Why use a 12-Month Calendar?

[Drax]

Hey,

I was staring at my calendar today, and I began to wonder: why do we use this weird 12-month calendar?

Weeks are 7 days long, making 4 weeks 28 days. As far as I know, calendars were first created following lunar cycles, which would make this 28-day month even more significant. Why have this odd calendar system where the days-per-month alternate from 30 to 31 every month, with February being 28/29 days? This doesn't fit the lunar cycles at all, and from what I can tell doesn't really fit anything apart from the four seasons, which are still stuck in on the 21^st/22^nd of every third month.

Why don't we simply have a 13-month calendar with 28 days per month, and we could stick the extra day on the 13th month, making it 29 days long, and the leap-day could be tacked on the 6th or 7th month? This would allow us to maintain our 365.25 day year, and would more-or-less synchronize our calendar to the lunar cycles.

Now, I know it's far too late for the calendars to be changed, as all of technology relies on this system, and our history for thousands of years also used this system. What I want to know is, what led to the adaptation of this 12-month year that we currently use?

PS. My apologies if this isn't Serious Business-worthy, please move it to the correct place if this is the case.

[Sharlos]

AFAIK, the way dates work on computers it shouldn't be too difficult to change the number of days in a month.

[Drax]

AFAIK, the way dates work on computers it shouldn't be too difficult to change the number of days in a month.

It wouldn't be too hard, but that's alot of updates for alot of machines. Even my little sister's tamagochi would be wrong.

That's not really the point though, I'm wondering more about why we adapted the 12-month calendar

[EmilyR]

The modern, Western, Gregorian calendar that has become the de facto international standard (obviously there are other calendars in use in plenty of places) is a revision of the Julian calendar, which was a revision of the Roman calendar of Numa, which was in turn a revision of the calendar of Romulus, which was probably based on early Greek lunar calendars.

The Greek lunar calendars, and the earliest Roman calendars seem to have (mostly) had 12 months of 29(ish) days with various methods (such as intercalary days) used to keep them in sync with the year.
The calendar of Romulus had only 10 months with intercalary winter months used to keep things in sync. This is why the last months of the year are still called seven-ember, eight-ember, nine-ember and ten-ember.
The calendar of Numa added Jan & Feb to take it up to 12 months and do away with intercalary months.

[Azrael]

Although the answer is multifaceted, I doubt they the facets you're looking for:

1) Because we have been since 1582.
2) Because it's hard to change.
3) Because no one has demonstrated a need to change, nor a benefit brought about from changing (never mind one that would outweigh the efforts).

The last time it changed, it did so because the lunar approximations were off, and the error had accumulated such that the calendar lagged the moon by several days. Also, leap years weren't as rhythmic, and most importantly it was hard to figure out when Easter was supposed to be. Which I guess were pretty troubling?

As for 12 months, it's divisible by 4. Gregory's adoption of Lilius' proposal dropped 10 days to re-align the calendar with the seasons.

[Drax]

Those are definitely some interesting facts, EmilyR. I've always wondered why Sept- through Dec- ember were named as such, instead of being named after deities. Heh, killed two birds with one stone I suppose.

Azrael, you're right, those aren't exactly the answers I was looking for, but they did bring a couple other things into mind for me. First, I had no idea it was only since 1582 that we've been using this calendar, I had wrongfully assumed that it was at least two thousand years old as every reference to the Bible that I know of has the same dating system. Have these dates simply been modified to fit our calendar over revisions of the Bible?
Yes, it's hard to change, and I could only imagine how long it would take to figure out holidays, special events, and birthdays with a newly adjusted calendar.

I did have a feeling it had to do with the seasons, though. It was really the only thing that made sense to me as I don't really know the history of the calendar. This brings up another question for me: Most, if not all of the world use the 12-month calendar now, right? If this is based off the Greek calendars, how exactly did the entire planet decide to modify and accept our now-standard calendar five hundred years ago?

Looks like I'll be doing some reading on Wikipedia this afternoon

[EmilyR]

First, I had no idea it was only since 1582 that we've been using this calendar

The Gregorian calendar of 1582 was very closely based on the Julian calendar which was in use since 45BC, the changes were relatively minor, so it's been about 2000 years, allowing for leap year tweaks.

Most, if not all of the world use the 12-month calendar now, right? If this is based off the Greek calendars, how exactly did the entire planet decide to modify and accept our now-standard calendar five hundred years ago?

The Gregorian calendar was named after the pope who put it into place, subsequently it spread throughout western Europe (first the Catholic nations, then in the Protestant ones a century or so later). From there it spread throughout the various European empires of the 16th-19th centuries. It's now the dominant international calendar for pretty much the same reasons that English is the dominant international language.

[Drax]

I just saw a bunch of that on Wikipedia.

Well, I guess I could say that my questions have been answered!

Thanks guys

[pabechan]

It took a really long time for the European countries to adopt the Gregorian calendar. Greece adopted it as late as 1923. I have always found amusing that the russian October revolution (Julian cal.) actually happened in our November(Gregorian cal.).

[smw543]

Azrael, you're right, those aren't exactly the answers I was looking for, but they did bring a couple other things into mind for me. First, I had no idea it was only since 1582 that we've been using this calendar, I had wrongfully assumed that it was at least two thousand years old as every reference to the Bible that I know of has the same dating system. Have these dates simply been modified to fit our calendar over revisions of the Bible?

Yes. They used the Hebrew calendar, and added the Julian calendar after the Romans conquered the region in 63 BCE (that is, they continued using the Hebrew calendar for everyday purposes, but also kept track of the Roman date for taxes and such).

[Lunchpuppy]

I did have a feeling it had to do with the seasons, though. It was really the only thing that made sense to me as I don't really know the history of the calendar. This brings up another question for me: Most, if not all of the world use the 12-month calendar now, right? If this is based off the Greek calendars, how exactly did the entire planet decide to modify and accept our now-standard calendar five hundred years ago?

It's also worthwhile mentioning that while it is accepted as the international calendar by most nations, many countries have their own domestic calendar system. Taiwan, for instance, not only uses the lunar calendar in parallel with the Gregorian calendar, year one for their solar calendar is actually 1912, when the Republic of China was founded. That means this is the year 99. North Korea also starts its Juche calendar in 1912, since that is the year of Kim Il-sung's birth.

[Andromeda321]

Also I don't think anyone mentioned where the number 12 came from specifically. That actually has its origins in Babylon, where the number 12 was sacred as a magic number divisible by 2,3,4, and 6. The Egyptians carried this over by having a 12 month calendar of 30 days each month, plus a week's celebration for the new year. They didn't worry about this not working out perfectly as the year was defined by the first day the star Sirius was seen, so if it was a day off sometimes so be it.

But yes, as others said then the Greeks took it, then the Romans, etc etc. But it's amazing how many things have such a long history when you get down to it!

[Le Téméraire]

During the French revolution, the French government proposed a decimal calendar and hour system. Even in this new calendar, a year consisted of twelve months (all the rest was decimal though, three décades in a month, 10 days in a décade, 10 hours in a day, 100 minutes in an hour, 100 seconds in a minute). The reason why they also ended with twelve months, was that they wanted to relate their months to the natural seasons. Every three months was a season. The seasons are related to the movement of earth around the sun, so in the end, you still have the non-decimal number of 356 days in a year

[Me321]

The months are based on the moons apearance from earth, while it takes 27.3 days for the moon to orbit the earth it takes 29.5 days for the moon to apear the same from the point of view of a person on the planet.

So 1 year is 365.25is days divided by 29.5= 12.38 so if you round up to 30 day months and account for the .25 by adding a leap day every 4 years you get 365 divided by 30 = 12 with a remainder of 5, so the logic would be 7 months of 30 days and 5 of 31 with an extra 31 month evey 4 years (or 12 months of 30 days with a 5 or 6 day (leap year) holiday each year), why it is not this I have no idea, but it is close.

[not the XKCD Rob]

The months are based on the moons apearance from earth, while it takes 27.3 days for the moon to orbit the earth it takes 29.5 days for the moon to apear the same from the point of view of a person on the planet.

So 1 year is 365.25is days divided by 29.5= 12.38 so if you round up to 30 day months and account for the .25 by adding a leap day every 4 years you get 365 divided by 30 = 12 with a remainder of 5, so the logic would be 7 months of 30 days and 5 of 31 with an extra 31 month evey 4 years (or 12 months of 30 days with a 5 or 6 day (leap year) holiday each year), why it is not this I have no idea, but it is close.

Google "Iranian calendar". Their year consists of six 31-day months, followed by five 30-day months, and then one month of either 29 or 30 days, depending on the year. No "thirty days hath September" for them!

[billyswong]

Why don't we simply have a 13-month calendar with 28 days per month, and we could stick the extra day on the 13th month, making it 29 days long, and the leap-day could be tacked on the 6th or 7th month? This would allow us to maintain our 365.25 day year, and would more-or-less synchronize our calendar to the lunar cycles.

If you want the calendar synchronous to the lunar cycle, you don't use leap days but leap months. For example, take a look at the Chinese calendar, which is a lunisolar calendar. But beware: when the calendar strives to follow all astronomical events precisely, the exact pattern becomes nearly impossible to remember for any laypersons... (Chinese calendar do not fix the leap month after a particular month. Well, because of those picky true solar time variation issue, even the period of leap months is not fixed too. So the only way to know if there is a leap month next year or where it's placed is looking up the calendar unless you are a expert in astronomy.)

[Wnderer]

Lunar cycles? It's the 21st century. We need a calendar that works well with computers. We need a base 2 calendar. Divide the solar year into 1024 days. I'll call them n_days. Each n_day would be about 8.56 hours. So three n_days would be 25.68 hours. That's close enough to a day. So it doesn't match the solar day, somebody had 28 hour day concept. You work one n_day and have two n_days off. You sleep about an n_day. Divide each n_day into 8 n_hours. Each n_hour into 64 n_minutes and each n_minute into 64 n_seconds. We get an n_second that's about 1.06 seconds long. We also can break the year into 16 n_months of 8 n_weeks of 8 n_days. Now the whole thing fits into a nice 25 bit number. Of course we'd have to start the days and months with zero. In octal, month/day/year hour:minute:sec form midnight on the last day of this year would be 17/77/3732 7:77:77

[Soralin]

Lunar cycles? It's the 21st century. We need a calendar that works well with computers. We need a base 2 calendar. Divide the solar year into 1024 days. I'll call them n_days. Each n_day would be about 8.56 hours. So three n_days would be 25.68 hours. That's close enough to a day. So it doesn't match the solar day, somebody had 28 hour day concept. You work one n_day and have two n_days off. You sleep about an n_day. Divide each n_day into 8 n_hours. Each n_hour into 64 n_minutes and each n_minute into 64 n_seconds. We get an n_second that's about 1.06 seconds long. We also can break the year into 16 n_months of 8 n_weeks of 8 n_days. Now the whole thing fits into a nice 25 bit number. Of course we'd have to start the days and months with zero. In octal, month/day/year hour:minute:sec form midnight on the last day of this year would be 17/77/3732 7:77:77

Bah! Minutes, hours, days, weeks, months, years, what do you need any of those for? Just go with seconds: http://en.wikipedia.org/wiki/Unix_time :

Unix time, or POSIX time, is a system for describing points in time, defined as the number of seconds elapsed since midnight proleptic Coordinated Universal Time (UTC) of January 1, 1970

Example: 1273239097 (2010-05-07 13:31:37Z)

[PAstrychef]

Lunar cycles? It's the 21st century. We need a calendar that works well with computers. We need a base 2 calendar. Divide the solar year into 1024 days. I'll call them n_days. Each n_day would be about 8.56 hours. So three n_days would be 25.68 hours. That's close enough to a day. So it doesn't match the solar day, somebody had 28 hour day concept. You work one n_day and have two n_days off. You sleep about an n_day. Divide each n_day into 8 n_hours. Each n_hour into 64 n_minutes and each n_minute into 64 n_seconds. We get an n_second that's about 1.06 seconds long. We also can break the year into 16 n_months of 8 n_weeks of 8 n_days. Now the whole thing fits into a nice 25 bit number. Of course we'd have to start the days and months with zero. In octal, month/day/year hour:minute:sec form midnight on the last day of this year would be 17/77/3732 7:77:77

So waht's more important, that the calendar match observed phenomena (days, etc) or that it conform to the needs of the machines we use? In plemty of places the lenght of hours changed depending on the length of daylight, so hours were shorter in winter and longer in summer.
People who spend time outside alot seem to like using lunar cycles as a way of telling months, however awkwardly they match up with a solar calendar.

[Durinthal]

Lunar cycles? It's the 21st century. We need a calendar that works well with computers.

It's easier to make a computer keep up with our crazy system (whatever it is) than vice-versa. It would be difficult to make a day be anything other than the ~24 hours that it takes for the planet to rotate once, because then our time system and circadian rhythms (our biological ~24-hour day cycle) would quickly desynchronize.

[tastelikecoke]

I've been wondering too. A stupid confusinating calendar, that's not exactly useful or anything.

Since the calendar is usually based on astronomical objects, It might be a good indicator on the olden times, but today where the speed of everything is determined to the last decimal point and where super special numbers like 3 or 12 doesn't glow anymore, the calendar is now a mere lag on forward technology.

[PM 2Ring]

About a decade or so ago, I devised a nice solilunar calendar, with the epoch on a solar eclipse that falls on an equinox (sorry, I can't recall the exact details at present and my calculations are on another computer). In my calendar, each month begins on a full moon, and over the long term it tracks the vernal equinox year better than the Gregorian calendar does, but I don't expect I have much chance of getting it adopted.

[ikrase]

In space, people probably will use the decimal seconds calender.

[Josephine]

In space, people probably will use the decimal seconds calender.

I'm not so sure. I think for exploration missions, they would use our calendar. but for settling another planet, they would probably have to devise their own to fit the day/night cycle.

[Iv]

So 1 year is 365.25is days divided by 29.5= 12.38 so if you round up to 30 day months and account for the .25 by adding a leap day every 4 years you get 365 divided by 30 = 12 with a remainder of 5, so the logic would be 7 months of 30 days and 5 of 31 with an extra 31 month evey 4 years (or 12 months of 30 days with a 5 or 6 day (leap year) holiday each year), why it is not this I have no idea, but it is close.

From what I know, this is used to be the case in ancient Rome. Several months had deities names and every month had 30 or 31 days. More important deities had 31 days in their month. Then comes human ego. Julius Caesar was deified. Considering that he was not a lesser deity, he estimated he was worth a month. And not a 30 days one. Therefore, a new 31 days month was created (July) and one day was removed from February. Later, Augustus Caesar made the same joke (with August). Consider yourself lucky that they stopped there. A few more and February would have had 24 days.

[Wnderer]

So 1 year is 365.25is days divided by 29.5= 12.38 so if you round up to 30 day months and account for the .25 by adding a leap day every 4 years you get 365 divided by 30 = 12 with a remainder of 5, so the logic would be 7 months of 30 days and 5 of 31 with an extra 31 month evey 4 years (or 12 months of 30 days with a 5 or 6 day (leap year) holiday each year), why it is not this I have no idea, but it is close.

From what I know, this is used to be the case in ancient Rome. Several months had deities names and every month had 30 or 31 days. More important deities had 31 days in their month. Then comes human ego. Julius Caesar was deified. Considering that he was not a lesser deity, he estimated he was worth a month. And not a 30 days one. Therefore, a new 31 days month was created (July) and one day was removed from February. Later, Augustus Caesar made the same joke (with August). Consider yourself lucky that they stopped there. A few more and February would have had 24 days.

No.
http://en.wikipedia.org/wiki/Roman_calendar
http://en.wikipedia.org/wiki/Julian_calendar

First there were ten months. I'm not sure the names of the first four (March Mars, June Juno?), but the last six translate to versions of 5, 6, 7, 8, 9 and 10 not gods.

Then Numa Pompilius changed it to twelve months by adding two new months at the end. Then they moved the new year from March to January to put the names in the wrong order. He made months of 29 and 31 because odd numbers were lucky except for February which was the month of purification so it got 28. This required a leap month of 27 days. There were no firm rules for the leap month. Caesar added days to the months and a leap year every four years to get 365.25 days in a year. The actual number of days in a year is 365.2425. The Gregorian calendar fixed that.

Augustus changed the names of both July and August. Whether it was ego or political savvy is debatable.

[Ivor Zozz]

Why use months at all? Couldn't we just number all the days in a year and denote events by their day number and year?

[Iv]

From what I know, this is used to be the case in ancient Rome. Several months had deities names and every month had 30 or 31 days. More important deities had 31 days in their month. Then comes human ego. Julius Caesar was deified. Considering that he was not a lesser deity, he estimated he was worth a month. And not a 30 days one. Therefore, a new 31 days month was created (July) and one day was removed from February. Later, Augustus Caesar made the same joke (with August). Consider yourself lucky that they stopped there. A few more and February would have had 24 days.

No.

To be more precise :
http://en.wikipedia.org/wiki/Julian_cal ... th_lengths
Thanks, for the knowledge update !

[Regulus]

It's been a very long time since I've been on the xkcd forum, but I just stumbled across this thread in the process of checking where my new 13 month dynamic calendar website (http://calendar13.com) indexed on Google searches.

My calendrical system is the standard 13 month, 28 day system: 13 months with 28 days each and each year having a "day zero" outside of it to mark the end/beginning of the year. In my system, the leap day occurs at the end of the year on leap years which corresponds to December 20. This is because I used December 21 as day zero as I felt the beginning of the year should have some astronomical significance. The winter solstice made a nice convenient place to move "New Year's Day", especially because it is only an eleven day adjustment from the Gregorian calendar. As an added benefit to this move, day 1.11 becomes January 1 and the Gregorian leap day of February 29 becomes the pi day: 3.14. These, among other conversions I'm sure, provide some useful points of reference along the calendar which makes it easier to translate between the two in your head.

One of the reasons why past 13 month calendar attempts failed to catch on (in addition to the obvious societal bias against overhauling concepts/systems with lengthy traditional backing), is that they often retained the Gregorian month names. I have dropped this aspect opting for a very simplified numerical date format that resembles the decimal system (year is the biggest measure of time so it comes first, followed by month and day). The problem with Gregorian month names, is that the "number" months don't actually match the month numbers. September should be the seventh month, but it is the ninth. October should be the eighth month, but it is the tenth. November and December suffer from this problem as well.

Any comments, questions, or suggestions would greatly appreciated. More features are on the way.

[CorruptUser]

Keep in mind that the leap-year is not 'every 4 years', but every 4 years except in century years that are not divisible by 400. That is, the years 1500, 1700, 1800, and 1900 were not leap years, but 1600 and 2000 were.

[Regulus]

CorruptUser,

Yes, that is the rule that is implemented for leap years. I just over-simplified my description.

[EdgarJPublius]

I like the idea of calendars with time 'out of the year' or otherwise not included in the regular calendar organization, such as the 'calendar13' "day zero". It brings to mind the 'timeslip' from the Red Mars trilogy where the Martian colonists rectify the discrepancy between the 24 hour clock with Mars' slightly longer day by freezing the clocks at midnight for the extra time, creating a sort-of natural 'party-time' where things that you wouldn't do during the normal day might become acceptable, like a temporal Las Vegas.

In a calendar, that sort of 'missing' day also becomes a great place to stick leap time, be it leap days or leap seconds or what-have-you. So during a leap-year you have a two day long New year celebration instead of one.

I also like that this system would make the IERS (and by extension, Astronomical science) responsible for telling everyone they can party for just a bit longer, even if it's only by a second, every once in a while.

[PM 2Ring]

CorruptUser,

Yes, that is the rule that is implemented for leap years. I just over-simplified my description.

If you're going to the effort of creating a new calendar, why make one that's merely as accurate as the Gregorian? It's not that difficult to implement a more accurate leap year rule that better tracks the vernal equinox year, or the tropical year. Eg, you could have 8 leap days every 33 years, like the Iranian calendar.

[AvatarIII]

It's been a very long time since I've been on the xkcd forum, but I just stumbled across this thread in the process of checking where my new 13 month dynamic calendar website (http://calendar13.com) indexed on Google searches.

My calendrical system is the standard 13 month, 28 day system: 13 months with 28 days each and each year having a "day zero" outside of it to mark the end/beginning of the year. In my system, the leap day occurs at the end of the year on leap years which corresponds to December 20. This is because I used December 21 as day zero as I felt the beginning of the year should have some astronomical significance. The winter solstice made a nice convenient place to move "New Year's Day", especially because it is only an eleven day adjustment from the Gregorian calendar. As an added benefit to this move, day 1.11 becomes January 1 and the Gregorian leap day of February 29 becomes the pi day: 3.14. These, among other conversions I'm sure, provide some useful points of reference along the calendar which makes it easier to translate between the two in your head.

One of the reasons why past 13 month calendar attempts failed to catch on (in addition to the obvious societal bias against overhauling concepts/systems with lengthy traditional backing), is that they often retained the Gregorian month names. I have dropped this aspect opting for a very simplified numerical date format that resembles the decimal system (year is the biggest measure of time so it comes first, followed by month and day). The problem with Gregorian month names, is that the "number" months don't actually match the month numbers. September should be the seventh month, but it is the ninth. October should be the eighth month, but it is the tenth. November and December suffer from this problem as well.

Any comments, questions, or suggestions would greatly appreciated. More features are on the way.

just out of curiosity, could you convert some of the major holidays to the 13 month calendar? like Christmas, obviously the 25th December, but with the 13 month calendar, would you put it 25/12 or 7 days from the end of the year? (22/13?) or 3 days after the winter solstice? etc.

[Regulus]

If you're going to the effort of creating a new calendar, why make one that's merely as accurate as the Gregorian? It's not that difficult to implement a more accurate leap year rule that better tracks the vernal equinox year, or the tropical year. Eg, you could have 8 leap days every 33 years, like the Iranian calendar.

You are right, the leap year rules could be improved from the Gregorian system, but I wanted to keep the leap year rules the same to make it easier for people to migrate to the 13 month system from the Gregorian system (dates translate consistently, except for dates after February 29 on a leap year which are offset one day). It might not be the perfect ideal, but compromising on a few things for ease of migration was a priority of mine (as many 13 month calendars have been attempted in the past with very little ease in date translations).

I think it would be an appropriate time to discuss improving the leap time corrective measures only after a critical mass of people adopted a specific 13 month calendar system. It must first be really easy to make the switch.

[Regulus]

just out of curiosity, could you convert some of the major holidays to the 13 month calendar? like Christmas, obviously the 25th December, but with the 13 month calendar, would you put it 25/12 or 7 days from the end of the year? (22/13?) or 3 days after the winter solstice? etc.

I think the holidays, for the most part, would remain on the same day (I couldn't get my family to move Christmas! ), with an obvious exception for New Year's Day which would move from January 1 to December 21. Christmas for Gregorian year 2011, under the Calendar 13 system, is on 2012.1.4.

[Tirian]

3) Because no one has demonstrated a need to change, nor a benefit brought about from changing (never mind one that would outweigh the efforts).

I've worked for quite a few companies that have used a variant of the IFC where they have thirteen months of twenty-eight days. (The difference is that the IFC has a leap day or two every year, while the models that I've seen save up days for a leap week whenever necessary.) It makes a world of sense for retail work -- you don't want to compare today's sales with the last ten November twenty-seconds, you want to compare it to the last ten forty-seventh Tuesdays for a more even comparison, and of course you can't compare November's sales with October's because they had a different number of days and it's not unusual for them to have a different number of weekends.

[exporito]

I've worked for quite a few companies that have used a variant of the IFC where they have thirteen months of twenty-eight days. (The difference is that the IFC has a leap day or two every year, while the models that I've seen save up days for a leap week whenever necessary.)

Interesting model. may be it can be used sometimes))

[Proginoskes]

I've always wondered why Sept- through Dec- ember were named as such, instead of being named after deities.

January and February were later additions. "September" means "seventh month", "October" means "eight month", etc. December was until the proper day for March 1. "July" and "August" were named after two Ceasers (Julius and Augustus).

What I've always wondered about is how (some of) the days of the week ended up being named after Teutonic deities: Tiew, Woden, Thor, and Freya?

[aldonius]

The Romans appear to have used the 5 'classical' planets plus the sun and moon and simply called each day 'Day of celestialBody'. Then, the Germanic tribes interposed their gods on some days (and there are associations between the respective planets and gods). The Germanic tribes later settled in Britain and the rest, as the Yanks say, is history.
Since this is SB, here's my source: http://en.wikipedia.org/wiki/Weekday_names