The West African Arabic Manuscript Database currently includes descriptive records of 20,000 Arabic manuscripts from West Africa. The database and search interface are bilingual. Comments, queries, and expressions of interest in adding data sets to this one are encouraged: please contact Prof. Charles Stewart.
return to search pageThe first version of this database (called “AMMS”) was created in 1987 as a finding aid for an Arabic manuscript microfilm project that preserved over 100,000 folios of material from the private library of Haroun o/ Sidiyya in Boutilimit, Mauritania. Our object then was to produce a bilingual hardcopy finding aid for that collection which consisted of diverse types of material from letters and notes to local histories and classical treatises in the Islamic sciences. Our goal was a simple and quick computer-generated entry system using un-transliterated Latin letters alongside Arabic entries that could be equally accessible to readers (and input specialists) using either Arabic or English. Our end product was to be a bilingual catalogue with indices that would be user-friendly in both languages. The original AMMS program was written using an early ARABDOS software to create 31 possible fields for entries about each manuscript and with an indexing capability to cross-reference and locate up to three fields in either language. Two years later the same software was employed to in-put a finding aid and generate indices for the Mauritanian national manuscript collection at the Institut Mauritanien de Recherche Scientifique. The possibility of expanding the number of entries to include other West African collections prompted a second version of AMMS, on the same platform, with the capability of merging files into a single database. Subsequently, in the early 1990s other published catalogues and hand-lists from West African collections housed in Niger, Paris, Timbuctu, and Evanston, Illinois were entered in the database. Taken together, over 19,000 records from these six collections were recorded in the AMMS vers.2 database, possibly a majority of the then-extant titles for the West African Sahel (excepting correspondence).
The research potential of a union index of authors, nicknames, titles and subject matter in these collections of West Africa's Arabic literary heritage, with the capability of expansion as other collections are uncovered, became obvious. AMMS provided us with a mechanism to re-unite a literary tradition represented by tens of thousands of Arabic documents across the West African Sahel that has been largely unknown beyond the work of a small band of local scholars and an even smaller cohort of Western-trained Arabists. Even where these materials were accessible to researchers in public repositories, the importance of this literary tradition has been well masked by the disparate systems used to record it and the dispersal of individual collections in Africa, Europe and North America.
This project seeks to bring together, in a single database, a sizeable cross-section of these Arabic materials, despite sometimes imperfect annotation and documentation, to provide users with an index to roughly two hundred years of Sahelian literary activity. It was in the early 1990s that the work of editing over 19,000 entries for consistent orthography and subject identification foundered on the magnitude of that task, an increasingly fragile software platform, and difficulties in imagining how such an unwieldy end product might be disseminated. One positive result of my inattention to the project for nearly a decade is that these problems have now been largely resolved, thanks to advances in computer technology. In summer 2002 all of the 19,778 records were ported onto a Windows platform, the screen was redesigned, and, most significant, a search engine was created that overcame many of the previous difficulties that had arisen from the diversity of input parameters. The present internet version of the database has been rebuilt yet a third time to use a search engine accessible on the WEB. With this internet version, we have finally reunited a representative sweep West African Arabic writing and scholarship from pre-colonial times. During the decade this project was on hold, new finds of manuscripts in private libraries in Mauritania and Mali continued apace, and the numbers of additional manuscripts now catalogued from ‘new' collections may well eclipse the number of initial entries in AMMSvers2; we welcome the addition of these new entries into the current data base as a resource for future generations of scholarship.
return to search pageThe subject headings below that identify the AMMS.3 entries were developed from the content of individual manuscripts identified by catalogers of manuscripts cited in AMMS rather than an externally-imposed set of classifications. The rubrics represent an editing of those subject entries, their consolidation and standardization, but occasionally the same or similar material may appear under more than one general rubric, following the best judgment of different catalogers.
Cross-references (see also: …) below, draw attention to the most common of these overlapping references; references that appear in italics as in (see Belief: theology), indicate the rubrics under which that material can be located.
The search engine is sensitive to individual words, irrespective of their placement in sub-categories in this listing. For example, a subject search for the word “oaths” will identify manuscripts in two different secondary headings under Jurisprudence as well as under Politics; a search for “conundrums” will identify records under Arabic language, Jurisprudence, Literature, and Qur'an; “genealogy” appears under three different rubrics; “slaves” and “captives” appear under four different rubrics.
The subject list below within these rubrics is not a comprehensive listing. A more thorough description of the subject headings is available as a rich text text document.
The principle rubrics in AMMS.3 (with numbers of records, effective 09/30/03) are:
Arabic language (1258)
Belief (1936)
Biography (213)
Conduct (105)
Devotional (1632)
Economy (554)
Education (174)
Ethics (424)
Geography (20)
Hadith (516)
History (488)
Jurisprudence (3934)
Literature (1841)
Logic (107)
Medicine (99)
Politics (572)
Prophet Muhammad (480)
Qur'an (854)
Reform (44)
Science (231)
Sciences, Esoteric (455)
Social matters (159)
Sufism (731)
The manuscripts that make up the Boutilimit collection come from the private library amassed by Haroun ould Sidiyya Baba (1917-1978) who had spent the last thirty years of his life reconstituting the library and letters of his great-grandfather, the Moorish savant Shaykh Sidiyya (d.1868) as well as his wider family's literary record (from Sidiyya's mentors in Timbuctu (Sidi al-Mukhtar al-Kunti and his son Sidi Muhammad), Sidiyya's son, Sidi Muhammad (d.1869), and his son, Sidiyya Baba (d.1926).
At his death Haroun left over 100,000 folios of manuscript material that was microfilmed and catalogued in 1987/8 (the beginning of the AMMS project). The description of that process can be found in the introduction to the catalogue as well as several journal articles that focused on the construction of our bilingual, computer-based finding aid.
The significance of this collection lies, first, in its very breadth – roughly 100 years of book collecting (ca. 1810-1910), interspaced with letters and treatises from and about Shaykh Sidiyya, his son, grandson and two of his great-grandsons. This representative work of four generations of scholars within the same family includes 700-odd pieces of their correspondence, in addition to the literary works they consulted.
Second, this four-generation ‘slice' of intellectual life is linked to one of West Africa's premier scholarly lineages, the Kunta savants in the Timbuctu region where Shaykh Sidiyya studied for twelve years (1810-23), and includes over 130 letters Sidiyya copied from his Kunta mentors.
One copy of the film is available for consultation at the University of Illinois Library Archives, filed under “C.C. Stewart Collection”; two other copies of the film were returned to Mauritania in the care of Baba Ould Haroun, custodian of his father's collection, for deposit in an appropriate national repository. The originals of these manuscripts remain in Boutilimit in the care of Baba Ould Haroun.
AMMS assigns serial entry numbers that appear as the individual record number; citations to the item's actual collection location appear as the “Collection #” referring to the reel and film location on that reel. Each reel is made up of 20 strips of 36-exposure film spliced together. Citations in AMMS refer to the reel and the spliced film, thus a citation “23/16” refers to reel 23 and the 16th strip of spliced film on that reel; a filmed ‘target' for each manuscript identifies the item on the film within that 23/16 location.
A manuscript that begins at reel 63, film 18 and continues through to the 5th film of reel 64 would be noted as “63/18-64/5.” All filming was conduced under natural light and without the intrusion of sophisticated technology resulting in the re-filming of about 5% of the first 86 reels; re-filmed items are noted in the catalogue. It is unlikely that the original collection remains today in the same order in which it was preserved following the microfilming that followed the sequence of filming.
The range of subject headings used in this finding aid marked the beginning of a process of refinement that continued with the Nouakchott collection and has, imperfectly, been applied to other collections entered in the AMMS. It is neither comprehensive nor definitive. An effort was made to arrive at classifications that would make sense to a student of the Islamic sciences and to be consistent across English and Arabic entries, but in a few cases there may be slight variance across these fields. Correspondence presented special problems that were solved, where the content was discernable from a title or by the annotator in the field, by grouping letters into three broad, not unambiguous rubrics: politics, economy and society.
return to search pageThe manuscripts that are listed under “Kano” are drawn from three collections housed at the Africana Library at Northwestern University that were entered in the AMMS by staff there in 1990/91. While they are not exclusively drawn from Kano, they are indicative of Northern Nigeria's rich literary heritage. AMMS # 2055-2614 come from the Paden Collection, acquired by John Paden for Northwestern in the 1970s; AMMS entries # 2615-5948, labeled “Falke,” come from the ‘Umar Falke library, also acquired by Northwestern in the 1970s; and AMMS entries #2949 – 6263, identified as “Hunwick” from Professor John Hunwick's collection.
The “Kano” collections have been surveyed in two articles: Abdullahi Mohammed and Richard Hay, Jr., “Analysis of a West African Library: the Falke Collection” in Benjamin Mittman, Personalized Data Base Systems (New York, 1975) 77-96; and the Paden collection is reviewed by E. Saad, History in Africa, 7, 1980, 369-72. By comparison to the compilations of manuscripts from ‘national' collections that appear in AMMS and which feature heavy concentrations of classical texts, the Paden and Falke private libraries tend to privilege contemporary, twentieth-century material heavily influenced by writings associated with the Tijaniyya tariqa. Saad estimated that about one-half of the Paden accessions were privately printed pamphlets and books from Kano, Zaria and Cairo. Like the private library that makes up the Boutilimit collection, the Falke entries hold special interest as the library of an individual bibliophile.
The Kano collections are available at the Africana Library of Northwestern University, Evanston, Il..
Serial entries of AMMS entries appear as the record numbers for the AMMS; citations to the actual manuscripts appear under “collection number” for the individual entries where the separate “Paden”, “Falke”, and “Hunwick” collections are annotated. Use of the “GO TO” feature for individual collections (see ‘Search Tips') permits researches to access any one of these sub-sets of the “Kano” listings. Unlike the other AMMS entries that were input at the University of Illinois and therefore benefit from some degree of consistency (inclusive, no doubt, of similar errors), this set of entries was independently entered in AMMS at Northwestern.
One result of the differing input venues is that the system for identifying authors in the “Kano” entries differs slightly from that used in the other collections consolidated in AMMS. Where classical texts and local authors could be easily identified from the Northwestern entry system an effort has been made to edit those entries to conform with other AMMS references. The ‘Kano' collections include a somewhat larger number of fragments and partial texts that have not been possible to identify by comparison to the other entries in AMMS.
The range of subject headings used in AMMS represent an effort to apply the subject headings developed for the Boutilimit and Nouakchott collections, but in cases where there was a minimal level of detail annotated, the subject headings have been reconstructed from titles where these are available. The identification of subject matter in the “Kano” listings, while roughly parallel to the subject headings employed in the other collections, tends not to be as complete nor in uniform conformity with the other AMMS.3 entries.
Subject rubrics are classifications that would make sense to a student of the Islamic sciences and to be consistent across English and Arabic entries, but in a few cases there may be slight variance across these fields. Correspondence presented special problems that were solved, where the content was discernable from a title or by the annotator in the field, by grouping this into three broad, not unambiguous rubrics: politics, economy and society.
return to search pageThe first several hundred manuscripts that make up the Niamey collection were recorded in the AMMS in 1990/91 from the mimeographed catalogue prepared in 1979 by the Institut de Recherche en Sciences Humaines (List des manuscripts en langues arabe et ajami à l'Institut de Recherche en Sciences Humaines, Niamey, Niger, Niamey, 1979). This was the least detailed of all of our input sources and presented a number of problems that are no doubt still evident in the AMMS listing.
This collection is drawn from regions to the north-west of Niamey (the Azaouad) across modern-day northern Nigeria, and, to the north, Ahir/Aïr, and it provides a link between material collected at Timbuctu and that identified as “Kano” in AMMS. All 2551 entries in the mimeo 1979 hand-list have now been entered in AMMS, but a number of these are only partially identified due to the incomplete information in that listing. Happily, with other partial references we were able to more fully identify a number of citations. Currently, the Institut in Niamey possesses an estimated 4,000 manuscripts in its collection.
The Niamey collection was begun by the then President of the Niger Assemblée Nationale, Boubou Hama, in the early 1970s as part of his personal collection in documentation of his writing on the region. A survey of what had become the IRSH collection was published in 1980 (Ahmed Mohammed Kani, “A New Source for Material on the Literary Activity of the ‘ulama of the central and western Sudan: the Niamey collection” in Bulletin of Information, Fontes Historiae Africanae, Evanston, 1984/5, 41-48), and in 1996 the collection description was up-dated by M. Ayman Fadel in Sudanic Africa (7, 1996, 165-169).
It includes some photocopies of original material that appear elsewhere, but the List annotations do not indicate provenance of individual entries, and there is a handful of non-Arabic materials (conscientiously given Arabic titles). It does appear to be especially rich in nineteenth-century correspondence, but what part of this consists of original manuscripts is not clear.
The Niamey collection is open to researchers at the IRSH, Niamey.
Serial entries of items appear as the record number for AMMS; citations to the actual hand-list numbers in the List appear under “collection number” in the individual records. Since the published catalogue was produced in Arabic, we have reproduced the entries (inclusive of some errors) and added full names and nisbas where this information is available from other citations to the same authors in AMMS. These entries have required appreciable editing which is still in progress as this version of AMMS is put on line.
The range of subject headings used in AMMS represent an effort to uniformly apply the subject headings developed for the Boutilimit and Nouakchott collections, but in cases like the Niamey entries where there was a minimal level of detail annotated in the original catalogue, the subject headings have been reconstructed from titles where these are available. These subject rubrics are neither comprehensive nor definitive and, in this collection, particularly incomplete.
An effort was made to arrive at classifications that would make sense to a student of the Islamic sciences and to be consistent across English and Arabic entries, but in a few cases there may be slight variance across these fields. Correspondence presents special problems that were solved, where the content was discernable from a title, by grouping letters into three broad, not unambiguous rubrics: politics, economy and society.
return to search pageThe manuscripts that make up the Nouakchott collection were recorded in the AMMS in 1988/89 from the hand-list of the national repository for Arabic manuscripts at the Institut Mauritanien de Recherche Scientifique (IMRS) in Nouakchott . The IMRS began purchasing Mauritanian libraries and individual manuscripts in 1975 and by the late 1980s has acquired over 3100 items that were entered in AMMS. A separate project at IMRS focused on the preservation of poetry and is not included in this list that primarily focuses on major literary works rather than ephemera (correspondence, individual legal decisions, etc.).
The manuscripts in this collection bear comparison to two other hand-lists of Mauritanian work by Mokhtar ould Hamidoun and Adam Heymowski in 1964/5 and Ulrich Rebstock's microfilmed selection of 2239 manuscripts (including over 600 from the IMRS collection) in 1985.
At the time this listing was compiled the IMRS collection was made up of manuscripts from 72 libraries, mainly from the region of Trarza in the southwest quadrant of the country and with a focus on manuscripts of local authorship. The AMMS listing includes a number of items (493) that had not been identified at that time, but the number of multiple copies in this collection point to the possibility that the contents may be broadly representative of at scholarly activity in the region adjacent to the right bank of the Senegal River.
The IMRS provides access to its collection by bone fide researchers who make application through the Bibliotheque Nationale in Nouakchcott.
Serial entries of items appear as the record number for AMMS; citations to the actual IMRS manuscript identifying numbers appear under “collection number” in the individual records. Compromises between the capability of our in-put system and Hassaniyya language usages, mainly in personal names and place names, were generally resolved in favor of an Arabized version of the Hassaniyya name.
However, the use of Latin characters did permit an insertion of a pronunciation help where Hassaniyya letters had no equivalent in Arabic.
The range of subject headings used in this finding aid represents an elaboration of the subject headings developed for the Boutilimit collection and has, imperfectly, been applied to other collections entered in the AMMS. It is neither comprehensive nor definitive. An effort was made to arrive at classifications that would make sense to a student of the Islamic sciences and to be consistent across English and Arabic entries, but in a few cases there may be slight variance across these fields.
Correspondence presented special problems that were solved, where the content was discernable from a title or by the annotator in the field, by grouping letters into three broad, not unambiguous rubrics: politics, economy and society.
return to search pageThe manuscripts that make up the Segou collection were recorded in the AMMS in 1988/89 from the catalogue prepared in 1985 by Ghali, Mahibou and Brenner, Inventaire de la Bibliotheque ‘Umarienne de Segou (CNRS, 1985). The collection has been variously called the “Ahmadou Library” and “Fonds Archinard” (the latter after the French officer who seized the manuscripts at the time of conquest in 1890).
References to these manuscripts had appeared previously but incompletely in a 1925 catalogue of Arabic manuscripts at the Bibliotheque Nationale, and in a 1976 catalogue based on a selective inventory of the collection done in 1947-52. These efforts both omitted much material of interest to West African historians, which led to the entire collection being re-catalogued, and microfilmed between 1979-82; the inventory cited above was published three years later.
The Segou collection is from the library of Ahmadu Seku, son of al-Hajj ‘Umar, who inherited his father's conquests and ruled a territory increasingly under siege from 1864 until the French conquest in 1890. It contains numerous fragments and single page items and a large, valuable body of correspondence.
Its disparate make-up lends it a certain air of authenticity as a working library, a repository of day-to-day writings not filtered by an owner self-conscious of his scholarly image or literary capital of classical works.
The Segou collection is available at the Bibliotheque Nationale de Paris, Section des Manuscrits Orientaux. Xerox copies and/or microfilm of the collection is available at Yale University Library, Africana Collection, and at the Centre de Recherche et Documentation Historiques Ahmed Baba, Timbuctu, Mali.
Serial entries of items appear as the record number for AMMS; citations to the actual BN manuscript identifying numbers appear under “collection number” in the individual records. Since the published catalogue was produced in French transliteration, the AMMS Arabic entries required a re-translation, greatly aided by access to M. al-Ghali and Mahibou's notes, kindly made available by Professor John Hunwick.
Some editing has been done to bring names and spellings in conformity with other entries of the same authors in the AMMS listings, and where partially-identified authors of classical works were listed, these have been filled out to facilitate their cross-listing with other collections.
The range of subject headings used in this finding aid represent an effort to consistently apply the subject headings developed for the Boutilimit and Nouakchott collections, but this was only possible where the level of detail annotated in the original catalogue allowed it. It is neither comprehensive nor definitive. An effort was made to arrive at classifications that would make sense to a student of the Islamic sciences and to be consistent across English and Arabic entries, but in a few cases there may be slight variance across these fields.
return to search pageThe manuscripts that make up the Timbuctu collection were recorded in the AMMS in 1990/91 from a photographed copy of the hand-list maintained at the Centre de Documentation et de Recherches Historiques Ahmad Baba (CEDRAB), Timbuctu, made available for this project by the then director Mahmoud Zoubair. At that time 5640 manuscripts had been recorded at CEDRAB, a compilation of locally gathered materials that has since grown to over 16,000 holdings. The Al-Furqan Islamic Heritage Foundation began printing a series of catalogues for CEDRAB in 1995 with rather more attention to individual manuscripts than was possible in the hand-list, and although the numeration of the first 5640 items in AMMS.3 roughly correspond to the printed volumes, there is some discrepancy and researchers will need to confirm the record numbers for the Timbuctu collection that are cited in AMMS.3 entries with the official published volumes.
These are:
Vol. I (ed. Sidi Amar Ould Ely, 1995) records 1-1500;
Vol. II (ed. ‘Abd al-Muhsin al-‘Abbas, 1996) records 1501-2999;
Vol. III (ed. ‘Abd al-Muhsin al-‘Abbas, 1997) records 3001-4500;
Vol. IV (ed. ‘Abd al-Muhsin al-‘Abbas, 1998) records 4501-6000;
Vol V (1998), records 6001-9000.
Correspondence between the order in which manuscripts are listed in these catalogues and the AMMS.3 numeration is not always exact. The following is a summery of the correspondence between AMMS.s entries and the published catalogues.
Volume I lists entries in a numerical order and inserts the actual collection numbers in (brackets); these bracketed collection numbers correspond to the numeration of the first 1500 items in AMMS.3.
In Volume II there is exact coordination between manuscript numeration and the AMMS.3 entries up to published catalogue entry 2672. Between it and # 2900 there is a discordance that is resolved by published entry # 3000 (which appears in AMMS.3 as # 3002 in the ‘Timbuctu' collection).
Volume III continues in exact coordination with the AMMS.3 entries but for a two-entry discordance noted above, and by entry #4467 both systems are again entirely synchronized.
Volume IV numeration and that of the AMMS.3 entries continue to be synchronized up to entry #5561 (the point where the CAB collection concluded in 1991) and the AMMS.3 entries now conclude.
The Timbuctu collection was built from various individual collections in Northern Mali beginning in the late 1970s. Among the most important local libraries that were incorporated into the CEDRAB collection was that of Ahmad Bul-Araf, a Timbuktu bibliophile of Moroccan ancestry. By the early 1990s, CEDRAB had become the major Arabic manuscript repository in West Africa, recipient of UNESCO and Al-Furqan funding and with facilities for researchers and manuscript preservation that are unparalleled across the Sahel.
Its early accessions reflect the bias found in other national repositories toward classical works and major local authors, although more recent acquisitions (as reflected in the published catalogues) include a large amount of correspondence and lesser works, perhaps more reflective of the locally-produced literary record of the region. This first on-line version of AMMS.3 has not been coordinated with the additional data relative to individual entries that appear in some of the published catalogues, and it must therefore be used as a rough, and partial, guide to the CEDRAB collections.
The Centre Ahmad Baba collection is open to researchers in Timbuctu, Mali. Officially, researchers require authorization from the Malian government to conduct research in the country. Application for research authorization should be directed to the Centre National de la Recherche Scientifique et Technologique (CNRST), BP: 3052, Bamako, Mali; Tel. (223) 219085; Fax: (223) 218446/216698.
The Centre Ahmad Baba can be contacted directly at the following address:
Le Directeur
Dr. Mohamed Gallah Dicko
CEDRAB
BP 14, Tombouctou, Mali
Tel: (223) 921081
Serial entries of items appear as the record number for AMMS; citations to the actual hand-list numbers in the CEDRAB collection (as they appeared in 1990) are under “collection number” in the individual records. The numeration of the first 5640 items from the CAB that appear in AMMS.3 roughly correspond to the printed volumes cited above, but there is some discrepancy (see above) and researchers will need to confirm the record numbers for the Timbuctu collection that are cited in AMMS.3 entries with the official published volumes (I-IV).
The range of subject headings used in AMMS represent an effort to uniformly apply the subject headings developed for the Boutilimit and Nouakchott collections, but in some of the CEDRAB entries (working from their hand-lists rather than the better-documented, published volumes now available) the subject headings were reconstructed from titles where these were available. These subject rubrics are neither comprehensive nor definitive.
An effort was made to arrive at classifications that would make sense to a student of the Islamic sciences and to be consistent across English and Arabic entries, but in a few cases there may be slight variance across these fields. Correspondence presents special problems that were solved, where the content was discernable from a title, by grouping letters into three broad, not unambiguous rubrics: politics, economy and society.
return to search page