Words that Burn: On the Manners and Implications of Oath-Taking Practices in Ethiopian Amhara Customary Law, Nineteenth–Twentieth Centuries

Between the spoken word, ritual action, and legal processes, the studies of oath-taking practices have developed a broad literature. This article provides an additional layer of materials and analysis on speech acts and ritual procedures involved in the manners of taking an oath in the Christian societies of Ethiopia, as recorded from the midnineteenth century to the early twentieth century. Some samples of Amharic discourse specific to the manners of oath-taking in the customary legal system of Christian Ethiopia are presented here through extracts from unpublished field notes recorded in the 1840s by the French traveller Arnauld d’Abbadie. This source is then compared to other ethnographic observations of oath-taking statements and rituals in the context of Ethiopian Christian societies. The implications of swearing an oath in Ethiopian customary law lead to the critical re-examination of the history of Ethiopian law in a comparative outlook, particularly with the canonical laws of Eastern and Western Europe.

put the meaning of its words into effect only by the performative virtue of its main verb. The French 'prêter serment' literally means 'to give an oath', adding an additional aspect of recipiency. Indeed, the fact that an oath is successfully enacted is not only a matter of delivering it as a statement in the first person: 'I swear that …'. It must be also heard.
After Austin, John Searle (1969) refined the logical component of the speech acts theory. What he calls the 'illocutionary force' of performative statements is effective only if there are recipients to acknowledge the speaker's intention behind the propositional content of the statement. Hearers, and particularly addressees, participate in the enactment of the illocutionary act until its completion. In short, the effect of the performative act is determined by the receiving side. This is wittingly encapsulated in the French political aphorism: 'promises bind only those who listen to them '. 4 In addition to utterance and reception, swearing an oath falls into the case of declarations defined by Searle as 'institutional facts' for they 'require special institutional contexts for their successful performance'. 5 For instance, priests or lay officers are officially invested with the capacity to transform lovers into spouses by pronouncing them husband and wife after having announced and ascertained the obligations involved by marriage as a sacrament or as a contract. In the same vein, the president of the jury is the only judge entitled to sentence a defendant found guilty at the end of a trial.
Whatever the context, ancient or contemporary, religious or secular, the full accomplishment of taking an oath involves specific procedures combining ritual acts and rhetoric formulas that add weight to the statement supported by the oath. Through these procedures the truth of the statement attached to the oath is bound to serious sanctions in case it is false, that is, perjury. This punishment, believed as inescapable, is triggered through a contract with the sacred or supernatural domain, beyond the temporal legal sphere dealing with factual proofs. The oath takes its value as a legal act from its property of reaching beyond the realm of ordinary justice. This procedural aspect of the oath as a performance that is incomplete as a single statement was highlighted by the linguist Émile Benveniste in the Introduction of a study on the expressions of oath in ancient Greek (ὅρχος) comparing the lexicon related to oath in Indo-European languages: The oath is not an institution defined by its own purpose and effectiveness. It is a particular form of assertion, which supports, guarantees, or demonstrates, but does not justify anything. Individual or collective, the oath […] prepares or completes a speech act that alone has significant content, but it does not state anything by itself. It is in fact an oral rite, often complemented by a manual rite of variable form. Its function consists not in the affirmation it produces, but in the relationship it establishes between the word spoken and the power invoked, between the person of the swearer and the domain of the sacred. 6 This comprehensive definition is applicable to cases presented below. If functional aspects can explain some common features in the performance of swearing and taking an oath in a wide variety of contexts, the wide dissemination of standard ways of performing an oath is linked to ancient historical roots transmitted mostly through sacred texts, and associated traditions. In a study of oath formulas in the Hebrew Bible, Blane Conklin identified five central patterns qualified as 'authenticating elements' that confer truthfulness to the act of swearing. These include (1) raising one's hand; (2) using verbs specific to the act of swearing; (3) invoking a third party to bear witness of the oath; (4) uttering the expression '(by) the life of X' where X is a sacred or superior entity, for instance God or the king; (5) and a final binding formula: 'thus will X do to Y'. 7 These biblical patterns, themselves rooted in older patterns of the ancient East, have had a long-lasting and farreaching influence in the formal manners of oath-taking throughout the world to the present. As we will see below, the liceity of the oath as codified in the Hebrew Bible was debated and reformulated in the works of Christian canonical law, allowing its recognition as a valid mode of proof in legal practices outside the canonical sphere.
Between the spoken word, ritual action, and legal processes, the studies of oath-taking practices have developed a broad literature. This paper provides an additional layer of materials and observations on speech acts and ritual procedures involved in the manners of taking oath in the Christian societies of Ethiopia as recorded from the mid-nineteenth century to the early twentieth century. Before reforms were undertaken to remodel the Ethiopian judiciary system in line with European standards, the processes of customary justice were primarily managed on a communal level. Any member of a neighbourhood, mostly males, could be involved in trials not only as witness, but even as arbitrators. 8 It was a necessary skill for any subject of law to possess the appropriate manners to defend their rights through spoken interactions. The ability of making official statements in the form of vow, pledge, or oath was particularly required for authenticating any kind of act with legal value. The idea that literacy was normatively superior has overshadowed, for many scholars working on Ethiopian history, the importance of orally transmitted information or orally performed statements. 9 Written acts had a high legal value, but their use was limited to the upper ranks of clerical and state administration. Ritualized speech acts, performed in a proper and efficient way, were thus required in all kinds of legal procedures.
Some 'fossils' of Amharic discourse specific to the manners of taking oath in the customary legal system of Christian Ethiopia will be presented here through extracts from unpublished field notes recorded in the 1840s by the French traveller Arnauld d'Abbadie. 10 This source will be compared to other descriptions of oath-taking statements and rituals published in later works on legal practices in the context of Ethiopian Christian societies. The implications of swearing an oath in Ethiopian customary law will lead to the opening up of certain perspectives for the critical re-examination of the history of Ethiopian law in a comparative outlook, particularly with the canonical laws of Eastern and Western Europe.

'Proto-Ethnographic' Notes on Customary Law in Christian Ethiopia
Born in 1815, Arnauld-Michel d'Abbadie was French-Basque on his father's side and of Irish descent on his mother's. He was the younger brother of Antoine d'Abbadie, who was a celebrated and accomplished scholar in the fields of earth sciences and philology and became one of the founding fa-8 Aberra Jembere 2000, 222. 9 In the Ethiopian context, the fieldwork research undertaken by Ivo Strecker and Jean Lydall among the Hamär of the South Omo area has been essential for including the voices of these people into a broader and multi-vocal perspective of Ethiopian history, by way of restitution of rhetoric practices and speech acts as genuine sources. See, in particular Strecker 2013, for the 'Iliad-like' narrative of Berimba, who was the spokesman of the Hamär when southern Ethiopian peoples were subjugated by the army and the administration of the northern Christian kingdom. 10 I thank Shiferaw Bekele, Ahmed Hassen Omer, Wolbert G. C. Smidt, Fesseha Berhe, Felix Girke, Sophia Thubauville, and Siena-Antonia de Ménonville for their advice on the very preliminary steps and first draft versions of this article.
thers of Ethiopian studies in France. Overshadowed by the fame of his elder brother, Arnauld is usually portrayed as an adventurer who played an auxiliary role in the research activities carried out by Antoine. His travel accounts, however, give an insight into his well-informed and highly penetrative knowledge of highland Christian Ethiopian society. 11 His still unpublished field notes, found in the d'Abbadie papers deposited at the Vatican Library (Biblioteca Apostolica Vaticana), show that his views are not only based on recollections from personal experience, but on very detailed handwritten records of interviews and first-hand observations. If not a scientist according to the academic criteria of his time, Arnauld prefigured a kind of social science more oriented towards the dense description of human characters through the understanding of their social positions and their life experience, rather than the systemic collection of factual and exact data processed through scientific instruments and formulas. Not only a bold horseman, Arnauld was trained in law. He had a keen interest in legal issues, political organizations, and military affairs. He recorded observations and first-hand information through a methodology that can be assessed as 'proto-ethnographic'. Before ethnography was institutionally established as an academically codified discipline, tools and guidelines for the systematic recording of social information, such as questionnaires, had already been designed and promoted by learned societies. 12 It is via these first instruments of fieldwork methodology that Arnauld d'Abbadie developed his own quasi-methodology of 'writing on the field'. 13 When comparing Antoine d'Abbadie's notebooks with those of Arnauld, it becomes clear the two brothers did not share the same methodological framework. Antoine's 11 The first volume was published in 1868 in Paris (Arnauld d'Abbadie 1868), it was reedited at the Vatican City in 1980 in the collection Studi e Testi of the Biblioteca Apostolica Vaticana (Arnauld d' Abbadie 1980a). The three remaining volumes, that were kept as manuscripts in the Vatican Library (representing 17 of the 19 boxes of d'Abbadie's papers) were published by Jeanne-Marie Allier in 1980, 1983, and 1999(Arnauld d'Abbadie 1980b1983;1999 Strecker and LaTosky 2013. notes are much more structured, indexed with keywords and symbols. It seems that Arnauld, five years younger than Antoine, had not been influenced by his elder brother's more rigorous methods for recording ethnographic information. Nevertheless, it was highly likely they exchanged views and information, before they quarrelled many years later. Antoine's article on 'procedure in Ethiopia', published in 1888, does not include remarks on oath making, 14 but the information and analyses it presents are very similar to observations recorded in his younger brother's notebooks. This indicates possible information exchange between the two brothers that should merit deeper investigations in the study of their, as yet, largely unpublished papers. Among the field notes preserved in Arnauld d'Abbadie's Vatican papers, the most extensive document (Vatican City, Biblioteca Apostolica Vaticana, Carte d'Abbadie 18, fols 340-403, henceforth BAV-ABB-18) is made up of 52 folios of large-size paper featuring a list of 450 questions followed by approximately 325 replies, 15 all related to different fields of the application of customary law in the Christian societies of Ethiopia (land issues, property transmission, rules of marriage and divorce, procedures in case of blood crimes, and so on). The first page bears the short title of Λογος (Logos) in Greek. The meaning of this title may refer to the general sense of knowledge acquired through investigation. It may also designate the logical arguments and supportive evidence to persuade an audience, whereas the Greek notion of nomos encompasses the conventional dimension of laws based on divine or social order. 16 Most of these questions are given short replies, which are a summation of personal observations or information collected from informants who are not named. This anonymization is more the result of a lack of rigour than the expression of a superiority complex. In his travel account, 17 Arnauld d'Abbadie particularly insists on acknowledging the invaluable advices and information he received from his hosts and companions such as the famous scholar Liq Aṣqu of Gondär or the horseman and independent warrior, Ḫaylu Ṣamru, who 'knew the genealogy of everyone, the history of various 14 Antoine d 'Abbadie 1888. 15 This number is approximate for not all questions have replies, and some replies do not refer to specific questions. 16 I thank the late Gianfranco Fiaccadori for having passed this idea on to me after I presented a paper on Arnauld d'Abbadie's field notes at the Hamburg international conference on Manuscripts and Texts, Languages and Contexts: the Transmission of Knowledge in the Horn of Africa, July 2014. 17 Arnauld d' Abbadie 1980a;1980b;1983;1999. provinces and he was accepted as an authority on the military and political events of the time'. 18

Translation and Edition Methodology
In d'Abbadie's questionnaire, three questions specifically relate to the various ways in which an oath is taken in Ethiopia and the legal value placed on this manner of solemn statement in trials. The first three extracts edited here combine the questions (numbered 80, 200, and 201 respectively) and their corresponding answers. In addition, the fourth text edited here is an extract from another smaller notebook (Città del Vaticano, BAV, d'Abbadie carte 19, fols 10-85, henceforth BAV-ABB-19) that contains lengthier comments on legal issues and records of orally transmitted narrations of trials.
Each extract is first presented in English translation. The French text, edited as close to the original notes as practicable, is annexed at the end of the article. The English translation is divided into numbered paragraphs, to facilitate reading and referencing. The original notes contain many Amharic sentences and words. In the English translation, full sentences in Amharic are spelt in Ethiopic script, 19 followed by their translation, whereas single words are mentioned in brackets.
One of the difficulties occurring in the translation from French to English was the pronominal forms linked to antecedents of indeterminate gender, such as 'plaintiff' or 'party'. I have opted for the 'singular they' widely used in contemporary English, whereas the French text uses masculine singular forms.

Extract No. 1
1.1 -Question no. 80: Can the defendant offer their oath instead of oral evidence? 20 1.2 -They cannot. In general, a person cannot decide alone to swear an oath. It is always the plaintiff who offers this option to the other party either because witnesses are lacking, or because they prefer to do so. After 18 Arnauld d 'Abbadie 1980b, 84. Despite acknowledgement of his information depending mainly on local sources, study of Arnauld d'Abbadie's field notes cannot avoid a critical examination regarding the extent of Eurocentric biases, as discussed below under the last section of this article. 19 Arnauld d'Abbadie's notes are mostly based on Amharic as spoken in Goǧǧam at that time. The rewriting into Ethiopic script is a reconstruction in today's standard Amharic, not exactly as it was actually pronounced in the 1840s. 20 Reference of § 1.1: BAV-ABB-18, fol. 342v, l. 5. having made the other party sworn, the plaintiff cannot revoke it, except in case of murder or land trials. 21 1.3 The preparatory formula for the oath is ዓጼ፡ ይሙት፡ እኔ፡ ስሞት፡ አ ትድረስብኝ፥ ዓጼ፡ ይሙት፡ ከሞትኩ፡ አልደርስብህም።, 'By the emperor's death, if I die don't come to my burial ceremony; by the emperor's death, I can't stand against you anyway after my death'. How could such a statement be contradicted? 22 1.4 In case of murder or even land issue, if evidence can be gathered to invalidate the oath, a parent can be called. It is said, የአልነበረ፡ ያፈርሳል፡ ቀዳ ዳ፡ ያፈሳል፡ ብሎ።, 'The one who was absent cancels everything, as a hole leaks'. If the solemn commitment (ፍጥም፡, fǝṭǝm) has been made (by the plaintiff), it cannot be reverted, but a relative (of the plaintiff) who was absent can take over the case.
1.5 If someone has to swear for land or for blood [crime], it is done by the so-called መደመድ፡, mädämmäd, 23 procedure: the one who takes the oath burns a fire and then provides water to the opposite party who extinguishes the fire completely by saying, ዘር፡ ማንዘሬን፡ እስተ፡ 7 ቤት፡ እንዲ መደመድ፡ እንዲያጠፋኝ።, 'Let me be destroyed and my ancestors up to the seventh generations perish'. 24 Sometimes seven fires are burnt, and in the same way the [opposing] party is asked to swear with six close relatives (brother, son, wife, mother, father, and so on). 21 Reference of § § 1.2-1.10: BAV-ABB-18, fol. 374r, ll. 1-26. 22 Here ዓጼ (ʿaṣe) is translated by 'emperor' to reflect Arnauld's use. See for instance the edition of another text from Arnauld's field notes, Ficquet 2018, 106. 23 This word is listed in Antoine d'Abbadie's Amharic dictionary (Antoine d 'Abbadie 1881, 109-110) under the verb መደመደ፡ defined as 'unifying', 'flattening (a piece of land)', 'levelling'. መደመድ፡ is defined as 'a very serious and meaningful oath (meaning): may God flatten me in the earth if …' Thomas L. Kane adds to the definition of the verb መደመደ፡ 'to clear (by chopping or felling trees), to cut down one after another (trees) […] to raze, destroy utterly' and defines መደመድ፡ as 'form of oath in which the swearer ignites a small bundle of hay, then puts out the fire while swearing' (Kane 1990, 322). Kane also gives an example of a swearing formula: መደመዴን፡ ይጥፉ።, 'may He destroy [all] trace of me [if I lie]!' (ibid.). 24 A double entendre wordplay may be read in this sentence that can also be understood in this way: 'Just like ripped-off amädmado weeds, let my ancestors up to the seventh generations perish'. This hypothesis was confirmed by Ahmed Hassen Omer, whom I thank for his advice. This interpretation is confirmed by the fact that the thin branches of this weed are used for torches (Kane 1990, 323). This form of swearing is also described by Mansfield Parkyns who travelled in Ethiopia in 1843-1846 (Parkyns 1853, II, 257). See below the full quotation.
1.6 For other issues, the oath that is deemed appropriate can be made. If one party swears an oath, the other party has to pay the court costs because they have lost the case.
1.8 Another way is to roll into a palm leaf burial mat (ሰሌን፡, sälen) by saying, እስተ፡ 7 ቀን፡ እንዲጠቀልልኝ።, 'up to seven days until my case is completed'. 25 1.9 It is also said, አሞራ፡  2.4 It is said, ተሺ፡ ምስክር፡ የታቦት፡ እግር።, 'The legs of the tabot (i.e. the priests who carry the altar in procession) 27 are worth more than a thousand witnesses'.
2.5 After having made sworn, it is not possible to call witnesses anymore. 28 25 Parkyns adds more details to the description of this oath in the sälen mat, which is used as a burial cloth (Parkyns 1853, II, 257). See below the full quotation. 26 Reference of § 2.1: BAV-ABB-18, fol. 346r, l. 1. 27 I thank Prof. Shiferaw Bekele for his investigation on the meaning of this proverb and for providing me with this explanation. Another translation of the same proverb is given by Walker 1933, 140, see below for the full quotation.

Extract No. 3
3.1 -Question no. 201: What are the different kinds of oaths, their respective values? 29 3.2 -For land issues, the oath consists of swearing by burning a torch (መደመድ፡, mädämmäd) or having seven relatives swear. For quarrels between husband and wife, excommunication is uttered. For other issues, it is as we like. Generally, it is the plaintiff who demands the oath. 30

Extract No. 4: Additional Observations on Questions 80, 200, and 201
4.1 If one of the parties has taken an oath extrajudicially and then the dispute goes to court, this first oath is null, for it was taken without commitment (ፍጥም፡, fǝṭǝm) or with commitment but without judge. 31 4.2 In pleading their cause one party says, እንካ፡ ትመርምረኝ፥ እሺ።, 'examine me-all right', then they demand the oath (to their adversary). If the oath has been made, and the opposite party brings witnesses to prove that the party who sworn is guilty, the latter has to pay the double court costs and penalty (እዳ፡, ǝda). After the oath called ተመርመር፡, tämärmär (that only exists in Amhara), if the (claimant) party lacks witnesses, they can impose a second oath, which is final and solemn. Before this second oath, the swearing party says, ውቤ፡ ይሙት፡ እንግዴህ፡ ወዲህ፡ ቁርጥ፡ ነው።, 'By Wǝbe's death, 32 isn't it immediately binding?' The other party then makes a solemn commitment (ፍጥም፡, fǝṭǝm). If a party makes this oath, the other will pay double the court costs and a penalty of two times the amount in dispute.
(H2) infamy (ነውር፡, näwǝr) refers to the self (ነፍስ፡, näfs) 34 for killing, for burning house, for secret action (መሥረታ፡, mäśräta). 35 4.4 In the case of a party who used the respectful mode of address (እርሶ፡, ǝrso) with its opponent, if they switch to the ordinary form (አንተ፡, antä) while pleading, and if it is proved by witnesses that they used to say እርስዎ፡, ǝrswo, before the quarrel, they are obliged to return to the polite form (እርሶታ፡, ǝrsota), by solemn commitment (ፍጥም፡, fǝṭǝm) before the judge. Therefore, they must pay the penalty attached to this commitment whenever they say አንተ፡, antä (the familiar form of addressing somebody) after having used እርስዎ፡, ǝrswo (the polite form). To be able to say አንተ፡, antä, somebody has to prove the infamy (ነውር፡, näwǝr) of the person they intend to downgrade. 36 4.5 It is the plaintiff who requires the oath from the defendant. If the defendant refuses to swear, when it is their turn they can challenge the plaintiff to take the oath. Then the plaintiff is obliged to swear. If they refuse to do so, their claim is dismissed and they have to pay the costs and expenses, as well as legal fees equal to the amount in dispute.
4.6 If the party who required the oath does not trust the party who takes it, the former can compel the latter that they associate their unmarried sons or their wife, whether she was married under the regime of community of property (ባለ፡ እኩል፡, balä ǝkkul) or with dowry (ባለ፡ ማጫ፡, balä mạ̌ a). This is in the case of a dispute for cattle or for blood crime. If the dispute is over land issues, the party who requires the oath can choose six men or women among the relatives of the party who will take the oath, altogether seven people can be forced to take the oath in common. In this case out of the house, married or not, they must swear.

Other Depictions of Oath in Studies on Ḥabäša Customary Law
These extracts from the original ethnographic materials gathered in the 1840s by Arnauld d'Abbadie can be completed and commented on via other observations of oath-taking rituals scattered in the few published works dealing with the customary legal practices of Christian Ḥabäša populations of Ethiopia and Eritrea. 37 Further historical perspectives will be then drawn through examination of the rules applicable to this practice in classical canonical texts.
A source closer to the historical and legal context depicted in Arnauld d'Abbadie's notes is the travel account by Mansfield Parkyns, who travelled in the highlands of Tǝgray from 1843 to 1846. He may have met Antoine d'Abbadie whom he mentions as an 'esteemed friend'. 38 However, Parkyns makes no mention of Arnauld, nor is there any mention of Parkyns in Arnauld's account. It is possible that Parkyns met Antoine alone in Tǝgray, while Arnauld was occupied elsewhere, or that the friendship between Parkyns and Antoine d'Abbadie developed later through correspondence, after their return from Ethiopia. Although it seems there was no direct exchange of information between Parkyns and Arnauld, their observations on legal issues are actually very similar. Like Arnauld d'Abbadie, Parkyns was not systematic in keeping a journal, but claims the materials he presents are based on first-hand observations: 'What I have described has been almost entirely what I have myself witnessed, or heard related on the spot'. 39 In Chapter 35 of his second volume, dedicated to observations on political and legal matters, Parkyns has two pages on oath: 'It is not uncommon to settle a dispute about money matters (such as small debts or other affairs), when 37 The cultural designation of ሐበሻ, Ḥabäša, encompasses several regional groups in Ethiopia and Eritrea, mostly Amharic and Tǝgrǝñña speaking. Regarding the Tǝgrǝñña-speaking populations of Eritrea, it is more relevant than the designation of Tǝgray or Tǝgrayan. See 'Ḥabäša', EAe, V (2014), 339a-340b (W. Smidt and É. Ficquet). 38 Parkyns 1853, I, 125. 39 Ibid. no witnesses can be produced, by one party's swearing to the validity of his claim.' 40 Three ways of taking oath are described by Parkyns. The first is the mädämmäd procedure by burning a straw and extinguishing it. This fits quite exactly to the description given by d'Abbadie ( § 1.5) and provides additional ethnographic details (italicized in the quotation): There are many forms of swearing, one of which, considered as very binding, is called 'Medammed'. In taking this oath the swearer lights a little straw, which is placed in his hand on a layer of cow dung. When the straw is well ignited he extinguishes it by pouring water over it, and at the same time expresses a wish that his family may be burnt and blotted out from the face of the earth for seven generations, if he should violate his promise. 41 We can also understand more precisely the whole ritual process of another form of swearing in a palm leaf mat d'Abbadie briefly mentions (see here § 1.7): The most impressive and solemn oath is that which is taken in the church, when, for some important question, a man's opponent requires of him to be sworn in that holy place. The man is taken into the outer circle of the church to the place where the bodies of the dead are laid previously to burial. He is there stretched on one of the mats which […] are used instead of coffins. Lying there, he makes his asseveration at the moment when the sacrament is being distributed and calls upon the Almighty to record it, and to grant as a testimony that, should he have sworn falsely, he may return, after the space of three days, or seven at the most, to the mat on which he is now lying, never to leave it more. 42 In the last forms of swearing depicted by Parkyns, the swearer is confronted to more or less symbolical representations of capital sentence: sword, weapon, or icon of the soldier-saint Saint George: Others swear by the sword. Unsheathing one, they pray that, as surely as it is thus drawn in witness of their word, so surely may the holy Archangel St. Michael draw his to their destruction if that word 40 Parkyns 1853, II, 257. 41 Ibid. 42 Ibid.
should prove false. Similarly, also by a gun or other weapon. Others again by the picture of St. George, placing their hands on his likeness, and calling upon him that should they prove faithless he should direct his lance against them as formerly against the dragon. 43 These ritual acts echo the practice of uttering an oath before the 'icon and the Cross' as mentioned by d'Abbadie ( § 1.6). Finally, Parkyns downplays the importance and solemnity of the oath by noting that many people are reluctant to take it: It is, however, always considered a rather disgraceful action to call thus upon the Lord, or even on his saints, in matters of 'filthy lucre', although, indeed, the cause be a just one, so much so, that many persons possessed of a good reputation, which they are scrupulous of in anywise sullying, in the event of being required to pay an unjust debt, or even an imaginary one, would place the amount in the hands of some trustworthy person to be paid over to the claimant, should he choose to perjure himself, preferring rather to risk their money than to be obliged to swear even to the truth. 44 This passage provides some clues that help clarify the rather elliptic and obscure passage in d'Abbadie's notes where the oath procedure is meant as a safeguard against false accusation, thereby linked to infamy marked by the shift from the polite form of address to the familiar one ( § § 4.3 and 4.4). We will return below to the questions of perjury and infamy codified in the treatises of canon law.
After the period when knowledge about Ethiopia was dominated by travellers' accounts, Italian scholars undertook more systematic studies on the customary law of the Christian populations of Ethiopia and Eritrea in the colonial context. The most prolific of whom, Carlo Conti Rossini, published a general description of the legal practices of the Ḥabäša populations of Eritrea (with some shorter chapters on non-Ḥabäša populations) in 1916. Among his sources he acknowledges he discovered the legal functioning of the kingdom of Gondär (i.e. when the seat of the kingdom was still established in Gondär) from his reading of the notebooks of Antoine d'Abbadie. D'Abbadie's influence on Conti Rossini may be seen, hypothetically, in the latter's description of the interactive process between the plaintiff and the defendant that leads to take an oath, which is presented in very similar way 43 Ibid., [257][258]258. to the analysis given by Arnauld d'Abbadie in the records we have edited here ( § § 1.2, 2.2, 4.5): Next to the witness evidence there is the evidence by oath. This second ritual is stronger than the first. The plaintiff may propose an oath to the defendant, but he cannot impose acceptance of his own oath: this is the defendant who can, in response, ask the plaintiff to take the oath or, on his own initiative, impose it on the plaintiff. But, having chosen the testimonial evidence, and, even more, if the depositions have already begun, the plaintiff no longer has the right to resort to this means. The oath always has a decisive virtue. It is admitted that if witness evidence appears insufficient, the judge can resort to this supreme remedy on his own initiative; but in many regions this is a controversial option. 45 The fact that Conti Rossini mentions that swearing an oath is a 'controversial option' may be in accordance with Parkyns's remark on the reluctance to take the risk of being suspected of perjury and its resulting infamy.
For Conti Rossini the most usual oath-taking ritual takes place in the church by reading the final part of Matthew 25 referring to the Judgement Day and the curse of the sinners, designated as goats and condemned to the everlasting fire of punishment. The oath is uttered by touching the corresponding page in the manuscript, 46 it is then repeated before each priest and concluded ritually by kissing their portable hand crosses. Besides this common oath ritual, Conti Rossini also gives a description of the oath taken for more serious cases by laying and covering the swearer in a shroud, corresponding to the palm leaf mat described in detail by Parkyns (see above) and alluded to in d'Abbadie's notes ( § 1.7): In the other type of religious oath, the intervention of the clergy is not necessary, and on the contrary it can be omitted to not worsen even more with it the responsibility of a perjury. Whoever has to swear is covered with the cloth used for funerals, then laid on the ground on four pieces of wood arranged in a cross and, in the presence of the judge's delegate, after the threats of excommunication and 45 Conti Rossini 1916, 511 (my translation). 46 The specific social use of the manuscript of the Gospels could be an interesting indication for codicological studies, by checking whether this passage shows wear marks or is highlighted by marginal annotations. perdition, with solemn rituals, is interrogated by the opposing party with a kind of form. 47 Another important contribution to the knowledge of the traditional legal system of the Ḥabäša was made by the Italian colonial officer and ethnographer Alberto Pollera. In 1913, he published the first study on the legal system and the customary procedure in Ethiopia and Eritrea, followed in 1940 by a posthumous book entitled L'Abissinia di ieri: osservazioni e ricordi ('Yesterday's Abyssinia: observations and memories'), to capture the traditional functioning of society before the reforms imposed on it by Italian imperial rule. 48 His description of the common oath-taking ritual in the church compound is very similar to Conti Rossini's. He also provides a description of the swearing ritual when covered with a burial cloth, which is combined with swearing by fire and water. According to Pollera this ritual is performed in the woods surrounding the churches that are inhabited by terrifying demons (ganen): The one who must swear is lying on the ground, covered by the blanket used for the dead and subjected to a sermon of circumstance by the priest who, at the end of this, presents him with a lighted candle inviting him to turn off with his breath repeating the formula: 'If what I have affirmed is false, or if I fail to live up to the promise I made, my life should be extinguished in sin as the flame of this candle is extinguished'. In the same way, the priest presents him with a bowl full of water, which the swearer must overthrow to the ground repeating the words: 'Let my kin be lost to the seventh degree; let their blood be absorbed by the earth, as this water disappears on the ground '. 49 In The Abyssinian at Home, published in 1933, Craven H. Walker, the British Consul at Gore, in western Ethiopia, from 1911 to 1928, gathered ethnographic observations on the rules of daily life based on the 'translation of Amharic notes, which are the statements of natives taken down in their actual speech'. 50 In the chapter on legal procedures, he describes the process that leads to making an oath if witnesses are lacking or if the accused party rejects the accuser's witnesses: 47 Conti Rossini 1916, 512-513 (my translation). 48 Pollera 1913;1940. 49 Pollera 1940, 124 (my translation). 50 Walker 1933, iii. If there is only one witness who saw or heard, he is sufficient, provided that the other party accepts him; but if the other party fears false witness and refuses him, the matter must be dismissed or they will be sent to the church to swear. So, too, if the accused counts out for himself two witnesses only, the accuser may cry, 'Than a thousand witnesses I have preferred the foot of the shrine-thou thyself!' 51 and the judge must send them to the church. 52 Walker goes on to describe a typical interaction between the plaintiff and defendant that precedes the formal act of the oath: If the accuser has no witness, he will say to the accused, 'Thou thyself (art my witness)!' and the other will ask, 'Who is to die?'-'Let Menilik die!' he the accused will cry again, 'Swear for me!' and the accuser will say 'Agreed! It (thy word) reached me. Let Menilik die! I will swear!' Which is the 80 bond, 53 and the judge will appoint a commissioner with four jurors to go with them to the church on a Sunday or on a day of festival to make the oath. 54 Similar formulas, preparatory to the act of swearing, are also found in d'Abbadie's notes, for which the commitments to take oath are made either on the death of the king (ʿaṣe) ( § 1.3) or on the death of DäǧǧazmǎWǝbe ( § 4.2) who was the most powerful regional leader of the time. 55 This way of certifying one's word by referring to the death of the highest authority is a common and well-established form of verbal commitment. Through this 'death sentence' carried by the oath, not only the life of the swearing party is at stake, but the whole legal order. This formula can be repeated many times between the accusing party and the accused. It is called ፍጥም፡, fǝṭǝm, with a semantic connotation of accomplishment and perfection as its binding character. Stanley Fisher, in his overview of the 'traditional criminal procedure', published in 1971, calls it a testamentary oath for it is 'employed by witnesses and the parties themselves to validate their evidence'. 56 51 Thus, exactly the same formula as that recorded by d'Abbadie ( § 2.4). 52 Walker 1933, 140. 53 The '80 bond' is not defined by Walker. It corresponds to ሰማንያ, sämanya, which is a strong and solemn commitment that involves a payment of 80 silver coins (bǝrr). 54 Walker 1933, 140. 55 For the expressions in oath, ንጉሥ፡ ይሙት፡, nǝguś yǝmut, and references to the invocation by the king in ancient sources (from South Arabia, Babylon, or from the Psalms), see Caquot 1957, 215. 56 Fisher 1971 The religious dimension of the oath-taking procedure calls for further comment. In the judiciary process, the taking of an oath marks a kind of boundary line between the civil and temporal authority of the judges and the spiritual responsibility of the priests in charge of preparing the faithful to the eternity after life. In the absence of sufficient evidence for the exercise of temporal justice, the oath option defers the decision and punishment to the trial of divine justice, deemed inescapable. The association and separation between the two realms have moved with time. In the observations recorded by d'Abbadie, the oath is fully part of the judiciary process: it is null if taken without the presence of a judge ( § 4.1) and involves the payment of double penalty fees by the non-swearing party ( § 4.2). There are even forms of oath that, apparently, did not require the presence of priests or the use of religious objects such as the mädämmäd ritual performed by fire and water. According to those observations of the early twentieth century (Conti Rossini, Pollera, and Walker) the oath-taking ceremony was usually held in the church compound. This may be an indication of transformations in the central codifications of justice (in colonial Eritrean or imperial Ethiopian contexts). The practice of taking oath had apparently shifted from a mixed civil-religious ground, to a more exclusive religious setting as a mode of conflict resolution before the case was brought to court. This hypothesis is supported by a recent ethnographic study by the anthropologist Andrea Nicolas in an Orthodox Christian Oromo community in the area of Bishoftuu in the East Šäwa zone of Oromia. She clearly demonstrates that the oath is not a matter of formal justice, but is a way of resolving an accusation made without enough evidence: Often it is sufficient, in a case of contradictory pleadings, to make both come to the church compound and to let them tell their stories again on holy ground. […] If an accused person still denies his guilt at church, the elders may request him to formally take an oath (qalä mäḥalla) of innocence.
[…] Particularly if the victim has no proof of his opponent's guilt that would legitimise a direct sanction against him, he may visit the soul father of the person under suspect and say to him: 'Qalä mäḥalla yǝfä emǝllǝňň!' ('He shall undergo the word of oath for me!'). The priest would go afterwards to his suspected soul child and confidentially ask him whether he had done it or not. If the suspected denied, the soul father would ask him to come to the church and to swear the oath of innocence there. […] If the suspected does not come on the appointed day, his guilt is taken for granted, but if he comes and swears that he is innocent, he may free himself from the accusation. […] This oath is then taken as the proof of his innocence and its accomplishment is communicated to the others so as to take the burden of suspicion from him. If he has sworn wrong, he will be subject to severe divine punishment, which is no longer in the hands of men. 57 In the analysis of constant patterns and variations in the practice of oathtaking, the original observations recorded by d'Abbadie in the midnineteenth century are advantageous for providing samples of oral statements directly in Amharic. In other published records, statements are usually translated and reduced to a proverbial-like rigid form. Standard formulas are part of the procedure, but do not fully account for the rhetoric variations that can be performed by the actors of a trial. In addition to general formulas, d'Abbadie adds two long oath-taking statements. The first ( § 1.9) compares the swearer with a vulture who, in case of perjury, is shot by a gunman (metaphor for the judges?) and then plundered by the ants (for the popular vindict?). The second ( § 1.10), instead of swearing on the Cross, or religious objects, proposes a worldly version of the oath by swearing on the different parts of the plough, which are essential to the survival of the farmer. These technical words are listed by two series of three, which may also be interpreted as an echo of a trinitarian Christian pattern.
Besides these indications of rhetoric variations in the manners of uttering an oath, there are several aspects of the legal procedure recorded in d'Abbadie's notes that are not found in other studies and that would deserve further analysis: -The party who accepts to take the oath, usually the defendant, wins the trial and double penalty fees must be paid by the opposing party ( § 4.2).
-A kind of double bind relation is established between the two parties: the accused party can refuse to take the oath when requested by the accuser. In return the accuser can be asked to swear, and they lose the case if they refuse ( § 4.5).
-A plaintiff cannot revoke an oath that has been taken on their request. If new evidence can be adduced to invalidate the oath, a parent can take over the case ( § 1.4).
-The most solemn forms of oath, for serious cases, can be taken collectively, by the defendant and six of their relatives ( § § 1.5, 2.3).
-By noting that the procedure called ተመርመር፡, tämärmär, in which a party requests to be submitted to the 'test' of the oath, is followed only in Amhara areas ( § 4.2), an indication is given on regional variations of the practice, despite the strong normative cohesiveness of legal and ritual codifications in which religious authorities are involved, whereas the rules under the control of local civil authorities are more subject to local differences.

Warnings and Instructions Regarding the Oath in the Written Legal Framework
Beyond procedural facts based on ethnographic observations, understanding of the legal issues at stake in the oath-taking procedure in Christian Ethiopia requires a focused investigation into the historical literary sources of the legal system. Resorting to the oath in the procedures of the customary law of Christian Ethiopians, is authorized and regulated by the written code of law known as the Fǝtḥa nägäśt, that is, the 'Law of the kings'. Believed to have been composed by the 318 Fathers of Nicaea, this text is a Gǝʿǝz translation of a thirteenth-century Arabic compilation of canon law translated from Greek for the use of the Egyptian Coptic Church. Since the fifteenth or sixteenth century, this book was adopted by the Ethiopian monarchy as a normative reference. 58 It is usually referred to as the supreme law, for it was ritually invoked as the central reference for judgments rendered by the king, as the last instance of appeal. There is discussion, however, over the real or presumed status of this text as a reference because of its obscure and impracticable nature. For Manfred Kropp, 'the translation [from the Arabic] was to a large degree unintelligible and required an Amharic commentary, which was orally transmitted and full of highly inventive interpretations'. 59 The same author sharply asserts, 'no, this compilation has nothing to do with Ethiopian law'. 60 The extent of the normative effect of the Fǝtḥa nägäśt, and other written sources, 61 on local legal systems can be questioned through the case of oath. Chapter 43 of the Fǝtḥa nägäśt deals with the judge and his relations with 58 See the Introduction to the edition and translation of the Fǝtḥa nägäśt by Paulos Tzadua and Peter L. Strauss and the historical study by Aberra Jembere; cf. Paulos Tzadua and Strauss 1968;Aberra Jembere 2000. 59 Kropp 2012, 259. 60 Ibid. 61 Another important written source for the legal history of Christian Ethiopia is the Śǝrʿatä mängǝśt, 'Order of the kingdom' defining the functioning and ritual rules of the royal court, including some judicial matters. I have not found any particular provision referring to the oath in the available edition of this text (Varenbergh 1915(Varenbergh -1916; note that the English translation of this book provided by Bairu Tafla and Scholler 1976 is a helpful introduction, but not an edition).
witnesses. The fifth section of this chapter focuses on the practice of the oath. The main point of discussion is to resolve a fundamental contradiction between the formal prohibition against swearing stated explicitly in the New Testament, and the very frequent use of this practice in the Old Testament: The ancient Law forbade swearing falsely and provided for the punishment of one who swore falsely; but in the new Law, an oath, according to the law of reason, should not be uttered, pursuant to what our Lord said in the Gospel: 'It was said to yours of old: "You shall not forswear yourselves," but I say to you not to swear at all.' 62 Between theological exegesis and legal pragmatism, this contradiction is addressed by a general theory on the discourse of a wider extension than the sole speech act of swearing. Any kind of statement must be uttered only when it is essential or necessary: Though Our Lord said in the Gospel: 'You shall not swear at all,' 63 He did not say [do not swear] in lawsuits, because all He said with respect to the proper way of speaking is: 'Let your speech be: "Yes, yes," or "No, no,"' and added: 'And that which is added over and above this comes of evil.' 64 The extra talk is unnecessary, while taking an oath during a lawsuit is essential. But it is proper that a man avoid taking an oath if he can, either by paying his money or without paying anything. 65 Following this textual commentary, the oath must be taken as a last resort, when no other option remains. These precautions explain that the oath cannot be taken on the defendant's own initiative, but through the request of the judge when no evidence can be obtained from witnesses: If the accused keeps silent, neither admitting nor denying the charge, the judge, to get an answer from him, shall say to him again: 'Answer! Otherwise you must take an oath, or produce a guarantor for the object you are accused [of having].' […] If he delays past the time [given] to produce the witnesses, he must either take an oath or confess. 66 The restriction on swearing spontaneously, in order to keep the oath as an extremely serious statement and as proof of innocence, is reinforced by a double bind mechanism between the accuser and the accused: the challenge of taking an oath can be turned by the accused against the accuser, as the closing act of the case. As mentioned earlier, this reciprocal dimension was noticed by d'Abbadie ( § 4.5), but no other observer after him mentioned this reversibility of the person who is required to swear. This is confirmed in the written law by a passage of the Fǝtḥa nägäśt: [I]f the accused denies the charge, the judge shall say to the accuser, 'Have you witnesses?' If the accuser says 'No,' he shall say in his turn to the accused 'Will you swear?' If the accused is afraid to swear and turns back the oath to his accuser, and if then the accuser takes the oath, the object shall be restored to him. 67 Beside these general principles, the Fǝtḥa nägäśt provides little empirical detail on how the oath is to be taken. This is left to the customary legal customs that are transmitted and regulated mostly by word of mouth.

Potential and Limits of the Comparison to Systems of Proof in Western European Medieval Law: the Case of Compurgation
The contradiction between the prohibition of oath in the New Testament and its actual persistence in the legal customs has been addressed since the early times of Christian thought. As well as for the original source of the Fǝtḥa nägäśt, the evangelical prohibition against swearing was reinterpreted and overturned in European medieval canon law, in particular through Gratian's Decretum compiled in the twelfth century. Admitting the practice of the oath within the framework of the legal interpretation of sacred texts was considered less damaging for the social order than letting it survive in the normative margins as a popular customary practice. The practice of taking the oath was admitted as a valid proof of innocence, 68 particularly to exonerate oneself against infamia, that is false accusation that could not be supported by witnesses.
The procedure of the oath taken with the support of 'oath-helpers' was known as 'compurgation': [A] defendant in a criminal matter against whom there was public suspicion, or a defendant in a civil case against whom a proper com-plaint had been made, would initially be summoned to appear before a judge. Upon the defendant's appearance, if he denied the allegation made against him, the judge made an award requiring him to swear an oath of his innocence and also to find and produce an assigned number of compurgators, more commonly known as 'oath helpers' in the parlance of the common law. These were men or women who were willing to support his oath by one of their own. 69 In England, compurgation was long practised as the main procedure for ascertaining guilt or innocence, until the eighteenth century. As put by the anthropologist William A. Shack: Apparently all that the court required of the defendants was to bring forth oath-helpers, who swore that the oath taken was pure and not false as the plaintiffs maintained. […] [W]hen the system of compurgation had been brought to its fullest development it became the basis for almost all judicial procedure in litigation. […] [I]n the law of Anglo-Saxon England, the oath was a vital sanction in public and legal life, and oathtaking was an integral duty in the social organization of the kindred. Oath-worthiness was requisite to 'normal' membership in society, for without this claim a man did not hold full 'folkright' in the community. 70 These few samples of description of the practice of collective oathing, or compurgation, in pre-modern European legal systems indicate parallels with the practice described by d'Abbadie of taking the oath collectively with close relatives as an option for the most serious cases (see § § 1.5 and 3.2 in our edition above). Beside this single observation, however, the extent to which compurgation was actually practised in Ethiopian Christian customary law cannot be confirmed by other studies.
This last remark leads to question whether d'Abbadie's records on oath statements and other legal procedures in Ethiopia were, to some extent, 69 Helmholz 2001, 83. 70 Shack 1979. The American anthropologist William A. Shack is known in Ethiopian studies for his study of the Gurage (fieldwork in 1958Gurage (fieldwork in -1959Gurage (fieldwork in , monograph published in 1966cf. Shack 1966). He authored the quoted essay on collective oath or compurgation in England, with comparison to African states, after a suggestion by Max Gluckman (Shack 1979). Shack's papers were discovered late in the process of finalizing this article. The views he developed open stimulating perspectives for a truly decompartmentalized comparative historical anthropology of law. The present article aims to be consistent with this kind of approach.
Eurocentrically biased. Due to his upbring, he grew up in different institutional and legal settings, in Ireland, England, and France. In France, as a young aristocrat who was to inherit and manage a large estate, he was well instructed about the coexistence and contradictions between the central regulations of the State, and the local rules of the Basque country, on the borders with Spain. In his interviews on customary law, he could have influenced his Ethiopian informants by asking them whether some manners of taking the oath in Europe were also practised in Ethiopia. In return, the informants might have been willing to confirm this was also the case, for instance by more or less consciously accepting the identification of common features between the legal systems of Ethiopia and Europe, with a view to consolidating individual friendship ties with the foreign traveller. This is a common bias in the process of ethnographic fieldwork, particularly when the ethnographer relies on 'privileged informants' who become 'attuned to his interests'. 71 The frequent comparisons with Roman or medieval history in the travel account published by Arnauld d'Abbadie in 1868 would tend to reinforce this hypothesis of a Eurocentric bias. For instance, after a long passage on the relevance of the concept of feudalism applied to the structures of the Ethiopian Christian kingdom, he says, Ethiopians are unaware of the historical existence of the Conscript Fathers of Rome as well as that of other patricians whose various denominations were more or less related to the word Father, who led the destinies of so many nations in Europe.
[…] Nevertheless they consider power or its representative not as a winner, as an enemy with a distinct interest, but as the summary of the interests of society and the highest political consecration of paternity. 72 By stating that Ethiopians have no knowledge of ancient Roman rules, d'Abbadie's intention is to highlight the independent political evolution of Christian Ethiopian societies and to disconnect them from the influence of 'theological misleading ideas'. 73 Nevertheless, his reasoning shows that his observations were at least implicitly based on an evaluation of the degree of resemblance and proximity between Ethiopian Christian society and Greco-Roman civilizational standards, as most observers did at that time (and still do, more or less implicitly, today). 71 Clifford 1983, 124. 72 Arnauld d' Abbadie 1980a, 112 (my translation There are some reasons, however, to minimize the extension of this presumed Eurocentric bias in d'Abbadie's notes. First, his book was published twenty years after he had completed field research. The analogies between some aspects of Ethiopian society and European history were worked out after a long period of maturation. This is evidenced by the multiple layers of correction and rewriting to be seen in the handwritten versions of his travel narrative. His first-hand information is not quoted as such but has been adapted and reinterpreted to suit the readership of the time. Secondly, the fact that field notes are jotted down directly in Amharic, with very few references to European legal concepts, tends to indicate that his information was genuine and the level of personal interpretation in his record was limited. Lastly, it has also been seen that most of the d'Abbadie's field notes were converging with observations made by other authors in different locations and at different periods.
After this critical detour, we can assume that d'Abbadie's records on the fact that 'the [opposing] party is asked to swear with six close relatives' are reliable and that forms of compurgation or collective oath were most likely practised in the customary law of Christian Ḥabäša Ethiopian societies. Confirmation of this practice and its extension would require deeper investigations either into the available records or into present-day local legal traditions, through their preserved practices or their memory. As the option of calling upon oath-helpers is not found in the written provisions of the Fǝtḥa nägäśt, its use in Ethiopian legal contexts-if confirmed-raises questions as to whether it arose from internal processes independent of remote influences, or was transmitted into the Christian Ethiopian legal system from sources of law that are still to be identified.

Interactions between Literacy and Orality
Another striking fact in d'Abbadie's records is the detailed transcription of the oral performance of oathing in Amharic. Some emphatic variants in the ways of pledging an oath (particularly § § 1.9 and 1.10) are much more elaborated rhetorically than shorter normative formulas recorded by later observers. This is an indication of cultural transformations oscillating between the enforcement of standard codifications and local processes of adaptation and creative reinterpretation. Part of the oath utterance as a legal action must be accomplished through precisely codified sentences, gestures, and attitudes. Their conformity is checked by institutional frameworks in order to acknowledge a proof value to the statement that is certified by the oath. In the Christian Ethiopian legal context these formal rules refer to principles preserved in the written books of law. Provided they are respected, a certain degree of improvisational skill or freedom is allowed in the verbal enunciation and ritual action, that can be transmitted through orality and gradually distort and transform established norms as far as setting standards recognized on a local scale.
Much advancement in textual scholarship is to be anticipated from the study of orality and how it may open up new ways of reading written sources. There is a spoken life of texts to be taken into account, that is, the processes of their oral activation from eyes to mouth and transmission from mouth to ears, until the loop is closed by the fixation of spoken words by the writing hand. At a lower level than the 'upper law' of the royal institutions, Smidt highlighted the existence of books of local laws, particularly in the northern regions. These books are written on the basis of orally transmitted and performed customary legal practices. The historical depth and regional extension of this local legal literature-on oath as well as many other aspects-would deserve deeper investigation. 74 The identification of oral layers in processes of textual transmission has been promoted by Manfred Kropp against a positivist and textualist orientation deeply rooted in the field of classical Ethiopian studies, by paying attention to 'low-intensity signals' (my expression), which form a network of hypothetical links conducive to gathering textual proofs that may go unnoticed. In this perspective, Kropp highlighted the 'spoken origin' of the text called Śǝrʿatä mängǝśt, that is, 'the order of the kingdom', which is a collection of rules of the royal court. 75 This text was recorded in Gǝʿǝz, presumably since the fifteenth century, to describe ceremonials that were held in Amharic, also known as lǝssanä nǝguś, the 'king's tongue'. The Gǝʿǝz text is quite unintelligible as it is written in a compact and elusive form, which is not self-explanatory. Its disambiguation and authoritativeness on practical issues used to be activated by way of interpretation through debates between clerics, who are both the holders of orally-transmitted knowledge and producers of new interpretations in case of disagreement. By studying Amharic translations that he initially considered as secondary by-products, Kropp noticed some hints of interpretations transmitted by word of mouth that made more practical sense than the elliptic statements in the Gǝʿǝz text.
Such processes of interaction between written legal frameworks, their spoken origin, and their applications through orally-transmitted interpretations and rituals may have played a role in the codification of oath, which is 74 'Law and judiciary: Traditional legal institutions', EAe, III (2007), 513a-516a (W. Smidt). 75 Kropp 2012. See also n. 61 above. strikingly stable in its performative structure over time and space in the Christian Ethiopian context, and maybe more widely. Some procedural details captured by d'Abbadie and later ethnographic observations, such as the case of compurgation examined above, or other aspects like the ritual simulating the burial of the oath-taker, can be seen as indications of patterns that may have been conveyed into the Christian Ethiopian customary law through sources yet unidentified. These were possibly of mixed nature between textual and oral ways of transmission, such as the Amharic translations and commentaries (tǝrg w ame) of the Fǝtḥa nägäśt code of law that are still awaiting in-depth study.

Conclusion
Through the edition of some original materials based on observations made in Ethiopia in the 1840s by Arnauld d'Abbadie, supplemented by a critical review of later sources, this article has been seeking to provide a consolidated picture of the manners of taking oath and their variations in the history of customary legal procedures in Ethiopian Christian societies since the mid-nineteenth century.
The speech acts and rituals involved in the performance of swearing an oath provide some grounds for showing that procedures of customary law were not isolated local traditions, for they could incorporate procedural elements refracted from legal frameworks of a wider scale. In the Ethiopian case, many aspects of oath-taking procedures can be related to a wide set of practices attested to historically in other Christian societies. Their transmission into local frameworks of customary law in Ethiopia seems to have been subject to little variation.
As the oath establishes a bond between mundane laws and divine justice, it can be assumed that this practice was codified and strictly regulated by the Church in order to avoid the resurgence of controversies and conflicting interpretations already addressed and resolved in canonical writings. However, these interactions between the ordinary skills practised at the level of local customary law and the church legal experts cannot be explained by referring to the most diffused official written compendium of canonical laws in Ethiopia, the 'royal code' titled Fǝtḥa nägäśt. Many details in the Ethiopian oath-taking procedures are not given in the Fǝtḥa nägäśt that provides only generic prescriptions. Other sources must have been used to codify and regulate the use of oaths as proof of innocence.