Thursday, 14 June 2018

The number of the beast

The Christian bible reports ‘the number of the beast’ as follows:

This calls for wisdom: let anyone with understanding calculate the number of the beast, for it is the number of a person. Its number is six hundred and sixty-six.
Revelation 13.18 (NRSV translation)

This number is a tremendously popular icon in modern culture — and in popular religion, too. Its popularity owes a lot to the repeating sixes: ‘6-6-6’. This snippet from Doctor Who relies on it, for example:

Doctor. To generate that gravity field, and the funnel, you’d need a power source with an inverted self-extrapolating reflex force of 6 to the power of 6 every 6 seconds.
Rose. That’s a lot of sixes.
The Doctor, about to encounter the Devil (Doctor Who, ‘The impossible planet’, 2006)

This six-ishness depends on a decimal counting system. And not just that: it has to be a Hindu-Arabic-style decimal system. You have to take a very specific kind of notation for granted.

Not enough of a myth for you? Well, how about this one. A fairly popular story floating around these days is that the number of the beast wasn’t actually 666, but 616. This is supposedly proved by the earliest papyrus of Revelation, which was published in 1999; the idea was popularised in a 2003 BBC documentary.

Well, none of that’s true. Not quite true, anyway. So, yes, there are some catches.

  1. Hindu numerals started to appear in the 7th century in India; Revelation was written about 600 years earlier in the eastern Mediterranean. Now, some ancient Mediterranean cultures did have decimal numerals — Greek Ionic numerals, and Hebrew numerals — but they were place-dependent. They had different symbols for 6, 60, and 600. You used different symbols depending on whether you were writing a number in the ones column, the tens column, or the hundreds column. The Greek notation for 666 is χξϛ, not ϛϛϛ; in Hebrew it’s םסו‬, not ‬‬‬‬‬‬‬‬‬ו‬ו‬ו‬. Repeating the symbol for ‘6’, ϛϛϛ or ‬‬‬ו‬ו‬ו‬, would be meaningless.
    A spectrogram: the computer game Doom (2016), premised on a demon invasion from hell, has the numerals ‘666’ embedded in part of the soundtrack. This wouldn’t work in any ancient numeral system.
  2. A prophetic interpretation, which sees the text as a genuine prophecy of a specific later time, is never going to have as much staying power as a historical-critical interpretation. This is an ancient Greek text which tells us explicitly that it contains a coded message: you’d need some very powerful and specific reason to imagine that the message should work just as well for, say, a 21st century American (who doesn’t speak ancient Greek) as for a 1st century Christian (who did). That way lies madness. But in a 1st century Greco-Jewish context, as we’ll see, the number is meaningful and based on well-attested practices.
  3. The 616 variant is certainly ancient: Irenaeus knew about it in the late 2nd century, though he strongly preferred 666. As we’ll see below, both readings are based on the same reasoning. But, so far as we can tell, neither number is older than the other. 616 is moderately well supported, but it definitely doesn’t have earlier support than 666.

The text

Here’s what the number of the beast looks like in two early manuscripts of Revelation:

Left: codex Sinaiticus, quire 90 fol. 6r (4th century). Right: Chester Beatty Library P. Bibl. 3 (= P47 Gregory-Aland), fol. 7r (3rd century).

The left-hand image is from one of the most important manuscripts of the New Testament, the famous codex Sinaiticus, written in the 4th century. Its version of verse 18 reads:

ὧδε ἡ σοφία ἐστίν· ὁ ἔχων [[ουσ]] 'νοῦν' ψηφισάτω τὸν ἀριθμ'ὸ(ν)' τοῦ θηρίου, ἀριθμ'ὸς' γὰρ ἀν(θρώπ)ου ἐστίν· ἑξακόσιαι ἑξήκ'ο(ν)'τα ἕξ.

Here’s a bit of cleverness. Let someone who has sense calculate the number of the beast, for it is the number of a person: six hundred and sixty-six.

(Key: [[...]] = crossed out by the scribe, '...' = inserted above the line, (...) = abbreviation.)

There are a few differences from the modern critical text. The scribe has made one error, crossed it out, and written νοῦν (‘understanding’) in the margin. The phrase ‘and its number is’ (καὶ ὁ ἀριθμὸς αὐτοῦ) is missing here. And most importantly, the number itself is written out as words, ἑξακόσιαι ἑξήκοντα ἕξ (‘six hundred and sixty-six’), not as the numeral χξϛ (‘666’).

The right-hand image is the earliest known copy of the passage: a fragmentary papyrus written in the mid-3rd century and currently held in Dublin, Ireland. Here verse 18 reads:

ὧδε ἡ σοφία ἐστίν· ὁ ἔχων νοῦν ψηφισάτω τὸν ἀριθμὸν τοῦ θηρίου, ἀριθμὸς γὰρ ἀνθρώπου ἐστίν· ἐστὶν δὲ χξϛ.

Here’s a bit of cleverness. Let someone who has sense calculate the number of the beast, for it is the number of a person: and it is 666.

This is a closer match to the modern text. But again, the words immediately before the number are different — ἐστὶν δὲ (‘and it is’), rather than the modern critical text καὶ ὁ ἀριθμὸς αὐτοῦ (‘and its number (is)’) —; and this time, the number is written with numerals, χξϛ (‘666’).

‘The number of a person’

It’s fairly widely known that the number was designed as a reference to the Roman emperor Nero.

1. Isopsephy. The most specific point is that the number 666 is plainly a case of isopsephy. Ancient Greek and Hebrew used letters of the alphabet as numerals: nine letters for 1 to 9, nine for 10 to 90, and nine for 100 to 900, for a total of 27 numerals. In isopsephy, the number of a name or phrase is obtained by treating its letters as numerals and adding up their values. In a Jewish context, this practice is known as gematria.

We have independent evidence of isopsephy in relation to Nero. Suetonius (Nero 39) quotes a Greek epigram where Nero’s name, transliterated into Greek,

Νέρων = Nerōn ‘Nero’
= 50 + 5 + 100 + 800 + 50
= 1005

is equated with

ἰδίαν μητέρα ἀπέκτεινε = idian mētera apekteine ‘killed his own mother’
= 10 + 4 + 10 + 1 + 50 + 40 + 8 + 300 + 5 + 100 + 1 + 1 + 80 + 5 + 20 + 300 + 5 + 10 + 50 + 5
= 1005

The epigram calls this a neopsēphon, a ‘new calculation’ — the same word as in Revelation 13.18 psēphisatō ‘let (him) calculate’. Isopsephy is attested in Judaeo-Christian contexts too, notably in a set of poems called the Sibylline Oracles; and a 1st-3rd century CE Jewish apocalyptic text known as 3 Baruch uses equations based on Greek words transliterated into the Hebrew script.

  • Sib. Or. 1.324-9 uses the number 888 = Ἰησοῦς, ‘Jesus’ in Greek.
  • Sib. Or. 5.12-44 gives a list of Rome’s rulers numbered from their initials: e.g. 12-13 ‘he will be the very first king, who will sum twice ten / with the start of his name ... he will have his first letter from 10’, that is, 20 = Κ(αῖσαρ), 10 = Ἰ(ούλιος), for Julius Caesar. The text carries on with Augustus = α = 1, Tiberius = τ = 300, Gaius = 3 = γ, Claudius = κ = 20, Nero = ν = 50; then three rulers with short reigns; then Vespasian = ο = 70, Titus = τ = 300, Domitian = δ = 4, and Trajan = τ = 300.
  • 3 Baruch ch. 4 discusses a dragon in connection with 360 rivers, where ‘dragon’ δράκων > דרקון drqwn, which adds up to 360; and a cataclysm involving 409,000 giants, where ‘cataclysm’ κατακλυσμός > קטקליסמס qtqlsms, or 409. (Bear in mind that the Hebrew script doesn’t have vowels.)
  • Koester 2014 has more evidence on the use of isopsephy/gematria in ancient graffiti, inscriptions, and epigrams.

Revelation follows the same procedure as 3 Baruch. It transliterates the Greek Νέρων into the Hebrew script and calculates with Hebrew numerals.

‘Nero Caesar’ = נרונ קסר Nrwn Qsr
= 50 + 200 + 6 + 50 + 100 + 60 + 200
= 666

In favour of this interpretation:

  1. The word ψηφισάτω (‘let him calculate’) plainly implies isopsephy.
  2. The same reasoning explains the variant reading ‘616’. It is based on the Latin form of Nero’s name: Nrw Qsr. (See below.)
  3. A papyrus from Qumran contains Nero’s name in Aramaic in the form Nrwn Qsr, so the orthography inferred here is legitimate, though not common.
  4. Two ancient discussions retain a memory of the link between Revelation and Nero, though neither mentions this isopsephic interpretation. The Liber genealogus (5th cent.; 194-6 ed. Mommsen) is aware that the number of the beast refers to Nero. Victorinus of Pettau (3rd cent.; Commentary on Apocalypse 17: 11) knows of a link between Revelation and Nero, and (at 13.18) knows both the 666 and 616 readings, though he explains the two numbers differently.
  5. The Greek word ‘beast’ can also be transliterated into Hebrew as θηρίον > תריון trywn = 666. The form actually used in Revelation 13.18 is a possessive, θηρίου, which gives תריו tryw = 616. With both 666 and 616, the text is equating ‘Nero’ and ‘beast’ in the same way that Suetonius’ epigram equates ‘Nero’ and ‘killed his own mother’.

Against this interpretation, there really isn’t much to say. One objection might be that Caesar/Καῖσαρ would normally be written in Hebrew as קי‬סר (Qysr), not ק‬סר (Qsr) (for a total of 676). On the other hand, a number of 1st-3rd century documents do attest the spelling ק‬סר (Qsr). Other isopsephies have been suggested, but they’re not strong. An abbreviated form of the emperor Domitian’s name, Α. Και. Δομετ. Σεβ. Γε. (for αὐτοκράτωρ Καῖσαρ Δομετιανὸς Σεβαστὸς Γερμανικός) adds up to 666. The emperor Gaius was justly detested by 1st century Jews, and his name adds up to 616 in Greek (Γάϊος Καῖσαρ). The Domitian one with the abbreviation looks like a stretch, though. And while Gaius was very objectionable to Jews, he wasn’t especially interesting to early Christians. Nero is by far the strongest candidate.

2. Nero as boogeyman. There is pretty good evidence that Nero was a figure of terror for late 1st century Christians. Early 2nd century sources blame him for the first persecution of Roman Christians, claiming that he made them scapegoats for the great fire of Rome in 64 CE. There is some doubt as to whether the persecution actually happened, but even if it didn’t, the 2nd century sources show that there was an early belief that it did. The sources are the Roman historian Tacitus (Annals 15.44.5-8; Latin text): he reports that Nero had Christians ‘clothed in the skins of wild beasts, and torn to pieces by dogs’, or ‘set up to be burned, so as to serve the purpose of lamps when daylight failed’ (these gruesome tortures are quite likely fictional). Suetonius reports on a persecution under Nero too, though in his version it had nothing to do with the fire (Nero 16).

3. Nero redivivus. Nero may have been a menacing figure for Christians, but overall he was a tremendously popular figure in the eastern empire. Suetonius claims that after his death some people acted ‘as though he was alive and was soon to return, to the great harm of his enemies’ (Nero 57). In other words, he was to have a second coming. We have reports of at least two, maybe three ‘false Neros’, who pretended that they were Nero, that Nero had never really died, and gained a significant following. (This wasn’t unique to Nero: Tacitus reports on other people being impersonated after their death, most notably Scribonianus in Histories 2.72.)

In 68 or 69 CE, shortly after Nero’s death, one false Nero appeared in Greece and, after landing on the Aegean island of Kithnos because of a storm, turned to piracy until he was captured and killed by Roman forces. (Source: Tacitus, Histories 2.8-9.)

A second false Nero, whose real name was Terentius Maximus, appeared during the reign of emperor Titus in 79-81 CE. He was active in Anatolia, Mesopotamia, and possibly Armenia, and sought refuge with the Parthians, apparently getting as far east as the river Euphrates. (Sources: Cassius Dio epit. 66.19.3 = Zonaras Epit. hist. 11.18; John of Antioch fr. 104 Müller.)

The third one is less well attested. The main evidence for him is Suetonius’ claim that there was a false Nero ‘twenty years after (his death), while I was a young man’ (Nero 57). This would suggest a date around 89 CE, during the reign of Domitian. However, Suetonius mentions the Parthians too: he states (1) that the Parthian king Vologaeses II (fl. ca. 78 CE) asked the Roman senate to honour Nero’s memory; and (2) that the Parthians supported the third false Nero ‘vigourously’ and ‘returned him reluctantly’. This sounds awfully like the second one, above: there could be a confusion of dates. One point in favour of Suetonius’ story is that Tacitus’ report, of the first false Nero above, refers to ‘results and attempts of others’, plural. That implies that he knew of at least three false Neros.

The neighbour of the beast: Pokey (Don McKellar) calls on a wannabe Devil (Earl Pastko) who can’t get his act together. (Highway 61, 1991)

Which number?

Some early copies give a different number. In particular, the Ephrem palimpsest (5th century) gives it as ἑξακόσιαι δέκα ἕξ: ‘six hundred and sixteen’.

This has a pretty strong heritage. As I mentioned above, Irenaeus knew both variants. And 616 works as an isopsephy of Nero’s name just as well as 666 does:

‘Nero Caesar’ = נרו קסר Nrw Qsr
= 50 + 200 + 6 + 100 + 60 + 200
= 616

The Latin form of his name, Nero, gives 616; the Greek form Nerōn (Νέρων) gives 666.

In recent years the 616 variant has enjoyed a lot of prestige, because of an ancient papyrus published in 1999. The papyrus, found at Oxyrhynchus and catalogued as P. Oxy. 4499, dates to the late 3rd or early 4th century. We have 26 fragments of the papyrus, containing bits of Revelation chapters 2 to 15. Fragment p is the relevant one.

P. Oxy. 4499 (= P115 Gregory-Aland), fragment p (Revelation 13.18-14.2)

This papyrus, all by itself, has generated two modern myths:

  1. This papyrus unambiguously supports the reading 616 (highlighted in yellow above: χιϲ).
  2. This papyrus is the earliest existing copy of the passage, or even the earliest existing copy of Revelation, and therefore its text carries more weight than any other manuscript.

These two myths, taken together, would suggest that 616 was definitely the original reading and 666 is a corruption of the text. However, they are both absolutely completely dead wrong.

First, here’s a transcription. I’ve copied the original editor’s text, with some minor adjustments because I think some of the letters are clearer than the original editor thought. (For technical side of how a papyrus text is represented, you can refer to this post from last month.)

Diplomatic text:

1               ]υ̣νψ̣ηφιϲ̣[
2                ]αρανου[
3                    ]ηχιϲ[
4            ]επ̣ιτοορ[
5              ]εχουϲα[
6               ]α̣υτουγε
7             ]α̣υτωνκαι
8              ]φω̣ν  ̣νϋ

Reading text as printed by the editor, Juan Chapa, in The Oxyrhynchus papyri:

1   [(13.18) ὧδε ἡ ϲοφία ἐϲτίν· ὁ ἔχων νο]ῦ̣ν ψ̣ηφιϲ̣[ά]-
2   [τω τὸν ἀριθμὸν τοῦ θηρίου, ἀριθμὸϲ γ]ὰρ ἀν(θρώπ)ου
3   [ἐϲτίν· καὶ ὁ ἀριθμὸϲ αὐτοῦ χξϛ] ἢ χιϛ. [ ]
4   [(14.1) καὶ εἶδον, καὶ ἰδοὺ τὸ ἀρνίον ἑϲτὸϲ] ἐπ̣ὶ τὸ ὄρ[οϲ]
5   [Ϲιών, καὶ μετ’ αὐτοῦ ρμδ χιλιαδε]ϲ̣ ἔχουϲα[ι]
6   [τὸ ὄνομα αὐτοῦ καὶ τὸ ὄνομα τοῦ π(ατ)ρ(ὸ)ς] α̣ὐτοῦ γε-
7   [γραμμένον ἐπὶ τῶν μετώπων] α̣ὐτῶν. (14.2) καὶ
8   [ἤκουϲα φωνὴν ἐκ τοῦ οὐ(ρα)νοῦ ὡϲ] φω̣νὴ̣ν ὑ-
8a  [δάτων ... ]

And my translation of Chapa’s text:

1    [(13.18) Here’s a bit of cleverness. Let someone who has sen]se calcu[l]-
2    [ate] the number of the beast, fo]r [it is the number] of a person:
3    [and its number is 666] or 616. [ ]
4    [(14.1) And I looked, and behold, the lamb standing] on Mou[nt]
5    [Zion, and with it were 144 thousan]d (people) who ha[d]
6    [its name and the name of] its [father] wr-
7    [itten on] their [foreheads.] (14.2) And
8    [I heard a voice from the sky, like] the voice of w-
8a    [aters ...]

Here’s the text and translation as they appear in modern editions. Before we get on to the number, some technical notes (skip over them if you want):

1. In Roman-era papyri, there’s often no visual distinction between the numerals ϛ ‘6’ (the letter stigma) and ϲ ‘200’ (sigma). However, it’s easy to tell which is which by looking at their position.

2. Chapa’s supplements stick to the standard text fairly rigourously, except for (a) numerals replacing numbers written out as words, in the missing parts of lines 3 and 5; (b) some normal abbreviations in the missing parts of lines 6 and 8. However, the result is that the number of letters per line is erratic, ranging from 28 (line 3) to 35 (lines 2 and 6). These could be evened up if we imagine some textual variants: for example line 2 could omit τοῦ, by analogy with ἀνθρώπου later in the line, bringing the line down to 32 letters; line 5 might have written out πατρὸς in full but omitted the repetition of τὸ ὄνομα, bringing it down to 31. Line 4 is also longer than I’d like, at 34 letters.

OK, on to line 3 and the numeral. Let’s see how Chapa arrived at this text. Take a look back at the image above, where the number of the beast, χιϲ ‘616’, is highlighted. Notice the letter immediately before the number? It’s the letter eta, η. That letter raises an important question of its own.

The problem is that there is nothing that eta can possibly correspond to in the standard modern text of Revelation 13.18. What is it doing there?

There are only a few things an eta before a word-break can be:

  • certain forms of a-stem (1st declension) feminine nouns and e-stem (3rd declension contracted) neuter nouns with endings in eta
  • a verb in a specific form with an ending in eta (the aorist passive 3rd person singular)
  • ἡ ‘the’, in a specific singular form
  • δή ‘indeed’
  • ἤ ‘or’

Revelation 13.18 has no occasion for any 1st or 3rd declension nouns or aorist passive verbs, and you can’t have a singular ‘the’ before the number 616. So it’s going to be δή or ἤ. The text is either

[ ... δ]ὴ χιϲ
[ ... ] indeed 616


[ ... ] ἢ χιϲ
[ ... ] or 616

The clearly correct interpretation, as Chapa realised, is that the eta before the number of the beast can only be the word ἤ ‘or’. The papyrus originally quoted both numbers, 666 and 616, as alternatives. And this is why Chapa’s edition correctly prints

[ ... καὶ ὁ ἀριθμὸϲ αὐτοῦ χξϛ] ἢ χιϛ.
[ ... and its number is 666] or 616.

So, no more talk of 616 being attested earlier than 666, please. We can be absolutely confident that they were both written in this papyrus.

And one last thing: there’s still that myth that P. Oxy. 4499 is the earliest existing copy of Revelation. Well, it isn’t. It’s the earliest existing copy of some parts of the book. But not this part.

The misunderstanding may be partly due to this phrasing used on the Oxyrhynchus Papyri website:

The newest volume of Oxyrhynchus Papyri contains a fragmentary papyrus of Revelation which is the earliest known witness to some sections (late third / early fourth century).

If you stop paying attention after the word ‘witness’ you could easily come away thinking this is the earliest copy of Revelation. Even if you carry on reading, you might think it’s implied that it’s the earliest copy of the ‘616’ passage. But it ain’t. That honour goes to P47, above, which is solidly 3rd century, and which has the reading χξϛ ‘666’.

Even if we didn’t have P47, we’d have Irenaeaus’ discussion of the passage — and he’s earlier than both papyri. He knew both the 616 and 666 variants, though he didn’t know the reason for either number. Both of them are very satisfactorily explained by the equations

Νέρων Καῖσαρ = נרונ קסר Nrwn Qsr = 666 = תריון trywn = θήριον ‘beast’

Nero Caesar = נרו קסר Nrw Qsr = 616 = תריו tryw = θηρίου ‘of the beast’

Further reading

  • Chapa, J. 1999. ‘4499. Revelation.’ The Oxyrhynchus papyri 66: 10-37.
  • Koester, C. R. 2014. ‘The number of the beast in Revelation 13 in light of papyri, graffiti, and inscriptions’. Journal of early Christian history 6.3: 1-21. [Taylor & Francis]
  • Parker, D. C. 2000. ‘A new Oxyrhynchus papyrus of Revelation: P115 (P. Oxy. 4499)’. New Testament studies 46: 159-174. [Cambridge Core]

Addendum (July 2022). This piece is still occasionally getting new readers, so I have made some edits to correct some links, tidy the html, and most importantly, to make it explicit that the reading text of P115 shown here is Chapa’s text, and not my own doing; I have also added a translation of Chapa’s text.


  1. While writing three sixes only works in Arabic notation, there is still a repeating of the element "six" when reading the number in base 10 languages like English or, for that matter, Greek (ἑξακόσιαι ἑξήκοντα ἕξ). As far as I know, Hebrew and Aramaic are base 10 as well, but I don't know them well enough to say whether the element 6 would be sufficiently unchanged in forming the decades and the hundreds that it would give a similar repetition effect (e.g., the repetition effect is diminished in Latin sescenti sexaginta sex due to the special shape of the numeral 6 used in "600".

    1. Quite right to point that out: I should have made that clear. Still, the repetition of the numerals is the most typical feature in representations of the number, so I do think my point holds up in spirit, at least!

  2. Thank you for this discussion. There's a lot of utter foolishness out there on the internet regarding this topic. I'm satisfied (for the moment) with your thoroughness and can make reference to this page in future consideration.

  3. Sorry, there seems to be some confusion here. This isn't what Chapa proposes at all. He reports a horizontal bar above the eta, possibly a deletion mark, and a trace before eta that would suit kappa (p. 34).