In the British Library collections is an eighth-century manuscript of a seventh-century text in Syriac, attributed to Thomas Presbyter. It’s a funny little jumble of geography, genealogy, chronicle and king-list. Why it survived I don’t know. But we’re very lucky it did, because it contains the earliest known mention in a non-Muslim text of the prophet Muhammad.
In the year 945 [=634], indiction 7, Friday 4 February at the ninth hour, there was a battle between the Romans and the Arabs of Muhammad in Palestine twelve miles east of Gaza. The Romans fled, leaving behind the patrician bryrdn(?), whom the Arabs killed. Some four thousand poor villagers of Palestine were killed there, Christians, Jews and Samaritans. The Arabs ravaged the whole region.
(Adapted from Palmer, West-Syrian Chronicles, 18-9; Hoyland, Seeing Islam, 120. Syriac text: Chronica Minora 2, 147-8).
There’s a lot to be said about this fascinating and troublesome little passage. What battle is this? Who is the patrician? Is bryrdn a name or a place? Are we to infer that Muhammad was still alive and leading an invasion, two years after (tradition tells us) he died?
But I’m not going to consider any of these questions today. Right now, I feel the need to rebut Robert Spencer’s maladroit interpretation of this source.
Robert Spencer’s book, Did Muhammad Exist?, was published 2012 by the American neoconservative Intercollegiate Studies Institute, to good reviews from other neoconservatives; as far as I can tell, it was ignored by most national press. However, in my experience, enough people have read his book and believe his claims that it’s worth engaging with them from time to time.
In Spencer’s view, this source “may not be an early reference” after all, perhaps having been added later; and besides, “it is not certain (although it is possible)” that the text really refers to Muhammad. It’s a rather small point within the book; but small mistakes can easily build to large, wrong-headed arguments. And the symbolism is great, since it’s one of our most important early sources that he so flippantly problematises. Best of all, he does so by misreading (genuine) scholars and then suggesting, bizarrely, that ‘Muhammad’ could not easily be misspelt by a foreigner.
It’s so ridiculous I just couldn’t help myself.
Seeing ‘Seeing Islam as Others Saw It’ as others saw it
Spencer starts off by questioning the date of the text.
“But some evidence indicates that these writings were revised in the middle of the eighth century, and so this may not be an early reference to Muhammad at all.”
He backs this up with a footnote quoting Robert Hoyland’s excellent study, Seeing Islam as Others Saw It, 119:
“A mid-seventh century Jacobite author had written a continuation of Eusebius and… this had been revised almost a century later when the lists of synods and caliphs and so on were added”.
That does sound problematic. But if Spencer had read Hoyland more carefully, he would have seen that the “list of synods and caliphs and so on” was the last folio of the manuscript, which looks like an addendum. The mention of Muhammad belongs to the earlier bulk of the text, which does not have any tell-tale signs of later editing. Andrew Palmer (West-Syrian Chronicles, 5) explains:
“…the last folio of the manuscript stands apart from the rest. This is a list of caliphs, translated from the Arabic… Nothing else in the manuscript need have been written after AD 640. Indeed, the scribe marked the last folio off as a new beginning by placing the rubric: ‘It is finished’ at the end of what preceded it.”
Palmer goes on to explain that the author of the earlier text shows no awareness of Heraclius’ death in 641, or of anything that happened thereafter, so we can take it that 640-ish is the date of composition. This is all perfectly sensible, and Spencer would have known it if he had read Hoyland carefully. In fact, if he’d followed Hoyland’s footnotes – like a professional – he would even have found Palmer’s clarification. Spencer has managed to convince laypeople that he’s some sort of scholar, but he doesn’t really know how scholarship is done.
By the way, a hallmark of the amateur revisionist is that Spencer picks up on what he (wrongly) thinks is a note of doubt, while neglecting Hoyland’s very positive assessment:
“This is the first explicit reference to Muhammad in a non-Muslim source, and its very precise dating” – 4 February was indeed a Friday – “inspires confidence that it ultimately derives from first-hand knowledge.”
You say potato…
But Spencer hasn’t finished with his obtuse reading of this source. He observes that Hoyland
“has translated tayyaye d-Mhmt as “the Arabs of Muhammad”; this translation and similar ones are relatively common. Syriac, however, distinguishes between t and d, so it is not certain (although it is possible) that by Mhmt, Thomas meant Mhmd—Muhammad.”
I should clarify what’s meant here. Syriac, like Arabic, does not usually represent short vowels, so don’t be thrown off by that. The issue is that the Syriac author has chosen an ‘emphatic’ t instead of an ‘unemphatic’ d. Since both languages have both of these sounds, why would there be any confusion between them?
Quite a few reasons, it turns out. It might simply be that the pronunciation Muhammad was uncomfortably foreign, even if the phonemes were individually native. My own name is a good example of this phenomenon: one by one the phonemes that make up Ian are pretty common, but their combination is pretty rare outside of English. I’ve been Éon in France, Eien in Germany, Eeenee in Egypt (no, I don’t know either) and Yan absolutely everywhere. In my name it’s the vowels that are twisted; between Arabic and Syriac it might be consonants.
Or perhaps Muhammad was too familiar, and the Syriac author was guilty of hyperforeignism: when speakers mangle a word so it sounds more like a foreign word ‘should’ sound. English speakers often do this with the Che in Che Guevara, turning it from chay into shay – perhaps influenced by the French chez. The fact that English contains both ‘ch’ and ‘sh’ sounds is neither here nor there. This is a surprisingly common feature of language. You may not know it, but you do this all the time.
Or maybe the emphatic t in Syriac is an attempt to mimic the d sound in Arabic for reasons of pronunciation that nevertheless aren’t clear from the script alone. We can see this in Pinyin, a popular system for transcribing Mandarin Chinese. Mandarin distinguishes between aspirated, ‘hard’ consonants and unaspirated, ‘soft’ consonants. English does have both aspirated and unaspirated consonants, but native English speakers aren’t consciously aware of the difference. However, we are more likely to aspirate unvoiced consonants (like t) than voiced consonants like d). So Pinyin uses unvoiced letters to stand for aspirated consonants, and voiced letters to stand for unaspirated consonants. This is why there are two ways of spelling Daoism-Taoism, one Pinyin and the other not, both sort-of correct. Spencer’s insistence that a spoken Arabic d be written with a Syriac d assumes a one-to-one correspondence of oral and written language that in reality is never so precise.
Neither Hoyland nor Palmer obsesses over the spelling: the difficulty of transcribing foreign words is taken for granted in philology. Spencer is placing an unreasonably high burden of proof on our medieval sources.
Mo money mo Mhmt
Any of these could account for the unfamiliar spelling. But there’s an even better reason to be sure that Mhmt is Muhammad. Spencer doesn’t mention – presumably he doesn’t know – that the spelling of Muhammad as Mhmt is very common indeed. It appears on Arab-Sasanian coins of the seventh and eighth centuries with inscriptions in the Pahlavi language, written in an Aramaic script closely related to the Syriac script of Thomas Presbyter’s manuscript. These coins bear the name Mhmt at the same time that Muhammad is first appearing in Arabic inscriptions.
Sometimes the Pahlavi Mhmt and the Arabic Muhammad are written on the same coin; like this one from Ohlig and Puin (eds.) Hidden Origins of Islam (2008), 56:
Not only that; we even have a very early coin (70 AH = 689 CE) that bears the familiar Islamic slogan “Mhmt is the messenger of God” in Pahlavi.
These Arab-Sasanian coins are fairly common knowledge, and they prove beyond reasonable doubt that Mhmt is another way of spelling Muhammad.
Seek knowledge, even in China; but not there
For students and enthusiasts of History, there are a few lessons to be gleaned from this episode. Read scholarly literature very carefully. Always follow the footnotes: if one scholar’s argument looks a bit thin, perhaps she cites someone else who gives a full explanation. Don’t neglect numismatics. Study languages, ideally the languages of your sources, and think about what happens when languages interact. Talk to historians: ask them why they read sources as they do. Be critical, be sceptical, but don’t be obtuse. And above all, please ignore Robert Spencer.
Edit (23/12/14): Thanks to some very clever readers, I can now explain exactly how these alternate spellings (probably) came about.
Robert Spencer, Did Muhammad Exist? (2012), ch. 1 and n. 5.
Andrew Palmer (ed.), The Seventh Century in West-Syrian Chronicles (1993), 5-24, 49-50.
Robert Hoyland, Seeing Islam as Others Saw It (1997), 118-20.
E.W. Brooks (ed.), Chronica Minora 2 = Scriptores Syri 3:4 (1904), 76-154 (full Syriac text traditionally attributed to Thomas Presbyter).