If you take a social constructionist view of mental health diagnosis, then diagnostic criteria will naturally change over time because they reflect social norms of the time.

For example: Imagine that the definition of “wealthy” in 1950 is owning one 1 TV, 1 car and an income of more than £5000. In 2014, the definition might be having a flatscreen TV in every room, owning 3 or more cars, and an income of more than £200,000. This is because the social norms and the context is different: income is generally higher, things like TVs are cheaper and more widely available, and expectations have changed. However, this doesn’t mean that we can say that a person in the 1950s isn’t wealthy because they don’t meet the 2014 threshold: in the context of 1950, the person would still be considered more well-off than his neighbours. Similarly, If someone has a diagnosis of schizophrenia in 1950, that diagnosis is “correct”* in the context of 1950s culture. Someone might need to meet different criteria in 2014, but then the culture of 2014 is also different. Either way, both diagnoses can be seen as “correct” relative to the context the person is in.