Shoot me for saying this, but the whole birchas hachamah thing bores me. I mean, it stands to reason that a berachah that shows up once every 28 years would attract some attention. But that doesn’t justify two lines of gemara and half a se’if in Shulchan Aruch getting book-length treatments. With that degree of bloviation a lot of nonsense is bound to be said.

Here is the short version. The gemara (Berachot 59b) cites a beraisa that when one sees the sun, the moon, the planets and the constellations at some (unspecified) special point in their respective cycles, one should recite the brachah “oseh [maaseh] bereishis”. Abaye explains that with regard to the sun, the reference is to the point once every 28 years when the vernal equinox is on Tuesday evening (i.e., the beginning of what we call “yom revi’i").

If not for Abaye, one could have taken the reference to be to the annual vernal equinox. But let’s work with Abaye. The simplest explanation of his remark is that, since as a matter of convention a solar year is defined as 365¼ days, any given point during the year will fall out on the same day and hour every 28 years. (28 times 365¼ is divisible by 7 and no number smaller than 28 will do the trick.) This is true in particular for the vernal equinox – a distinguished astronomical point – and so we recite the berachah at the vernal equinox every 28 years. (Rashi frums it up a bit by tying the particular spot in the 28-year cycle to the original point of creation, a flourish not mentioned by the gemara or, subsequently, the Rambam. Why modern (?) commentators insist on seizing upon this flourish as the starting point for the discussion is a mystery to me. Whatever.)

By this account, the 365¼-day year is merely a convention, a sort of canonical approximation. There’s nothing to see out there every 28 years. And this shouldn’t bother anyone. After all, by all accounts, the 28-year number is crucially dependent on the seven-day week, which itself is merely a convention. And Hazal point out that our holidays are defined by decisions of the beis din, not by astronomical events (“eileh moadei hashem asher tikreu osam” – al tikri osam ela atem). In halacha, astronomical events are anchors that prevent unlimited calendrical drift, but they aren’t constitutive.

This approach has a certain post-modern cachet and neatly avoids commitment to (very very) bad astronomy. But still. You’ve got to wonder how people who know better – and many great rabbanim did and do know better – can mark a virtual vernal equinox 18½ days after the actual equinox (which is today, Friday, at 11:44 GMT).

I think the answer lies in the dual definition of equinox used by Hazal (and many others). The official (modern) definition of equinox is the (twice-a-year) moment when the line connecting the sun to Earth is perpendicular to the earth’s axis (the line from the South pole to the North pole). At that moment, the sun is moving east to west right along the equator.

But there’s another, closely related, definition that was more commonly used in the olden days and which is used by the Rambam: it is when the sun enters (the part of the celestial sphere identified with) the constellation Aries (taleh). The two definitions were once indistinguishable. But here’s the weird thing. The tropical year (the duration of one orbit of the earth around the sun) is a bit over 11 minutes short of 365¼ days, so that the vernal equinox by the standard definition is 18 days earlier than the virtual equinox Jews use. But the sidereal year (the time it takes for the sun to get back to the same spot relative to the constellations) is actually just over nine minutes longer than 365¼ days. And the definition of the part of the celestial sphere identified with Aries is a bit fuzzy. So while the virtual equinox has been visibly different than the actual equinox according to the standard definition going back to Amoraic times, the sun did continue to rise in Aries at the virtual equinox right up until about the year 1500. (The sun shifts relative to the constellations by about one degree every 70 years and each constellation is identified with 30 degrees worth of celestial sphere, so any point in the tropical year coincides with a given constellation for about 2000 years.) So by the sidereal definition of equinox, we were doing more or less okay for a while. By now, though, the actual (tropical) vernal equinox is well into Pisces and even our virtual equinox has slipped into Pisces. The sun actually enters Aries these days on April 15, a week after our virtual equinox. (Somebody tracks this because Indian astrologers use the sidereal cycle to do whatever it is Indian astrologers do. At least there are a few true believers left in this world.)

So in the end it’s all about the completion of a virtual cycle. That’s it. Move along, move along. Nothing to see here.

The other events mentioned in the same beraisa, the completion of some sort of cycles involving the moon, the planets and the constellations, have all gone the way of the dodo bird. Nobody seems to have lost any sleep over any of them. If you get my drift...