Hacker News Reader

Comment by idlewords | Hacker News Reader

Polish orthography is a disaster partly because of the choice of Latin over Cyrillic. The latter would have been a much better fit for the sounds in the language.

Making matters worse, there are homophones (ż and rz, u and ó, ch and h) that depend on when the word entered the language.

Like so many things wrong with the country, you can comfortably blame the Catholic Church for this orthographic train wreck.

View on HN

Polish letters on top of Latin are no worse than German (ä, ö, ü, ß), or French (é, à, è, ù, â, ê, î, ô, û, ë, ï, ü, ÿ, ç). Cyrillic is not the answer, as each Cyrillic country has own variations, own letters. See https://en.wikipedia.org/wiki/List_of_Cyrillic_letters

Regarding homophones, this is not 'when word entered the language' but rather, the change of pronunciation no longer matching the phonetic writing.

The 'ó' vs 'u' is homophone because the sound now is the same. However, historically it wasn't. Additionally, since historically it was different sound, it also had different rules for how it morphed during declension.

Similar situation is with 'ż' and 'rz'. There are even cool words that due to declension we can see where they originated from. For example two declensioned words have same pronunciation: 'każe' and 'karze'. The first comes from root word 'kazać' as in tell people what to do. The second comes from root word 'karać' meaning punish.

I would wager that 'ch' and 'h' is the most recent homophone. Some people still alive were taught to use correct hard 'H' vs soft 'CH' when pronouncing words.

Thanks for this correction!

I would say citing all those French diacritics is cheating a bit. The diaeresis is just used to indicate separate syllables, and ù and ÿ are not a thing in standard French.

Polish is uniquely bad not because of the diacritics, but because of the special phonetic behavior of digraphs ('si', 'ci', 'zi', 'rz', 'sz', 'cz', 'ch', 'dż', 'dź', 'dzi').

To strengthen your argument there is another digraph 'ni'.

But this is now approaching discussion that writing doesn't match phonetic. Pronunciation shifts. Words change. Brother Grimms (famous for fables) were linguists studying how consonants use has changed. They talk how 'pater' (like paternity) shifted to 'father'. But Polish has own such shifts. Take a word like 'jabłko', which many people pronounce 'japko' or 'śliwka' that is pronounced as 'ślifka'.

There is even a branch discussing how people using correct pronunciation are HyperCorrect https://en.wikipedia.org/wiki/Hypercorrection#Polish

I wouldn’t call it “using correct pronunciation”—the whole idea of hypercorrection is that the result is not correct, because some perceived rule was mistakenly over-applied, like in English * “He gave it to Sam and I”.

(Plus I’m guessing that sound change happened pretty early on, because as far as I know, all Polish consonant clusters obey that rule of voicing & devoicing obstruents.)

Digraphs aren't unique to Polish, though. Take Hungarian as an example ('cs' or 'zs' to start with).

There's at least one word that uses ù in French: où (where, as opposed to ou, or)

I drove today next to the "nonstandard" Haÿ-les-Roses (south of Paris)

The curious thing about Cyrillic is that, at least in Slavic languages, the correspondence between phonemes and letters is much more uniform than in Slavic languages that use Latin.

For example, "ч" is always the "ch" or "tsh" sound, whether we're talking about Russian, Ukrainian, Bulgarian or Serbian. OTOH, in Polish you use "cz" for it, while in Czech and Serbian it's "č".

It's not an inherent advantage of Cyrillic, of course, it's just that, by virtue of being designed specifically for Slavic languages, it had certain letters designated from the get go, while adopting Latin allowed for a lot more leeway in deciding how to represent sounds, and different languages did it differently.

It's not only that Cyrillic is designed specifically for Slavic languages, but also that the general tradition around Cyrillic is to make up new letters when adapting it to a new language, as oppposed to the Latin alphabet where the two traditions are to use digraphs and to add diacritics. This choice between two possibilities is already causing the problem of "cz" vs. "č", but within those there are still the choices of "cz", "ch", "tsh", "tch", ... and "č", "ç", "ĉ", ...

I think the line between "making up new letters" and "adding diacritics" is rather blurry, and really hinges on your definition of "new letter", which is largely defined by convention. Most (all? I can't think of any that don't) Cyrillic alphabets have some letters with diacritics - ё and й in Russian; й, i and ї in Ukrainian; j in Serbian etc.

But speakers of those languages don't consider the diacritics to be a modifier in this case - we treat them as completely separate letters, that just have a disjoint element in their overall shape. I don't see why e.g. c vs č can't be treated in the same exact way as и vs й (and indeed, I wonder if Czech don't already do that?).

And yet - somewhat hilariously, Polish is one of the easiest languages out there in terms of knowing how to pronounce the written word.

Spend a week or two practising a pronunciation guide and that's pretty much well it.

Reading something out loud for the first time? Native level accuracy 95% of the time, easy.

Compared with the absolute crap shoot which is English or - my nightmare - French, reading Polish out loud is an absolute walk in the park.

You're completely right. There's also a predictable syllable stress pattern. I taught a friend once to read Polish out loud fluently without understanding it.

My mouth actually hurts when I speak French. Such unusual shapes for a native English speaker :\

In a famous polish novel "Krzyżacy" French language is compared to shaking of tin bowls, so even Polish speakers have troubles.

English is indeed a crap shoot, but it's not entirely fair to put French in that bucket as well. Much as you describe for Polish, there is a highly consistent set of writing-to-sound rules for French that you can learn quickly. French's downfall is in the other direction: with multiple ways to write some phonemes and many contexts where letters are silent, hearing a word, even perfectly, is not nearly enough to know with any certainty how to spell it.

But reading French out loud? Anybody with a week or two of practice should definitely be able to know how to pronounce well over 95% of what they read, even if they don't know what they're saying.

Irish and Korean are also extremely regular. Hangul (Korean) is probably the easiest alphabet in the world to learn to pronounce. Irish suffers from the use of digraphs in modern use, but the older style of placing a dot on top of the letter instead of an h afterwards is much clearer IMO.

Who do we blame for Vietnamese adopting the Latin alphabet? That makes Polish seem pretty straightforward.

For some reason, they really didn't like China!

You can blame the French. It was a good choice though. Vietnamese and Korean are probably the easiest East Asian languages to learn to read.

I feel like Indonesian and Malay are far easier than either of those, at least for English speakers.

Could have done what Serbian (and formerly Serbo-Croat) does: freely added letters to the Latin alphabet to match the required set of phonemes, so that there's a one-to-one correspondence between the Latin orthography and the Cyrillic orthography.

(It doesn't solve the problem of missing letters and missing diacritics when trying to write your language using English letters only — but Serbs, Ukrainians etc. can't write their language correctly using only Russian letters either.)

We just have more of this crap than other Slavic languages. Specifically:

ś—sz

ć—cz

ź—rz/ż

dź—dż

Mandarin of all languages has this phonetic distinction, but I don't think any other Slavs kept the full set, on top of the nasals mjd wrote about.

Wikipedia says Mandarin has all of /t͡s/ ⟨z⟩, /ʈ͡ʂ/ ⟨zh⟩, /t͡ɕ/ ⟨j⟩, /t͡sʰ/ ⟨c⟩, /ʈ͡ʂʰ/ ⟨ch⟩, /t͡ɕʰ/ ⟨q⟩, /s/ ⟨s⟩, /ʂ/ ⟨sh⟩, and /ɕ/ ⟨x⟩, all contrastive. The fricative and affricate sounds in English that are kinda close to these are only /t͡ʃ/, /ʃ/, /s/, /t͡s/, /ʒ/, and /dʒ/.

I think these contrasts represent a very underacknowledged difficulty for English speakers learning Mandarin, because we're used to thinking of tones as the only hard thing. I'm still struggling to properly pronounce a friend's name that I think begins with /ʈ͡ʂʰ/. Maybe Polish speakers can deal with this easily!

My pronunciation of these consonants was improved a lot by reading the respective Wikipedia articles and imitating the described tongue position. That got me interested in the rest of the IPA, and for a while I spent idle time making strange sounds with my mouth.

Retroflex consonants are still awkward, but at least I can produce them reliably. What's really hard for me are uvulars, they feel a bit like choking every time.

I spent a long time learning to say [k'], and when I succeeded I had shirts made for myself and a friend who was practicing with me.

They say:

"I pronounced the velar ejective consonant [k'] and all I got was this [t']-shirt."

Everyone in this thread was a Linguistics major then?

I was a dropout, but I wish I had been a linguistics major!

There was a post on Language Hat or Language Log where a professor recounted meeting a student who had linguistics "as a hobby", which was astonishing to the professor -- and the student explained that this was possible nowadays because of Wikipedia. (To which I could add, for some people because of the conlang community, or because of linguistics blogs!)

Anyway, Wikipedia's coverage of linguistics topics is generally really excellent and detailed (maybe strongest in phonology, perhaps because of a few super-obsessed editors).

Yes, Polish is the worst. Serbian has ć—č and the voiced equivalent spelled đ—ǆ, but it doesn't have ś—š and it doesn't have nasals.

I "think" these should be pretty close to Cyrillic ш/щ, ч and ж.

It's a trap! Roughly speaking, the ones on the right are those. The ones on the left are different and they correspond historically to something like сь, чь, зь (though the pronunciation is rather different from standard Russian pronunciation).

well, sure, it's not uncommon even for ppl who speak the same language and use the same letters to pronounce things differently.

the point is I don't think Latin vs Cyrillic in Poland's case was driven by Polish language phonetics per se. I would venture to guess this had more to do with political/religious affiliations etc.

Absolutely. Mieszko I chose to convert himself and his realm to the Roman faith for both political and religious reasons. Had he chosen to accept Christianity from the East (remember, this was pre-schism), it would have entailed the adoption of Cyrillic.

Half of them are, the other half are the 'matching' pairs that don't exist in Russian (for example). Soft ш and ж, hard ч and щ.

Czech also did a great job of this (unfortunately the word Czech itself seems to come to English through the some other spelling system). A few years ago I tried to learn a bit of the language before travelling to Prague, and I was pleasantly surprised by how easy it was to learn to read and write.

Prior to changes introduced by Jan Hus (in this case, the introduction of the háček), Czech shared digraphs that are still used in Polish. "Cz" is was ultimately replaced with "č". It is possible that the archaic form somehow made it into English and survived (though my sources tell me that the spelling has only been present in English since the 17th century, or after Hus' reforms).

"The latter would have been a much better fit for the sounds in the language."

The Cyrillic alphabet as originally developed for Old Church Slavonic lacks several sounds of Polish, namely the velarized /l/ (which has now become a semivowel /w/ except in peripheral dialects), and the palatalized affricates /ź/ and /ć/.

If you are a Russian speaker, you might think that Cyrillic could represent palatalized sounds by use of the soft sign, but that it not what the soft sign was actually used for originally. It was meant to represent the front reduced vowel /ĭ/ before the fall of the yers. So, Russian choose to represent its phonology by extending the Cyrillic alphabet in a way that was originally unintended, while Polish chose to represent its phonology by extending the Latin alphabet. How is either of these choices better than the other?

Worth noting that Serbian did its own thing, and added new Cyrillic letters for palatized L and N as separate letters Љ and Њ, (which are obviously digraphs of Л and Н with Ь). Although that's a late 19th century invention.

Yes, modern Serbian orthography only did so in the end because Vuk Karadžić was intentionally following the example of Russian.

Ah, but it's not really an example of Russian in that sense, since he made them separate distinct letters rather than reusing Ь as a modifier, as Russian did. So semantically a rather different approach that just looks similar because it appropriated the shape.

Polish orthography is weird (subjective opinion!) mostly because of its weird digraphs, IMO. Czech and Serbian also use Latin, but they are easier to parse.

I'm not a linguist, but looking at the differences between the corresponding alphabet, it feels like Polish one had a stronger German influence. The use of W rather than V stands out in particular (and makes no sense in an alphabet that doesn't use V at all!). But also, Germans love their digraphs and trigraphs.

This may be true, but may I suggest two things:

1. Any suitable choice of orthography will become unsuitable eventually. Pronunciation is not static.

2. There is, and has been, quite a bit of dialectical variation within the Poland and its diaspora. What may be suitable for one group, may not be for another.

Reply on news.ycombinator.com