(Would be interesting to know the details of that lawcase. I investigated some of those incidents in Germany and in most cases they turned out to be quite different from the initial aggregations that I read in public).
https://www.lepoint.fr/societe/eric-zemmour-condamne-pour-pr...
The logical relation between (a) and (b) is the important detail here.
If the majority of drug dealers are either black or arab, this does not logically conclude, that the majority of blacks and arabs are drug dealers!
The only logical reasoning for racial profiling would be, if there was a significantly higher probability to catch a drug dealer if you randomly pick someone from that group.
The math:
Let's assume a population consisting of 20 percent group A and 80 percent group B. 0,1 percent of the population is drug dealers. 60 percent of the drug dealers belong to group A, 40 percent belong to group B. Group A therefore makes the majority of drug dealers.
With the majority of drug dealers in group A and only 20 percent share of the population, there is a six times higher probability that a random pick of group A will be a positive hit. In absolute numbers: the chance to make a positive random hit in group A is 0,3 percent, in group B it is 0,05 percent.
But: the likelihood to make a negative hit in group A is 99,7 percent (99,95 percent in group B), so even with a six times higher probability for a positive hit, the overall change for a positive hit - on a random basis - in both groups is still extremely small.
The small chance to catch a drug dealer on a random pick out of a population (not regarding race) does not qualify for an effective police procedure – to begin with. The small difference in probability of 0,25 percent between the groups does not qualify for racial profiling either. Any other visible attribute of a person that correlates with drug dealing with a higher value than 0,25 percent (clothing, cars, peer groups, haircut, jewelry, behaviour, slang, provenance and and and) is a better qualifier for random picks than racial profiling.
So, back to the case:
- France has good reasons, to forbid racial profiling under its law. It IS discriminatory, because you cannot define 99,7 percent of a group by 0,3 percent of that group.
- (b) does not conlude from (a), as it does not significantly rise the success rate, but at the same time feeds prejudices and harasses innocent people.
- Insisting on (b) clearly shows the will to ignore data and a will to feed prejudices and having innocent people harassed, so government decides to stop this behaviour.
Did they really sentence him for (b), or was he rather obliged not to repeat that statement?
It may be that certain groups are pursued more vigorously, less likely to mount a strong defence (therefore more likely to be convicted), or just more narrowly focused on.
