Is the defense spending increase that President Reagan wants for next year 3 percent, as the administration contends? Or is it 8 percent, as congressional budget experts argue? Or is it 12 percent, as a literal reading of the numbers would indicate?

The answer is yes -- or yes but -- to all three questions, a paradox that helps explain why arguments over the federal budget are often so arcane to spectators -- and why the results are often so frustrating to the players themselves.

At the heart of the dispute is something called a budget baseline, which, when used in connection with a particular spending program, shows what it would cost to carry out existing policy in a new year. When services are expanded or cut, it is used as the benchmark for measuring the change.

The problem is that the administration and Congress can set their baselines anywhere they want, within certain limits of credibility. Depending on where the baseline is set, a cut can look like an increase, or vice versa.

This year's defense numbers illustrate the point:

*The administration bases its claim for a 3 percent increase on the congressional budget resolution passed last summer before cuts were made in the appropriations process and in the first cutback under the Gramm-Rudman-Hollings budget-balancing law.

*The 8 percent calculation is based on actual spending authority for the military, as adjusted for inflation.

*The 12 percent figure represents the increase over existing spending authority without an inflation adjustment.

Translating this into everyday life, your personal family baseline could be exactly what you spent last year. Or it could be what you spent plus inflation. Or, it could be what you wish you had spent plus inflation.

It is tempting -- and perhaps accurate in the abstract -- to dismiss the baseline numbers game as the budgetary equivalent of arguing over how many angels can dance on the head of a pin. In the ideal world, if $280 billion, or $300 billion or $320 billion is right for defense, what difference does it make how big the cut or increase is?

But politics is influenced by symbolism at least as much as reality, and so baseline juggling is a critical early bargaining maneuver for both the White House and Congress. "It the baseline has very important connotations, even if they are symbolic," says Congressional Budget Office Director Rudolph Penner.

By staking out a position early and bringing Senate Republicans along with it, the White House won last year's baseline skirmish. But Congress moved early on the issue this year and appears to have the upper hand for this year's battle.

Last year, the administration insisted that defense cuts be calculated from a baseline pegged to a "Rose Garden" agreement it reached earlier with Senate Republicans that called for continued growth of the military budget. Democrats argued in vain that the baseline should reflect a continuation of official policy rather than a presidential wish-list.

Congress wound up cutting way back on Reagan's defense spending request, but some argue that use of the higher "Rose Garden" baseline for defense helped prevent even deeper cuts.

"It helped, no doubt about it," said House Armed Services Committee Chairman Les Aspin (D-Wis.). "It had a lot of effect last year," said Rep. Dick Cheney (R-Wyo.), chairman of the Republican Policy Committee.

It also enabled Congress, with a proud parent-like wink from the White House, to claim that it made more than $50 billion in deficit cuts even though about one-third of the total was more attributable to baseline massaging than to actual cuts from existing program levels.

So there was less pressure for cuts in spending as a whole, including domestic expenditures, according to House Minority Whip Trent Lott (R-Miss.). "Defense got more, but so did domestic spending ," he said.

This year, the CBO, which tends to take a purist's view on baselines, decided to go back to the original text and say the baseline should represent what it costs to continue current policy as reflected in actual appropriations, adjusted for inflation.

The administration backed off the wish-list concept but stuck with the budget resolution as its interpretation of current policy. Hence the difference between 3 percent and 8 percent.

The difference, although largely an accounting one, will undoubtedly help frame the budget debate because of the prevailing view in Congress that defense must bear a greater share of the budget-cutting burden.

An 8 percent defense spending increase after accounting for inflation would be a lot harder to sell than a 3 percent after-inflation increase when many other programs are barely keeping pace with the cost of living. Some lawmakers say even a 3 percent increase will be hard to digest.

But the Gramm-Rudman-Hollings law changes the equation in such a fundamental way that baseline arguments -- for all the energy that will no doubt be devoted to them -- may well turn out to be irrelevant at the end of the process, according to some lawmakers.

The law says the deficit must be $144 billion for fiscal 1987, regardless of how big a spending cut it takes to get there. And, even if the Supreme Court upholds a lower court decision last week that knocked out its automatic enforcement provisions, Congress is required under the law to put the deficit reductions into law. In any case, the focus will be more on actual deficits than on the size of the cutbacks.

"This year it's going to be harder to play the game," said Aspin. "This year the numbers are real."