The internet, though, has been a mixed blessing for Esperanto. While providing a place for Esperantists to convene without the hassle of traveling to conventions or local club meetings, some Esperantists believe those meatspace meet ups were what held the community together. The Esperanto Society of New York has 214 members on Facebook, but only eight of them showed up for the meeting. The shift to the web, meanwhile, has been haphazard, consisting mostly of message boards, listservs, and scattered blogs. A website called Lernu! – Esperanto for the imperative “learn!” – is the center of the Esperanto internet, with online classes and an active forum. But it’s stuck with a Web 1.0 aesthetic, and the forum is prone to trolls, a byproduct of Esperanto’s culture of openness to almost any conversation as long as it’s conducted in – or even tangentially related to – Esperanto.
But there’s hope that the internet can give the language new life. Wikipedia and its 215,000 pages was a first step, and yesterday, Esperanto debuted on Duolingo, a virtual learning app with 20 million active users – far more people than have ever spoken Esperanto since its invention.
This article is the perfect mix between two of my favourite subjects – technology, and language. A highly recommended read.
The link to the article is wrong, has an extra double quote at the end. The real link is this:
http://www.theverge.com/2015/5/29/8672371/learn-esperanto-language-…
Read the article, then inevitably went to DuoLingo and took the first 10 XP of the Esperanto class. Yep, it looks pretty darned easy to learn. I’ll probably stick with it as long as I stuck with the lojban class, though… 🙁
Lojban… *sigh*
I don’t know the first thing about Esperanto, and to be honest my first impression was that this language hasn’t got a chance. However a regular, predictable, concise language not only makes the language easy for humans, but it makes it easy for computers to handle too.
“Evolved” languages are often irregular and it takes years to learn all the nuances without sounding like a noob. Even someone who knows a language well can have trouble converting this knowledge into software. Using something simple like esperanto to convey concise/unambiguous meaning is a brilliant way to communicate clearly with computers as a base language.
One concern if this language were to actually gain widespread adoption would be people trying to break the rules on, purpose.
http://www.osnews.com/thread?611446
From your 2nd paragraph, the first sentence (which I agree with) is the reason why the 3rd sentence is unrealistic. Esperanto would have to be a superset of all existing languages. For one example, it would have to be able to represent all the different registers of politesses that exist in Japanese and Arabic, and it would have to be able to represent all these registers in all the ways that those languages represent those registers (choice of verb, conjugation of verb, choice of subject, particles…).
Or a simple, concrete example: Most European languages have two ways to say “you” (2nd person singular), (English is an exception) and the distinction is usually described as something like more-formal and less-formal. Supermarket signs in Brussels are usually bilingual with the French text using French’s more formal form of “you” while the Dutch text uses Dutch’s less-formal form. Both languages agree that the more-formal form is used for addressing professors, and the less-formal form for addressing children, but there are situations where they diverge on whether situations require more-formal or less-formal. To make a meta sentence which would correctly translate, in this situation, to the more-formal form in French and to the less-formal from in Dutch, you would need a meta language that has a “you” that represents the concept of “this is the word ‘you’ in the form appropriate for the context of a sign in a supermarket addressing customers”.
That’s a very advanced challenge for machine translation, and we’ll surely get their some day, but it would be massively more complex (maybe impossible) to create a language this rich for translation software which could also be used by humans.
Esperanto, like all human languages, is too limited. It could only contribute to the production of very approximate translations (and we can already do that via English (or via simple English), which has advantages such as massive corpuses of texts which have been translated by humans from practically all existing languages into English, which can be used for training the translation software.)
You is the polite form. The English were just SO formal they ran out of contexts where the non-formal form was used (except biblical commands, and shakespearean love).
Ironically this over formality is now influencing other languages to drop their formal you, and go informal all the way, since that is the de-facto result of the overformal English when you lose the distinction.
Languages are funny.
The important distinction is that where English (modern English) uses “you”, French uses “vous” and “tu”, Dutch uses “u” and “jij”, Spanish uses “usted” and “tú”…
(not to mention the regional variants of these languages.)
You is actually the collective “you”, second person plural. Most European languages use this model. The singular is for close friends, relatives and children, the plural is used as the polite form.
Some don’t though… Polish uses “Pan” (male, lit. “Mr/Sir”) and “Pani” (female, lit “Mrs/Madam”) and declines the verb in the third person: Czy pan ma paszport? which translates as “Do you have a passport?” but more literally means, “does sir have a passport?” Informally, it would be “Czy masz paszport?” I think… though I’m still learning, so that might not be right. The verbal aspects are mind bending in Polish… 😉
My grandparents spoke Polish and it’s tough, I spent a few weeks in Poland and reverted to toasting and grunting, Voda, nastrovia! good memories.
sorry i have nothing to add, just to encourage you to continue with polish, it can be really beautiful. have you read/seen Pan Tadeusz? amazing.
You was the polite form — since now there is only one second person pronoun, it gets a default politeness level. And while the default could still be the polite register, I do not think the usage patterns justify this. Let’s call it “neutral”.
ciaran,
Of course a compelling reason to use English as a base would be it’s popularity, but otherwise I’m not convinced that English is a particularly good choice. All things being equal the consistent regular formation of Esperanto would be a strong advantage for the purpose of AI.
I agree with you that other languages have different words to convey “registers” (ie. politeness). Another example is the Eskimos having dozens of words for “snow” (*1). But from a technical programming point of view I don’t think it’s a good idea to have a vocabulary of words with overloaded meanings. To the extent that these are important, I think that such registers would be better handled with metadata on top of the base grammar.
This metadata would help prevent loss of context and even help convey the context intelligently across languages. For languages that don’t convey registers via overloaded word combinations, you could just drop the metadata, or perhaps you (polite) could display it in parenthesis. This approach might not be perfect, but I still think there may be merit in using a special language for computers to be able to work with it easily. It could really reduce the barriers to AI.
*1 – http://www.washingtonpost.com/national/health-science/there-really-…
Edited 2015-05-30 14:21 UTC
That could work for certain applications, such as translating instruction manuals or use as a common language in a certain organisation, but it would exclude Esperanto from other applications such as literature. Literature wouldn’t be much fun to read if social aspects were filtered out in translation or if sentences were peppered with constructs such as “(polite)”.
ciaran,
That was just an example of how the metadata could show what got lost in a translation. However you are correct it would not be appropriate for literature. Sometimes translators are quite liberal in rewriting things to flow better in the target language, I see this most often is closed captioned / dubbed movies. This would definitely pose a formidable programming challenge (in part because the translators need to factor in on screen lip syncing even at the cost of accuracy). Still, I think more metadata could help here too. This could contain syllables from the source language, and the computer could attempt many permutations in the target language to better match the flow. This would be an advanced feature though, probably not till V3. V2 might output a number of possible permutations but force the operator to select the best. V1 would be an abysmal failure
Edited 2015-05-31 06:13 UTC
Yeh, I think the snow problem is one of the least harmful, there are ways for the snow translations to get hairy alright.
AFAIK, Google’s current approach to avoid having to make a language mapping for each pair is that they make mappings for the main pairs and do the other languages via English.
Machine translation is one of those problems that if you think about it too much, you just don’t know where to start anymore.
Google constantly falls down. It really only works for a few languages. If you pump in some words, it will give extremely bad results… for example, I wanted to know the difference between móc, umieć, potrafić in Polish, so I threw them in to my Android phone, only to be told they all mean “able”… Facepalm. Yes, they do… but, from what I now understand, the nuance of usage means that they are not simply interchangeable.
umiem kupić. “I can buy” according to Google, except it actually means more like “I know how to buy” – “able as in “I have the knowledge to do something”.
mogÄ™ kupić. “I can buy” according to Google, and that is pretty much correct from what I now understand.
And potrafić is the perfect aspect of the imperfect móc.
Google translate does indeed a poor job translating single words. But it has good chance picking the right translation if you build a simple sentence with it.
And I never use Lang1 <-> Lang2 where both != English. Google T only really works translating from / to English.
henderson101,
I’ve also found that it has trouble with some translations. Once my wife plugged in a birth announcement written in french and the translation was simply hilarious.
Here I took the article, translated it to french and back:
Ironically google fails twice to translate the word “Esperanto”, somehow substituting “Dutch” and “English”. But other than this glitch, the significant overall improvement would suggest to me that Esperanto maps better to english or maybe even google’s internal translation mechanism.
Edited 2015-06-01 16:11 UTC
Yes! That used to be a game we played at a previous job during down time. If you translate through 3 or more languages and back to the source language, the results are somewhere between gibberish and mind blowing.. 🙂
Unfortunately, Japanese and Korean were the main culprits, and they would seriously change the meaning in amazing ways at the time. They still do to a certain extent, but Google has got a lot better over time.
Hmmm. The quality is surprisingly good. I tried Google translate with a random Spanish news article and the quality is worse:
https://translate.google.com/translate?sl=es&tl=en&js=y&prev=_t&hl=f…
(And Spanish->English should be among the most advanced language pairs.)
Here’s the Esperanto Wikipedia’s article on GNU translated into English:
https://translate.google.com/translate?sl=eo&tl=en&js=y&prev=_t&hl=f…
The quality’s actually not far from the Spanish news article.
Translating English text to another language and back mightn’t be a fair test since the intermediary text is in Google Translate’s own version of that language.
Edited 2015-06-01 22:51 UTC
Try any non Western European language. In fact, try Finnish, as Scandinavia is often lumped in with western Europe.
The English word ‘you’ no longer differentiates between singular and plural, objective and subjective or singular and plural.
http://www.oxforddictionaries.com/definition/english/thou
Correct.
English has one word for the 2nd-person-singular subject. Most European languages have two.
(And, the English word “you” also has meanings other than 2nd-person-singular, but I’m not talking about those meanings.)
The problem with Esperanto from a Germanic language view is that it relies too much on particularly romance languages and almost not at all on Germanic languages. Personally I’d prefer Folksprak, being much easier from a North Germanic perspective.
Esperanto will never succeed as long as the learning curve is so steep as it is in reality. The steps should be making a common Germanic language, a common Romance language, a common Slavic language and so on.
When people have switched to these common tongues, a modern Indo-European language can be made, and when people have switched to that, a new tongue created from 6-8 main languages can be introduced. It has a greater chance of success, and makes the learning curve less steep.
Interlingua is similar to what you describe:
https://en.wikipedia.org/wiki/Interlingua
It’s mostly based on Romance languages (and IMO, Romance languages are clearly simpler than Germanic languages, and probably clearly simpler than Slavic languages if what I’ve heard is correct).
To check the readability, here’s an article from the Wikipedia in Interlingua:
https://ia.wikipedia.org/wiki/Projecto_GNU
I think it’s the best of the constructed languages, but the advantages aren’t strong enough to motivate me to learn it instead of Spanish. Spanish has so much music, literature, travel destinations, people to talk to…
They aren’t hard, they are just different.
Difficult question. What method to use…
IMO, if we compare a language with complex grammar from each family, such as French and German, the Germanic language is the more difficult. And if we compare an easy language from each family, such as Spanish and Dutch, the Germanic language is again the more difficult.
I’ve studied law at universities through three European languaes and I speak an Asian language at home. As a hobby I added Spanish and German.
I don’t dispute your experience, but mine has been completely the opposite. Romance languages use a far less parallel grammar to English, what with reversing the noun and adjective, formation of the possessive and highly inflected verbal system. Take Swedish, which I’m somewhat familiar with.. other than the plural and definite article, most of the rest of the grammar only diverges where English has innovated (the “do” auxiliary verb, for example, and the present progressive tense, i.e. “I am running”.) English has a lot of Latin, Greek and French loan words, but quite often their meaning has diverged and it isn’t always possible to use the word from the English usage to understand the Latinate. I was in Italy recently, and Italian was very alien to me.
Which is closer to English?
I have been to Italy. It was very warm there.
Jag har varit i Italien. Det var mycket varmt dar.
Sono stato in Italia. Era molto caldo li.
Caveat: I trusted Google Translate, but I purposely chose simple sentences. The Swedish is almost a word for word translation of the English. Var = was/were, and in varit, the -it is a bit like -ed in English past tense. Mycket is the neuter of mycker, which is cognate with “much”. So the Swedish is a bit like “I have be’d in Italy. It were much warm there”, which, okay is poor English grammar, but we all know that isn’t how language translation works anyway…
Edited 2015-06-01 09:52 UTC
I’m not a fan of that sort of test. It’s still subjective, plus it reminds me of programming language fans bragging that Hello World can be done in one line in their language (I don’t care, I want to know how a programming language handles big complex tasks).
So I did a few web searches and found:
http://www.languagesurfer.com/2012/08/21/how-long-does-it-take-to-b…
It says the US government gives Spanish, French, Italian, and Portuguese a difficulty of one, and German a difficulty of two. (Japanese and Chinese are four.)
The “Foreign Service Institute” has a similar ranking, but they have some Germanic languages in category one:
http://www.effectivelanguagelearning.com/language-guide/language-di…
And another:
http://www.languagetesting.com/how-long-does-it-take
One trend is that Romance languages are always in category one. Another is that German is always in category two. The rankings that have other Germanic languages, put them all in category one.
It’s not really fine grained enough, nor complete enough, to settle our disagreement. One possible interpretation is that Germanic and Romance languages are roughly equivalent in difficulty, except German which is a total pain in the arse.
(Two minor notes: that French sentence in your test is slightly wrong because “nom” is family name. Also, “My name is X” is a little strange in French, people usually say “I’m called X”. The “I’m called X” idiom is common in Spanish too, but maybe “My name is X” is also normal.)
Icelandic has the most complex grammar of a living Germanic language. German is by no means as complex, but it has a grammar closer to the complexity of Icelandic.
I believe Romanian has the most conservative grammar for a Latin daughter, but all still tend to inflict verbs for the various persons in speech.
When you look at most other Germanic languages, German is very much harder because it is more conservative. This is not universal though. Modern Swedish has:
Ha – to have
Jag har – I have
Du har – you have
Han/hon har – he/she has
Vi har – we have
Ni har – you have (plural/polite)
Dom har – they have.
True. As a Spanish speaker Esperanto is like Chinese to me but I completely understood that Interlingua page.
Actually not exactly true. While the root words are mostly romance, I’d actually say the grammar is far closer to the germanic languages. It relies on pronouns rather than altered verbs to show person and number where most romance languages change the verbs themselves (French being the obvious exception as it does both), and basic word endings (far simpler than most germanic languages actually) to denote parts of speech (a for adjectives, o for nouns, add and n for direct object, j for plural, etc). It only relies on verb conjugation to show tense making it similar, in that respect, to modern Swedish though it’s far more regular in the basic conjugations used than Swedish is. Save some of the accented consonents and frequent use of “j” sounds (equivalent to “y” for English-only speakers), I don’t actually see much Slavic influence in the language at all.
I took an online course on it once. I’ve never been taught formal grammar even for English, but Esperanto was simple enough (and regular). I didn’t continue on with it due to time and lack of practice, but it wasn’t so difficult.
The point of Esperanto is that it isn’t designed to be a replacement language. It’s designed to be a supplemental language for international communication. So to the concerns about it not covering the nuances that appears in natural languages like Japanese and Arabic is missing the point.
And Japanese and Arabic people don’t find it an imposition to learn languages that don’t match their language any more than an English speaker would to learn other languages.
Learning another language is difficult, but Esperanto is pretty easy, considering.
Great, but cultural aversion isn’t the obstacle that’s keeping Esperanto down. The problem is that Japanese and Arabic people, and all other peoples, don’t see learning Esperanto as a good use of their time.
An Esperanto optimist could say that this is a mere chicken-and-egg problem, and when some event finally causes a critical mass of people to learn Esperanto, others will then have the motivation to learn it and it will snowball and spread around the world.
The Esperanto pessimist will point out that if this event hasn’t happened between 1887 and now, it must be a very unlikely event (and maybe the practical advantages are being wildly overestimated).
I don’t think I’ve ever tried to argue that Esperanto will or should be successful. My whole comment was about my limited experience with it and how there’s a bit of misunderstanding from people about what it was supposed to be.
Your analysis of the situation is appreciated but quite unnecessary, like telling people that FIFA is corrupt.
You claimed that Japanese and Arabic people wouldn’t be opposed to learning it. I don’t know what this is based on, but since the population of Esperanto speakers in those groups is tiny, I pointed out that you might be overestimating Esperanto’s chances.
No I didn’t. I claimed that Japanese people who learn other languages had no problems with it, so they wouldn’t have any more problem with Esperanto than they had with other languages. They’d probably have an easier time because there’s no cultural and semantic nuances to deal with.
Edited 2015-05-31 02:55 UTC
Not at all. We already have translation technology. I’m not rejecting progress, I’m saying I’m not convinced adding Esperanto will improve things. Companies are investing good money to get smart people working on this, and if they have no interest in Esperanto it’s probably not because they all never gave it a try. It’s more likely that they know the problem well enough to see that Esperanto doesn’t help.
Which companies are these? Because the only company that seems to have been successful at just parsing deep meaning from just one language is IBM with Watson. I don’t think any other company has thrown more money at the problem than has gone into Watson.
And sometimes, even when two people are speaking the same language, they can completely fail to communicate.
Esperanto was designed to be a politically neutral (Latin derived) language for people stuck between Russia and Germany. It is no longer relevant and should be laid to rest.
I don’t think it was ever relevant. I tried learning Esperanto as a teen thanks to Red Dwarf, but of course it never came in handy and we didn’t have the same type of multimedia on the Web that is available today. When I started learning Spanish (you never stop learning a language), I found that I would confuse the Esperanto word with the Spanish word; the Esperanto got in the way.
A few years ago, I visited a live demonstration of Esperanto from speakers of different languages with my wife, a linguist. It was TERRIBLE. While the speakers had taken the time to learn the vocabulary and grammar, they each pronounced words as if they were speaking their native language. The wife said something about the speakers retaining native morphemes. It would be as though an English->Spanish speaker never got the hang of the “ll” or “ñ” sound, except everybody was doing it as there was no reference. I left wondering if the Esperanto speakers would have really understood one another without context (as so often happens when conversing with someone in another language in a foreign country).
Some Esperanto websites blame the lack of success on the US State Department spending $$$ to keep American English as the language of international business. I think this creeps into conspiracy theory; Esperanto doesn’t seem like a viable alternative so far, in its hundred thirty years of existence.
Yeah. While much fault can be laid at the feet of our government, the rise of English isn’t one of them. English and French were already on the rise thanks to the colonial periods, and the international language of business probably would’ve come down to one of the two anyway.
That would be phones. She might’ve said “phonemes”, but she’d be wrong, as they are the underlying, theoretical, unique elements the language’s sounds (phones) are derived from.
Esperanto has many problems (see e.g. http://miresperanto.com/konkurentoj/not_my_favourite.htm), originating from the fact it’s creator wasn’t a linguist. Also, it is not completely regular (see e.g. here: http://www.zompist.com/kitespo.html).