Actually, it would follow that there could be no knowledge.
Unlike deer, knowledge only exists in our mind. It is not an actual thing. Without humans (or other sentient beings), there would be no knowledge.
Right. But that's not the question. Ex hypothesi, for there to be a case of knowledge, there must be at least one knower. But as per the last man example, for him to know, neither him nor anyone else (since there is no one else) needs to identify or judge his knowledge as knowledge. Thus, an inability to identify or judge knowledge as such does not entail the non-existence of knowledge. It could be that it exists, but we can never recognize it as such. Of course, ala
Willemena's point, its existence would be a difference which makes no difference, from an epistemic point of view. But the point I'm making is just the trivial one that logically, the one doesn't follow from the other.
As it happens, I think we
can identify knowledge as knowledge, and so the point turns out to be moot- the whole argument against identifying knowledge as knowledge is that we can never be certain, the possibility for error is always present. But if we recognize that certainty is a false criteria, for the reason's already mentioned (that it pertains to deductive knowledge in particular, and is neither a necessary nor sufficient condition for knowledge in general), this argument loses its force.
We decide what is knowledge and what is not knowledge. We must define something as knowledge in order for it to be knowledge.
We decide how to use the word "knowledge", and so we must define the word before we can say this or that is an instance of knowledge. But that doesn't mean that, given a definition or usage for this word, certain things won't
satisfy this definition irrespective of whether anyone is there to say, "hey, that's a case of knowledge!" To claim this, one has to commit to a bizarre metaphysic indeed.
Now, if we stipulate that something is knowledge if and only if it is true, then we must know whether something is true in order to define something as knowledge.
Since, as you concede above, we are unable to determine truth with certainty or infallibility, then we are also unable to define anything as knowledge.
We aren't stipulating that, that seems to be how the word is used. But yes, in order to consider or identify (not define) something as knowledge, to identify knowledge as knowledge in a particular instance, we have to see whether it is true. But your conclusion only follows if we agree that certainty or infallibility is part of our criteria for "is true".
If there is nobody to define anything as knowledge, then knowledge does not exist.
This is ambiguous. If nobody is there to define the word "knowledge", then the term is meaningless, and doesn't denote anything- so in a sense, knowledge would not exist. And as above, if there is nobody to
know in the first place, then knowledge doesn't exist; knowledge is contingent upon a subject, a knower.