The question is why sometimes we report certain psychiatric conditions as being discrete (henceforth categorical) and sometimes as being dimensional (or in a spectrum).
The answer is related to the history of psychiatric classifications. By the time DSM-III was designed, there was an increasing resistance to psychoanalysis in America. Psychoanalysis tended to organize disorders in a dimensional fashion (have a look at this podcast. So the final product (DSM-III) translated that resistance by organizing all conditions in categories, increasing significantly their number as well.
In the past 10 years the lack of success (this is questionable, but psychiatry did not developed as well as, say, oncology) of the field made some scholars rethink about the exclusive use of categories.
It seems that categories are suited health insurances and legal objectives. For instance, judges are normally interested to know whether someone has or not a condition in order to decide about responsibility (necessary and sufficient conditions). The same with health insurance companies.
The problem is that psychiatry should be thinking on advancing the field, instead of satisfying the legal system or the health insurance companies (these are very important issues as well, don't get me wrong, but they are secondary).
Now the question is: are mental disorders really dimensional things? What are the evidences for thinking this way? I will only mention on argument and then I will suggest some readings to read all the other findings that suggests so.
The level of co-occurrence of mental disorders (co-morbidity) are very high. So that alcoholism is commonly accompanied by anxiety and depression. Depression and anxiety are themselves commonly found together, and the same for almost all the other hundreds of conditions. That means that perhaps a dimensional approach would be a better one to treat theses concepts.
Why do we accept that personality and intelligence are the type of things that should be treated as dimensional, then? Because the psychologists have been working without giving as much attention to politics than the psychiatrists in the last half-century. And they developed the statistical scaffolding for their theories on personality and intelligence, quite independently from psychiatrists. That is why these subjects are treated today as dimensional, whereas mental disorders are treated as categorical (note though that personality disorders also exists in the classification systems and DSM-5 attempts to introduce some dimensional approach on the subject).
Finally, why Autism has been taken as a spectrum condition? Simply because its study has gained traction only recently, and also because the main researchers in the area used methods from psychology to develop their ideas.
So the answer is that your perception that disorders are discrete is something inherited by the long use of classification systems like DSM, but that it will change shortly.
References:
- Bentall, R. P. (2004). Madness Explained: Psychosis and Human Nature. Penguin.
- Volkmar, F. R., & McPartland, J. C. (2014). From Kanner to DSM-5: autism as an evolving diagnostic concept. Annual Review of Clinical Psychology, 10, 193–212. http://doi.org/10.1146/annurev-clinpsy-032813-153710
- APA. (2013). DSM 5. American Psychiatric Association. http://doi.org/10.1176/appi.books.9780890425596.744053
- Zachar, P. (2000). Psychiatric disorders are not natural kinds. Philosophy, Psychiatry, & Psychology, 7(3), 167–182.