Think about haskell. haskell uses normal order evaluation (evaluate from the outermost parens) and CL is believed to use applicative evaluation (evaluate from the innermost parens). Not actually.
Evaluation is all about rewriting and transformation. Then remember that CL macros expands from the outermost parens. Thus, it can be said that CL supports normal order evaluation in a form of macro, which runs in compile-time only, while in runtime it uses applicable eval with functions & special operators. Haskell, on the other hand, also use normal order eval in compile time, but hides every runtime clutters inside the (fog of) monads and lazy evaluation.