Let’s Conlang – Part II “Let’s Get Phonetical”

In the last instalment of Let’s Conlang, I went through the basics of what conlanging actually is, then gave you a very basic primer on phonology, phonetics and all that good stuff. You might want to open that in another tab now just to have the IPA chart on hand, but just because I’m an awesome guy, I’ll also link you to the Wikipedia article on each group of sounds we use as we start the epic* process of…

Building a Phonology

The first step in building a phonology is to decide which distinctions are going to be meaningful. Will, say, the difference between [pʰ] and [p] be enough to create a minimal pair¹ (as in Ancient Greek φόνος /pʰónos/ “murder” vs. πόνος /pónos/ “work”) or will it be allophonic² as in English?

There are infinity and one different distinctions you could use building the phonology of your conlang (don’t question it; I’ve counted them all), but a lot of them are unlikely to ever occur in a natural language and a lot of those which do occur in natural languages can be very hard to pronounce.

Sure, there’s nothing stopping us from making a language with two hundred consonants and no vowels or something equally outlandish, but while that could be fun, for our purposes here, I think it would be nice to create a language that we have some chance of being able to pronounce. So let’s begin, starting with…


1. Plosives

Let’s start with plosives, since pretty much every language has at least one of these. The most common plosives cross-linguistically are /p/, /t/ and /k/³, so let’s start with those. Now a very common distinction for plosives is voicing, so let’s add voiced variants of our plosives: /b/, /d/ and /g/.

Now we’ve got six plosives, but that’s a little boring. We want to imitate natural languages, but we don’t want to be sheep, blindly conforming to the corrupted norms of a soulless society, lead to the slaughter by the consumerist mass media and politicians whose only allegiance is to rapacious corporations, drowning in the fetid taint of world rotten to its core! … or something. So let’s add another distinction: prenasalisation.

A prenasalised plosive is basically exactly what it says on the tin: there’s a nasal sound before the plosive (much like the sequence nd in hand but functioning as a single unit). We could double our plosives here, by having prenasalised voiceless and voiced plosives, but let’s just go for one prenasalised series. Since nasals are naturally voiced sounds⁴, it makes sense to have a series of prenasalised voiced plosives, giving us /mb/, /nd/ and /ŋg/ for a total of nine plosives.

These prenasalised plosives are going to have some interesting implications for the distribution of some sounds in our language, depending on how we decide they could have come about, but we’ll get to that later. For now, we’ve got our plosives sorted, so let’s get onto another method of articulation: fricatives.

2. Fricatives

Less than 10% of languages in the real world lack fricatives (with most of these being native languages of Australia and New Guinea), so it’s probably a safe bet to include some in our conlang. Languages often have fricatives in the same places of articulation as their plosives⁵, so let’s give ourselves some fricatives: /f/, /s/ and /x/ (the ch in loch or Bach), corresponding to /p/, /t/⁶ and /k/ respectively.

But we wouldn’t want to stop there. So let’s add ourselves another. When it comes to fricatives, we’re really spoiled for choice: there’s a whole bunch of them and they’re mostly not too difficult to pronounce. For our purposes, one more fricative should be all we need, so let’s pick something nice and typical: /ʃ/ (the sh in ship).

Now, we have a choice here whether we make voicing a phonemic feature of our fricatives or not. Unlike with our plosives, here I’m going to go with no voicing distinction. As in Old English, the fricatives will be voiced intervocalically and when adjacent to voiced consonants, but the voicing will be entirely allophonic.

That might sound like an odd decision, but there are two good reasons for it: firstly, it’s different to what we’re used to in (modern) English, making it more interesting, and, secondly, the distinction is actually only found in about a third of the world’s languages, so we’ve got fairly good grounds for not making it here.

So, those four fricatives bring our total consonants up to thirteen, which is a lot more than Central Rotokas’ six, but still a far cry from English’s twenty-four⁷. So where to from here? There’s the sonorants, which we’ll get to soon, but there’s one more set of obstruents left to look at first: affricates.

3. Affricates

Affricates are funny little things. With only a cursory glance over an IPA chart, you might even miss them completely; they tend to get tucked away off to one side, or relegated to a footnote, but affricates are pretty common cross-linguistically and they’ll be an important part of our conlang. An affricate is basically just a plosive which is released as a fricative and diachronically they often develop from plosives which have been mutated by various processes⁸. How our affricates developed is not that useful to know right now, but it will come to be important later, so it’s good to keep it in the back of your mind.

A whole range of affricates are possible, but most of them are fairly rare. The most common by far are those with /t/ as their plosive element, so let’s roll with that. I know I said it doesn’t matter where the affricates came from, but it actually kind of does, it’s just not something we’ll be getting into quite yet, so just take my word for it that we’re going to want three affricates: /t͡ʃ/ (the ch in chair), /d͡ʒ/ (the dge in bridge) and /nd͡ʒ/ (similar to the nge in hinge), corresponding to our three types of plosives (voiceless, voiced and prenasalised).

OK, so all the obstruents are finally done. Now it’s time for sonorants.

4. Nasals

Nasal consonants are extremely common; there are virtually no languages without at least one⁹. In fact, /m/ and /n/ are two of the most common phonemes in the world. With that in mind, our language is going to have /m/ and /n/, for sure.

Now which other nasals we’re going to have (if any) is much more interesting. About half of the world’s languages have /ŋ/ (the ng in fornicating) and since we have /g/, and /ŋg/, it makes sense to have it in our language. Another cross-linguistically fairly common nasal is /ɲ/ (the ñ in piñata); we’ll take that too, giving us four nasals: /m/, /n/, /ɲ/ and /ŋ/.

Easy. Next?

5. Laterals

Lateral consonants are basically L-ish consonants¹⁰, so it should come as no surprise that the most common lateral consonant is /l/ (as in the L and possibly the l in Lando Calrissian¹¹). We’ll take one of those and, since we have a palatal nasal, it makes sense for us to get the palatal lateral /ʎ/ (which is not a lambda (λ), but a y turned 180°) as well.

The above two sounds are lateral approximants. There are also lateral fricatives, affricates with lateral releases, lateral flaps and even lateral clicks and ejectives. Most of these are pretty rare (though the fricatives aren’t too so) and, since we’re not going to be having any of them in our conlang, I won’t bother going into them here.

6. Approximants

So we’ve just seen some lateral approximants, now it’s time for the regular kind of Approximants. Approximants are odd; they’re like consonants having identity crises. Some of them overlap a lot with voiced fricatives and in some places of articulation, no language bothers distinguishing the two. Don’t worry; we’re not going to be using those crazy buggers. We’ going to use the crazy buggers who can’t quite decide if they’re consonants or vowels: semivowels.

Semivowels are sounds like the y in you and the w in will do as I command. Now which semivowels we have in our language is really going to depend on our vowel system, so unfortunately, we’re going to have to come back to the semivowels next post, but trust me: it’ll be worth the wait.

And with that, we’re onto our last consonant:

7. The Rhotic

If a lateral is an L-ish consonant, a rhotic is an R-ish consonant. Nobody’s entirely sure what makes a rhotic a rhotic. Without throwing a whole bunch of new terms at you, the main unifying property is that they tend to pattern similarly to lateral approximants, but there’s not much else that all rhotics have in common. It’s all very interesting, but not very relevant, so let’s get to it. Rhotics are common cross-linguistically (though there are quite a few languages, most notably in North America, which lack them), with most languages that do have a rhotic having only one. For our conlang, we’ll go with that and we’ll just leave it as a rather ambiguous /r/ for now. Once we get into allophony in a couple of posts, we’ll be revisiting our rhotic and deciding on the particulars of its pronunciation(s).

And that’s our consonants (almost) done. We’ve got a good list now of nine plosives (/p/, /t/, /k/, /b/, /d/, /g/, /mb/, /nd/ and /ŋg/), four fricatives (/f/, /s/, /ʃ/ and /x/), three affricates (/t͡ʃ/, /d͡ʒ/ and /nd͡ʒ/), four nasals (/m/, /n/, /ɲ/ and /ŋ/), two laterals (/l/ and /ʎ/) and a rhotic (/r/), giving us a grand total of twenty-three consonants, plus whatever semivowels we decide to use. That’s fairly respectable.

Before finishing off with the consonant inventory of our language, however, there’s just one thing left to do. You’ll probably notice that we’ve got a pretty nice, orderly system going on here, but natural languages are rarely so tidy (though they certainly can be) and, as some guy said once, I feel like destroying something beautiful… or something like that. So let’s take our nice column of velars (/k/, /g/, /ŋg/, /x/, /ŋ/) and mess it up. A good target is /x/, which can be a volatile sound; instead of having it be a velar fricative, we’ll have it debuccalise to /h/ in syllable onsets¹², palatalise to /ç/ (the h in human) in syllable codas after front vowels¹³ and stay /x/ elsewhere. And that should give you a taste of the allophony we’ll get to start playing with once the basic phonology is in place!

Now, I’ll leave you with this pretty table¹⁴, neatly summarising all the work we just did (click to embiggen).

Isn't it beautiful?

In the next post, we enter the exciting world of VOWELS!

– Ben xoxo


¹ Two words differentiated only by a single sound (e.g. cat /kæt/ and pat /pæt/ whose only meaningful difference is the PoA of the initial voiceless plosive: /k/ vs. /p/).
² Conditioned by its environment or just varying freely like some kind of crazy loose cannon phoneme.
³ With /p/ being slightly less common (having been lost in Arabic and a bunch of other languages from around Northern Africa).
⁴ A number of languages (such as Welsh, Icelandic and Burmese) have voiceless nasals, but they’re uncommon cross-linguistically.
⁵ This is both a sweeping generalisation and a vast oversimplification, but at least gives us a starting point.
⁶ We could take /θ/ (the th in think) as the fricative corresponding to /t/, but since it’s a pretty rare sound cross-linguistically, I chose /s/ (probably the most common fricative) instead.
⁷ There are twenty-six consonants if you count the marginal phonemes /ʍ/, which only appears in certain dialects, and /x/, which optionally occurs in a small number of words. To stretch it further, you’d probably have to include interjections like uh-oh/ˈʌ̆ʔ˦əʊ˨/, tut-tut /ǀ ǀ/ or phew /ɧu˥˩/. Accepting consonants in interjections like these would also require you to say that English is tonal and has click consonants, which, though not entirely untrue, would be a misleading to say the least.
⁸ As in the English /t͡ʃ/ (the ch in chair), which developed from palatalised /k/. (In some dialects, the sequence /tj/ as in Tuesday also becomes /t͡ʃ/, while in American English the /j/ is usually dropped.)
⁹ Only about 2% of languages lack nasal consonants and, of these, most have nasal vowels (as in French bon /bɔ̃/), prenasalised plosives or both.
¹⁰ Not to be confused with hellish said by a French or Cockney person.
¹¹ The l in Calrissian is not necessarily [l]; in most dialects of English, an /l/ in the syllable coda is often realised as [ɫ] (a velarised lateral approximant), often called “dark l” (presumably due to its connection with the dark side of the force). /l/ in syllable codas may even lose its lateralisation completely, merging with /w/ or vocalising to [o] or [ʊ].
¹² We’ll get into syllables later, once the basics of our phonology are done.
¹³ We’ll get into vowels next post. Jeez, gimme a break.
¹⁴ Note for the table: /h/ is not technically a fricative, but it is traditionally grouped with the fricatives, so we’ll do so here.

* Opinions may vary as to the epicness of phonology-building.

2 thoughts on “Let’s Conlang – Part II “Let’s Get Phonetical”

  1. Pingback: Let’s Conlang – Part I “Crash Course Phonology” | The Brain Exchange Podcast

  2. Pingback: Let’s Conlang – Part III “Solemn Vowels” | The Brain Exchange Podcast

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s