Reflections on Working with Mandinka Ajami Manuscripts

By Ousmane Cisse

I came across Mandinka Ajami thanks to my maternal grandfather Moussa Drame’s work; may his soul rest in peace. My grandfather was the Imam of Jumaa kutoo (Gnafoulene, Ziguinchor) in the 1980s. He passed away when I was a little boy and left a legacy behind. The legacy consists of a collection of written prayers from the teaching of the Prophet Mohammad (PBUH) along with some Quranic verses he used to read for spiritual healing and spell removal for people who sought his help. He also left behind a record of his transaction and some manuscripts in Mandinka Ajami. His personal items (clothes, shoes, and other valuable items) were distributed among my uncles, but my grandmother kept his work, including his manuscripts. As I grew up in the house, I was taught to read and memorize the Quran as well as some Hadiths in Arabic. Later, during the summer, my dad used to send me to a local school commonly known as a franco-Arabic school, where I improved my Quranic reading skills and acquired more knowledge of the Arabic language. These skills allowed me to read and write Mandinka Ajami.  The first Ajami scripts that I read were the papers my grandfather used to scribble the days of the week in Mandinka Ajami for his own planning. This memory of my grandfather triggered my interest in Mandinka Ajami. That is why I was so excited to join the Mandinka Ajami team when Prof. Ngom presented the project to me.

The linguistic and Arabic knowledge that I previously acquired have immensely helped me perform the cross-checking tasks. What helped me the most as a linguist are the morphological skills in particular, for there are instances where two words are merged together in the Ajami scripts; the knowledge of Mandinka morphology made it easy for me to draw the morphological boundaries of the words and separate them in the romanized transcript. Regarding the Arabic quotations in the Ajami manuscripts, had it not been my knowledge of Arabic and Quranic verses, I would not have been able to accurately transliterate the Quranic quotations. Although Arabic has significantly influenced the Mandinka language due to the Islamization of its people, it cannot be assumed that a native speaker of Mandinka can write or read Mandinka Ajami in Arabic script.

While cross-checking the transcriptions, I encountered quite a few speedbumps. The nature of these speedbumps pertains to the inaccuracy of the transcriptions, and the lack of clarity in the scripts (some of the original documents were not clear due to poor preservation).  The most recurrent speedbumps are (see for more details):

Alternation between two methods of recitation/writing in the manuscripts: Warsh and Hafs methods of writing.

e.g.: inconsistency between  Qaf / Faa

                     Warsh Qaf:     ڧ             Warsh Fa: ڢ

                     Hafs    Qaf:     ﻕ                Hafs    Fa: ﻒ

There are very noticeable impediments to accurate transcription, such as scribbled handwriting, unfamiliar fonts or writing styles, and poor conservation of the manuscripts that is caused by superimposition of two pages, allowing some scripts to reflect over other scripts.

Besides the challenges mentioned above, I also found the following complications affecting the quality and accuracy of transcription:

  • Dialectal variations (subtle regional differences in the varieties of Mandinka).
  • Interpolation of the reader’s comments in the transcription (as a result of the audio-based transcription instead of reference to the original text).
  • Inclusion or transfer of the reader’s mistakes /errors in the transcription.
  • The Ajami reader’s skipping some words or phrases, which is also reflected in the transcriptions.
  • Inclusion of Arabic phrases, and some loanwords (quotes from Quran and hadith, loanwords from French), which requires some background knowledge in those areas.
  • The transliteration of Arabic using Brill’s standards (EI3 style).
  • The representation of different vowels by similar diacritics (I call it diacritical ambiguity)

e.g.: /e/ and /i/   are written using Kasra        ب   ِ        بِ

/o/ and /u/ are written using Damma       ب ُ          بُ

  • Inconsistency in how some consonants/sounds are written (/p/ vs /b/, /ŋ/, /g/, etc., )

e.g.: /ŋ/  ݞ    by writer #1                   /ŋ/  ﻉ by writer #2

   feŋolu ‘things, clothes’                                     fulanjaŋo ‘the second one’

Given these challenges, I recommend the following suggestions for improvement:

For preservation of the manuscripts: give away document folders/binders with sleeves to Ajami writers for a better preservation of their manuscripts (initiative).

For transcription accuracy: Organize a workshop on transliteration using BRILL’s EI3 style, provide training on transcription, translation and raise awareness of the recurrent challenges constituting some obstacle to transcription.

For consistency in the writing methods: promote the widely used method of writing (either Warsh or Hafs) (although this effort could take decades to be successful) or at least encourage writers to use one consistently throughout a manuscript.

For diacritical ambiguity: Come up with a conventional way to represent each vowel by one diacritical sign so that there will not be any overlap or confusion: for instance, because of this diacritical ambiguity one needs to exhaust the different readings and decide which meaning fits the context (for example, fuloo ‘the two’ vs fuloo ‘Fulaani’ vs foloo ‘the first’; keloo ‘a fight’ vs kiloo ‘an egg’, and kiloo ‘kilogram’, etc.).

Moreover, it will also be very helpful to strictly follow the guidelines below that include the previous ones for transcription and cross-checking the Ajami texts:

  1. Read the metadata of the manuscript to familiarize oneself with the text.
  2. Check the region the manuscript was collected from and check what dialect is dominant in that area.
  3. It would be helpful during the fieldwork to include the information about the dialect in which the manuscript is written.
  4. Be mindful of the interpolations (addition of Ajami text reader’s comments).
  5. Use footnotes to add personal comments on the transcribed document.
  6. Have a separate Excel sheet in order to list the challenges and their in­-text references.
  7. Highlight unknown and undecipherable words in red.
  8. Highlight unclear words in green after guessing their meaning.
  9. Make sure the transcription’s writing style (whether it is written with rhymes or prose) aligns with that of original text.
  10. Follow the transcription writing convention (fonts type, font size, space, etc.).
  11. Keep track of time for the purposes of payment.

Overall, cross-checking the transcription of Mandinka Ajami manuscripts was a rewarding experience. As a reader, I realized that it is not straightforward and smooth to switch from one writer to another. Reading Mandinka Ajami could be effortless if all Mandinka Ajami writers followed a conventional system of writing Mandinka Ajami (same punctuation, same alphabet along with one diacritical sign per vowel). The writing could be simplified and organized in order to faithfully convey the meaning encoded in the manuscripts. Having this conventional writing system in place could encourage more people to write more about health, philosophy, medicine, teaching, tales, etc. Thus, not only will the manuscripts be more accessible to people, but they will be more engaged and will interact more with the text.

One big takeaway while cross-checking the transcription is that I learned a lot about Mandinka poetry and the importance of religion in Mandinka culture. While religion is at the heart of Mandinka society, poetry on the other hand tends to be used in the chants to illuminate one’s soul while praising the prophet Mohammed (PBUH). The religious chants are reminiscent of my childhood when I used to spend sleepless nights surrounded by friends, family members, and neighbors relentlessly chanting until dawn. I appreciated the fact someone took the time to write down these chants because I almost forgot some of them. Although I enjoyed reading most of them, there are some that I found difficult to follow due to the inclusion of unfamiliar words and style of writing. In contrast, cross-checking letters of correspondence, records of historical events, and proverbs were very easy to follow and fun to read. Another takeaway for me is the overall skill that I acquired in the process: the practice of transliteration, the familiarity with the different writing styles in Mandinka, and Mandinka poetry in general.

To conclude, I would like to express some words of gratitude to Fallou Ngom. Had it not been Prof. Ngom’s guidance and insightful comments, working in the Mandinka Ajami project would not have been easy. I also would like to thank Bala Saho and Ablaye Diakite for paving the way and laying down the foundation so that I was able to do the cross-checking. Special thanks go to Dr. Daivi for her hard work, perseverance and patience. On an appreciative note, I also would like to thank the IT team for their invaluable and creative work. I wish to end this reflection by thanking everyone who is closely or remotely involved in the Ajami project.

Al niŋ baara !

Ousmane Cisse graduated from Université Gaston Berger de Saint-Louis with a Master’s degree in Linguistics (Phonology/Sociolinguistics) and was granted a Fulbright scholarship by the US State Government to pursue his studies in the United States. He received another Master’s degree in Linguistics (Syntax/Phonology) from Eastern Michigan University, and is currently pursuing his Doctoral studies at Boston University in the Linguistics Department.