Comparison with the IPA

On this page, I'm going to explain why Musa is not only a great practical script for any natural language, but is also a great phonetic alphabet: much better than the International Phonetic Alphabet (IPA).

If you think about it, that shouldn't be surprising. Musa essentially is a phonetic alphabet, but one that was designed from the ground up. The IPA, on the other hand, was an adaptation of the Roman alphabet used by all of its founding members. One of the guiding principles was to use Roman letters as much as possible - and why shouldn't they? But as we also see with Unicode and many English spelling reform proposals, the legacy compromises made in the name of easing adoption often reappear as weaknesses later on.


The IPA is an extension of the Roman alphabet to cover the sounds of all the world's languages. The current version consists of 79 consonants and 28 vowels, for a total of 107 letters. It also includes 32 diacritics, 9 suprasegmental marks, and 24 tone marks in two systems. That's a total of 172 signs, 173 counting the space.

Musa, on the other hand, includes only 22 vowels, 6 accents for tone, and 24 punctuation symbols, including the space. However, it offers many more consonants - 138 so far - for a total of almost 200 signs, a few more than the IPA.

The IPA has about half of the consonants that Musa has. What's missing? Well, letters for affricates, aspirates, breathy and murmured sounds, labial-velar sounds, voiceless laterals, and ejectives, to mention broad categories. Those are all reasonably common in the world's languages, and the diacritics used to compensate for their absence make IPA look hieroglyphic. They also violate the IPA's own principles.

Of course, the IPA can write all the sounds you need, but the result looks like a chemical formula. For example, the IPA for an alveolar lateral ejective affricate, as occurs in many American languages, is tɬʼ; the Musa is just . The IPA for a voiceless lateral velar fricative is ʟ̝̊; the Musa is . Musa is just more efficient, and much prettier!

On the other hand, Musa has slightly fewer vowels. Musa doesn't consider central vowels to be distinguished from front rounded or back spread vowels, and the acoustical data supports that analysis. Musa doesn't make the fine distinctions that the IPA can make - has to make - between ɘ ɤ or ɜ ʌ, for example. Is there a language that distinguishes either of these pairs as phonemes? I'm skeptical - they're hard to distinguish in isolation!

Some of the IPA letters are featural. For example, the letters for retroflex sounds all feature a right-swinging tail: ʈ ɖ ɳ ɽ ʂ ʐ ɻ ɭ. The central nasals are all variants of n: ɳ ȵ ɲ ŋ. But many of the IPA letters are simply "turned" versions of other letters, since it was easy to reuse the slugs of 19th century typesetting by simply turning them over. In some cases, the turned letters denote variants, like ɹ ʍ ə ɒ ɥ. But in others, like ɯ ʌ ɔ, the turned versions have nothing to do with the originals.

In contrast, Musa is completely featural: all the sounds articulated in the same position share the same bottom, and all the sounds articulated in the same manner share the same top.

The IPA uses the Roman alphabet, which should be an advantage for those of us who use it for our own languages. However, this advantage is reduced or even reversed if the letters don't stand for the sounds we expect. For example, IPA j q x and many vowels don't stand for the same sounds they do in English, but at least there are some languages that spell those sounds that way.

But look at the letter c. It stands for a k or s sound in English, French, Catalan, Portuguese, and Italian. In German, Polish, Hungarian, Croatian, Albanian and even Chinese pinyin, it stands for ts. It has a few other values: th sh ch dj, even a click or a glottal stop. But in the IPA, it stands for the voiceless palatal plosive, the sound written ty in Hungarian. Who writes that with a c, the Dinka?

In many cases, it's actually an advantage of Musa that is doesn't use the same alphabet as the languages it's trying to describe. In one of John Wells' books, he tells the anecdote of slipping the trap word /hæpy/ into an IPA reading exercise for his graduate students: in IPA, the y letter stands for a close front rounded vowel, but almost all of them read through it as English happy.

The IPA has diacritics to mark tone, but as with letters, their use doesn't match other romanizations. For example, the name (in standard Chinese) of chess grandmaster Dīng Lìrén 丁立人 is spelled tíŋ lîɻə̌n in IPA. Here are those two romanizations again, large enough for you to see the tone marks clearly:

Dīng Lìrén   tíŋ lîɻə̌n

In contrast, Musa writes stress with vowel height, tone with simple accents, and intonation with punctuation. The Musa notation for prosody is so natural that it's not omitted (a big problem in African languages, for instance). Think how often you see IPA without stress and break marks.

Musa offers a number of smaller advantages:

This website offers you a couple of tools to help you switch. There's an IPA Transcriber, with which you can convert IPA to Musa. And there's a page of IPA charts showing the correspondences.


Ironically, despite the scope of the IPA, linguists working with indigenous languages of the Americas, Caucasus, and India; Slavic, and Semitic languages often use a different alphabet, the APA Americanist Phonetic Alphabet (there was also a SPA Slavistic Phonetic Alphabet). Linguists working on Uralic languages often use the UPA Uralic Phonetic Alphabet, while those working on Afro-Asiatic languages use Afrasianist Phonetic Notation. The existence of all these different phonetic alphabets is testimony to the shortcomings of the IPA.

Take a look at the image below. Which notation do you prefer?

The IPA, like Unicode, solves a real problem, and we're all better off for that. Unfortunately, it doesn't solve it very well, and that's a shame. The creators of the IPA were competent and smart, and they worked hard with the best intentions, but it's like the old joke that a camel is a horse designed by a committee: the IPA isn't faithful to its own guiding design principles, it tries to please all constituencies, it gives too much respect to legacy solutions, and it ended up as a patchwork of ad hoc fixes. At least the IPA, unlike Unicode, makes changes every once in a while.


< Questions Principles >


© 2002-2022 The Musa Academy musa@musa.bet 13jan22