[coliveira]

This is not something that can be "proven" in the sense that all early calendars had drifts that had to be "manually" fixed over time. But we know that the Roman year started officially at the Spring equinox, so that was a calibration point [1]. The first serious attempt to fix the calendar was done by Julio Cesar, but by then he was stuck with the Winter Equinox at Dec 25th (our traditional date for Xmas). When the church introduced the Gregorian calendar their goal was to go back to the dates established for Easter around the year 300 CE, that's why they decided to jump ~10 days in the calendar.

[1] https://en.wikipedia.org/wiki/Roman_calendar#Calendar_of_Rom...

[famous]

OK, the piece really isn't about how pre-Julian calendars defined the seasons or how and when they may have gone off the rails. It's an argument for why [1] meteorological seasons, starting a few weeks before the solstices/equinoxes, is the way to go, and [2] even if you are adamant about strictly astronomical seasons, beginning them on cross-quarter days makes more sense than the system we use now. One system used over 2,000 years ago doesn't figure into it.