Beyond ‘aam in Ethiopia: a short Note on an Arabic-Islamic Collection of Texts written in Ethiopian Script (fidäl)

: As in many other countries in Sub-Saharan Africa, Muslims in Ethiopia have produced a substantial amount of literature in their local languages, mostly using the Arabic script rather than the Ethiopic syllabary ( fidäl ), which has been connected with the Ethiopian Christian Church and state for centuries. Recently, after fidäl was adopted to write the language of the Harari Muslim people, a manuscript was written in which texts in Arabic and Old Harari were copied in Ethiopic script. I have analysed the manuscript in this paper, highlighting the strategies that the copyist followed in order to transcribe Arabic into fidäl , and have attempted to place the work within the general framework of Islamic-Chris-tian relationships in Ethiopia.

Some marginal paratexts among Christian Ethiopian manuscripts in Gǝ'ǝz and Amharic have been found to include Arabic words, sentences and also short fully-fledged texts written in Ethiopian and/or in Arabic scripts.3Christian Ethiopian astronomical and calendar treatises (e.g. the Baḥrä hassab, 'The sea of calculation'),4 contain many Arabic words and expressions.In the famous 16 thcentury Christian treatise Anqaṣä amin (the 'Door of the faith'), written in Gǝ'ǝz, Arabic-Islamic religious terms are quoted and transcribed using fidäl script.5Also, the 20 th -century Amharic Christian Sylloges of šayḫ Zäkaryas contain several words and phrases taken from Arabic.6During my recent examination of MS EMML 6239, I identified a bilingual Arabic-Amharic version of šayḫ Zäkaryas' work written entirely in Ethiopian fidäl.7 Arabic-Islamic words are also scattered throughout Christian Ethiopian magical literature, where they function as 'abracadabra' to enhance the mystical power of the texts.8The writing down of Islamic texts in Arabic or Amharic or their transliteration into one of these languages was usually carried out by or for the sake of Western scholars working in Ethiopia.9 Moving from the Ethiopian Christian to the Ethiopian Islamic cultural landscape, the existence of Arabic-Islamic texts written in fidäl becomes less vague, at least during the modern era.A cultural and religious prejudice identifying Ethiopian script with Ethiopian Christianity may have hindered the usage of fidäl among Muslims at an early stage.However, since the mid-20 th -century emergence of Islamic books printed in Amharic and Tigrinya, Ethiopian script has also begun to be used to write scattered words and phrases in Arabic.After the fall of the Socialist regime in 1991, a more or less religiously oriented Muslim book production began to blossom in Ethiopia, and fidäl became more and more widespread among Amharic-and Tigrinya-speaking Muslims who make use of it to transliterate passages from the Qur'an, Hadith and other fundamental texts of the Islamic tradition for books and booklets published in local languages.The subsequent birth and diffusion of a kind of Islamic calligraphic practice in Ethiopian script came to confirm a fairly generalised acceptance of fidäl among many Ethiopian Muslims.
An especially remarkable position in the general picture I am sketching here is occupied by Harari, the Semitic language of the city of Harar, spoken by the ethnic group of the Hararis.Traditionally written in Arabic script since at least the beginning of the 18 th century,10 Harari texts written in Ethiopian script have been found to have emerged at the end of the 19 th century, after the walled city was conquered by Mǝnilǝk in 1887 and incorporated into the modern Ethiopian state.11The connection between the Ethiopian writing system and the Harari language became even tighter after the fall of the Imperial state (1974) and under the Socialist regime.Harari texts in Ethiopian alphabet were published abroad12 and passages compiled by aläqa Tayyä for Eugen Mittwoch in Berlin who published them in 1906 in Ethiopian script.10 For a general description of Harari literature in Arabic script, see Banti 2005Banti , 2010.The language of the texts written in Arabic script has been labelled 'Old Harari', because it shows some quite substantial differences from modern Harari as spoken and written in Harar nowadays.11 See for example the texts written by the secretary of ras Mäkwännǝn for Casimir Mondon-Vidailhet and published by Carlo Conti Rossini in 1919 (see Wagner 2004, 355, note 5 with other examples).12 See for example the Suwār malasāyāčč ṭabā ('Voice of the revolutionary youth'), a handwritten ideological journal produced by the Harari Students' Association in Egypt (Harari ardāwigāčč ahadǝnnat mugād misrābe).On a website entitled Everything Harar (http://www.everythingharar.com;last accessed 04/02/2018), which is run by Harari individuals from the Diaspora, I found two issues of the periodical (vol. 3 no. 2, January 1980 andvol. 3 no. 3 [originally 2], February 1980) which I was able to retrieve and download.The periodical was apparently produced in Egypt, but no information is available or traceable about its origin and periodicity.The content of the texts is extremely interesting from both a politico-cultural and from a purely linguistic point Unauthenticated Download Date | 4/20/19 1:20 PM in Ethiopia,13 fostering the diffusion of fidäl among Hararis (first and foremost among the intellectual elite).
After the pro-Soviet regime collapsed in 1991, a relatively short period of incertitude followed: different options (Latin, Arabic, and Ethiopian scripts) for the writing of the Harari language were discussed, until eventually the choice officially fell on fidäl.14The appearance in 1992 of the first Harari-Amharic dictionary published entirely in fidäl paved the way to further development of the connection between the Harari language and the Ethiopian script.15 Fidäl is nowadays well established among Harari speakers and readers and has progressively become more deeply rooted in the writing and reading practices and in the literary production of Hararis.At the same time, Latin script has not completely been entirely abandoned and still seems to be quite widespread among Hararis living outside Ethiopia,16 and it is extensively used on the web.17 What was probably also instrumental in this impressive success of the Ethiopian script among the Hararis was the diffusion of Amharic in the city of Harar and especially in the Harari diaspora.Numerically substantial and economically lively communities of Harari speakers can be found in Addis Ababa, Dire Dawa and many other cities of Ethiopia.These scattered groups of Hararis (in particular those living in the capital) are more and more exposed to Amharic, are practically bilingual and tend to use written Amharic for their daily business and in connection with official and bureaucratic issues.The combined effect of the shift to fidäl to write Harari and the wider and more intense exposure to Amharic has reduced the knowledge and usage of Arabic script among the Harari people and the knowledge of Arabic itself.
As a matter of fact, while Arabic is known as a language of culture by a very restricted circle of learned men and is used as a spoken medium of communication by Harari emigrants returning from jobs in the Gulf, the usage of Arabic script to write Harari seems to be vanishing completely.Moreover, the capability to read and understand traditional Old Harari texts written in Arabic script is apparently declining dramatically.Manuscripts in Arabic script (both in Arabic and in Old Harari language) are not copied anymore but are substituted by computer-typed books and electronic documents.The recent production of fidäl reprints of the Kitāb al-farā'iḍ ('The book of obligations'), probably the most famous piece of Harari literature, proves the sorely felt need for editions of traditional texts targeting the growing number of readers who are unable to read Arabic script.18It is this general sociolinguistic framework which forms the context for the origin of the manuscript I am going to deal with in the rest of this chapter.

Arabic in fidäl: the case of a Harar manuscript
The manuscript I am about to discuss was photographed in Harar on 21 September 2003 by Dr Simone Tarsitani (Durham University)19 during an ethno-musicological research mission he was conducting.The item is basically a collection of different texts written down by the late Mr Abdi Abubakar Sufiyan20 with a blue ballpoint pen (some parts are in red) on a personal organiser (15cm x 20cm) for the year 1990 of the Ethiopian Calendar (i.e.year 1997-1998 of the Gregorian calendar; the Ethiopian year starts on 11 September).The manuscript pages lack proper numeration, therefore in the following I will locate the texts according to the dates shown on the daily agenda (every single page is devoted to a different day, except for the weekends: Saturdays and Sundays share the same page).
The texts are written almost continuously from the first day of the year (11 September) until 6 April.The rest of the organiser is blank, except for the pages of 4 September and 8 September (the pages in between these two dates have been torn out) and the page of the weekend 9-10 September (the last two days of the Ethiopian year).The manuscript contains texts in three languages written in Ethiopian script: Old Harari, Arabic and Oromo.Besides, English in Latin script and Arabic in Arabic script (8 September, sūrat al-kahf, Qur'an 18, verses 1-10) are also copied in the agenda.
The texts copied into the manuscript belong to what I have elsewhere referred to as the 'Harari Mawlid collection',21 that is, the constellation of poetical and prose texts (mainly in Arabic, but interspersed with sustained wide sections in Old Harari) which is usually recited by Hararis to solemnise the feast of the birthday of the Prophet and on other important public and private occasions of a religious but sometimes also a secular nature.The collection appears in a wide diffusion of manuscripts (both relatively ancient and very recent) in Ethiopia.
While no comprehensive analysis of the manuscript tradition of the 'Harari Mawlid collection' has as yet been carried out, the following data are available: In the library of the Institute of Ethiopian Studies at the University of Addis Ababa, the following eight manuscripts contain the 'Harari Mawlid collection': 264:2, 273:3, 1855:2, 2662:2-3, 2663:2, 2664:2 (incomplete), 2665:2, 2666:2  The 'Harari Mawlid collection' has also been printed several times both abroad and in Ethiopia.
In Egypt, the work has been printed twice: in 1350/1931 (127 pages) and in 1366/1947 (100 pages), under the title Mawlid šaraf al-'ālamīn ('The birth of the honour of the universe').Two editions published in Ethiopia (around 1992Ethiopia (around -1993Ethiopia (around and in 2000) )  This conspicuous number of testimonies proves the high esteem that the text collection enjoys among Hararis both inside and outside the city of Harar; however, despite this, research on the origins and the first diffusion of this text collection is still scanty and insufficient.24 One of the most remarkable features of the 'Harari Mawlid collection' is that while it is structured according to a relatively stable general framework, the verses and prose sections which actually make up the textual constellation and the sequence in which they are put together show a high degree of variation from manuscript to manuscript and from book to book.
Among the testimonies of the 'Harari Mawlid collection' I am aware of I was not able to identify a direct model for the manuscript I present here.
Moreover, for the specific purposes of this chapter, I will only focus here on a number of main Arabic texts which I managed to identify, leaving to another occasion a full description of the unknown or unidentified Arabic texts and the analysis of the Oromo and Harari sections.
When lacking a well-established title, poetic texts are identified by their incipit; for cross-reference, I will use the September 2005 edition of the 'Harari Mawlid collection'.
Here is a list of the texts I managed to identify in the fidäl-Arabic part of the manuscript: -6-8 October: poem  , a member of the Burhāniyya brotherhood.The text copied into the manuscript under discussion extends to the verse: Šafī'u al-ḫalqi fī ḥayratihim ('Intercessor of the creatures in their bewilderment').28 I am here transcribing the phrase as it should be according to the rules of Classical Arabic: the transliteration system of Arabic used in the text is specifically analysed in part 3 of this chapter.3 Fidäl for Arabic: an easy adaptation?

Unauthenticated
The relative proximity of the Arabic phonetic system to that of many of the Semitic languages of Ethiopia and Eritrea could generate the idea that the usage of fidäl to write Arabic might be a relatively simple adaptation process.I believe that the following few observations on the way Arabic has been transcribed in the manuscript under analysis only partially confirm this idea.First of all, the linguistic interference represented in the manuscript is that between Arabic and Harari.To briefly highlight the main differences between the phonologies of these two languages, Wolf Leslau's contribution on the Arabic loanwords in Harari is useful as a reliable reference:39 Harari, on the one hand, possesses four phonemes which are absent in Arabic: a voiceless palato-alveolar affricate /č/; an 'emphatic' (ejective, glottalised) voiceless palato-alveolar affricate /č̣ /; a palatal nasal /ñ/; and a voiced velar stop /g/.Arabic, on the other hand, has seven phonemes which do not exist in Harari: a voiceless (inter)dental fricative /ṯ/; a voiced (inter)dental fricative /ḏ/; an 'emphatic' (pharyngealised) voiced alveolar fricative /ẓ/; an 'emphatic' (pharyngealised) voiceless alveolar fricative /ṣ/; an 'emphatic' (pharyngealised) voiced dental stop /ḍ/; a voiced velar fricative /ġ/; and a voiced pharyngeal fricative /'/.
Moreover, there are four further important differences 1) the so-called 'emphatic' consonants are pronounced as ejectives (glottalised) in Harari, while they are basically pharyngealised in Arabic; 2) in Harari the voiceless velar fricative [ḫ] appears only as an intervocalic variant of the voiceless velar stop /k/ as a consequence of spirantisation and thus has no phonemic status.In Arabic loanwords, /ḫ/ passes to /k/ but is sometimes preserved; 3) Harari has a voiceless pharyngeal fricative /ḥ/, but no voiceless glottal fricative /h/.Arabic /h/ is thus rendered as /ḥ/ in Harari, but in some cases the original /ḥ/ is preserved; and, finally, 4) the vocalic system of Harari is still an object of discussion among linguists, but latest research points to the phonemic value of vowel quantity.40It thus seems that in terms of both vowel quantity and vowel quality the phonology of Harari can easily accommodate Arabic words.
In the end, Leslau's analysis clearly demonstrates that the passage of a relatively substantial amount of Arabic loanwords (about 300) into Harari vocabulary did not take place according to a stable correspondence among phonemes of the two languages.The inconsistent phonetic adaptation is apparent when considering the way the Arabic phonemes (in particular those unknown to Harari) are rendered in Harari.41It can be surmised that the picture is complicated by the fact that the Arabic loanwords' mode of entry into Harari is twofold: 1) oral, through one of the forms of spoken Arabic to which the speakers of Harari are exposed;42 and 2) written, 39 See Leslau 1956, andLeslau 1957, which contains a comprehensive description of the differences between the Arabic phonological system and that of other Semitic languages of the Horn of Africa.40 For a detailed discussion on this, see Garad-Wagner 1998, 157-168, and for a more general description, see Wagner 1997, 487-488. 41 For the details, see in particular the recapitulative table in Leslau 1956, 21. 42 Leslau (1956, 22-23), indicates a number of Arabic dialects as possible sources for at least Unauthenticated Download Date | 4/20/19 1:20 PM through a learned milieu, especially that of Arabic teachers and Islamic scholars who are well acquainted with the Classical Arabic of the Qur'an and the theological tradition.43 In the case of the manuscript I am presenting here, the issue of the Arabic dialects remaining behind or underneath the words is not relevant.Nevertheless, the idea that an oral dimension of the texts has influenced their transcription cannot be excluded from these considerations.
The main features of the transcription system used in the manuscript for depicting Arabic words in Ethiopian script can be briefly sketched as follows: -voiceless (inter)dental fricative /ṯ/: written as ሰ <s> for voiceless alveolar fricative /s/; e.g.ሱቡቱ for ṯubūtu ('His immutability'); -voiced (inter)dental fricative /ḏ/: written as ዘ <z> for voiced alveolar fricative /z/; e.g.ቢዛቲሒ for bi-ḏatihi ('By His essence'); -'emphatic' (pharyngealised) voiced alveolar fricative /ẓ/: written as ዘ <z> for voiced alveolar fricative /z/; e.g.ዛሐራ for ẓahara ('It appeared'); -'emphatic' (pharyngealised) voiceless alveolar fricative /ṣ/: written as ሰ <s> for voiceless alveolar fricative /s/; e.g.ሲድቁን ṣidqun ('His truthfulness'); -'emphatic' (pharyngealised) voiced dental stop /ḍ/: written as ደ <d> for (nonemphatic) voiced dental stop /d/; e.g.ዲያኡን for ḍiyā'un ('Lights'); -voiced velar fricative /ġ/: written as ኽ < ḫ > for voiceless velar fricative /ḫ/, not only in intervocalic position but also in initial position; e.g.ኸኒዩን for ġaniyyun ('Rich'); -voiced pharyngeal fricative /'/: written as አ <'> for voiceless glottal stop /'/; e.g.አሊዩን for'aliyyun ('Sublime'); -voiceless velar fricative /ḫ/: written as ኽ <ḫ> for voiceless velar fricative /ḫ/, not only in intervocalic position but also in initial position; e.g.ኸረቅቱ for ḫaraqtu ('I tore'); -voiceless pharyngeal fricative /ḥ/ and voiceless glottal fricative /h/: mostly both written as ሐ <ḥ> for voiceless pharyngeal fricative /ḥ/; sometimes the Arabic voiceless glottal fricative /h/ is 'correctly' written as ሀ <h> with the some of the Arabic loanwords in Harari, without being able to localise all of them in one single dialect.His conclusions are somewhat ambiguous.After remarking that 'one is inclined to think that not one single Arabic dialect is to be taken as the source of Harari, but various Arabic dialects', he says a few lines later: 'As for the historically possible source of the Arabic loanwords in Harari, a South Arabic dialect seems to be more likely to be considered'.43 Leslau (1956) did not at all consider the possibility that at least some of the Arabic words he discusses could have entered Harari vocabulary from a learned, written source and not from a spoken variant of the language.An assessment of the loanwords taking this possibility into account might provide some fresh insights into the topic.
The quality of the Arabic vowels /a/, /i/, /u/ is always reproduced accordingly in the text in fidäl.As for quantity, only /ā/ is sometimes rendered with the fourth order of the Ethiopian syllabary (<Ca>), which in the orthography of modern Harari also represents the long a (short /a/ written with the first order of the syllabary (<Cä>), e.g.ኸረቅቱ for ḫaraqtu ('I tore') and ዋሒዱን for wāḥidun ('Unique'); however the fourth order long /a/ can also be used for an Arabic short /a/, e.g.ዛሐራ for ẓahara ('He appeared').44As for Arabic /i/ and /ī/, both are always written with the third order of the Ethiopian syllabary (<Ci>), as well as /u/ and /ū/ being reproduced with the second order of the fidäl (<Cu>).
Diphthongs /aw/ and /ay/ are more or less consistently rendered in fidäl (e.g.ለው for law ['If']), but many times the writing in fidäl seems to point to an assimilated pronunciation of the /a/ to the following /w/ or /y/ e.g.ዮውም yowm for yawm ('Day'), and ኼይሩ ḫeyr(u) for ḫayr(u) ('Benefit').
It is worth noting that in many instances the word division does not correspond to the one used in Arabic; thus forms like ቢዛቲሒ for bi-ḏatihi ('By His essence') can be found together with ፈቀልቢል ዮውመ (falqalbil yowma) for fa-qalbī al-yawma ('Then my heart today').
To sum up the collected data, the rendering of the Arabic texts in fidäl can be considered as being only partially accurate.Failure to represent many Arabic phonemes makes the transcription to a great extent imprecise: it is true that the Ethiopian syllabary lacks specific signs for some of the Arabic phonemes and for distinguishing the vowel quantity of /i/ and /u/, but no creative effort has been made by the copyist to fill the gaps in the writing system he was using and create a one-to-one sign set of transliteration.
The written rendering in fidäl of many of the Arabic sounds which are absent in Harari follows the same pattern as the phonetic adaptation of the Arabic loanwords into Harari.The word division in Ethiopian script is quite inconsistent and often only oddly corresponds to the Arabic original.The diphthongs are written in a way that hints at an assimilation of the short Arabic /a/ to the following semi consonant.
These elements possibly indicate that the writing process of the manuscript was carried out while hearing someone from Harar reciting the texts or while the copyist himself was reading them aloud.As a matter of fact, the writing down of the Arabic texts does not adhere to any written model: the copyist does not even try to rewrite in fidäl a written Arabic Vorlage, but mostly transcribes the texts the way they would sound, if they were read by a Harari speaker.The copyist does not seem to have any intention to produce a well-thought-out system of transliteration of Arabic with the Ethiopian syllabary, he just aims at providing the faithful who are unfamiliar with Arabic script with a tool to access the texts and recite them on different religious occasions.45 To my knowledge this is so far the only known example of a substantial amount of complex Arabic texts written in Ethiopian syllabary.The general cultural trends in the Islamic communities in Ethiopia will tell us whether it will remain only an individual, isolated effort or the first step of a linguistic and literary development among the Ethiopian Muslims.
under the title Kitāb 'unwān al-šarīf bi-al-mawlid al-šarīf ('The book of the noble title on the noble birth') are actually photomechanical reproductions of two manuscripts.The 1992 edition (160 pages) contains a manuscript dated 1412/1992-1993 and written by an anonymous copyist.The 2000 edition (171 written pages) reproduces a manuscript dated 26 ramaḍān 1421/22 December 2000 and written by the famous copyist Ibrāhīm Muḥammad Wazīr.A third Ethiopian printed edition was published in Dire Dawa at an unknown date with the title Mawlid šaraf al-'ālamīn ('The birth of the honour of the universe') at the expenses of Mahdī ḥāǧǧ 'Abdallāh.It seems to be a reprint of one of the above-mentioned Egyptian books.Finally, in ša'bān 1426/September 2005, a computer-typed text (175 pages) was circulated in Addis Ababa under the title Kitāb mawlid šaraf al-'ālamīn ('The book of the birth of the honour of the universe') by the Harari scholar Abū Bakr Ṯābit (Sabit).