Adapting the Arabic script to Swahili phonology

Starting from the later half of the 19th century, continuing into the 20th century, and going on in the 21st century, a process of "Swahilization" of the Arabic script has been under way by Swahili scribes and scholars. The first of such attempts was done by Mwalimu Sikujua, a scholar and poet from Mombasa. Below are Mwalimu Sikujua's innovations in the adaptation of Arabic script for Swahili phonology:^[4]

1. In his script, he distinguished alveolar [t] and [d] sounds from their dental counterparts which is the norm in Arabic phonology. He did so by adding 4 dots to the letters tāʼ (ت) and dāl (د), creating the letters ٿ and ڐ. This is the same as what was done in Urdu alphabet to denote retroflex consonants, where later, the 4 dots evolved into looking like the letter ṭāʾ (ط)
2. The introduction of new letters to represent the sounds [p] and [v] by adding three dots to letters bāʼ (ب) and fāʼ (ف), same as what's been done in Persian alphabet,^[5] creating the letters پ and ڤ.
3. For representing aspirated consonants, Mwalimu Sikujua wrote a tiny letter hāʾ (ھ) on top of the aspirated letter. For example, an aspirated alveolar [t] would be written with ه on top of the four dots of the letter ٿ, producing "ٿ(ه)".
4. For representing prenasalized consonants (sounds such as [ⁿd], [ⁿɡ], and [ᵐb]), Mwalimu Sikujua wrote a tiny letter mīm (م) or nūn (ن) on top or bottom of the letter. For example, the sound [ᵐb] would be written as "بۭ" and the sound [ⁿd] as "دۨ".
5. As explained, Arabic only has vowels for [a], [i], and [u], whereas Swahili has five vowels, the three aforementioned ones and [e] and [o]. For showing the vowel [e], Mwalimu Sikujua modified the existing Kasrah diacritic used for showing [i] (◌ِ), by changing its angle slightly and by adding markings to modify its shape slightly, thus producing a shape resembling "◌̼ ". As for [o], Mwalimu Sikujua inverted the orientation of the existing Ḍammah diacritic used for showing u (◌ؙ), thus representing it with a diacritic resembling "◌ٗ".

The poetry and texts written by Mwalimu Sikujua were published by W.E. Taylor, a Swahili literature scholar of British origin, in the late 19th century. This was probably the first time ever that Swahili in Ajami script was printed and published as opposed to being handwritten. However, the spread of a standardized indigenous variation of Arabic script for Swahili was hampered by the colonial takeover of East Africa by the United Kingdom and Germany. The usage of Arabic script was suppressed in German East Africa and to a lesser extent in British East Africa. Nevertheless, well into the 1930s and 1940s, rural literacy rate in Arabic script as well as a local preference to write Swahili in this script was high.^[4] But it's important to note that literacy at the time was in unmodified Arabic script, and not the modified scripts that were thus far proposed by the likes of Mwalimu Sikujua.^[4]

In the mid 1940s, 3 prominent Swahili literature scholars of British origin noted on the writing of Swahili in the urban centre of Mombasa. They noted that at the time, a standardized reformed Arabic script had been adopted by writers of the city. While they only made use of the three original vowels, they had consistency in indicating the stressed syllable by writing Arabic long vowels using alif, wāw, and yāʼ (ا, و , ی). Colonial administrators as well as prominent Swahili scholars, despite recognizing the need for implementation of reform in the script, citing local opposition and conservativity, were anxious to do so and impose it on the populace in a top-down manner.^[4]

Competing standards of the Swahili language

The process of standardization of Swahili Ajami script continued into the 20th century. However, scholars didn't always follow or propose the same standard. For example, Hajj Nur b. Muhammad Hajj Nur, being from Barawa in Somalia, worked based on the local Chimbalazi dialect of Swahili. In his proposal, he maintained the use of tiny letters for representing prenasalized consonants just as Mwalimu Sikujua had done. Whereas, Mu’allim Sheikh Yahya Ali Omar from Mombasa decided to do what is done in the Latin script, namely to just write the letters mīm (م) or nūn (ن) as part of the word. For example, the number two (mbili) will be written as مبِيلِ‎ and بِۘيلِ‎ respectively in either standard.^[1][4]

Furthermore, on top of orthographic variations, it is noteworthy to mention the dialectical variation within Swahili. For example, traditionally, dialects from Lamu (on the Kenyan Coast) have been the basis for Swahili literary heritage, and dominates Swahili poetry.^[6] This dialect was the basis for Swahili Ajami script. But this came to be challenged by Kiunguja dialect from Zanzibar City. This dialect was by Christian missionaries and British colonial officials to be the basis of the Latin Script.^[1]

In the most recent and most widely acknowledged orthographic standard, devised by Mu’allim Sheikh Yahya Ali Omar, the dialect of his hometown Mombasa has been chosen as the basis. This is, according to Yahya Ali Omar himself, because this dialect has historically been affected by all vernacular varieties of Swahili and it has formed the basis of literary Swahili. This dialect is in his opinion, best fitted for accurate Swahili prose.^[4]

Letters

Letters in Yahya Ali Omar's orthography are based on the dialect of Mombasa. There are several consonants that represent sounds present in Mombasan dialect but not necessarily elsewhere in the Swahili-speaking world. There are 48 consonants in Swahili Ajami. Digraphs are counted as separate letters.^[4]

Although the Roman orthography does not distinguish between syllabicity and prenasalized sounds, both Sheikh Yahya's manuscripts and Yahya Omar's convention make a distinction between a syllabic nasal followed by a voiced plosive (e.g. [m̩ɓ]) and a prenasalised voiced plosive (e.g. [ᵐb]). Both of these are written as mb in Roman orthography. The former is seen in Swahili noun class 1 (the M-wa class), and the first letter of the word is written as mīm (م), such as mbrazil مْبرَزِل‎ (Brazilian person). The later is seen in Swahili noun class 9. For this class, in Roman script, either a prefix of m or n is used, reflecting pronunciation. In Ajami script, nūn (ن) is used all the time, reflecting grammatical consistency. An example being mbazi نْبَازِ‎ (beans).^[4][2]

Aspirated as opposed to non-aspirated consonants are also marked in Swahili Ajami, with a "two-eyed" hāʾ (ھ) similar to what's done in the Urdu alphabet. This is not indicated in Swahili Roman script.

Dental as opposed to alveolar [t] and [d] consonants, sounds unique to Mombasa Swahili, are also marked in Swahili Ajami. The dental forms are represented with Arabic letters tāʼ (ت) for [t̪] and dāl (د) for [d̪], whereas the alveolar ones are represented with unique letters, similar in shape to Urdu letters Ṭe (ٹ) for [t] and Ḍal (ڈ‎) for [d]. Dental versus alveolar [t] and [d] are not distinguished in Swahili Roman orthography, nor in unmodified Arabic script.

Unfortunately, online script conversion tools are unable to distinguish between the above nuances unique to Ajami script, and they render all text as they'd correspond to Roman. Manual editing of text will be required.^[2]

General conventions

^[4][2]

However, these diacritics cannot be written on their own and independent of a letter. When a word begins with a vowel, a letter alifu is used as the carrier, with a hamza on the top (for [a], [u], [o]) or on the bottom (for [e], [i]). The writing of Hamza in handwriting is optional.^[4][2]

In Swahili Ajami, vowels in the middle of the word are shown differently depending on whether the syllable is stressed on unstressed. In Swahili, stress patterns are predictable, as almost always they fall on the second-to-last syllable of a word. The exceptions to this rule are extremely rare, and are usually found in words borrowed from other languages, mostly Arabic (for example, maalum).

Vowels in stressed (second-to-last) syllable of the word are marked with diacritic as well as a carrier letter, namely alifu (ا) for vowel [a], yee (ي) for vowels [e] and [i], and waw (و) for vowels [o] and [u]. The diacritic for the vowels [a], [i], or [u] can be omitted, but this is not recommended. This practice of indicating the stressed syllable also helps to delimit individual words in the Ajami script.

Below are some sample words where the above principles can be observed.

Vowel sequences

In Swahili Ajami script, to denote vowel sequences, hamza and either alifu (ا), yee (ي), or waw (و) are used. There are different conventions depending on whether the vowel sequence occurs in the middle of the word or at the end of the word.

If the vowel sequence occurs in the middle of the word:

• if the second vowel of the sequence is [a]
  • if the first vowel of the sequence is also [a], an alifu-hamza letter (أ) is used and the [a] diacritic (◌َ) is placed on it.
  • if the first vowel of the sequence is [e] or [i], a yee-hamza letter (ئ) is used and the [a] diacritic (◌َ) is placed on it.
  • if the first vowel of the sequence is [o] or [u], a waw-hamza letter (ؤ) is used and the [a] diacritic (◌َ) is placed on it.
• if the second vowel of the sequence is [e] or [i], a yee-hamza letter (ئ) is used and the [e] or [i] dacritic, (◌ٖ) or (◌ِ), is placed on it.
• if the second vowel of the sequence is [o] or [u], a waw-hamza letter (ؤ) is used and the [o] or [u] dacritic, (◌ٗ) or (◌ُ), is placed on it.

And in addition to above conventions, if the second syllable in the vowel sequence is the stressed (second-to-last) syllable of the word, alifu (ا), yee (ي), or waw (و) are also written in correspondence with the vowel of the syllable.^[2]

Below are some sample words where the above principles can be observed.

If the vowel sequence occurs at the end of the word, meaning that the first syllable in the sequence is the stressed syllable of the word:

• if the first vowel of the sequence is [a], an alifu letter marking the stressed syllable followed by a hamza (اء) is used and the hamza is marked with one of the five diacritics.
• if the first vowel of the sequence is [e] or [i], a yee-hamza letter (ئ) is used and is marked with one of the five diacritics. There no longer is a need to add a yee letter (ي) to mark the stressed syllable; an example being كُپٗكٖئَ‎ (kupokea, to receive).
• if the first vowel of the sequence is [o] or [u], a waw-hamza letter (ؤ) is used and is marked with one of the five diacritics. There no longer is a need to add a waw letter (و) to mark the stressed syllable; an example being كُپٗؤَ‎ (kupoa, to cool).

Below are some sample words where the above principles can be observed.