Who invented score voting? A History

[What would have happened in many previous elections if they'd used score voting? See these polls.]
Sometime between 20 and 200 million years ago
Honeybees start using score voting to decide new nest location each year.
≈700 to 146 BC with interruptions during 220-146; also 146 BC to 396 AD as partially-autonomous Roman province
Although we generally hear in Civics Class about how Athens was the birthplace of democracy... it turns out that the Ancient Spartans used score voting to make important collective decisions... sort of. The Spartans employed these democratic methods both before Athenian democracy started (and also before the Northwest Indian Buddhist period democracies in 600 BC to 200 AD) and after it ended, and thus appear to be both the earliest and the longest lasting substantially-democratic government in world history.
1268-1797 AD (and also democratic during some part of 726-1268 using earlier rules)
In Venice they used score voting with a 3-point scale to elect Dogi (leaders). [Marji Lines: Approval voting and strategic analysis, a Venetian example, Theory & Decision 20,2 (1986) 155-172; Miranda Mowbray, Dieter Gollman: Electing the Doge of Venice: analysis of a 13th Century protocol (pdf) , IEEE Computer Security Symposium (July 2007) Venice Italy.] The Venetians' procedure was complicated and involved eliminations and various stages, but the final stage, which was the one that mattered, was a range3 vote.
Note this is over 500 years of Range voting experience for both Venice and Sparta (each) – each more consecutive years of experience than any other election method has had. Venice and Sparta were apparently the two longest-lasting governments incorporating some democratic elements ever. (Compare: Roman republic lasted only about 460 years before the Roman Emperors took over. Athens democracy only lasted about 200 years. USA so far: only about 220 years. The English might qualify as champions if you date their "democracy" to the Magna Charta in 1215, but I would not consider them as comparably democratic to Venice until the 1700s. Carthage's government involved both democratic and checks-and-balances principles and would be a legitimate contender for the longetivity crown, since it lasted from Carthage's founding in 814 BC(?) until its utter destruction in 146 BC by the Romans, except that Carthage apparently only became substantively democratic during the 400-300 BC century. Because the city and all government and other records were destroyed, we know little about it. For example, what sort of voting system(s) did the Carthaginians use? That seems unknown (based on reading several books about Carthage).
1294-1621 AD
Approval Voting was used for selecting 41 Popes, in a system instituted by Pope (and saint) Celestine V, who then became one of the few Popes (perhaps the only one) ever to resign before Benedict XVI (2005-2013) who resigned a few years after I wrote this just to screw up this web page. While the papacy can hardly be considered a democracy, it has been probably the longest-lasting human institution on the face of the earth, and during these years it was an extremely powerful position. These conclusions summarize my attempt to study all the approval-voting popes.
Poundstone in Gaming the Vote speculates this Pope-system was probably inspired by the Venetian system and cites a paper by J.M. Colomer & Iain McLean: Electing popes, approval balloting and qualified-majority rule (pdf), J.Interdisciplinary History 29,1 (1998) 1-22; reprinted in Politics and Political Change (MIT Press 2001, R.I.Rotberg ed.) 47-68. They in turn cite the "Ordinarium Sanctae Romanae Eclesiae" by Cardinal Jacobi Gaytani. Typical elections had about 40 voters and 20 candidates. Secret-ballot approval voting but with a 2/3 supermajority requirement was employed, with re-voting as many times as necessary (while also introducing new candidates into the race as necessary) until the winner exceeded that threshold. This could take a very long time – Celestine V's election in 1294 famously took 2 years starting in 1292, which presumably inspired his reforms of the election system – and while that was improved by the use of approval rather than vote-for-one plurality voting, the duration still seems to have been a big reason the old system was abandoned in 1622. (Electoral speed was encouraged by such rules as these: if it took longer than 3 days, the cardinals were restricted to 2 simple meals per day; if it took longer than 8, their diet was reduced to bread and water only; and the conclave was usually held in austere quarters with armed guards. One of the new rules after the approval-based system was abandoned was that the kings of France and Spain – and later Austria – were granted veto power over the election of any pope they found displeasing, albeit they only got to exercise one veto per election. That rule operated until about 1910.) Another problem was that during repeated re-votes the totals for different candidates could and sometimes did change by a lot as new information arose and/or as the strategic environment changed. (For example, one Pope initially got zero approvals; and at least three times, pope-candidates who got over 50% approval failed to win – since of course the ultimate winner got still more approval.)
1788-1800 AD
The USA's first four presidential elections were conducted using a variant of approval voting.
1900 AD
The 1900 Olympics (Paris) and 1904 (St. Louis) both used score voting by a jury to choose gymnastics winners.
Apparently the scores at that time were integers; the use of decimals and the 0-10 scale had apparently not yet been introduced. E.g. in 1900 Paris, the "Pointage" for the "APPAREILS ET SERIES" was an integer in the range [0,20] with 6,7,8="mediocre", 15,16,17="tres bien" and 20="parfait"; the gymnast with the highest point-total won. In St. Louis the point system differed from the Paris system. (Some other gymnastics sub-competitions were objectively judged, such as "rope climbing" where it was about time and distance. But they don't do that one anymore.) I do not know how the original 1896 (Athens) Olympics handled the gymnastics competition. The Ancient Greek Olympics (held 794 BC to 394 AD) had no subjectively-judged events. It also excluded female athletes, although there were some female horse-breeder/owner entrants in the chariot races, such as Cynisca (or Cynistra?) of Sparta (who won in 396 and 392 BC) and there was a separate women's Olympics (the "Heraea Games") held less-frequently at the same site.
Early 1950s
John C. Harsanyi (winner of the 1994 Nobel Prize in Economics) points out that Arrow's impossibility theorem does not apply to score voting. He is largely ignored. (The same observation was rediscovered by at least three more people.)
1968
Guy Ottewell invents approval voting and then at least 5 political science professors also independently invent it during the 1970s. (Meanwhile the United Nations, founded in 1945, had been using a system essentially the same as Approval all along to elect their secretary general. This was an excellent – and in the opinion of all highly universally-successful – part of the UN's political design ...unlike the UN's and the European Union's "veto power" rules which were an awesomely bad political design that served as a surefire recipe for paralysis.)
1983
S.J.Brams & P.C.Fishburn publish book "Approval Voting."
1987-1990
Approval voting employed in thousands of elections in the Soviet Union in experiment in democratization started by M.S.Gorbachev.
1995
World Figure Skating Championships decides not to use score voting. Major screwup results.
1999-2000
Warren D. Smith conducts the first Bayesian Regret study comparing different voting systems and including score voting as a contender. Finds score voting superior to all others tried, very robustly to changes in simulation parameters.
2004
Jan Kok discovers how to do score voting with ordinary voting machines designed for plurality voting (no modification or reprogramming required).
2005
Kok & Smith co-found RangeVoting.org. (By 2010 this website has grown to contain over 1000 pages.)
2008
The Boston Tea party holds some elections using approval voting. For example they used approval to nominate Charles Jay for USA president. (Their system has repeated-ballot rounds under certain circumstances, which did not arise here.) Jay received a grand total of 2422 votes in the later general election (only got on ballot in three states). It is depressing that tiny parties mostly regarded as "fringe" or "strange" are intellectually well ahead of the "mainstream" USA parties when it comes to voting systems.
2008
William Poundstone publishes popular-science book Gaming the Vote. (Buy it on Amazon. Read the New York Times's review.)
2010
The Pirate party of Germany (so-called because of their main issues: internet freedom and repeal of copyright laws) holds elections using an approval voting system. (See especially rules 4.3.3.2-4.3.3.5.) For example they elected Jens Seipenbusch chair. In the Sept. 2009 general election this party received 845904 votes. In the 2011 Berlin state election, the Piratenpartei won 8.9% of the votes and 15 out of the 141 seats in in the Abgeordnetenhaus (German state parliament).

Both the Boston Tea and Pirate Party rules allow any voter to vote for any subset of the candidates as well as for "NOTA" (None Of The Above). Most approved candidate wins.
2010-2011
HB240, sponsored by representatives Dan McGuire, Cameron DeJong, Seth Cohn, and Spec Bowers, will switch New Hampshire over to approval voting. But as of March 2011, this bill had failed to win passage. There have also been other approval-voting-enaction bills in other US state legislatures with other sponsors in other years, but up to 2011, none have succeeded.
2012
The Center for Election Science finally succeeds in getting official tax-exempt status from IRS.
2013
HB 2518 (enabling Approval Voting in cities) passed the Arizona house by 31-26. Will it pass their Senate?

Return to main page