From Anonymous Dicta to the Prophet’s Sunna

By Mathieu Tillier

This is part three 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.

The history of Islamic law and that of ḥadīth are closely connected. As I recalled in my previous posts, prophetic authority as expressed in Muḥammad’s sunna became one of the pillars of legal hermeneutics from the third/ninth century onwards. Although a ḥadīth is presented as a written record of an uninterrupted oral transmission beginning in the first/seventh century, it remains to this day difficult to date its circulation due to the circular character of isnād analysis. Yet, as I have argued in my previous posts, it is possible to compare a sample of ḥadīth with dated epigraphic sources. The invocations or supplications to God (duʿāʾ) that Muslims engraved on stone from the earliest days of Islam do indeed use formulae found in ḥadīths relating to that form of prayer. Insofar as invocation to God followed normative practices – what Joseph Schacht called “living tradition” – I hypothesized that a comparison between these epigraphic practices and ḥadīth could reflect, more generally, the dialectical relationship between practices responding to either social or legal norms, and the expression of those norms as ḥadīths.

Comparing the corpus of epigraphic inscriptions that include invocations to God with ḥadīth literature has previously led me to conclude that some ḥadīths mentioning formulae that overlap with the inscriptions of the first century AH can be considered as historically plausible. This means that the Prophet and his Companions might have used them to address God in their prayers. Conversely, other formulae early attested in epigraphic inscriptions are absent from ḥadīth literature. For instance, the invocation for Jesus found in the Dome of the Rock has no ḥadīth counterpart. In the second/eighth century, Jesus quickly lost the centrality he enjoyed at the time the Dome of the Rock was built in 691-692 CE.[1] This absence of invocations for Jesus means that ḥadīth literature actually offers only a selection of first-century phrases. Some formulae that the Prophet or his Companions may have uttered never entered the body of Islamic traditions, partly because of dogmatic changes since the early days of Islam. In addition, we have seen that many invocatory traditions found in the ḥadīth must be dated to later periods insofar as their formulae do not match those attested in first-century inscriptions.

In his seminal work on the historical formation of Islamic law, based on literary sources and excluding any documentary material, Joseph Schacht argued that legal traditions referring to the Prophet’s authority did not emerge before the first half of the third/ninth century. According to him, the anonymous “living tradition” was first projected backwards onto post-prophetic authority figures – that of Successors, and subsequently of Companions – before it was eventually transformed into prophetic sayings.[2] Similarly, Intisar Rabb has recently shown how the “doubt canon” in legal procedure remained rooted in anonymous practice until it was transformed into a prophetic ḥadīth in the third/ninth century.[3] So far, I have deliberately avoided distinguishing between speaking authorities, that is, those to whom ḥadīth literature ascribed the formulae that are attested on stone. I use the term ḥadīth in its broader sense of a saying ascribed to any authority, whether or not this authority is the Prophet. It is now time to introduce those authorities into the picture, in order to better understand how formulae were chronologically ascribed to different persons. This comparison, using dated epigraphic documents as a reference point, should thus make it possible to assess the validity of Joseph Schacht’s conclusions.

When one compares the earliest epigraphic invocations with the earliest ḥadīth collections or, at least, with those that preserve the widest range of attributions to authorities (among which, the Muṣannafs of ʿAbd al-Razzāq al-Ṣanʿānī [d. 211/827] and of Ibn Abī Shayba [d. 235/849]),[4] it appears that these formulae are often ascribed to anonymous or biblical figures (also cited in the Qur’ān) such as Abraham, Moses, David and Jesus, but also to supernatural creatures such as the angel Gabriel or the houris of paradise.[5] The authoritative value of sayings attributed to anonymous speakers became extremely weak when the ḥadīth reached its maturity in the third/ninth century. Biblical figures did still enjoy some authority, but to a much lesser extent than the major authorities of early Islam, probably because the transmission of their sayings offered no guarantee of reliability. The weaker authority of these anonymous or biblical sayings led subsequently either to their exclusion from the canonical ḥadīth corpus or to their re-attribution to more authoritative figures. One can therefore conclude, as Joseph Schacht did, that anonymous traditions represent the earliest layer (or one of the earliest layers) of ḥadīth literature. We may add that this early layer also included mythical figures whose authority was later disqualified.

At a second stage, which can also be reconstructed from the same sources, invocatory formulae were attributed to non-prophetic speakers (Companions and Successors), in what scholars called ḥadīth mawqūf and ḥadīth maqṭūʿ.[6] One is struck by the strong chronological convergence, at the end of the first century, between the appearance of a formula on stone and the time when authorities to whom it was attributed were alive. For example, the phrase Allāhumma … urzuq (“O God, … give sustenance”), which is first attested on stone in 78/698,[7]  is ascribed in early ḥadīth to roughly contemporary figures such as Abū Muslim al-Khawlānī (d. 62/682), al-Ḥasan al-Baṣrī, Ṭāwūs b. Kaysān (d. ca. 106/724-725), Makḥūl al-Shāmī (d. ca. 113/731-732) and ʿAṭāʾ b. Abī Rabāḥ (d. ca. 114/732). This convergence suggests that these post-prophetic authorities could plausibly have uttered these formulae, and that their later dissemination in the form of non-prophetic traditions may have originated in the memory of words historically uttered by these figures.

Other traditions state that these Companions and Successors approved invocatory formulae that are also found in anonymous or biblical versions.[8] This phenomenon indicates that these Companions’ and Successors’ ḥadīths correspond to a second stage in the formation of Islamic tradition, which was meant to enhance the credit of earlier traditions lacking sufficient authority.

The third stage is, finally, that of the “prophetization” of invocatory traditions, that is, their transformation into prophetic ḥadīths, directly linked to the Prophet. This prophetization is already noticeable in the second half of the second/eighth century in the Muwaṭṭaʾ of Mālik (d.179/975), who ascribed to the Prophet some formulae attested on rocks as early as the first century.[9] Conversely, Mālik never ascribed to the Prophet invocations that appeared on rocks after 92/710. In his work, such second/eighth-century formulae are only quoted in the form of āthār going back to Companions or Successors.[10] These formulae were still new in his time, and had not yet had time to be retroactively attributed to the Prophet – which happened in later ḥadīth collections.

A profound paradigm shift indeed occurred in the first half of the third/ninth century, which I would associate primarily with the Musnad of Aḥmad b. Ḥanbal (d. 241/855). This paradigm shift resulted not only in a restriction of authority to prophetic ḥadīth (non-prophetic material is excluded from the Musnad and relegated to secondary collections), but also in ascribing to the Prophet invocatory formulae that were previously non-prophetic. This process, which I call “prophetization,” took different forms. The Prophet was sometimes made the speaker of a formula that had previously been ascribed to a Companion, who thereby lost his status as a speaker and became a mere transmitter.[11] In other cases, the Prophet was inserted as a listener of an invocation pronounced by someone else, which made it possible to legitimize the invocation with the Prophet’s tacit or explicit approval.[12] In a third scenario, the Prophet was introduced in the narrative to justify a secondary theological issue appearing in the invocatory tradition; the invocation remained non-prophetic, but the corresponding ḥadīth became adorned with prophetic legitimacy.[13]

Ibn Ḥanbal’s Musnad finally contains prophetic ḥadīths that did not previously exist in any form, whether prophetic or non-prophetic. Such is the case, for example, in a ḥadīth containing the formula Allāhumma … ḥāsib … ḥisāb (“O God, ask [him] for accounts…”), attested on stone from 225/840 onwards, that is, during Ibn Ḥanbal’s own lifetime. This formula, unprecedented in ḥadīth literature, was directly prophetized and included as part of the sunna.[14]

It is nevertheless worth noting that the prophetization of invocatory formulae was limited. As evidenced in fig. 1, the number of epigraphic formulae that were transformed into prophetic traditions decreased over time. This means that prophetization could only occur if and when a formula had become commonly accepted. The more recent a formula was, the less likely it was to be unanimously accepted and thus to become part of the prophetic sunna. The prophetization of old and accepted material went therefore hand in hand with a gradual closing of the canonical corpus of prophetic ḥadīth, whereby part of the invocatory material (especially some of the most recent one) was excluded.

Studying the authorities to whom ḥadīth literature ascribes invocatory formulae attested on rocks thus allows for a reconstruction of the historical dynamics that led to the formation of prophetic sunna. Standard formulae that were initially attributed to anonymous people or to Biblical/Qur’ānical or mythical figures were, at a second stage, reattributed to Successors or Companions in order to grant them higher authority. The last stage, which flourished in the third/ninth century, was the projection of these words onto the Prophet. The results I have just presented take as reference point dated inscriptions offering a relatively reliable chronology for the introduction of new invocatory formulae. This dating thus provides a historical starting point for the integration of a formula into the normative system under construction, of which the ḥadīth became a key material. Therefore, these results do not only confirm Joseph Schacht’s conclusions regarding the emergence of legal ḥadīths and their progressive incorporation in the legal discourse, but also provide a documentary basis for them. How generalizable are these results? One should of course remain cautious not to over-interpret them, since they rely only on a limited sample of ḥadīths. The convergence between my results regarding the progressive transformation into ḥadīths of social norms regarding invocation, and those of Schacht, which relied on the analysis of ḥadīth with legal content, suggests, however, that my own conclusions are potentially extensible to a large part of the ḥadīth corpus, whether or not it was eventually used for legal purposes.

Yet, a nuance to Schacht’s conclusions is necessary. While he believed that a search for a core of ḥadīth dating back to the Prophet was pointless,[15]  I argued in my last post that some ḥadīths preserved a lexical core potentially dating back to the middle of the first/seventh century and that the Prophet may have used some expressions occurring in the ḥadīth literature. Similarly, the temporal convergence between the appearance of invocatory formulae on stone and the life spans of some Successors to whom they are attributed suggests that these formulae were associated with them, if not during their lifetime, then at least shortly after their death, and that they may very likely have uttered them. This confirms that Successors’ sayings, in the first half of the second/eighth century, if not as early as the end of the first/seventh century, represent one of the earliest roots of Islamic law.


[1] See Mathieu Tillier, “‘Abd al-Malik, Muḥammad et le Jugement dernier : le dôme du Rocher comme expression d’une orthodoxie islamique,” in Les vivants et les morts dans les sociétés médiévales. Actes du XLVIIIe Congrès de la SHMESP (Jérusalem, 2017), ed. Société des historiens médiévistes de l’Enseignement supérieur public (Paris: Éditions de la Sorbonne, 2018), 348, 351, 363 (online version:

[2] Joseph Schacht, “A Revaluation of Islamic Traditions,” Journal of the Royal Asiatic Society of Great Britain and Ireland, 2 (1949): 145; Joseph Schacht, The Origins of Muhammadan Jurisprudence (Oxford: Clarendon Press, 1950), part 2.

[3] Intisar A. Rabb, Doubt in Islamic Law: A History of Legal Maxims, Interpretation, and Islamic Criminal Law (Cambridge: Cambridge University Press, 2015), 48-59.

[4] ʿAbd al-Razzāq al-Ṣanʿānī, Muṣannaf ʿAbd al-Razzāq, ed. Ḥabīb al-Raḥmān al-Aʿẓamī (Beirut: al-Maktab al-islāmī, 1983) (; Ibn Abī Shayba, al-Muṣannaf, ed. Muḥammad ʿAwwāma (Beirut: Dār al-qibla-Muʾassasat ʿulūm al-qurʾān, 2006) (

[5] See for example Ibn Fuḍayl, Kitāb al-duʿāʾ, ed. ʿAbd al-ʿAzīz b. Sulaymān b. Ibrāhīm al-Buʿaymī (Riyad: Maktabat al-Rushd, 1999), 280 (#102); ʿAbd al-Razzāq, Muṣannaf, III, 119 (#4989); V, 149 (#9218), 256 (# 9538); Ibn Abī Shayba, Muṣannaf, V, 348 (#8181); X, 261 (#19697); Ibn Ḥanbal, Kitāb al-zuhd, ed. ʿAbd al-Raḥmān b. Qāsim, (Beyrouth : Dār al-kutub al-ʿilmiyya, 1983), 72, 101; al-Nasāʾī,ʿAmal al-yawm wa-l-layla, ed. Fārūq Ḥamāda (Beirut, Muʾassasat al-risāla, 1985), 200 (#137).

[6] See for example ʿAbd al-Razzāq, Muṣannaf, II, 158 (#2889), 159 (#2892), 187 (#3010), 206 (#3082), 449 (#4042), 568 (#4489); III, 110 (#4968), 116 (#4982); Ibn Abī Shayba, Muṣannaf, III, 50 (#3050); V, 313 (#8049), 314 (#8053), 350 (#8188); VI, 83 (#8930); VII, 199 (#11301), 246 (#11477, #11478), 247 (#11480), 249 (#11484).

[7] Frédéric Imbert, “Califes, princes et compagnons dans les graffiti du début de l’Islam,” Romano-Arabica, 15 (2015): 69.

[8] See for example Ibn Fuḍayl, Kitāb al-duʿāʾ, 212 (#47).

[9] Mālik b. Anas, Muwaṭṭaʾ al-imām Mālik (riwāyat Yaḥyā b. Yaḥyā al-Laythī), ed. Bashshār ʿAwwād Maʿrūf (Beirut: Dār al-gharb al-islāmī, 1997), I, 227 (#441), 233 (#456), 234 (#457).

[10] Mālik b. Anas, Muwaṭṭaʾ, I, 313 (#609); II, 524 (#2700).

[11] See for example Ibn Abī Shayba, Muṣannaf, XIX, 346 (#36269) versus al-Tirmidhī, al-Jāmiʿ al-kabīr, ed. Bashshār ʿAwwād Maʿrūf (Beirut: Dār al-gharb al-islāmī, 1996), V, 425 (#3424).

[12] See for example Aḥmad b. Ḥanbal, Musnad al-imām Aḥmad b. Ḥanbal, ed. Shuʿayb al-Arnaʾūṭ and ʿĀdil Murshid (Beirut: Muʾassasat al-risāla, 1997), VII, 40 (#3925).

[13] See for example Aḥmad b. Ḥanbal, Musnad, XXXVI, 27 (#21697).

[14] Gaston Wiet, Catalogue général du Musée arabe du Caire : Stèles funéraires, IX (Cairo: Institut français d’archéologie orientale, 1941), 104, #3366. Cf. Aḥmad b. Ḥanbal, Musnad, XL, 260 (#24215).

[15] Joseph Schacht, “A Revaluation of Islamic Traditions,” 147.

(Suggested Bluebook citation: Mathieu Tillier, From Anonymous Dicta to the Prophet’s Sunna, Islamic Law Blog (Jan. 20, 2022),

(Suggested Chicago citation: Mathieu Tillier, “From Anonymous Dicta to the Prophet’s Sunna,” Islamic Law Blog, January 20, 2022,

Leave a Reply