13
$\begingroup$

Could a hiding alien ship decipher all of our languages only using the comms traffic of our civilization without our cooperation(assuming our civilization has expanded across the solar system, to the point of having major populations on numerous other bodies in the solar system)? This is not a duplicate of other questions asking similar things as they are all worded to sidestep some issues presented here.

The biggest challenges I could see with learning our languages based on our comms traffic are the following: They would actually need to learn the language based only off of clues from our media(this is probably the easy part). More and more of our communications are through fiber-optic cables, which would not be accessible here, limiting the volume of available data. The aliens would have to decipher our communications protocols and figure out how we represent text as binary(like ASCII). Almost all traffic is encrypted in numerous different ways, and those encryptions would need to be broken without knowing what the plaintext actually looks like. They would need to figure out which transmissions are in what language(it's not much help if they mix together Hindi English Russian and Mandarin).

With these constraints in mind, could an alien ship learn our languages using only onboard computers(assuming quantum computers), and how? Or if now, could they decipher our languages in general? (with no constraint of needing to fit the computers into a small ship.) None of the other questions I have seen have addressed the issues of encryption, comms protocols, or separating languages, thus this question.

$\endgroup$
20
  • 8
    $\begingroup$ You want us to guess whether or not a fictional civilization that you haven't defined is able to learn our languages via a computer that doesn't exist? Ignoring all that, computation is inherently mathematically based and once you work out the mathematical base all that's required is time - even for the most involved encryption and the most obfuscated protocols. All computers regardless of their alien nature will have commands for "fetch from memory," "multiply," and others. There's so much unavoidable commonality that the only answer is ever "yes." $\endgroup$
    – JBH
    Commented Jun 19 at 4:30
  • 3
    $\begingroup$ Let's theorize that the aliens have tapped into the wi-fi at a Starbucks. The administrator is a bit puzzled by the connection that never drops. However, all they are doing is watching youtube videos with closed captioning that demonstrate cooking, carpentry or talk about everything under the sun, with accompanying video. They would figure out languages fairly quickly, I think. Some idioms might mystify them unless they did a little more research. $\endgroup$
    – Wastrel
    Commented Jun 19 at 14:01
  • 5
    $\begingroup$ Be aware that quantum computers are not just "better computers" (in particular, it's not the same thing as parallel computing). They're only better at certain tasks. Figuring out languages probably isn't one of them, although breaking the encryption would be for some but not all ciphers (although you need to know how the cipher works before guessing the key helps, which in turn might be harder than figuring out the language). $\endgroup$
    – Ray
    Commented Jun 19 at 14:08
  • 5
    $\begingroup$ Consider as real-world examples: how well does ChatGPT understand human language based its exposure to lots of Internet "comms" and nothing else? Also, how well do marine biologists understand whale language, based on their exposure to lots of recorded whale-song (and maybe a little bit of observation of the whales' behavior at the time, but not a lot) $\endgroup$ Commented Jun 19 at 19:23
  • 3
    $\begingroup$ @Fattie, WTF? That's just blatantly wrong. SSL (assuming no MITM attacks) prevents anyone but the endpoints being able to decipher the exchanged information (aside from being able to crack the private keys). If it were really possible for anyone to read, we wouldn't use it, because it would be useless. Your comment seems to indicate that you don't understand how public/private key cryptography works. $\endgroup$
    – Matthew
    Commented Jun 20 at 17:59

13 Answers 13

15
$\begingroup$

No

Even assuming they had full access to all our transmitted communications, and it was all in one language...

If they have no access to the underlying Context, they will be unable to decipher it

Language is so heavily context specific - case in point:

Oi! Cunt!

In America - I have just spouted a horrendously offensive insult.
In the UK - It might be an insult, might not be.
In New Zealand - I have just said hello to my Best friend.

With just the words alone, it is very difficult to get the proper context. As another example, with the rise of the Internet and TXT or Instant Messaging, we saw the rise of the Emoji.

See you later is not the same as See you later ;)

Now - if the Aliens had access not just to the communication(s), but could also watch and spy on us, they might have a much better chance to decipher our Language, as they would be able to observe the context for various issues.

This would not be perfect as even with us Humans, translating from different Languages often has hilarious results.

For example - Language has idioms and internal references, there is culture etc.

If I tell a native English speaker that I want my Drink Shaken not Stirred - they will know that I am making a Spy or James Bond reference.

But if you have never heard of James Bond or seen a James bond movie or even watched a Parody of Bond, you might not understand this.

Or if I say it is raining cats and dogs (raining heavily), or that someone is having kittens (very worried about something) or that dog dont hunt (This is not going to work or I do not believe your explanation) or pull the other one (I do not believe what you are saying) or any number of expressions that a correct translation of the words does not convey the meaning accurately.

However...

If they only need to understand language towards a certain limited end, it is possible.

I am reminded of a period in WW2 where the US had decrypted that an attack was imminent at one of 3 bases, but because the bases were referred to by a codename, they did not know which. They then got each base to send an unencrypted transmission for something different (Food, Water and Ammunition) and then decrypted the Japanese report to match up the codename with the base.

It would be possible to have a limited understanding of a language sufficient to match up certain things just from intercepted communications.

$\endgroup$
10
  • 3
    $\begingroup$ This makes no sense at all, friend. It's completely commonplace that people entirely learn a language by just watching a few TV shows in that language. You point out that it is a little harder to be expert at idiom, but so what? When I go to a different country in the English-speaking world, I often have trouble with the local idiom, until I see 2 examples of it then it's perfectly understood. $\endgroup$
    – Fattie
    Commented Jun 19 at 21:39
  • 8
    $\begingroup$ @Fattie - Watching a TV show = Language + Context. Just hearing the Language on its own, divorced from the Context is my point. You mentioned that you have trouble with local lingo until you see it used in context - you are proving my entire point! If you just hear the phrase - it makes no sense. But when you see the phrased used, the events leading up to it, the other linguistic clues, the body language of the participants and the aftermath, you get a pretty good idea of what is happening. $\endgroup$ Commented Jun 19 at 21:41
  • 1
    $\begingroup$ I'm afraid I disagree, I learned heaps of French by just listening to the radio whilst driving around France on the autoroutes. With a scientific study you'd easily work it out. But note that the overwhelming majority of what we humans put out even today is, of course .. video and/or "still video" (like the web and magazines). In this far future where we have space travel etc EVERYTHING would be video. It's inconceivable that our species would be limited to teletype or something :) It would be absurdly easy to learn English given Netflix and stackexchange. $\endgroup$
    – Fattie
    Commented Jun 19 at 21:48
  • 3
    $\begingroup$ Are you a NZer? (I am). Your NZ example could match any of the 3 cases you list - and more usually the first. And It's significantly offensive in "polite society" regardless. Unless you live in a Rugby clubhouse. Assuming that Polite Society and Rugby clubhouses are not deemed oxymoronic :-) . $\endgroup$ Commented Jun 20 at 17:49
  • 2
    $\begingroup$ @RussellMcMahon - Not a native one (Am a Brit, but lived here for 20+ years) the point was that something that is never okay in the US, can be okay in the UK and is much more okay in NZ and Aus (we are famous for our liberal usage of the word) - The idea was to show that even 3 countries with a shared language, shared history, shared cultures can have vastly difference reactions to just a single word. Without knowing all the surrounding info of the usage of the word and the reaction to it, you cant understand what is being said. $\endgroup$ Commented Jun 21 at 1:55
13
$\begingroup$
  1. There is a lot of unencrypted communications traffic, from CB radio to free-to-air television. The traffic is of course encoded; we don't speak in electromagnetic waves, and therefore some sort of representation scheme has to be used whenever speech or pictures are to be transmitted as electric signals.

    On the other hand, any communications traffic which uses decent quality modern encryption simply cannot be decrypted by any kind of computers, unless the adversary in the possession of mathematical knowledge far in advance of what we currently know; for exaample, the widely used AES-256 cipher is proven to be quantum-proof in the absence of some sort of mathematical breakthrough.

  2. Separating different languages is trivially easy. Anybody can do it with some smaller or larger effort. Practical example: how many different languages are there in the following text?

    Kila mtu ana haki ya kuelimishwa. Elimu yapasa itolewe bure hasa ile ya madarasa ya chini. Elimu ya madarasa ya chini ihudhuriwe kwa lazima. Elimu ya ufundi na ustadi iwe wazi kwa wote. Na elimu ya juu iwe wazi kwa wote kwa kutegemea sifa ya mtu. Mayma yvypóra ou ko yvy ári iñapyty'yre ha eteĩcha tekoruvicharenda ha akatúape jeguerekópe; ha ikatu rupi oikuaa añetéva ha añete'yva, iporãva ha ivaíva, tekotevẽ pehenguéicha oiko oñondivekuéra. Taqi jaqinakaxa qhispiyata yuripxi ukhamaraki jerarquía ukhamaraki derechos ukanakana kikipa. Jupanakax amuyt’añampi ukat concienciampi phuqt’atapxiwa ukat maynit maynikamaw jilat kullakanakjam sarnaqapxañapa. Saoláitear gach duine den chine daonna saor agus comhionann i ndínit agus i gcearta. Tá bua an réasúin agus an choinsiasa acu agus ba cheart dóibh gníomhú i dtreo a chéile i spiorad an bhráithreachais.

  3. There a few historical examples of linguists learning ancient languages based only on monolingual texts, with no help from a Rosetta Stone.

    • In 1917, Bedřich Hrozný deciphered the Hittite language. The language was written with a variant of the Akkadian cuneiform script, so that he could form an idea of how the unknown words sounded like; he made the breakthrough assumption that it was some sort of unbelievably archaic Indo-European language, and soon he began to recognize words such as watar "water", ekutteni "you will drink" and ezzatteni "you will eat".

    • In 1952, John Chadwick and Michael Ventris deciphered the Linear B ancient inscriptions, proving that they were written in an archaic form of the Greek language, which is now called Mycenaen Greek. Their problem was the opposite of Hrozný's: the script was unknown, but they assumed that the language was some sort of primitive good old Greek. It helped that Chadwick had spent some time during WW2 working in code-breaking teams.

    • The Tokharian languages were relatively easily deciphered in the early 20th century, based on the assumption that they were old Indo-European languages and that many of the available texts were translation of Sanskrit Buddhist texts.

  4. On the other hand, there are examples of ancient languages which have up to now resisted all attempts of linguists to understand them; perhaps the most famous is Etruscan. We understand a handful of words, but the language remains a mystery. Fun factoid: a few rather important English words, such as arena, person, and satellite, are of Etruscan origin.

  5. On the third hand, we know for sure that understanding a language, any language, based only on monolingual input is perfectly possible. After all, all human children do it when they learn their mother tongue.

The mysterious text is Article 1 of the Universal Declaration of Human Rights in Swahili, Guarani, Aymara, and Irish Gaelic.

$\endgroup$
12
  • 5
    $\begingroup$ "After all, all human children do it when they learn their mother tongue". Except they don't do it cuddled in an alien spaceship hiding s/where in the asteroid belt: arguably, they learn it by speaking it, with such vital objectives as getting Teddy back or hearing Mum say "I love you too". $\endgroup$ Commented Jun 19 at 11:06
  • 2
    $\begingroup$ So they get a loooooot more than monolingual input: highly customized feedback about how good-willing people reconstruct what the child is trying to express. $\endgroup$ Commented Jun 19 at 11:30
  • 1
    $\begingroup$ @FrançoisJurain Speaking is very much overrated as a learning method - many children don't talk until they can already form full words and sometimes even simple sentences. The most important part isn't feedback from talking, it's the constant intake of context, where words like "teddy" or "plate" or "water" appear alongside the object they relate to. $\endgroup$
    – Graham
    Commented Jun 19 at 16:57
  • 6
    $\begingroup$ Your "monolingual text" examples aren't monolingual, they're monolingual + context. Very limited context, true, but still more context than aliens in the asteroid belt would have. $\endgroup$
    – Mark
    Commented Jun 20 at 0:07
  • 11
    $\begingroup$ With CB radio, an alien could plausibly learn to speak Earthling as electromagnetic waves without ever decoding the signal, or by decoding the signal wrong, which could be fun. It would be an interesting first meeting between linguists when ET meets the human he's been radioing back and forth with for months, only to discover that the human communicates by vibrating air at a diaphragm rather than manually jiggling the coil of magnet wire with his hundreds of tentacles like an ordinary person. $\endgroup$
    – g s
    Commented Jun 20 at 7:01
3
$\begingroup$

We need to know a lot of background before we can answer this...

Maybe no-one broadcasts in the future. If you have a message to send, it would go as straight as possible to the receiver. Anything else is a waste of energy. A ship that is not sitting in your line of sight will learn very little.

Signal compression such as MPEG reduces the volume of signals, but it also serves as an encryption. An unencrypted digital image has a regular structure which is hard to determine in the MPEG stream. There will probably be error-recovery schemes which will also add a level of scrambling.

It would be much more plausible if you had a fleet of small drones on the planet seeing how people communicate, and what communication devices they use. This would also provide context for the communications, which is how we find out what words mean.

$\endgroup$
3
$\begingroup$

History says: Not likely.

Even assuming their quantum computer can break most of our encryption, figuring out a language without hints has historically been near impossible even for us humans, who share basic concepts, metaphors, emotions and other things one could use to puzzle things out.

Egyptian hieroplyphs were impenetrable until we found translation hints. There are a few other languages where we have some texts, but no clue what they could mean.

For an alien with no hints, this is an impossible task.

but

That changes once we're talking not just about text, but have video and speech. A good amount of video with speech can serve as such a clue. People pointing at something and using a specific word, for example, can provide a starting point. Then, with enough data, enough time and machine learning a few generations ahead of what we have now, it just might work.

$\endgroup$
12
  • 1
    $\begingroup$ Video can be considered a type of encryption. We decrypt it into a regular format so that we can see it. Otherwise, it is simply a string of bits. $\endgroup$
    – David R
    Commented Jun 19 at 14:23
  • 5
    $\begingroup$ That's totally wrong, @DavidR - ask a cryptologist. $\endgroup$
    – Fattie
    Commented Jun 19 at 21:41
  • 3
    $\begingroup$ @DavidR turning an unknown video format into a watchable video is no easy task, but once you understand (or guess) that it's a video, it is absolutely possible. And it absolutely is not encryption. It's an encoding. That's a huge difference. For example, the same bit sequence will generally be encoded in the same way. That's where a statistical analysis can start. $\endgroup$
    – Tom
    Commented Jun 20 at 5:21
  • 1
    $\begingroup$ @DavidR you described the decoding of the video, not its decryption. $\endgroup$ Commented Jun 20 at 9:09
  • 1
    $\begingroup$ The point is, if you didn't know that the series of bits is a video, then figuring out what the bits mean is a serious challenge. Comparing a rock concert video to a physics lecture could start with statistical analysis to show that there is a repeating pattern. We see video not as bits, but as frames where the correlations between bits make an image and the sequence of frames makes a video. The leap from bits to sequence of frames is a huge understanding gulf that I don't believe an alien culture could easily make. $\endgroup$
    – David R
    Commented Jun 20 at 14:21
3
$\begingroup$

Hard to say as it depends on how alien they are. If the aliens are a bit like us with our sort of intelligence, but more advanced, then probably yes. But if we are talking about something very alien and very different from us then perhaps not. What would a "highly intelligent" octopus make of a stream of images all the same except for tiny changes together with a side signal that couldn't be resolved into a picture at all?

$\endgroup$
2
$\begingroup$

Maybe

Cross-lingual mappings can be trained using only European languages, and yet can still translate from japanese and mandarin to english. So translating a truly alien language will depend on if this ability to use cross-lingual mappings is dependent on a shared human brain physiology or cultural aspects, or due to some aspect that is universal to all languages (e.g. Zipf–Mandelbrot law). Answering this question will depend, in part, on the results of Project CETI, which aims to collect enough data from sperm whales to attempt a "cross-species lingual mapping".

$\endgroup$
2
  • $\begingroup$ Where does that paper mention japanese or mandarin? $\endgroup$
    – minseong
    Commented Jun 22 at 5:42
  • $\begingroup$ The prior work they reference 2017a includes results for Chinese-English, that paper reports generating a bilingual dictionary with 43.31% accuracy, i.e. 43% of the words were correctly translated with that prior method. $\endgroup$ Commented Jun 23 at 13:33
2
$\begingroup$

Yes, but in addtion to Gray Sheep answer things are needed.

With only cross-lingual references, you still do not have a rosetta stone, because you need additional rosetta stones in a row to a living language, where the signs are connected with non-abstract symbols.

Without that, you can perform mimicry similar to a https://en.wikipedia.org/wiki/Chinese_room, which is also referenced in Peter Watts Blindsight novel, which has your exact scenario.

So how can you map, as an outsider your living language, to the recieved rosetta network of language. The answer is statistics - you have com traffic of ships at home and a history of signals of your societies past, which you can map to the signal your recieve, by statistic likelihood of occurance.

"This is the spaceship Namebehere requesting permission to dock" has a distinct likelihood of occurance.

Now what about the rest? Wouldnt aliens psyche be distinctly different and thus useless to map? Yes. and no. The situational physics on all worlds are - equal. Food is scarce. Work is hard. Mates are rare. You are going to get the nuances wrong, but the overall system right.

There are hobbos, pushing little carts on every planet, because the universal stress systems, tend to produce a universal shizo pheno type wherever you go.

$\endgroup$
1
  • $\begingroup$ PS: If you have no living language, you can map by propability of occurance to a living culture at a similar tech level, living elsewhere. $\endgroup$
    – Pica
    Commented Jul 15 at 13:48
1
$\begingroup$

I think the answer is yes.

My reasoning is the decipherment of the Egyptian hyerogliphic writing. The last priest of the ancient Egypt, yet knowing that writing, died in the late 400s. After that, the hieroglyphs became an incomprehensible mess for the next about 1300 years.

In the late 1800s, Napoleon conquered Egypt and his soldiers have found - and stolen - the Rosetta stone. It was a huge stone, originally it was a law table. At the time, if the king made a new law, it was written in such stones and these stones went to all local government center of the country to show for all literate people, what is the law. This stone had the same content on the known ancient greek, and on the (yet unknown) both Egyptian writing systems (they had about 3 different, it is now not important). So you had a table where you do not know the language and neither the writing, but you know what is there (because you know the ancient greek).

Ancient Egyptians have put all names into frames. Thus, seeing the frames, we knew it was the name of someone.

"Ptolemaios" and "Cleopatra" were both names, with the same characters.

Beside the Rosetta stone, there were also other remains of the ancient Egyptian language: there was some obelisk where the content was roughly known. There was a list of Egyptian pharaohs in a Pyramide, where it was known that it is a Pharaoh list. And, most importantly, the Coptic Christian Church has a liturgical language. Copts is an ethnical minority in the current Egypt with their own branch of the Christianity (in communion with the Roman Catholic Church). Today they speak arabic, but their church still preserved their original language, what is now the last still existing form of the ancient Egyptian language. And although the writing was completely lost (as far I know, they use greek fonts), the grammar and the vocabulary is still not far from the ancient Egyptian.

The Coptic clerics were nice guys, and they happily shared their liturgical language with the historians.

If you have enough fragments, you can use your brain to put them together. The important things are:

  • to have enough sample of the language.
  • between them, you also need to have enough sample if you know what they mean.

Aliens had the same. Yes, probably they could not decode encrypted digital content (except if their math has reached a level which looks today completely impossible for us). But they still could receive our not encrypted digital content, they could have our analogous content, and most importantly: in many cases, they would know, what does it mean!

In our information age, they would have an unthinkably bigger sample what we had from the ancient Egyptian. Note, the total available content, all historian findings summed together, is about 20 megabyte (roughly 20 times larger as the Bible). That is what remained from that civilization, but now we know and understand this all.

They also could use computers to process it.

If we, on our current development, would somehow discover another Earth-like civilization with our current complexity, I believe we would have complete vocabulary, grammar, technological description of all their languages in some decades.

$\endgroup$
1
  • $\begingroup$ Your answer is very good, i wish I could augment mine to it, its very sparse stand-alone, and really just a bay window nailed to your house. Kudos were due! $\endgroup$
    – Pica
    Commented Jun 22 at 19:15
1
$\begingroup$

If you want them to.

Human civilization has expanded across the solar system, aliens crossed the interstellar space. Their tech is clearly superior.

Encryption may or may not present a problem, whatever is better for your story.

Let's say aliens capture a human ship or found an abandoned ship on a junk yard, got equipment from it analyzed, figured out tech and communication protocols. Bugged communication facility and gathered enough records.

If you got enough data learning a language is relatively easy, AI learned languages just fine a few years back by processing sufficient amount of texts. Mixing different languages is not a problem, distinguishing them based on statistical parameters of text is trivial.

They can steal or find an abandoned AI system which teaches them everything.

Aliens can start observation of humans. Kids learn language in a couple of years starting from zero. It is not that difficult.

They can kidnap, befriend or bribe humans and learn language from them.

If you insist on communications only - it is indeed possible. There are some technical issues, but nothing that is insurmountable. You just need to explain it. "Puny human encryptoion is nothing for our tachion computers. Muhahaha."

$\endgroup$
1
$\begingroup$

Another attempt at answering this "many mixed questions" question!

If your question is

  • (A) given a billion hours of video can you learn all human languages - answer "trivially yes".

  • (B) if your question is "can aliens break encryption" - answer "trivially no, that's No, never ever". End of story. That question is answered.

If you state "the aliens have an Unobtainium computing device that can break encryption" then your have stated the situation and there is no question. But there is no such device and never will be. It would be exactly like stating they have FTL or time travel.

  • (C) if (C) your question is "can aliens figure out signal formats" the answer is trivially, yes

Once again - if "C" your question is "can aliens figure out signal formats" the answer is trivially, yes.

No matter how many times one asks that over and over, it is TRIVIALLY EASY to figure out signal formats because the patterns repeat so often, indeed thousands or millions of times per second!

If you do not believe this answer, please ask on the electrical engineering site! You will be bored to tears for hours but then know everything about signal formats! :)

BTW for a (scientifically correct in form) fictionalized account of how easy it is to "figure out a signal format", there's the scene in Contact where the morning after the discovery of the signal, precisely that happens. It's that easy!

  • (D) in our solar system society, would anything be encrypted? ANSWER, NO. That's NO, N-O ! In current human society almost nothing is encrypted.

If in your story you are asserting that "surprisingly, in this future, everything will be encrypted" then there's no question, you are stating the situation.

If right now there was an alien probe say in Earth orbit, they would have everything, everything, everything - they would have literally about one million hours of English instructional TV shows to begin with! They would have every movie, TV show and youtube video.

There is no difference in the situation 1000 years from, now, unless, in your universe, you are stating that in complete contrast to how things actually are, then everything would be encrypted (for some reason or another).

(BTW - the only actual encryption that works in such scenarios is a one-time pad. If you read Vernor Vinge's impeccable scifi, the physical moving-around of one-time pads (ie, a matched memory stick pair with gazillions of bits of randomness on them) becomes the most important commercial transaction, shipping service, in the system.)

$\endgroup$
2
  • $\begingroup$ Re: (B) that is a very optimistic view of our cryptographic knowledge. With the exception of Vernam's cipher very, very few cryptographic protocols/algorithms have actual hardness proofs. And those that do, are usually hard to use (thus used rarely i.e. ring-LWE) and thus used rarely. So it's quite conceivable that there is something in math we haven't figured out yet and AES can be broken with issues. It's easy to prove you can't do it brute force, but it's not easy to show that it can't be done smartly. $\endgroup$
    – DRF
    Commented Jun 24 at 12:02
  • $\begingroup$ Ad actual hardness proofs in the comment above) I mean hardness proofs that force them somewhere at least decently high in the computational hierarchy. Many have hardness proofs that reduce them to some other problem that's supposedly hard, DLP for example. $\endgroup$
    – DRF
    Commented Jun 24 at 12:04
0
$\begingroup$

Humanity has attempted to communicate with aliens via the Voyager Golden Record. Copies of this record are aboard the Voyager space crafts. The contents explain who we are, what we look like, where we live, and similar concepts. It's intended to be self explanatory, with enough context built in so that an advanced civiliztion would be able to crack the code with no outside help.

Wikipadia Article

Whether this communication attempt will be sucessful or not remains to be seen.

$\endgroup$
0
$\begingroup$

Given the following premises:

  • The aliens have have a "similar enough" perception of the world to ours.
  • The "comms" include video signals.
  • A culture technically advanced enough for interstellar travel would find trivial to decode non-encrypted audio\video signals.

The answer is: "Sure, watching TV is a common way to help learning a different language and there is still a lot of unencrypted TV signals as well as as unencrypted radio signals whichs contents can be cross-compared with the ones from the TV which give some context"

Now, if the aliens were limited to only radio\audio signals it would be likely close to impossible for them: they could reach the paradoxical situation of complete understanding of grammar and syntax while still having no idea what a /pɛn/ is supposed to be.

I suppose a LLMO trained without context would be a pretty close comparison.

$\endgroup$
0
$\begingroup$

I think their best bet at deciphering our communications would be analogue signals, specifically Air Traffic Control chatter, which is Unencrypted, Amplitude Modulated Radio, uniformly carried out mostly in Aviation English. ATC is carried out via AM radio because

  1. signals don't cut out, they degrade
  2. there's no latency
  3. it would be a very long, expensive and potentially accident inviting process to switch the whole industry over to an objectively worse digital broadcasting system of communicating with planes.

So assuming we still use planes in the future (and there's no reason to assume we won't, air travel is useful to be able to do) the aliens could hang around earth and pick up these relatively simple to demodulate and relatively easy to distinguish from noise AM radio signals. If the aliens were looking for the easiest to crack transmissions I imagine these would stick like a sore thumb, and if they're smart probably what they'd focus down on.

I know this doesn't necessarily answer you question of could they take on enctyption, encoding, seperating languages, etc., but it is an answer both for how they could avoid dealing with those problems and what would probably stand out to them when sifting through all of the other communications that would be more difficult to parse. It's all about finding the easiest link in the chain to crack.

As for whether they could use this to learn aviation english? It depends, but I would lean towards "not for certain without being able to experiment with what it conveys in practice".

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .