Unfortunately it is true that taxonomy is not an 'absolute' science. If it was not, why the need of a committee - we would simply need to apply a set of rules and have a simple check of the results by a reviewer.
I feel that the last three pages illustrate the need to emphasize that taxonomy isn’t ANY kind of science. In ornithology, we use science to inform our taxonomy, which is only an organization system.
The Tobias system was an attempt to make that organization systematic, objective, uniform, and criterion based… which are all characteristics that science has. But as has been pointed out, there are still subjective criteria. As I want to point out, there is practically no way to test hypotheses, illustrate significant difference, etc. in what is essentially a naming system after all. We can do science with things like genes and investigating their differences, or with any of various phenotypic traits. But stating which of those differences makes a category or not - this is absolutely not an issue of hypothesis testing and all to do with naming and structuring.
Or to take the inverse perspective - let’s imagine that we are investigating “taxonology.” This would mean running some sort of experiment in which we can test a single parameter to measure (somehow) the categorization. So for example, if we define a “good” taxon as one that is “not confusing” and present different systems of Eurasian/Green-winged Teal taxonomies to scientists or birders (in statistically robust numbers of course), then measure confusion with some proven test to decide whether it should be split. If we define a “good” taxon as having a certain genetic difference - well, do we simply deal with gulls and ducks measured the same way as owls and warblers (regardless of any other differences/similarities), or do we need a different taxonomy for each… and how do we “test” those? Or we could dive into Tobias groundwork and test which criteria produce categories that produce whichever (desired?) result of categorization… but we first need to somehow find an objective and testable way of defining what “makes” a category. How does one measure and test the difference between behavioral, physical and genetic difference against one another, alongside communication needs, past and present usage, and any number of other variables which relate to how we organize things? Do we first need to study what factors humans use to categorize birds… in some sort of controlled unbiased island devoid of influence from Clements, IOC, and Birdlife? If its found that the most important thing is what color and size the bird is and genetic relationship be damned, is that really the kind of communication we want?
Subjectively, this is very easy - we just decide. Objectively - nearly impossible to compare these apples with those oranges and pretend it is not violating basic principles of scientific method. Needless to say, we don’t engage in this absurdity for every species, genus, family, and so on - we subjectively decide.
Taxonomy is a system of word use - in a sense it is a language. Beyond “describing what is used and what isn’t” I don’t believe there is much room for objective truth without introducing pretense. And I believe that’s okay, as long as we aren’t pretending it is something it isn’t.