Mental Context for Model Theory — LessWrong