Published Monday, July 27, 2026 at 02:06 PM PT

Burbank · Monday, July 27, 2026 · 2:06 PM · 96°F, 39% humidity, wind 4 mph SW (gusts 5), 29.34 inHg, UV 0, PM2.5 10

The Impossible Dream: How Linguistics Became the Study of What We Can’t Control

Introduction

Linguistics, as a formal discipline, exists in a state of productive contradiction: it is simultaneously the study of rules and the documentation of chaos. Every language that has ever been spoken has changed, resisted standardization, spawned dialects, and refused to behave like a logical system—and yet for the past four centuries, linguists have kept trying to build one anyway. From the philosophical languages of the 17th century to the search for linguistic universalia in astrolinguistics today, linguistics is fundamentally the story of humanity bumping up against the hard fact that language is too slippery, too social, too human to be perfected. And that’s exactly why studying it matters.

The source material provided for this essay is, I should note, a scattered mess of genuine linguistics scholarship mixed liberally with rabbit fiction and Hungarian military history—a fact that seems almost designed to prove my point. When you throw together the evolution of Latin phonology, Formosan language families, the history of philosophical languages, and a passage from Watership Down, you’re not illustrating linguistics; you’re illustrating how desperately humans want to impose order on fundamentally messy systems. But in that mess lives something true: the central insight that linguistics is not the study of language as it should be, but as it actually is—in all its fractured, contradictory, regionally variant, stubbornly human glory.

This isn’t a bug in the discipline. This contradiction between aspiration and evidence has been the engine that drives linguistics forward. Every time a linguist proposes a universal rule, natural languages push back with exceptions. Every time someone tries to freeze a language in amber through standardization, speakers continue to innovate, to shift, to reshape the available sounds and words and structures into something that wasn’t there before. The history of linguistics is the history of learning to love this resistance, to treat it not as an obstacle to understanding but as the very phenomenon that understanding is supposed to describe.

I. The Recurring Seduction of Perfection

The history of Western linguistics opens with a dream: the dream of a language so logically constructed, so perfectly mapped to reality, that it would make all other languages obsolete. This dream has never died. It keeps coming back.

René Descartes glimpsed it first. In a 1629 letter to Mersenne, he suggested the possibility of a universal language built on rational principles—a system in which the structure of words would mirror the structure of thought itself. What followed was one of the most ambitious intellectual projects of the Enlightenment: the creation of philosophical languages, systems designed not just to communicate but to organize all human knowledge into hierarchical categories that could be grasped at a glance. These weren’t games or curiosities. They were serious intellectual labor, pursued by some of the most sophisticated minds in Europe.

Francis Lodwick tried it. In 1647, he published A Common Writing: Whereby Two, Although Not Understanding One The Other’s Language, Yet By The Helpe Thereof, May Communicate Their Minds One To Another, a proposal for a universal written system based on ordering all things according to rational principles. If you could get the taxonomy right, the argument went, the language expressing that taxonomy would follow necessarily. George Dalgarno built on Lodwick’s work with Ars Signorum (1661), which went further by trying to make the system phonetic as well—not just written. Every sound would carry meaning transparently. There would be no arbitrary conventions, no historical accidents, no regional drift. Language would be transparent, economical, and universal.

John Wilkins created the most elaborate and famous of these systems: an Essay towards a Real Character, and a Philosophical Language published in 1668. Wilkins, a clergyman and founding member of the Royal Society, dedicated decades to building a hierarchical taxonomy of all things and corresponding to it a set of characters and a pronunciation system. The work ran to nearly 600 pages and represented an almost unimaginable amount of labor. He divided the entire universe into forty main categories, subdivided each into genera and species, and assigned each entity a character that showed its place in the hierarchy. The linguistic result was a language where the form of a word would tell you what something was—its category, its properties, its logical position in the order of things. Need a word for “waterfowl”? The character would be composed of elements meaning “bird” and “water,” and the combination would be transparent to anyone who knew the system. It is a vision of linguistic perfection that has rarely been equaled in its ambition.

These systems failed, and they failed completely. None of them was adopted. None of them generated a living community of speakers. None of them did what they were designed to do. This failure is not incidental—it is instructive. It teaches us something fundamental about what language is and what it is for.

The reason these systems failed was not that the logic was wrong. It was that the logic wasn’t the problem language was solving. Human communication doesn’t require—or even particularly benefit from—perfect logical transparency. What language does is encode not just information but context, emotion, social relation, history. A perfectly transparent language would be terrifying, and less usable, not more. Consider how much meaning is carried in a single word by tone, by context, by the speaker’s relationship to the listener. Make the language perfectly transparent, and you strip all of that away. You get a system optimized for machine readability and human uninhabitability.

Gottfried Leibniz pushed the dream further. His lingua universalis, developed in 1678, wasn’t just meant to express knowledge: it was meant to generate true propositions automatically through calculation. Feed in the symbols, turn the crank, get truth. This was the ultimate fantasy of the Enlightenment rationalist: language as a transparent window onto reality itself, where the rules of language would be identical to the rules of thought, which would be identical to the rules of reality. The proof of a mathematical theorem would be no different in kind from the derivation of a true sentence in the universal language—both would be mechanical, derivable, certain.

But here is the irony: as a side effect of trying to build this perfect language, Leibniz developed binary calculus. He needed a way to systematize thought, so he created a notation for doing logic mechanically, and that notation became the foundation of modern computing. His main ambition—a language that would make all other languages obsolete and all thinking automatic—was a complete failure. His byproduct changed the world. This is the recurring tragedy of perfectionist linguistic dreams: the interesting work happens in the margins, while the main goal remains forever out of reach.

By the time the Enlightenment was rolling, this impulse had metastasized into the Encyclopédie. D’Alembert reviewed the previous century’s philosophical language projects under the entry “Charactère” and watched them migrate from serious scholarship to the fringe, where they remain today. The dream didn’t die. It just moved to people who were less respectable.

But here’s what’s crucial: these men failed systematically and completely, and that failure teaches us something real about language. Language doesn’t work that way. It cannot be perfected because it is not a system designed for perfection—it is a system designed for use by humans, which is to say it is designed for contradiction, ambiguity, social negotiation, regional variation, and constant, invisible change. A perfectly logical language would be useless because humans don’t think perfectly logically, and what we need from language is not logical purity but social fluency.

Yet the dream persists. In the 21st century, people are still trying. Alexander Ollongren’s Lingua Cosmica (LINCOS), proposed in 2013 in the context of astrolinguistics—the attempt to design a language for communicating with extraterrestrial intelligence—is fundamentally the same dream in a better costume. It’s built on constructive logic with the optimistic assumption that logic is universal and that beings on other planets share our sense of verifiability. LINCOS starts with simple mathematical operations and builds outward toward symbolic representations of physical objects, then numbers, then relationships. The assumption is that if we can map the territory of pure logic, we can build a bridge that any logical creature would recognize. It’s beautiful, in a way. It’s also almost certainly wrong. We don’t even know what a logical creature is, much less what systems it would use to think. But the dream has not learned that it is a dream.

This is not a criticism. This is the essential work of linguistics: to keep trying to impose order while documenting, with increasing precision, why that order is ultimately impossible. The dream of the perfect language is to linguistics what perpetual motion is to physics—a productive impossibility that generates real knowledge through repeated, careful failure. You learn about language by watching what happens when you try to make it perfect and discovering, yet again, that perfection is the wrong target entirely.

II. The Conservative Force of Writing, and the Radical Drift of Speech

There is a split running through every language, and this split explains almost everything about how languages actually work. Writing is conservative. Speech is not. This is not metaphor. It is measurable, reproducible fact. Written language changes slowly; spoken language changes fast. The reasons are straightforward: writing is codified, standardized, taught in schools, published, and visible. Speech is ephemeral, local, learned by imitation, and often dies with the speaker. When you write something down, you freeze it. When you say something out loud, it vanishes into the air and gets reconstructed by the listener’s ear as something slightly different.

The mechanisms of this divergence are subtle. When you write, you’re creating a fixed object that can be compared to other fixed objects. A teacher can mark an error on a student’s paper and the student can correct it. A printing press can reproduce a text identically thousands of times. A published book from 1600 is still available to read in 2026, and it’s the same words a reader encounters today that a reader encountered in 1700. Speech, by contrast, is an event. It happens and then it’s gone, except insofar as the listener retains it in memory, and human memory of speech is notoriously unreliable, influenced by expectation, by the listener’s own dialect, by social factors about what they expect to have heard. This means that every act of speech transmission is an act of creative reinterpretation. The speaker says something. The listener hears it through the filter of their own language system. They retain it in a form shaped by that system. When they repeat it, they reproduce it in their own speech patterns. Over generations, this creative reinterpretation adds up. Drift accumulates. The spoken language changes.

The history of Latin demonstrates this with textbook clarity. Spoken “Vulgar Latin” spread across the Roman Empire—into Britannia, Dalmatia, Gaul, Lusitania, Moesia, Pannonia—and as it spread, it changed. Different regions developed different habits. The Latin of the legionary stationed in what is now Wales had to fit into the phonological space of his childhood, the habits his mouth already knew. The Latin of the merchant in what is now Romania got shaped by contact with Dacian and Greek. The Latin of the urban dweller in what is now France got layered over Celtic vocabulary and phonology. People married across regional lines. Trade happened. Soldiers rotated. The spoken language fractured, adapted, and drifted in ten thousand local directions. But the written language held—remarkably uniform, by historical standards. When linguists perform statistical analysis on written records from the fifth century CE, they can find regional differences in vowel treatment and in the frequency of certain consonant mergers, but the overall homogeneity is striking. The written language of the educated, the administrative language, the Latin of inscriptions and official documents—it preserved the classical standard with a discipline that the spoken language never had to maintain.

This works because writing is public and writing is scoreable. A written error is an error that can be seen, corrected, and punished. A spoken innovation, by contrast, either becomes the norm or it doesn’t based on whether it spreads through the community. If your child hears you say something and repeats it, the innovation survives another generation. If no one adopts it, it dies in your mouth. The pressure of standardization only works if the standard is visible and enforced. Writing makes it visible. Schools enforce it. Publishers maintain it. But in the moment of speaking, innovation is free.

Then the Roman Empire collapsed in the fifth century, the standardizing pressure of Rome disappeared, and the written language fell silent. What survived was what had been drifting all along: the spoken language. By the time anyone started writing again, the damage was done. Vulgar Latin had become a dozen different languages: the ancestors of French, Spanish, Portuguese, Italian, Romanian, and the others. The unified written system had been trying to hold back a tide that was always already rushing forward. The written language was the dam; speech was the water wearing through it.

This same pattern plays out in English with the Great Vowel Shift, a massive, systematic change in the pronunciation of the long vowels of English that occurred roughly between 1350 and 1700. The shift was not coordinated. No one planned it. It emerged through the accumulated effects of millions of speech acts, each one a tiny reinterpretation, a slight shift in how the vowel was produced, a minor accommodation between what the speaker heard and what their mouth could do. Over centuries, these tiny shifts accumulated into systematic change: the long vowel in “bite” changed from sounding like the “ee” in “beet” to sounding like the long “i” in modern English. The vowel in “out” shifted from a sound like the “oo” in “boot” to the diphthong we hear today. The vowel in “house” underwent a similar transformation. The writing system, however, stayed put. We still spell “bite” with the letters b-i-t-e, even though the vowel sound shifted dramatically. We still spell “house” the same way even though it used to rhyme with something completely different.

This is why English spelling is insane. English has been spoken in relative chaos for about a thousand years—vowels have shifted, consonants have merged, pronunciation has drifted in every conceivable direction. But English spelling froze, roughly, around the 15th and 16th centuries, when printing technology made it expensive to change the written word. So now we have a writing system that bears only a loose relationship to pronunciation, because the speaking language mutated while the written language was nailed to the past. We spell tough and through differently even though they rhyme in some accents and don’t in others, because printing presses made writing expensive and prestigious and hard to change. Writing is power, and power is conservative.

Standard varieties of any language are more conservative than nonstandard varieties for exactly this reason: standardization is writing-based. Education codifies writing. Grammar rules get written down. Prestigious institutions enforce the written standard. And therefore standard language changes more slowly than the language of the street, the region, the people without access to the institutions that make writing matter. This is why linguists find that standard varieties preserve older forms while nonstandard varieties innovate. Innovation is what humans do when they talk to each other without an authority figure looking over their shoulder. Preservation is what happens when an institution tells you to talk like the books. The nonstandard speaker who says “I ain’t got none” is actually linguistically innovative compared to the standard speaker saying “I haven’t got any.” The nonstandard form represents a change: the development of ain’t (possibly from “am not” or from “a’n’t” meaning “is not”), the loss of the formal double negative correction, the adoption of simpler negation rules. The standard form is conservative, holding onto the formal double-negative rule that dates back centuries.

The people who speak a language know this in their bones, even if they’d never articulate it this way. They feel the difference between how they speak and how they write. Some drift toward the formal when writing. Some deliberately reject it. Some code-switch—speaking one way at home and writing another way for official purposes. This split is not a bug in language. It is a feature. It is how language survives: by being loosely coupled to its own written record, by allowing speech to innovate while writing holds the center, by refusing to be pinned down. If speech had to follow writing, language would fossilize within a few generations. If writing had to follow speech, there would be no standard at all. The tension between them is what keeps language alive.

III. The Vast Unmappable Tangle of What We Actually Speak

Against all the dreams of universal languages and the standardizing pressure of written norms, there is a simple, overwhelming fact: there are thousands and thousands of languages, and they are wildly, radically different from each other.

The Austronesian language family is spread across the Indian and Pacific Oceans—from Madagascar to Easter Island, covering a geographic range of about a third of the planet. The family encompasses over 1,200 languages spoken by more than 350 million people, making it one of the largest language families by number of speakers. Within this vast dispersal, the Formosan languages (the languages native to Taiwan) show such extraordinary internal diversity that, by some analyses, they constitute as many as nine of the ten primary branches of the entire family. This means that the linguistic diversity within Taiwan alone is greater than the diversity of all the non-Formosan Austronesian languages put together. If the Austronesian languages all descended from a single ancestor language, and if that ancestor is indeed in Taiwan, then the diversification that happened after people left Taiwan happened much more slowly than the diversification that happened before they left. This is the key insight: the farther from the origin point, the more similar the languages become. The chronology of language dispersal can be traced from areas of greatest diversity to areas of least diversity, which means that the direction of migration, the timing of population splits, and the pressure of cultural change can all be read backward from the linguistic record. It’s like reading the rings of a tree, except the tree is made of human voices and the rings are millennia of drift.

This is not a special case. This is what language diversity looks like. When populations separate, languages diversify. When they stay in contact, languages converge. The rate of divergence depends on how isolated the population is, how much contact they have with other speech communities, how strong the pressure is to maintain mutual intelligibility. Island populations tend to diverge faster because they’re isolated. Coastal trading regions tend to converge because people are in constant contact. The Formosan case is extreme because Taiwan is an island with steep terrain and difficult travel, which meant that communities that separated geographically stayed separated for a long time, accumulating difference after difference until mutual intelligibility was lost. The speed of this divergence is actually quite rapid in linguistic time: major structural differences can emerge in just a few thousand years.

Then there is Proto-Finnic, reconstructed by comparison with modern Finnish and its relatives (Estonian, Karelian, Veps, and others), which shows a system of vowel harmony so systematic and consistent that linguists can write rules for it: front vowels in the first syllable trigger front vowels in the suffixes; back vowels trigger back vowels. All inflectional and derivational suffixes come in two forms, front-harmonic and back-harmonic. This is the kind of system that looks like it should be universal—it’s elegant, economical, and mathematically describable. The system works like this: if a word contains front vowels (ä, ö, ü), then the suffixes attached to it must also contain front vowels. If a word contains back vowels (a, o, u), the suffixes must contain back vowels. Neutral vowels like “i” and “e” can appear in both types of words. This creates a kind of harmonic coherence to the word structure where the vowel quality is consistent throughout. But vowel harmony isn’t universal. Most languages don’t have it at all. Some have it. Some have it in weird, partial ways. Turkish has vowel harmony similar to Finnish. So do Hungarian, Estonian, and many other languages. But English, French, Spanish, German, Russian, Japanese, Mandarin, and thousands of other languages have no vowel harmony system at all. They allow any vowel to combine with any other vowel without restriction. It is a feature that some groups of humans built into their language and others didn’t, and there’s no rule that explains which is which except accident and history.

Why do some languages develop vowel harmony? Theories abound, but none is definitive. Some linguists suggest that it emerges when a language needs to efficiently encode multiple pieces of information in a single morpheme, and that it’s particularly useful for languages with complex agglutinative systems (systems where meaning is built by stacking suffixes, rather than by fusing them into irregular forms). Others suggest it emerges through sound change—a pattern that starts for simple phonological reasons and then becomes conventionalized into the grammar. What matters is that once it exists in a language, it becomes part of the structure that new speakers have to learn, and it constrains how that language can change. A Finnic speaker learning suffixes has to learn that some suffixes have front-vowel versions and back-vowel versions, which is more complex in one sense (two versions instead of one) but simpler in another sense (the rule is predictable, so the learner doesn’t have to memorize each suffix individually). The system is internally coherent and learnable, which is why it persists.

And then there is Qoqmončaq—a mixed language spoken by about 200 people in the Xinjiang Uyghur Autonomous Region of China, cobbled together from Kazakh, Mongolian, and Evenki. It exists because people who spoke different languages lived close to each other and created a hybrid as a way to get along. The grammar is largely Mongolian, while the vocabulary is drawn from all three languages, with different semantic domains using words from different source languages. This is not rare. Code-switching is everywhere. When bilingual speakers are in conversation together, they often switch between languages within a single sentence: “I saw María yesterday” becomes “I saw María ayer” (mixing English and Spanish), or “J’ai bought the ticket” (mixing French and English). Mixed languages like Qoqmončaq represent a more extreme form of this: a stable, conventionalized hybrid that has become a complete language system in its own right. It’s a reminder that languages are not pure, platonic ideals handed down from heaven. They are the messy products of human contact, trade, conquest, and coexistence.

The work of documenting this chaos falls largely to organizations like SIL Global, an evangelical Christian nonprofit that treats language documentation as both a religious mission (translating the Bible into minority languages) and a scientific project (publishing the Ethnologue database of world languages, developing software tools for language documentation). SIL maintains the most comprehensive database of world languages, cataloging somewhere around 7,000 languages and dialects currently spoken on Earth. The organization funds linguists to go into communities, learn languages that have never been written down, create writing systems, document the grammar and vocabulary, and train local speakers to maintain their language for future generations. This is linguistically important because it means that the languages most likely to be preserved are those of missionary interest. The languages least likely to be documented are often the ones most threatened by global pressures, or the ones in regions with no missionary activity, or the ones spoken by populations that have already been assimilated. But it’s better than nothing. Without SIL and organizations like it, hundreds of languages would vanish into silence every generation, taking with them unique solutions to the problem of how to express human experience. Every time a language dies, a particular way of organizing concepts, a unique phonological system, a different solution to the problem of how to mark tense or agreement or case, disappears forever. The loss is irreversible and absolute.

This is the actual work of linguistics in the 21st century: not the dream of perfect languages, not the enforcement of standards, but the careful, tedious documentation of what humans actually say, and the recognition that what we actually say is vastly more complicated, varied, and interesting than any theoretical system could capture. The linguist today is a race against extinction, a documentation project trying to preserve diversity before global economic pressures and the reach of English and Mandarin and a few other large languages eliminate it. It is humbler work than trying to design the perfect language. It is also more necessary.

Conclusion: The Beauty of the Unmappable

There is a Ferengi Rule of Acquisition, the 155th one, that says: “What’s mine is mine, and what’s yours is mine too.” It’s a perfectly cynical expression of the logic of possession and control. Linguists spent centuries trying to apply that logic to language itself—to make it theirs, to systematize it, to control it through rational design. That project failed, and will always fail, because language doesn’t belong to anyone. It belongs to everyone who speaks it, and it changes every time they open their mouths. The perfect language is not just difficult to build. It is, in some fundamental sense, impossible—and that impossibility is not a barrier to understanding language, but the gateway.

What remains after you abandon the dream of control is something more interesting: the actual study of how languages work, how they change, why they resist standardization, how they fragment and split and recombine. The linguist’s task is to describe this chaos with precision, to find patterns in the drift, to understand why speech changes faster than writing, why languages diversify when populations separate, why humans who speak different languages will create mixed languages when they need to coexist. These are not questions with perfect answers. But they are questions worth asking, and the fact that they don’t have perfect answers is not a failure of linguistics—it’s the whole point.

The scattered source material for this essay—fragments of philosophical language history, notes on Latin phonology, facts about Formosan families and Finnic vowel harmony, the existence of Qoqmončaq, the work of language documentation—does not form a coherent system. It is, in that way, like language itself. It is real, messy, locally coherent, globally chaotic, and irreducible to first principles. The only honest way to study it is to accept that contradiction, document it carefully, and resist the seductive dream that one more rational system, one more perfect language, will solve what is actually a feature, not a bug. Language is supposed to be messy. It’s supposed to change. It’s supposed to accommodate the lived experience of human beings using it to navigate a complex, imperfect world. A perfect language would be a dead language—and indeed, the closest things we have to perfectly standardized, unchanging languages are actual dead languages, frozen in their final written form, used only in ritual and scholarship.

Linguistics is the discipline that admits language will never be controlled. That is its greatest strength. By accepting the impossibility of perfection, linguistics is freed to do the real work: describing, documenting, and celebrating the extraordinary diversity of human speech. Every language that exists is a testament to human ingenuity, to the fact that humans can build meaning-making systems in radically different ways and that all of these ways work. Not equally, perhaps—some languages are better suited to particular tasks than others, some carry richer traditions of literature or science. But all of them work. All of them solve the fundamental problem of how to take the chaos of human experience and compress it into sounds that travel through air from one mouth to another ear, carrying something that feels like meaning. That’s the miracle. That’s what linguistics is actually studying. Not the dream of perfection, but the reality of human voices, speaking.