It said in the papers Sunday that the Electoral College was about to elect "a candidate who did not receive the most votes from the American people for the first time in 112 years."
Yesterday, as the electors were gathering in the state capitals to elect George W. Bush, who did not by any count get more popular votes than Al Gore, the papers again reported that 1888 was "the last time it happened." Over the past five weeks, newspapers, magazines, radio and television reported that bit of conventional wisdom at least 100 times.
Like much conventional wisdom, it's wrong.
The last time it happened was in 1960, when a significant popular vote in Alabama and Mississippi went to unpledged Democratic electors.
The South in 1960 was still solidly Democratic, the heritage of the Civil War. Most voters in the 11 states of the old Confederacy had heard hand-me-down family stories from the war or the reconstruction era; some had personally known a Civil War veteran. They could not abide voting Republican. They blamed Abraham Lincoln's Republican Party for the war.
In 1952 and 1956, Republican presidential nominee Dwight D. Eisenhower got 35 percent and 43 percent of the votes in Alabama in elections in which he won in landslides nationally, with 55 percent and 57 percent of the vote. In Mississippi, Eisenhower got 40 percent and 42 percent of the vote in those two elections.
When conservative Democratic voters in the Deep South wanted to go against the national party's presidential candidate, they preferred to follow a local hero rather than the Republican alternative.
For example, in 1948, outraged with the Democratic Party's civil rights initiatives, Democrats in several states formed the States' Rights Party, also known as the Dixiecrats. South Carolina's Strom Thurmond (the same) was its presidential nominee.
His slate of electors got 87 percent of the vote in Mississippi to the regular Democratic slate's 10 percent and Republicans' 2 percent. And in Alabama, Thurmond's electors got 80 percent of the vote. The regular Democrats didn't even have electors.
By 1960, when the civil rights revolution had high visibility, the Democrats nominated Sen. John F. Kennedy, who was seen by many white Southerners as more pro-civil rights than the Republican presidential nominee, Vice President Richard M. Nixon.
Mississippi Democrats were cool enough toward the national party that in addition to a regular slate of Democratic electors, they put on the state ballot a slate of officially unpledged Democratic electors.
In Alabama, the 11 Democratic elector candidates ran not as a slate but as individuals. Six ran as unpledged, and five pledged to support Kennedy if they won. Alabama voters could vote for one or all in each group, or for none.
In Mississippi, the unpledged electors beat the pledged ones by about 8,000 votes out of just more than 200,000 Democratic votes cast (and Republican electors got 75,000 votes).
In seemingly every vote count chart published in 1960 or since (including those offered by the Associated Press, the clerk of the House of Representatives and Congressional Quarterly, which uses data provided by the Inter-University Consortium for Political Science Research at the University of Michigan and others), the total national popular vote for president included for Kennedy the 108,362 votes cast for the pledged Democratic electors in Mississippi, but not the 116,303 cast for the unpledged ones.
That is the right decision, obvious and logical.
But in Alabama, the popular vote was assigned in a different way by news organizations and, eventually, by the academics who archive the presidential election votes. They counted the number of popular votes cast for the leading elector candidate pledged to Kennedy and put that number in the Kennedy total.
That would be wrong even if that elector candidate had been the leading overall Democratic vote-getter. But an unpledged elector ran ahead of the loyalist, 324,050 to 318,303. (In the authoritative "History of American Presidential Elections," edited by Arthur Schlesinger Jr. and others, the chart accompanying a narrative of the election written by Theodore Sorenson wrongly gives Kennedy the 324,050.)
Surely, the obvious and logical way to assign popular votes in such an unusual election would be to give Kennedy not the top vote of either set of Democratic electors. The Alabama Democrats who voted for six unpledged electors were not for Kennedy, and the voting public knew it.
Those electors, like Mississippi's unpledged slate, voted for Virginia Sen. Harry Byrd when the Electoral College met that December.
Kennedy's popular vote total should include only 5/11ths of the Democratic vote. Depending on how you determine the overall total, he would receive not the 318,303 or the 324,050 but either 147,255 or 144,680.
The final national totals showed Kennedy beat Nixon by 34,221,344 to 34,106,671. That's a margin of 114,673. Properly assigning the Alabama popular vote, Nixon beat Kennedy in the national popular vote by 50,589 or by 47,974.
Kennedy went on to become one of the most popular presidents in the history of public opinion polling. That doesn't mean that George W. Bush will, only that it can be done, that he can.
That the South has gotten over its Civil War heritage of not wanting to vote for Republicans was certainly obvious in this year's presidential voting.
In the 11 states of the old Confederacy, George W. Bush beat Al Gore in the popular vote by 56 percent to 44 percent, a reversal of the Eisenhower vote of the 1950s. The change has been even sharper when the wide disenfranchisement of blacks in the 1950s is taken into account.
In the Electoral College, Bush beat Gore in those 11 states 147-0.
In the states that weren't part of the Confederacy, Gore beat Bush 52 (and a fraction) percent to 47 percent, and 267 electoral votes to 124. Two landslides.