Imploring God and the “Living Tradition”: A Relative Chronology of Epigraphic and Traditional Invocations

By Mathieu Tillier

This is part two in a series of four posts on the historical formation of the Sunna, with a focus on methodological reflections on the emergence of Prophetic authority.

Stating that the sunna of the Prophet represents a major source of classical Islamic law may appear as self-evident. Many legal rulings are supposedly rooted in the words and deeds of Muḥammad, which Muslims considered not only exemplary, but also invested with a normative value.[1] The emergence of the Prophet’s authority, however, has a history, which includes the formation of the ahl al-ḥadīth movement in the second half of the second/eighth century, followed by the success of their methods in the third/ninth century. This movement eventually resulted in stressing the exclusive importance of authoritative sayings attributed to the Prophet Muḥammad. The affirmation of a universal type of prophetic authority, superior in value to the opinions of local post-prophetic figures, had a major impact on legal hermeneutics. This phenomenon, which researchers call “traditionalization” of fiqh, led in the third/ninth century to the development Islamic legal schools that formally relied to a large extent on prophetic authority.[2] In order to understand the early stages of the formation of Islamic law, tracing the historical emergence of authoritative sayings (ḥadīth), and understanding the mechanisms by which specific types of dicta emerged in early Islam, appear therefore as crucial issues.

Yet, ḥadīth dating methodologies, on which depends the reconstitution of this historical process, have so far led to circular reasoning. This realization led me, in my last post, to propose an experimental method based on a comparison between early inscriptions on rocks and normative sayings collected in ḥadīth books, focusing on invocatory inscriptions (duʿāʾ). At first sight, the issue of duʿāʾ might appear distant from legal matters, insofar as it never became a major subject of fiqh – although it did make its way into legal chapters. Addressing God, even in a more informal way than in ritual prayer, responded nonetheless to specific rules. Muslims who engraved invocations on rocks used ready-made formulae, following social and/or religious norms. Addressing supplications to God was thus regulated by what Joseph Schacht called “living tradition,” that is, generally accepted customs or practices.[3]  For that reason, however minimal the interest of classical jurists in this subject, the relationship between early practices of duʿāʾ and its theory as recorded in ḥadīth books may reveal more general dynamics concerning the production of norms and legal rulings in Islam. How did the living tradition of epigraphic invocations relate to authoritative sayings collected and recorded in ḥadīth works? In order to answer that question, I systematically compared invocatory formulae found on rocks with a corpus of 26 collections of ḥadīth composed up to the sixth/twelfth century.[4] In what follows, I shall refer exclusively to the Hegira calendar, which provides homogenous one-century periods that correspond well with the history of ḥadīth. For the record, the first, second, and third centuries AH respectively match the following timespans: 622-718 CE, 718-815 CE, and 815-912 CE.

Let us first look at the epigraphic corpus of dated invocations. The number of inscriptions containing an invocation to God increased dramatically after the first century AH. Simultaneously, the themes of supplications diversified considerably. During the first century, supplications consisted mainly of individual requests for God’s forgiveness or blessing for the speaker,[5] and, more rarely, for Jesus or for Muḥammad – but only from the end of the first century onwards, especially in the inscriptions of the Dome of the Rock.[6] During the second and third centuries, supplications became more diverse, with the introduction of explicit prayers for eternal salvation or spiritual guidance, requests to die as a martyr or to be exempted from the tortures of the grave.[7] Invocations concerning Muḥammad also flourished at that time.[8] In addition, it is worth noting the cumulative nature of these themes. New topics and formulae did not erase earlier ones, but were rather added to them. Thus, alongside new themes, individual requests for forgiveness continued to be expressed on the rock in the second and third centuries AH. Finally, and perhaps most importantly, the formulae (i.e. words and phrases) for expressing supplications to God multiplied in an exponential manner. Between the end of the first and the end of the third century AH, the number of formulae increased sixfold. These results reveal a major development in the invocatory genre through the second and third centuries AH.

If we now place epigraphic invocatory formulae side by side with formulae found in collections of prophetic and non-prophetic sayings (ḥadīth), what do these two corpora have in common? We notice that between 49% and 75% of epigraphic formulae find a parallel in ḥadīth, which suggests a strong correlation between inscriptions and ḥadīth (see fig. 2).[9] This result could at first sight agree with the results of the sanguine approach of ḥadīth history.[10] One could indeed conclude that words historically uttered by religious authorities served as a model, through ḥadīth reports, for the inscriptions that survived on rocks. This would indicate that the repertoire of authoritative sayings found in ḥadīth collections is very ancient indeed, as claimed by traditional scholars.

However, this interpretation immediately raises a serious problem. If ancient sayings uttered by religious authorities served as prototypes, why did the themes of invocations multiply over time, as well as formulae that were used to convey them? Even if we consider that many formulae remained only oral (and were never recorded on stone) in the first century AH, we should expect, if these formulae actually went back to the first-century authorities mentioned in ḥadīth literature, to find all these formulae popping up on rocks at the same time, that is during the lifetime of these authorities or immediately after. However, innovation in the creation of new formulae was not only continuous, but it also increased over time and even accelerated in the third century AH, at a time when religious authority was increasingly restricted to the Prophet and excluded contemporary figures. One could adduce historical arguments to explain this phenomenon, such as the fact that words heard or uttered by Companions may have reached the provinces of the Islamic empire only gradually, due to the geographical distance between Medina and the rest of the empire. However, this would mean that only a very small number of Companions used these formulae (i.e. not all those who participated in the conquests), and it would also imply the existence of an intellectual and linguistic gap between the elite of the Companions and other Muslims. We must therefore reject the hypothesis that epigraphic formulae that appeared in the second and third centuries AH were inspired by sayings from the first century AH, and thus reject the positivist conclusion that ḥadīth constituted an authoritative corpus early on. The alternative conclusion is that, if an invocatory formula does not appear on rocks at a certain time T, it means that this formula was unusual or unknown at that time T, and that no corresponding model existed in ḥadīth at that same time T. This conclusion suggests, with some nuances to which I shall return, that epigraphic invocations and invocatory ḥadīths that included both the same formulae must have appeared at about the same time, and therefore that the dates of the inscriptions match, more or less, the time when the corresponding models found in ḥadīth literature came into circulation.

Moreover, one can observe that the proportion of epigraphic formulae that find a parallel in ḥadīth gradually diminished (fig. 2). One must conclude that invocatory formulae on rocks were not necessarily influenced by ḥadīth. On the contrary, it seems more likely that the creation of new invocatory formulae, of which traces can be found on rocks, preceded their consecration in the form of ḥadīth. Finally, many formulae attested on rocks do not appear in ḥadīth literature, which indicates that ḥadīth literature only canonized (or abstained from canonizing) formulae that had spread at an earlier stage.

Yet, these conclusions should be qualified according to time periods:

(1) In the first century AH, when invocatory formulae on rocks agree with the same formulae as canonized in ḥadīth literature, a historical link between the two may be considered. Since these formulae are attested on rocks as early as the first century AH, they are likely to have been uttered by the authorities of the same period. For example, the frequent occurrence of the phrase Allāhumma ighfir (“O God forgive”) in first century inscriptions suggests that Muḥammad could historically have uttered these very same words. These early formulae attested on rocks may therefore represent the earliest “core” of invocatory ḥadīths. This historical “core,” which have been desperately sought after by modern scholars, would therefore be lexical in nature. Nevertheless, this convergence does not allow us to ascertain that corresponding ḥadīths spread as early as the first century. Because of the cumulative character of formulae that I mentioned earlier, ancient formulae were still engraved on rocks during the following centuries. A ḥadīth that quotes the formula Allāhumma ighfir could therefore have been put into circulation in the second or third century, simply because Muslims continued to use this expression.

(2) In the second century AH, as ḥadīth began to develop in scholarly circles, one cannot exclude that some authoritative sayings influenced epigraphic fashions. Nevertheless, the continuous invention of new formulae in epigraphic invocations calls for caution, and suggests that an invocatory genre continued to develop independently from ḥadīth.

(3) In the third century AH, however, the relationship between inscriptions and ḥadīth became increasingly sophisticated. Some formulae appeared on rocks after they had been first attested in large ḥadīth collections. This suggests that from that time onwards, ḥadīth may have provided invocatory models and thus preceded the epigraphic diffusion of some formulae. Nevertheless, many formulae that did not match any ḥadīth model continued to appear, which shows that the innovation process continued in an autonomous manner, without being hindered by the emergence of ḥadīth or by its canonization.

At this point, our comparison between epigraphic and ḥadīth corpora relating to invocations shows that some formulae mentioned in ḥadīth literature (prophetic and non-prophetic) actually date back to first/seventh-century historical practices, and could plausibly have been uttered by Muḥammad. However, the number of such formulae is very small. The vast majority of invocatory formulae appeared at a later stage, and multiplied in an exponential manner in the second/eighth and third/ninth centuries, which suggests that their appearance in ḥadīth collections is anachronistic. Rather, one should conclude that traditions employing these formulae were put into circulation around the time of their first dated appearance on stone.

What can these results tell us about the history of Islamic law? First, they reveal that, when a comparison between documentary (epigraphic) and literary (ḥadīth) sources is possible, a lexical convergence between first-century practices and ḥadīth may occasionally appear. However, this convergence is extremely limited, as most of the formulae endorsed by ḥadīth reports were only engraved on rocks from the second/eighth century onwards. If one extends that observation to other normative fields represented more in fiqh literature, it suggests that a tiny fraction of ḥadīths employed by Muslim jurists to infer rules or defend norms they adhered to may have resorted to archaic expressions historically in use in the first/seventh century. This concordance, in addition to being limited, does not allow us to infer at this stage that the Prophet historically pronounced the rules associated with such archaic formulae, nor, more importantly, that his words were faithfully transmitted. The preservation of an archaic lexical core within certain formulae could indeed turn out to be an illusion created by the conservatism of the first generations of Muslims. It is therefore necessary to remain cautious and confine ourselves to assuming the historically realistic character of a limited part of the vocabulary used by certain legal traditions.

The antiquity of legal traditions, namely that they date back to the time of the Prophet, could be all the more doubtful since the preceding results suggest that in the second/eighth century, practices still generally preceded their theorization in the form of authoritative sayings. This may therefore confirm Joseph Schacht’s hypothesis that the formation of ḥadīth primarily consisted in projecting practices that had become accepted as normative onto past authorities.[11]  In other words, and insofar as my findings regarding the very narrow field of duʿāʾ may be generalizable, the expression of legal rules in the form of raʾy and prophetic ḥadīth, and the subsequent formation of Islamic law, merely reacted to practices emerging from the living tradition – to either endorse or, on the contrary, condemn them.


[1] See for example al-Shāfiʿī, al-Risāla, ed. Aḥmad Muḥammad Shākir (Cairo: Maktabat dār al-turāth, 1979), 85-103 (#282-307).

[2] See Christopher Melchert, “Traditionist-Jurisprudents and the Framing of Islamic Law,” Islamic Law and Society 8 (2001): 383-406.

[3] On living tradition, see Joseph Schacht, The Origins of Muhammadan Jurisprudence (Oxford: Clarendon Press, 1950), 58.

[4] The most important of these 26 works are: ʿAbd al-Razzāq al-Ṣanʿānī, Muṣannaf ʿAbd al-Razzāq, ed. Ḥabīb al-Raḥmān al-Aʿẓamī (Beirut: al-Maktab al-islāmī, 1983); Abū Dāʾūd, Sunan Abī Dāʾūd, ed. Shuʿayb al-Arnaʾūṭ et Muḥammad Kāmil Qurra Balalī (Damascus: Dār al-risāla al-ʿālamiyya, 2009); Aḥmad b. Ḥanbal, Musnad al-imām Aḥmad b. Ḥanbal, ed. Shuʿayb al-Arnaʾūṭ and ʿĀdil Murshid (Beirut: Muʾassasat al-risāla, 1997); al-Bayhaqī, al-Daʿawāt al-kabīr, ed. Badr b. ʿAbd Allāh al-Badr (Kuwayt: Ghirās, 2009); al-Bukhārī, Ṣaḥīḥ al-Bukhārī, ed. Muṣṭafā Dīb al-Bughā (Damascus-Beirut: Dār Ibn Kathīr-al-Yamāma, 1987); Ibn Abī Shayba, al-Muṣannaf, ed. Muḥammad ʿAwwāma (Beirut: Dār al-qibla-Muʾassasat ʿulūm al-qurʾān, 2006); Ibn Fuḍayl, Kitāb al-duʿāʾ, ed. ʿAbd al-ʿAzīz b. Sulaymān b. Ibrāhīm al-Buʿaymī (Riyad: Maktabat al-Rushd, 1999); Ibn Ḥibbān, Ṣaḥīḥ Ibn Ḥibbān bi-tartīb Ibn Balbān, ed. Shuʿayb al-Arnaʾūṭ (Beirut: Muʾassasat al-risāla, n.d.); Ibn Māja, al-Sunan, ed. Shuʿayb al-Arnaʾūṭ et Muḥammad Kāmil Qurra Balalī (Damascus: Dār al-risāla al-ʿālamiyya, 2009); Ibn Wahb, al-Jāmiʿ fī l-ḥadīth, ed. Muṣṭafā Ḥasan Ḥusayn Muḥammad Abū l-Khayr (Riad: Dār Ibn al-Jawzī, 1995); al-Maḥāmilī, Kitāb al-duʿāʾ, ed. Saʿid b. ʿAbd al-Raḥmān b. Mūsā al-Qazaqī (Beirut: Dār al-gharb al-islāmī, 1992); Mālik b. Anas, Muwaṭṭaʾ al-imām Mālik (riwāyat Yaḥyā b. Yaḥyā al-Laythī), ed. Bashshār ʿAwwād Maʿrūf (Beirut: Dār al-gharb al-islāmī, 1997); al-Maqdisī, al-Targhīb fī l-duʿāʾ wa-l-ḥathth ʿalay-hi, ed. Abū Yūsuf Muḥammad b. Ḥasan (Cairo, 1991); Muslim b. al-Ḥajjāj, Ṣaḥīḥ Muslim, ed. Muḥammad Fuʾād ʿAbd al-Bāqī (Cairo: Dār al-ḥadīth, 1991); al-Nasāʾī, Kitāb al-Sunan al-kubrā, ed. Ḥasan ʿAbd al-Munʿim Shalabī et al. (Beirut: Muʾassasat al-risāla, 2001); al-Nasāʾī, ʿAmal al-yawm wa-l-layla, ed. Fārūq Ḥamāda (Beirut, Muʾassasat al-risāla, 1985); al-Ṭabarānī, Kitāb al-duʿāʾ, ed. Muḥammad Saʿīd b. Muḥammad Ḥasan al-Bukhārī (Beirut: Dār al-bashā’ir al-islāmiyya, 1987); al-Ṭayālisī, Musnad Abī Dāʾūd al-Ṭayālisī, ed. Muḥammad ʿAbd al-Muḥsin al-Turkī (Cairo: Dār Hajar, 1999); al-Tirmidhī, al-Jāmiʿ al-kabīr, ed. Bashshār ʿAwwād Maʿrūf (Beirut: Dār al-gharb al-islāmī, 1996).

[5] See for example H. M. El-Hawary, “The Most Ancient Islamic Monument Known Dated A.H. 31 (A.D. 652, from the Time of the Third Calif ʿUthman,” The Journal of the Royal Asiatic Society of Great Britain and Ireland, (1930): 322.

[6] Max Van Berchem, Matériaux pour un Corpus Inscriptionum Arabicarum. Deuxième partie : Syrie du Sud, Jérusalem « Haram », II (Cairo: Institut français d’archéologie orientale, 1927), 231 (#215). For an online picture of the facsimile edition, see

[7] See for instance M. Abū l-Faraj al-ʿUshsh, “Inscriptions arabes inédites à Djabal Usays,” Les Annales Archéologiques de Syrie, 13 (1963) : 233 (#36); M. Sharon, Corpus Inscriptionum Arabicarum Palaestinae, III (Leiden-Boston: Brill, 2004), 180; H. Hawary and H. Rached, Catalogue général du Musée arabe du Caire : Stèles funéraires, I (Cairo: Institut français d’archéologie orientale, 1932), 3 (#3).

[8] See for instance R. W. Hamilton, “An Eighth-Century Water Gauge at al-Muwaqqar,” Quarterly of the Department of Antiquities in Palestine, 12 (1946): 70.

[9] For a detailed comparison between inscriptions and ḥadīth, see my forthcoming article in Der Islam.

[10] See Mathieu Tillier, “Early Fiqh and the Issue of Ḥadīth Dating,” Islamic Law Blog, January 6, 2022,

[11] See for instance Joseph Schacht, The Origins, 66, 82, 156, 204-18.

(Suggested Bluebook citation: Mathieu Tillier, Imploring God and the “Living Tradition”: A Relative Chronology of Epigraphic and Traditional Invocations, Islamic Law Blog (Jan. 13, 2022),

(Suggested Chicago citation: Mathieu Tillier, “Imploring God and the “Living Tradition”: A Relative Chronology of Epigraphic and Traditional Invocations,” Islamic Law Blog, January 13, 2022,

Leave a Reply