…no machine has been constructed that will deduce more than one conclusion; yet it has been shown that all possible general conclusions can be arranged in serial order and as soon as anybody wishes to defray the not extravagant cost, the specifications will be ready for a machine that will actually turn out new theorems from a given set of premises, one after another, as long as they continue to have any interest. But though a machine could do all that, and thus accomplish all that many an eminent mathematician accomplishes, it still cannot properly be called a reasoning machine, any more than the sort of man I have in view can be called a reasoner. It does not reason; it only proceeds by a rule of thumb.
