Increasing research efforts try to identify biological markers in order to support or eventually replace current practices of diagnosing mental disorders. Inasmuch as these disorders refer to subjective mental states, such efforts amount to their objectification. This gives rise to conceptual as well as empirical challenges: What kind of things are mental disorders? And how to deal with situations where subjective reports, clinical decisions, and brain scans contradict each other? The present paper starts out with a discussion of recent efforts to objectify beauty. Such attempts to quantify and localize psychological constructs in the brain are compared to earlier examples from the history of psychology. The paper then discusses personal and social implications of the objectification of subjective mental states, including mental disorders. The construct of Major Depressive Disorder, one of the most prevalent mental disorders, is then analyzed in more detail. It turns out that this is a very complex construct probably associated with highly heterogeneous actual instances of the disorder. It is then shown that it is unlikely to replace these symptoms' descriptions with patterns of brain activations, at least in the near future, given these patterns' empirical lack of specificity. The paper then discusses which of the disorder's core symptoms are more or less amenable to behavioral or neuroscientific investigation and analyses whether the heterophenomenological method can solve the problem. The conclusion is that the disorder construct is neither entirely subjective, nor completely objectifiable, and that clinical experts do well by continuing to take a pragmatical stance.