Thursday 28 July 2022

Latin spells in classic D&D video games

Video games based on the rules and world of Dungeons & dragons have been a staple for a long time. Three games released in 2000 to 2002, widely regarded as classics of the genre, draw on the traditional pairing of magic and the Latin language: Baldur’s Gate II (2000), Icewind Dale (2000), and Icewind Dale II (2002). Ancient Romans didn’t accidentally set off fireballs every time they had a conversation. But in these games, when characters cast spells, you hear the verbal component of the spell as an incantation in Latin.

The traditional D&D response to aggression: a fireball in the face. (Baldur’s Gate II: enhanced edition, 2013)

This isn’t typical. The original Baldur’s Gate (1998) and the Neverwinter nights series (2002, 2006) have spoken incantations, but with nonsense syllables, not in Latin. A re-release, Baldur’s Gate: enhanced edition (2012), switches to the Latin incantations of the 2000 sequel. Most other games from that period — Planescape: Torment (1998), Ruins of Myth Drannor (2001), Temple of elemental evil (2003), and more — have no spoken incantations at all. Neither does a new game shoehorned into the older series, Baldur’s Gate III (in early access since 2020).

Fans of the classic games have made attempts to decipher the incantations. Unfortunately a shaky knowledge of Latin, on the part of both the voice actors and the fans, has led to uneven results. This is just a note to give some annotations.

Note. Attempts at transcription/translation: ‘BelgarathMTH’, 2013; ‘Magpie Randoms’, 2017.

First, some links with the actual sounds:

  • Latin incantations in Baldur’s Gate II, Icewind Dale, and Icewind Dale II: YouTube
  • Nonsense syllables in Baldur’s Gate (1998 version): YouTube
  • Nonsense syllables in Neverwinter nights: YouTube
  • Nonsense syllables in Neverwinter nights 2: YouTube

The Latin is inconsistent and a bit sloppy. Different voice actors pronounce words differently. Scio ‘I know’ comes out as either see-oh or ski-oh. Alia ‘other things’ gets stressed randomly on the first syllable (correct) or the second syllable (wrong). Some syllables are muffled by reverb or other sound effects. In one incantation the first word is omitted in some sound files, included in others. In another incantation, the first word isn’t a real Latin word.

A nasty trap for some Frost Giants: a wizard in the process of casting Death Fog, a spell belonging to both the Alteration and Evocation schools. (Icewind Dale: enhanced edition, 2014)

The D&D rule system divides spells into eight ‘schools’ of magic. In the video games, each school has its own incantation, with different voice actors depending on whether the spellcaster is male or female, wizard or cleric.

The incantations can be accessed directly and exported from the game with a tool called Near Infinity. The relevant files are in the game’s WAV directory: CHA-FM01.WAV to CHA_FM08.WAV and CHA_FP01.WAV to CHA_FP08.WAV (FM = ‘female mage’, FP = ‘female priest’); and CHA_MM01.WAV to CHA_MM08.WAV and CHA_MP01.WAV to CHA_MP08.WAV (‘male mage’, ‘male priest’). There are also some additional files with voice only, without magical wooshing noises.

Each incantation consists of three words. They’re not sentences. They have no syntax, and they aren’t meaningful as sentences. It’s better to see them as combinations of ‘power words’, to use D&D terminology: three independent magical utterances. The incantations are as follows:

Spell school Incantation Translation
Abjuration Manus; potentis; paro Hand(s); of someone powerful; I prepare
Conjuration / Summoning [Facio;] voco; ferre I do; I call; to carry
Enchantment / Charm Cupio; virtus; licet I want; merit; it is permitted
Divination Scio; didici; peto I know; I have learnt; I seek
Illusion Veritas; credo; oculus Truth; I believe; eye
Invocation / Evocation Incertus; pulcher; imperium A doubtful man; a beautiful man; authority
Necromancy Vita; mortis; careo Life; of death; I lack
Alteration Praesi; alia; fero [Nonsense word]; other things; I carry

Some of these look like they may be mistranslations of English words. Imperium, in the Evocation incantation, is probably intended to mean ‘power’, but really it means ‘authority, position of command’. A better word to suggest magical energy would have been potestas or vis. And the moral connotations of virtus seem a poor fit for Enchantment spells: perhaps it was intended as a translation of ‘glamour’ or something of the kind

The only ones that could possibly be construed as whole sentences are the incantations for Abjuration spells — something like ‘I prepare powerful hands’ (taking manus and potentis as accusative plural, rather than genitive singular as in the table above) — and for Necromancy — ‘I lack a life of death’ (taking vita as ablative). These don’t make much sense and they’re not good matches for their spell schools. It really is best to interpret all of them as a sequence of three separate ‘power words’.

Summoning a water elemental to fight the BBEG: a Conjuration/Summoning spell. (Icewind Dale: enhanced edition, 2014)

Here are some additional notes on each incantation.

Abjuration. See above on interpreting this as a complete sentence. The sense ‘I prepare powerful hands’ is a bad fit for Abjuration spells, which are about dispelling magical effects or protecting against hostile spells. The fact that the syntax actually works is presumably just a coincidence. In any case, this sense would require both manus and potentis to have a long vowel in their final syllables; and that isn’t how they’re pronounced in the games.

Conjuration/Summoning. The first word facio is partially or completely omitted in mage sound files. In Baldur’s Gate II it is completely omitted by female mages, and partially omitted by male mages; in Icewind Dale it is completely omitted by all mages. Clerics pronounce the word in full in both games. The second word is pronounced by female spellcasters as vaco, which would mean ‘I am empty’ and is certainly not the intended meaning. Ferre ‘to carry’ is the infinitival form of fero ‘I carry’, which also appears in the Alteration incantation.

Enchantment/Charm. The intended meaning of virtus isn’t obvious: neither ‘moral merit’ nor ‘proficiency’ (the usual meanings) are a good fit for Enchantment spells; see above. Licet receives a variety of pronunciations: sometimes with a hard c, sometimes a soft c; the i is sometimes as in fight, sometimes as in fit. Both pronunciations of c, /k/ and /tʃ/, are legitimate for different varieties of Latin. The same is true of didici, in the Divination incantation below.

Divination. Scio is variously pronounced see-oh or ski-oh. Didici ‘I have learnt’ is unique among these incantations as the only past tense verb form. On the pronunciation of the c, see on ‘Enchantment/Charm’ above.

Illusion. The third word can easily be misheard as oculos (accusative plural form) in some sound files.

Invocation/Evocation. In authentic Latin the first two words are gendered adjectives, which imply a male entity (‘a doubtful man, a beautiful man’. This is probably not intended. It is easy to mishear the last syllable of the noun imperium, making it a verb, imperio ‘I command’; but the final -um is clear enough. See above on the inappropriate sense of imperium.

Necromancy. Mortis (genitive singular ‘of death’) is pronounced mortes in some cleric incantations (‘deaths’, plural). Careo may be a mistranscription: it may have been intended as a non-Latin variation on caries ‘rot, decay’ or cariosus ‘rotten’, which would be a much better fit for Necromancy spells. Then again, some voice files clearly pronounce the word with a stressed e. Stressed e must be careo ‘I lack’; unstressed i suggests a non-Latin variation on caries ‘decay’.

Alteration. In all versions the first word is praesi, which isn’t a real Latin word. It was presumably intended as a form of praeses ‘protector, guard’, which is what some Alteration spells do (Stoneskin in Baldur’s Gate; Wraithform; etc.). Alia ‘other things’, representing spell effects like shape-changing, is sometimes mispronounced with the second syllable stressed; the correct stress is on the first syllable (AH-lee-ah).

Wednesday 20 July 2022

Jesus’ empty tomb and The matron of Ephesus

The story of the ‘empty tomb’ in the New Testament gospels uses many motifs that also appear in a Roman morality fable, first attested just a few decades earlier: The matron of Ephesus. The Matron story is best known to modern readers from Petronius’ bawdy novel the Satyrica.

If the parallels are legit, does that mean the one story is derived from the other? Modern observers often find that tempting. I’m sure you’ve met people who love explaining that bits of Genesis are based on Gilgamesh or the Enuma elish.

But ‘A is based on B’ explanations are usually jumping to conclusions. It’s almost always more robust to assume that both stories are working with a shared pool of story conventions and motifs.

Possibly an illustration of The matron of Ephesus, ca. 54–68 CE. Detail from Dandré Bardon, Costume des anciens peuples (Paris, 1772), cahier 8 plate 12, claiming to reproduce a bas-relief found in the ruins of Nero’s palace; the bas-relief is now lost. Note the bowl of hot food, aligning the bas-relief with Petronius’ version of the story, and against Phaedrus. Bardon’s retelling of the Matron story is at 71–72 (probably under the influence of La Fontaine’s poetic version). See further Hansen 2002: 274 with 279 n. 18.

I’m not the first to notice the parallels in the Matron story, but it was only in the late 2010s that anyone started taking notice of them. I’ve found only one competent scholarly discussion, by Prof. Robyn Walsh. Walsh’s position is that Petronius and the gospels are in dialogue with one another: 1st–2nd century CE Christians understood stories in the gospels, including the ‘empty tomb’ episode, through the lens of topical storytelling conventions — conventions like the ones in novels of the time, such as Petronius, and like the ones in the Matron fable.

I prefer a middle ground. ‘A is based on B’ is reckless. But it’s too tentative to limit the influence of novelistic and fabulistic motifs to the gospels’ readers. (Not that Walsh does this, exactly, but it’s best to be explicit.) We can be confident in saying that these motifs are part of the storyteller’s toolkit. The Christian story isn’t derived from Petronius. But it is a cousin.

Note. Walsh 2020: 363–367 = 2021: 146–149. For an extensive bibliography on the gospels as texts firmly welded to the Greco-Roman literary tradition, see Walsh 2021: 134–135 n. 1. On the use of Hellenic-Roman story-telling conventions specifically in the ‘empty tomb’ narrative, see Miller 2010; Cook 2018: 598–601.

The matron of Ephesus

Here’s the oldest surviving version of the story, from Phaedrus.

A woman lost her beloved husband of many years
and laid his body in the ground.
Nothing could tear her away,
and in his tomb she filled her days with weeping.
She became famous as the very model of a chaste maiden.
      Meanwhile, men who had pillaged Jupiter’s temple
were crucified, to pay the penalty to the divinity.
To stop anyone removing their remains,
soldiers were stationed as guards for the bodies,
next to the tomb where the woman was shut in.
      One day, one of the guards was thirsty and
asked (the widow’s) slave for water at midnight.
As it happened, the slave was helping her mistress
prepare for bed. She was sitting by lamplight;
she had kept her vigil late.
The door was open a crack. The soldier looked in
and saw a woman both sad and of great beauty.
His heart was fired with lust
and slowly the shameless man’s desire flamed up.
His inventive shrewdness found a thousand reasons
to see her again and again.
Ensnared by the daily habit,
she gradually became more receptive to her visitor.
Soon an even closer bond subdued her heart.
      While the guard passed his nights in love,
a body went missing from one of the crosses.
Distraught, the soldier explained this to the woman.
That holy woman said, ‘There’s nothing to fear.’
She gave her husband’s body to be tied to the cross,
so the soldier wouldn’t be punished for his lapse.
      In this way wickedness besieges a place of praise.

We have three Roman or Roman-derived variants of the Matron story. Phaedrus, above, is the oldest, in iambic verse: Fabulae Aesopiae, ‘Appendix Perottina’ 13 (ca. 10s–30s CE). A much longer and more famous variant appears in Petronius’ bawdy novel the Satyrica, 111–112 (probably ca. 60 CE, but there are scholars who want to re-date it to the 2nd century). And a much later and shorter prose variant appears in one of the mediaeval fable collections that go under the name ‘Romulus’, the Romulus ordinarius 3.9 (ii.208 ed. Hervieux).

The story almost certainly originates in the fabulist tradition. Petronius’ version is traditionally linked to Aristeides’ Milesiaka, rather than the Aesopic tradition. But looking at the full set of variants makes it clear that, even if it is not strictly Aesopic, it is at least para-Aesopic. If you want you can see it as ‘Milesian’ and Aesopic at once, of course. The matron’s closing words in Petronius, ‘I would rather make use of a dead man than kill a living one’, probably echo a fable with a concluding moral. The moral in ps.-Romulus plays on a similar theme of life and death: ‘The dead have something to grieve, and the living something to fear’ (reading habent).

Petronius and ps.-Romulus have several differences from Phaedrus, mostly minor. Petronius makes the tomb an underground vault, supposedly Greek-style (in hypogaeo Graeco more); and the matron’s lamp is, uh, lampshaded into the story early on. Petronius and ps.-Romulus specify that the guard is placed because the crucified man has relatives or friends who want to remove the body; and the soldier’s repeated visits are initially to console the matron, not to seduce her. In ps.-Romulus the love affair is only hinted at, and the matron has no slave. In Petronius, the soldier offers food to the grieving matron, and the matron’s slave encourages her first to accept the food, and later, to succumb to sexual desire — quoting from Vergil’s Aeneid both times (reenacting the role of Anna to Dido).

Note. For the text of Phaedrus, we may now refer to the edition of Zago 2020. The Appendix Perottina is a compilation made by the 15th century humanist Niccolò Perotti; several of Phaedrus’ fables survive in no other form. Schmeling (2001: 427) draws attention to an argument that the Petronius variant is based on Phaedrus, rather than both being based on a lost fabulist tradition. That is certainly wrong. First, as I point out in my introduction above, ‘A is based on B’ is usually a reckless assumption. In addition Petronius includes several motific elements that are absent in Phaedrus: the quasi-moral at the end; the placement of the tomb underground; the more active roles played by the widow and her slave; the explicit involvement of the crucified man’s friends or family.

The fable is misogynistic through and through. Its central twist has the matron vacillating from extreme loyalty to her dead husband to extreme irreverence for his body. Its message is that even the most chaste woman will become promiscuous when put under any pressure. And it’s based on the sexist premise that a woman’s merit is decided by her sexual availability to men.

I mention the story’s misogyny because I want to emphasise, up front, that I don’t mean these underlying messages necessarily apply to any story that contains similar motifs. Below, we’re just talking about the motifs.

As well as the three Aesopic variants, we have many other more divergent variants from a number of cultures, outlined in detail by William Hansen in his book on folktale traditions in the Greco-Roman world (Hansen 2002: 266–279; folktale type Aarne-Thompson-Uther 1510). There’s a comical variant in an ancient Life of Aesop, where a ploughman stands in for the soldier and is the butt of the joke. In the Middle English Seven sages of Rome, both the dead husband and the soldier are knights, and the living knight rejects the woman after realising her inconstancy. In a 19th century Tunisian variant, the widow has vowed never to re-marry, but when she encounters a vizier who is distraught because he cannot capture a thief, she immediately offers him her husband’s corpse and her hand in marriage. Hansen reports several more.

It’s the three Aesopic variants that share the most motifs with the Christian ‘empty tomb’ story. Unlike most of Hansen’s variants, the Aesopic variants are consistent in giving the guard’s role to a soldier; and they lack the motif of mutilating the dead husband’s body to disguise his identity.

The Matron on the silver screen. Left to right: the dead husband, the matron, and the soldier (Fellini Satyricon, 1968)

Jesus’ empty tomb

Five early Christian texts include an episode set following the crucifixion of Jesus: he is buried in an underground tomb, and two days later Mary Magdalene visits the tomb and finds that the body has disappeared. The story appears in all four New Testament gospels, and in the fragmentary gospel of Peter.

Note. Mark 15.42–16.8; Matthew 27.55–28.15; Luke 23.49–24.12; John 19.38–20.18; Peter 23–57 ed. Mara. For convenience I have drawn up elsewhere a tabulation of the full text of the variants in translation.

Here’s a summary of their shared elements:

  • Jesus dies by crucifixion.
  • On Preparation Day Joseph of Arimathea, a follower of Jesus, gets permission from the Roman governor Pilate to bury Jesus’ body. (In Peter, Joseph gets the body from ‘the Jews’.)
  • Joseph wraps the body in linen and places it in an unused tomb which is carved into rock, and with a large stone serving as its door. In John and Peter the tomb is located in a garden.
  • In Matthew and Peter, soldiers are posted as guards to prevent Jesus’ disciples from removing the body.
  • The morning after the Sabbath Mary Magdalene, and another woman or women, visits the tomb. In John she goes alone.
  • They see that the tomb is open and Jesus’ body is missing. In Matthew and Peter the miraculous opening of the tomb is witnessed, by different groups of people.
  • They see an apparition of one, or more often two, men in white clothes. (In John, Mary also encounters Jesus himself.) The apparition tells them that Jesus has risen from the dead, and that they should report this to Jesus’ disciples.
  • In Luke and John, Peter visits the tomb too — along with the ‘beloved disciple’, in John’s case — and finds it empty. Luke and John differ in the details and placement of this episode.
  • In Matthew and Peter, the guards conspire to conceal what they have seen.

The synoptic gospels are normally understood as reworking material in Mark, but all five variants match up fairly well. Matthew and Luke have some distinctive elements not present in Mark, some of them overlapping with John and Peter. The soldiers posted as guards appear only in Matthew and Peter. Matthew and Peter have people witness the tomb being opened by divine intervention, though the witnesses are different groups. Matthew and John have Mary Magdalene encounter Jesus in person, in different locations. Luke and John have Peter visit the tomb too.

Biblical scholars have differing views on the age of the ‘empty tomb’ story. The earliest mention of Jesus’ burial, in Paul, 1 Corinthians 15.3–8, has little overlap with the ‘empty tomb’ — no Mary Magdalene, no cave with a stone for a door, no soldiers, no linen, no apparitions in white clothes. The overlaps are that he refers to burial — though burial doesn’t imply ‘in an underground chamber’ any more than it does today — and to returning to life on the third day.

Note. For a comprehensive overview see Cook 2018 (passim); but be aware that Cook thinks 1 Corinthians 15.4 ἐγήγερται necessarily implies a mausoleum (see also Cook 2017).

Parallel motifs

Here’s a summary of the motifs shared between the Matron story and the ‘empty tomb’ story.

  • The tomb is an underground chamber
  • A woman grieving for a dead man, with one or more women as company
  • Mourning lasts multiple nights
  • A soldier or soldiers appointed to guard the body of a crucified man
  • Danger that the friends/family of the crucified man will remove his body
  • Importance of the tomb’s door as a barrier
  • Body goes missing — empty tomb and/or empty cross
  • Discovery in the morning
  • Guard(s) conspire to conceal the truth

We mustn’t overstate the similarities. Even where they’re similar, there are important differences. The Matron story has:

  • Two dead bodies; the theme of mistaken identity
  • The matron grieving inside the tomb
  • Love affair with the soldier; themes of consolation, infidelity, misogyny
  • Pithy moral at the end

And the ‘empty tomb’ story has elements that are absent in the Matron:

  • An extra character (Joseph of Arimathea) who performs the burial
  • A large stone for the tomb’s door
  • Miraculous story (resurrection)
  • Supernatural figures appear and explain what has happened
  • Claim to truth — the emphasis on eyewitnesses to the linen cloths left behind
‘La matrone d’Éphèse’: etching by Jean-Baptiste Oudry, from a drawing by Pierre-Alexandre, for an edition of La Fontaine’s Fables (1755–1759)

What exactly do the parallels imply?

Just to repeat, this is not a case of literary plagiarism. There is no reason to think the gospel writers are ripping off Petronius or Phaedrus.

Note. And Petronius isn’t ripping off the gospels either. Rather adorably, two very inexpert discussions assume that’s the only imaginable explanation for any parallels: Blocker 2016; Godfrey 2019. Blocker even thinks that the similarities require dating the Satyrica later than the Neronian Petronius (but Phaedrus, apparently, doesn’t need to be re-dated). If you do choose to waste time reading either of these, take care not to be confused by Blocker’s misspelled names, or by Godfrey’s bizarre mulishness.

The central plot points of the fable, the love affair and the concealed identity of the husband’s body, are both missing in the gospels. The misogynistic themes of the fable are also not present (or at least not in the same way). So we’re not talking about the ‘empty tomb’ story being a close adaptation of an oral fable, either.

Still, some of the common motifs are particularly striking. Especially important is the soldier(s), who are posted to prevent a crucified man’s body being removed by his friends, and who conceal the truth of its disappearance; and secondly, the tomb itself. In the gospels the tomb is hewn into rock, and in Petronius, it is ‘underground in Greek fashion’, using a Greek word for ‘underground’ (in hypogaeo Graeco more). And both stories have a strong focus on the tomb’s door.

Just to be clear, tombs in caves are not a standard thing. [NOTE, added later: see Addendum at bottom. I was too dismissive of cave-tombs here, but some additional complications arise when you start looking at real cave-tombs.] The ‘underground’ or ‘rock-hewn’ tomb is a totally artificial motif. It may perhaps be loosely inspired by things like the underground oracular shrine of Trophonios, or the story of Pythagoras revealing his teachings in underground chambers: these are both stories that enjoyed a moderately popular appeal in the early principate.

The only competent scholarly discussion I’ve found of parallels between the Matron and the ‘empty tomb’ is the one by Walsh, as I mentioned at the start. Walsh draws on a long-standing scholarly position that ancient readers understood the gospels in terms of Greek storytelling conventions. She makes good use of another article by Richard C. Miller, which focuses specifically on Mark’s ‘mimetic use of the Greek classical canon’ in the ‘empty tomb’ episode (Miller 2010). The episode carries special weight in Mark because the end of the episode is also the end of the gospel.

Note. Mark 16.8 is the last authentic verse in Mark, and this isn’t remotely controversial. The alternate endings typically printed in modern Bible translations — the ‘longer ending’ (Mark 16.9–20) and the ‘shorter ending’ (usually confined to a footnote) — don’t exist in the oldest copies or in the earliest witnesses to the text; they were added later.

Miller’s precedents for the ‘empty tomb’ involve only one motif, namely the missing body itself. The parallels he cites are the disappearances of the bodies of Romulus, Aristeas, Heracles, Amphiaraus, and others. Even that has been enough for some biblical scholars to conclude that the ‘empty tomb’ story is artificial.

The parallels in the Matron story are much more extensive. But it’s important to remember that none of this involves copying, plagiarism, or parody. No one thought, ‘Hm, how shall I tell the story of Jesus’ burial? I know, I’ll recycle The matron of Ephesus.’ What we’ve got is multiple stories, in different contexts, with some similar motifs.

And similarity breeds similarity. The more motifs stories have in common, the more prone they are to cross-contamination. The soldiers guarding the tomb, who conceal the truth of what happened, are a clearcut case of contamination: they appear only in Matthew and Peter, and their role — failing to prevent the body’s disappearance, and concealing the truth — is exactly analogous to the Matron fable. Most probably, older variants of the ‘empty tomb’ story — Mark, and oral stories — omitted the soldiers, but were still similar enough to the Matron story to attract the soldiers into subsequent retellings.

The tomb itself is an artificiality too. Jesus’ stone-hewn tomb has close analogues in John 11.38–44 and in Petronius: it’s a motific element, and motific elements tend to imply artificiality. We can’t trace a line of descent: the fact there are two rock-hewn tombs with a large stone for a door, just within John, means that we can’t say that one was based on the other.

And that prompts me to finish by emphasising that it’s important to avoid the ‘A is based on B’ mindset. That is the mindset of survivorship bias. Our evidence — the corpus of surviving stories — is skewed by the fact that only some stories survive. We don’t get to pretend that whatever does survive is necessarily based on other things that survive. There’s no unbroken chain reaching back from Luke to Phaedrus.

But they are related. And yes, it is fair to say that replicated motific elements are a sign of artificiality. When biblical scholars do defend the historicity of the ‘empty tomb’ story, they sometimes highlight the fact that in Matthew, the authorities try to spread disinformation with a fake story of Jesus’ disciples stealing the body, ‘and this story is still told among the Jews to this day’. But that’s only in Matthew; it’s based on the testimony of the guards; and it’s precisely the premise of the Matron fable. Out of all elements of the ‘empty tomb’ story, that one is the most artificial, by far.


  • Blocker, D. 2016. ‘The relationship between the Satyricon’s “Tale of the Ephesian widow” and texts associated with early Christianity.’ Jesus granskad (Roger Viklund), 24 Apr. 2016. [Internet Archive]
  • Cook, J. G. 2017. ‘Resurrection in paganism and the question of an empty tomb in 1 Corinthians 15.’ New Testament studies 63: 56–75. [DOI]
  • —— 2018. Empty tomb, resurrection, apotheosis. Tübingen.
  • Godfrey, N. 2019. ‘Is the satirical Widow of Ephesus story an attack on Christianity?’ Vridar, 4 June 2019. [Internet Archive]
  • Hansen, W. 2002. Ariadne’s thread. A guide to international tales found in classical literature. Ithaca, NY.
  • Mara, M. G. 1973. Évangile de Pierre. Introduction, texte critique, traduction, commentaire et index. Paris. [Internet Archive (borrowable)]
  • Miller, R. C. 2010. ‘Mark’s empty tomb and other translation fables in classical antiquity.’ Journal of biblical literature 129: 759–776. [JSTOR]
  • Schmeling, G. 2011. A commentary on the Satyrica of Petronius. Oxford.
  • Walsh, R. F. 2020. ‘The Satyrica and the gospels in the second century.’ Classical quarterly 70: 356–367. [DOI]
  • —— 2021. The origins of early Christian literature. Contextualizing the New Testament within Greco-Roman literary culture. Cambridge.
  • Zago, G. 2020. Phaedrus. Fabulae Aesopiae. Berlin/Boston.

See also: a tabulation I have drawn up of the story of Jesus’ burial as related in the gospels, with the text of all five in parallel translations.

ADDENDUM, 28 July 2022

In light of some counterpoints raised elsewhere, I need to add some qualifications.

In the first place, ‘empty tomb’ is a misnomer, as pointed out in a recent article by Mark Goodacre. The earliest gospel, Mark, makes no suggestion that the tomb is unused. That’s something that only appears in the later gospels, as a secondary development. That isn’t a case of contamination from the Matron story: in fact it has no bearing on the Matron story, since the Matron story doesn't have an empty tomb either. There, as in Mark, a body goes missing from the tomb, but there’s no suggestion that only one person has ever been deposited there.

Second: above, I was too dismissive of cave-tombs. It was a normal practice in ancient Judaea for bodies to be deposited in a cave-tomb — but specifically an ancestral cave-tomb, with a model in the Hebrew Bible, Abraham’s family tomb: this is the cave that Abraham purchases in Genesis 23 (‘in a field’, in the Septuagint version, conspicuously similar to the garden in John and Peter), and which appears again as the place where Jacob’s body is deposited in Genesis 49.29–33.

Ancestral cave-tombs continued to be used in Talmudic times, alongside grave burial. References from elsewhere in the New Testament — particularly 1 Corinthians 15.3–8, which I mentioned above, but also Acts 13.29 — don’t mention a cave-tomb, and I’m not persuaded by Cook’s argument that 1 Corinthians implies a cave-tomb. Acts, moreover, isn’t independent of Luke, so even if it did imply a cave-tomb, that would have to be taken in conjunction with the ‘empty tomb’ narrative in Luke 23–24 anyway.

The fact that Matthew, Luke, and John emphasise that the tomb is new and has never been used, has to be read against the normal paradigm of ancestral cave-tombs. That is, what they are really emphasising is that Jesus’ tomb isn’t an ancestral tomb. And, given that Matthew and Luke also emphasise Jesus’ divine ancestry — by adding nativity narratives, for example — that looks likely to be the most salient context for their emphasis on Jesus’ tomb not being ancestral. That is, Jesus’ cave-tomb can’t have had mortal bodies deposited in it, because Jesus’ ancestry has something non-mortal.

This is an alteration from the story as told by Mark, but again, not a case of contamination from the Matron story. A more pared-down version of the parallels, limited to the oldest version of Jesus’ cave-tomb burial — that is, the parallels between the Matron story and Mark alone — would produce a shorter list of parallels. Exactly how much shorter is something that needs further work. For the record, I don’t think I’m likely to be the one to do that further work.

Friday 1 July 2022

Was Sirius once red?

Back in 2020, when I wrote a piece about ancient Greek colour terms, two comments raised an interesting side-question: what colour was the star Sirius in antiquity?

Did Sirius look red to ancient astronomers? First, be warned that the colours above aren’t real. Colours in space photos are always heavily enhanced — or, as in this picture, totally fake.

Sirius is the brightest star in our night sky. In antiquity it was also known as the Dog star; it still is today. It looks white. Or maybe blue-white: its spectrum has its peak at blue frequencies, and photographs tend to create blue halos. But really Sirius floods the entire range of visible light.

Over the last 250 years a question has been simmering, because — supposedly — some ancient writers describe it as ‘red’. Astronomers want to know: is it even possible that Sirius changed colour between the 100s and the 1700s CE?

The modern debate

But, to this rule there seems to be one exception, and that in a remarkable star: for old authors mention the Dog star, which is now white, and not at all inclined to redness, as being then very much so; as in the following places: ...
Barker 1759: 499

The colour change theory was first proposed in 1759 by Thomas Barker, an English country squire, meteorologist, and astronomer. He cited several pieces of ancient testimony, looking directly at the Greek and Latin terminology in each case:

  • Aratos, Phainomena 326–328 (3rd cent. BCE): the Dog star is ποικίλος (poikílos).
  • Cicero, Aratea fr. 34.107–108 ed. Pellacani (1st cent. BCE): the Dog star ‘shines with rutilus light’ (rutilo cum lumine claret).
  • Horace, Satires 2.5.39-40 (1st cent. BCE): refers to rubra Canicula.
  • Seneca, Quaestiones naturales 1.1.7 (1st cent. CE): refers to the rubor of Canicula.
  • Ptolemy, Almagest 7.5, 142 ed. Heiberg (2nd cent. CE): ‘the star in the mouth [of the Dog constellation, i.e. Canis Major] is the brightest, called the Dog; it is ὑπόκιρρος (hypókirrhos)’.
  • Hyginus, Astronomica 2.35 (1st cent. BCE/CE): refers to the ‘brightness’ or ‘whiteness of its flame’ (flammae candorem); Barker thinks (1) candor means strictly ‘whiteness’; so (2) this reference ‘expressly contradicts’ the others; so (3) Hyginus should be disregarded.

You’ll notice I’m leaving some Greek and Latin terms untranslated. That’s because their meaning is precisely what we’ll explore below.

I won’t cover the full history of the debate, which was argued by astronomers such as Jérôme de Lalande, Theodor von Schubert, and John Herschel. For the full story, see Ceragioli (1995).

One key development was in 1850, when Alexander von Humboldt published volume 3 of his book series Kosmos. Humboldt delved into the philology of the terms used by the ancient writers, with special attention to Ptolemy’s colour term, hypókirrhos.

The expression ὑπόκιῤῥος, which Ptolemy employs indiscriminately to designate the six stars named in his catalogue [i.e. the six stars named as hypókirrhos], implies a slightly marked transition from fiery-yellow [feuergelb] to fiery-red [feuerroth]; it therefore refers, strictly speaking, to a fiery-reddish colour. ... Κιῤῥός is, according to Galen, (Meth. med. 12,) a pale fiery-red inclining to yellow. ... Sirius is said by Seneca (Nat. Quæst., i. 1) to be redder than Mars ...
Humboldt 1851: 176 n. 46 (tr. Otté) = 1850: 204 n. 46

This report of the terminology is still widely treated as authoritative, because astronomers don’t typically learn Latin and ancient Greek these days. But as we’ll see, Humboldt is entirely wrong.

Roger C. Ceragioli, who trained as a classicist, wrote a 1995 article covering the modern debate and exposing the deluge of misinformation. Ceragioli handily disposes of most of the supposed evidence for ‘red Sirius’, but he thinks Barker’s references to Ptolemy, Horace, and Seneca stand up.

There is no doubt about what Ptolemy means. He bluntly says ‘reddish’ and other ancient sources corroborate him ...
Ceragioli 1995: 187

As a result, they still need explanation: either it’s because Sirius scintillates in various colour-flashes when close to the horizon (Ceragioli 1995), or because hypókirrhos in Ptolemy is an interpolation (Ceragioli 1996).

Note. The only other substantive discussion of the matter by a modern classicist that I’ve found is by Pellacani (2015: 156–157), who is much more willing to indulge the ‘red Sirius’ theory, is unaware of Ceragioli’s articles, and is selective in his reading of another key source, Avienius. We’ll return to Avienius below.

In fact none of the supposed evidence for ‘red’ Sirius stands up to scrutiny:

  • Hypókirrhos doesn’t mean ‘fiery red’ or ‘reddish’: it means ‘pale yellow’.
  • Horace’s and Seneca’s rubor primarily denotes ‘heat’. Its meaning as a colour term, ‘red’, is uncommon.
  • Cicero’s word rutilus primarily means ‘bright, shiny’, especially when used of celestial bodies. And anyway, when he uses the word he’s not talking about Sirius.
In fiction, Sirius is usually shown as noticeably blue to the naked eye. (Elite Dangerous, 2014)

Ptolemy: hypókirrhos

For modern advocates of ‘red Sirius’, Ptolemy’s testimony carries the most weight.

Κυνὸς ἀστερισμός.
ὁ ἐν τῷ στόματι λαμπρότατος καλούμενος Κύων καὶ ὑπόκιρρος ...
Constellation of Canis (Major).
The (star) in the mouth (of the Great Dog); the brightest; called the Dog; and hypókirrhos...
Ptolemy, Almagest 7.5, 142 ed. Heiberg

Hypókirrhos is ‘somewhat kirrhós’. But neither hypókirrhos nor kirrhós means ‘red’. A more correct translation is ‘cream, light yellow’.

Hypókirrhos and kirrhós are very rare words in most contexts. In extant classical-era texts (5th–4th century BCE) they appear only twice, both in the Hippokratic corpus. The etymology is unknown. Even at the best of times, interpreting Greek colour terms is tricky: here, extreme caution is needed.

The argument for interpreting hypókirrhos as ‘reddish’ rests entirely on these two points:

  1. Ptolemy describes six stars as hypókirrhos: Arcturus, Aldebaran, Pollux, Antares, Betelgeuse, and Sirius (Almagest 7.5: pages 50, 88, 92, 110, 132, and 142 ed. Heiberg). Of these, Aldebaran, Pollux, Antares, and Betelgeuse are noticeably golden or orange-ish to the naked eye.
  2. According to Humboldt, a passage in Galen defines kirrhós as ‘a pale fiery-red inclining to yellow’ (ein blasses Feuerroth, das in Gelb spielt; Galen, De methodo medendi book 12, x.832 ed. Kühn).

Humboldt cherry-picks his Galen, and even the passage he cites has problems. Galen’s exact words are

εἰ δ' ἄλλως ἐθέλεις ὀνομάζειν τὸ κιρρὸν χρῶμα, δύναιο ἂν λέγειν πυρρὸν ἢ ὠχρόν.
If you want to refer to the colour kirrhós differently you could say pyrrhós or ochrós (‘pale’).

In the 19th century, dictionaries treated pyrrhós as referring straightforwardly to pŷr ‘fire’. Humboldt seems to have in mind Pape’s definition (1st edition 1842):

feuerfarben, feuerroth, röthlich, in verschiedenen Abstufungen der Farben, bis zum Blonden hin ...
fire-coloured, fiery red, reddish, in various colour grades ranging to blonde ...

(Cf. Passow: ‘feuerfarb, feuergelb, feuerroth, goldgelb, überh. röthlich’.) In fact the exact linguistic relationship between pyrrhós and pŷr is unclear (Beekes 2010: 1264). Even to the extent that they’re connected, fire ... um, isn’t red.

Evidence from ancient usage of pyrrhós is a mixed bag. A pseudo-Aristotelian work implies a link to the colour of foxes; one late lexicographical source treats pyrrhós as a synonym for erythrós. Then again, another late lexicographical source defines erythrós as mélas, and pyrrhós as xanthós, keeping them as two separate colour categories. And Plato says pyrrhós is what you get if you mix xanthós (‘golden’) and phaiós (‘grey’) pigments.

The point is, there’s nothing simple about this.

Note. Ps-Aristotle, Physiognomonica 812a (xanthós-haired people have a bold spirit like lions, pyrrhós -haired people are villainous like foxes); Hesychios 6084, ii.201 Latte (ἐρυθραίνετο· ἐρυθρά ἐγένετο, πυρρά); Souda ε.3101 (ἐρυθρόν· τὸ μέλαν; similarly ε.3092), π.3235 (πυρρός· ὁ ξανθός); Plato, Timaios 68c. I am grateful to @SartrixMartiana and @oliveratlantis for pointing out the ps-Aristotle, Hesychios, and Souda examples to me. The new Cambridge Greek lexicon (2021) avoids ‘red’ in its definition of pyrrhós, except — paradoxically — when used of gold. The same lexicon omits kirrhós altogether.

Back to kirrhós. Setting aside Humboldt’s Galen passage, other contexts don’t support ‘red’. The most common use of kirrhós, in fact, is for a colour of wine — and it is not red wine.

Kirrhós wine is a white wine with relatively heavy colour: think chardonnay rather than pinot gris. It is invariably opposed to red wine, which is called mélas (‘dark red, deep brown, black’; mélas is also regularly used of blood).

Bear in mind that wine wasn’t backlit in antiquity. Ancient wine-drinkers saw their wine in the bottom of a ceramic kylix, not through the side of a transparent glass. This may help explain why they called ‘red’ wine mélas. A wine would have to be very pale to look erythrós (‘red’) in a kylix — more like a rosé.

Instead of kirrhós wines, use red ones [mélasin] ...
ps-Hippokrates, De mulierum affectibus 115
Wines are white [leukós], kirrhós, or red [mélas]. ... And Mnesitheos of Athens says: ‘Red wine is very good for growth. White wine is a very good diuretic, and is the thinnest. Kirrhós wine promotes dryness, and is good for digesting food.’
Athenaios 1.32c–d (= Mnesitheos fr. 46 ed. Bertier)
... keep watch on the testing of the wines, and always choose the one lightest in substance — the colour that Hippokrates usually calls kirrhós. You could also call it golden [xanthós]. Pale [ochrós] is also good, in between golden and white. In fact if you want to mix golden wine with white, you will produce a combination of both that is pale ...
Galen, De sanitate tuenda vi.335 ed. Kühn
Note. In Mnesitheos/Athenaios, ‘dryness’ (ξηρός) has nothing to do with the modern sense of a low-sugar wine: the word for that is αὐστηρός. See Boulay 2015: 277–279 on the vocabulary for the tastes of wine. ‘Dry’ here means foods that supposedly reduce phlegm, in the ancient medical theory of the four humours.

Here Galen sorts ‘white’ wines by the intensity of their colours, ranging from leukós, to ochrós, to kirrhós or xanthós. That is, he groups kirrhós firmly among white wines. He expands the range to red wines elsewhere:

You won’t find any wine that is thick and sweet that isn’t red [mélas]. ... There is no sweet wine that is white [leukós], but some are dry [austerós] and thick, some watery and light. Golden [xanthós] and kirrhós wines are moderately sweet, like Hippodamanteian wine, and the Faustian Falernian. Some are not sweet at all. Light red [erythrós] wines are thicker than these, just as others are (thicker) than them, as they approach the reds [mélas] in colour.
Galen, De bonis malisque sucis, vi.800–801 ed. Kühn

We can’t review kirrhós exhaustively. Galen uses the word over 100 times, far more than any non-medical writer. Almost every time it refers to wine. Let’s just highlight a few more sources that use kirrhós for things other than wine.

  • A Hippokratic text (5th cent. BCE) describes the symptoms of a sick woman, and says her faeces were hypókirrhos.
  • A fragment of Aristotle (4th cent. BCE) describes a fish as kirrhós, and the fragment is repeated elsewhere. But unhelpfully, the fish is of unknown species. (Liddell & Scott suggest it is a Labrus or wrasse, without basis.)
  • An excerpt of a lost work by Aristophanes of Byzantion (3rd cent. BCE) states that the best milk is the thickest and ‘most kirrhós’; white [leukós] milk isn’t as nourishing.
  • Dioskourides (1st cent. CE) says the best beeswax is hypókirrhos. Elsewhere he refers to three varieties of frankincense, one white [leukós], one hypókirros, and one ‘more kirrhós’; Frankincense is typically pale yellow.
  • Galen (2nd cent. CE) refers to bitter vetch, acacia, and mastic as kirrhós. Acacia is yellow; mastic is pale yellow. Bitter vetch flowers are lilac, but since he’s talking about consuming it roasted, he must be talking about the seeds, which are yellow. (Another less helpful passage says that kirrhós is synonymous with two dialectal words ... both of unknown meaning.)
  • Aetios (6th cent. CE) uses kirrhós as a synonym for hypóxanthos, ‘somewhat golden’, and assigns various foods to that colour including wheat and chickpeas — that is, light yellow.
Sources. ps-Hippokrates, Epidemics 7.11; Aristotle fr. 307 Rose = Athenaios 7.281f; Aristophanes, Historia animalium epitome 1.94; Dioskourides 2.83 (beeswax), 1.68 (frankincense); Galen, De remediis parabilibus xiv.366 Kühn (bitter vetch), xiv.533 (mastic, acacia), Glossary xix.129 (‘Zenodotos ... says the people of Sikyon refer to kirrhón as péllon’); Aetius i:proem = viii.1:29,18–21 ed. Oliveri.

In short, Ptolemy does not call Sirius ‘red’. A safer translation would be ‘cream-coloured, pale yellow’. There may still be a question of why he calls it that when it’s clearly white, but at least ‘red’ is off the table. One of Ceragioli’s explanations may still be needed for the ‘pale yellow’.

Photos reportedly taken in 2018 by Amanda Cross, an English amateur astronomer, using an out-of-focus camera to illustrate how a star’s colour varies when close to the horizon. Sirius is white when higher in the sky; the exceptional range of its colours close to the horizon lends support to Ceragioli’s argument. Source: EarthSky, March 2018.

One final point I’d better address is another place where Ptolemy uses hypókirrhos, in the astrological Tetrabiblos. There he claims that solar eclipses effectively put a colour filter over the world; and he links each ‘filter’ to one of the five planets. Hypókirrhos is the colour he associates with Mars (Tetrabiblos 2.90).

But once again, this doesn’t imply ‘red’. If you look at the whole list, it’s obvious he isn’t describing the colours of the planets themselves:

Planet Associated colour in eclipses Approximate translation
Saturn (1) μέλας (mélas) or (2) ὑπόχλωρος (hypóchloros) (1) black/dark red; (2) somewhat chartreuse
Jupiter λευκός (leukós) white
Mars ὑπόκιρρος (hypókirrhos) somewhat kirrhós
Venus ξανθός (xanthós) golden/tawny
Mercury ποικίλος (poikílos) multicoloured/dappled

Jupiter looks white to the naked eye, so that works. But Venus only starts to look a bit yellowish if you have a really good telescope. And it’s hard to see Saturn as ‘black’ or chartreuse, or Mercury as ‘dappled’. The Tetrabiblos passage adds no new information.

The sequel to The hundred and one dalmatians: Sirius, the Dog Star, addresses the dogs of the world from the top of Nelson’s Column and invites them all to come and live in space. Hey, man, it was the 60s. (Dodie Smith, The starlight barking [1967], ch. 10)

Roman sources: rubor and rutilus

With Roman writers, we don’t need to mess about with wine colours. As we saw earlier, Cicero says the Dog star ‘shines with rutilus light’; Horace refers to rubra Canicula ‘the red Dog star’; and Seneca refers to the rubor of Canicula. (Cicero, Aratea fr. 34.107–108 ed. Pellacani; Horace, Satires 2.5.39-40; Seneca, Quaestiones naturales 1.1.7.)

And rutilus, ruber, rubor certainly mean ‘red’ — in terms of their etymology, anyway. They’re cognate with Greek erythrós ‘red’, Sanskrit rudhirá- ‘red, bloody’, and Germanic rot, rood, red. Rubor regularly refers to blushing and inflammations. So at first sight, it does look like there’s a legitimate case for ‘red Sirius’ here.

First we’d better set aside one doubt about the Horace and Seneca: Canicula ‘small dog’ can also refer to Procyon, the most prominent star in Canis Minor, the ‘little dog’. Prokýon is in fact Greek for ‘dog in front’. And a recent scholarly commentary specifically identifies Horace’s Canicula as Procyon, not Sirius (Freudenburg 2021: 210). But I’m going to say that we’d better not assume that’s right. First, Procyon also looks white. Second, Sirius was proverbial in antiquity for scorching heat, rising as it does in the ‘dog days’ of summer, so Procyon should never be the default interpretation. For now at least, let’s assume there is a real question over Sirius.

The more substantive problem is that, though the words are etymologically linked to ‘red’, and though these words can sometimes mean ‘red’, they hardly ever actually refer to a literal hue. To illustrate, here’s the earliest attested appearance of rubor in Latin:

frigit saetas rubore ex oculis fulgens flammeo.
His hair roasts, he shines a flaming rubor from his eyes.
Accius, Meleager fr. 4 ed. Ribbeck

The metaphorical meaning ‘heat’ is much more salient here than a literal red hue.

And this isn’t a one-off. I’ve checked all appearances of rubor before the 1st century CE, and a few references in the 1st century, and only once does rubor refer to a literal hue. Its primary meaning is ‘shame, modesty’. Additional senses include ‘fiery heat’ (Accius, above); in the medical writer Celsus, it can mean either ‘blush’ or ‘inflammation’; once Cicero uses it for a blush cosmetic, evoking ‘shame’ still more metaphorically. And once, and once only, Vergil uses it to refer to a literal colour: the rubor of Tyrian dye.

Note. rubor = ‘shame, modesty’: Plautus, Captivi 962; Publilius Syrus R 8 (= 576 ed. Woelfflin); Catullus 42.16, 65.24; Cicero, De oratore 2.242, Topica 52, De re publica 4.6 fr. 1, Tusculae disputationes 4.(8).19, De natura deorum 1.(27).75, fr. 15.2; ps.-Cicero, Ad Herennium 4.(10).14; Tibullus 2.1.30; Vergil, Aeneid 12.66. rubor = cosmetic, standing for shame: Cicero, Orator (23).79 (‘let all cosmetics of candor and rubor be abolished (from oratory)’). rubor = ‘blush, inflammation’: Celsus, frequently. rubor = red dye: Vergil, Georgics 3.307.

Some of these things are (or can be) literally red, but the real theme connecting the primary meanings — ‘shame, blush, inflammation’ — is heat. Fire, an embarrassed blush, and an inflammation are each hot; Vergil makes it even clearer, with a rubor running through someone’s ‘heated face’ (rubor et calefacta per ora cucurrit).

In short, the most common meaning of rubor is actually ‘heat’. And for Sirius that makes perfect sense. Remember, Sirius was and still is proverbial for the scorching heat (or ‘dog days’) of midsummer.

In Seneca, the context is that he’s paraphrasing a discussion by Aristotle of atmospheric phenomena, explaining that ‘fires’ in the sky, such as comets, are caused by ‘evaporations’ from the earth, and these evaporations manifest as different phenomena depending on their heat. Seneca then gives Sirius, Mars, and Jupiter as celestial parallels to the terrestrial phenomenon. And he straight-up tells us that while Jupiter converts its brightness into ‘pure light’ without heat, in Sirius’ case it becomes rubor.

As for Horace, unfortunately, we have no context. His reference to rubra Canicula is a quotation from another poet, Furius Bibaculus, whose works have been lost.

Note. Seneca’s discussion is paraphrased from Aristotle, Meteorologica 341b; the reference to Sirius is Seneca’s own insertion. On Horace’s quotation of Furius, see Freudenburg 2021: 210–211.

We have one last ‘red Sirius’ passage to deal with. It’s in the fragments of Cicero’s Aratea, a Latin translation of Aratos’ Greek poem the Phainomena. Aratos’ Phainomena is a poem about astronomy, dating to the 3rd century BCE, which spends some time describing the major constellations.

Aratos himself doesn’t call Sirius ‘red’: he refers to the constellation as a whole as poikílos, ‘variegated’, in reference to the uneven brightness of its stars.

And behind [Orion’s] towering back, what a guardian
appears! A Dog walking on both his hind legs,
variegated [poikílos], however, not shining all over ...
... The tip of its
fierce chin is marked with a star which scorches [seiriáei]
more keenly than any other, and people call it
Aratos, Phainomena 326–337

The ‘scorching’ is another reference to the midsummer heat associated with Sirius. Cicero turns this passage into:

For beneath [Orion’s] feet there shines with rutilus light
the fiery Dog, shining and reflecting the light of the stars. ...
But not from all its body does it breathe flame
and make summer fires burst out with its strong breath:
all the blazing shines from its mouth and is cast upon mortals.
Cicero, Aratea fr. 34.107–112 ed. Pellacani

Two points to observe here:

  1. Just like Aratos’ poikílos, Cicero’s ‘rutilus light’ refers to the constellation as a whole. Sirius itself is the star at the Dog’s mouth, the totus ... micans ... ardor in line 112.
  2. While rutilus is linguistically linked to ruber ‘red’, it is routinely used to describe stars. And when describing stars, it doesn’t mean ‘red’: it means ‘bright, shining’.

It’s instructive to compare what Avienius does with the same passage, in another translation of Aratos dating to the 4th century:

[The Dog] burns, studded with rutilantes stars,
but its intensity isn’t the same all over ...
... The heavy blazing flows from its chin,
and burns the aether under a terrible title: Sirius.
When the sun turns its rutilos poles that way,
what suffering threatens people’s bodies, threatens the fields!
Avienius, Phaenomena 730–735

Avienius makes it clearer that rutilans doesn’t refer to Sirius, but to ... well, every other star in the constellation. And in line 734, it’s the sun itself that has a rutilus axis, not Sirius.

Can we hammer the point home even harder? Oh, you betcha. Avienius mentions Sirius again later on:

Meton set the very start of the year at the season
when Phoebus’ rutilus star scorches Cancer,
when the sea carries away Orion’s belt,
when Sirius burns with its blue [caeruleo] star.
Avienius, Phaenomena 1373–1376

Here again it’s the sun that’s rutilus, ‘the star of Phoebus’. And Sirius is — well, what do you know. He calls it blue. How about that.

And that isn’t a one-off either. Another astronomical poem by Manilius (1st century CE) refers to it as ‘blue’.

A great assurance of [the Dog’s] power are its colour, and the twinkling
of fire in its mouth — hardly less than the sun, except that it’s fastened
far off. Cold are the beams it brandishes in its blue [caeruleo] face.
Manilius, Astronomica 1.407–409

The paradox here isn’t that Manilius calls it blue, it’s that he calls Canis Major cold. For every other ancient author, Sirius is the symbol par excellence of midsummer sweltering.

The upshot

To sustain ‘red Sirius’ on Greco-Roman evidence, you’d have to ignore all the metaphorical flavour of rubor, which was really its primary meaning; you’d have to throw away every reference to kirrhós wine, milk, beeswax, chickpeas, acacia, and so on. And you’d have to ignore the fact that when Avienius uses the word rutilus, he doesn’t use it to refer to Sirius, he uses it to contrast with Sirius.

As well as that, you’d have to imagine that Sirius was red when Horace and Seneca were active, in the 1st centuries BCE and CE, but it turned blue by the time Manilius was active, also in the 1st century CE. Then it turned red again for Ptolemy in the 2nd century, then blue again when Avienius came along in the 4th century.

Stuff and nonsense. No, Virginia, ancient Greco-Roman astronomers did not describe Sirius as red.

Pale yellow, maybe. And if you really want to push it, maybe you can make that a problem. But given that that’s based entirely on a single throwaway line in Ptolemy ... I don’t fancy your chances.


  • Barker, T. 1759. ‘Remarks on the mutations of the stars.’ Philosophical transactions 51: 498–504. [JSTOR | Internet Archive]
  • Beekes, R. 2010. Etymological dictionary of Greek. Leiden/Boston.
  • Boulay, T. 2015. ‘Wine appreciation in ancient Greece.’ In: Wilkins, J.; Nadeau, R. (eds.) A companion to food in the ancient world. Oxford. 273–282.
  • Ceragioli, R. C. 1995. ‘The debate concerning “red” Sirius.’ Journal for the history of astronomy 26: 187–226. [Sci-Hub]
  • —— 1996. ‘Solving the puzzle of “red” Sirius.’ Journal for the history of astronomy 27: 93–128. [Sci-Hub]
  • Freudenburg, K. 2021. Horace. Satires book II. Cambridge.
  • Holberg, J. B. 2007. Sirius. Brightest diamond in the night sky. Chichester.
  • Von Humboldt, A. 1850. Kosmos. Entwurf einer physischen Weltbeschreibung, vol. 3. Stuttgart. [Internet Archive]
  • —— 1851. Cosmos. A sketch of a physical description of the universe (trans. E. C. Otté; English translation of the above). [Internet Archive]
  • Jouanna, J. 2012 [1996]. ‘Wine and medicine in ancient Greece.’ In: Greek medicine from Hippocrates to Galen. Selected papers. Tr. N. Allies. Leiden. 173–193. Orig. publ. as ‘Le vin et la médicine dans la Grèce antique.’ Revue des études grecques 109.2: 410–434. [JSTOR]
  • Pellacani, D. 2015. Cicerone. Aratea. Parte I: Proemio e catalogo delle costellazioni. Bologna.