In fact Dick is not assuming that reality is communicable. All he assumes is that any concept that is communicable (ie, any word) can be tagged with a number. That notion in itself is not as controversial as Harv thinks it is; all Dick is saying is that any explanation of anything can be converted into a series of numbers. That is exactly how we use computers to communicate after all.
Let me say that this concept is controversial - very controversial. Whenever we talk about the human experience of reality, I agree - and have always agreed - that we can represent any concept (or object, process, etc) as a series of numbers. BUT, this is NOT the same thing as saying that reality is representable by a series of numbers. This is what we do not know. Reality, for all we know, may exist far outside the scope of being captured by quantitative methods (and there's good reason to suggest this), and therefore I reject this naive notion by Dick.
However, let's assume for argument's sake that Dick is talking about the communicating our human experience and not representing reality itself. The quantitative model that he holds as restrictive to human model construction is still naive since to represent our experience of reality by numbers doesn't mean anything unless you can show how the so-called restriction to our model making conforms to new observables. This is what science is vastly composed, finding how our new theories conform to new observables such that the new theories can described as a restriction in our observations. Stafford scoffs at this requirement, however mathematical models have shown continual fallibility of being non-restrictive to our experience (i.e., wrong), and therefore just being a mathematical model is not a high enough standard to gauge whether a new model is restrictive or non-restrictive to human experience such as what Dick would like everyone to believe.
Dick is not really concerned about reality, he's concerned about the kinds of relationships you can find between the symbols used to communicate a self-consistent explanation of anything, including "reality" and "bananas". That's the bit Harv doesn't get, mostly because Dick doesn't know how to explain himself.
Dick is all over the map here. At times he wants us to believe he has found an ontological restriction and at other times he wants us to believe he has found an epistemological restriction. Both cases fail. As an ontological restriction his model fails because he cannot show that reality is necessarily mathematical. And, as an epistemological restriction his model fails because he cannot show that our experience of reality is necessarily dictated by every conceivable math model. In fact, quite the opposite is the case, and the only fully verifiable means we have to know whether our experience is in anyway constrained to certain observations is through the scientific methods and subsequent models that emerge as a result. Since Dick only has ridicule for those epistemological methods, he is very far from having a clue as to why his model cannot be seen in any serious light.
We were born without knowing a single word in English (or whatever your mother tongue happens to be), and you had to learn that language by listening to people talking in that language. How do we do it?
This problem is far from solvable, and I hope you aren't suggesting an answer.
Btw, nice to see that you are back!