I’ve always been fascinated by imaginary and constructed languages and alphabets. When I was in primary school, I used to send my classmates pieces of paper with encrypted messages they couldn’t read. The letterforms were completely invented, and I created translation tables to correspond between the made-up shapes and the Latin alphabet, allowing the message to be decoded. While it made me feel clever, it wasn’t anything innovative; all I did was write the Italian language with an alphabet that worked exactly like the Latin one but replacing the shapes with strange variations of lines and points. I later found out that many children's fantasy books have games like that to interact with young readers.
Adulthood has brought much more complexity to the topic. Not only did I learn that complex languages built from scratch exist – the most famous of which I believe is Esperanto – but I also became aware of the infinite world of writing systems. There are constructed scripts for natural languages, constructed languages that use existing scripts, and constructed scripts for constructed languages.
While all existing writing systems, from Latin to Chinese to Arabic, are human-made constructs, we typically refer to natural scripts when they evolve gradually from preexisting scripts, as opposed to being the product of a single individual. In most cases, a language is initially written in the script of another language, and over centuries gradually develops features that are unique to its new surroundings. Instead, we talk about constructed scripts when they are created in a much shorter time by a single individual or team. For example, the Korean alphabet – Hangul – was created in 1443 CE by King Sejong the Great in an effort to promote literacy by acting as a substitute for the previous system, based on Chinese characters.
In addition to Hangul, other constructed scripts for natural languages include Native American Cherokee, Canadian Aboriginal Syllabics, West African N’Ko script, and others. The International Phonetic Alphabet is one famous example of a constructed script developed for practical purposes, but perhaps the most well-known and intriguing scripts are those developed for fictional languages. Tengwar and Cirth scripts by J.R.R. Tolkien are extremely complex and have achieved unparalleled fame, but there are many others, including the Aurebesh from Star Wars, the pIqaD script for the Klingon language in Star Trek, the D’ni from the Myst video game series, and the Orokin language script from the video game Warframe. More recently, a newly constructed script was developed for the High Valyrian language in George R. R. Martin’s universe.
Tengwar, as we mentioned, is certainly one of the most well-known fictional writing systems and is a constructed script also in the Middle-earth universe, where it was created by the elf Fëanor and initially used to write the Elven languages Quenya and Telerin. Later, the Tengwar script was used to write a large number of Middle-earthen languages, including Sindarin. In reality, Tolkien created it in the late 1920s or early 1930s, although the full overview of the Tengwar was originally published in Appendix E of The Lord of the Rings in 1955. The majority of letters are made up of two fundamental shapes: a vertical stem, which could alternatively be ascending or descending, and one or two rounded bows, which can be on the right or on the left of the stem, and can present and underscore.
Designing a typeface for a constructed writing system has the potential to be a fascinating undertaking for enthusiasts, but what is the typographic perspective that we should consider in approaching this? To start, it is important to consider the purpose and function of the writing system. Is it used for a specific language or culture? What are the phonetic and grammatical rules that govern the language? These factors will play a significant role in determining the design of the letters.
The design decisions that need to be made are basically the same for any other script, and it is useful to understand how the type designer approaches the specificities of different main writing systems to design a font.
When designing a typeface, it’s important to pay attention to details such as letterforms, kerning, and spacing. Consistency is key – each character should have a unique but cohesive design that fits within the overall aesthetic of the writing system.
The design decisions that need to be made are basically the same for any other script, and it is useful to understand how the type designer approaches the specificities of different writing systems to design a typeface. It is essential to start with the story of that script: we need, for example, to understand how a letterform is constructed when it’s hand-written, where the thick and thin strokes go, how the stems end and how much is the axis inclined. We also need to understand which tools the characters were originally shaped with: are we looking for a more calligraphic touch or a more geometrical construction? Is the contrast based on expansion, translation, or a completely new set of rules?
For example, one of the most common font categories in the Chinese writing system is Songti. Readers can associate it with high-contrasted Latin serif characters, in some ways, for its use in textbooks, formality, and history, but it has developed from totally different premises. In the Song Dynasty (960-1279 CE), when Chinese woodblock printing was at its peak, early Songti scripts started to be adopted. The horizontal grain of the wood used to make the woodblocks made it simple to create and thin out horizontal lines, while vertical lines had to be made thicker to prevent breaking. Flourishes were also added to the endpoints of the horizontal lines to make them thicker so they would survive longer because they were easily worn away. This is how the Chinese “serifs” were born.
These thick and thin strokes are generally completely absent in the second biggest group of Chinese characters called Heiti, which we can assimilate to Latin sans-serif. Appearing in print media in the early 1900s, Heiti typefaces are a relatively modern invention, born from the encounter between Eastern Asia and Western movable type. The so-called “Gothic” (Gochikku, subsequently known as Goshikku) typeface was developed by Japanese type foundries to adapt Chinese characters to the monolinear, uniformly thick Latin Grotesque script.
Clearly – like in the example of the Chinese script – the history of the writing system strongly influences the stylistic elements of a typeface, but what should be done for writing systems that have little to no historical background? We can try to guess its development from the elements we have. Going back to Tengwar, it has been hand-written by Tolkien many times during its creation. Some hypothesized Tengwar shapes were inspired by the insular minuscule (used to write Anglo-Saxon) or the Gothic textura and cursive scripts (used to write Middle English), which Tolkien knew by studying medieval manuscripts as a professional linguist.
Many Tolkien admirers have already tried their hand at creating Tengwar fonts. The attempt to capture the calligraphic nature of handwritten Tengwar has been the most common approach, from what I saw. We can assume that if elves were still part of our current society and adopted printing before, and computers later, Tengwar would most likely change and possibly absorb influences from Latin Roman styles. However, the type designer should make sure that Tengwar individual components that make up letters originate from Tengwar script strokes rather than being translated from Latin. This, of course, would not prevent the style from developing geometrical attributes or having different stroke modulations.
I believe in fact that the easiest mistake to make would be to substitute Tengwar in-strokes with Latin features, such as serif-like endings. This is a fairly uninventive solution to create entry strokes in Tengwar, where such a development never occurred in its history. It is crucial to emphasize the unique characteristics of Tengwar and set it apart from Latin. For instance, Tengwar frequently uses horizontal strokes that are typically thicker than those found in Latin, has a more angled axis, and relies less on vertical strokes. There is also a rhythmic repetition of rounded shapes. Let’s take as an example the first sentence inscribed in the One Ring, translated as “One ring to rule them all”. This phrase is written in the Black Speech language – spoken in the evil realm of Mordor – using the Tengwar script.
I believe in fact that the easiest mistake to make would be to substitute Tengwar in-strokes with Latin features, such as serif-like endings. This is a fairly uninventive and trivial solution.
To begin with, we must recognise the fundamental strokes employed in this writing system and produce a framework, or skeleton of the letters. In the books, J.R.R. Tolkien has already presented various writing styles of the script, with some being more angular and others more upright: the inscription on the One Ring features very long ascenders and descenders and a very steep angle of slanting. Having different styles is helpful in determining the core motion of the letter and instead, what constitutes a decorative element. When starting from calligraphy, as in this situation, it is simpler to deduce the direction and movement of the strokes. After constructing the skeleton, we can proceed to add thickness and contrast, and then gradually incorporate more intricate details. I designed a Tengwar version of Messapia Regular, which is a serif, high-contrast, display typeface with stroke modulation based on expansion and with clear cuts at the end of strokes.
To create Messapia Tengwar, I changed the type of contrast to translation to reveal a more calligraphic nature and to ensure that the shapes stand out enough as a different script. I avoided using serifs, which are too connected to Latin history and instead chose different entry strokes that are already used in some of Messapia’s glyphs. The contrast is still extremely high and the proportions are wide, as in the original cut.
A sans-serif-like look is more straightforward to achieve because stroke modulation and contrast are almost absent. I designed the same sentence in Mattone Regular by going back to the skeleton and starting from there. The look and feel are completely different, but it's still easily recognisable as Tengwar and legible.
Another very recent writing system that I find incredibly interesting is the one created by the American linguist David J. Peterson for the High Valyrian language, described by The Guardian as the richest linguistic universe since J.R.R. Tolkien’s. Peterson has constructed many languages for television series and often created writing systems for them. The script for High Valyrian was not included in Game of Thrones, but George R.R. Martin mentioned a written form of it in his books and, ultimately, it was commissioned specifically for its spin-off House of the Dragon.
David J. Peterson describes the writing system in an Instagram post as a “mixed script that features an alphabetic component, an iconic component, and a paradigmatic component, with all three elements often used simultaneously.” Peterson also comments that he wasn't inspired by existing scripts but only by his knowledge of the series, although he admits that the Latin script was important for the final look: “The point was to give the effect of the Roman script while having the function of a logographic one. Frills and curls wouldn't fit the setting. It’s the song of ice and fire, not the lay of elves and hobbits.” The examples from the author came already in a font loaded with a massive presence of serifs, a low contrast and almost no stroke modulation. Looking at the script, the first thing you notice is that the lines are very straight, with a few rounded shapes, many diagonals, and almost no ascenders/descenders.
It’s hard here to find some reference material. Some of these letters remind me of characters used in the Cyrillic script, such as the letter Ge or Yus, while some others are closer to simple Chinese radicals or runes. For example, the Ingwaz rune is very similar to the High Valyrian character for “rope”.
Once again, I tried to understand the scheme at the base of every glyph, using the phrase “Valar Morghūlis” – all men must die – as a model. A High Valyrian version of Apfel Grotezk, sans-serif and low contrast, is an easy step from the skeleton just by using it as a starting point and adding some optical adjustments in place. We need to consider the same old things as always: rounder shapes appear smaller than straight ones, so they need an overshoot; horizontal lines need to be a bit thinner than the vertical ones in order to appear the same, and so on. When it comes to designing a typeface that encompasses multiple scripts, it's important to consider the proportion of the letters in relation to each other. For instance, how should the letters align with other scripts, such as Latin? Should they have the same vertical proportions as the Cap height or the x-height of the Latin script, or perhaps a new dimension altogether? High Valyrian script seems to have just one case for its letters, and I decided to align their top with the Latin cap height, as they seem to complement Latin upper-case letters more effectively than lowercase. The script features also some ascenders that can simply align with the Latin line of ascenders. Another noteworthy design decision concerns the second character of the sentence, which was originally designed by the author to feature two additional serifs in the middle of the letter. In this sans serif typeface, I opted to include a middle stroke that exits from the letter to replicate this design element, although I guess could be possible that these serifs may be omitted in other sans serif versions.
I believe that the cursive handwriting version of High Valyrian would likely differ significantly from the typographic rendition. To explore this notion, I tried to imagine what High Valyrian might appear as when set in the Mazius Display typeface—a high-contrast serif font characterized by a distinct calligraphic mood inspired by chancery hand. Prior to crafting this representation, I embarked on a calligraphic attempt, beginning with the skeleton structure and using a broad nib pen to establish the contrast of thick and thin strokes. I also added a bit of chancery flair to the pen movements to reference the medieval setting.
In the case of Mazius Display, the designer – Alberto Casagrande – made the Horizontal lines very thin, so we could try to apply the same criteria for High Valyrian too. In this instance, I aimed to enhance the upward stroke of the third character by aligning the letters with the x-height, resulting in a look that closely resembles Latin small capitals. However, as I said before, I’m uncertain if this approach is more effective than the proportions I assigned to Apfel’s version. It's worth noting that the significance of this aspect diminishes when the script is used in isolation and not in conjunction with other scripts.
Once the font is complete, it’s time to test it. Make sure to test it in various contexts, such as different sizes and on different screens. It’s also a good idea to get feedback from others to ensure that the font is legible and functional. Ultimately, when everything is ready we need to make these fonts work in a digital environment, so it’s fundamental to understand how computers can read these shapes as a proper writing system.
Various encoding systems have been developed to represent artificial scripts, such as Tengwar, in digital media. One of the earliest attempts was made by J.R.R. Tolkien himself, using a system of letters, numbers, and punctuation marks. This system was relatively crude, but it laid the foundation for later attempts to encode Tengwar in digital media.
In the late 1990s, Michael Everson created a Unicode proposal for Tengwar, which was accepted by the Unicode Consortium in version 3.0. Unicode provides a comprehensive encoding system for Tengwar, including a full set of characters, but it does not always capture the unique features of individual scripts. To address this issue, some encoding experiments have been conducted, such as the TengScribe project, which uses a custom encoding system to represent Tengwar more accurately than Unicode. TengScribe aims to work seamlessly with software applications like Microsoft Word and Adobe Photoshop, allowing users to create high-quality Tengwar documents and images.
Tengwar script is also included in the unofficial ConScript Unicode Registry (CSUR) in the Private Use Area. Various typefaces and fonts have been designed to support Tengwar using the Unicode standard or CSUR. Some, like Kurinto, use their own mapping system and are not directly compatible with Unicode or CSUR, but they provide an alternative option for using Tengwar in digital media.
Other encoding experiments have focused on creating encoding systems for other artificial scripts, such as Klingon and Dothraki. In conclusion, encoding artificial scripts in digital media is crucial for their wider recognition and use. While Unicode provides a standardized encoding system, specialized encoding systems like TengScribe offer a more accurate representation. It is likely that more encoding experiments will be conducted to create even more advanced and specialized encoding systems for these fascinating scripts.
Peterson, D.J. The Art of Language Invention: from Horse-Lords to Dark Elves, the Words behind World-Building. NY, NY, Penguin Books, 2015