An abjad is a writing system in which only consonants are normally written, is the opposite possible?

I've recently discovered that English actually has far more vowel-sounds than we have vowel characters (namely ~20) this is very similar to the number of consonant sounds (~24) so it seems to me that a writing system with only vowel should lose about as much information about word pronunciation as a writing system with only consonants.

obviously you would need more symbols for describing many different vowel-sounds rather than the English approach of ~ 5 vowel symbols

is there something i'm missing here which would prevent an anti-abjad from developing or is it viable?

  • 3
    The reason it works in Semitic languages is that the consonants carry the vast majority of the semantic content and some of the syntactic content, and the vowels mostly only contribute syntactic content (at least at the periods when the abjads developed — the silent aleph in Modern Hebrew, for example, means that the vowel attached to a given aleph now carries its meaning...). If a language were to use a reverse abjad, it would also need this relationship reversed. Sep 10, 2018 at 22:59

3 Answers 3


This conlang seems to fit the bill:

The Qohenje writing system is a "reverse abjad" (like the logical opposite of the Arabic or Hebrew writing systems, for example), with the dominant symbols being those showing vowels (+ tones), and with consonants shown by diacritics written above or below the vowel (see below). The vocalic core is considered the “main” part of the syllable (or more precisely the "glyph", see below), which accounts for the range and complexity of the Qohenje vowel inventory: there are 90 vowel symbols in Qohenje!

You might also be interested in vowel based abugidas:

Vowel-based abugidas

In a few abugidas, the vowels are basic, and the consonants secondary. If no consonant is written in Pahawh Hmong, it is understood to be /k/; consonants are written after the vowel they precede in speech. In Japanese Braille, the vowels but not the consonants have independent status, and it is the vowels which are modified when the consonant is y or w.



For a natural human language? Almost certainly not.

Just having a similar number of vowels and consonants does not mean that vowels and consonants carry the same amount of information. That depends on a many other factors, like relative frequency of use. In English, for example, even though we have a large number of vowel phonemes, mst f thm dn't crr mch nfrmtn. You can mutate or drop vowels quite freely in speech and in writing and still be understood. O e oe a, eai ou ooa ae i ioeeie. (On the other hand, leaving out consonants makes it incomprehensible.)

Of course, different languages put different functional loads on their vowel vs. consonant inventories. Some languages really do stuff a lot of information into vowels. And vowel-centered abugidas are certainly possible, with core characters representing vowel sounds and consonants added as diacritics. But, to the best of my knowledge, there are no natural languages where the vowel stream is actually more distinctive than the consonant stream, and all those natural language abugidas still require writing the consonants! (Which is why they are called abugidas, and not reverse-abjads, in the first place.) While this sort of thing is still an open area of research in psychoacoustics, cross-linguistically it appears that vowels are more important for carrying intonation contours that help identify phrase and sentence boundaries, while at the word level consonants tend to be more informative, largely because consonants are more reliably distinguishable than large inventories of vowels--and that intonation contour information that is uniquely suited to being carried by vowels is precisely the sort of information than tends to be left out of writing systems, or indicated with punctuation, anyway, so whether or not you write the vowels doesn't actually matter as far as that goes!

So it may be possible to design a functional, comprehensible language which puts the majority of its contrasting information in the vowel stream, such that consonants can be recovered by context, analogous to recovering vowels in the consonant stream of an abjad from discourse context... but it probably won't be practical for any natural human language.

  • The example is not very good because it uses the illogical and non-phonemic English orthography for the vowel symbols. Sep 13, 2018 at 20:24
  • @VladimirF It works equally well if you use phonemic transcription, but that would be less accessible to the average reader. Sep 14, 2018 at 15:51
  • There are some languages that can be whistled.
    – Joshua
    Jul 5, 2020 at 18:20

It is possible that by asking this question, you have increased the probability that someone will make omission of consonants be a feature of their next conlang. If you are looking for plausible future writing systems for natural languages, such a system would be unlikely. A writing system that omits useful information is less likely to develop compared to a system that includes that useful information. Typically, languages have more consonants than vowels, so if you are going to drop all consonants, or drop all vowels, you are more likely for messages to be difficult to interpret if you stop the consonants, compared to dropping the vowels. That is, the overall informational value of consonants is higher than that of vowels. The most likely outcome would be that a new writing system will be alphabetic.

It is possible that a new writing system would be invented based on an existing and culturally relevant writing system. For example, writing for Evenki in Russia is based on Cyrillic, and in China it is written in Latin and experimentally in Mongolian: it is not written in Arabic, because there would be no cultural motivation for that. To the extent that a new writing system is based on existing language writing systems (and none mark only the vowels), it is unlikely that a new writing system would mark only vowels. Alternatively, a new alphabet could be made up from scratch, as in the case of Vai writing. The most common quasi-modern ex-nihilo writing systems are syllabaries, suggesting that script-inventors prefer system that retain information about the pronunciation of words, rather than throwing it away.

Ai a oo eao o ie a ai-aa, e ae i a ea o oie o eai a e ii e, i i uie a a e a a ooa ou ee e aoe. Lacking a good reason to invent an anti-abjad, when faced with a wealth of choices for creating a new writing system, it is unlikely that a system that lacks consonants would ever be adopted. I predict that a new writing system invented by an illiterate person would be a syllabary. Of course, fictional systems can be really strange.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.