letter R, you’ll probably be faster with that one;

and if you learn with another style of the letter R, you’ll be faster with that one.

If even when we got past all those challenges,

it can be a little bit difficult to figure out exactly how you should measure this.

Well, I can’t go so far as to say that the serif hypothesis is definitively untrue;

what I can say is that there’s a lot of evidence to support that it actually is real.

And Alex Poole does a great job of describing this on his web site:

He writes that “What initially seemed a neat dichotomous question of serif versus sans serif

has resulted in a body of research consisting of weak claims and counter-claims,

and study after study with findings of ‘no difference’.

Is it the case that more than one hundred years of research has been marred by [repeated] methodological flaws,

or are serifs simply a typographical ‘red herring’?”

It does seem that mixed-case type like you get in a sentence — upper and lower mixed together —

is faster to read than all caps, presumably because

you get greater vertical variation with the ascenders and descenders in lowercase letters,

and that gives your eye more information as it’s reading.

Also, in general, proportional fonts seem to be faster than monospaced fonts, like a typewriter.

So, typewriter font[s] — slower to read — [but] that stuff only matters if you have to read a whole lot.

But the simplest summary is that, in practice, legibility simply amounts to what you’re accustomed to.

What do these words say? The bottom half is obscured.

As many of you guessed, “I am a jelly donut”.

How about this one here? Here the top half is obscured.

Seems a little harder?

“jfk spoke in berlin.”

Turns out that more information seems to be encoded in the top half of texts than in the bottom half.

And what does this say?

[I] bet many of you saw it as “THE CAT”,

and fewer people, “TAE CHT”.

And that’s because expectation plays a really important role in what we read.

And if you’re really perceptive,

you may have realized that those two middle letters are actually exactly the same letter,

so the only difference is what you’re expecting to see in each case.

Your prior probability influences your interpretation.

So, if you put all that together and return to our original question of “Which typeface should I use”,

the answer is “It depends.”

If you’re putting down a long block of text,

it’s best to use a typeface that many people will be familiar with;

and if you’re working online, use something that has a relatively high x-height.

On the other hand, if you’re making a logo or a sign,

or something that’s short and intended to be dramatic,

then you probably want to use something a little bit funkier as your typeface —

something that’s going to catch attention and be unique, memorable.

And typefaces — like everything in life — [build] up reputation.

So, over time, certain typefaces come to mean certain things.

And here’s a great example from Hoefler and Frere-Jones, writing on their blog.

They talk about how Baskerville seems like it’s out of a literary magazine

and Optima seems to remind us of cheap 1950’s hygiene products.

Here are two easy exercises you can use to improve your typographical eye:

The first one is: Look around you for all of the different ways that the same typeface gets deployed;

that will give you a sense of its range.

And the second one is: Look at how the same text, with different typefaces, really changes its meaning.

In this class we have a lot of ground to cover quickly,

and so unfortunately we can’t spend too much time on typography;

but it’s a fascinating subject, and if you’d like to learn more about this,

there are many great books and design resources online, and they’re growing rapidly.

Here are just a few that I think are particularly good.