Show HN: Eskéndereyya – Learn How to Read and Write Arabic in Latin Script | Hacker News Reader

Show HN: Eskéndereyya – Learn How to Read and Write Arabic in Latin Script

github.com

72 points_samihasan_68 comments

I'll play devils advocate here. Their solution is too complex and tries to fix a non-problem.

For someone learning Arabic the letters are the first of many challenging steps. I learned the Arabic letters in 1 day with flashcards. That single day was easy compared with everything that came afterwards. Using a romanized scripts makes all these subsequent steps with grammar and accents much harder.

I can see how this would be helpful in accurately learning new accents, but accents are notoriously inaccurate and flexible, the Arabic letters leave that flexibility, outside of academic text Romanized letters would require so many exceptions that it be like nailing jello to a wall.

It would be useful for news organizations if this system was applied across many languages so that we can always spell and pronounce names correctly.

The use of non-english keyboard letters creates complexity. I much prefer the use of number replacements for letters that teens use when texting. sa7?

Author here.

Thanks for the feedback.

I might disagree with your characterization that the problem I'm addressing here as a non-problem as writing systems are there for a reason and so is mine.

Eskéndereyya doesn't tackle the problem of Arabic letters memorization as this is a very simple problem to crack. It instead addresses pronunciation and reading and to a lesser degree writing as you may know that the common way to write and serve text in Arabic is without diacritics, and for this reason, it becomes hard for beginners to practice and improve their skills without aid and this is where I envisioned Eskéndereyya to fill the gap.

I'm not sure what you're trying to convey with the part of accents in your comment. Are you referring to Arabic regional dialects/languages?

Re complexity, doing nothing will always be less costly that doing something in terms of energy and effort but this point of view overlooks the gains or return expected on the energy expended, and if the return turns out to greater than the costs, the endeavor is determined to be profitable and vice versa.

So, you may think that this layer of complexity is unnecessary but I will take your word "sa7" as an example to prove how I think otherwise.

"sa7" is "صح" in colloquial Arabic but it's still ambiguous and confusing as the letter "s" in Arabizi can mean both "س" or "ص" depending on context but when the word is transliterated as "šaħ", the ambiguity disappears with no room to error.

You might counter and say that there's no "سح" in Arabic but there actually is but it's less common since it's slang but featured in a well-known old Egyptian folk song titled "essaħ eddaħ embú السح الدح امبو"

To summarize, there's always a trade-off and it's up to you to decide which you to go.

> I'll play devils advocate here. Their solution is too complex and tries to fix a non-problem.

I'd say you are on the wrong side. All Arabs continue to write in Arabizi. I have lived in this region for years, as a non-native speaker. When I first arrived as a student for my first rodeo in 2006, I once wrote SMS in Arabic letters. That Egyptian kid later asked why I would do that. He was terrified it was a plant or a trap or something, that is how rare it was then.

However, I will give you a point. None that we are moving beyond basic SMS encodings and the limit of standard cell phones and on to Unicode and emoji-rich smart phones, WhatsApp and others have enabled a renaissance in Arabic writing amongst some.

But many continue to write to me in English characters. And re your point learning in one day: the reading is fun, the writing on a keyboard for a society where Arabic literacy (not English, mind you) is largely de-emphasized and many cannot read/write formal Arabic, the desire for Arabizi hides deficienies people would not rather admit.

>Using a romanized scripts makes all these subsequent steps with grammar and accents much harder.

This is the exact problem with learning Japanese using roumaji (rouma ji = roman character; ローマ字), because you of course start to read the latin script with your starting accent (assuming of course your first language is written using the latin alphabet).

However some textbooks persist with using it, despite being excellent otherwise (Japanese: The Spoken Language being one). I can't stress enough the need to move off reading Latin characters. It is also useful to learn the characters by sound rather than their roman equivalents. i.e instead of learning that ロ is "ro", you could learn it by listening: https://youtu.be/aLEtZ2CRoho?t=1m53s

The mental association is everything.

Well, I must say that I fail to see the rationale of this argument as I speak other European languages and I haven't encountered this problem before.

The mental process goes like this; I identify the language of text let's say Spanish and then like a switch in my brain is turned on for the Spanish pronunciation and then I proceed to read the text using the rules of the Spanish language while English is totally disabled.

This is not like unique to me as I observed other students with the same process. I can't really say that the issue you described is a universal issue for all language learners worldwide.

I think the case is different for languages that use the same script, such as a European language learner learning another European language. The learner must learn to switch "modes" and this is reinforced all the time by any kind of reading, because the learner has no choice but to switch modes.

However this mode switching is not reinforced in the case where the script (in this case the Japanese script) is not latin-based.

I think I'm trying to say that native speakers of latin-based languages who are learning latin-based languages have to learn to change their mode of pronunciation, whereas if they were to learn a non-latin-based language they don't (and shouldn't) learn this at all, because in the long run it isn't useful.

Except romaji isn't used to teach pronunciation, it's meant to help the student read while they are still getting familiar with kana; much like furigana for kanji.

Japanese characters have one pronunciation in all circumstances (excluding diacritical modifications). Respectively in romaji, consonants and vowels will always have the same pronunciation. Compare that to American English where accents have been removed and even native speakers can have trouble pronouncing new words.

>Except romaji isn't used to teach pronunciation

I know, but students will continue to treat it as a pronunciation guide anyway, unless told specifically how to pronounce otherwise. It's especially bad when the Japanese is supported by romaji way into the course, which fosters laziness in the student.

unvoiced vowels at the end of sentences sometimes get dropped in Japanese 'です' (more 'dess' than 'desu'). So it's not entirely consistent.

I studied Russian for one semester when I was in middle school. I didn't learn very much and can remember even less: only a few simple words like яблоко (yabloko -- "apple").

But I can still sound out words in the Cyrillic alphabet without any difficulty. Learning another alphabetic script really doesn't take that much time or effort.

Same here, but in college. Got the Cyrillic alphabet in a couple weeks but it was by far the easiest part of a very difficult semester.

I switched to French.

I've learnt azbuka purely from our national TV news (in central Europe), during the course of a year, and I was not trying at all - one day I realized I could read the sentence they showed (even though I couldn't understand it).

Cyrillic is easy until you see it written in cursive :)

>For someone learning Arabic the letters are the first of many challenging steps.

Exactly, for most other languages as well. I only needed flashcards with Hebrew (as an Arabic speaker, I only needed to remember shapes because sounds are similar). To learn Cyrillic was more figuring out what changes and connecting it to Greek alphabet (р is rho, г is gamma, д is delta, п is pi, etc).

Transliteration for most languages is just another layer of complexity in my opinion given the tiny amount of time I needed to acquire alphabet (the letters, not Google's parent company).

Here's something I started. Need to get back to that:

https://github.com/jhadjar/Language-Acquisition

I've been learning Russian for two years, and I can only agree with you. Learning to read and write Cyrillic script was the easiest part of this adventure.

And I think learning all sorts of nuances of the language would be more complicated using Latin script. Latin script also hides the fact that, even though the letters are similar, they really are different sounds. I feel like it would only hinder pronunciation.

It really isn't fair to compare Cyrillic script with Arabic. Cyrillic is basically the same script as Latin, even characters are alike. Hell, even something like Japanese hiragana is pretty much the same stuff as Latin alphabet, even though it looks nothing like that and is built upon syllables, not letters. And hence it is a bit harder to learn, but not by much.

However Arabic script is a whole different thing. It has a whole lot of diacritics, characters change their shape depending on the position within a word, characters are notably less distinctive for an unaccustomed European than Latin letters, Cyrillic letters, hiragana or runes.

The comparison of Arabic with Russian is not sound as Arabic is written with diacritics which in and of itself is a big plus if you're a native speaker or an experienced one but not so for inexperienced or novice speakers while Russian to my understanding is more explicit about that part.

Do I understand you right, that Arabic uses diacritics when written in Arabic script? Or did you mean Arabic when written in Latin script?

Using diacritics is usually a sign, that the language uses a script that does not map very well to the target language, so the latter would be understandable.

Slavic languages would the same in this regards. Those, that are written in Latin script, have to use diacritics. Those that use Cyrillic, do not need to.

All Cyrillic alphabets I am aware of do use diacritics. Й, Ў, Ё, Ґ..

While you are of course right, that there exist such characters, Belorussian and Ukrainian are not all Cyrillic alphabets.

But then, I'm also not an expert on all Slavic languages, I just remember when I was taught Russian and azbuka years ago, there was none and that caught my attention.

> Belorussian and Ukrainian are not all Cyrillic alphabets.

LOL Please, tell us an «all Cyrillic» alphabet. Old Church Slavonic by Saints Cyril and Methodius maybe?

Why Belorussian and Ukrainian are «not all Cyrillic alphabets»? (I'm Ukrainian).

There's the difference between 'some' and 'all'. Ukrainian and Belorussian are 'some', not 'all'.

Language may have some local specifics - Serbians have Ћ, which all other Cyrillic-using languages do not have. You have І instead of И, etc.

Й and ё are in Russian alphabet. In schools it is never highlighted that those are diacritics, so to many it doesn't register.

Are those diacritics? I tend to think of them as separate letters just as å and ä are separate letters in Swedish. Of course, ё is usually written as е so I may well be wrong about that.

This is different from è, é, ë, and ê in French where these are all e, but with different diacritics.

That's my impression of the statement as well. Even if technically they can be called diacritics it doesn't make much sense, not within this discussion. "Й" sounds absolutely unlike "и" and has pretty much no connection to it. "Ё" is a bit more tricky, but nevertheless — there's no another letter in Russian alphabet we could put umlaut on and get the result similar to putting umlaut on "е".

I guess "ь" has more right to be called diacritic, because it genuinely has 1 purpose: altering the other sounds. But even that is treated as a separate letter.

It's nothing like diacritics in French, German or even Latvian, let alone Arabic.

Й is absolutely linked with и, and it arose as alphabetic distinction from it. It appeared and disappeared for certain periods; in Russian it reestablished as a letter only in 20th century. It is still commonly collated with И in indexes and dictionaries.

> "Ё" is a bit more tricky, but nevertheless — there's no another letter in Russian alphabet we could put umlaut on and get the result similar to putting umlaut on "е".

This is a bizarre distinction: diacritics in natural languages aren't supposed to be freely recombinable.

Those letters use no diacritics. They are just letters.

These are the letters with diacritic marks.

No. іе -> ё, іі -> ї, иі -> й. оу -> ў. They are just shortened, to write less.

Well, ў for once is not a shorthand for оу (e.g. слаўны, ўзяў). But it's beside the point, all diacritics are used to "write less".

> But it's beside the point, all diacritics are used to "write less".

That may be the reason for misunderstanding. In my language (Slovak), diacritic is not used to "write less". It is used to express different sounds, i.e. s = с, š = ш, c = ц, č = ч.

Sure, but just as you demonstrate you have fewer distinct characters when you use diacritic than when you don't. And arguing whether Е and Ё are same character with diacritics, or just different letters that happen to have diacritics by the virtue of canonizing in alphabet is really splitting hairs. They serve same purpose, in the same circumstances, with the same visual arrangement. Quacks like a duck etc.

Agreed. I have some proficiency in MSA and Levantine Arabic after studying for 4 years in college and the alphabet is the least challenging part.

I was going to ask "why?" before I read in the overview the reason being helping non Arabic speaking learners of the language. It still defeats the purpose of learning a language to be able to read literature written in original script and communicate in writing with the language readers. The only case where I see this useful is when verbally communicating using words memorized using this Latin script as tourists do when visiting a foreign land.

As an Arab I also hope that the motivation behind this is not ideological. Like how the Turkish language was transformed to use a Latin-based script system.

Edit: As of the claim of it being useful to Arabs, please no! Arabs when use "Arabish" tend to use it badly with no vowels making pronouncing words hard and sometimes ambiguous. The "Arabish" trend was popular in the 90's (in the infancy of Internet or at least Internet penetration) and has declined dramatically in the 2000's. Arabs lazy enough to not write in Arabic script would also be lazy enough to not learn the proper usage of this Latin script.

Author here.

Thanks for the feedback.

Regarding the concern of this substituting the Arabic script, I addressed this in the intro section of the project where I laid it out plain and clear that the target is novice learners of Arabic as I noticed that they suffer with pronunciation and reading skills and I wanted to help them overcoming this gap to make the transition to the Arabic script smoother and less painful.

As far as I can tell, the project was not ideologically or politically motivated but to be honest with you, the Turkish language was an influencing factor on the development of this project whether which parts to use and which parts to pass but I am not sure if you know that Turkish throughout its history was not written only in Arabic script but in other scripts like for example Armenian but the Ottomans adopted the Arabic script for their official communications until came Mustafa Kemal and introduced his reforms and switched to Latin which was a good idea since writing Turkish in Arabic script was a nightmare.

Re your last concern, like I said in that part, it's for regional languages/dialects not the standard variety and I outlined the reasons for that.

Hi. I applaud your motivation, but to be honest, when I was learning Arabic, the only trouble I would regularly run into would be with the tashkeel, which I find difficult to memorise. On the other hand, it was much easier for me to associate each letter with their respective sound(s).

This is a good idea, and one that I could certainly have used when learning Arabic— but its real utility is not for Modern Standard Arabic- it's for 'amiya/3amiya (spoken Arabic), which, to those not in the know, differs from region to region and is vastly different from both written and spoken forms of Modern Standard Arabic.

As someone suggested above, learning Arabic script is the first-- and smallest--of many challenges for those trying to learn formal Arabic.

There are few good systematized sources available for people trying to learn spoken Arabic, particularly if they're not particularly interested in reading the news or classical texts. I'm thinking of aid workers, diplomats, vagabonds, whatever.

A system such as this has great utility to these people— except it already exists in much of the Arabic-speaking world, particularly the Levant and Egypt, where numbers are used to represent sounds not found in the Latin alphabet.

For example: "You will speak Arabic soon" (Levantine) - "إنت رح تحكي عربي قريبا" - can also be rendered as "inta ra7 ti7ki 3rabi 2areeban". This makeshift system is used widely in texts in Lebanon and elsewhere. Utilizing this existing method will be easier and have wider applicability.

If you are Turkish and opted-in for religion classes in the school, you learn some parts of the Qoran reading them in Latin alphabet. Because we don't know arabic alphabet, but Turkish one (Latin alphabet with 3-4 extra letters). Actually one of the letters in "extended Eskéndereyya", "ş" is a Turkish letter, created for Turkish language. I don't understand why Eskéndereyya is different than the alphabet we used to use in high school.

This is not a good idea. From a language learning perspective, it is essential to become immersed in the sounds and shapes of a language as soon as possible. Approximations will not do. Arabic is equipped with sounds that foreign speakers are not able to produce without significant effort and coaching.

This will make it more difficult to become proficient.

Why not https://en.wikipedia.org/wiki/DIN_31635 or https://en.wikipedia.org/wiki/ISO_233?

Super slick and exicitng. Since this is my wheel house I will check it out later and report back something more constructive.

Thanks for the work, sami. I love to see people bringing Arabic to the open source world. It needs love.

People might also like Qalb (heart or template, not sure which romanized as tired, validating comments later in this thread), which was an awesome idea to me.

https://en.wikipedia.org/wiki/Qalb_(programming_language)

I like this idea. But I am not sure about its ease of use. There is a much less precise but significantly speedier way to convey Arabic words using the Roman alphabet on a QWERTY keyboard using the Arabic chat alphabet. [1]. The more difficult sounds in Arabic are conveyed with numbers. For example ح is 7 - and ع is 3. In both cases the Arabic letters are visually similar to the character that represents them.

[1] https://en.wikipedia.org/wiki/Arabic_chat_alphabet

The problem with this system, besides the appearance and inherent ambiguity to name a few, is that it doesn't cover essential parts of the language like "stress" and "tonality" where other writing systems do.

This should certainly be adopted by official/serious projects (such as albums, movies, events, etc.) with Arabic names written in Latin script. The use of the "chat" form (e.g. "7" for "ح" and "3" for "ع") is just horrid and makes any serious work look childish.

Trying to explain glottal stops with English examples is pointless - "Apple" doesn't begin with a glottal stop in English; "fear" and "ear" differ only in the absence of /f/.

German does introduce a epenthetic glottal stop before word-initial vowels, but not English.

Also, attempting to explain /ɪ/ as "Fin" -> "fen" is also unhelpful.

I've seen it described as the sound or throat action in the middle of "uh oh!".

I found that kinda useful.

At least in my accent, the 'tt' in written is a glottal stop.

There's a glottal stop between book and case in bookcase, though.

For anyone interested in this space one of my buddies made a site / apps for helping write Arabic transliterations: http://www.arabictypist.com/

Not sure if he's active on HN so posting it for him and forwarding Eskéndereyya to him. Great work OP!

Has anyone ever tried writing English in Arabic script? From working with complex text layout, I know Arabic letters have 4 forms (isolated, initial, medial, final). It would be kind of cool if I could spot check the shaping engine is working properly by testing with English in Arabic script...

Come to the middle east. Plenty of English words written in Arabic. McDonald's and subway immediately come to mind.

Always got a chuckle when Canadian tourists would take pictures of Tim Hortons written in Arabic: http://cdn.torontolife.com/wp-content/uploads/2012/11/tim-ho...

I prefer to write Tim Hortons like this "تِم هُرتُنز" in Arabic but that's just me.

Aside from that, Spanish, Catalan, and Occitan have historically been written using both the Arabic and Hebrew alphabets, by their respective user communities. Same for Yiddish, linguistically a Germanic language, which was often written using Hebrew script. Turkish seems to flip from Arabic to Latin and vice-versa every few decades.

English written in Arabic script looks surprisingly natural to me.

Is there a simple algorithm for Latin letters to Arabic ones (meaning logic input letters, not shaped ones)? Or are there lots of complexities? My understanding of the Arabic language is that most vowels are omitted in writing, but that might be wrong.

It would have to depend on what language you're trying to transliterate, because the phonetic meaning of the Latin letters is different depending on the language they're used to represent (and for many languages, depending on the particular word). For example, ti- will usually be /tʃi/ in Brazilian Portuguese and /tɪ/ in English. Typically each orthography is specific to a language and a script doesn't encode sounds in a consistent language-independent way, so you have to know something about both the source and target languages to devise a transliteration system.

This looks a lot how many Arabs write using Latin keyboards.

Example : "أهلاً بكم", becomes "ahlan bekom"

I find it a cook 'quick-an-dirty' trick to start speaking/writing with locals without worrying much about proper grammar (which is EXTREMELY complex)

Well, "أهلاً بكم" which means "Welcome" for second person plural in English is transliterated into "Eskéndereyya" as "ählan bekom".

Note the umlaut mark on the letter "a" as it's a Hamzat Qaŧў همزة قطع and should be pronounced with glottal stop. Of course, in colloquial speech across many dialects, it softened to just a plain "a".

Some discussion from a month ago: https://news.ycombinator.com/item?id=12956885

You may also like this universal script, Shwa. http://shwa.org/

That sounds like "Alexander". Interesting.

... not sure how well this system will fare, given that "Eskéndereyya" is 'literally' the arabic name of the city of Alexandria.

Just a minor correction, "Eskéndereyya" is Alexandria in Egyptian Arabic not MSA. In MSA, it's "al Iskándareyya" using this writing system rules.

Languages (or language families) that are influenced by Arabic (Turkic family[0], for example) `interpret` Alexander as such, albeit with slight modifications (Iskender[TR], İsgəndər[AZ] and so on). If you listen to the pronunciation[1] of Alexander in Arabic you could witness that this is indeed the case.

[0] https://en.wikipedia.org/wiki/Turkic_languages

[1] https://translate.google.com/#en/ar/alexander

Edit: fixed spacing

Yes.

In MSA, it's "Iskandar" and in Egyptian Arabic it's "Eskandar".

Your case endings in the exercise are wrong, it should be "Rȃbbi assamáu," "ära alħayáta"

That's right in abstract but in the context of poetry, the case ending can change arbitrarily and rules are bent to serve the medium, and that's what exactly Shakira did in this song. :)

So cool! Language translation standardization projects are great teaching tools

Reply on news.ycombinator.com