Home

Emergency Dialect

Arrival is a movie about making peace — by spreading the Anthropocene war of extinction to the stars

In Arrival, the film adaptation of Ted Chiang’s Nebula award-winning “The Story of Your Life,” a seven-limbed alien species arrives on Earth in twelve mysterious monoliths spread around the globe. As the nations of the world set their soldiers and scientists to studying the aliens, looking for anything that will gain them the upper hand in the planetary standoff, Louise Banks (Amy Adams), a linguist, is recruited by Colonel Weber (Forest Whitaker), acting on behalf of the U.S. Army, and finds herself in the middle of a foggy field in Minnesota, leading a team charged with deciphering the heptapods’ language.

Banks, who in the preceding scenes had been lecturing sleepy students about early medieval Portuguese and drinking wine alone at home, quickly learns the following: the heptapods, like us, have both spoken and written language, which she names Heptapod A and B. However, unlike with humans, the two differ significantly from each other in appearance and behavior. Heptapod A is a bit like cetacean song, but raspy and deeper. Heptapod B is an ephemeral, coffee stain-like script that encodes meaning in a self-arranging dark smoke projected by the heptapods from one of their limbs. The two are so at odds they appear to be entirely unrelated, as alien to each other as the heptapods are to humans. Banks and theoretical physicist Ian Donnelly (Jeremy Renner) seek to deduce the points in common between the aliens’ phenotype — their form of organic life — and our own, going beyond their sevenfold asymmetry and general anatomy, to questions of the heptapod’s way of thinking, and the structure of their mind.

Human language use isn’t amenable to crisp snapshots — what matters isn’t a perfect specimen, but that we grasp the mechanisms that produce them

To begin to answer the heptapod question, we have to look to the evolutionary history of life on Earth. By understanding the specification of our own species phenotype we’ll have a bit of solid ground from which to orient ourselves to the newcomers. Because this is science fiction, not fantasy, it’s safe to assume heptapods, like humans, evolved to be suited to a particular environment, in a specific world. But Banks isn’t an organic chemist, she’s a linguist, a specialist in a kind of theoretical biology that seeks to describe the computational processes underlying language, an evolutionary adaptation different from all other known forms of animal communication and unique to our species. At least until the heptapod’s arrival.


Linguists and computer scientists use a rubric known as the Chomsky hierarchy, first put forward by Noam Chomsky in 1956, which seeks to describe the major classes of formal grammars — the rules that define the possible sentences of a language. There are four types, ranked by computational power, with Type 3 being the simplest and smallest family of grammars, and Type 0 the most powerful. Any programmer is aware that some higher-level languages are more powerful than lower-level ones, but that lower-level languages are often easier to use for certain dedicated tasks that require verbose solutions in more powerful languages. The same is true for communication systems produced by evolution. Gestural systems like those found among primates are simple and highly effective: they’re based on individual signals, each associated with broad meanings like “food” or “danger,” but with no regular relations between signs, which are instead produced in an unordered and unstructured, “stream of consciousness” manner, even by primates who have been taught to sign by humans. Human language, and only human language, exhibits properties from Types 1, 2, and 3.

Heptapod B doesn’t play by the rules we’re used to. Its symbols are semasiographic, depicting meaning rather than sound, and don’t appear to follow any linear order — each expression in the script is rendered as features projecting from different segments of a semicircular backbone, like so many pools and threads of ink spreading adventitiously from a central spill. Yet, Heptapod B obviously has more going on than do monkey howls, which are also unordered; or emojis, which also encode meaning rather than sound. In the universe of Arrival, the heptapods identified something in human language that is unique to the human biological endowment among all the Earth’s intelligent species. This something is popularly known as UG, or Universal Grammar.

The theory of universal grammar has come to dominate the study of language since it was first proposed by Noam Chomsky over 60 years ago, at the dawn of the cognitive revolution. Elaborated and developed over the decades under a number of names, it is now known as “the minimalist program,” but the term “universal grammar” seems to have struck a chord with people, so the name, while somewhat misleading, has stuck. At its core, the theory is predicated on a simple hypothesis: all human populations exhibit language, all babies can learn any language they’re exposed to, and all languages appear to have deeply rooted commonalities despite their numerous differences. Human languages exhibit higher-order computational characteristics found nowhere else in the natural world, in even the most cursory of exchanges. Scientists have theorized, not uncontroversially, that the human language faculty is at least in part genetically encoded — that language, in all its mystery, complexity, and beauty, is part of our phenotype, a species-wide phenomenon that, operating from the core of what we may call human nature, has had indelibly marked the history of our species.

Like other parts of the human organism, the virtual organ responsible for language develops over the course of each person’s growth and development — in the course of normal development, the human brain naturally implements the computational processes that power language, just like it implements those processes that interpret sensory information or control our fine motor skills. According to this view, because the basic properties of language are genetically encoded, a newborn’s brain need only be provided with the correct stimuli in an appropriate environment and, like the seed of a vine planted in fertile ground, the organism will self-assemble. Indeed, linguistic and cognitive science research increasingly suggests that there is only a single human language — the language of thought, of which every other language is simply a type of dialect.


Over the decades, the list of features posited as universal to human grammar has been laboriously reduced to an almost catchy formulation: human linguistic expressions are linearized, recursive hierarchical structure, with differences in structure associated with differences in interpretation, and, in principle, no limit on the depth of hierarchical structure. Recursion is the better known and easier to grasp part of this definition: the computational feature that allows us to form arbitrarily long statements by adding new elements, e.g. “the alien > the long alien > the long alien’s spaceship > the mysterious long alien’s dark spaceship > …”

The concept of “linearized hierarchic structure” seems a little more obscure, though its effects are likely just as intimately familiar. Any fluent English speaker can tell you there is a difference between [[a mysterious alien]’s spaceship] and [a mysterious [alien’s spaceship]] — one might be an old fashioned space shuttle, belonging to a mysterious alien individual, while another is a spaceship of mysterious character, belonging to an alien whose pedestrian personality might be known to you. The linear ordering of the word remains the same; what changes is the underlying syntactic structure, which doesn’t care about linear order at all. Words follow each other one by one; when we hear them, our brain immediately begins to interpret them and try to arrange them into meaningful clusters, or phrases, which are themselves interrelated throughout the sentence. Phrase structure, and therefore syntax, is non-linear. Instead, it looks a bit like an Alexander Calder mobile.

The heptapods in Arrival, for all their seven-sided symmetry and imposing technology, have eaten from the same tree of knowledge we have

Like a Calder, what matters about the phrase structure of a sentence is the way it’s constructed — the way its components depend on each other, each word connected to another by a semantic scaffolding; sometimes complexly nesting dependencies like an exploded matryoshka, other times elegantly spare. All language users are masterful artists, capable of producing a discrete infinity of virtual sculptures out of meaning. When we pronounce a sentence, our mind, like a museum gift shop worker, “packages” the concept-structure by assigning it the appropriate wrapper — be it sound forms, hand signs, or written symbols — then  “ships” it, or as linguists prefer, “externalizes” it, from thoughtform to physical manifestation, neatly linearized and ready for our interlocutors to unpack and reconstruct, and then to enjoy the original virtual structure in their own minds. Linguists call this metaphorical packaging “spell-out,” and the pronounceable, packaged representation of a sentence “phonetic form,” or PF.

When Weber naively presents Banks with an audio recording of Heptapod A, the alien’s spoken language, he’s giving her raw data: phonetic form and nothing more. Like anyone hearing an unfamiliar language for the first time, it’s basically impossible to interpret. What’s more, unlike with human languages, which often come to resemble one another through historical or geographic association, no one on Earth has heard spoken Heptapod before. It’s with good reason, then, that Banks insists on engaging with the aliens directly, and soon thereafter, on trying written, rather than verbal, communication. If the brain is so well adapted to picking out human voices, it’s because speech is a literally “noisy” channel through which to transmit a signal. Written language is a cleaned up version of the same signal, with none of the background noise. At least this is how human language works — it’s an encoding of the same linear string that gets externalized at spell-out. Even in systems that don’t directly represent sound, like Chinese script, readers can pronounce the words they read on a screen.

It comes as a great surprise, then, when Banks discovers that Heptapod B, which should by rights be a crisp, standardized register of Heptapod A, neatly linearized, turns out to be a splotchy, disorganized mess, an inky coffee stain ejected from a tentacle and suspended in mid-air, with no beginning or end — unlike any known language, written or spoken. But language processing doesn’t stop at spell-out, and phonological form is not the only product put out by the human language faculty. If the original thought-form somehow survived spell-out and linearization, order-free yet structured, an object of pure meaning, this would begin to look a lot like Heptapod B.

This linguists call LF, or “logical form.” If phonetic form is like exiting through the gift shop at the language museum, logical form is like touching the artwork. Whereas phonetic form needs to be linearized before being passed on to the sensorimotor systems that let our mouths and hands make the correct sequence of sounds or signs, logical form is subject to different requirements — instead of being passed on to the brain’s sensorimotor systems, it’s passed on to our conceptual-intentional systems. Phonetic form is language as spoken, logical form is language as understood.

What logical form is has been a question of debate since Bertrand Russell first used the term. For him, it was an altogether different meaning system than natural language — Russell believed natural language was misleading and inappropriate for the type of logical reasoning needed to understand fundamental questions of mathematics and abstract reasoning. Now we know, though it’s hard to understand, that there are fundamental limits to what is expressible in any language; no matter what rules or atomic elements it contains, no matter what it “looks” like, there are true statements that cannot be articulated in it. Because the theory of logical form is still developing, it might be better to think of it not as any particular language, but as more of a limit that all languages approach, an imaginary “perfection” akin to that of a crystal as it forms. Representations of logical form are as varied as the point they’re trying to make, deploying mathematical tools like the lambda calculus and higher order logical notation. They look more like equations than English or Kikongo.

In reality, human language use isn’t amenable to the kind of crisp snapshots we would like it to produce — we speak in incomplete sentences, interrupt each other, and mishear. Like crystals, what matters isn’t so much that we have a perfect specimen, but that we have a grasp on the mechanisms that produce them; but, like all knowledge of the natural world, the knowledge wouldn’t come with a transcendent revelation of meaning and purpose. Logical form, like our digestive or visual system, is only a component of what makes us human. What we do with it, or how we could do something different, may be a question too difficult for human minds to understand — we may be capable of describing emergent complex behaviors without therefore being capable of understanding how to go beyond them.


The computational processes that link clusters of meaning together, chaining concepts to each other in nested hierarchies, are both eerily intuitive and strikingly exotic, and the subject of ongoing research. In the film, Banks gives Donnelly a piece of advice for speaking to laypersons about linguistics: dazzle them with the basics, she says. And indeed, the more we learn about the basics of language, the more dazzling and alien it begins to seem. So much, that if we take another look at the splotches and curlicues of Heptapod B, it becomes more and more familiar, more and more human.

We may be capable of describing emergent complex behaviors without therefore being capable of understanding how to go beyond them

This is why the heptapods came for us: not because humans are good, special, or virtuous. They came because we’re similar to them, because our world, our technology, our languages, bear the traces of a historical, convergent evolutionary trajectory that, on two different planets across the vastness of the interstellar medium, brought our two species’ phenotypes miraculously close. The heptapods, for all their faceless, seven-sided symmetry and imposing technology, have eaten from the same tree of knowledge we have. As far as the film shows us, this turns out for the best. Planetary and interplanetary war is successfully averted with the help of the transtemporal cognitive abilities unlocked by Banks after learning Heptapod B, triggering a sort of second mental infancy. Like a child learning the magic of language, or a bird first taking flight, Banks is in a sense not becoming like the heptapods, but becoming more human, more herself.

Humans call our species Homo sapiens sapiens. The second sapiens, often elided, is there for a good reason — it marks us, contemporary humans, as behaviorally distinct from our anatomically identical ancestors. The break, hard to pinpoint with precision due to the imperfection of the archaeological record, came about 80–60,000 years ago, when homo sapiens were only one of various human species. At this time, for reasons anthropologists are still investigating, the language faculty made its first appearance in our species, likely as an instrument of cognition, or a language of thought that allowed us to chain thoughts of previously impossible complexity, reason with unparalleled precision, and, ultimately, conquer the Earth.

When we evolved the capacity for recursive, hierarchically structured thought, the one true human language underlying all languages, is when our species truly became itself. It’s from the earliest signs of human symbolic behavior in southeastern Africa that the story of our life begins, and the story of countless nonhuman lives begin to end. As we spread out across the globe, diversifying our languages and expanding our technologies, developing controlled fire, perfecting tool use, becoming seafarers, harvesters, builders, and agriculturalists, we quickly drove all our human relations, our closely allied human species, to extinction. We wiped out most megafauna, permanently transformed the ecosystems we inhabited, and began that great dying-out we now know as the Anthropocene. Often, we won battles without needing to fight. Armed with complex thought and a means to externalize it, our plans and schemes have spread their effects without the need for our direct intervention. Species are doomed from the moment we look to a new landscape and begin to think of the possibilities. The judgment is passed down before we utter a word.

This is not to say we would be better off without language, or that there is a moral lesson to be gleaned from better knowledge of our biolinguistic nature. It hardly follows that any single component of our species doomed us to the current state of affairs, and it may be the case that no language, no moral or logical argument, will suffice to alter the net result of human activity on Earth. Like termites that build temperature-regulated nests because they evolved to do so, without any intent or understanding of the fluid dynamics that govern air circulation, we may be, as individual organisms, fundamentally incapable of grasping the processes we have set in motion since our long night of conquest began all those tens of thousands of years ago. And so it likely was with the heptapod homeworld.

Language isn’t an omen, but a gate, a virtual organ that opens up our organic being to what Heidegger called “the open.” At the threshold of this gate, we can capture a discrete infinity of possibility — no single fate, but an uncountable plurality of fates. That the gate allowed the heptapods to reach out to us means something, not about the timeless fate of species, but about the atomic, finite choices we make out of what we have been given. No right-thinking or peace-loving alien species would choose us, humans, for alliance. The heptapods chose us because we, like them, are consummate destroyers, so skilled at war we wage it from a distance, almost invisibly, speaking softly about co-existence. When the heptapods turned whatever sensory organs they may have to the stars, they encountered potential comrades in arms, biologically endowed with untapped capacities needing only to be nudged into activation. Born into this, we continued to become more fully ourselves.

Mathematicians can describe various types of infinities, some larger than others, each with properties that blur or push the limits of comprehension and common sense. We know language is powerful, but even an infinitely generative machine may have infinities it can never access, incommensurable truths it can never articulate. The boundaries of the human aren’t only sketched out by language, but by the whole of our being. Beyond these boundaries, our notions of right and wrong fail as they approach the limit of comprehension and our sense, so deeply held, that we are logical, moral creatures, collapses. And this is perhaps the darkest thesis of the film — disabused of the illusion of choice, and brought out into the bright open light of timeless time, it may be the case that humans would continue, earnestly and full of hope, like any animal on the hunt, to choose the future we have made for ourselves. Whether or not Arrival ever receives a sequel, we, like Banks, know quite well what darkness such a future holds.

Paco Salas Pérez is a writer, sci-fi enthusiast and computational linguist based in Brooklyn. They’re a contributing editor at the New Inquiry.