Politics is, in part, a teaching activity, and Gov. Hugh Carey of New York, responding to a recent column of mine on capital punishment, has sent me a scholarly article and a letter recommending it as an antidote to my errors. He says two of my erroneous propositions are these:

First, the burden of proof is on those who argue that the fear of death cannot be a significant deterrent to significant crimes. Second, there may be a moral symmetry between capital punishment and certain crimes, such as assassinations. A capital punishment law can express, and may thereby strengthen, public sentiments that should be encouraged, such as outrage about assaults on the social order.

Carey believes the scholarly article shows that capital punishment, far from deterring murder, actually stimulates it. And although capital punishment may be intended to express salutary sentiments, it nevertheless stirs dangerous sentiments in some people: those susceptible to momicidal impulses.

The article is "Deterrents or Brutalization: What is the Effect of Executions?" by William J. Bowers and Glenn L. Pierce of the Center for Applied Research, Northeastern University. They believe they have demonstrated that executions in New York have "contributed to" (a necessarily imprecise formulation) two to three homicides in the month immediately following.

Since 1890, New York has conducted 695 executions, more than any other state. The most that the authors' research (and somewhat similar research concerning other jurisdictions) can demonstrate is a correlation that may or may not represent a causal relation. It is a correlation between executions and "unexpected" short-term increases in homicides.

As the authors acknowledge (and purport to compensate for, statistically), factors such as war complicate attempts to isolate the effect of a single cause, such as an execution. Since 1890, four wars have taken large numbers of young men -- a class that accounts for a disproportionate share of homicides -- out of normal society, exposed them to often brutalizing experiences, and then released them to readjust, often with difficulty, to civilian life. In such periods, it is hard to say what homicide rate is to be "expected."

The authors' analytic model is an ingenious as an algebraic representation of complicated social processes and relationships can be. But the authors also rely on some psychological theories. They say they have documented a "contagion" from executions, a phenomenon explained by theories of psychological "suggestion" and "identification."

The deterrence theory of punishment holds that the potential murderer will identify with the executed criminal, and hence will bedeterred. But the authors say most murders are "acts of passion between angry or frutrated people who know one another," and many murderers feel they have been injured -- betrayed or dishonored -- by the people they then kill. Many murderers may consider themselves victims, and may be partially unbalanced by the grudges they nurse.

Provoked by reports of an execution, a potential murderer may "identify" with the state as executioner, and may identify his victim with the executed criminal. The fact that the execution occurs in a context of due process, from trial through appeals, may be lost on the potential murdered, for whom the only message of an execution is that lethal violence is an appropriate response to those who offend.

Perhaps this theory of "suggestion" and "identification" explains the authors' data. But it also is possible that both sides in the debate about capital punishment have focused excessively on the crime of common murder. Importing heroin is a more serious assault on society than many murders, and often has lethal effects. And it is a purely premeditated crime. The traditional deterrence argument has obvious plausibility in relation to that crime, and the theories of "suggestion" and "identification" seem irrelevant to it.

By citing research, Carey implicitly accepts my point that one's position for or against capital punishment cannot be "categorical" -- meaning immune to evidence. We urgently need evidence about the minds of assassins and terrorists. Perhaps assassination and terrorism are premeditated crimes but are underterrable by any stipulated punishment. But that is not known, and even were it known, it might not be conclusive as an argument against capital punishment.

The feelings Maureen yreagan expressed -- cold rage, clear and controlled -- were worthy not only of a daughter but of any citizen. Perhaps (I will not put it stronger) capital punishment is justified as a clear and controlled means for a nation to express feelings that not only are justified but are indispensable to civilization: feelings such as implacable rage about assaults on the social order.