(o-Ar[t] = 41-

Tree-recognizers were introduced as generalizations of finite automata independently by J. E. Doner ([40], [41]) and by J. W. Thatcher and J. B. Wright ([114], [115]). They can be defined for both finite and infinite alphabets. Our definition and results so far have been for hyperrecognizers with a finite alphabet. But if we extend our definition to the case of a countably infinite alphabet X, it turns out that recognizability and hyperrecognizability are no longer equivalent.