In recent years, metapsychology has been effectively destroyed by a series of critiques, here summarized. Clinical psychoanalysis, its heart, is a testable scientific theory and need not be trivialized by being reduced to a hermeneutics, but it has been exposed by Grunbaum and Rubinstein as scriously lacking in empirical verification. Its genetic hypotheses are extremely difficult to test; the clinical case study is useful only as a means of generating hypotheses. As Rubinstein has shown, however, the clinical theory can be systematized and stated in probabilistic propositions testable by statistical research. Its fundamental propositions can be tested only by nonpsychoanalytic data, however. Object relations and self psychology have had a large vogue but do not address the fundamental theoretical problems. Those threaten the survival of psychoanalysis, but are being complacently ignored. Some possible solutions are discussed.