(1) Cause-and-effect exists within society, but as an interconnected web, not as a linear chain. Everything causes everything else.
(2) The concept of cause-and-effect comes from the material world where, at the macro (but not the sub-atomic) level, it is valid.
However, even here, as the chaos theorists have shown, what may be analyzable in principle may not be in practice. (Even our Solar System, once held up as the very model of Newtonian determinism, is chaotic.)
(3) Unlike planets, human beings are conscious. This means that attempts to intervene and change their behavior run into the problems that do not occur with non-conscious entities. For example, if I observe that in a poor neighborhood with a violent crime rate 100% higher than the average, the crime rate goes down if the government builds a Youth Center with free recreational facilities, would it be a wise idea to announce that, for any neighborhood whose violent crime rate exceeds 100%, the government will build a Youth Center with free recreational facilties?
This same concept in insurance is known as "moral hazard" -- once someone is insured against the consequences of certain acts, there is a danger that they will be more tempted to commit those acts.
(4) Because human society is so complex, and because human beings are self-aware, we must be extremely cautious in assigning causes to effects, and vice versa. It is a conceit of human intelligence to believe that a social system can be analyzed like a physical system.
This is why one strain of conservative thought is .... conservative. We are skeptical about schemes for human improvement which make simplistic assumptions about how easy it will be to bring about radical alterations in human behavior. We are doubly skeptical when these schemes stem from abstract propositions about "rights" and "social justice" rather than empirical observations.
An example of incautious attempts to assign causes to outcomes can be found in the two books[ame=http://www.amazon.com/Freakonomics-Revised-Expanded-Economist-Everything/dp/0061234001/ref=pd_bbs_sr_1/103-9655811-7155832?ie=UTF8&s=books&qid=1191046349&sr=8-1] Freakonomics[/ame] and [ame=http://www.amazon.com/Freedomnomics-Market-Works-Half-Baked-Theories/dp/1596985062/ref=pd_bbs_sr_1/103-9655811-7155832?ie=UTF8&s=books&qid=1191046445&sr=1-1] Freedomnomics[/ame]. The authors, one side sort-of liberal and the other conservative, duel over the question: did Roe vs Wade cause a drop in the crime rate two decades later (by causing more babies of crime-prone poor people to be aborted)?
Each adduces various statistics to bolster his case. But the fact is, we don't really know. Society is too complex to assign such simple causes to effects.
Although conservatives should be more receptive to this idea than liberals (the latter, after all, are keen to get in there and change society for the better), in fact, conservatives often fall into the trap of simplistic assignment of cause to effect, via "common sense" reasoning.
While common sense is not to be sneered at, in reality what our common sense tells us is often wrong.
Does getting more welfare money for each new baby cause welfare mothers to have more babies? Maybe, but don't assume it. Better try to do some empirical research on the issue before deciding.
Will the oppressed and downtrodden Iraqi people welcome the liberating American troops with flowers and dancing? Maybe, but don't assume it. Better have a "Plan B" in case they don't. (Well, too late for that one. But the point is valid.)
