Davide Castelvecchi writes for Nature about a subject dear to the heart of this old science fiction fan, what is sometimes called xenolinguistics:
Sheri Wells-Jensen is fascinated by languages no one has ever heard — those that might be spoken by aliens. Last week, the linguist co-hosted a day-long workshop on this field of research, which sits at the boundary of astrobiology and linguistics.
The meeting, at a conference of the US National Space Society in Los Angeles, California, was organized by Messaging Extraterrestrial Intelligence (METI). METI, which is funded by private donors, organizes the transmission of messages to other star systems. The effort is complementary to SETI (Search for Extraterrestrial Intelligence), which aims to detect messages from alien civilizations.
There follows an interview with Wells-Jensen, which of course touches on the recent movie Arrival (see this LH post), but I want to feature this interesting exchange:
What was discussed at the conference?
The piece that was underpinning everything there, I think, was the extent to which human language is innate. If language has a necessary innate piece, then two civilizations might have a good chance of understanding each other: that was the Chomskian approach represented in some of the papers presented at the conference. Others expressed the sense that third factors — body shape, what your planet is like — would have more to do with language and that little has to be innate. If that is so, then we’d have a better chance of understanding aliens that are similar to us than of understanding those that aren’t.
Needless to say, I am firmly on the anti-Chomsky side. Thanks, Trevor!
Even if you believe in Chomskianism here on Earth (and, as a non-linguist, I take no position apart from a gut feeling that it’s implausible) I cannot imagine why even a die-hard Chomskianist would imagine that an alien language would have very much in common with Terra-speak. Different environment, different evolution, different brains, different biochemistry, for all we know.
SETI enthusiasts tend to argue that if we can communicate with aliens, it would be through shared knowledge of common facts–prime numbers, wavelengths of hydrogen, and so on. It seems like a huge leap of faith to suppose that the elements of language would belong to this class of phenomena.
Wait, I would actually draw the opposite conclusions. If language is innate in a Chomskyan way, then it might have evolved totally differently on another planet, and we just wouldn’t have the brains to understand theirs. But the alternative is that human languages have evolved their similarities by optimizing some communicational parameters, in which case alien languages would have gone through a similar optimization process (though some of the parameters might be modified by the different setting.)
This is where I bring up the Natural Semantic Metalanguage. Its major claim is that humans have a set of approximately 65 irreducible semantic primes that form the basis of all meaning and syntax in all human languages. It’s very definitely not Chomskyan, if only because it prefers pragmatic simplicity where Chomsky sees theoretical complexity, and it starts out with semantics where Chomsky starts out with syntax.
It’s possible that an alien species might not have concepts that are expressed in English as “I”, “if”, “one”, “two”, “many”, “do” or “be”. I have a hard time imagining that, though.
ref: https://intranet.secure.griffith.edu.au/schools-departments/natural-semantic-metalanguage .
On the first day of Alan Hein’s Animal Behavior class at MIT, he always asked the class why it should make sense to treat animal behavior as a single unified subject at all. Why could we learn anything useful about the behavior of cats from studying the behavior of periwinkles (taking two examples from Hein’s own research)? The answer, he said, was evolution—and for two distinct reasons. First, all the animals are ultimately related, descended from our ultimate sponge forbears. Second, evolutionary processes had shaped the behavior of different kinds of animals, and the same kinds of evolutionary pressures in one animal lineage might lead to the development of analogous physical or behavioral adaptations in other lineages.
With aliens, we will not share the ultimate common origin, so many of the similarities that we see between human language and other animal species’ communicative behaviors would not be expected to appear in aliens. However, it is still possible that evolutionary processes could mold alien language into something not totally unlike human speech.
FM: Wait, I would actually draw the opposite conclusions.
My reaction too. But the supposition that alien advanced life would use the auditory (or visual) channel for communication at all seems shaky. There must be many more ways of sensory communication that can develop than we can imagine with our limited senses.
(And I don’t mean the simple alien mind-reading trope.)
Obligatory: comparative xenolinguistics.
Like everyone else here, I thought this sentence was going to turn out the opposite.
Pirahã doesn’t have “one” or “two”…
With aliens, we will not share the ultimate common origin
But we would share many of the same evolutionary pressures. Jack Cohen makes the distinction between universal and parochial traits; universal traits like fur, eyes and wings have evolved several times on Earth independently, as solutions to the problems “how can I keep my insides at their optimum temperature in cold conditions”, “how can I gain information about my surroundings” and “how can I move fast and avoid ground-based predators”. So it’s reasonable to assume that these would also evolve among aliens. Parochials, on the other hand, evolved only once on Earth – feathers, for example. So it’s less likely that aliens would have evolved them too.
In terms of communication, universals would be things like visual signals, auditory signals, communicating through patterns of movement. Unfortunately we only have one example of “language” – human speech – to work from…
The journalist seems to be confused about what linguistic innatism actually is. It’s pretty much the opposite of the doctrine that logical and real-world constraints are likely to play a great role in determining what real languages, however exotic, are going to be like – which is what he seems to think “innate” means.
FM and everyone else are surely right, that innatism à la Chomsky would, if true, make it much less likely that we could communicate meaningfully with aliens.
It’s one of the most depressing consequences of the (fortunately completely fanciful) theory that it entails a sort of ultra-Sapir-Worfism: if language is determined in all significant respects by our genetic equipment, then, given the place of language in thought, human beings would be utterly unable to think outside the box, ever (pending genetic changes on an evolutionary timescale: but even that would just be one box mutating into another box.) There could never be any real new thoughts (I gather Chomsky actually does think that human culture is essentially a closed system with no possibility for meaningful development.)
In its Minimalist guise, admittedly, Chomskyism makes so little in the way of potentially falsifiable claims that you could argue that it actually has been reduced to simple logical constraints. I dare say you could forcibly interpret some feature of Martian as representing Merge if you’d a mind to.
Unfortunately we only have one example of “language” – human speech – to work from…
Petitio principii, surely? Only a diehard Chomskyan would call that “one” example. To say nothing of all the different sign languages …
I have a hard time imagining that, though.
Of course you do. It is impossible to imagine things we have no experience of; we can only extrapolate from what we already know. See the laughable attempts a century ago to imagine life in the future: it’s pretty much the same as life then, except with flying cars (or whatever) and slightly different women’s fashions. It boggles my mind that people think they can figure out anything whatever about aliens based on present knowledge; I realize we can’t help trying, because we’re humans and that’s what humans do, but it’s all just pure fantasy, and if we ever do meet actual aliens we’ll have to throw it all out and start fresh (and probably toss our ignorant imaginings in the memory hole and pretend they never happened). Aliens are alien, they’re not space-faring Belgians or raccoons. I remember this being impressed on me by Algis Budrys’s Rogue Moon when I were a lad. And for the general impossibility of imagining nonhuman experience right here on earth, there’s Thomas Nagel’s famous “What Is It Like to Be a Bat?“
Oh, and I had meant to quote this exchange from the link:
Never mind about referring to a signal we don’t know; things could be much worse than that. See Peter Watts’ terrifying Blindsight
.
“toss our ignorant imaginings in the memory hole and pretend they never happened” — of course you don’t mean that. Reading 1920 predictions for 2020 tells us nothing about 2020 but much about 1920.
Unfortunately we only have one example of “language” – human speech – to work from…
Petitio principii, surely? Only a diehard Chomskyan would call that “one” example. To say nothing of all the different sign languages …
Sorry – one example of language evolving, I should have said. Just as there are lots of different taxa of feathered animal in the world, but they all share a common ancestor; but there are several winged taxa which evolved wings independently (insects, birds, bats, flying fish and so on).
If octopuses had language, then we’d have two examples and could draw useful conclusions.
Or if we knew that some human languages had developed independently from others; if we knew for sure (somehow) that the original human settlers of the New World had not had language, for example. Then we’d have two independent clades of language – Old World and New World.
Stanislaw Lem’s books – especially Solaris and His Master’s Voice – really took apart the idea that we could find a way to communicate with an alien species.
It boggles my mind that people think they can figure out anything whatever about aliens based on present knowledge; I realize we can’t help trying, because we’re humans and that’s what humans do, but it’s all just pure fantasy, and if we ever do meet actual aliens we’ll have to throw it all out and start fresh
The justification for at least some of this is SETI; if we’re looking for life with big telescopes, what should we look for? So it’s reasonable to think about “what aliens might look like” because you can reach conclusions like “we shouldn’t be looking for green things, because plants might not be green; we should be looking for atmospheres in chemical disequilibrium”.
@ David Marjanović
Determining whether a language has all the postulated universal semantic primes is a significant task taking significant resources by trained semanticists, since different languages express primes in radically different ways. Just for starters, there is polysemy. Not all primes are words in a given language: in English, some are short phrases. In other languages that have been studied, some seem to be inflections. They seem to swap parts of speech with wild abandon.
I haven’t heard of that work being done for Pirahã. At least, I’ve never seen Pirahã on the list of languages that have been studied thoroughly enough to be reasonably sure that they fit (or don’t fit) the criterion.
@languagehat
As I’m sure you’re aware, it’s a very difficult job to prove a negative. It can be done, but I’m doubtful that the statement “It is impossible to imagine things we have no experience of” is true.
As I mention to ajay below, we have experience of lots of animals that can’t pass the mirror test, so presumably don’t have any equivalent to the semantic prime “I”. Social insects are a prime example. What would a space-faring intelligent species of social insects look like?
@ajay
Interesting question: do octopi pass the mirror test? If so, that’s suggestive evidence that they have the semantic prime “I”.
Even if human languages evolved independently, the preconditions for language didn’t. They would have evolved in closely related populations and within the same set of genetic, physical and behavioral constraints, the co-occurence of which are, among other things, path-dependent. There must be innumerable interesting paths to language not taken in the history of life on earth. And that’s just non-alien life.
John Roth: What would a space-faring intelligent species of social insects look like?
I was thinking of a species with semantically complex chemical signals co-functioning as a horizontal-transfer based reproduction system. They could be insects, but what if we crowd-source intelligence itself? Like fungi or swamps reallocating to change the configuration of the brain for different tasks.
And that’s just chemistry-based life.
I imagine aliens contact us continually, in many different ways, but we don’t know how to recognize (let alone read) their messages yet, and until we can do that they’re not terribly interested in us. We’re at that awkward age when we desperately want someone to be interested in us but we’re not yet very interesting.
we have experience of lots of animals that can’t pass the mirror test, so presumably don’t have any equivalent to the semantic prime “I”.
Peter Watts, already mentioned above, is deeply interested in the difference between intelligence and self-awareness and makes it central to the plots of Blindsight and its sequel Echopraxia. Both are worth a read. He’d definitely argue that you could have a space-faring species that couldn’t pass the mirror test.
but what if we crowd-source intelligence itself? Like fungi or swamps reallocating to change the configuration of the brain for different tasks.
The trouble here is bandwidth. If you put the whole of the brain in one skull, all the bits can communicate at high bandwidth (because there are lots of interconnections). If you put different bits in different skulls, then the link between the skulls is going to be a problem. One reason why Vernor Vinge’s Tines never quite rung true; ultrasound is high-bandwidth, but it’s not that high.
Latency might not be too much of a problem; speed of sound in air is about three times as fast as the speed of a nerve impulse. Which is to say that if one of your ears were several metres away from your brain on a long stalk (and for some of you that may indeed be the case; the Internet makes no distinctions on appearance) and someone made a noise right by it, you’d hear the noise with your other ear, the one still attached to the side of your head, first.
Pirahã doesn’t have “one” or “two”…
Aliens speaking a language like Pirahã would never be able to get into space because they wouldn’t be able to have a workable launch countdown. You just don’t get the same drama and tension from a voice going “many, a few, one or two, lift off!”
“toss our ignorant imaginings in the memory hole and pretend they never happened” — of course you don’t mean that. Reading 1920 predictions for 2020 tells us nothing about 2020 but much about 1920.
I don’t mean they’d literally be destroyed and forgotten, just that they’d be pored over by cultural historians for information about the period they were made (as you suggest) and studiously ignored by people trying to pretend they knew all along what aliens would be like.
The justification for at least some of this is SETI; if we’re looking for life with big telescopes, what should we look for? So it’s reasonable to think about “what aliens might look like”
Well, yes, but we don’t need justification — we do it automatically. My point isn’t that we shouldn’t do it but that it’s fruitless in terms of yielding actual information. If it makes us feel better for the time being, that’s justification enough.
As I’m sure you’re aware, it’s a very difficult job to prove a negative. It can be done, but I’m doubtful that the statement “It is impossible to imagine things we have no experience of” is true.
And I’m absolutely confident that it is, based on my (quite extensive by now) life experience with imagining things, seeing other people imagining things, and seeing how it all turned out. But I realize it’s injurious to our human pride to accept that negative conclusion. As TSE said, human kind cannot bear very much reality. Mind you, I’m not saying that nothing we say about aliens could possibly be true, just that if something does turn out to be descriptive of whatever actually turns up, it will be sheer coincidence, just like predicting the weather a year from now.
Maybe we first need to refute/debunk (to our own satisfaction) the famous terrestrial claim that “If a lion could talk, we could not understand him,” before moving on to speculating about the possibility of communication with non-terrestrial non-mammals.
Problem with the mirror test: gorillas refuse to look at a mirror. Looking someone in the eyes is a threat, and why would you threaten yourself? That makes no sense.
Point a video camera at a gorilla at any other angle, show the image to the gorilla on a screen, and the gorilla passes the test with flying colors.
What I can tell for sure is that the words that were at first thought to mean “one” and “two” turned out to mean “very few” and a bit more than that. They’re not numerals. No numerals have so far been found, and there’s plenty of evidence for their absence.
ajay: The trouble here is bandwidth. If you put the whole of the brain in one skull, all the bits can communicate at high bandwidth (because there are lots of interconnections). If you put different bits in different skulls, then the link between the skulls is going to be a problem. One reason why Vernor Vinge’s Tines never quite rung true; ultrasound is high-bandwidth, but it’s not that high.
I thought a little about that between the insect-like creatures with chemical signals and the sponge-like with reallocating cells.
Chemical signals may have high bandwidth in the same way as a containership full of disks. And I don’t see why coding and decoding need to be slow in a highly evolved chemical sensatory system. In that case I think enzymes or pieces of DNA (or whatever serves that purpose in our alien life form) could be used to communicate between individuals in a way we do with language. The question is how that would develop. My idea was a co-opted reproductory system. Or co-evolution with one.
The civilization of blobs of reallocating cells would communicate in different ways on different levels. Between cells we may imagine a chemical system governing physical reconfiguration and an electrical system doing sensing, reasoning and imagination. But none of these would be language-like, since the semi-autonomous cells would lack consciousness on their own. On a higher level, maybe visually with shape-shifting? But why would blobs develop the capacity for conscious reasoning in the first place?
It is impossible to imagine things we have no experience of; we can only extrapolate from what we already know.
Hm… I have no experience of winged iguanas covered in purple polka-dots, yet I can certainly imagine them..
More directly, given that what you’re actually critiquing is the idea of correctly imaging things we have no experience of — if that were true, then most science and mathematics would be impossible. No one has ever “experienced” gravitational waves; yet Einstein was able to imagine them, and they have since been shown to exist. And all sorts of strange things have been imagined in mathematics; some of them have even turned to be “useful” (e.g., the Riemannian geometry Einstein based General Relativity on, or the bits of number theory used in cryptography).
Fair points, but I’m not really talking about imagining physical and mathematical facts, which are after all extrapolations from what we already know. I’m talking about imagining something like an entirely new branch of science, one with no relation to the math or sciences we already know. How can you imagine something like that? You can’t. No more can you imagine alien life, sez I.
The idea of what kind of new science we can imagine reminds me that one of Edmund Husserl’s philosophical goals was to identify how various scientific fields could have their own forms of rigor. By “rigor” (not Husserl’s word, obviously; I don’t remember what word he used in German), he meant the ability to derive new science without direct recourse to experimental data. By this standard, math has been rigorous for millennia, physics for centuries, and chemistry for decades. Biology might get there in my lifetime.
The “rigor” word in Husserl is “Strenge”. See the 1911 essay Philosophie als strenge Wissenschaft.
What is “new science”?
New scientific results. Husserl’s model sciences are logic and mathematics. He is interesting but dead.
Husserl’s model sciences are logic and mathematics.
Hm. So, not really science per se.
Mathematics is applied logic; science is applied mathematics; medicine and engineering are applied science.
Mathematics is an abstraction & generalization of how physical objects behave; logic is an abstraction & generalization of how mathematical objects behave.
What you just wrote – is that an example of abstraction of logic ? It is what it is, but what is it ?
It’s actually a ready-made critique of pure reason that the topic reminded me of.
More to the point: math & logic don’t produce “new science” because they aren’t science in the first place. Mostly anyway (rabbit hole down from Gödel).
Hm… I have no experience of winged iguanas covered in purple polka-dots, yet I can certainly imagine them.
“Anyone inheriting the fantastic device of human language can say the green sun.” —Tolkien, “On Fairy-Stories”
But this is just recombination of known parts. We can imagine a human being with just one eye, or one with wings. But we cannot even imagine a square circle, in the sense of something truly square and circular at the same time. Nor can we truly imagine an entirely new order of living things, and when we discover one, it is always a shock.
(These remarks of course apply to you normal folks who can actually visualize. For me, it’s all abstract unless I’m actually looking at it already.)
I’m not sure about that.
“But we cannot even imagine a square circle, in the sense of something truly square and circular at the same time. ”
Heisenberg’s ‘schnitt’ – the discrepancy between the quantum and classical views of matter – is an example of a concept we currently cannot imagine a solution to, though it may be reconciled in the future, and may already be understood by an alien intelligence. Likewise a solution to the mind/matter problem isn’t directly knowable within our current mode of thought, or a recognition of whether all our thought is physically determined. Aliens may well have gone beyond this, and may think of it as a kind of simple ‘mirror test’ of mental development.
Nor can we truly imagine an entirely new order of living things
The whole history of science fiction, not to mention mythology and religion, suggests that you might be off the mark with that one.
We can’t imagine a square circle because it is a contradiction in terms. (Personally, I can imagine a square circle, but I also can’t imagine it.)
The whole history of science fiction, not to mention mythology and religion, suggests that you might be off the mark with that one.
I entirely disagree. The whole history of science fiction, not to mention mythology and religion, suggests that we can only imagine extrapolations of what we already know. (God = big bearded man in sky; alien = person with extra limbs, odd-colored blood, or unusual social system.)
The Snaiad vertebrates that David linked to look like Earthly vertebrates with small changes. (By the way, I wasn’t using “order” in the technical Linnaean sense.)
Likewise a solution to the mind/matter problem isn’t directly knowable within our current mode of thought, or a recognition of whether all our thought is physically determined.
I am not so sure of that.
We can’t imagine a square circle because it is a contradiction in terms.
Pretty much this.
That said, even theoretical nonexistence of an object probably isn’t a sufficient condition for being unable to imagine it; as far as I can tell, it’s easy enough to imagine a dissection of a square into an odd number of triangles of equal area, even though such a dissection does not actually exist (as proved by Paul Monsky in 1970).
Ah, but look a bit closer. (By which I mean “read the text”.)
“Mind is what the brain does.”
If we define “circle” as all points on a plane equidistant from some point and “square” as a quadrilateral with equal sides and equal angles than some squares will be circles in Manhattan metrics.
@D.O. I also thought of the taxi cab metric making circles of squares, but I decided that my personal definition of a circle required that the distance be taken using a Euclidean metric in a flat two-dimensional subspace of a manifold.
but I decided that my personal definition of a circle required that the distance be taken using a Euclidean metric in a flat two-dimensional subspace of a manifold
I think my own definition would probably have allowed it being taken using a Euclidean geodesic metric in a sufficiently smooth, but not necessarily flat, (locally) two-dimensional surface.
Weirdly, I honestly have no idea whether that would have been enough to allow a square (it probably depends on the degree of smoothness).
[EDIT: it probably could if we allow piecewise smooth surfaces, so in that version, I actually could imagine a square circle. If I think about it for a bit I might even be able to mock up a plausible formula for it… not sure, however.]
Brett, obviously in imagining tetragonocycle something‘s gotta give. The question is what it can be that on listening to the proposal we could reasonably say “yeah, you have a point”. Like people who thought that marine mammals are essentially fish could conceivably have said “swimming in water is the most important thing, but yeah, they are mammals all right, so yes, if you insist on grouping them together with rhinoceroses, we can tolerate that”. Maybe some weird quantum mechanical construction, or thinking about a circle not as a set of points, but of a single point that traces the circle and defining a square through projections on two coordinate axes and then using this construction to measure areas. Or simply something completely banal like taking a great circle on a sphere and dividing it on four parts, if it makes somehow more sense in that world than flat geometry. Or taking an algebraic route and saying that completing the circle in four turns is somehow similar enough to doing a continuous rotation that it warrants a single term and fundamental enough that it makes sense to start your basic math with such an object.
The point is that we do have a number of dimensions in which we can extend or change our mathematical intuitions and, as matter of fact, we sort of doing it for a reasonably long time, so people trying to find an example of the limits of imagination have to do a better job than look at a preschool math.
So there’s hope for you yet, you educated stupid fools! But don’t stop at a square circle. The Wisest Human can imagine a cubic sphere! Seek in haste to attend a lecture by him.
Fools worship mechanics of language.
The things you find on the interwebs…
Don’t tell me you didn’t know it. I remember measuring Internet craziness in Timecubes (on a logarithmic scale) ten years ago.
Nope, I see this for the first time. The internet is a big place…
Apparently not on line is Shel Silverstein’s cartoon “Egg of the Greel”. The Greel is a basically spherical bird, seemingly flightless but with legs, beak, etc. sticking out past the sphere. The egg is about the same size, but perfectly cubical.
“This is the egg of the Greel.
If it makes you feel funny just to look at it,
Imagine how the Greel feels!”
@January First-of-May: I think a figure in a sectionally curved 2-manifold is acceptable to me as a 1-sphere, but not a circle.
On real life alien linguistics, specifically grey parrots.
Yes. This is obviously where alien linguistics (exo-linguistics?) should start. But in an exo-planetary perspective even parrots and humans are close kins.
“Fortunately, my ancestors spawned in another ocean than yours. My blood salts are *pause, assuring facial expression* quite different.”
– Spock, before the writers had settled how much human ancestry he had
Y: On real life alien linguistics, specifically grey parrots.
I’ve read it all now. It’s refreshing with a paper where most of the footnotes and references are YouTube links.
Noting that most research has been focused on what animals can be trained to do in a laboratory, she argues that it’s more interesting what they achieve more or less by themselves when growing up in a “language-rich environment”. Parrots, like small children, practice language by acting out dialogue for themselves, testing elements to produce different meanings. She also finds it significant that parrots and humans are the only vocal-learning species known to have a sense of rhythm and be able to (or enjoy to) follow a beat. Parrots are generally monogamous to the point where mates spend their whole life in close proximity of eachother, in a constant vocal exchange, and when stuck with humans they transfer this to a human mate.
And this is the really exo-linguistic perspective, how parrots use their linguistic abilities to communicate between themselves in the wild. She doesn’t go much further into it, since there’s not been much research on parrots in the wild, but it’s known that African Greys have unique identification calls, and at least one South American species of parrot has been shown to give unique naming calls to their chicks.
We know that parrots have a variety of calls and that they can learn to use arbitrary combinations of sequences of sound to produce a dialogue of sophisticated messages. It would be cool if some species of parrot actually had something resembling language. But even if they have a capacity for it, it doesn’t mean it’s culturully transmitted and used to store information on a community level.* It seems to me that the linguistic abilities of parrots may have developed as a bonding mechanism. Maybe each parrot couple develop a system through their life together, like a secret code, and that it dies with them.
*) There’s a lot of anecdotal evidence on community learning in corvids. One example is the guy here in town who was put in charge of removing crow’s nests from parks and cemeteries. A couple of days after he started he was greeted by angry crows everywhere he went. The story was more complex than that, involving different clothing and gear, but I haven’t been able to find the original article in the local paper. But obviously, like all anecdotal evidence it’s subject to all sorts of bias.
Typo – I mean “settled on”. I’m pretty sure they hadn’t decided yet.
As far as I’m concerned, the two wordings are equivalent.
Huh. I thought “settled” alone implied the writers had settled a dispute, probably among other people, while “settled on” meant they made up their minds on this, likely before anyone else ever thought of the question.
You can settle an issue whether or not there is debate on it.