The Phase of the Year: Something the Gregorian Calendar Gets Very Wrong

In the first article of the series, we concluded that days and years are the appropriate periods of our calendar. In the second article, we determined the optimal phase of the day. Each day should begin and end on midnight because of its astronomical relevance and biological convenience. That leaves us with the phase of the year. The period of our year is already synchronized with the seasons. We just need to decide where among the seasons we should divide one year and the next. This is the first time, though not the last, that we will find that the Gregorian calendar falls over completely.

Agriculture

We are not a hibernating species, so unlike the day, there is no obvious coordination point where we are all sleeping in which we could put the start of the year. That being said, the earliest calendars were mostly maintained for agricultural reasons. As such, calendars from around the world tend to flip over sometime before sowing and after reaping.

The growing season varies wildly across the Earth. This goes beyond the fact that the seasons of the northern and southern hemispheres are six months out of sync with each other. Even in the same hemisphere, the agricultural cycles can be many months out of sync. The ancient Roman calendar starts in March, to reflect the time of planting in Italy. The Jewish calendar starts in October to reflect the start of the sowing season in Israel. The Chinese calendar starts in January or February, just before the spring sowing season in China.

One could argue that agriculture is no longer a relevant factor in building a civil calendar. The fraction of our species's labor committed to agriculture has fallen substantially since the Industrial Revolution. With the USA as an example, 95% of its population worked in agriculture in the 1700s, but that has fallen to 2% in the 2000s. Agricultural considerations no longer drive our civilization, and we may be prudent to drop it from consideration like we did the phases of the moon in the first article.

Seasonal alignment

There are four obvious alignment points for a solar calendar: the winter and summer solstices and the spring and autumn equinoxes. This is not without precedent. The Iranian and Indian calendars start on the spring equinox. The French Republican calendar starts on the autumn equinox.

Solar calendars don't have to start with any astronomically meaningful event. The Gregorian and Bangladeshi calendars are two such examples. As long as they have a good leap year system, they'll repeat whatever random day they started with, but the main point of this article is that this is such a waste of a new year.

Zodiacal alignment

Because we have already established that we want to align our calendar with the seasons, I am only going to briefly mention one other way to align a calendar and that is with the zodiac. Under such a system, the new year starts when the sun hits a particular point in the sky. The Thai lunar calendar, despite the name, is one such calendar, starting when the sun enters Aries. The Tamil calendar is another such calendar, starting when the sun enters one of their constellations.

As discussed in the first article in this series, calendars aligned with the zodiac will drift relative to the seasons due to the precession of the equinoxes. Because the seasons are what is important to living on Earth, we won't consider the sun's movement relative to the stars in finding an optimal starting point for the year.

Gregorian calendar

January 1. It's the new year of the Gregorian calendar. Other than that, it's a boring day. It's not the start of winter or spring or summer or fall. It's not even the date the sun enters some constellation or rises with some notable star. It's just an astronomically uninteresting date adrift between the waypoints of the seasons.

The ancient Roman calendar began March 1, just before spring in the northern hemisphere and actually pretty close to the modern March 1. This is sowing time in Italy. The calendar ran for ten months through the harvest. In the winter, the calendar simply did not exist until spring came around again. It was a lunisolar calendar—the year started on a new moon.

Eventually, the Roman calendar filled out with the months of January and February and the start date of the Roman calendar was moved to January 1. This didn't matter for agricultural purposes because the division of the year was still between reaping and sowing, but some governmental activities started in winter, so it made sense to put the winter at the front of the calendar rather than the rear.

It was still a complex lunisolar calendar at this time. It takes a lot of work to keep the calendar in sync with the seasons while also syncing up with the moon. Relying on astronomical observations to determine if this year's officials would get an extra month in office also allowed some political shenanigans. As a result, Julius Caesar jettisoned leap months and established a simple solar calendar on the advice of the Sosigenes, who wisely agreed with the first article in this series in thinking that lunisolar calendars were not worth the effort. One leap day every four years would keep the calendar in sync with the seasons.

Well, not exactly, but we'll get to that in a minute. So what did they choose as the start of the New Year? The reason was not recorded, but if you run time backward, the new moon landed 45 minutes into January 2 on 45 BC. That's probably within the margin of error at that time to call January 1 the new moon—the first new moon after the winter solstice. They were smart enough to know that a solar calendar would not remain synced with the lunar cycle, but it looks like they gave themselves one last hurrah.

If this every-four-years leap year system were perfect, then the new year would start 7 days after the December 25 winter solstice every year forever. Instead, the Julian calendar drifts by 3 days every four centuries. It drifted enough that, by the time of the Council of Nicea in 325, the winter solstice had shifted to around December 21. No one did anything about this then, but this date is important because the Gregorian calendar is going to undo all the drift after this date, but not the drift before.

By the time 1582 rolled around, the drift had increased by about 10 more days. The winter solstice now fell on around December 11. The Church was concerned about the spring equinox, which was used to calculate Easter, but all seasons shifted equally. The fact that Easter jumps around by more than a month each year because of its insane formula (first Sunday after the first full moon after the spring equinox) apparently bothered no one, but the fact that it consistently occurred 10 days later than it did a millennium before did bother them.

So the Gregorian calendar was born. It fixed further drift, but also undid the ten days of drift since the definition of Easter was given. October 4, 1582, was immediately followed by October 15, 1582. While that must have been a massive administrative headache for everyone, I am genuinely impressed by the lengths they went to in order to get the calendar to behave the way they wanted.

And after all this, January 1 still means nothing. The first one was a new moon, seven days after the winter solstice. Then it drifted to twenty days after the solstice. Then it got reset to ten days after the solstice. And here it has been stuck for over 440 years.

The best we can do

Assuming we are making a tropical solar calendar from scratch, there are really only four sane options for the new year: the two solstices and the two equinoxes. If existing solar calendars were voting, the equinoxes would win. I could not find a single solar calendar that started on a solstice. Is this just coincidence? There aren't that many solar calendars, after all.

Is the equinox easier to observe than the solstice? If you know your latitude and the tilt of the earth, both are pretty easy to observe. Only a tiny fraction of Earth's population has ever lived beyond the arctic or antarctic circles, where the winter solstice is impossible to see. Maybe the equinoxes are easier to measure because that's when the angle of the sun is moving fastest?

Astronomically, the solstices feel like more natural starting points. The equinoxes are transition points between the axis of the earth pointing away from the sun to pointing toward the sun. The two equinoxes are basically the same. The sun moves in the sky the same way on both days, namely the day is half daylight and half nighttime. The only difference between a spring and an autumn equinox is that one is transitioning from winter to summer and the other from summer to winter. I'll concede that this is largely a matter of taste.

Even if we chose a solstice or an equinox, we still need to choose which one. Remember, the winter solstice of the northern hemisphere is the summer solstice of the southern hemisphere. The spring equinox of the northern hemisphere is the autumn equinox of the southern hemisphere. About 68% of Earth's landmass and 90% of Earth's population is in the northern hemisphere. That makes it relatively easy to prioritize the northern hemisphere, but that only helps us if we have a reason to pick one season over the other, such as agriculture. But even if we thought that agriculture was still important enough to drive our world calendar, only the temperate regions with their inhospitable winters consistently have growing seasons stretching from spring to fall. The tropical regions can be totally different.

Spring is culturally associated with new life and a new beginning. But this is undoubtedly a consequence of the English-speaking culture's roots in the temperate regions of the northern hemisphere. In the absence of an objective reason to pick among options, I would have no problem with deferring to the world's dominant culture. But the English-speaking world is not going to be comfortable with the new year starting in the spring anyway. Our cultural inertia is with the new year starting in the dead of winter.

With that, there is simply no principled way to choose one of the four seasonal waypoints over the others. That leaves just one final consideration to break the tie: status quo. The Gregorian calendar already starts really close to the winter solstice—just ten days after. If we dropped ten days from one year, the Gregorian calendar would start on the winter solstice every year (or thereabouts) and each season would get three whole months (again, thereabouts).

This is surprisingly fixable. We have lopped off 10 days from our calendar before. Do we have the courage to do it again? It would be a one-time cost in the form of the biggest time-keeping headache of the industrial age, but it would pay off as thousands of years of a calendar not just synced, but also aligned, with the seasons.