However, this power comes at a cost. If you call a method which does not exist, it would be a run-time error, not a compile-time error. This isn’t a big deal for interpreted languages - they don’t have a compile-time. However, the ability to find these errors before the code is run is a key benefit of compiled code.

Like most things in software engineering, there is a tradeoff.

I think having it optional is a good idea. You can normally take advantage of the compile-time method existence checks… but if you have a specific solution which would benefit from dynamic typing, that technique is still available.