Talk:French Republican calendar/Archive 2

The article currently states: "The calendar was abolished in the year XIV (1805)." But it does not give a clear day of when another calendar took its place. It's clear that it went in use on 1792-09-22 (Gregorian) as that equals the first day of the first year. This site: 1789-1815 calendars - table de concordance implies that it was used until a day in the 3rd month (frimaire) of the year XIV. However, due to the lack of links beyond year IX, we can't see the exact date. Is there any source providing an official day when the switch happen?
Liggliluff (talk) 07:35, 14 August 2020 (UTC)

The artIce states: "As the

epoch

of 22 September 1792."

It then states: "Napoleon finally abolished the republican calendar with effect from 1 January 1806 (the day after 10 Nivôse Year XIV)…"

--Nike (talk) 05:00, 15 August 2020 (UTC) (Octidi 28 Thermidor an CCXXVIII à 2 heures 6 minutes décimales t.m.P.)

As per this page the decree in question is dated 22 Fructidor XIII, or 9 September 1805, to take effect 1 January 1806. Arcorann (talk) 15:00, 9 September 2020 (UTC)

I note from French_Republican_calendar#History that "It was, however, used again briefly in the Journal officiel for some dates during a short period of the Paris Commune, 6–23 May 1871 (16 Floréal–3 Prairial Year LXXIX)." I assume that the Communards must have made a decision as to which of the above systems they would use in order to decide that 6 May 1871 was equivalent to 16 Floréal Year LXXIX. Would it be worth mentioning this in the article?

Sorry - I note that this was discussed at Talk:French_Republican_calendar/Archive_1#Use_in_1871 and Talk:French_Republican_calendar/Archive_1#The_leap_year_rule and according to the latter "the dates work out the same that year". Alekksandr (talk) 21:36, 25 December 2020 (UTC)

I get that deciding when leap years occur in the “real” French Republican calendar is not easy.

As I understand the history:

The decree of Saturday 5 October 1793 establishing the French Republican calender contradicts itself on the issue of the occurrence of leap days.

• The first articles describe how the first day of the year is the day on which the autumnal equinox falls at the Paris Observatory.
• But articles X and XVI describe a leap day, called Day of the Revolution, every 4 years. A period of 4 years ending in a leap day is called a Franciade.

A controversy ensued. Yet a general consensus formed at the time, with the equinox rule prevailing over the a-leap-day-once-every-4-years rule.

But this consensus led to another issue. Mr. Jean-Baptiste Joseph Delambre, an astronomer, pointed out after the introduction of the French Republican calendar that according to his calculations the equinox of Monday 23 September 1935 was to take place about 20 seconds before midnight, with an uncertainty of several minutes. So it wasn’t yet clear to him, an astronomer, whether year CXLIV would start on that day or the next. (It turned out that autumnal equinox took place on Monday night 23 September 1935 at 23:48 in Paris, local time. So Mr. Delambre was about 11 minutes off, but his calculations had the day right nonetheless.) Clearly, an arithmetical rule would be preferable for the calendar to be more predictable.

Hence Mr. Charles-Gilbert Romme, the head of the commission that introduced the French Revolutionary calendar in the first place, proposed already in year III a new rule: French Republican years would be leap years if they are divisible by 4, except not if they are divisible by 100, except too if they are divisible by 400, except not if they are divisible by 4000. According to this proposal, the first leap year would be year IV. His project of decree was discussed in committee on 19 Floréal III (Friday 8 May 1795) (less than 5 months before the end of year III which would change from leap year into common year, if the proposal were enacted).

But Mr. Romme supported the demands the Montagnard uprise of 1 Prairial III (Wednesday 20 May 1795). The revolt failed, and Mr. Romme was among those sentenced to death. Before the sentence could be executed, he committed suicide on 29 Prairial III (Wednesday 17 June 1795). His project of decree amending the leap year rule died with him: it never got enacted into law.

Without the enactment of Mr. Romme's project of decree, year III was the first leap year, because of the equinox rule. With some creative thinking, this decision could still be explained in accordance to the once-every-4-years rule: the first Franciade simply started one year before the epoch. Thereafter, years VII and XI were leap years: neatly and orderly 4-year Franciades apart.

The Bureau des Longitudes held on to the consensus that the equinox rule prevails over the every-4-years rule, and published official guidance that years XIX and XXIII would be common years, and years XX and XXIV would be sextile years.

But on 22 Fructidor XIII (Monday 9 September 1805), emperor Napoleon Bonaparte sings the senatory consultary that abroges the Republican calendar on 11 Nivôse XIV (Wednesday 1 January 1806), and restores the Gregorian calendar that day.

In Paris, the French Republican calendar was briefly reinstated for a short period in 1871 (year LXXIX).

So, how to decide which years are leap years from XIV onwards?

• Since the equinox rule was in force immediately prior to the abrogation of the French Republican calendar, this rule is the most obvious.
• Or: one could follow Mr. Romme’s rule, even though it wasn’t followed during the time that the French Republican calendar was in actual use (at least from the Day of the Revolution in year III onwards).
• Or: one could follow the late-19th-century proposal of Mr. Johann Heinrich von Mädler, another astronomer, in which years from year XX onwards are leap years if they are divisible by 4 but not by 128. Prior to year XX, years III, VII, XI and XV are leap years (and year XIX is a common year).
• Several concordance tables did not agree with the equinox rule prevalence over the once-every-4-years rule, and predicted years XIX, XXIII, XXVII, etc. to be leap years. Especially lawyers and notaries with a preference for ease and predictability used such concordance in contracts, usually mentioning explicitly that future dates in their documents are to interpreted according to the once-every-4-years rule, even if this does not agree with the common or legal calendar interpretation in that future time.
• Added 7 November: I came across an approximation to the equinox rule by Jean-Baptiste Joseph Delambre with vernal equinoxes that occur every 365,2420463 days exactly. — Preceding unsigned comment added by Adhemar (talk • contribs) 19:40, 7 November 2020 (UTC)

The article describes another possibility: the “Continuous” rule. What this rule comes down to is: the leap years according to the Continuous rule are the years prior to the leap years according to Mr. Romme’s rule.

I can’t see where that rule comes from. I find absolutely no reference to such a rule in the cited sources (either Renouard (1822) or the archived project of decree of Romme at prairial.free.fr). — Adhemar (talk) 22:53, 5 October 2020 (UTC)

As one can see in the references, the so-called “Continuous” rule was followed in some concordances, and by various modern-day users, which is what you call the "once-every-4-year" rule, even though Romme also added a leapday once every 4 years. Unlike Romme, this rule is continuous with historical leapyears. --Nike (talk) 02:52, 8 October 2020 (UTC) ( Septidi 17 Vendémiaire an CCXXIX à 1 heure 25 minutes décimales t.m.P.)

Thank you, Nike. I appreciate your reply. I’m afraid I did not express myself clearly enough. I do not dispute that there are concordances such as this one (upto year XXV) and this one (upto year XLII) printed before or shortly after the abrogation of the French Republican calendar that followed the continued “once-every-4-year” rule, with leap years in years where (French Republican year number) mod 4 = 3. But using that rule, year CCXXIX would start on Thursday 24 September 2020, not on Tuesday 22 September 2020 as the table in the article states. The difference of two days is explained by the exceptions for years XCIX and CXCIX. Under Romme’s rules, years C and CC would not have been leap years. What I can’t find is a reference to a concordance that follows or a document that describes the “Continuous” rule with the every 4-but-not-100-but-too-400-but-not-4000 exception — or some other form of “most century years” exception, as it is described and applied in the article. (Other than the ones that describe Mr. Charles-Gilbert Romme’s project of decree, which provided leap years in the most years with year numbers divisible by 4, rather than the years preceeding them, that is). — Adhemar (talk) 16:31, 8 October 2020 (UTC)

The rule you are referring to is not a rule de jure, but de facto. It is, in point of fact, the rule most often used by modern users of this calendar. I am not aware of any concordances or annuaires for years after IC, but this method has been used by software applications since at least the early third century. You might try searching software manuals. I only know of one example of the 128-year method. --Nike (talk) 07:28, 6 November 2020 (UTC) (Sextidi 16 Brumaire an CCXXIX à 3 heures 17 minutes décimales t.m.P.)

Thank you. I must use different softwares than you do, because the different softwares I know do not support this rule — nor the 128-year von Mädler rule, for that matter. The only ones I have come across to are: (1) the astronomical rule; (2) the Romme rule; (3) the every-4-year rule without century year exception (which is only exactly as accurate as the Julian calendar); and (4) approximations for (1) such as the one by Jean-Baptiste Joseph Delambre. Formulas for only (1) and (2) are included in Calendrical Calculations by Edward M. Reingold and Nachum Deshowitz, which is one of the standard works for calendar studies.

But thank you for confirming that the “Continuous” rule with the 4-but-not-100-but-too-400-but-not-4000 exception has no basis in historical decisions or documents, but something that somehow only found its way in modern software (that I do not know of). After all, “since the early third century” is since the mid-1990s or even the 21st century, Gregorian. If this is truly a rule used by modern users of this calendar, can references to modern software be provided? — Adhemar (talk) 19:31, 7 November 2020 (UTC)

I would like to see an example of description of rule (3), without century exceptions. Software I have seen uses the "continuous" method with Romme's modifications. I have not seen any historical sources which describe how future century years are to be dealt with in this scheme. I have also not seen any modern examples, which would differ from "continuous" dates by two days. Since the topic is converting contemporary dates, methods which are nowhere used now are irrelevant. Further, "continuous" dates are the only ones I have observed which are consistent with historical dates continuing after the Revolution, which were all in the First Century. If you have historical sources which say otherwise, please let me know, but the "continuous" method seems be one in use today.

BTW, the term "continuous" is original to this article, and may not be the best one. Perhaps "consistent" would be better, since it is consistent with historical dates given both before and after 1805. --Nike (talk) 05:07, 9 November 2020 (UTC) (Nonidi 19 Brumaire an CCXXIX à 1 heure 86 minutes décimales t.m.P.)

An example of software using this method may be found at Brumaire. It has been online over 20 years, I think. From the Brumaire manual:

Le logiciel respecte parfaitement la période historique (an 2 à an 14, pour les généalogistes), et applique le projet de Romme décalé d'un an, sans plus tenir compte de l'équinoxe vrai, ce qui provoque d'éventuelles différences d'un jour à partir de l'an 19, pour certaines années... Les années bissextiles ne sont pas les années multiples de 4 (comme prévu dans le projet de Romme), puisque dans la période historique, c'étaient les ans 3, 7 et 11. Les années bissextiles correspondent donc aux années précédant un multiple de 4.

The method is also used in their iOS app. --Nike (talk) 02:17, 10 November 2020 (UTC) (Decadi 20 Brumaire an CCXXIX à 1 heure 1 minute décimale t.m.P.)

If you are talking about the intercalary rules, the only valid rules for this calendar was the astronomical rule (The year starts on the autumn equinox.). In 1795, Romme was finally convinced to change to an intecalary rule. Delambre was tasked with presenting options, and in the end Romme chose to use the Gregorian rules with an extension where years divisible by 4,000 would be common years. He wrote up report and a draft decree for the convention to approve, but before any action was taken he was caught up in an uprising and got sentenced to death. The report was scheduled to go to the convention the day he was sentenced to death and took his own life with a smuggled knife rather than face the guillotine. The matter was dropped. Delambre had computed the time of autumn equinoxes for 400 years, so there was really no need. And, there doesn't ever seem to have been any desire to re-examine the issue.

Delambre was the person set to come up with all the intercalation options. His own preference for an idea was originally devised by Barbara Oriani in 1785. In that scheme, one divided century years by 900, and if the remainder was either 0 or 400, it was a leap year, otherwise century years would be common years. The "128 year" rule mentioned in the Wiki article was his second choice choice, but he didn't like it. It was too difficult for the layman to do the math by hand. He mentioned a possible extension where years divisible by 28800 would be a leap year to improve accuracy, but that made it just that much more difficult. The "Persian" intercalation of 8 leap years in 33 years was mentioned but not presented as an actual option. Romme apparently didn't like any of these ideas, so Delambre suggested using the Gregorian rules but extending them so that every 3600 years leap year would be skipped. Romme like this a lot better, but he and Delambre made the math easier by changing the 3600 years to 4000 years. It cost some accuracy though.

Note: References Are Jean Baptiste Joseph Delambre - Astronomie théorique et pratique, Volume 3, Michel Froeschlé - À propos du calendrier républicain : Romme et l'astronomie, James Guillaume - Procès-verbaux du Comité d'instruction publique de la Convention, Volume 6, and the French Wikipedia article (https://fr.wikipedia.org/wiki/Calendrier_r%C3%A9publicain#Les_ann%C3%A9es_sextiles) for the chart (It is masked and must be expanded to show all predicted Gregorian dates for the equinox.) Terr1959 (talk) 19:38, 26 May 2022 (UTC)

Oh, the so called Romme - 1 scheme is purely a modern idea to reconcile the concordances written assuming leap year was every 4 years, with the first at year 3. Terr1959 (talk) 19:44, 26 May 2022 (UTC)

And if you want to duplicate Delambre's calculations, start with year 1 equinox at 09:19:24 (24 hour time) and add 5h 48m 32s to the next year to get that equinox. Do this for each successive year, subtracting 24 hours if it reaches or exceeds 24:00:00. The year BEFORE the one you have to subtract 24 hours from the time is the sextile year. Delambre would have done it by hand in hours minutes and seconds, but with calculators it is probably easier to convert the units to Seconds. 5h 48m 32s = 20912s; 9h 19m 24s =33564s; 24h = 86400s

Note: 365d 5h 48m 32s is the value determined for the autumnal equinox year at that time, and Delambre was well aware that the equinox and solstice years were all different from each other and the tropical year. I have not been able to retrace my steps and find the document that gave me that particular figure. It might just be buried in his Astronomie Théorique et Pratique someplace. Using these figures you will come up with the same dates and sextiles as Delambre (see chart in the French Wikipedia article). If you use Lalande's tropical year (365d 5h 48m 48s) or Romme's (365d 5h 48m 49s) or other "tropical years", it will start to deviate by year 15. Per Delambre's math, Year 15 is a sextile. Using the tropical year Year 16 is the sextile. Terr1959 (talk) 19:32, 27 May 2022 (UTC)

Méthodes pour trouver les sextiles du calendrier français Bureau des longitudes;

Joseph Delambre (1797), Dupont (ed.), Connaissance des temps à l'usage des astronomes et des navigateurs pour l'année sextile VIIe de la République, Paris, pp. 318–347 [1]

I removed the references to ahistorical calendars. Since the Julian and Gregorian calendars are considered to be two distinct calendars, despite differing mainly in the calculation of leap years and the beginning of the year, the two versions of the Republican Calendar can be considered as two distinct calendars. I have encountered both in current use. Some call the revised calendar after Romme, but the original calendar could just as well be attributed to Romme, and it was Delambre who first proposed the revision. -- Décade III Quartidi Nivôse an CCXXXI à 9 h. 38 m. decimal t.m.P. Nike (talk) 22:27, 14 January 2023 (UTC)

“It turned out that autumnal equinox took place on Monday night 23 September 1935 at 23:48 in Paris, local time. So Mr. Delambre was about 11 minutes off, but his calculations had the day right nonetheless.”

I think you forgot to take into account the equation of time. The law clearly states true time be used, as was usual back then, not mean time. At the equinox, that is about 7.5 minutes. Delambre gave the decimal time as 0.9972390, which was 23:56:02 Paris true time. 23:48 + 7.5 minutes is between 23:55 and 23:56. —Nike (talk) (Décade III Quintidi 25 Nivôse an CCXXXI à 6 h. 37 m. decimal t.m.P.)

As my finger slipped for my edit summary. Here it is. There is a discussion about a an almanac but that was not the reason for decimalising the calendar and the article doesn’t say it was. I think the reasons given when I deleted the text we’re sufficient. Anyone wanting to revert me again must show sources saying that the change was partially for religious reasons and add a section to the body of the article to justify it being in the lead. Doug Weller talk 20:22, 13 February 2023 (UTC)

Please put your replies in the main thread above (which fully, even extravagently, answers this point) Scarabocchio (talk) 16:12, 14 February 2023 (UTC)