Is the United States entering a recession or isn't it? The sages, here and abroad, differ.

In his congressional testimony this week, Alan Greenspan said the risks of recession had been increased by the confrontation in the Persian Gulf, but he still saw no downward tilt to the economy.And he said the Federal Reserve would not ease monetary policy until the administration and Congress reached agreement on a long-term plan for cutting the federal budget deficit.

The International Monetary Fund, on the eve of its and the World Bank's annual meetings, sounded more bullish this week than Greenspan. It said the United States would avoid a recession this year and growth would pick up in 1991.

But, on the bearish side, 58 percent of the 50 economists and forecasting groups surveyed by Blue Chip Economic Indicators predicted that a recession would start this year or in 1991.

Robert J. Eggert, Blue Chip's editor, said that in the past two months its consensus forecast for growth in real gross national product in 1991 had declined to 1.2 percent, from 2.2 percent, one of the sharpest drops in the survey's 13-year history.

What explains the differences among these people who peer into the future? Not economic indicators or other facts and factoids available to all economists, but rather three other factors: semantics, assumptions and policy.

Greenspan's position that it was "too soon to indicate" whether the odds on recession are greater or less than 50-50 reflected both his semantics and policy.

On the semantic side, he shied away from the "conventional" definition of a recession - two consecutive quarters of decline in real GNP - suggesting that a more useful definition was that a recession is a decline so serious that it begins to feed on itself.

If the Blue Chip consensus forecast of 1.2 percent growth in 1991 is about right, and if the Fed essentially shares it, as it apparently does, that forecast could be consistent with a recession, defined as two quarters of negative growth.

Such a recession could start, for instance, in the fourth quarter of this year, and last only through the first quarter of next year - or even through the first half of next year.

But that would imply a recession that "did not feed on itself," followed by a relatively good recovery in the second half of 1991. In Greenspanian semantics, that would be no real recession, so a recession is not yet in sight.

On the policy side, the chairman of the Fed has two concerns: one, to keep pressure on the administration and Congress to deal with the budget deficit; two, not to relax monetary policy too soon, lest the markets regard this as an inflationary response to political pressures.

That in turn might perversely raise interest rates and give the economy a double blow of higher inflation and lower growth. Cutting the budget deficit would reduce public claims on private savings, ease rates and set conditions in which the Fed could relax monetary policy without fear of pushing up inflation.

The major difference between the more cheerful forecast by the IMF and the others lies in its assumption that the world oil price, now around $35 a barrel, will average $26 for the rest of this year and come down to $21 by the fourth quarter of 1991.

This prognosis is based on the assumption that the four million barrels a day lost to the world oil market from Iraq and Kuwait will gradually be replaced by extra output from Saudi Arabia, Venezuela and other countries, as well as the further underlying assumption that there will be no war or greater loss of oil.

How good are these assumptions? They are certainly wishful. It is true that the Organization of Petroleum Exporting Countries has not changed its reference price of $21 a barrel that was in place before the Iraqi invasion of Kuwait, and that an OPEC representative said last week in an interview in Antalya, Turkey, that OPEC still assumes the oil price for the coming year will stay in a range of $21 to $25 a barrel.

But that would represent a steep fall from present prices. Some analysts are making more cautious assumptions. Antonio Costa, special adviser to the president of the European Community, is assuming that oil prices will range from $23 to $27 a barrel during the next 18 months.

That would mean, he said, a worldwide transfer of $160 billion to $170 billion in gross revenues from oil consumers to producers; net, after deducting the flowback of payments for higher imports, the increase would represent a shift of $95 billion to $110 billion from consumers to producers - a shock big enough to give the world a new monetary recycling problem.

Costa's estimate assumes a stalemate, rather than a worsening conflict, in the gulf.

Even on relatively hopeful assumptions, however, the gulf crisis, as long as it lasts, will mean somewhat higher inflation, somewhat slower growth and higher unemployment for the United States and the world economy.

Policy will counteract those effects but is unlikely to eliminate them.