Correct. There would be no epistemology as we know it if we had no word 'knowledge' or a calque of it. This seems blindingly obvious. Just trying to remove that word from existing treatises on epistemology, let alone words defined through that word, should show you that epistemology without that word would have to be very different.
I must say, that I highly disagree with you on this issue. I believe that words are not just tokens to be played with in a game, but that they actually refer to things in reality. There are real differences between knowledge and ignorance, or between truth and untruth, or between justice and injustice.
Say we succeeded in enforcing a ban on using the word "knowledge" (or any in philosophical term of your choice). My prediction is that philosophers would not stop writing treatises. They would just find or invent new words to describe what used to be described by the banned word(s). That's because they (maybe not all, but many of them) are not primarily interested in the word
"knowledge", but int the phenomenon
of knowledge we talk about by using that word.
Of course there is a difference between knowledge and ignorance. But which things are 'ignorance' and which things are 'knowledge' is a matter of knowing how and when those words are used correctly. As Tarski said, "snow is white" is true iff snow is white - and thus conversely, snow is white iff "snow is white" is true. And that's a fact about how we use the word 'white', and the word 'snow'.
Is there a difference between knowledge and ignorance? Yes, we call one thing knowledge and the other ignorance - and since we can be observed to use these words in similar ways to one another (ie what one man calls knowledge, another man is likely also to call knowledge, with the exception of some intentionally troublesome borderline cases), there must be some difference that we are distinguishing. But we cannot describe that difference except by making some distinction in our language, and hence we cannot refer to the difference in contradistinction to the distinction through which the difference is expressed. That is, we cannot distinguish the fact that something untrue is not a known thing from the fact that if a thing can be truly said to be 'untrue' it is not a known thing. The positivists attempted to define these criteria of use in terms of phenomenal experience, but phenomenal experience itself not only cannot be described without language, but is not itself prior to language-like structuring [we do not in reality experience seeing some hands and some arms and a torso and thus conclude that we are seeing a man... rather, we experience seeing a man, and thus conclude that we must also be seeing some hands, some arms and a torso... but the concept of 'man' is a linguistic or protolinguistic concept that cannot be extracted from our phenomenal experience].
Ultimately, I think, the meaning of words is fixed not in experience but in action - words are, as you put it, counters in a game, and the nature of the game (our mode of life) determines the counters we use and the relations between those counters. (We must assume there is some external constraint on our game, a reason why it takes the forms it takes and not others, but it is precisely that - external, unspeakable, not what we are ever talking about.
There are, I think, two different kinds of philosophical inquiry: ethical and lexicographical. Lexicographical inquiries take the counters that we use, and attempt to describe them and their relations more clearly. Ethical inquiries instead use
those counters in a reflexive process of action. [There is a third potential type - the type that encourages use to use the counters differently. This we might call 'therapeutic'. All lexicography is ultimately therapeutic, but not all is self-aware in being so].
What would happen if the word 'knowledge' were banned? Well, as you say, first of all people would use a different word or series of words to stand in place of the banned word - they would have to, because their mode of life would not have changed, and that mode of life requires the concept of knowledge as we now think of it. It's as though we banned the use of cards in poker - people will just use tiles instead, in place of the cards - the functional demand remaining, an alternative counter is introduced into that function. This is why I originally said "or a calque
". These expressions would be calques.
However, if our form of life were to be changed such that 'knowledge' no longer fulfilled a necessary function, we would by definition not need to replace it. In that case, lexicographical inquiries into the meaning of 'knowledge' would go away - there would be no word to investigate. Ethical inquiries would remain, and it is indeed hard to imagine an ethics without knowledge, which is why the whole thought experiment of a form of life without 'knowledge' seems so improbably. And yet, while ethics may seem to us to require what we might call a knowledge-type counter, that counter not only need not be the word 'knowledge', it need not be anything functionally equivalent to knowledge in its modern-day philosophical sense. Indeed, it's arguable I think that 'knowledge' as philosophers define it is not very indispensible at all, and that the way the word 'knowledge' is actually used in practical ethics is quite different. Or rather: philosophical discussions of knowledge are a political project (like all discussion) that advances one particular family of ethical theory. [Specifically, when a philosopher asks 'what is knowledge' he mostly seems to be asking about an ethical system in which rationality is a cardinal virtue and knowledge is its reward - Gettier problems are problematic because they are cases where "knowledge" as defined simply is awarded too easily, as the result of luck rather than of epistemological virtue - but at the same time a system in which epistemological virtues are not promoted per se, and are seen as defined by external, objective, facts of the matter. Outside these post-Kantian ethical commitments, it's hard to see why Modern Philosophical 'Knowledge' (as opposed to some other, differing or more vague, concept within a knowledge-like semantic field) is of any moral significance].
OK, I think i've strayed, and I'm making this too complicated (true, but too complicated). My simpler objection: sure, epistemology could carry on as normal without the word 'knowledge', but not without a calque of that word. This, to me, is cheating, because what I am talking about with the word 'word' is not the phonological surface form, but the functional identity, the place in the language. Epistemology as we know it depends on a language that has a word for 'knowledge', which is to say that the language shares with Modern Philosophical English a functional position, the position that in english is filled by the utterance 'knowledge'.
This is why we are, more or less, able to talk philosophy with Germans (or Austrians, at any rate) - because our languages have, as it were, analogous functional positions, which may not be purely identical, but which are sufficiently similar to allow translation. Imagine a language as an irregular grid of dots, and a philosophical treatise as a line-shape that joins up some of these dots - 'close' languages can translate because a line shape in one can be imposed on the other grid by joining up analogous dots, with minimal distortion. But when you try to translate into english from, say, ancient chinese, you run into problems, because the dots in the two grids are arranged so differently that joining up the analogous dots gives a totally distorted figure... or even because it is no longer clear which dots are really 'analogous' to one another.
Of course, as I said before, basic parts of our language usage will be easy enough to translate - they're part of any ethical system we could want. It's hard to imagine living for long without a concept of heat, for instance, so heat, in its basic physical sense, will be something we can talk about easily from one language to another. But more complicated ethical concepts, like 'knowledge', or 'chastity' or 'possession' will often be more specifical to one culture (and hence one language) than another, so translation will be difficult without describing a huge chunk of the surrounding form of life.
OK, this seemed very simple, so I'm going to resist trying to explain it again, though I'm sure there must be a better way that I've missed, because I'm making things sound complicated, I think...