America's 28.6 million blacks -- 12 percent of the population -- still lag far behind whites in every measure of economic and social well-being. Although the majority of blacks are doing markedly better than they were 20 years ago, one-third are below the poverty line. And for a core group of 2 million to 3.5 million chronically poor and alienated blacks, conditions seem to be deteriorating with no improvement in sight.

This group at the bottom, described by many specialists as an "underclass," seems beyond the reach of existing social programs and may be growing. Its plight can be captured in part in a series of stark statistics:

*In 1950, 16 percent of children born to blacks and other minorities were born to unwed mothers; by 1983, 58 percent of black infants were born to unwed mothers, compared to 12.8 percent of white newborns.

*In 1960, about 21 percent of black families with children were headed by a woman with no husband present. By 1985, half of such families were headed by women, compared to 15 percent for whites.

*In December 1955, blacks and other minorities had an unemployment rate of 8.2 percent, compared to 3.7 percent for whites. Last month, black unemployment was 14.9 percent, compared to 5.9 percent for whites. For black teen-agers the rate was 41.6 percent, versus 15.9 percent for white teen-agers.

*In 1960, about three of every four black children under 18 lived with two parents. By 1984, the figure was 41 percent. Among whites under 18, 81 percent lived with two parents.

*In 1983, about 46 percent of persons sent to prison were black. In most cases, the victims were black.

According to many experts, the impact of these statistics is disproportionately concentrated in a small group of blacks who make up the underclass. Concentrated in bleak urban ghettos such as the South Bronx, this group consists not of all the black poor, but of a subgroup that lives primarily off welfare or crime, whose members are increasingly cut off from mainstream American life and unlikely to break out of the poverty cycle.

Available statistics indicate increasing polarization of income levels within the black community. For example, while the underclass has festered, black families that include a married couple living together show steadily improving income levels. Their median income measured in constant 1984 dollars was $18,212 in 1967 and $23,418 in 1984, rising from 68 percent to 78 percent of the figures for white couples.

But the median income of female-headed black families (including those without children) was only $9,380 in 1967 -- and fell to $8,648 in 1984.

William J. Wilson, sociology department chairman at the University of Chicago and a leading authority on the black poor, said in an interview that "there are inner-city groups whose behavior contrasts sharply with middle-class America." He said these groups have high levels of "welfare dependency, violent crime, illegitimacy."

In the past few decades, he said, there has been a tremendous social transformation of the black urban community -- brought about, paradoxically, by the economic successes that many blacks had so painstakingly achieved.

"Once, the black inner city was vertically integrated," he said, explaining that people of various income, education and achievement levels lived there -- doctors, lawyers, teachers, entrepreneurs and religious people, as well as low-income groups. Now, Wilson said, there are fewer role models for blacks in the community. "The black middle class no longer lives in the inner city," he said. "Even the stable working class is moving out.

"Today it is populated almost exclusively by the most disadvantaged -- people outside the mainstream, criminals, families with long-term spells of poverty and welfare dependency. So today the ghetto is different. It is increasingly isolated from mainstream patterns and norms of behavior."

The poor economic situation of black men is seen by Wilson and many others as a primary reason for the high rate of illegitimacy. With many black males having poor prospects for a decent living, the observers say, black girls who become pregnant have little inducement to marry. With the number of available black men further reduced by incarceration, drug addiction and a high incidence of early death, there is also lowered opportunity for women to marry.

Wilson has studied what he calls the "male marriageable pool," looking at the ratio of employed men to all women of their race.

In the 1960s, he found, there were 70 employed black men for each 100 black women between age 20 and 24, and 70 employed white men for 100 white women. By 1980, however, the black figure had dropped to 50 men per 100 women while the figure for whites had gone up to 75. Among blacks age 18 and 19, the figure had dropped to 35 men per 100 women.

"The males are not out there," Wilson said.

Another school of thought holds that a hard-core group living off welfare and crime has been fostered by the government's efforts to wipe out poverty.

Charles Murray, author of "Losing Ground," a controversial critique of U.S. social programs that was published in 1984, said in congressional testimony and in an interview that a new atmosphere was created by social changes of the 1960s and 1970s. Different sanctions for crime and creation or expansion of government support programs, he said, challenged traditional American ethics of hard work and family solidarity.

"All the changes in the incentives pointed in the same direction," Murray said. "It was easier to get along without a job. It was easier for a man to have a baby without being responsible for it, for a woman to have a baby without having a husband. It was easier to get away with crime . . . . Because it was easier to get away with crime, it was easier to support a drug habit. Because it was easier to get along without a job, it was easier to ignore education. Because it was easier to get along without a job, it was easier to walk away from a job and thereby accumulate a record as an unreliable employe. In the end, all these changes in behavior were traps."

Murray's theory is rejected by many economists and sociologists who work on poverty issues. They say the availability of welfare could have had some influence, but not a major one.

David Ellwood, associate professor at the Kennedy School of Government at Harvard, has done a number of studies of the impact of welfare on behavior. "I don't think welfare has done a great deal of damage," he said.

A frequently cited major study by Ellwood tests the notion that the availability of welfare is an incentive for unmarried women to have more children. If that were the case, birthrates to unmarried women would be highest in states offering the highest benefits.

The study found that "the availability of higher benefits does not produce higher birthrates. In Mississippi, where the benefits are astonishingly low, birthrates to unmarried women are about the same as in higher-benefit states."

Charles Willie, a professor of education and urban studies at the Harvard graduate school of education, said, "Welfare doesn't create a class of subnormal people. . . . We are now calling compassion a spur to depravity."

Ellwood pointed out that many of the black poor don't live in the inner city and that others are disabled or elderly. He said that if there is an underclass, it consists of "no more than 20 percent of the black poor," or less than 2 million people.

Wilson and research associate Robert Aponte said the figure could be higher, maybe 3 million to 3.5 million, consisting of perhaps one-fifth of all blacks in the central cities, including those who are persistently poor but otherwise share mainstream American attitudes and are not inclined to commit crimes.

A study by Greg J. Duncan at the University of Michigan Survey Research Center found that most welfare recipients have collected benefits for short periods, not for years. Only one in four children from families heavily dependent on welfare ends up as a chronic welfare recipient after leaving home, Duncan found.

Many students of the underclass contend that the high crime, illegitimacy and welfare dependency rates have roots in a broad pattern of social dislocation.

Sen. Daniel Patrick Moynihan (D-N.Y.), who has written extensively on this subject, said some groups moving into American urban society seemed to have an "everybody up" experience of social mobility, while others, like Irish immigrants after the great famine, split, with some going up and others down, producing a "destroyed lower class."

For the Irish, he said, "Some made it, others became criminals and whores. The Irish cop and the Irish burglar. It took a long time" before that "destroyed" element began to move up again.

The groups that seemed to split the worst, Moynihan said, have often been those whose deprived peasantries were shifted into an urban society -- as happened to American blacks in the past half century.