A constructed language or “conlang” is a language that an individual or group has deliberately created with a purpose in mind, as opposed to languages that have arisen naturally. Today, when most people today hear this term, they think immediately of languages used in works of fiction, such as J. R. R. Tolkien’s Elvish language Quenya (which he created long before he used it in his novels The Hobbit and The Lord of the Rings), Klingon from the television series Star Trek, or Dothraki and High Valyrian from Game of Thrones.
Many people may be surprised to learn that the oldest examples of what we might call constructed languages were not created for fictional worldbuilding at all, but rather as a philosophical or theological exercise. Such examples are also far older than many readers may realize, dating all the way back to the Middle Ages.
The search for an ideal language
To understand the medieval impetus for creating constructed languages, we must go back to the very roots of western philosophy. The ancient Athenian philosopher Plato (lived c. 428 or c. 424 – c. 347 BCE) has a prominent role to play in this story, as he does in so many others.
Plato’s dialogue Kratylos, which scholars traditionally consider one of his middle dialogues, addresses the question of whether words and names for things are “natural” (i.e., determined by the essential qualities of the things they describe) or simply arbitrary. In the dialogue, Plato’s Socrates argues that, in the same way that arts such as painting and sculpting seek to depict the essential qualities of their subjects, words should also capture through their sounds the essential qualities of the things they describe.
Plato’s Socrates even identifies specific sounds as conveying certain natural qualities. For instance, he says that the letter ρ (which makes a trilled ‘r’ sound) signifies motion, the letter ν (‘n’) signifies inwardness, the letter ι (‘i’) signifies small things that easily penetrate, and the letters α (‘a’) and η (long ‘e’) signify big things.
Throughout the ancient and medieval eras, European thinkers widely agreed with Plato’s Socrates that an inherent ontological connection exists between concepts and the words that describe them. For premodern thinkers and philosophers, words and the sounds they contain were not arbitrary or meaningless; they were inherently imbued with ontological, philosophical, and theological significance. A natural implication of this thinking is that certain ideal words for concepts must exist and that it is possible for humans to discover such words.
ABOVE: Photograph from Wikimedia Commons of a Roman marble portrait head of Plato, based on an earlier portrait made by the Greek sculptor Silanion c. 370 BCE
Hildegard of Bingen’s lingua ignota
The earliest constructed language that is reasonably well-attested is the lingua ignota or “unknown language” of the German Benedictine abbess Hildegard of Bingen (lived 1098 – 1179 CE), who easily stands among the likes of Leonardo da Vinci, Benjamin Franklin, and Johann Wolfgang von Goethe as one of the greatest polymaths in human history. As a prolific mystic visionary, theologian, musical composer, physician, natural scientist, and polyglot, Hildegard is one of four women whom the Catholic Church has officially recognized as Doctors of the Church.
In contrast to modern conlangers, Hildegard does not describe herself in her writing as having created a new language, but rather as having simply “brought forth” a perfect language that God himself had created and kept hidden from humanity until he chose to reveal it to her. Hildegard also “brought forth” the litterae ignotae or “unknown letters,” the alphabet of twenty-three letters used to write the lingua ignota.
Hildegard produced a glossary of 1,006 unique words (almost all nouns) in the lingua ignota (which are mostly glossed in Latin) as well as a hymn mostly in Latin with a few words of lingua ignota sprinkled in. These survive in the Wiesbaden Codex held in the RheinMain University and State Library. Although the lingua ignota uses invented vocabulary, it follows the same basic grammar and morphology as Latin.
ABOVE: Frontispiece illustration from the Rupertsberg manuscript of Hildegard of Bingen’s Scivias (1151 or 1152 CE) depicting her receiving a vision while simultaneously drawing on a wax tablet and dictating to her scribe Volmar
Voynichese?
In addition to Hildegard’s lingua ignota, another example of a possible constructed language from medieval Europe is the language of the Voynich manuscript, an illustrated vellum codex that has been carbon-dated to the early fifteenth century CE and is written in what appears to be an unknown language, dubbed “Voynichese,” which no one has ever been able to decipher convincingly.
Scholars have proposed a wide range of hypotheses about the Voynich manuscript and some have argued that Voynichese may be a constructed language. This hypothesis, while plausible, runs into the obstacle that, given the relative simplicity of other medieval constructed languages (such as the lingua ignota discussed above), it is difficult to explain what could have driven a medieval person to create one as elaborate as Voynichese seems to be, why the apparent language bears many features uncharacteristic of any known language, and why modern computer-assisted codebreaking has been unable to crack it, despite its immense volume of text and the decades scholars have spent trying to understand it.
I will not completely rule out the constructed language hypothesis for the Voynich manuscript and remain open to the possibility that it may be correct, but I find the assessment of Dr. Justin Sledge, who argues that the apparent “writing” in the manuscript is actually a medieval hoax meant to look like writing without actually encoding any meaning, highly compelling, since this is the only hypothesis that explains Voynichese’s bizarre features and why it remains undeciphered even in this day of computer codebreaking (i.e., because it isn’t a real language at all, so there is nothing to decipher).
Even if the Voynich manuscript is just an elaborate hoax for profit by medieval scribes, though, I think it still deserves a place in the history of constructed languages, given that creating a script that looks like a language without bearing meaning still bears many similarities to creating an actual language.
ABOVE: Scan of the Voynich manuscript, folio 66r, showing the unreadable script, which may be a constructed language or a hoax meant to look like a language without actually having any meaning
John Dee and Edward Kelly’s angelic language
Moving beyond the Middle Ages into the Early Modern Period, the English occultists John Dee (lived 1527 – 1608/9) and Edward Kelley (lived 1555 – 1597/8) record in their private journals conversations that they claim to have had with angels, in which the angels revealed to them an angelic language (and accompanying script), which was also purportedly the original language of all humanity.
Dee himself calls this language “Celestial Speech,” “the Holy Language,” “the Language of the Angels,” or the “Adamical” language, but subsequent authors have referred to it as “Enochian” as a result of Dee’s claim that the patriarch Enoch from the Book of Genesis was the last mortal to speak it. The purpose of Dee and Kelley’s Enochian angelic language is clearly theological and esoteric, similar to that of Hildegard’s lingua ignota.
ABOVE: Portrait of John Dee (left) and image of Dee’s journal entry recording the angelic script (right)
Conclusion
Constructed languages have a far longer and more complicated history than many may realize, which encompasses far more than just languages constructed in modern times for books and television. It is a history that stretches long before Tolkien to Hildegard of Bingen and even has roots in the ideas of Plato.
(This blog post is an expansion of this answer I originally wrote to a question in r/AskHistorians.)
This was a delight after … your announcement. As I have oft indicated you have a gift for communicating about the past and I would hate to see that lost. And many scholars of the past lived off of their royalties when they couldn’t find jobs.
As I said in my previous post, I don’t plan to end the blog, but it is going to have to be something I do for fun in my spare time. I don’t want to be reliant on content creation as an occupation because it is not going to be a dependable source of income. I also think I will have to write shorter articles than I used to—more articles the length of this one, rather than longer ones requiring more extensive research. I’m just not going to have time to write insanely long articles in the foreseeable future.
I have several more posts in my drafts that I am hoping to publish before I go off to law school next month, so stay tuned!
I’m still foolishly hoping for the Voynich manuscript to be a remnant of some lost Pre-Indo-European language that was only spoken by a few peasants in an isolated mountain valley, and that one of those peasants joined a monastery and created a script for his people’s dying language.
It’s a foolish, romantic hope I know, but it seems so captivating, doesn’t it?
Yes, that’s definitely the coolest thing that the Voynich manuscript could turn out to be! Unfortunately, it’s also extraordinarily unlikely, considering that, if that hypothesis were correct, we would expect to find some kind of other evidence aside from the Voynich manuscript for the Voynichese language. It is very unlikely that a language with its own unique script would be completely unattested anywhere except for one long manuscript dating to the 1400s that was clearly produced by professional scribes (as evidenced by an array of factors, including the style of the illustrations, the preparation of the vellum, the consistency of the script, etc.). The pre-Indo-European hypothesis would be much more plausible if, instead of a long scribal-produced manuscript in a unique script, we were dealing with a shorter inscription in an unknown language written using a version of the Latin alphabet or another known script.
There is actually a really astounding example of a surviving manuscript written in an otherwise poorly-attested pre-Indo-European language: the Liber Linteus Zagrabiensis or Linen Book of Zagreb, which is a linen scroll bearing the text of a ritual calendar in the Etruscan language dating to the third century BCE that was preserved through being reused in the wrappings for a mummy in Ptolemaic Egypt.
Four important factors, however, set the Liber Linteus apart from the Voynich manuscript. The first is the fact that, although the Liber Linteus is the longest surviving text in the Etruscan language and the only surviving Etruscan linen scroll, it is not the only surviving text in that language; scholars estimate that around 13,000 Etruscan-language inscriptions have been found, although most of them are very brief. The second factor is that the Etruscan alphabet is clearly identifiable as an adaptation of the Greek alphabet which also served as the basis for the Roman alphabet that we still use today, meaning we can easily fit the script into the known history of European scripts, whereas the script used in the Voynich manuscript does not closely resemble any script known from outside that single manuscript.
A third difference between the Liber Linteus and the Voynich manuscript is that the Liber Linteus dates to the third century BCE, a time period in which we know that pre-Indo-European languages like Etruscan were still spoken in many regions of Europe, whereas the Voynich manuscript dates to the fifteenth century CE, around sixteen to seventeen and a half centuries later, by which point Etruscan and most other pre-Indo-European languages other than Basque are thought to have been long extinct.
The fourth and final separating factor is that, although scholars are not yet able to fully translate the Etruscan text of the Liber Linteus, scholars know enough about the Etruscan language at this point that they can identify the meanings of many of the words and have some concept of the grammar. By contrast, no one has ever been able to decipher a single word of Voynichese.
In my personal opinion, the weight of evidence strongly points to the Voynich manuscript being either a deliberately constructed language or medieval hoax meant to look like writing without actually having any meaning. If it were a known natural language written using some kind of code or cipher, I’m certain that computer-assisted cryptography would have cracked it by now. Meanwhile, I find it highly implausible that an unknown natural language would be unattested anywhere except for one manuscript of clearly professional production from the 1400s, for all the reasons I discuss above. The text also has so many bizarre features, such as its extremely high rate of repetition of the same apparent “words” (even multiple times within the same line), its extensive use of repeated apparent prefixes, midfixes, and suffixes, and its very low entropy signature, which are all very unlikely to occur in a natural language.
So, yes, the idea that it might be a rare, otherwise-unattested pre-Indo-European survival language is very cool to imagine, but very, very unlikely.
That’s so cool!
Thanks for the post
Thanks! I’m glad that you enjoyed it!
Thank you Neki, for offering a comment that provoked more of Spencer’s fascinating explanations! What a great and unexpected topic!
Plato’s letter categorizations remind me of the bouba–kiki effect.
Indeed! The bouba/kiki experiment is an example of how, even today, linguists are still debating to what extent innate associations shape language. It is a very ancient debate that goes all the way back to Plato’s Kratylos, perhaps even further back to Herodotos’s story in his Histories 2.2 about Pharaoh Psammetikhos (Psamtik)’s legendary experiment with the two children to discover which language is the oldest.
An example of my own: It’s surely at least partially coincidental, but I lately noticed how, in English, words signifying largeness often begin with “m” – many, much, multi-, mega-, million, etc.