Thanks to
bamalip I'm now learning about
Subobject Classifier.

Let's play with it a bit. If we define a single issue as the whole purpose of American politics, then we can consider it to be a terminal object *. For example, when we declare "jobs for Americans" the ultimate goal for the president anybody (∀X) who is for "jobs" satisfies the Classifier Ω and is a patriotic American U. Furthermore, if we structure the debate as a Boolean Topos: "jobs vs environment", or "jobs vs self-driving cars", or "jobs vs immigration", or "jobs vs China", or "jobs vs trade", or "jobs vs Amazon", etc., everybody on the other side of "jobs" is the enemy. The simplified, "terminal" structure of the debate defines the outcome and political polarization.
upd. On a somewhat related note, the sex vs gender debate has a similar structure. Category "sex" has a terminal object, e.g. a definite chromosome characteristic, therefore it can be easily understood as a Boolean Topos. Category "gender" doesn't have a terminal object, therefore people get confused.