Will Knowledge Organisation Systems kill diversity?

The release of GBIF's "Recommendations for the use of Knowledge Organisation Systems by GBIF" made me think of whether the implicit drive to create the ultimate information system is not a wrong ultimately damaging approach. Such as system assumes that it models our world properly, and, ultimately that there is only one way to do so, hence the KOS for biodiversity. Essentially, this will mean, that anything that can not be packaged within the ontologies building the KOS can not be integrated - or from another perspective, we begin to look at the world from a very restrictive view - in German this would be called Scheuklappen, the little piece of leather mounted on the sides of horse eyes to avoid distractions.
This all would be not really relevant, wouldn't the organisation behind this recommendation utlimately strive to create the the bioinformatics technology standards (TDWG - Taxonomic Data Working Group, more recently changed to Biodiversity Information Standards) we all are using - better will have to use - to become part of all we hope, the seamless knowledge space.
I find this even more questionable at a moment, where this community can not even deal with something as trivial as a bibliographic citation, not to speak to build up a database of all the citations.
Or in the realm of taxonomic digital literature where we encounter on the one hand something very simple like treatments, but then are not able to define what it is - something that is now crucial to create schemas, DTDs or similar to model this domain for the purpose of creating semantically enhanced documents. The legacy data shows clearly, that there is a huge variety in how treatments are being communicated - but our goal now is to create a standard - eventually defined in one of the TDWG vocabulary that ought be more restrictive to allow reasoning and machine enabled tools to help to work through the huge amount of data we hope to open up with that move. May be, one might have to consider a definition less as something all inclusive but rather as a concept that allows a lot of flexibility in its application?

Another point is: Knowledge of what and what do we want to do with it? Bowker and Star put it this way:

Each standard and each category valorizes some point of view and silences another. This is not inherently a bad thing - indeed it is inescapable. But it is an ethical choice, and as such it is dangerous - not bad, but dangerous.

Though the others have in mind racial classification, I would argue, that we need to keep this issue of exclusion in mind.

In other words: What is the limit of what we can do with the system we create? What is the limit of the material regarding creating meaningful (as opposed to artefactual) knowledge?

This of course touching upon something very different and very negelected: Quality Control of our input data.

I wonder whether the authors provide an answer to this. The why and what questions seems to me the main stumbling block for biodiversity informatics in general.

With this in mind, I will read through this new recommendations.