Tuesday, April 08, 2008

One year away!

[CORRECTION (8/12/08): The sources say that the sun (along with the moon and stars) was created at the beginning of the fourth day, i.e. Tuesday night -- NOT sunrise on Wednesday. The blessing is said at sunrise because that's the first time that the sun is visible. So everything in the analysis below should be shifted 12 hours earlier, but the result of a 28-year cycle is the same.]

(Crossposted to Jewschool.)

One (solar) year from today, Wednesday, April 8, 2009, is birkat ha-chamah, the blessing for the sun that is said only once every 28 years!

Why every 28 years?

The short version: It's based on bad science, but it's still cool to have something like this that only happens a few times per lifetime. We should brainstorm about how to take advantage of this opportunity!

The longer version:

There are two Jewish calendars. One is the lunar calendar (with leap months added to keep it generally in sync with the solar year), which determines the dates of all the holidays. The other, much less well-known one is the Jewish solar calendar, which is used for only two purposes: birkat ha-chamah (the subject of this post), and the date to begin praying for rain in Babylonia and by extension the rest of the Diaspora (in this century, December 4).

The Jewish solar calendar is in essence the Julian calendar: it assumes that a solar year is exactly 365 1/4 days (thus adding a leap day every 4 years). This approximation is off by about 3 days every 400 years, which is why the world has switched to the Gregorian calendar, which eliminates 3 leap days every 400 years to account for this discrepancy (1600 was a leap year; 1700, 1800, and 1900 were not; 2000 was; 2100, 2200, and 2300 will not be; etc.).

This error in the Julian calendar is the reason that the date to begin praying for rain in the Diaspora is on December 4, despite the statement on Ta'anit 10a that it should be 60 days after the equinox (which would be around November 20): because of the error that has crept in over the centuries, "November 20" on the Julian calendar is now December 4 on the Gregorian calendar. In 2100, the error will increase by one day, so the date to begin praying for rain in the next century will be December 5, and after 2200, it will be December 6, and so on.

So, birkat ha-chamah.

According to the biblical creation story, the sun was created on the fourth day of creation, i.e. Wednesday. What time? At sunrise, of course. What time of year? The vernal equinox (according to Abaye on Berachot 59b).

So we say birkat ha-chamah every time the sun returns to the position it was in when it was created, at the anniversary of its creation: Wednesday at sunrise on the vernal equinox. Let's calculate that anniversary, using the Jewish solar calendar, in which 1 year = 365 1/4 days = 52 weeks 1 1/4 days. If we start with Wednesday at sunrise and add 1 year, we can disregard the complete weeks and just add 1 1/4 days, so we get Thursday at noon. Add another year and we get Friday at sunset. Add a third year and we get Saturday night at midnight. Add a fourth year and we get Monday at sunrise. So now (after 4 years) we're back at sunrise, but on the wrong day of the week. We have to cycle through all 7 days of the week before we can get back to Wednesday. So we get back there every 4 x 7 = 28 years!

The "vernal equinox" currently occurs on April 8 on the Gregorian calendar, due to the aforementioned error that increases over the centuries. So during the 20th and 21st centuries, birkat ha-chamah is always on Wednesday, April 8, at sunrise. The last one was on Wednesday, April 8, 1981, and the next one will be Wednesday, April 8, 2009! After 2100, it will be on Wednesday, April 9; after 2200, it will be on Wednesday, April 10; and so on.

There's no way to "fix" this error in regard to birkat ha-chamah, because once you get rid of the fiction that a year is exactly 365 1/4 days, you lose the entire premise (outlined above) for the 28-year cycle. In real life, there's no reason to think that the number of days in a year (i.e. the ratio of the earth's revolution around the sun to the earth's rotation about its axis) is any simple rational number. So birkat ha-chamah is what it is.

So the reason I'm posting this now is so that we can start thinking about what we're going to do next year to celebrate this once-in-a-generation occurrence. What are your plans for birkat ha-chamah? If you did anything for it in 1981 or 1953, what did you do? What are your creative ideas? What are our communities going to do? Start commenting away!

Things get a little bit sticky when we cross-reference Wednesday, April 8, 2009 with the normal lunar Jewish calendar. Turns out it's Erev Pesach. (Not surprising, given the connection to the vernal equinox.) So the Jewish community will be busy. But we should see this, too, as an exciting opportunity. (Say the blessing over the sun, and then take out a magnifying glass and use it to burn our chameitz? Anything is possible!) So if we're traveling for Pesach, we can participate in the awesome birkat ha-chamah event we're going to come up with in the city where we live early in the morning, and then hop a plane to wherever we're going for Pesach. (Heck, some of us will have to take the day off anyway to accommodate Pesach travel, which frees us up to do birkat ha-chamah stuff in the morning.)

Bonus question: For those who observe Ta'anit Bechorot (I don't, for various reasons, though I'm a firstborn), is birkat ha-chamah itself sufficient justification for a se'udat mitzvah so that people can break the fast? This question would have last come up three cycles ago, on Wednesday, April 8, 1925, which was also Erev Pesach.


  1. "The "vernal equinox" currently occurs on April 8 on the Gregorian calendar"

    Did you mean "Julian"? Because I'm pretty sure it fell on 3/20 in the Gregorian.

  2. Sorry I wasn't clear. The "vernal equinox" (scare quotes meaning that it's not an actual astronomical event anymore) is on March 20 on the version of the Julian calendar used for birkat ha-chamah, which corresponds to April 8 on the Gregorian calendar. Yes, the real vernal equinox is still on March 20 on the Gregorian calendar.

  3. I'm confused, though. My understanding is that the Julian and Gregorian calendars are currently 13 days out of alignment, which is why Julian December 25 (Orthodox Christmas) is Gregorian January 7. But that would mean that Julian March 20 is actually Gregorian April 2. Are we just working with a bad estimate of the equinox?

  4. Oh, one more thought: 5769 = 1 mod 28.

  5. I was using shorthand in the post -- I didn't mean the actual Julian calendar (as used in the Roman and Christian worlds), I just meant "a calendar with an average year length of exactly 365 1/4 days", which can in principle have any offset from the Gregorian calendar, depending on how long ago they diverged.

    That said, I share your confusion. Even if the Jewish solar calendar isn't the same as the real Julian calendar, it should at least be consistent with itself.

    (April 8) - (March 21) = 18 days, indicating that the calendars diverged around 2400 years ago.
    (December 4) - (November 21) = 13 days, indicating that the calendars diverged around 1700 years ago.

    Any explanations for this difference?

    Part of it may be mistaken dates for the astronomical equinoxes. The actual equinoxes and solstices aren't precisely evenly spaced, due to mumble mumble. (Eccentricity of the earth's orbit???) Using the data from this year, the spacings are approximately:
    March to June: 92 days 18 hours
    June to Sept: 93 days 15 hours
    Sept to Dec: 89 days 21 hours
    Dec to March: 89 days

    In contrast, the tekufot (equinoxes/solstices on the Jewish solar calendar) are exactly 91 days 7.5 hours apart (365 1/4 days divided by 4).

    If we assume that the vernal equinox is March 20 and add (91 days 7.5 hours) * 2, we get September 18 for the autumnal equinox, and November 16 for the 60th day starting at the autumnal equinox.

    (December 4) - (November 16) = 18 days, the same as the difference between March 20 and April 8! Aha! We have consistency! So you've solved it. Thanks for letting me think out loud here. :)

  6. Excellent analysis, BZ.
    This needs further exploration (which I can only pose, not answer).
    An error of 13 days corresponding to 1700 yrs ago is *very* close to ~360 CE, which is when Hillel II put the calendar in writing. This seems a logical date for the error to have started creeping in, as opposed to 2400 yrs ago, ~400 BCE, corresponding to what?
    Hopefully those more well-versed in this field than I can shed light.

  7. I'm guessing based on this historical argument that the 13-day error is the correct one, meaning that they were using the correct date for the autumnal equinox, but observing the vernal equinox around March 25 (due to the error of assuming that each season is exactly 91 days 7.5 hours).

    This is the opposite of what I assumed in my previous comment - that the vernal equinox was (correctly) on March 20 and the autumnal equinox was observed on September 18.

    But it could also be anywhere in between.

  8. One way to reconcile the tradition with the actual length of the solar year which is not a simple rational number would be to observe birkat hachamah whenever the calculated moment of the equinox falls during the first quarter of a (Jewish) Wednesday i.e. the last quarter of a Tuesday (between sunset and midnight) — er, but it's always dusk somewhere, so we'd have to pick a location, with Israel being the obvious choice.

    This would produce an approximately 28-year cycle while keeping birkat hachamah aligned with the actual equinox. You'd lose the fiction of the sun being in the "exact same spot", but we've already lost that by noticing the the vernal equinox isn't on April 8th! Also, this system would do intriguing number-theoretical things when it did end up adjusting itself.