Working off Kelsen, logic and rules of inference, as well as other rule based systems, are normative, "ought" based systems. If this is true, or even if it isn't, what reason do we have to take that logical rules are reasonable? In other words, why should one accept that rules of valid inference (of any system) as actually generating true responses from true premises?
To test a rule of inference, you can try to find counterexamples to it, cases in which the rule lets you derive a falsehood from true premises. Professor Vann McGee offered a well-known (and controversial) such attempt in this article . But there's no getting around rules of inference entirely. Even as you test one rule of inference you unavoidably rely on others. Because any attempt to answer the question "Why should we trust rules of inference at all?" will rely on reasoning, it will trust some rules of inference, whether or not those rules are made explicit in the reasoning. There's no way to get "outside" all rules of inference and see how they measure up against something more trustworthy than they are.