Welcome back to Let’s Conlang! Last post, we worked hard on giving our conlang some totally radical consonants, but our language doesn’t have any vowels yet and without vowels a language is just lngg …. so, uhh… yeah. Vowels. Let’s get some of those.
Before we start talking about vowels, I want you to forget everything you know about vowels in English, because English vowels are an absolute mess and generally spelt in the most unintuitive way possible¹. Second, go listen to these vowel samples. It’s not really that important for this post, but it is pretty fun trying to make them sing hot cross buns or something.
Now, in natural languages, vowel systems can range from the very simple (e.g. Arrente, an Australian Aboriginal language with only two vowels) to the very complex (e.g. Danish, which has infinity-million). For our conlang, I think it would be nice to do a goldilocks and go for something in the middle, so let’s get started.
If you went to that link I linked or spent some time looking at the vowel section of your IPA chart, you might have noticed the words close, mid and open along one axis and front, central and back along the other. These are basically describing the position of your tongue when you articulate² the sound.
The first set says how high your tongue goes while you articulate the vowel. For example, /i/ and /u/ are close (or high³) vowels because the tongue is held high or close to the roof of the mouth (depending on which terminology you prefer) when you say them. Similarly, /a/ and /ɑ/ are open (or low³) because the tongue is held low or such that the mouth is more open (hence doctors telling you to “say ah”). And, between the two extremes, we get our mid vowels like /e/ and /o/.
The other axis which decides what a vowel sounds like is that of front-/backness. On this axis, /i/ is front because it’s pronounced with the tongue pushed forward and /u/ is back because it’s pronounced with the tongue pulled back. In English the central vowels are reduced vowels like /ə/ (the unstressed e in cromulent⁴).
As with most matters of phonetics, the best way to quickly understand what’s going on is to try it out. Try saying beat and compare the vowel to that in bet and that in bat. You should notice that your tongue gets progressively lower for each. Now compare beat and boot and you should feel your tongue moving back in your mouth.
OK, enough tonguing; it’s time to give the lips some attention. The third way that vowels can be differentiated is by their rounding. When articulating a rounded vowel, like /u/, the lips are – you guessed it – rounded. And when articulating an unrounded vowel, like /i/, they’re not.
In English, back vowels are rounded and front vowels are unrounded, so we don’t really have to worry about distinguishing vowels only by their rounding, but a lot of languages do make the distinction and, fortunately, it’s one that is quite easy to pick up, even for English speakers.
Right, so that’s enough introduction. Let’s get to picking our vowels. Let’s start with the most common vowel around: /a/⁵. Now, not all varieties of English have this vowel, but in Australian English (my dialect), it’s the vowel in cut. If you’re familiar with Spanish, French, Italian, Japanese, German, Polish, Turkish, or pretty much any other language, it’s the sound of the vowel a.
(Note that the English examples in the next few paragraphs are generally approximations. Refer to the vowel samples linked above if you want a more precise idea of what the vowels sound like.)
As you’ll recall from our work with consonants, phonology is all about which differences are going to be meaningful and the most commonly used distinction in vowels is height, so let’s go with that first. We’ve already got an open vowel (/a/), so let’s grab a mid vowel, /e/ (the e in bed), and a close vowel, /i/ (the i in bid), to go along with it. Sweet.
We could go on to have a distinction between close-mid and open-mid vowels, like a lot of languages do, but in the interests of making this language easier to speak (and write down), we’ll just stick to the three heights.
Now, let’s use our second axis and make some back vowels: a mid one, /o/ (the o in bod), and a close one, /u/ (the oo in boot (not the u in bud)). You’ll probably notice both of these are rounded. That’s because it’s not just English that likes to round its back vowels; it’s a pretty common feature of most languages.
Now, most languages also tend to just leave rounding to the mid-to-high back vowels, like English does, but that’s a bit boring, so let’s also add ourselves some rounded front vowels, namely /y/ (the rounded version of /i/) and /ø/ (the rounded version of /e/)⁶.
But how the hell do you say those? It’s not as bad as you might think. Just say /i/ and /e/ but with your lips rounded at the same time. If it helps, /y/ is the u in French zut and /ø/ is the eu of sacrebleu.
Cool. We’ve got seven vowels now, but hang on. A language is a system where the sounds are there for a reason, not just because someone threw them in there⁷. Why would we have these rounded front vowels? My solution: palatalisation.
Without going too far off-track, let’s just say that /y/ and /ø/ evolved from historical /u/ and /o/ preceded by a palatal consonant. Don’t worry if that doesn’t make sense to you now; we’ll probably get into it in more detail next post.
Next up are long vowels. If you didn’t remember to forget everything you know about English vowels, you might be thinking a long vowel is like the i in bite. That’s not a long vowel; it’s a diphthong. (Be patient. Diphthongs are coming up soon.) You might also think that a long vowel is like the ee in feet. That one is a long vowel, but it’s not a long e, it’s a long i.
All the confusion with English spelling might make long vowels seem complicated, but they’re not really. A long vowel is exactly the same as the short version of that vowel, but you keep saying it for longer. For example, English bit has a short /i/ and English beat has a long /iː/. (Once again, English let’s us down, because the i in bit is actually slightly centralised and lowered to /ɪ/, but let’s ignore that.) The “ː” bit here is what tells you the vowel is long. If you look closely, you’ll see it’s not quite a colon⁸, but if you use a colon instead, everyone will know what you meant.
In our conlang, we could just let every vowel be either short or long, as is the case in Latin, but let’s instead tie vowel length in with stress. We’ll get into stress when we start working on the prosodic features of our language, but for now we’ll just leave it as a vowel is long if it’s in a stressed open syllable and short otherwise.
With both short and long vowels handled, you might think we’re almost ready for diphthongs, but before we get into them, I’m going to jump back a bit to our consonants. Specifically, the approximant consonants which we looked at last post, but couldn’t finish until we’d got our vowels sorted.
Semivowels are basically the consonants that you can make by putting your tongue in the same place that you would for their corresponding vowel. English has two semivowels: /w/ (the w in water, corresponding to /u/) and /j/ (the y in you, corresponding to /i/). These are the two most common semivowels and, since our conlang has both /i/ and /u/, it makes sense that we’ll have /j/ and /w/ too.
You may have noticed that these semivowels correspond to our close vowels and our conlang has another close vowel: /y/. You will be pleased to find out that, indeed, there is a semivocalic version of /y/ and it’s represented in the IPA by /ɥ/ (an upside-down h). To figure out how to say that, just say /j/ but with your lips rounded.
Now, what about the other vowels? Are there semivocalic counterparts to /e/, /ø/, /o/ and /a/? Well…. yes, there are, but most of them don’t occur in many (if any) natural languages. Semivocalic /e̯/ and /o̯/ do occur in Romanian, but being so similar to /j/ and /w/, they are very rare cross-linguistically. It’s a similar story for /ø̯/, which is too similar to /ɥ/. As for /a/, it’s hard to say. It’s possible to have a semivocalic /a̯/, but I’ve not come across any evidence of a natural language using it.
And finally, we reach diphthongs. I love diphthongs, mostly because they give me an excuse to use my favourite ever sequence of sounds: /fθ/⁹. You can also pronounce it “dipthongs”, but what kind of horrible monster would even contemplate such a heinous act?
Diphthongs, from Ancient Greek di- “two” + phthongos “sound”, are literally that: two vowel sounds put together into one super-powered-mutant-vowel-thing! Examples in English are /ai̯/ (the igh in sigh), /au̯/ (the ou in sound) and /ei̯/ (the ey in grey). These are “falling” diphthongs, so called because their second element is less prominent. “Rising” diphthongs such as /i̯a/ also exist, but since we won’t be having any in our conlang, I won’t go into them here.
The above English examples are all also closing diphthongs, so called because they end in a closer vowel than they started with. In addition to these, non-rhotic dialects of English (such as Received Pronunciation or Australian English) have centring diphthongs like /iə̯/ (as in fear) and /uə̯/ (as in pure). Once again, we won’t be having these in our conlang, so I won’t go any further into it.
(Note that the little bump¹⁰ below the vowel indicates that it’s the less prominent element of the diphthong. Another way to show this is by using the symbol for the semivowel for the less prominent element; that’s what I’ll be doing from here on.)
What we will have in our conlang are diphthongs which are both falling and closing; that is, they’ll all end with one of our close vowels. The first three will be the ones starting with /a/: /aj/, /aɥ/ and /aw/. Now, moving up to start on the mid vowels, we’re going to avoid mixing /y/ and /ø/ with the back vowels (because it wouldn’t make sense from the sounds’ histories), so we get: /ew/, /ow/, /ej/, /øj/, /oj/, /eɥ/ and /øɥ/.
So there we go: seven monophthongs (our short vowels) and ten diphthongs. That’s our vowels done and the last of our consonants done too.
And, with that, we can make some nice pretty tables illustrating all of our phonemes:
Next post, we’ll come up with an orthography and start setting out the rules for how these phonemes can be put together to make our language’s words. Stay tuned!
– Ben xoxo
¹ Yes, I’m aware that I’m being somewhat hyperbolic here, but English is certainly one of the least sensibly spelt languages in the world (slightly behind Icelandic). In case you’re interested, the main reason our spelling makes so little sense is that most things are spelt as they were said in the late fifteenth century, so, on the bright side, at least it’s easier to figure out how people spoke during the Wars of the Roses…. I guess.
² Or “say”, if you insist on using confusing technical terminology.
³ A lot of linguists like to avoid using the terms high and low to describe tongue position because they think it will make the low vowels feel bad. Also, it can get confusing when you’ve also got high and low tone going on.
⁴ Note that cromulent can also be pronounced with a syllabic n, instead of the sequence /ən/ and, most of the time, is. But it’s a good word, so whatever. Sue me. (But don’t, ‘cos I have no money.)
⁵ The IPA symbol a can stand for any vowel between front /a/ and central /ä/. The difference isn’t that big and for our conlang, we don’t really care which it is.
⁶ The kind of rounding we’d find on these front rounded vowels would probably be with the lips compressed, rather than the protruded rounding of back vowels, but that’s a minor detail which we needn’t really worry about.
⁷ Which is, technically, a reason, but not a very good one if we’re trying to imitate natural language. Work with me here.
⁸ Although Unicode calls it “modifier letter triangular colon”, it’s clearly actually “two tiny triangles pointing at each other”. There’s also a half-long symbol which looks like this: ˑ. Just one tiny triangle on that one. Yep.
⁹ Another good word for the /fθ/ combination is phenolphthalein, which is even better because it has both in the onset of the same syllable. Just say it with me; phenolphthalein. Beautiful, isn’t it?
¹⁰ Unicode calls it a “combining inverted breve bellow”, but what would Unicode know? It thinks the length sign is a colon. Silly Unicode. (Sorry, Unicode, I still love you, baby.)