Was there no festival of Sol on 25 December before 324 AD?

Most of us are aware that the 25th December is labelled as the “Natalis [solis] Invicti” in the Chronography of 354; specifically in the 6th part, which contains the so-called “Calendar of Philocalus” (online here), listing the state holidays.  Sol Invictus was introduced into Rome by Aurelian in 274 AD as a state cult, and it seems reasonable to suppose that this state holiday was introduced at the same time.   The Chronography also lists the saints’ days, in another calendar dating from 336 (online here), including Christmas on 25 December.  It is often supposed, therefore, that the date of Christmas was selected precisely to coincide with this solar holiday.  This theory was advanced by H. Usener in his book Das Weihnachstfest (1889, rep. 1911) with a follow-up in his posthumous article on Sol Invictus in 1905.^[1]

However I have lately seen claims that, far from Christmas being located on the date of a pagan holiday, the truth is that Julian the Apostate (or someone) established a solar festival on the pre-existing date of Christmas!  These claims seem to derive from an interesting article by Steven Hijmans, “Usener’s Christmas”.^[2]  Hijman is a revisionist, so it is necessary to be wary, but I thought that it might be useful to review some of the evidence.

In the Chronography of 354, in the “Filocalian calendar”, some holidays – all associated with emperors or gods – are marked by chariot races (circenses missus).  These are also in multiples of 12 races, with one exception.  The sole exception is the entry for 25 December:

 N·INVICTI·CM·XXX

Which is the natalis of Invictus (rather than Sol) and 30 races, rather than a multiple of 12.  It is, therefore, an anomalous entry.

Hijmans makes some very interesting points about this.

• Firstly, he argues that celebrating festivals with chariot races rather than sacrifices was an innovation of Constantine, introduced after Constantine defeated Licinius in 324.  It’s not an ancient thing.  So all these chariot races were introduced then.
• Secondly, since all the ancient festivals were multiples of 12, it is clear that no festival of Sol existed on 25 December at that time.  If it had, it too would be a multiple of 12.  Therefore it is a later addition; as the irregular naming also indicates.
• Thirdly he speculates that this entry may not even have been present in the original copy made in 354, but added later.
• This leaves the first definite mention of a solar festival on this date to Julian the Apostate’s Hymn to King Helios, in December 362.

This is an interesting argument indeed.  What do we make of it?

Hijmans does not detail his first point, merely referring to M. Wallraff, Christus Verus Sol (2001), p.132, “citing Eusebius”. Unfortunately the Wallraff volume is inaccessible to me.   So we have to leave this point unchecked.

The second point relies on the accurate transmission of numerals in copies of the Chronography.  I am not clear whether this is actually reliable, or whether the text printed by Mommsen – which is the basis for the online version – is a critical text or not.   The Dec. 25 date could really have read “XXXVI” for all we know.

Obviously speculation, as in the third point, is not evidence.  I would suggest that we should not infer interpolation without need.

All the same this is a very interesting point.  Is it really possible that this was the case?

Update (5th August 2023): In August 2020 Dr Hijmans kindly responded to this article, and gave me permission to post his comments.  I am slightly ashamed that it took me so long to do so.  He wrote:

I have a few comments to supplement my arguments which you summarize so clearly in that blog.

1. Irrespective of whether we take the calendar (354) or Julian (362) as the first mention of a solar festival, it is interesting to note that there is no evidence before the mid 4th c. AD for a solar festival on December 25th or indeed on any of the astronomically significant days (if we take Dec. 25 to be the winter solstice). All traditional feast days for Sol are on astronomically random days. Thus there was no latent expectation, in antiquity, that the winter Solstice should be celebrated in honour of Sol. In other words the evidence we have for the celebration of Christmas (330s) is about a generation earlier than the evidence for a festival for Sol on that day.

2. The calendar mentions every celebrated emperor or deity explicitly by name, including Sol for the multi-day festival in October. Why is December 25 the sole exception?

3. Julian clearly states that there were two separate festivals of Sol in 362. One was the “newish” multi-day festival held every four years, and the other was an ancient one-day festival celebrated around the time of the winter solstice, established by Numa. If the newish, four-year festival is not the one founded by Aurelian, which is it then? If it is the one founded by Aurelian (as it surely must be), then on the evidence of Julian it was not celebrated on or around December 25, as that was the date of the annual one-day festival. The entry in the Calendar of 354 for a multi day ludi Solis on October 19-22 confirms this (which ludi were these, if not the ones of Aurelian). As the ludi were first celebrated in 274, they would also have been held in 254 (calendar, 20th games) and 262 (Julian, 22nd games). I really do not see any other way to read this evidence.

4. I think that the fact that Julian attributes the annual winter solstice Sol-festival to Numa is simply to give it pedigree. There is no evidence for such a festival, even in the late Republican and early imperial fasti (in which Sol is well-represented: 8/9 August, 29 August, also 11 December (if we accept Lydus).

To this I replied:

I do like your theory, and it would be convenient for me personally in arguing with the bat-witted “Christmas is really pagan” element among us. But for the same reasons I’m on my guard against it.

I like your argument that Julian knew that the October multi-day festival was newer; that would fit with Aurelian. But what, then, is the Natalis Invicti on 25 Dec? It works as the supposed anniversary of the dedication of the temple of Sol Invictus by Aurelian. But can we say that Aurelian created two festivals, and that Julian knew only of one and supposed the other to be from the days of Numa, i.e. traditional (I wouldn’t see this wording as anything but a rhetorical flourish meaning “very ancient”)? Do we see it as some form of dressing up of the solstice that ordinary people celebrated anyway, (was it with torches?)? If so, any 3rd century emperor could have created it. Maybe even a 4th century emperor. Does it have to have a deeper significance? You make a good point about the absence of mention of either emperor or deity – is that why?

Your point about the Roman failure to mark astronomical events is fascinating. Hmm!

I can also see Julian creating a fake festival, to undermine a Christian one. He was an intelligent persecutor,and his methods have been adopted ever since. But I don’t know how much time he even spent in Rome. And … would he care? It would only affect Rome, after all. I’ve never looked at the data for Christmas in the 4th century – was it widespread? If not, why would he bother? I think of him as mainly interested in the Greek east.

But at root, I don’t much like hypothesising an interpolation of this stuff into the Philocalian calendar. The Chronography was, after all, a physical book – a splendid artwork. It could well have had something added into it at an early stage, before any copies were made. Just as Jerome added material into the empty spaces in Eusebius’ Chronicon, someone could have added in the material about DNSI. It’s true. But I don’t like it. It feels way too much like the lazy German scholarship of the 19th century – I almost wrote “the last century”! – which treated inconvenient data as something to be excused. There’s no evidence of this. We have so little evidence, that we can not afford to discard any of it.

So … I am hesitant.

Dr H. kindly responded:

I am hesitant too to conclude with certainty hat Christmas preceded a pagan winter celebration of Sol. I (try to) state simply that there is no evidence for it. The evidence against it is essentially an argumentum e silentio.

Julian states unequivocally that the multi-day, quadrennial festival that he did not celebrate around the time of the winter solstice was “newish” and that the one-day festival which is at the heart of his hymn to Helios goes back to Numa. He does not say when the multi-day festival was celebrated, but I think the October date given by the Calendar makes that quite clear.

I would qualify that as “established Roman religion connected with Sol” did not show particular interest in astronomy (which – given the state of the Roman calendar up to the Julians, should not really surprise us, in hindsight.) But even Roman imperial religion shows no real sign of this, at leas as fat a Sol is concerned. To the best of my knowledge even Mithraism had not particular festival on the winter Solstice. I would have to check, but I thought the Tienen evidence placed the major Mithraic celebrations (as established by buried remnants of feasts) in the Spring? That they celebrated December 25 is not, I believe, supported by any actual primary sources…

Julian’s hymn to Helios sounds to me like a work written specifically to promote the solar celebration of December 25. I don’t think that necessarily means he ‘invented’ it. There are various ways in which one can imagine such a festival to have arisen in the fourth century, with collective “memory” being a very likely one. I think it is very well possible that Julian believed the convenient “fact” of a solar festival on the 25th. His insistence on the importance of that festival is striking.

As for interpolation. We need to explain the major anomalies of this entry (number of chariot races, wording of the entry). The calendar gives no hint of what those reasons were. I would hazard that a later interpolation is the most likely interpretation – not in the Calendar of Filocalus himself, but in the “mother calendar” which it copied. An entirely pagan official calendar Rome would not be unthinkable in the 350s, but may have prompted somebody to pencil in – say – the birthday of Christ (in Rome around the 320s, much late in most other cases we can identify)

Very useful points indeed.  Thank you!

1. ^[1]H. Usener, “Sol Invictus”, RhM 60 (1905) pp. 465-491.↩
2. ^[2]Steven Hijmans, “Usener’s Christmas: A contribution to the modern construct of late antique solar syncretism”, in: M. Espagne & P. Rabault-Feuerhahn (edd.), Hermann Usener und die Metamorphosen der Philologie. Wiesbaden, Harrassowitz, 2011. 139-152.  Online here, although the online version appears to be a draft.  However Hijmans’ full thesis, with extensive plates, is online here.↩