Orthography ... yes, more about human languages

How does your written language represent the sounds of your spoken language? How does that mess up the pronunciation of someone not from your language, but with an alphabet in common, reading your language?

Take the Dutch tennis star Kim Clijsters. The “IJ” in the middle of her family name is the Dutch IJ digraph and makes a sound that rhymes with “bay” not with “eye”! I still cringe every time I hear someone (who ought to know better) mangle the poor girl’s name.

Similarly, Irish Gaelic. Their language uses the Latin alphabet but not as an English speaker expects. Beware reading Gaelic to a Gaelic speaker if you don’t know the pronunciation rules as you may find their laughter disconcerting.

There’s a good introduction to the vagarities of orthography on Wikipedia.

It’s good that we got to type our instructions into the computer in the early days before lots of RAM, lots of gigahertz and speech recognition. I wonder what kind of languages we’d be using now if we had had to speak all our instructions.

1 Like

Dutch ij does not rhyme with “bay”. (The Wikipedia recording is misleading.) Instead, it is a diphthong of the vowels heft and eek. Dutch strictly distinguishes between these sounds: (English loanword) hey is a greeting, whereas hij means “he”, and the difference is obvious to a Dutch ear.

I cringe every time I hear an anglophone ‘pronounce’ Huygens. Where does that o come from?! I’d understand it if the u were mispronounced as woo or huh.

Like English, Dutch has a somewhat complicated vowel system. Also like in English, Dutch vowel spelling is outdated – on top of the Latin alphabet not being a great fit. (Besides the vowels, Dutch spelling is much more sensible than English spelling.)

  • All of a, e, i, o, u denote two different vowels, depending on context, termed ‘long’ and ‘short’ but actually qualitatively different as well.
  • Actually, e also denotes a third vowel, which is schwa-like but not exactly a schwa.
  • aa, ee, oo, uu unambiguously denote the ‘long’ variants.
  • The ‘long’ variant of i is ie. But it is actually short. (Contra in English!)
  • oe is like romance u, i.e. woo, but short.
  • au is exactly like romance au. The digraph ou is pronounced the same way.
  • ei in Standard Dutch is pronounced the same as ij. The two are distinguished in various accents and dialects.
  • eu and ui… after months of trying to figure out how to explain them to foreigners, I still have not figured them out. (ui’s spelling offers no hints at all; eu’s maybe a little.) See Wikipedia, I guess :stuck_out_tongue_winking_eye:
  • Standard Dutch is not the only Dutch: just like in English, the qualities of the vowels shift among accents and dialects. For example, Flemish ij is like a long ‘short’ e from Standard Dutch.

Polish workers in the Netherlands end up asking for hoersubsidie (whore subsidies) instead of huursubsidie (rent subsidies). Probably not merely a spelling issue: Polish lacks the uu vowel.

LaTeX struggles with ij! Or actually, it mostly doesn’t: it is recognized as a digraph and therefore not hyphenated. However, it turns out there is (at least) one Dutch word that should be hyphenated there: bijectie (bijection).

I have heard foreigners (not sure where they were from) pronounce Den Bosch as /dɛn ˈbɔʃ/, maybe analogous to German. It should be /dɛn ˈbɔs/. I imagine they would do the same with fantastisch (fantastic) and the like. Final -sch, weirdly, is always pronounced /s/.

Maybe later I’ll be able to come up with more – especially non-vowel issues.

I stand corrected. Thank you. And thank you very much indeed for the other glorious examples of mispronunciation and misunderstanding. The rent one is classic! As for Clijsters …

So how should one pronounce ‘Dooyeweerd’?

Modern Turkish with the Latin alphabet was a pretty much recent (1924) and “designed” language. There’s a pretty consistent mapping between graphemes and phonemes. It’s designed to be easy to learn (literacy rate in Turkey jumped from around 10-15% to 85-90% within 10 years of its introduction)

The funny side is when a native Turkish speaker tries to learn almost any other language, they usually expect a similar near-perfect mapping of what they see and what they voice, and hilarity ensues.

The inverse funny detail is the letter ı (unicode: dotless i). I guess no other language outside the Turkic family of languages has that sound or anything similar, and the way most foreign speakers of Turkish contort their throats to unsuccessfully imitate it is usually a funny sight :smiley: . And if they happen to ignore the lack of the dot, then a word like sıkıntı (boredom/problem) becomes sikinti (f**kery, almost)

1 Like

/doːjəʋeːrt/. You’d have to ask the man himself whether the stress should go on the first or on the second long syllable. Normally it would go on the second one (“dead hamlet”), but surnames can act weird.

The y in Huygens is to be read as an i, i.e. Huigens. Such y’s occur only in old words and surnames.

Wikipedia suggests I will produce it when I imitate a Flemish accent. Do you agree? (There is a recording in the table.) (This was about the close back rounded vowel instead of the unrounded one.)

Though ı is the close back unrounded vowel… almost.

Oops. I’ll retract it.

1 Like

I’ve also never managed to get foreigners to understand eu and ui (whereas the g sound is relatively easy once French speakers get it’s similar to their r), but I really lose them on eeu in eeuw and sneeuw and plural uien or worse uieren . ooi, oei, aai, ieuw don’t have the same pain in explanation in my experience.

Neither sound (eu and ui) seem to have close resemblance to other sounds in English, but the former is easy again for French speakers as they have it in lieu and deux. Germans? Try your word for hairdryer! The ö in Föhn is alike our eu. For anglophones I often try this by changing soy. soysuy → now stop pronouncing the uy transition by not pushing your tongue against your teeth. Alternatively take a word like or I try it with with a word like hurtur and now don’t pronounce the r (again, don’t press that tongue against those teeth), and round your lips.

On the topic, my biggest hardship with foreign friends who are learning the language is to make them realize that like some other western languages, the way we make the first person form of a verb is by taking the infinitive and cutting of the “end”, or taking the infinitive and adding zero to a few letters, but unlike those other western languages, you then need to make sure it still has the same sound. So as @MatthijsBlom indicated we have long versions and short version of various vowels, and that context is often “are the next two letters two consonants or a consonant-vowel pair?”. If it’s the later, the sound is long.

Take those two things together and this is where my friends struggle most in the first year.

For example:

The verb for to walk in Dutch is lopen (the o is pronounced as oa in boat, it’s long because it’s followed by a single consonant and then a vowel.). The first person form isn’t lop. That sound would be a short vowel, and be pronounced as the o hop or lot. The only way to keep that sound the same is by adding another o, so it becomes loop.

I walkik loop
He walkshij loopt
We walkwij lopen
to walklopen

Visually to me the context of vowels (followed by how many consonants_) is super clear, but in practice seems hard for many non-native speakers. The conjugation of especially older verbs makes it harder to write if you don’t know how to speak (or what the rules are).

But … can you explain it by using this pink trombone simulator for speech?

1 Like

(This is true for @SleeplessByte’s regional accent, but not generally. The pronunciation of Dutch g/ch is diverse; a very noticable example of this variation is ‘hard’ vs. ‘soft’ g – not to be confused with a similarly named phenomenon in English and other languages.)

Isn’t German ö just like Dutch huh? eu seems a diphthong to me, somewhere near uh⁠→⁠uu. (Similarly, I estimate ui at schwa→⁠uu.)

Anglophones apperently often struggle with French’s rue, which Dutch shares (as uu). I have heard it instructed as

  • First, pronounce oo.
  • Then, while keeping your face exactly the same, pronounce ee.

I agree: the vowel story here isn’t complicated, it’s the spelling that is confusing, if the spelling is your starting point.

Found some more potential confusion, but is is shared with English: some letters are pronounced differently in different positions. Examples: (English) level, (Dutch) lepel. Here, the second l is velarized whereas the first is not. In Dutch, you can velarize the first one, but then you’ll sound like you are from one of a few specific cities or neighborhoods. However, weirdly enough, if you unvelarize the second one you might not be understood at first (even though it is not at all ambiguous): it sounds weird. In the case of l, anglophones do this correctly. Some other letters (may) do similar things, but I cannot think of any that stand out.

Strictly an orthography issue: obstruents at the end of words become unvoiced, but the spelling does not reflect this. It is wij hebben /ʋɛi̯ ˈɦɛbə(n)/ (“we have”), but ik heb /ɪk ɦɛp/ (“I have”). You will not be misunderstood when you do this wrong, but you might stand out (depending on circumstances).

Neat! I’m still having trouble generating diphthongs (by mouse), and I can’t find a record/export functionality.

1 Like

I disagree in the sense that it doesn’t matter. Getting them to a familiar sound makes it work in my opinion :stuck_out_tongue:. I’ve tried to make them pronounce “kei gaaf” and you can imagine how that was much much harder than a hard g :sweat_smile: .

Isn’t German ö just like Dutch huh? eu seems a diphthong to me, somewhere near uh⁠→⁠uu. (Similarly, I estimate ui at schwa→⁠uu.)

Interestingly, Germans have, just like us, short and long sounds. Words like to open (öffnen) or spoon (Löffel) have the ‘short’ sound and that’s like our u in huh. But their longer variant as in oil (Öl) and pretty (schön) are like the eu sound!

[…] but you might stand out (depending on circumstances).

Yeah, both ways, as you indicate this falls well into accents and dialects. In the same vain, so does it when you do reflect it in writing which isn’t considered “correct” but absolutely happens and can be frowned upon when done, and in some groups even when you don’t.

Do we in Dutch have two vowels that have the same quality but different quantity? My mental model does not contain such pairs. (E.g: had is short but is qualitatively different from haat as well. Moreover, a long form of the former and a short form of the latter do not exist.)

Some leafing through the Wikipedia teaches: apparently there is regional variation here; in some places eu is a diphthong, in some others a long monophthong and the same as German ö.

Do we in Dutch have two vowels that have the same quality but different quantity?

I was “forced” to learn about language evolution and etymology in English class in High School, including proto- and indo-european, and the only things I can recall that vaguely resemble what you’re asking about is that in “older Dutch” (but not old Dutch) people would sometimes write vowels that are long by the single consonant + vowel follow up rule with the double / long variant, e.g. roken and rooken.

This is a very nice article about this piece of history in our language AND they also point out some of the regional differences.

In middle (and perhaps old, but I don’t recall) we wouldn’t necessarily write double-the-same-vowel, but rather add an e or i, think jaer instead of jaar. In this case it’s the same amount, but different pairs :slight_smile: .

Unfortunately it’s been very very very long since I studied this material so I’m certain a lot of other examples (and nuance) has been lost between what was taught then and what I reproduce now :stuck_out_tongue:

apparently there is regional variation here;

Myeah. There’s always going to be that with vowels in most western languages.

To circle back to @axtens original question, perhaps you want to highlight the lovely thing where in some regions we completely switch the vowels around, but only in some combinations (twents etc.), or add vowels where there are none…

Oh I think I just came up with two that sorta are interesting.

ie in Words we “lend” from French such as opticien (optician) or lesbienne (lesbian) are not pronounced like the earlier ie examples.

I guess we also have that for words where the vowel combinations are across a syllable, such as geuit where we have e and ui and not eu and i, or ooievaar where we have oo and ie, but ie is pronounced as je.

But … can you explain it by using this pink trombone simulator for speech?

Neat! I’m still having trouble generating diphthongs (by mouse), and I can’t find a record/export functionality.

I guess the record functionality would have to come from a video camera, or screen recording, or some other mechanism. (Perhaps using a tool that is programmable, such as xdotool, or something similar, even.)

UPDATE: I wonder if that simulator is able to use a multi-touch interface.

Pink trombone is great :star_struck: Just wishing I could link to particular settings here.

This is the Turkish ı sound.

That’s pretty cool. How does it represent the voiceless alveolar lateral fricative, where the airstream goes over the sides of the tongue, rather than down the middle (“ɬ”)?