The origins of the 5 daily Islamic prayers: Jonathan Brown on Tom Holland’s “In the Shadow of the Sword”

In light of my recent posts and research interests, this could not come more timely: An article written by Jonathan Brown on Holland’s ideas on the origins of Islamic ritual practice of the Salaat, or the five daily prayers that Muslims.

Brown is a published scholar on the authenticity of Hadith and other related topics. I highly recommend that those interested in the Islamic historical critical tradition of authenticating and falsifying Hadith start with his book “Hadith: Muhammad’s legacy to the Medieval and Modern World”.

The article in question:

Tom Holland, the Five Daily Prayers and the Hypocrisy of Revisionism

On a radio show on BBC 4 a couple of weeks ago, Tom Holland raised his claim (made originally in his book In the Shadow of the Sword on the origins of Islam) that the famous five daily prayers in Islam were not originally part of the Prophet Muhammad’s teachings. They were actually imported into the religion from Zoroastrianism well after the death of the Prophet Muhammad and after the Muslims had conquered the greater Middle East. Holland sets his story in the environs of the city of Kufa in southern Iraq in the mid eighth century. The Muslim practice of praying five times a day, he argues, resulted from Islam, in effect, imitating Zoroastrian practice. More specifically, Zoroastrian converts to Islam in Kufa brought with them practices such as the five daily prayers from their own religion (Holland, In the Shadow of the Sword, 405).

Holland’s evidence for this is an observation supposedly made by Rav Yehudai Gaon, who was the senior Jewish scholar of the Suru rabbinic academy near Kufa from 757 CE until his death in 761. Rav Yehudai is quoted as remarking that Zoroastrian converts to Islam retained some aspect of their previous religion. Holland writes, quoting Rav Yehudai, “The hearts of those mowbeds [Zoroastrian priests] who had ‘converted to the religion of the Ishmaelites,’ so he reported, were still not entirely clear the trace of their former beliefs, even down to the third generation: ‘for part of their religion still remains within them.’”

Holland belongs to a school of historical thought known as revisionism, which criticizes mainstream Western scholarship on early Islamic history for relying too much on historical sources that 1) are written by Muslims, and therefore biased towards Islamic orthodoxy, and 2) postdate the events they describe by many decades or centuries, during which time the ‘true’ description of events must have been adjusted by the faithful to fit with the orthodox sacred Islamic history that had gelled during the intervening period. So we should not believe that the Muslim scholar Ibn Ishaq’s (died 767) famous biography of the Prophet is a historically reliable source for the events of the historical Muhammad’s life because 1) Ibn Ishaq was a Muslim writing a sacred history of a figure who had become a fixture in his religious tradition and was not being approached objectively, and 2) Ibn Ishaq was compiling his biography some 140 years after the death of the Prophet. To take Ibn Ishaq as reliable would be like historians a century from now writing the history of the American Civil War based on documents written in 2015 by Americans who all celebrate the victory of the North over the South.

To solve this problem of historical sources, revisionists have proposed relying on non-Muslim sources for the early Islamic period, some of which do date from early on (for example, the writings of the Armenian bishop Sebeos come from the 660’s CE) and are not colored by pro-Islam bias.

I could go into a long critique of the revisionist approach here, but that is totally unnecessary. Holland’s argument on the point of Muslim prayer is so feeble that one need only hold it to the standards of the revisionist school itself for it to collapse entirely.

1) Why is Holland putting words in Rav’s Mouth?:

The first problem with Holland’s argument is that Rav Yehudai does not actually mention the daily prayer as an example of Zoroastrian religious influence on Islam. Instead, Holland reaches this conclusion by open speculation. He asks, “What evidence might the rabbi have had for making such a claim?” Well, Zoroastrianism included a five-times-a-day prayer, he notes, so that was probably what Rav Yehudai meant. But speculation is unnecessary, since Rav Yehudai actually said exactly what he meant in the passage Holland cites. His comments about Zoroastrian converts to Islam involves how they tend not to give up drinking wine immediately, sometimes continuing to imbibe into the third generation.

2) How on Earth does Holland think this is a reliable historical source?

Let’s just pretend that Rav Yehudai was actually talking about the daily prayer practices of Zoroastrians/Muslims (which, of course, he never mentions at all). And let’s just assume that Rav Yehudai was making an accurate, fair observation about the practices around him in southern Iraq in the 750’s CE. The writings of Rav Yehudai must be more historically reliable than Muslim ones like Ibn Ishaq, right? What book of Rav Yehudai is Holland citing? Actually, he doesn’t cite any work by Rav Yehudai. His endnote cites the Sefer Ha-Eshkol, a work attributed to Rabbi Abraham ben Isaac, a rabbi living in southern France in the twelfth century. Nor does the Sefer Ha-Eshkol cite Rav Yehudai directly. Instead, Rav Yehudai’s report about Muslim converts comes via a senior rabbi of the Pumbedita rabbinic academy in Babylon who lived some three hundred years after Rav Yehudai, Rav Hayya Gaon (died 1038 CE).

So, Holland is saying that, in order to overcome the problem of Muslim sources like the biography of the Prophet by Ibn Ishaq, which was compiled in Baghdad a century and a half after Muhammad’s career, we should turn to a source written in France five centuries after Muhammad’s career? But, Holland might reply, Abraham ben Isaac was drawing on earlier reports and historical works, which we should trust. But this is exactly what Muslim historians like Ibn Ishaq claimed to be doing in their works, and the central criticism made by revisionists like Holland is that we can’t just trust that historians are reliably passing on earlier material.

But let us be charitable. Let’s assume that during the five centuries between the lives of Rav Yehudai Gaon and Abraham ben Isaac, five centuries of religious polemics and warfare between Christians, Jews and Muslims across the Mediterranean, that Rav Yehudai’s observation remained intact to be preserved for us in the Sefer Ha-Eshkol.

The problem is that the Sefer Ha-Eshkol itself is unreliable. As has been discussed for over a century by rabbinic scholars and scholars of Judaic studies, the 1868 Halberstadt edition of the Sefer that Holland relies on was a forgery produced by the famous nineteenth-century Rabbi and literary scholar Zvi Benjamin Auerbach (died 1878). This has led some leading scholars of rabbinic literature to conclude that the book “should not be used for historical purposes.”

3) So what sources should we rely on for the origins of the five daily prayers?

Let’s indulge revisionist skepticism about historical sources written by Muslims. Let’s forget that the story of how and when the Prophet Muhammad instituted the five daily prayers, which Muslim scholars concluded either happened in 617 CE or soon before the Prophet’s emigration to Medina 622 CE, was recorded in major Muslim historical collections from the late 700’s and early 800’s. The earliest attested book in which this story appears in the Muwatta’ of Malik bin Anas (died. 796), which was compiled in Medina in the mid to late 700’s. Malik includes a report transmitted via a chain of narrators from the Prophet, who said, “Five prayers God has ordained for His servants, and whoever does them without treating them lightly, God has given that person a promise to grant them entrance into the Garden….” (Muwatta’: kitab salat al-layl, bab al-amr bi’l-witr). If we indulge in revisionist skepticism and assume that Malik was making up the whole transmission that he claims came from the Prophet, we would still know that, at least during Malik’s own lifetime in Medina, there was the clear idea that a core part of Islam was five daily prayers.

And then we could indulge more revisionism and insist on relying on non-Muslim sources. Since Rav Yehudai never mentions the Muslim prayer, why not look at a non-Muslim source that does? We could look at the T’ung tien, a Chinese Tang court work of history and geography that was published in 801 CE. It contains a description of Kufa by a Chinese soldier who was taken prisoner at the Battle of Talas in 751, spent years amongst the Muslims in Iraq and Iran, and returned to China in 762. One of the few observations that this Chinese soldier recalls of Kufa, which was the Abbasid capital at the time (Baghdad not being built until the 760’s), was that the Muslims there would pray five times a day.

So between the Muwatta of Malik and the T’ung tien, we know that Muslim communal practice in Medina and Kufa in the mid 700’s included the five daily prayers. This despite the fact that the two regions of Medina and Kufa had dramatically different traditions of Islamic law. So both regions must have inherited the prayer practice from a common, earlier practice, and there thus must have been some common origin for the five prayers. This would push the historical attestation for the practice back at least one generation to at least the early 700’s, only seventy or so years after the death of the Prophet.

4) When arrogant historians tell us that being critical means ignoring the data

So we have a choice. We can believe Holland’s claim, based on an unreliable nineteenth-century forgery of a supposedly twelfth-century work from France quoting an eleventh-century rabbi in Baghdad quoting an eighth century rabbi from near Kufa, that, because Zoroastrian converts to Islam still liked to drink wine, that therefore the Muslim practice of praying five times a day, which, like wine drinking, Zoroastrians also did, must also have been imported into Islam from Zoroastrianism by Zoroastrian converts.

That, or, we can believe, based on historically attested Muslim and non-Muslim sources, which paint a reliable overall picture of Muslim practice in Kufa and Medina decades before Rav Yehudai supposedly made his observation, that the five daily prayers were widely accepted as a core practice of Islam by at least the early 700’s, only seventy or so years after the death of the Prophet Muhammad.

Actually, Holland’s claim makes even less sense when we remember that the tradition of Islamic law in Kufa, where Holland has all these Zoroastrian practices and five-times-a-day-praying Zoroastrian sleeper converts supposedly influencing Muslims, actually argued for there being SIX required daily prayers (the sixth, the witr prayer, is still considered required in the Hanafi school that originated in Kufa). The Muwatta of Malik, on the other hand, written in Medina where Holland would have us assume that there were many fewer Zoroastrian converts wandering around, rejects the sixth prayer and insists on the supposedly Zoroastrian-based five times.

Brown and Holland were also featured in a recent episode of BBC’s “Beyond Belief” podcast, which is what prompted Brown’s refutation of Holland’s theories. The audio can be accessed here:


The Hadith Critical Methodology: A brief look at how Hadith are authenticated in the Islamic tradition.

Some time after the death of the Prophet, a need arose among the Muslims for an authentic source for the beliefs and ritual practices of the Prophet Muhammad (ص). The Qur’an, which was undoubtedly considered authentic and authoritative among all Muslims, did not provide much information about Muslim practice and Law. Indeed, it is the Hadith, that is, the sayings and actions of the Prophet Muhammad (ص), that provides the bulk of the raw material for Islamic jurists past and present. After the death of the companions and the generation who succeeded them, there arose a new challenge facing the ranks of the Muslim jurists. All those who had personal contact with the Prophet had passed away: There was no longer any early figures to tell the Muslims what the authentic teachings of the Prophet was. Forgery also became a huge problem in the Islamic lands: Sectarian and political turmoil produced an environment where putting words in the mouth of the Prophet to support one’s own cause was common. So how did Muslim scholars of Hadith, the Muhaddithoon, sift through the masses of Prophetic reports to single out the authentic ones?

1. Demand a source. 

The first step to filtering out useless traditions is to ask the narrator for a source on his Hadith. Ideally, the man (or woman) should be able to provide an Isnaad. The word “Isnaad” in Arabic means “support” and it is in effect a chain of narrations from the reporter back to the Prophet. Lets look at an example. There is a hadith found in Sahih Al-Bukhari, which goes by this text (called the matn):

كَانَ النَّبِيُّ صلى الله عليه وسلم يُعْجِبُهُ الْحَلْوَاءُ وَالْعَسَلُ‏.‏

“The Prophet, may the blessings of God and peace be upon him, used to like eating sweets and honey.”

An isnaad for this text is provided by Al-Bukhari:

حَدَّثَنَا عَلِيُّ بْنُ عَبْدِ اللَّهِ، حَدَّثَنَا أَبُو أُسَامَةَ، قَالَ أَخْبَرَنِي هِشَامٌ، عَنْ أَبِيهِ، عَنْ عَائِشَةَ ـ رضى الله عنها

‘Ali ibn Abdullah narrated from Abu Usaamah who said Hishaam told me from his father who reported from ‘Aishah. 

Perhaps it is more intuitive to represent this isnaad on a diagram: Isnaad chain 1 If the narrator is able to provide an isnaad like the one seen above, he has passed the first layer of the critical method. Note that right now his report is only a little better than hearsay. Did the Prophet Muhammad (ص) really have a sweet tooth? We still don’t know: Several other steps remain for us to take before declaring the report as useful.

EDIT 1/1/15 : Hisham’s father’s name is ‘Urwah, and the next diagram reflects this change.

2. Criticize the source

As you probably can already tell, asking for a source doesn’t deter a forger. Anyone could make up an isnaad. The names could be just something the forger thought up of on the spot- or, if he were taking a bit more care, looked at examples of good narrators and simply strung them up in a somewhat believable order of narration. Enter the next step of Hadith evaluation: Isnaad Criticism. The critic’s job was to scrutinize the isnaad itself to see whether it was actually plausibly represented a real transmission of the report. They did this in several ways:

1. Biographical evaluation. 

A very useful weapon in the Muhaddith’s toolbox was the massive biographical tomes that historians and hadith scholars produced. They listed the birth/death dates of narrators, a brief description of their lives (if available), their geographical locations, their students and their teachers, among other useful information. The usefulness of these books has been stressed by many scholars past and present. Says Hassan b. Zayd, “We never used against the forgers any device more effective than the Ta’rikh (historical/chronological information).” From these books, a scholar will find out whether:

1. There is any information confirming that these people in the isnaad actually existed. If the narrator is not found in any of the biographical dictionaries, then there is a risk that the name is simply invented; the chain would therefore be considered weak. Another way of confirming the existence of someone is when he is quoted by name by several well-known sources of Hadith.

2. The lives of the people in the Isnaad actually overlapped. It is very difficult to place confidence in a chain where one of the narrators was born 30 years after the death of the person he supposedly heard the report from!

2. Their biographical situations make it plausible that these transmitters in the chain would report from each other.

In summary, what scholars want to see is whether the isnaad actually makes sense. Ofcourse, some scholars would have stricter standards than the points mentioned above: Al-Bukhari, for example, wanted positive proof that the narrators met from those they said they heard from: it was not enough for him that they lived in the same general area, or that their lives overlapped.

2. Transmitter evaluation

The individual reliability of the narrators was another thing that the scholars of Hadith were on the look out for. The hadith scholar had to go through all the reports of the narrators in the isnaad and check whether they were reliable sources of information.

Lets look at Abu Usaamah as an example. How do we find out whether this man is reliable? One of the primary methods is looking for ‘multiple attestation’ among his reports. This means, if Abu Usaamah heard a Hadith from Hishaam, do we see that same Hadith being reported by the other students of Hishaam? It would only be logical that this be the case, if Abu Usaamah is telling the truth. Lets say for example, you’ve missed a class on one of your course subjects. You ask one of the attendees, named Zayd, this question: “Was there anything important in the lecture?” to which he replies, “X topic won’t appear on the final exam”. You then ask five other attendees about this matter- and you see that they don’t say anything of the sort; even though they should have actually heard this if the lecturer said it. Zayd’s report doesn’t spark much confidence now: we should play it safe and say Zayd is probably lying or he misheard the lecturer.

The scholar Muslim bin al-Hajjaj says this plainly:

“[A weak hadith transmitter is] when his narrations are compared to those of people known for preservation of Hadith and uprightness of character, his narrations do not concur with their narrations, or do so only rarely. If the majority of his hadiths are like that then he is rejected and not used in hadith.”

Other criterion for transmitter evaluation is the religiosity and piety of a transmitter. One cannot at any cost accept the report of someone who has already been shown to be a liar and forger of Hadith. For a very good isnaad, all the narrators would ideally be scholars who are known for their practice and devotion to Islam. In this case the critics are somewhat flexible, after all, an expert forger would try his best to look like a devoted Muslim. Furthermore, even religious people can easily make mistakes- or forge material “for a good cause” as is the case for many ascetic forgers. Pious forgery is a well known phenomenon in religious traditions outside Islam too. Al-San’ani points out that accuracy (based on the multiple attestation criterion above) is far more important than uprightness.

3. Look for corroboration of the source

This is arguably the most important part of Hadith criticism, and it is after this stage can a report of the Prophet actually inspire some confidence in the eyes of the hadith scholar. The scholar now looks at all the isnaads of the Hadith and see whether there is a lot of different sources saying the same thing. If lots and lots of people attest to a report, and all these people’s reports really do trace back to the eyewitnesses, then we can have good confidence in the report. In Western historical evaluation this is called the criterion of multiple attestation This is probably best explained via demonstration. Lets look back at our diagram earlier. We will now introduce some more sources. These are by no means exhaustive, I’ll only be bringing a few more isnaads in for the sake of demonstration. Isnaad chain 2 The new diagram introduces some isnaads found in the Sunan of Ibn Majah as well another Isnaad found in Sahih Al-Bukhari. As I have stressed earlier, there are far more isnaads than this in reality. We see that the Hadith is well corroborated at the point of Abu Usaamah. Four people claim to have heard directly from him. It probably is the case that Abu Usaamah really did say this. Can we move up further? Probably: Hishaam, too, is multiply attested. A certain ‘Ali also corroborates Abu Usaamah’s report from Hishaam. Just by looking at these isnaads alone, we have some confidence that Hishaam said this. What remains now is to look for some sort of corroboration for Hishaam’s hearing from ‘Urwah. This will not be done in the present article: but if it is found, it will give some credit to ‘Urwah’s statement too. Since ‘Aisha alone would probably be the one who narrates about the Prophet’s preference in food, since she was one of his closest wives and the only one of them that became a major Hadith transmitter, we will probably take her word for it without looking for any corroboration from the Prophet’s other wives. Sunni dogma aside, it is unlikely that she would lie about the Prophet’s love of sweet food. A more complete example of the transmission of Hadith can be seen in this diagram from Al-Azami’s book. This time it concerns the love that God has for the believer that fasts and abstains from minor sins during it for His sake. big isnaad chain P

The Isnaad bundle that ‘Azami has rendered above gives us very good confidence in a Hadith report. We can see a large amount of attestation: it is hard to believe that this report is forged. The hadith appears in Madinah, Basra, Kufa, Mecca, Wasit, the Hijaaz, and Khurasaan at the 3rd level of the isnaads. It is widely multiply attested; some scholars would give it the “mutawaatir” grade.

Perhaps I’ll edit this article to look for more Isnaads to corroborate our report on the Prophet’s liking for sweets, but for now, I hope this has given the reader a brief overview as to how Hadith criticism is done. And God knows best.


Al-Azami, Mustafa. Studies in Hadith Methodology and Literature. 2002.

Zubayr Siddiqi, Muhammad. Hadith Literature: Its Origin, Development, Special Features and Criticism. 2006

Brown, Jonathan Hadith: Muhammad’s Legacy to the Medieval and Modern World. 2009

Al-Azami, Mustafa. On Schacht’s Origins of Muhammadan Jurisprudence. 1996

Notes: This article approaches the topic of discussion the format of the “Three-tiered method” of Sunni Hadith Criticism that is introduced in Jonathan Brown’s “Hadith: Muhammad’s Legacy to the Medieval and Modern World”, but I have used more books (listed in the bibliography) to supplement this text. The so called three tiered method is only presented this way for ease of presentation, the Muhaddithoon don’t consciously call it that.

On the origins of the Isnad: Ibn Sirin’s statement and Schacht’s interpretation of it

One of Schacht’s underlying contentions behind his magnum opus, “The Origins of Muhammadan Jurisprudence”, is that hadith literature did not exist in the first century of Islam. This obviously means that the isnads – the chains of narration that link a hadith report back to the Prophet – are almost completely fictitious. They cannot trace information back to the first century and consequently are not reliable sources of information on the Prophet himself or even his companions.

On the other hand, we have the statement on the origins of the isnad from Ibn Sirin (d. 110AH) who spent the early years of his life growing up in Ali’s caliphate and lived until the dawn of the second Islamic century. He says:

They did not use to ask about the isnad, but when the fitna arose, they said, “Name us your men.”[1]

Schacht interprets this statement in light of his preconceived conclusions. He says:

We shall see later that the fitna which began with the killing of the Umaiyad Caliph Walid b. Yazid [A.H. 126], towards the end of the Umaiyad dynasty, was a conventional date for the end of the good old time during which the sunna of the Prophet was still prevailing; as the usual date for the death of Ibn Sirin is A.H. 110, we must conclude that the attribution of this statement to him is spurious. In any case, there is no reason to suppose that the regular practise of using isnads is older than the beginning of the second century A.H.[2]

Essentially, this is Schacht’s logic: The “fitna” refers to the killing of the Umayyad caliph Walid b. Yazid, so the isnad started being used in the second century, not the first. Also, because Ibn Sirin wasn’t even alive at this time, the statement is forged anyway!

The problem with this interpretation is obvious: the “fitna” probably does not refer to the death of the Umayyad Caliph[3]. It is more reasonable to suppose that Ibn Sirin was talking about the greatest fitna of early Islam- the fitna between ‘Ali and Mu’awiyah, only decades after the death of the Prophet Muhammad (pbuh). As other events happened during the life of Ibn Sirin that were also called by the word fitna, he probably meant the most significant one.

Harald Motzki [4] also rightly points out that if the statement was falsely attributed to Ibn Sirin then the forger would have taken care to specify which fitna the attribution is referring to. This is because, again, there were many such events- had the forger wanted to actually achieve his goal (of proving the Isnad to be a 2nd century invention) he would have made Ibn Sirin more specific, otherwise readers would potentially be confused as to which fitna was being spoken of.

Schacht’s argument only really works if we have very good reason to believe that the isnad did not originate in the first century. Both the works of Harald Motzki and Mustafa Al-Azami go far to dispel this claim. Further information can be found in their books, “The Origins of Islamic Jurisprudence” and “On Schacht’s Origins of Muhammadan Jurisprudence”.

[1] Al Azami, Mustafa. On Schacht’s Origins. Islamic Texts Society. 1996. p 155

[2] Schacht, Joseph. Origins of Muhammadan Jurisprudence. Oxford University Press. 1967. p 36-37

[3] Al Azami, Mustafa. On Schacht’s Origins. Islamic Texts Society. 1996. p 168

[4] Motzki, Harald. The Origins of Islamic Jurisprudence. Brill. 2002