What's a double dip? No one really knows.

FORTUNE -- Here's what we know about double dip recessions: The economy shrinks. Then grows. Then shrinks again. It's essentially a hilly economic recovery (or battle) where the ebb and flow of growth strangely takes a W-shape.

Beyond that, no one can agree on what a so-called double dip is, whether or not we're about to have one, or even if we've ever had one before.

Even the most sophisticated economists can't agree what the term "double-dip recession" technically means. The National Bureau of Economic Research (NBER), which includes a committee of academic economists that officially declares the beginning and end of recessions, does not have a separate designation for the phenomenon. Economics 101 courses don't spend much time studying it, either. According to a Factiva search, the term first started appearing regularly in the media in 1980.

Despite its sketchy historical record, economists, politicos, columnists, and just about everyone with an opinion on the economy is speculating about a possible double dip today. We're heading into a double dip recession! We're going to avoid a double dip for now!

This week, economist Nouriel Roubini tweeted that there was little difference between a US double dip and "growth so anemic (1.5%) that [it] will feel like a recession even if it's not formally one."

Many economists define a recession as two consecutive quarters of negative GDP growth. Can it be possible to have economic expansion - albeit small, as Roubini notes - during the second part of a double dip? NBER, which has yet to formally declare an end to the U.S. recession that formally started in December 2007, believes a double dip would essentially be one continuous recession with a period of growth occurring and then back to a downturn.

The historical record

As if that's not vague enough, the double dip gets even murkier. Follow if you will.

According to the NBER, a double dip has never actually happened. It says the closest the US economy has been to one was in 1980 and 1981. The bureau found that those periods were marked by two separate recessions that occurred close in time.

Not surprisingly, some economists disagree.

Economics professor Sung Won Sohn of California State University, Channel Islands acknowledges that while the definition of double dip isn't always clear, the U.S. economy did undergo a double dip in the early 1980s partly due to spikes in oil prices and the U.S. Federal Reserve raising interest rates before the economy had a chance to fully recover.

Sohn says a double dip occurs when an economy recovers briefly before it begins to fall into a recession again. It's not necessarily one continuous recession with a blip of growth somewhere in between, as NBER says. In fact, he believes a double dip also happened around the time of the Great Depression. Just as the U.S. economy was recovering in 1936 and 1937, it slid back into a recession as interest rates rose. Sohn believes there is a 30% chance today's economy will slip into a double dip.

Clearly, the game of declaring a double dip is fuzzier than calling a recession, which has its own quirks. The NBER considers several economic factors when it identifies a recession, but most (not all) downturns have included two or more consecutive quarters of declining real GDP.

Many economists agree the U.S. came out of a recession in June or July of last year when the economy returned to growth, but NBER has yet to corroborate that. Nariman Behravesh, chief economist at IHS Global Insight, says the dating of recessions is a "judgment call."

He says a double dip recession today, whereby GDP falls back to negative levels, is unlikely. He gives it no more than a 20% chance.

Nevertheless, growth will likely cool next year amid an already weak economy. Behravesh predicts growth will likely slow to 3% to 3.5% this year, and then to about 2.5% to 3% in 2011 as government stimulus spending wanes and the U.S. dollar strengthens amid Europe's debt problems.

The slip in GDP may not put us in double-dip mode. But then again, what really does?