Wikipedia:Reference desk/Archives/Humanities/2022 September 3

= September 3 =

Where did Pitt marry Hester?
Our articles Hester Pitt, Countess of Chatham and William Pitt, 1st Earl of Chatham say they were married on 16 November 1754 at her house on Argyle Street, London. This is referenced to. This seems unlikely to me, as my understanding of the Clandestine Marriages Act 1753 which came into force on 25 March 1754 is that marriages had to take place in a church or chapel, not a private home. Can anyone clear this up? Thanks, DuncanHill (talk) 15:01, 3 September 2022 (UTC)
 * I think the restrictions to churches and chapels only applies to marriages performed after the publication of banns. The language of the act is a bit complicated, but it seems to allow marriages elsewhere if married by license, and it certainly allows such a marriage by Special Licence ("nothing herein [...] shall be construed [...] to deprive the Archbishop of Canterbury and his Successors, and his and their proper Officers, of the Right [...] of granting Special Licences to marry at any convenient Time or Place") . --Stephan Schulz (talk) 15:27, 3 September 2022 (UTC)
 * Ah, good call, and well read. I find in Williams, Basil. The Life of William Pitt, Earl of Chatham (2 vols, 1915) vol 1 online that "On Saturday November 16, 1754, two days after the opening of the session, William Pitt and Lady Hester Grenville were married in her lodgings in Argyle Street. Dr. Ayscough, brother- in-law of George Lyttelton, and an old friend of both families, married them under special licence. A few friends only were present at the ceremony..." I shall use that to clarify our articles. DuncanHill (talk) 15:38, 3 September 2022 (UTC)

Why exactly is the Condorcet winner criterion incompatible with later-no-harm?
The relevant articles about those voting system criteria say this is the case, and articles around the Internet agree, but I have never found an actual explanation for this. Where is the contradiction between the two? ± Lenoxus (" *** ") 23:13, 3 September 2022 (UTC)


 * There is a proof in this paper: https://www.rangevoting.org/Woodall97.pdf, assuming the Axiom of Discrimination. The relevant part of the proof is right at the page break between page 86 and 87. Here is a vague idea: it's possible to have an election between candidates A, B, C, which is almost exactly tied, but in which A just barely wins, and also there is no Condorcet winner. In such a case, if you had lots of voters who voted simply: A (no second or third place votes), it's possible that changing their votes to ABC will make B into the Condorcet winner.
 * Here is the example from the paper:
 * Consider these ballots, where n:XYZ means n people ranked X, then Y, then Z (X is ranked on top).
 * 3:A
 * 3:B
 * 3:C
 * 2:ACB
 * 2:BAC
 * 2:CBA
 * Since these ballots are symmetrical with respect to A,B, and C, the result must be a 3-way tie. Then by the Axiom of Discrimination, there is an arbitrarily close set of preferences for which A is the winner. Maybe something like:
 * 301:A
 * 300:B
 * 300:C
 * 200:ACB
 * 200:BAC
 * 200:CBA
 * Now you can check that if the A votes all change to ABC, then B becomes the Condorcet winner. In this case our method cannot satisfy both later-no-harm (in which A must win) and Condorcet (in which B must win). Staecker (talk) 11:57, 4 September 2022 (UTC)