He isn’t defining the word that way. He is saying subjective listening can be sighted (open to non-sonic influence) or controlled (not open to non-sonic factors influencing our perception of sound). Then he discusses what goes wrong with sighted subjective listening.
Also, sighted subjective listening it is not “incredibly narrow”: it is general practice to listen that way.
And so it should be (gasp). That does not invalidate objectivism or any of the criticisms of sighted listening.
Let's try and unpack some of this.
When you're buying a system or components, the system is effectively defined as the contraption that takes the source information and spits out the soundwaves. That makes you the observer of the system. You need to observe the whole experience - you need to control the source, use the remote, see what is there (because sight is the strongest of our senses, it influences our understanding of the sound), you need to let your whole brain operate because your memories, learning and understanding all feed into what you hear. I'd define the room here as part of the environment of the system - if you change the environment, you also can also change the outputs of the system, so I think we're OK with that, though you could define a treated room for a particular setup as part of the system.
But let's change our focus to that of someone wishing to learn from a system owner's description of a component change. When we wish to study that experience, we need to view the system in a different way.
Now, we include the listener and the room as part of the system. The person reading the description - let's say, you - are an outside observer of the system. The output of the system, as far as you the reader of the description - isn't the sound waves at all. The output is the listener's response, as described. What we know is as good as the description and no more.
The point is that we miss the difference between these two things. It's too easy to assume that we know about the sound waves - either to assume that the report will apply to the same hardware in a different room to a different listener, or to assume that "understanding and fixing the sound waves" will improve the system, when better results come from treating "the part of the system with the ears".
When you are looking to change or evaluate components for the purpose of improving your own system, you are actually looking to improve the system of which you are a part. So, of course, you have to do it sighted, to experience everything, not just the sound waves.
I guess, the ability to be the dispassionate observer armed with objectivity, and the listener in the moment, is what arms you against snake oil products and allows you to give permission to yourself to be subjective at the same time.
It also tells us that if we want to improve audio, we have to observe and improve the system that includes the listener. And that allows for the research that gives us the various Harman curves, doesn't it?