I get asked this a lot. “Why not apply your methods and skills to the question of whether Muhammad actually existed or not?” My answer is always the same: I will not likely ever be able to do that, because it would require getting a whole second Ph.D. in Medieval Arabic Studies & Languages (and those languages include not just Medieval Arabic, but also Syriac at the very least). And I must emphasize, I would need not just a good command of the languages (as I do the Greek and Latin required for studying the origins of Christianity), but also “a strong grasp of the historical, cultural, political, social, economic, and religious context” of the origins of Islam (Proving History, p. 18), and that happens to include numerous relevant yet distinct cultural contexts (not just of the Middle East and North Africa, but Byzantine as well).

And I’m not going to do that. Because that period bores the shit out of me. If I ever get a second Ph.D., it will be in contemporary philosophy, to improve myself even further in that area, or a modern science. But Medieval Arabic Studies? Sorry. No. I have all these skills with respect to the origins of Christianity, so I am well qualified to produce peer reviewed studies of the historicity of Jesus (hence my book On the Historicity of Jesus, published at the University of Sheffield). But not so for Islam. We need someone to do that who has parallel skills in that field. I can competently communicate the findings of experts in the field. But on this topic, those findings are confusing and disputed at present.

Okay. So. Lacking my ability to test the question myself, or fully vet anyone else’s case as an expert, can I at least answer whether there is any plausibility to the claim that there was no Muhammad, that he was invented by early Arabic military leaders to give a name to a text they cobbled together to inspire their soldiers and build a new civilization on? In other words, that he was invented in exactly the same way and for exactly the same reasons as the militaristic Jews invented Moses?

I don’t know.

This is the most I can tell you…

The Basic Problem

There is a Wikipedia page on the subject (Historicity of Muhammad), but as one can tell from the similar page for Jesus (Historicity of Jesus), it’s mostly crawled over and meddled with by fundamentalist apologists and secular historicity fanatics, so it can’t really be relied upon to give you an honest picture. (Indeed, honest Jesus historians should be embarrassed by that page for Jesus. So I fear the same could be true of honest Muhammad historians for his correlating page.)

Experts (at least non-fundamentalists) do agree that:

  • The Quran can’t be used non-circularly to challenge such a hypothesis. (Nor, conversely, can a recent manuscript find prove such a hypothesis, since even that was written on parchment dated to precisely when the Quran claims to have been written, not before, and its stylistic features strongly suggest the current ink was not even placed on that parchment until many decades later.)
  • The Hadith contains much that is fabricated (so in fact Muslims were inventing Muhammad tradition, in abundance). Discussion and bibliography on that point can be found in Robert M. Price, “The Abhorrent Void: The Rapid Attribution of Fictive Sayings and Stories to a Mythic Jesus,” Sources of the Jesus Tradition (ed. R. Joseph Hoffmann).
  • No literature about Muhammad, that adds information not in the Quran, appears to have been written (or if written, none survives) until a century or more after his purported death, a situation in fact worse than for Jesus.
  • Mentions of Muhammad, and minor details about him, do exist within decades of his death, even from non-Muslim sources, but they all appear to be repeating what is said or implied in the Quran, or by Muslims using the Quran as a source. There are no eyewitness sources, nor any contemporary sources.
  • There is no archaeological corroboration (coins, inscriptions, or attesting manuscripts, of documents or literature, dating to within his life or very near it, other than the Quran). The earliest coins mentioning Muhammad start in 685 A.D., and the earliest inscriptions mentioning Mohammad start in 691 A.D. (dates that are fifty to sixty years after his purported death), but both reference him only in a creedal declaration (known as the Shahada), the existence of which is already entailed by any minimal non-historicity thesis. Similarly all subsequent inscriptions (e.g. on the Dome of the Rock, inscribed a year later; in fact, that just quotes the Shahada and the Quran).
  • Until recently the “best” attempt to argue Muhammad was a fiction was Robert Spencer’s Did Muhammad Exist? (2012), but that is not a peer reviewed university publication, and from what I can tell, it is not very convincing to those qualified to assess it. (Spencer might also be something of an Islamophobic crank. And according to Wikipedia at least, he has only an M.A. in “religious studies,” and even that only specializing in the history of Catholicism.)

This leaves us with certain first-stage questions that need to be answered by a qualified expert:

  • (1) Does any of the later literature cite or describe earlier sources that they used? As is claimed to be the case; but unfortunately I can’t vet those claims. Even some comments or notes in non-Muslim Syriac texts, which seem to derive from eyewitnesses within years of Muhammad’s purported death, are problematic as to authenticity and date, and are again in Medieval Syriac, requiring a different set of skills to evaluate than I possess, and on top of all that, all they do is mention that Muslim invaders follow the writings of the prophet Muhammad. Which the non-historicity thesis already grants. None of them are witnesses to the verifiable existence of Muhammad.
  • (2) Has anything been overlooked? For example, by taking historicity for granted, there may be evidence yet supporting it that no one has gathered up yet. And even the evidence already acknowledged requires a more definitive and unbiased critical analysis. The Syriac notes, for example, often don’t get mentioned by historicity deniers, or treated critically by historicity defenders. And there are complicated arguments surrounding the Charter of Medina, for example, that could yet yield a case for historicity. And that is also well beyond my ability to assess. And who knows what else there might be?
  • (3) What is the normal expected literature, document, and epigraphy survival rate from the 7th century? (The century when Muhammad is to have lived and died, and his followers first began expanding their conquests.) Because that affects how likely we should expect better evidence to still exist if Muhammad did. For instance, we have contemporary inscriptions attesting the existence of Caliph Umar from within ten years of Muhammad’s purported death, and Umar was reputedly a “close companion” of Muhammad and major player in governing the new Muslim state; and we have inscriptions attesting other players as well; but so far, never Muhammad.
  • And related to both questions, (4) are there any events essential to Muhammad’s existence that should be attested by better evidence?

Further questions would pertain (5) to prior probability: in that time and place, how commonly were fictive authors created to market holy books or other propaganda (since Mohammad is not in any of the myth-heavy reference classes Jesus belongs to); and then (6) to hypothesis testing: how would one prove likely an alternative origin story for Islam and the Quran. Because, remember, possibly is not probably (this is what I have named the possibiliter fallacy, as described in Proving History, Axiom 5, pp. 26-29).

Robert Spencer

Spencer’s theory, for example, is that one can adduce internal and external linguistic evidence that the Quran is just a redaction of a pre-Islamic Syriac holy text (or texts) actually foundational to a Middle Eastern sect of Jewish Christians centuries before (this is also now argued by others, including Karl Ohlig, discussed below). And certainly, Islam is actually just a sect of Christianity (as much as Muslims—and Christians—would be horrified to admit it; Islam is in many ways just the first Mormonism), and clearly descended from the obscure but more original Torah observant wing of Christianity (hence their common requirement of circumcision, their common prohibition against consuming pork, and other details). Other scholars have proposed similar theories (as even acknowledged on Wikipedia), that the Quran is not really all that original, and Spencer is drawing on their previous work—indeed, more experts continue to make similar arguments. But crucially, most others who do, didn’t and don’t conclude this entails Muhammad didn’t exist.

Indeed, Spencer’s theory is compatible with historicity, since Muhammad might have been the very one to do that adaptation and pass it off as an original work—adding, perhaps, details about himself and his use and gradual development of the treatise during his life. Although Spencer claims, “we can glean nothing” from the Quran “about Muhammad’s biography,” because such details are too vague, and it is not “even certain, on the basis of the Qur’anic text alone, that these passages refer to Muhammad, or did so originally” (p. 19), that is a claim that requires a well-qualified expert to assess. Similarly his extensive arguments for there being problems with the earliest evidence, which he argues supports the conclusion that Muhammad was invented later. Assessing all that also requires expertise I don’t have.

Likewise, while even adding a historical Muhammad back into Spencer’s theory would still be damning to fundamentalism (which requires the text to have been revealed by an angel of God and not a half-assed plagiarism of heretical Christian scriptures), even that claim requires vetting by a qualified expert. And since Spencer’s case for it depends on comparing Arabic with Syriac (although he isn’t alone in noting this evidence), the skillset required is not a simple thing to acquire. Spencer himself relies on other purported experts in the language. But that doesn’t excuse their work from the need of vetting.

For those who want to explore Spencer’s case further, the most authoritative critique I’ve found so far is that of Ian Morris, a doctoral student in Medieval Studies who exhibits significant expertise in Syriac (a crucial skill for this debate). His article “Misspelling Muhammad: Why Robert Spencer Is Wrong about Thomas Presbyter” (2014) is not comprehensive, but important to challenging key elements of the Spencer thesis, especially centering as it does on one of the most important evidences. More importantly, it shows the kinds of skills and background knowledge that are needed to engage in this debate overall, and thus why I can’t contribute to it (since I don’t have those skills, and won’t likely ever). Likewise the critical analysis “Muhammad: Man or Myth?” by J. Mark Nicovich at The Catholic World Report, who is a fully qualified professor in Byzantine Studies, and thus brings another key area of expertise to the discussion.

Even so, there are some useful favorable summaries of Spencer’s case to examine, too: most useful I think are: an article by Neil Godfrey at Vridar; a syndicated review by Aymenn Jawad Al-Tamimi (published in various places, including The Middle East Forum and The American Spectator); and an article at Answering Islam by Dallas Roark, which also links to a debate on the topic between Spencer and David Wood. (Although be aware, I think that’s this David Wood, whom I consider an established liar for Christ; the same guy also recently debated ex-Muslim Heina Dadabhoy, and was kind of a douche with them in a subsequent vidcast. So he is not, IMO, the best rep for the historicity side of the Muhammad debate.)

Patricia Crone & Yehuda Nevo

Another book often cited is Hagarism: The Making of the Islamic World by Patricia Crone and Michael Cook, published by Cambridge University in 1977, which argues for similar conclusions about the Quran being an adaptation of earlier Jewish and Christian texts, and the minimal involvement of Muhammad in originating Islam, but they didn’t argue for his non-existence. Their book also remains highly controversial and has not persuaded mainstream experts. Although as we know from Jesus studies, bias could be operating on them more than sense, I am not qualified to ascertain that. But the fact that the authors themselves have evidently backed off some aspects of their thesis after some severely critical reviews is reason enough to approach it skeptically, even though their book is a peer reviewed university publication, and Crone and Cook both are fully qualified PhD’s in the subject (Crone in Islamic Studies, Cook in Middle Eastern Studies). So even their work is not trouble free assistance to anyone who would use it in a case against historicity.

Even so, other scholars have argued for the non-existence of Muhammed, most famously Yehuda Nevo and Judith Koren, in their 2003 book Crossroads to Islam: The Origins of the Arab Religion and the Arab State (Prometheus), which contains a survey of the earliest literary and archaeological evidence available (Nevo was a somewhat qualified archaeologist of the period; Koren, their research assistant), and has had some positive reception from experts, though none are convinced by it. On the other hand, those hostile to its thesis have equated it to Holocaust denialism, the same ridiculous hyperbole found in the Jesus debate, the use of which immediately discredits the opinion of anyone who utters it. And yet even they still make sober enough points that have to be addressed. And again, assessing that, or the merits of other criticisms (e.g. this), requires skills I don’t have.

Ultimately, Crone’s argument (likewise Nevo’s) was that the Quran was assembled later than claimed, which does indeed eliminate Muhammad as its author, and that’s minimally the only thing that his having done renders him relevantly historical. But certain evidence can challenge that (such as recent manuscript finds; the Syriac notes I mentioned; etc.), which then leaves the question of whether their thesis can be modified to account for that. All well beyond my ability to assess. Crone, meanwhile, does not see her theory as entailing the non-existence of Muhammad, as she sees him as just one of many compilers, and that view would work just as easily with Nevo’s position as well.

Muhammad Kalisch

Other historicity challengers now include Muhammad Sven Kalisch, a relatively qualified expert (I believe he at least has full credentials as a professor of Islamic theology), who is writing a book to make his case, and has, so far, produced one article summarizing it. It is only available in German, as “Islamische Theologie ohne historischen Muhammad,” “Islamic Theology without a Historical Muhammad.” In it Kalisch states his position as a historicity agnostic, similar to several well-qualified professors with respect to Jesus. In fact, Kalisch agrees the historicity of Jesus is also doubtable (“Islamische,” p. 9), and he thinks Muhammad might have begun the same way, as an imagined revelatory being, not an actual prophet. On this view the Shahada developed the same way as the Nicene creed: as a shibboleth to test sectarian allegiance.

The destruction of opposition literature in the first two centuries of Islam (“Islamische,” pp. 3-7) does give cover to such a sequence of events. Likewise, claims of descent from the prophet would then be another contrivance to the same end, just as apostolic pedigrees were invented in attempts to seize control of Jesus tradition (“Islamische,” pp. 8-9). Kalisch also sees the late appearance of a Muhammad tradition in coins, epigraphy, and papyri odd on the theory that he existed (“Islamische,” pp. 9-12), in particular as Allah was featured alone for half a century before that on coins, epigraphs, and documents. Except, apparently, some evidence Kalisch claims of a “Muhammad” figure in pre-Muslim Jewish and Christian materials (“Islamische,” pp. 10-11). Again, I lack the skills needed to vet that claim. Although that the name Muhammad means, essentially, “The Ever-Praised One” and thus can be the title of almost any exalted being is indeed a problem for historicists.

Kalisch then goes on to challenge other assumptions about early Islamic history, such as some of its traditional claims of sustained military conquest, in parallel to the now doubted claims of a sustained Hebrew military conquest of Palestine (“Islamische,” pp. 13-14). He proposes a more symbiotic and migratory evolutionary change (albeit with some military events), in which Islam results as an evolution from a newly dominant Torah-observant Christian sect under the Caliphate, with Muhammad originating as in fact a title for Jesus, the cosmic revealer of the “teachings” that came to be assembled into what is now the Quran (“Islamische,” pp. 14-17).

Kalisch attempts to lay out how this could become transformed into a historicized Muhammad for political and sectarian convenience, in order to demote external Christian traditions and set-up a superior source of authority (“Islamische,” pp. 17-18). The new “Muhammad” thus gets the teachings direct from Gabriel, God’s right-hand archangel, bypassing Jesus—perhaps adapting a Gabrielite revelatory tradition that dates all the way back to pre-Christian times, and seizing on the by-then-established precedent that the most recent revelation is God’s truest. Thus what the Christians used to supplant Judaism, the Muslims used to supplant Christianity.

Kalisch then reasons…

Muhammad ist für die Muslime der ideale Mensch, das vollkommene Vorbild. Wenn man es genau nimmt, dann haben sich Muslime niemals wirklich für den historischen Muhammad interessiert. Die Idee einer Befolgung des prophetischen Vorbilds war eigentlich immer schon eine fromme Illusion. In Wirklichkeit namlich hatten Theologen aller Zeiten immer ihr Bildvom “wahren” Muhammad im Kopf, das sie dann als den historischen Muhammad betrachtet haben.

Muhammad is for Muslims the ideal man, the perfect model. Look closely, and you’ll see Muslims have never really been interested in the historical Muhammad. The idea of following the propheticrole model’ has always in fact been a pious illusion. In reality, theologians of all eras just always have a true’ Muhammad in their minds, which they then simply regard as the historical Muhammad.

Thus the “biography” of Muhammad evolved (“Islamische,” pp. 18-20). In much the same way and for the same reasons as the Gospels, in which Jesus is again the model all missionaries are to follow, in their own missions as well as in the stories they tell to explain their understanding of the world.

Kalisch admits he cannot prove this. His argument is more sensibly that it fits all the evidence, particularly evidence the traditional theory does not fit so well, so that at most we cannot rule it out. Sacred histories are simply not reliable. Ever. And yet that’s all we have for Muhammad as a historical man (as opposed to a revelatory being). And there is some odd stuff left over (such as a strange disinterest in Muhammad for the first half century or so, and evidence of a pre-Muslim Muhammad tradition). So we are left with enough uncertainty to have a reasonable doubt.

Kalisch’s case thus sounds more credible and well thought out to me than Spencer’s, but I am not qualified to know for sure, particularly at key points where he makes claims about the surviving evidence or the lack thereof. One doubt I have, for example, is that Kalisch does not address some of the best evidence for historicity in his article, such as the Syriac notes. Hopefully he will address that in his book. And hopefully his book will make it into English.

And More

Then there is Hans Jansen, also a relatively qualified expert (he has a doctorate in Arabic; though he appears to have specialized in contemporary, not Medieval, studies), but his reasons for doubting the historicity of Muhammad do not appear well thought-out. Although I am only basing this judgment on his summary; I have not read his book, Mohammed: Eine Biographie. But in the linked article, it doesn’t look good. He confuses evidence against the later and definitely mythical Mohammad as evidence against the original, minimally historical Muhammad (out of whom the myths were constructed). And he too casually leans on the argument from silence, not addressing the complications attending that (such as the fact that the record might not be so silent, e.g. he does not seem aware of the Syriac notes; and that we still have to resolve the question of evidence expectancy for minimal historicity; and so on).

I should also mention Ibn Warraq. Even though I cannot ascertain that he has any qualifications beyond a B.A. in Islamic studies, he has written or edited several books critically examining the subject of Muhammad’s historicity, in turn citing other scholarly work, or assembling it in one place, so these are probably crucial starting points for examining the broader debate. Most importantly is his anthology The Quest for the Historical Muhammad (Prometheus 2000), although also relevant are Which Koran? Variants, Manuscripts, Linguistics (Prometheus, 2008) and Koranic Allusions: The Biblical, Qumranian, and Pre-Islamic Background to the Koran (Prometheus, 2013). Notably, Warraq also wrote the new introduction to the paperback edition of Spencer.

There are also two other edited works that collect articles by several experts doubting (or at least willing to doubt) the historicity of Muhammad, which engage in some detailed analysis of key items of evidence, and lend support to key arguments of Kalisch, so I think anyone who wants to explore the matter certainly must consult them: The Hidden Origins of Islam: New Research into Its Early History (Prometheus, 2009), edited by Karl Heinz Ohlig and Gerd R. Puin; and Early Islam: A Critical Reconstruction Based on Contemporary Sources (Prometheus, 2013), also edited by Ohlig (and summarized here). Ohlig is a retired professor of religious studies; Puin, an established expert in Arabic manuscripts. Their contributors are likewise qualified experts. So this is not coming from amateurs. But I would still need to see competent and honest responses by well-qualified opponents to assess it all.

The Case for Historicity

For contrast, I do not know what the best books are that argue for the historicity of Mohammad. I have found reference to only one that even focuses on the question, and that not even directly, and it is not available in English: Tilman Nagel’s Mohammed: Leben und Legende (2008), which sounds a lot like your average “historical Jesus” book—based on presumptions of historicity, rather than demonstrations of it. And we need a book that directly takes on the question.

The closest I have found in English is a short but recent article by the very Patricia Crone, “What Do We Actually Know about Mohammed” (2008), which is good, but hardly comprehensive (and not sufficiently critical of the Syriac notes to be sure she has properly vetted them as to date and content). Likewise the about ten pages surveying the matter in the introduction to The Historical Muhammad by Irving Zeitlin, which is probably worth reading as being representative of the standard mainstream view and its basis (much like the handful of pages in Van Voorst are for the Jesus question: hence for comparison see my discussion of the latter in On the Historicity of Jesus, pp. 4-7).

But perhaps equally important is an articulation of the most defensible hypothesis of minimal historicity for Muhammad, since the doubter of historicity must render improbable the most defensible alternative. Patricia Crone, again, has provided that, along with another expert co-author Martin Hinds, in God’s Caliph: Religious Authority in the First Centuries of Islam (Cambridge University, 1987). There the idea is maintained that Muhammad was just one prophet among many in the production of the Quran, and simply the last one the list was winnowed down to later when a singular authority was more convenient to claim.

The most defensible minimal mythicism for Muhammad, meanwhile, is any basic revelator-authority model as followed by the Jews in creating Moses, and the Christians in creating Jesus. The fewer assumptions attached to it the better.

Conclusion

In the end, best I can tell (and I am not qualified to tell with much confidence), it is at least significantly more probable than not that a guy named Mohammad existed, and cobbled together the Quran, perhaps adapting earlier writings from a Torah observant Christian sect, and perhaps not alone, and perhaps even at someone else’s behest (e.g. Crone & Cook propose he was simply working as an assistant to Umar in this respect, and elevated to prophet status later for convenient propaganda). But that’s at most.

And here I am imagining a contest between minimal historicity and mythicism for Muhammad. That is, what it would take for him to merely have existed and constructed (at least some of) the first Quran. The rest of his biography could still be bullshit. But proving that wouldn’t be enough to conclude he didn’t exist at all. Just as for Jesus, where one must separate the obviously mythical Jesus of the Gospels (as all mainstream experts agree, that guy is a myth) from a modern historical reconstruction of the most probable actual Jesus (which mainstream experts still struggle to do and debate the details of), so also one must separate the probably mythical Mohammad of the later Sira and Hadith, from a needed modern historical reconstruction of the most probable actual Mohammad. And the latter could well consist of nothing credible beyond that he lived and composed some of the Quran.

Whether one can go further and flip the probability the other way, or close enough to 50/50 to entail agnosticism, depends on two things: (1) Whether you can establish a low rather than a high prior for Muhammad’s existence (initially, I would say there is not a strong enough precedent for fabricating a scribal “prophet” for the Quran, e.g. Umar could just have claimed to have received and transcribed the revelation himself, so I would assign a high prior to Muhammad’s historicity, although again I only barely know enough to do this—Kalisch and Ohlig, for example, do seem to make a good case regarding the history of Muhammad’s name, which would put him in a more myth-heavy reference class); and/or (2) Whether you can show that the evidence is less likely on historicity, or at least no more likely.

And the latter can only happen once all the evidence has been fully collected and analyzed. And it does not appear that this has happened. For instance, there is no mainstream peer reviewed monograph for Mohammad equivalent to Van Voorst’s for Jesus. Nor has a good survey been done, that I know of, of what evidence we should expect but don’t have, and how likely we should expect to have it (e.g. why we have contemporary inscriptions for major leaders of Islam who were later said to be companions of Muhammad, but no inscriptions for Muhammad). And even then, once all that is done, the question must be asked: How likely is that evidence’s collective existence (including what is absent) if Mohammad minimally existed, and how likely is it if he didn’t? The ratio between the latter two probabilities gives us the likelihood ratio (or Bayes factor) favoring historicity. When that ratio is multiplied by the ratio of the prior probabilities (for and against historicity), you get the odds that Muhammad minimally existed.

No one has done that yet. And I am not qualified to do it myself.

Discover more from Richard Carrier Blogs

Subscribe now to keep reading and get access to the full archive.

Continue reading