Composition and Division

A recent discussion drew my attention to a rhetorical turn that is becoming common in our political discourse.  It is worth a bit of reflection because it has the consequence of allowing us to talk past each other about a great many conflicted issues.  A few examples will make it easier to characterize this rhetorical turn.

     Women who attempt to explain the concept of rape culture to men are often met with the response that not all men commit rape.  This is true, of course, but it misses the point.  The point of the claim that there is a rape culture in the US is, as I understand it, that there are cultural institutions and practices in place that facilitate and encourage rape, and that many (most/all) men implicitly support these institutions and practices.  The fact that many (most) men do not commit rape doesn’t speak to this claim very constructively.

     Anyone who suggests that the way we train and equip police leads to a lot of unnecessary violence and death is likely to be met with outraged protestations about the integrity and courage and sacrifice of police officers.  But the fact that many (most) police officers are people of integrity does not speak very constructively to the claim that we are training and equipping them for the wrong job.

     People who criticize religion—either a specific one or just in general—for the violence and mistreatment and misery it perpetrates meet the same kind of response.  They are admonished that most Christians/Muslims/Jews/whatever are “not like that.”  Again, true or false, this is not a very constructive response to criticism of a body of doctrine or a set of institutions and practices.  To say that a body of doctrine encourages people to behave badly is not the same as to say that all (or even most) who subscribe to the doctrine behave badly.

     The response in all three of these examples is a version of what logicians would call the fallacy of composition.  A fallacy is an inference that may be convincing but is not sound.  The fallacy of composition consists in arguing that because the parts of some whole have or lack some characteristic, the whole itself must have or lack this characteristic.  Two very simple examples:  since neither oxygen nor hydrogen has the property of being wet, neither can water; since each of us wants to pay lower taxes, society as a whole wants everyone to pay lower taxes.  

     These simple examples help to clarify the more complex examples above.  Men who reject claims about rape culture by pointing out that many (most) men do not commit rape are making an inference that looks very much like the fallacy of composition.  The implicit inference is that because most men lack a certain characteristic—the inclination to commit rape—society could not be encouraging rape.  The same can be said of those who answer criticisms of the police with protestations of the integrity of officers.  The implicit inference is that the system cannot be violent or corrupt if the people in it are not inclined to violence and corruption.  And of those who answer criticisms of religion with protestations of the goodness of believers.  The implicit inference is that if most of the people who subscribe to a doctrine are good people, the doctrine could not be the cause of evil.

     The corollary of the fallacy of composition is the fallacy of division, which consists in arguing that because some whole has or lacks a certain characteristic, each of the parts must have or lack this characteristic.  It is certainly possible to find people who argue that because there is a rape culture, all men are rapists; or that because policing involves much violence, all police officers are violent thugs; or that because religious doctrines are vile, so are all religious folk.  The fallacy of division is rather less common in our political discourse, in part, I think, because it does not cohere as well with individualism as the fallacy of composition.  But it is no less problematic for all that.

     Notice that both composition and division are what logicians would call “informal” fallacies.     Unlike inferences that are unacceptable because of their logical structure, inferences that embody informal fallacies are sometimes cogent.  There is no logically necessary connection between rape culture and the inclinations of individual men, but there may be a causal connection.  And this causal connection may run in either direction:  rape culture may influence the inclinations of men, or the inclinations of men may undermine rape culture.  And the same sort of thing can be said about the other examples above.

     The point is that whether or not an informal fallacy is actually fallacious will depend in part on the factual context in which the inference is made.  Sometimes there is factual support for the claim of a connection, or a disconnect, between part and whole; sometimes there is not.  It is the facts, and not logical structure that determines the cogency of the inference.

     Of course, facts, or their absence, don’t count for much in our political culture these days; if someone puts forward facts that undermine your reasoning, you can always cobble together some alternative facts.  And then we can continue to talk past each other, at the top of our lungs, in high dudgeon.  It’s so satisfying.