The most spectacular defeat in last Sunday's French parliamentary election, former cultural affairs minister Francoise Giroud remarked yesterday, was that of the polling organizations.

For months the polls have been forecasting a leftist victory as a foregone conclusion. But in the balloting, the Socialists and Communists fell far short of what had been predicted. The question that the French were asking yesterday is, why?

The newspaper Le Monde, which has never been a big fan of polls spread out all their major results since January in a chart. With impressive monotony, the polls gave the left a 5 to 8 percentage point lead though early March. A single exception, giving the left a lead of only three points, was by the least respected of the major pollsters.

The first poll, in January, gave the left 51 percent and the government parties 45. The last poll gave the left 52 and the incumbents 45. Voters, though, gave the left 44.5 percent to 48.4 for the government parties.

The French are - or were - hooked on polls, which continued to circulate after the ban on publication. One of the top polling organizations gave a top weekly a late poll showing 53.5 percent for the left and 43.5 percent for the government forces.

On the day before the vote, a representative of that Polslter's main competitor gave a Western diplomat the results of his organization's 3-day-old poll giving the left 53 percent.

Did something happen to change the minds of a million voters? The last three days of the campaign were not particularly different from the previous three months. Or were the polls just plain wrong all along?

The Socialists say the last-minute shift was the fault of Communist polemcis against them. But that had been going on since September and the tempo dropped in the last week. The communists blame Socialist recalcitrance in meeting Communist demands. That, too, was going on since September.

Giscardists say the shift resulted from the last-minute appeal by President Valery Giscard d'Estaing. But studies made by - well, by pollsters - have shown that similar speeches by far more dramatic French leaders like Charles de Gaulle had never changed the outcome measurably. The Gaullists say they never believed the polls anyway.

The pollsters are now saying for public consumption that it, is the fault of the ban on publishing election-eve polls and that later polls would have revealed the trend. But too many reputable editors, politicians and diplomats are ready to testify to the exact opposite.

One explanation is that there were a lot of polls being done on the cheap, that while a valid sample is 2,000 persons all over the country, many French publications settled for cut-rate samples of 1,000 persons only in the Paris region. But both kinds of polls showed the same results.

The head of IFOP, the oldest and most respected French polling organization, said it was indeed the fault of the Socialists, because they put on a poor show during their last week of free state television time. So, said Jean-Marc Lech, a million voters that the Socialists should have gotten decided to teach the party a lesson.

If you believe the pollsters, said Pierre Viansson-Ponte of Le Monde, it was the fault of the ban on publishing polls, the politicians, the voters - but certainly not the pollsters themselves.

The pre-election publication ban, he said, is a pointless way to regulate the polls and only encourages rumors when effective regulation should be applied to the techiniques and to the rigor of their presentation in the media.

That would be asking France's media to be far more serious than its politicians. The Interior Ministry, as it does in nearly every election, presented the results in a complicated way that understated the vote of the left.

Said one Socialist Party worker, "What happened to our votes is the big question. It's a mystery. But if we start looking too hard for an explanation inside the party now, there is a big risk that people would start blaming each other. The explanation will have to wait for after the runoffs" on Sunday.

One explanation that keeps recurring is that the temperament and behavior of the French people are far too complex to be recorded by any known scientific methods.