Criticism of racism addresses at least two distinct problems attached to racist claims: Ethically, that they propagate hate; epistemologically, that they propagate falsehoods. When a racist says, "We must kill the Jews, because they are greedy liars and poison our nation.", he is both hateful in proposing to kill Jews, and wrong in claiming that they are greedy liars. Note that the epistemological problem is not that no Jews are ever greedy or liars, because they probably are, just like anyone else. The problem is in the generalisation that posits a group of people with common character traits in the first place. The problem is with the concept of race itself. Today, many people only look for the criticized ethics of racism. They believe that we should be nice to people of all races, basically because being nice is generally, well, nice; perhaps they also refuse any claims that would attribute obviously negative characteristics to a certain race. (The latter is in fact somewhat illogical if you do believe that races exist. If they are a group with shared traits, some of them might be undesirable, unless all traits are equally desirable.) But many of us seem to forget that the very concept of race is wrong, even if you don't use it to argue for hate. Others do argue for hate but think they are not racist because they think they aren't using the concept of race. They think that on the one hand, they hate a certain group; and on the other hand, race is a valid description of a different grouping of mankind, which they aren't using as a basis for their hatred. I think many of these people believe that race is in fact a valid biological concept. This is wrong in two ways. Neither is the concept valid, nor is it originally from biology. The word stems from Italian razza in the 16th century, which means dialect or language (the etymology is unclear; it might be connected to radix, root). You belong to a certain race depending on which language you speak. Soon, the concept was widened to encompass cultural or geographical shared qualities, where you could be said to come from different razzas if you came from two different cities, or from the same city, but one of you was e.g., a part of the Jewish community in that city with its distinct cultural and lingual properties, whereas the other was a "gentile". By the late 18th century, 'racist' had become a derogative term. This was because two things had become obvious with regards to the concept of race: That it was used to engender hatred; and that predictions on the basis of race failed unless tautological. When you had a group of people, all of whom spoke Swedish as a mother tongue, you could somewhat clearly predict what words they'd use -- Swedish words. Any prediction beyond that (character, competence, metabolism, etc.) regularly proved to be wrong. It is at this point that the rise of biology as a science seemed to offer a saving grace for the concept: Surely, if we could see that traits could be inherited, then that would give a sound basis for assuming that groups with shared heritage could have shared traits, right? And that thinking is not logically false, as far as plausible general hypotheses go. Descendants of humans do tend to be human. It just doesn't apply to all kinds of traits. Being greedy or a liar, for instance, is not something we can accurately predict by knowing which continent your ancestors hail from. But note that at this point it had long also become clear that we can't predict the same depending on culture, language or location, either. Biological racism is just the fourth and last among these doomed concepts to be proven wrong. Biology's short (in historical terms) love affair with the concept of race is over; the concept is hardly ever used any more, and when it is, it is employed in an emphatically non-essentialist, tautological manner (we observe that all these peas have angular ends; let's call any pea with such ends part of the same race, until we can figure out the actual context of the phenomenon, but remember that this does not suggest other shared traits). Some countries still have laws that speak of races; the US, notably, classifies its citizens by race, which can possibly but only possibly be defended by pointing out that citizens classify themselves, and the classification is used for political measures that are supposed to counteract racism. So the use of the concept is both separate from the original meaning (biological racism assumes that you can't set your race by consciously deciding your affiliation, and the same is true for language, location, and culture) and reactive (it's not there because the law wants to treat races unequally, but as a reaction to unequal treatment that the law is intended to equalise -- ideally, the endgame is abandoning the classification). So we hope that it isn't ethically racist, when used well, but it does carry traces of the epistemological falsehood of the concept of race. So what is a racist? At the very least, you are a racist if you hate a group of people and have created the concept of that group of people in your mind by using one set of observable characteristics, and then deceiving yourself into thinking they also share another set of characteristics that you haven't observed as applying to them specifically. To limit racism to abused biology is to accept as valid one of the false defensive strategies that racism adopted around 1800.