Course available: English Grammar: The Missing Course

Courses
The complete guide to conlang vowel inventories

There are patterns that exist from language to language in how vowel inventories are organized. This is important for naturalistic conlangers because a language whose vowel inventory breaks these patterns may end up looking unnatural. In this article I will outline the most common patterns in the structure and organization of vowel inventories to help you design naturalistic vowel systems.

Number of vowel phonemes

Let's look at the vowel space in the table below:

Front Unrounded Front Rounded Central Unrounded Central Rounded Back Unrounded Back Rounded
High i y ɨ ʉ ɯ u
ɪ ʏ ʊ
Mid e ø ə ɤ o
ɛ ɶ ʌ ɔ
Low æ a ɑ ɒ
Common vowels, in IPA notation.

Every different position on this chart represents a different vowel quality, from [i] in the top left corner all the way to [a] in the bottom right. As you can see, there are a lot of potential vowel qualities. But no language distinguishes between all of them. Instead, languages pick a subset of vowel qualities which they use as phonemes. And, in fact, the number of contrastive vowel qualities a language has is often quite low. Let's take a look at a few examples of relatively common inventories. You can take any of these as a starting point when developing your phonology.

5-vowel inventories

The most common size for a vowel inventory is 5. Spanish is a good example of a 5-vowel system, with the vowels /a e i o u/, as shown in the following table.

Front (Unrounded) Central (Unrounded) Back (Rounded)
High i u
Mid e o
Low a
A typical 5-vowel system, e.g. Spanish (Indo-European, Romance).1

When inventories are small, as the Spanish inventory is, it is very common for all the front and central vowels to be unrounded and for all the back vowels to be rounded. We will return to this topic below.

3-vowel inventories

Vowel systems do come in smaller sizes as well, although the vast majority of languages has at least 3 vowel phonemes. Kalaallisut is a good example of a 3-vowel system, contrasting the vowels /i a u/, as shown in the following table.

Front (Unrounded) Central (Unrounded) Back (Rounded)
High i u
Mid
Low a
A typical 3-vowel system, e.g. Kalaallisut (Eskimo-Aleut, Inuit).2

9-vowel inventories

On the high end, it's uncommon for a language to distinguish more than 9 vowel qualities. A good example of a 9-vowel system is Lika, which contrasts the vowels /i ɪ e ɛ ə a ɔ o ʊ u/, as shown in the following table.

Front (Unrounded) Central (Unrounded) Back (Rounded)
High i u
ɪ ʊ
Mid e o
ɛ ɔ
Low a
A typical 9-vowel system, e.g. Lika (Atlantic-Congo, Bantu).3

There are languages which go above 9 – in fact, the variety of English I speak distinguishes 14 vowels and diphthongs.4 You can see the vowel inventory for Canadian English5 in the following table.

Front (Unrounded) Central (Unrounded) Back (Unrounded) Back (Rounded)
High i u
ɪ ʊ
Mid (eɪ) ə (oʊ)
ɛ ʌ (ɔɪ)
Low æ (aɪ) (aʊ) ɑ
The vowel inventory of the author's variety of Canadian English, including diphthongs in parentheses.

I recommend you stay within the 3–9 vowel range if you're early in your conlanging career. This will help to keep things manageable.

Symmetrical and asymmetrical inventories

The most basic vowels that almost every language has are /i a u/, i.e. the inventory of Kalaallisut shown above. Notice how they're roughly evenly spaced in the chart. That's no accident. Vowel inventories seem to space themselves out as much as possible, as if to maximize contrast.

If a language contrasts more than 5 vowels, that language is 100% certain to have /i a u/. Occasionally, with smaller vowel inventories, one of these three will be missing.

We've seen that vowel phonemes like to distribute themselves evenly in the vowel space. Let's call that characteristic of vowel systems dispersion. Another characteristic of vowel inventories is symmetry, specifically symmetry between front vowels and back vowels. As more vowel qualities get added, we tend to see them evenly distributed between front and back vowels, with the exception of the low vowels, which are often left with a single unpaired /a/.

We can see this principle at work in the inventory of Holoholo (see the table below), which contrasts /i ɪ e/ in the front with /u ʊ o/ in the back. And here too, /a/ stands alone.

Front (Unrounded) Central (Unrounded) Back (Rounded)
High i u
ɪ ʊ
Mid e o
Low a
The symmetrical vowel system of Holoholo (Atlantic-Congo, Bantu).6

When there is asymmetry between front and back vowels, it's usually in favour of the front vowels. That is, it's more common to have an extra unpaired front vowel than an extra unpaired back vowel. An example of a front-weighted asymmetric system is Ukrainian. Ukrainian has a 6-vowel system, with three front vowels /i ɪ ɛ/ and two back vowels /u ɔ/, along with unpaired /a/.

Front (Unrounded) Central (Unrounded) Back (Rounded)
High i u
ɪ
Mid
ɛ ɔ
Low a
The front-weighted asymmetrical vowel system of Ukrainian (Indo-European, Slavic).7

Non-peripheral vowels

As the number of vowels in an inventory increases, vowel systems start to include non-peripheral vowels, vowels which occupy the centre of the space.8 The non-peripheral vowels include, most commonly, schwa /ə/9 and barred I /ɨ/.10 But they also include the front rounded vowels and the back unrounded vowels. More on this in a minute!

Let's take a look at how the addition of a central vowel works. Malay is a good example of this. Malay is a six-vowel system which contains the same five vowels as Spanish, in addition to a schwa in the middle of the vowel space: /i e a ə o u/, as shown below.

Front (Unrounded) Central (Unrounded) Back (Rounded)
High i u
Mid e ə o
Low a
The 6-vowel system of Standard Malay (Austronesian, Malayo-Polynesian).11

Another type of non-peripheral vowel is front rounded vowels. The reason they're considered non-peripheral arises from their acoustic properties, which is a bit beyond the scope of our discussion. Suffice it to say that front rounded vowels, like /y/ and /ø/, tend only to occur when there is a corresponding front unrounded vowel in the system already. So you don't see /y/ without /i/.

Finnish is a good example of a system which makes use of front rounded vowels, as shown in the table below. The Finnish inventory has five front vowels, three unrounded /i e æ/ and two rounded /y ø/, along with its three back vowels /u o ɑ/. Here we also see an exception to the rule that low vowels are often unpaired.

Front (Unrounded) Front (Rounded) Central Back (Rounded)
High i y u
Mid e ø o
Low æ ɑ
Finnish (Finno-Ugric), a system with front rounded vowels.12

Another kind of non-peripheral vowels are the back unrounded vowels. These are like the front rounded vowels, but in reverse. Because back vowels are rounded by default, unrounding them makes them non-peripheral. Although back unrounded vowels do sometimes occur without the corresponding back rounded vowels, it's more common to see systems like that of Lao (see the table below), where we have the back unrounded vowels /ɯ ɤ/ in addition to the back rounded vowels /u o ɔ/.

Front (Unrounded) Central (Unrounded) Back (Unrounded) Back (Rounded)
High i ɯ u
Mid e ɤ o
ɛ ɔ
Low a
Lao (Tai-Kadai, Lao-Thai), a system with back unrounded vowels.13

Vowel quantity

Apart from vowel quality, there is another important dimension we have not yet considered: vowel quantity. Vowel quantity refers to distinctions in the length of the vowel, literally how long the vowel is held for. Many languages have what is called phonemic vowel length, which means that you can distinguish two words just by holding the vowel longer.

One example of language with phonemic vowel length is Classical Latin. The word malum, with a short /a/, means 'evil'. The word mālum, with a long /aː/,14 means `apple'. Quite important to keep separate!15

When a language has a length contrast, it can double the number of vowels in the inventory. This is a good way to increase the number of possible syllables you can make without increasing the number of vowel qualities you need to keep separate.16

There are other ways in which languages create contrasts in vowel systems, including nasalization, as well as different things you can do with the vocal folds, like creaky voice,17, but I'll save those topics for another article.

Conclusion

Establishing a language's vowel inventory is one of the earliest and most important decisions you will make in creating a conlang. If you're looking for a naturalistic vowel system, free to use any of the systems described in this article as a starting point, or use the principles of symmetry, dispersion, and peripherality to create your own.


References

  • Clynes, Adrian & David Deterding (2011). Illustrations of the IPA: Standard Malay (Brunei). Journal of the International Phonetic Association 41: 259–268. Cambridge University Press.
  • Coupez, A. (1955). Esquisse de la langue holoholo. Tervuren.
  • Harris, James W. (1969). Spanish Phonology. MIT Press.
  • Kutsch Lojenga, Constance (2008). Nine vowels and ATR vowel harmony in Lika, a Bantu language in DR Congo. Africana Linguistica 14: 63–84.
  • Moran, Steven & Daniel McCloy (2019). PHOIBLE 2.0. Max Planck Institute for the Science of Human History. https://phoible.org
  • Morev, L. N., A. A. Moskalev, & Y. Y. Plam (1979). The Lao Language. Nauka.
  • Pugh, Stefan M. & Ian Press (1999). Ukrainian: A Comprehensive Grammar. Routledge.
  • Rischel, Jørgen (1974). Topics in West Greenlandic Phonology. Copenhagen.
  • Schwartz, Jean-Luc, Louis-Jean Boë, Nathalie Vallée, & Christian Abry (1997). Major trends in vowel system inventories. Journal of Phonetics 25: 233–253.
  • Suomi, Kari, Juhani Toivanen & Riikka Ylitalo (2010). Finnish sound structure: Phonetics, phonology, phonotactics and prosody. Puhe ja kieli 30: 121–124.

  1. Harris (1969), as cited in Moran and McCloy (2019). The notation "Spanish (Indo-European, Romance)" indicates the name of the language, followed by the names of the language family, and notable branch of that family, to which the language belongs. You can read this as "Spanish is part of the Romance branch of the Indo-European family.
  2. Rischel (1974), as cited in Moran & McCloy (2019).
  3. Kutsch Lojenga (2008), as cited in Moran & McCloy (2019).
  4. The exact count of English vowel phonemes depends on how you count them, but it's a large number no matter how you slice it.
  5. The inventory for Standard Canadian English is the same as many varieties of American English, in particular those varieties that participate in the "cot-caught" merger, whereby the contrast present in other dialects between the vowels in cot and caught are neutralized to something like /ɑ/.
  6. Coupez (1955), as cited in Moran and McCloy (2019).
  7. Pugh & Press (1999), as cited in Moran & McCloy (2019).
  8. See Schwartz et al. (1997) for extensive discussion of the distinction between peripheral vs. non-peripheral vowels.
  9. Schwa /ə/ is the mid-central unrounded vowel.
  10. Barred I /ɨ/ is the high central unrounded vowel.
  11. Clynes & Deterding (2001), as cited in Moran & McCloy (2019).
  12. Suomi et al. (2010), as cited in Moran & McCloy (2019).
  13. Morev et al. (1979), as cited in Moran & McCloy (2019).
  14. Vowel length is marked in IPA using the triangular colon symbol [ː].
  15. Incidentally, this shows how the tradition of the forbidden fruit in the Garden of Eden being an apple may have arisen from a pun.
  16. When a language has contrastive vowel length, it is not always the case that every short vowel has a long counterpart, and vice versa. In some cases, a particular vowel quantity may appear only as a short vowel or only as a long vowel.
  17. Creaky voice is known more popularly outside of linguistics as "vocal fry".

Subscribe to my conlang newsletter

Updates on articles, resources, and exclusive course discounts.