Humanities desk
< June 12	<< May \| June \| Jul >>	June 14 >

Welcome to the Wikipedia Humanities Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

June 13

Ancient Romans and Eight-fold Division

The ancient Romans had weeks of eight days, called nundinae, miles consisting of eight stadia or furlongs, and their smallest military unit was the contubernium, consisting of eight men. What is the reason for this ? Using only the four fingers of both hands for counting, to the exclusion of the thumbs ? Computational ease or convenience, based on repeated division by two ? Or something else entirely ? — 79.113.198.41 (talk) 10:36, 13 June 2017 (UTC)[reply]

In some cultures it comes from counting using the gaps between fingers - see Octal. There seems to be no way of knowing whether that applied to the Romans. Wymspen (talk) 12:19, 13 June 2017 (UTC)[reply]

Yes, I was aware of that. — 79.113.239.167 (talk) 02:42, 14 June 2017 (UTC)[reply]

79.113.198.41 -- I would tend to doubt whether there's a great prevalence or predominance of the number eight in ancient Roman culture. Note that the word nundinae comes from words meaning "nine days", due to the ancient method of inclusive counting. The calendar of the Roman Republic was structured so that there were sixteen days between the "Ides" of a month (originally representing the astronomical full moon, though by the time we know much about the ancient Roman calendar the months were no longer in phase with the moon) and the "Kalends" of the next month (originally representing a thin crescent visible at sunset, not the same as astronomical new moon or conjunction), and that might have had some influence on the idea of 8-day intervals. The "Nones" were scheduled to be 8 days before the Ides in the calendar, so all the main monthly landmarks were defined by 8-day or 16-day intervals... AnonMoos (talk) 13:47, 13 June 2017 (UTC)[reply]

AnonMoos took the words right out of my mouth. The sixteen days did not include the ides and the kalends themselves, and it didn't apply to February anyway. The time from the nones to the ides was exactly one nundinum, February not excepted.79.73.131.8 (talk) 14:20, 13 June 2017 (UTC)[reply]

I guess you're right -- it's the last day of the month which is 16 days after the Ides, while the Kalends is the next day. The year originally began in March, so the end of February was always where adjustments were made that caused February to be somewhat inconsistent with the rest of the months (as still today)... AnonMoos (talk) 08:02, 14 June 2017 (UTC)[reply]

The fact that a quarter of the moon (73⁄8 days) has a time-length between 7 and 8 days might have indeed contributed to adopting precisely such a time unit; however, moon cycles (of 291⁄2 days) fit better with 28-day periods (four Babylonian weeks) than with 32-day periods (four Roman weeks). — 79.113.239.167 (talk) 02:42, 14 June 2017 (UTC)[reply]

79.113.239.167 -- There wasn't any division of the Roman month into four 8-day periods. Rather, the main landmarks of the Roman month were structured so that the Nones was 8 days before the Ides, while the last day of the month was 16 days after Ides. The interval before the Nones was left adjustable -- the Nones were sometimes on the 5th day of the month, sometimes on the 7th day (never on the 8th). As I said before, the Ides-Kalends interval was probably not an approximation to half a lunar month (which would be 14.77 days), but rather an approximation of the average time between an astronomical full moon and the visibility of a thin crescent at sunset (I said 16 days above, but actually 17 days)... AnonMoos (talk) 08:02, 14 June 2017 (UTC)[reply]

But why approximate 17 by 16 = 2×8, and not by 18 = 2×9 ? The Welsh and Lithuanian calendar, for instance, are known to have had nine-day weeks. — 79.113.239.167 (talk) 14:52, 14 June 2017 (UTC)[reply]

79.113.239.167 -- 17 days is the approximation, not the thing being approximated. I'm not sure what the actual average interval between a full moon and the following first observation of a thin crescent moon at sunset is -- it would depend on latitude and other factors. AnonMoos (talk) 15:18, 16 June 2017 (UTC)[reply]

As our Latin teacher told us:

March, May, July, October,

These are they,

Make Nones the seventh,

Ides the fifteenth day.

The fact that August later acquired 31 days (as did January and December at the same time) gave rise to the story that Augustus jealously filched a day from February to make his eponymous month the same length. Most of the months (apart from February, which always had 28) had 29 days. The Ides - Kalends relationship is more likely to be related to the Romans' abhorrence of even numbers than assiduous moonwatching by the pontifices. 86.176.19.17 (talk) 12:46, 14 June 2017 (UTC)[reply]

8-day weeks would make it easier to assign shifts, such as alternating working 4 days on and 4 days off. 7-day weeks are a bit messy, if you want everyone to work the same number of hours as everyone else (you can do 7 days on then 7 days off, but 7 days of work in a row can be a bit much). StuRat (talk) 15:07, 13 June 2017 (UTC)[reply]

In other words, your answer is equivalent to the `computational ease or convenience` mentioned above, right ? — 79.113.239.167 (talk) 02:42, 14 June 2017 (UTC)[reply]

Yes, and specifically modular arithmetic is involved when dealing with repeating time periods. StuRat (talk) 03:27, 14 June 2017 (UTC)[reply]

I think the more typical usage would have been for farmers to come into town from their surrounding fields to attend a market held every eighth day. AnonMoos (talk) 16:42, 13 June 2017 (UTC)[reply]

And why exactly did farmers held their markets every eighth day, as opposed to every seventh, ninth or tenth day ? — 79.113.239.167 (talk) 02:42, 14 June 2017 (UTC)[reply]

Probably no very deep reason. A seven-day astrological week started spreading from the Greek population of Egypt to other parts of the Roman realm in the 1st century B.C. or A.D. (before there was much direct Jewish or Christian influence on Classical culture), eventually supplanting the earlier 8-day cycle. AnonMoos (talk) 08:02, 14 June 2017 (UTC)[reply]

Why do you call it an "astrological" week? I am interested in reading more about the origin of the week. I have a hunch that dividing the month into 4 may have been a parallel to dividing the year into 4, but I haven't seen much good scholarship on these questions. Itsmejudith (talk) 11:19, 14 June 2017 (UTC)[reply]

It doesn't seem to be very clearly explained in our Week article, but you can read about it in a number of books, such as Mapping Time: The Calendar and its History by E.G. Richards (ISBN 0-19-286205-7 p. 268). Probably around the 1st century B.C. (though the dating isn't known very specifically), some astrologically and/or mystically inclined Greek-speakers in Egypt assigned every hour of every day a planetary "ruler" from among the 7 classical planets; the cycle repeated continuously every seven hours. (For a significant number of people in Hellenistic civilization, Babylonian-influenced astrology had pretty much replaced religion for many purposes.) If you say that the planet which rules the first hour of the day rules the day as a whole, then you get a 7-day cycle. It was this form of the week which was spreading within the Roman Republic / Empire at a time when Christians and Jews didn't yet have much influence on classical culture. It's also the astrological week which gives days of the week planetary names such as "sun day", "moon day" (which never existed in Jewish culture or religion). AnonMoos (talk) 12:07, 14 June 2017 (UTC)[reply]

The four periods of the day (sunrise, noon, sunset, midnight), the month (new moon, half-moon, full moon, half-moon, new moon again), and the year (the four seasons: spring, summer, autumn, winter) arise naturally within each. They do indeed parallel one another, but not because of any artificial intent or specific desire to force such an effect. — 79.113.239.167 (talk) 14:38, 14 June 2017 (UTC)[reply]

Seasons shows several different systems: 2 seasons (dry and wet), 3 seasons (dry, wet, mild), 3 seasons (flood, growth, and low water), 4 seasons (aligned to equinoxes or cross-quarter days or solar terms or calendar months), 6 seasons (with spring in February or March or April depending on system). Rmhermen (talk) 15:49, 14 June 2017 (UTC)[reply]

In temperate climates, the most obvious division is summer and winter, plus the two in between. — 79.113.239.167 (talk) 22:07, 14 June 2017 (UTC)[reply]

Stu do you have a reference for ancient Romans working 4 days shifts, or did you just make that up? CodeTalker (talk) 17:08, 13 June 2017 (UTC)[reply]

I never claimed that Romans worked 4 day shifts. Stop making things up. StuRat (talk) 17:09, 13 June 2017 (UTC)[reply]

If your comment wasn't related to ancient Roman practices, how is relevant to this question? CodeTalker (talk) 20:13, 13 June 2017 (UTC)[reply]

The Q asked about the advantages of groups of 8, and specifically 8-day weeks, and did not ask us to restrict answers to ancient Rome. StuRat (talk) 02:40, 14 June 2017 (UTC)[reply]

Actually, that's exactly what I was doing, as is clear from both the title and the text of the question. — 79.113.239.167 (talk) 02:45, 14 June 2017 (UTC)[reply]

Well, that's not how I read it. Feel free to ignore my answers then. StuRat (talk) 03:20, 14 June 2017 (UTC)[reply]

...Is that permission specific to this thread or are you giving it generally... ;)

Absolutely, everyone should feel free to always ignore everything I say. Note however, that "ignore" means you won't be making disparaging remarks about them, either, and, since making disparaging remarks appears to be to main purpose in life for most Ref Desk regulars, I doubt if I will get many takers. StuRat (talk) 13:37, 14 June 2017 (UTC) [reply]

That is in itself a disparaging remark about your beloved brethren. We grieve for your prodigality, but pray ceaselessly for your return from the outer darkness where there is weeping and gnashing of teeth, to the bosom of your ref desk family where fatted calves are waiting impatiently to be slaughtered and eaten and vats of wine are brimming over in their haste to be quaffed in voluminous quantities, to celebrate the return of the lost one who has come home at last. -- Jack of Oz ^{[pleasantries]} 22:08, 14 June 2017 (UTC) [reply]

So let's plug in some numbers (data from [1]):

Full moon                    New moon sighted             Location               Time interval
May      10d 21:41           May      26d                 Sri Lanka              15d 15h 
April    11d 06:06           April    27d                 Brisbane               16d 02h
March    12d 14:52           March    29d 03:32           California             16d 12h 40m
February 11d 00:31           February 27d 18:45           Saudi Arabia           16d 15h 14m 
January  12d 11:31           January  28d 22:55           Virginia               16d 11h 24m 
December 14d 00:03           December 30d 11:46           Java                   16d 11h 43m

How was the seventeen days approximation calculated? 2A02:C7F:BE2D:9E00:C049:10E7:D61B:C61C (talk) 18:16, 16 June 2017 (UTC)[reply]

For starters, they observed the thin crescent following the new moon, and not the actual darkening itself (which they probably regarded as an end, not as a beginning, for the same reason they also [initially] considered night as the end of day, and winter as the end of the year). Furthermore, although the moon is visible during daytime (something I've noticed since I was a small child), people don't pay attention to it until night time. So, adding 24 hours to the new moons, and rounding all those occasions on which the moon gains its fullness during the day to the nearest night, we easily get 17 or 18 days. — 79.113.241.132 (talk) 11:13, 17 June 2017 (UTC)[reply]

Average will be one half synodic month plus how long it takes after New Moon to see it again. For 17 days to be the better approximation this has to be at least 16.5 days. 16.5 days minus 14.7653 days is 1.7347 days. Minus slightly under .5 days for the average interval between first visible crescent at your latitude to when you can see it and it implies an average of over 1.2347 days till the Moon is visible anywhere on your latitude. That seems a bit high but maybe the very early Romans just weren't very diligent moonsighters (their calendar would've had to have switched to non-lunar very early and they weren't into astronomy much. When they wanted to invent something as simple as the Julian calendar they had to ask a Greek guy.) Sagittarian Milky Way (talk) 23:22, 16 June 2017 (UTC)[reply]

2A02:C7F:BE2D:9E00:C049:10E7:D61B:C61C -- 17 days is the interval between the Ides of one month (except sometimes February) and the Kalends (first day) of the next month in the calendar of the late Roman republic. If the choice was limited to an odd number (since the Kalends, Nones, and Ides were always on odd dates, and the lengths of the months other than February was always either 29 or 31), then 17 was a better approximation than either 15 or 19... AnonMoos (talk) 02:33, 17 June 2017 (UTC)[reply]

The way I see it, counting inclusively, there are always (excepting perhaps February) exactly 9 days from nones till ides, and exactly 18 days from ides to kalendae (2×9 = 18). The latter half of each month always has 16 days (2×8 = 16). As an aside, the semester between the day/night `evening` (equalization) of March 17, and the day/night `evening` of September 25, can easily be regarded as six months of 32 days, and the semester between the day/night `evening` of September 25, and the day/night `evening` of March 17, can easily be regarded as six months of 29 days. Since Romans ignored much of the latter interval, and since a (moon)quarter of 32 is 8... but this is just idle speculation. — 79.113.241.132 (talk) 11:13, 17 June 2017 (UTC)[reply]

Not "sometimes February". February had either 23 or 28 days (the missing five days correspond to the epagemonai in solar calendars where twelve months of thirty days totalled only 360 days). By dividing the month into portions of 23 and 5 days the Romans sought to avoid the bad luck associated with months containing an even number of days. The ides were always on the thirteenth, so there was never a seventeen - day interval between them and the calends of March. A fifteen - day interval happened when the mensis intercalaris, which was sometimes inserted after the 23rd of February, had 27 days. 2A02:C7F:BE2D:9E00:A2:C382:E0D6:2FD7 (talk) 11:09, 17 June 2017 (UTC)[reply]

Inclusive counting doesn't work like that. A good analogy is a fence with posts and rails. An eight - panel fence (representing one nundinum) has nine posts. A sixteen - panel fence (two nundina) has seventeen posts. This is why the French call two weeks (what we call a "fortnight") quinze jours). 92.8.219.72 (talk) 13:46, 17 June 2017 (UTC)[reply]

On the same principle, when a cycle of seven variables, (e.g. days of the week, notes in a musical scale) returns to its starting point this is known as an "octave". It extends to poetry - we have an editor "Ottava Rima". I thought this was a Canadian name but it turns out to be Italian. 92.8.219.72 (talk) 14:05, 17 June 2017 (UTC)[reply]