You list a bunch of claims that mostly seem over-optimistic and that show no sign that you have thought critically about whether they are actually true.
Being able to communicate with everyone means we can travel and study and work anywhere.
No, immigration law is a stronger barrier to travel and work permits than languages are.
Since education is the key for getting out of the malicious poverty cycle, this will bring fundamental change to a major portion of the world’s population.
Education is a lot about signaling. Imaging that Western education itself without the signaling value is able to change the lives in developing countries doesn't have a good base.
When there's a language that everyone is proficient in, contracts will be much clearer, resulting in more trust in oversea partners and confidence in foreign trading.
Do you have any evidence that language of making contracts is any barrier to business? When it comes for example to China, the way contracts are enforced seems to be much more central than the wording of the contracts when it comes to trust.
But the world's economy will actually benefit the most from the leaps-and-bounds advances in science and technology. The IAL fosters much better collaboration between scientists from different nations, now that they can understand their colleagues' research published in a foreign magazine.
English works for that purpose and most fields get centered in a way that makes the important communication English.
More generally, if you would introduce a new IAL, you wouldn't start with it being spoken by everyone. To get it adopted you need to have it be useful if not everyone speaks it.
Education is a lot about signaling. Imaging that Western education itself without the signaling value is able to change the lives in developing countries doesn't have a good base.
From what I've seen studied, a base level of education (eg. basic literacy and numeracy) is absolutely significant in changing lives in developing countries, but thankfully today that's only applicable for a relatively small and decreasing number of people.
Signaling matters a lot in developing countries for access to jobs as well.
But I do grant that base literacy and numeracy are important. There are some features of Esperanto that make it easier to achieve literacy in it than most natural languages but that's quite distinct from the arguments that were made in the OP. I don't see why we would expect much better base numeracy resources.
You can learn mathematics from Khan Academy, there are popular mathematical channels on YouTube, etc. If enough people speak whatever language, there most likely will be a localization of Khan Academy in that language, and probably also at least subtitles for many of the popular channels.
This is not an argument for everyone speaking the same language. Having ten major languages would work almost as well. It just sucks to be a native speaker of a language spoken by few people (or by many people but all of them in developing countries), because then you need to master a foreign language before getting an access to the international wealth of educational resources, so you cannot e.g. learn math from Khan Academy when you are 6 years old.
There are many "open source" resources for all kinds of things, and yet a language barrier can deprive you of them. To make it worse, your social environment is probably also behind the same barrier.
One more point along those lines: you say these advantages will come from everyone speaking the same language. Well, we already have one language that's approaching that. Wikipedia says "English is the most spoken language in the world (if Chinese is divided into variants)" and "As of 2005, it was estimated that there were over 2 billion speakers of English."
From reading your post, I bet you have glowy happy thoughts about an IAL that wouldn't apply to English. If so, to think critically, try asking yourself whether these benefits would arise if everyone in the world spoke English as a second language.
Huh? Also from Wikipedia, I read that "English has between 450m & 2B speakers", so an unbiased figure should hover around 1B.
My imaginations consider a great IAL, one that has many advantages over English. So while they do still apply to English, the magnificence of each facet is limited. Taking the snowball effect into account, it's like putting a handicap of -75% on Benefits. Why settle on something inferior?
Besides, there's a reason why English can't ever be a true IAL - which I may address in the next part of the series. To put it in your scenario, it's likely the question: if everyone in the world spoke English as a second language, then what first tongue are the Americans, British and Australian... speaking?
It’s dubious if a language can be significantly better than English or any other natural language*, since there’s the fundamental limitation that human brains have to learn and produce it. Some parts can be made simpler like irregular conjugations, but natural languages are already pretty close to as optimal for humans as possible - when a language loses a feature (like noun classes), speakers will naturally start adding new features (like more morphemes) or speaking faster to maintain a constant rate of information transmitted. The obvious simplification gains would help people who are learning it for the first time, but they wouldn’t really help child learners so it only saves you one generation of teaching everyone English (or any other language).
On the contrary, I believe it's possible to improve any existing language significantly. I've already have some idea to increase the rate of transmitted information, and that's only me. When we get more brain power into this project, even better stuffs will invariably be found.
The obvious simplification gains would help people who are learning it for the first time, but they wouldn’t really help child learners so it only saves you one generation of teaching everyone English (or any other language).
Could you explain more? I don't think I really grabbed what you said, but the great IAL will not necessarily simplify everything. On some facets, it will be more sophisticated. The goal is to have a much more efficient language, not just a simple one... And that has nothing to do with saving 1 gen of teaching everyone IAL, no?
When a child is learning their native language, they don't have the same difficulties with irregular verb conjugations that second-language learners have. So getting rid of irregular verb conjugations would make it simpler for second-language learners. But once everyone has learned the language and is teaching it to their children as a native language, no one is better off because the language doesn't have irregular verb conjugations.
On top of that, there's evidence that irregular verb conjugations (and other irregularities) actually make a language easier for humans to speak, since they help with error-handling. Human hearing and speaking are lossy conversions, so redundancy helps decipher meaning. If you hear a word and think that it might be two different verbs, hearing either an irregular or regular ending can tell you which verb you actually heard. Noun class (like whether a word is "masculine" or "feminine" in Romance languages, but some languages have many more types) and grammatical cases are a couple other language attributes that can help with this.
When a child is learning their native language, they don't have the same difficulties with irregular verb conjugations that second-language learners have.
They don't have a difficulty understanding, but I think they keep making mistakes when talking... and the adults keep correcting them, day after day, and after a few years the kids finally get it reliably right. It just all happens at a small age and is quickly forgotten.
This is very interesting! Could you point me to some research links about irregular conjugations, noun class and other redundancy help with clarity in conversations? I've tried googling to no success. Anyway, if with "hear a word and think that it might be two different verb", you were referring to homonyms, then I believe there's at least a solution for that while not compromising the simplicity of a language.
But once everyone has learned the language and is teaching it to their children as a native language, no one is better off because the language doesn't have irregular verb conjugations.
This is assuming that people will deliberately try to make their next gen use IAL as 1st language, which is absolutely not what an IAL is meant to be. The core idea is for a particular person to use native tongue when conversing with the family, tribe, and people in the same country in general. When they meet a foreigner, then they'll both switch to IAL. So typically a child would learn their mother tongue exclusive for the 1st 3-4 years of life, and only starting to get IAL at age 2 at the earliest.
Of course, there will be a few parents who teach their infants IAL, as the case with some Esperanto fanatics has shown. But I don't think of this whole IAL endeavor as an individual race to be better off, i.e. "I have to learn this stuff to get ahead and step on that guy's head" (terribly sorry if I'm misinterpreting your words here). I envision a great IAL as an excellent way to dramatically enlarge the cake, and therefore bringing bigger slices for everyone.
The relevant topic in linguistics is redundancy. This article ("The role of redundancy in language and language teaching", Darian 1979) is a decent introduction to the topic, and it also talks about its role in language learning.
This article ("Redundancy Elimination: the Case of Artificial Languages", Chiari 2017) seems quite relevant to your purposes (full PDF).
if with "hear a word and think that it might be two different verb", you were referring to homonyms
I was referring more to if you hear a word that is similar to another but not identical, and then you're trying to figure out which it was. If I say "John hit the ball", you might hear instead "John hid the ball".
One example in English is the redundancy in the plural ending. If I say "Alice read the three books", there's redundancy because the "-s" ending on "books" indicates that there are multiple books, while "three" also indicates that there are multiple books. If you mishear me and either hear "Alice read the three book" or "Alice read the books", you still know that there are multiple books. If we got rid of the plural ending in Englsih, you might hear either "Alice read the three book" or "Alice read the book", and then you don't consistently know that there are multiple books.
Going further abroad, you can see an example with noun classes in German. You can compare "Er sah den Bär" ("He saw the bear") versus "Er sah das Bier" ("He saw the beer"). In English, the only difference between the sentences is the vowel in the final word, so if you misheard that vowel then you'll get a completely wrong idea. In German on the other hand, the articles are marked with the noun class ("Bär" is masculine while "Bier" is neutral), so you have two pieces of evidence to tell you what noun you heard.
If you start adding in adjectives which agree with the nouns, there's even more evidence in German but still only the one bit of evidence in English. Compare "Er sah den schwarzen Bär" ("He saw the black bear") versus "Er sah das schwarze Bier" ("He saw the black beer").
Another feature of redundancy displayed in German is noun case. In German, you can say "Der Mann sah den Bär" or "Den Bär sah der Mann" to mean "The man saw the bear". In English though, if you say "The bear saw the man" then you mean something different. This works in German because in both of those sentences "Mann" is marked as a subject while "Bär" is marked as an object.
German can have the object before or after the verb, letting you emphasize either the subject or the object by putting it in the first place. English always has to have the subject before the verb and the object after the verb though, since otherwise it's unclear which noun is supposed to be the subject. English in turn lets you emphasize a noun by putting extra stress on it.
English has redundancy built into the word order while German has the redundancy built into the noun cases. If you look at the history of Germanic languages, you can actually see quite clearly that word order became stricter in English at the same time that it lost noun classes and case. The speakers of Middle English stopped using noun classes or case but needed some extra source of redundancy, so they started to say their sentences only in a certain order. (source: "Syntactic Reconstruction and Proto-Germanic", George Walkden 2014)
It's even more obvious looking at Romance languages. Latin had a free word order, few prepositions, and six cases while the modern Romance languages have a relatively strict word order, many prepositions, and no cases. (Note that they're all the same because they all evolved from Vulgar Latin which happened to have changed that way; there's no fundamental rule that free word order always evolves in that way).
Thank you! The links you provided are valuable. I can't access the full Darian article, but with Chiari there seem to be some issues with her approach. She tried to defend redundancy, but by only citing previous (very old) works and providing some comparison examples between languages. IMHO, if one is to prove something, she'd have to set up experiments. Like, recording people having conversations using a conlang with little to no redundancy, compared to using a natural redundant tongue. Then asking them to rate the level of clarity after the talks, and combining it with analysis of the recorded videos, etc. Then repeating the experiment with different pairs of language... In other words, her article is not convincing at all. Granted, maybe the presence of numerous grammatical errors in that supposedly professional linguistic paper contributes quite a bit to undermine her message.
Nevertheless, as her article suggested, the lack of any constructed language with absolutely zero redundancy may point to it being necessary for speaking. I have nothing against redundancy and try to hold a stance of 'blank slate' when it comes to IAL ideas and opinions. The goal is to build an IAL as easy to learn and effective as possible, and if some redundancy can help, then I can see no reason not.
The problem is that, not many linguists are also "LWer"s. They can be quite biased toward their own studied tongue and can't see some brilliant ways which other languages employ to solve their own one's problems. Case in point, your Germanic examples help me open my eyes to a lot of interesting stuffs. Yet at the same time, I can already formulate some ideas to eliminate a few of those issues, inspired by my native tongue. It's just that my linguistics knowledge is too limited right now to correctly express them. Well, that's why getting serious education on the topic is the 1st step of my plan :)
Your last sentence perfectly describes the main obstacle a supposed IAL will face in the process of emerging into global usage. It must prove way more effective than the next candidate in order to persuade.
immigration law is a stronger barrier to travel and work permits than languages are.
That's true. However, I'd say that if the total difficulty is 100%, then immigration law contributes around 55% and language barrier 45%. If you can eliminate the latter, then you effectively make it twice as easier to do it.
Moreover, IMO immigration law is kinda ephemeral. Trump was elected and put up some barriers. Biden then went razed them down. Meanwhile, something as basic as language is much more rooted - the US hasn't changed its official language since its birth.
Western education itself without the signaling value is able to change the lives in developing countries doesn't have a good base.
I didn't emphasize Western education anywhere in my post. Chances are you're biased against Eastern education?
Whether education is about signaling or not, isn't it just better to have more education? I think we can agree on this point.
When it comes for example to China, the way contracts are enforced seems to be much more central than the wording of the contracts when it comes to trust.
Agree! I hadn't thought about it thoroughly enough, now that you mentioned it. How about a different point? Companies always have to spend resources on the translation process when dealing with foreign partners - in your example, China. But now they don't have to do it anymore, and thus have more capital to spend on other projects.
English works for that purpose and most fields get centered in a way that makes the important communication English.
English can do the job at an 'huh, OK' level. It was patchily built throughout many centuries with not-for-scientific-research mindset. Not to mention when scientist actually go to their colleague's foreign lab for collaboration, they will face even more difficulties, because in every-day purposes, English is also far from great.
Gain = (Benefits − Costs) ∗ Probability
Would be more like gain = benefits*probability of those benefits - costs*probability of those costs, especially if there are failure modes that exist. I'd also try to avoid framing it as "benefits are almost unlimited while costs are finite;" while an IAL is great, the benefits of an IAL are just as finite as the costs are.
That being said, I think that if you can make an IAL that is exceptionally good on many dimensions and get enough interest/funding behind it, it would be an extremely worthwhile project.
Thank you! I think that's a better formula. But I don't understand the phrase after "especially if..."
Do you have 1st hunch about what kind of outside funding sources a project like this can more likely draw from?
English is the current "language Z" currently, but I agree that the language Z could be better. Are you planning on designing an IAL? I only do loglangs, so I can't help (or help to a meaningful degree), but I'm interested in reading your plans if you do have plans.
Absolutely! My plan does heavily involves designing an IAL, but it will NOT ever be an one-man project. Instead, it must seek help from all talents from a wide range of disciplines.
I'll write more about the failure of Esperanto and how I think we can do better.
In my opinion, the difficult part of things like this is coordination. There are already hundreds, maybe thousands, of constructed languages. Most of them probably have less than dozen actual speakers, i.e. people who can fluently keep a general conversation. And yet, many people decide that the best way to solve this situation is to create yet another constructed language. (Definitely not to support the one that already has a few orders of magnitude more speakers than the remaining constructed languages combined.)
So it seems to me that you either assume that coordination is easy in general, or that making the right design choices in your language will have a very powerful effect on all potential speakers. The former seems disproved by evidence; the latter kinda by outside view... consider the hundreds of authors of the other constructed languages -- they probably also believe that the impressive features of their languages will irresistibly draw humanity to their project -- what exactly is it that makes all of them wrong and you alone right?
Your opinion is correct. Coordination is actually 1 of the hardest things for a successful IAL. That's why before talking about how to choose the best design, we'll have to solve the coordination problem.
The most salient thing I've observed from conlang fora is that practically all of the constructed tongues there are built by individual authors. Most of them are full of themselves, and thus, as you precisely put it, believe that their language is so impressive every single one on Earth should run over and beg to learn. Those biased people are not the kind I'd want to invite into the great IAL project.
You've probably noticed that your post has negative points. That's because you're clearly looking for reasons why an IAL would be great, rather than searching for the truth whatever it may be. There's a sequences post that explains this distinction called "The Bottom Line". Julia Galef also wrote a whole book about it called "The Scout Mindset" that I'm halfway through, and is really good.
That said, having an excellent IAL would obviously be a tremendous boon to the world. Mostly for the reasons you gave, scaled down by a factor of 100. And Scott Alexander and I think also Yudkowsky have written about the benefits of speaking a language that made it easier to express crisply defined thoughts and harder to express misleading ones---which is an entirely separate benefit from "everyone speaks it".
One of the biggest pieces of advice I would give my past self is "start small". I find it really easy to dream of "awesome enormous thing", and then spend a year building 1% of "awesome enormous thing" perfectly, before realizing I should have done it differently. When building something big, you need lots of early feedback about whether your plans are right. You don't get this feedback from having 1% of a thing built perfectly. You get much more feedback from having 100% of a thing built really haphazardly.
Putting that all together, my advice to you---if you would accept advice from a stranger on the internet---is:
Thank you JP. There's a high chance that I don't have the Scout mindset yet, so finding the truth is hard. But to my defense, until now no comment whatsoever has been able to point out a cons of IAL, nor express the concern about the cost. So probably it's true that the IAL scenario is (very) desirable. The negative score of the post doesn't speak much: it can be a high karma member overwhelming lesser members, or it may simply reflect their opinion about my mindset, not the idea of IAL.
What you wrote about a crisp language that helps prevent misleading ideas is what I'd forgotten to include in the main post - it's actually 1 of the goals of the excellent future IAL! If other veterans have explicitly mentioned it, then that means the problem with current languages is significant, and improving it is very possible.
You don't get this feedback from having 1% of a thing built perfectly. You get much more feedback from having 100% of a thing built really haphazardly.
That's an insightful advice. My plan, I think, more or less align with that view. I may attempt with something similar to Toki Pona - that is, a simple project not with IAL in mind. I'm a bit afraid that if I did a miniature version of the IAL, I would be anchored to it, seeing it in all glory and thus not be able to cooperate properly with fellow linguists & scientists & experts in other fields when it comes to the real deal. Something like the first-born child almost always has the most love. Meanwhile, my vision of the great IAL is that it must be a big collaboration right from the very start. But I'm not too worried, because I do have ideas of smaller projects that will test my enjoyment for elements needed for IAL.
As a note on this:
I reason with myself that, since the Benefits to the world are almost uncountable while the Costs are finite, even if the Probability of success is small, the Gain for humanity is significant. Something along this line: (9e15 − 7e13) * 1% = 89.3e12 ($)
This is not valid upon reflection because an identical argument holds for any other action, say, everyone buying a tomato. If actions have their effects multiplied by a large number, then we should look for relatively good actions. If the chance of buying a tomato having big effects is small but the gain is really large, then we should look for more effective paths to big gains.
... Uh, I don't really get your example with tomato. In that case, the cost is relatively small and the benefit is also just marginal. The Probability of getting everyone in the world to buy a tomato is decent - say, 50% - because it's easy and many of us already bought it in our life. It's quite different from IAL, yes?
Therefore I couldn't follow your line of argument. Could you provide examples or elaborate more on 'effective paths to big gains'?
In summary, I disagree that large expected gains is an argument for action when applied like this. Normally, this heuristic is fine, but here you are multiplying by many instances, meaning there are other likely actions.
In summary of my summary: opportunity cost will sting at scale.
I was delighted to run across this post; I joined lesswrong so that I could reply to it!
There have been some pivotal changes since the push was made to adopt Esperanto as an IAL; perhaps the idea should be explored again.
One of these changes is AI. It should soon be possible to make an "AI friend" -- basically a glorified tablet, perhaps with a simple arm - that could serve as a baby monitor and speak to the infant in the IAL. The child would learn to speak the language effortlessly if it is well-designed and completely regular. The AI friend could also teach the child to read the language, as well as some arithmetic, colors, animals, measurements, etc.
This would greatly reduce the enormous costs of teaching children an IAL in school, while doing a better job of it. One of the problems with Esperanto back in the day was that even if we had succeeded in teaching every kid in the world Esperanto, it would have been spoken a bit differently in each country, and there would have been further drift from there. But every child learning an IAL at a young age from an AI friend has the same teacher, eliminating regional differences and drift.
Such a dramatic reduction in cost changes the calculus. Another change to the calculus is provided by climate change and the low birth rates in many countries, both of which will stimulate massive immigration. Immigration is vastly more difficult if the immigrants do not speak the host country's language; the costs are enormous. For one thing, immigrants often are less economically productive because of the language barrier, so they are poorer, and they pay less taxes to the host country. (As an example, I know a surgical nurse who is immigrating to the US but will likely clean toilets with her mother because she is having a hard time learning English.) When we are talking about tens or hundreds of millions of immigrants, the costs of such reduced economic productivity are enormous.
Moreover, immigrants who don't speak the host language tend to live in enclaves rather than assimilating into the larger society. There are many such Spanish-speaking enclaves in the US. This too has consequences.
I'd love to chat more about this; I have written privately about the costs of teaching the world's children hundreds of different natural languages and then dealing with the difficulties that causes, but am looking to engage with others with the same desire to see the world learn a common language.
Background
The IAL scenario
Suppose that tomorrow, every single person on Earth wakes up and somehow knows at least 2 languages: a mother tongue and a particular "Z" - an IAL. Z can be any existing or imaginary language. How different is that future world, compared to what we had yesterday?
My imagination
(The IAL can save many trees in another interesting way IMO. There are at least 407 tongues spoken by one million or more people. A simple calculation reveals that we’ll need 165.242 different direct cross-language dictionaries between each pair of them. If we take into account other, not-as-popular-but-still-are-spoken-by-many-people tongues, the number explodes into an astronomical value. All of those dictionaries consume an enormous amount of trees, definitely more than the number that the search engine Ecosia planted. Meanwhile, consider the case when IAL is around. We only need 814 types of dictionaries to cover everything)
Related case in point: LW is all-English. That, of course, doesn’t mean non-English speakers are all irrational. Now, there are roughly 1 billion people on Earth who speak English, which leaves 6.8 billion others. Just think of a world where LW has 7 times more brain power, and about 4 times more resources! What will we able to achieve with such potential?
My attempt of assessment
What I'm looking for
The whole rationale for me to pursue this dream is based on the notion that it's worth the effort. It can be somewhat described by the 'formula':
Gain = (Benefits − Costs) ∗ Probability
I reason with myself that, since the Benefits to the world are almost uncountable while the Costs are finite, even if the Probability of success is small, the Gain for humanity is significant. Something along this line: (9e15 - 7e13) * 1% = 89.3e12 ($)
Of course, all of these figures are a totally arbitrary guess with no base at all. To improve my accuracy, I ask all of you guys to give your own numbers, so that I can recalculate. But aside from that, you can help me in many more ways, such as:
What's next
I do have a plan for how to achieve this dream. However, depending on the updated assessment after this post, the IAL may or may not prove worthy. If it's the first case, I'll write more about the failure of Esperanto and how I think we can do better.
And hey, a big thank you for reading, especially if you decide to chime in! I have a small request: since my native tongue is not English, when you comment, please write one level down.