There is work to be done within any science that is pre-scientific. Categorization is sometimes one of those things. There are assumptions behind the way one chooses to classify things, but it may not be theoretically based. You may be assuming that your language contains names for the right parts of things, for instance, so that you can describe the structures in a clear way. You may be making assumptions about simplicity and completeness in observed patterns, that are not really theories, just intellectual biases.
Until you can put your assumptions into a testable form, a lot of folks would not classify classification work as scientific, because if there were an alternative classification scheme, there would be no basis on which to compare their merits until you rooted around in your unconscious assumptions for justifications that you had not actually explicitly used when you did the work.
From a Popperian/Kuhnian point of view, such work is pre-scientific, and not scientific. So things like descriptive chemistry (the ability to identify compounds by looking at them or testing them in traditional ways, largely handed down from Alchemy), or the parts of the original Linnaean taxonomy that have not yet been reconciled with evolutionary theory, are chunks of chemistry and biology that are still pre-scientific. You still have to know what chemicals you are using, and you still need to name birds consistently across time, whether or not they are dinosaurs.
So the now-traditional demarcation criterion creates a chicken-and-egg problem. Not all of science can be scientific, some of it has to be done on the way up to a theory. And that part can be a lot of work. And it is still important and engaging: Darwin loved classifying worms, slowly working out his criteria over time. Mendeleev spent a lot of time just comparing abstract patterns of elements before he could present an overall structure that could then be tested by new entries. Clinical Psychology, as the ongoing pre-science with occasional scientific pretensions, spends tons of time arguing over what things are or are not diseases -- e.g. transsexual feelings, sex-avoidance, sex addiction, homosexuality, and pedophilia -- and which diseases are really different from one another in what ways -- Borderline Personality Disorder vs Explosive Impulse Control Disorders vs the Complex Multiple Minor Trauma version of Post-Traumatic Stress Disorder... Those judgments can only be scientifically theorized about properly once their statistical stability allows testing something.
So to a large degree, saying that is not derogatory, it is just necessary.
Some corners of different sciences, biology and psychology especially, are more full of such necessary, pre-scientific content than others.
