Social progress is possible because social learning, although spasmodic, is cumulative. The record seven consecutive years of reducing the national crime rate is, in part, the result of learning about policing.

But how big a part? In two cities on opposite coasts, here and in New York, opposite, or at least sharply contrasting, approaches to policing have caused -- or have they? -- comparable reductions of crime.

Sorting out social causation is problematic, and the recent consensus has been that policing cannot be the primary cause of dramatic reductions of crime. Policing, it is said, primarily displaces crime, causing it to relocate, and catches criminals after the fact. The ebbing of the crime wave is partly due to economic and demographic factors -- high employment and a decline in the size of the cohort of males 18 to 24.

Intellectuals -- professional sorters-out of social causation -- are particularly reluctant to credit increased rates of incarceration. Remember the (unintentionally) hilarious New York Times headline "Crime Keeps on Falling, but Prisons Keep on Filling."

New York, with a police officer for every 198 citizens, is comparable to Chicago (one per 202) and Philadelphia (one per 192). NYPD's stunning success (homicides down 70 percent and major felonies down 46 percent from 1993 to 1998) has coincided with a change from reactive (the "911 paradigm") to proactive policing, based on the "broken window" theory.

It holds that disorder is a contagion: If a broken window remains unrepaired, soon there will be many broken windows. Such things as graffiti, aggressive panhandling, noisy boom boxes create a menacing sense of anarchy, driving law-abiding citizens indoors and ceding the streets to predators. Hence NYPD's "zero tolerance" of "quality of life" offenses. Aggressiveness has permeated NYPD operations.

San Diego, too, has had remarkable success reducing crime: nine consecutive years of decline, a three-fourths decline (better than New York's) in homicides since 1991. And San Diego has its own approach to policing. It is one of the most lightly policed (the size of its force, relative to population) of America's major cities, but Police Chief David Bejarano is not feeling overmatched by the bad guys.

In 20 years with SDPD, he has been a patrol officer and a member of a SWAT team, has dealt with violence along the border with Mexico, the drugs that come across that border, and the drugs made here (San Diego County is the "source country" for America's methamphetamine, and one-third of all the people arrested in San Diego are under its influence). And there are 5,000 documented members of at least 75 gangs in this, the nation's sixth-largest city.

Bejarano's officers (about 2,000, compared with New York's more than 37,000) are supplemented by 1,100 volunteers who have donated 100,000 hours of service in the past six months. RSVP (Retired Senior Volunteer Patrol) uses retired people -- they have uniforms and patrol cars, but no weapons or arrest powers -- to check houses of vacationers, write parking tickets, identify abandoned vehicles ("broken windows") and multiply the number of eyes looking for suspicious behavior.

Soon there will be 32 major traffic intersections where cameras photograph the cars and license plates of drivers who run red lights. This will enable SDPD to economize manpower. Bejarano remembers that 20 years ago he would start a shift by loading into his patrol car a duffel bag of forms and books -- the criminal code, the vehicle code. Now all that is in every officer's laptop.

How much credit can policing take for reducing crime? John DiIulio of the University of Pennsylvania explains two theses.

One is the "re-administering" thesis about crime reduction -- that reduction results from changed policies, such as those about policing and sentencing. The other is the "re-norming" thesis -- that America's "social capital" is being replenished by "the rebirth of certain social norms," such as the reassertion of community considerations over individual self-interest, and moral standards over moral relativism.

The re-norming thesis has interesting historical evidence. James Q. Wilson, DiIulio's mentor, notes that by the beginning of this century the growing Sunday school movement included 60 percent of young people, so a majority of the young were receiving an hour or two of religious instruction every weekend. (And YMCAs and YWCAs burgeoned then, too.) That this coincided with a steep decline in crime is surely more a matter of causation than coincidence.

These academic debates are of urgent practical interest. Although the crime rate is almost down to the 1974 level, it remains 2 1/2 times the 1960 level, proof that progress, like social learning, is spasmodic, and that much re-norming remains to be done.