A Voice through Type

The typeface is the voice of a text—so they say. But what kind of voice? On closer inspection we are mostly talking about well-behaved ones. For corporate logos they need to have enough general appeal to appease the largest possible audience. For most articles they are more comparable to ambience in order to serve the message. Wouldn’t it be fun to take this statement more literal and apply it to (fictional) people? Giving the textual representation of speech a visual indicator regarding the character of the speaker and the quirks of their voice. Sounds fun—at least as an experiment.

A limit to our fun

A few limitation beforehand. What I propose here is not a method. It is more like a thought-process that I hope can help to decide on a typeface representation of a character. Voices are wonderfully complex. We change our voice depending on speaking context or emotion. Picking various fonts or designing a super-family that would feel cohesive while also being meaningfully different for each situation would be super cool, but also very overwhelming. We are talking more of a proof of concept here. So I focus more on a general representation. Broad strokes often are a good thing when we are talking about details of details (the thing letters are in the context of a page or screen). I’m also gonna brush over the intricacies of characterization of people. To keep it simple let’s say we work mainly with a adjective-based framework that can be extended by whatever association comes to our mind in regard to the character.

It’s not a rigid system but a fun experiment after all. Talking about fun: The limit to our fun isn’t the sky, but legibility and readability—to a degree. Namely how well the letters are decipherable and how easy it is to read the text based on the merits of the chosen font. If we pick something very out there and expressive we might jeopardize the legibility. Same goes for readability. If the font causes discontent in the reader, they might hate it when the character speaks. On the other end: Maybe that is exactly what you want. Another way to circumvent these issues would be to limit the expressive font to the speakers name. The actual dialog could then be set in a more general font. The name would then function as a kind of logo that by proxy colors the words of the speaker. But it would also introduce a form of abstraction, as the typographic voice wouldn’t be as immediately connected to the words. Also, you are missing out on a lot of cool glyphs with this method. As a time-saving alternative it is still interesting.

The establishment

Now to the actual task at hand: Drawing a connection between typeface and characterization. In essence this process comes down to dealing with readers expectations. That doesn’t mean you need to meet all these expectations. I’m operating under the assumption that typographic connections are somewhat arbitrary. Arbitrary in this case doesn’t mean random. Readers’ associations can be wildly different, based on their personal experiences and preferences. A blackletter font can look edgy-cool to one person, to someone else it might look traditionalistic. One person might find Helvetica timeless, someone else might find it overstayed its welcome. A Font that looks exciting to one viewer, someone else might find pretentious. As designers we can’t look into other peoples’ heads. But we can make an educated guess. 

Because even though everyone is an island, no one actually is an island. Our associations with typefaces are based in a common learned visual language that is shaped by the context in which we see fonts used. For example for the longest time Apple used lightweight sans serif typefaces to promote their technical gadgets. This resulted in the association of light sans with luxury. On the opposite site the usage of Comic Sans by many small businesses resulted in an association by the wider public with affordability. Futura still can be associated to futurism and space-travel because of “2001: A space Odyssey.” 1 Blackletter fonts are perceived either iffy due to the appropriation for nazi propaganda or traditionalistic due to the usage in traditional restaurant signage (at least in Germany). Typographic associations are of course subject to change. This can come in the form of trends or simply new use-cases. 2 To use Blackletter as an example again: The young folks (that definitely doesn’t make me sound old) first association with this letter type is more likely to be with urban street wear.

For our purpose established associations can serve as shorthand. Similar to the way stereotypes are used for writing characters. When we use typography tropes in such way, the shorthand should not be the end-goal. Just as a stereotype in character writing is not the end-goal for characterization but a “conversation opener.” The more interesting character—typography connections come from subverting expectations or building onto them.

Spiky means angry

Another well to draw from are haptic and kinetic associations. A letter with spiky features will probably evoke the haptic feedback of touching a spiky object. In contrast, a font with very round features will cause associations with smooth and round surfaces just by proximity. A light typeface can be perceived as flimsy or as filigreed. Bold letters can be associated with heaviness. A font with a wide letterform can be perceived as steady and immovable. Oblique weights often imply a form of movement. For our little experiment we could use these associations to refer to unique features in a character’s appearance or personal traits.

A final aspect to consider when associating typefaces with characters is the “well-it-just-looks-cool” aspect. As with most things in design picking something you like because you think it looks cool, is a valid reason. One could argue that what looks cool to a designer is not just personal preference but (also) the result of research and exposure to different designs. To pick something that looks cool in this case is also a good way to create new typographic associations. So that is an added bonus.

While all these aspects are important when considering font associations, more often than not typographic nuances go right over most readers heads. 3 This means on the one hand to be blunt with your choices. Tiny letter details that you consider important to convey a characters personality are likely to get overlooked. If not due to the text size than because typographic nuance is often not perceivable to laypeople. The latter aspect could also be considered freeing, as you can try a lot of type shenanigans that go under the radar.

Words into actions

Enough overthinking. Time to put these concepts into action. For that we need a volunteer. May I propose we go with everyone’s favorite android songstress Hatsune Miku.

Hatsune Miku is a fictional character that was created for the software synthesizer ›Vocaloid‹ in 2007. The software is able to not only interpret notes with different pitch and velocity but also with vocals—hence the name. The sound is based on the voice actress Saki Fujita’s voice. The resulting sound is a very distinct robotic, and high-pitched voice, fitting for idol pop music. The software found a wide-spread following—which can in large parts be attributed to the iconic character design by Kei Garō. The concept for the character was that of a singer-diva android in a future where all songs are lost and need to be reinterpreted. As the character started to gain a life of its own through a large fan community and motion-captured life concerts with her as a hologram, the character is a bit more of a Japanese idol than a diva.

Hatsune was imagined as an android—a machine build in the image of humans. Considering this, references to the digital would be out of place, as she is meant to exist in the physical world of the future. Existing typefaces associated with technological progress have a tendency to look dated. Dated, in the sense that you can tell what idea of future shaped its design (think for example of ›Eurostyle‹). Going the “2001” route—using Futura as a shorthand for futurism—wouldn’t fit due to its mechanical construction principle. A more contemporary mix-style of geometric and grotesque sans could fit the “machine-imitating-a-human“ theme. Than again, to me this seems a bit too sterile, considering Miku’s implied extroverted-ness. Therefore I think focusing on this aspect is more promising.

The typographic style that caries the most diva-ness for me is the modernist serif form principle (Didot). It is often used in the context of high fashion, where it communicates carries an air of aloof-ness, quality and beauty. Typefaces in the modernist style emphasize verticality, and feature constructed shapes, as well as extreme stroke contrasts. 

I want to draw a connection between the constructed letter shapes and the artificial construction of an android like Miku and her singing voice. I think the modernist form principle transports this better than geometric type, as it is less streamlined in its construction and allows for more decorative elements. Also the high contrasts read to me as eccentric and therefore fitting well with the diva aspect. Because androids are modeled after humans but are still distinct from them, I think it could be cool to add a detail that implies human behavior was imitated but in a wrong way. I thought mirroring the shadow axis of the “o” in a transitional serif type could be a nice nod in that direction. That will put the font more in the category of transitional Serif instead of modern, but that’s fine. I like the transitional Serifs more anyway. Also I want to add prominent ball terminals—just to be extra.

In case I would look for a font—not make one from scratch—I would need to make more compromises. But working with existing fonts also creates other association possibilities. Like, seeing some feature in a font and come up with an association afterwards.

Anyway, this process like I said is very biased and subjective. But, I still think it is a fun way to approach typeface selection. I’m also sure there is a certain applicability outside of characterization of fictional people. Why not use this thinking process to pick fonts for a logo?

Notes

1 — Funnily enough “2001” doesn’t get nearly as much flag for its rather literary connection to its title font than “Avatar.”

2 — Speaking of trends: Type Campus’ Whitepaper “The 2022 Type Trends Lookbook” and “The 2023 Type Trends Lookbook“ do a very good job outlining recent font trends and putting them in the contemporary and historic societal context.

3 — Jeanne-Louis Moys was able to show in her survey that far more important were how the text was spaced, what weight was applied, or if the text was set in cursive font. Jeanne-Louis Moys. (2011). Typographic Voice: Researching Readers’ interpretations. (p. 14–15). In: Technical paper 6, Simplification Centre. (accessed April 2022).