A typical psychiatrist asked to describe something will give it a name. A typical psychologist will give it a number. The first approach is "categorical". It is the simplest and most natural way people sort things and is the method used with few exceptions (eg hypertension) throughout medicine. The second is "dimensional" and works best to describe phenomena that are continuous, lacking in clear boundaries, and reducible to numerical measurement.
In every day modern life, categorical and dimensional descriptions are complementary and both are essential. We use numbers to preserve accuracy in labelling things like height, weight, temperature, and IQ. We use names when a simpler, more vivid shorthand will do - saying something is red or blue is certainly imprecise but in most situations outside the physics lab is more convenient than giving an exact wavelength.
The Diagnostic and Statistical Manual currently in its fourth revision (DSM4) uses a categorical system naming the mental disorders, despite the fact that this loses information- since the phenomena it describes (like the color chart) have very fuzzy boundaries. For at least three decades, I and many others have suggested that a dimensional approach might have important advantages. These are particularly apparent in the diagnosis of the personality disorders because these so obviously lack clear boundaries with normality, one another, and with many of the other conditions diagnosed in the manual.
When, in 1987, I was appointed Chair of the task force preparing DSM4, I was hopeful we could begin to introduce the dimensional approach into the official nosology. After years of disappointed consideration, we decided that this would be premature for two reasons: 1- busy clinicians don't like dimensions and tend to ignore the ones already in the system and; 2-there are no available dimensional systems that are easily tranportable to the diagnostic manual. This latter problem does not arise from a lack of systems. During the past hundred years, academic psychology has produced literally thousands of measures -far too many,rather than too few. In this cacophony, there are no standout dimensional systems so widely accepted by the field as to be ready for enshrinement in an official nosology.
The first draft of the next revision of diagnostic manual (DSM5) posted at wwwDSM5.org. DSM5 has as a stated ambition to be "paradigm shifting" through the introduction of three different types of dimensional rating for: 1- personality disorders, 2- the severity of individual disorders and, 3- a list of "cross cutting" symptoms that would be rated across all disorders. Unfortunately, the reach of DSM5 far, far exceeds its grasp. Only by going to the website yourself and reviewing the DSM5 dimensional suggestions can you get a feel for just how remarkably ad hoc, idiosyncratic, and cumbersome they are. I have discussed the suggestions for dimensional personality disorder ratings with a number of experts (and this is also my area) and none of us could decipher the proposal, much less conceive of its ever being workable. One described it as an example of "too many research cooks spoiling a clinical broth".
The provision of severity ratings for individual disorders was actually tried in DSM3R, but was neglected by clinicians. The simple, generic ratings of severity and level of functioning already provided in DSM4 suffer a similar fate. The only chance for the acceptance of a dimensional system rest on its being simple, quick, universally acclaimed, and clinically useful and face valid. Even then, the odds at the moment remain low.
It is absolutely clear that DSM5 dimensional suggestions will not be paradigm shifting.
Indeed, in their current dense, inconsistent, and sometimes even incoherent form, the DSM5 dimensional proposals would constitute a grave setback for dimensional diagnosis. I wrote a paper in 1994 titled "Dimensional Diagnosis for Personality Disorder-Not Whether But When". I still believe that dimensional diagnosis offers advantages in accuracy that will eventually prevail over their unfamiliarity- but this will take time, consensus building, and education. The current DSM5 proposals are so badly done they may poison the well.