First, no, we use digital machines because manipulating analog quantities is an impossible mess. The essence of an analog quantity is that it is unquantified, and therefore there is no real way of telling if two quantities are the same or not, or whether they would have been the same if one of the two hadn't been perturbed by something. Whenever you write down (or emit in any way) analog information, you're losing part of it; it can and will be altered by any amount of noise, and altered again when it's picked up by a reader or receiver. Not altered as in "a bit of noise but we can discard that". Altered in the sense of "it's impossible to tell the information from the noise". There is not much processing or storing you can do with it.
As for the fact we're "symbolic creatures", perhaps we are, but again- is it because our brains are natively digital or because we really put a lot of effort in trying to be symbolic? (I don't know whether our brains can be called digital- neural networks seem more analog than digital, but there might be quantization going on at some levels to manage the complexity and the noise of handling analog information.) Maybe symbolic reasoning emerges despite being really hard to master, because it actually manages to grasp some fundamental aspect of reality?
You may need to quantize data for error correction (as in "a trigger that differentiates symbol A from symbol B"), but you don't need numbers for that (as in "an arbitrary chaining of the sequence operator", which is the base of natural numbers).
You can have logic -binary or not- to define rules of inference, without defining a particular rule of inference for numbers (measuring quantities).
As someone else sort of implied elsewhere [1], if aliens use mathematics at all, it would probably be easier to reason with them in terms of topology rather than number theory.
As for the fact we're "symbolic creatures", perhaps we are, but again- is it because our brains are natively digital or because we really put a lot of effort in trying to be symbolic? (I don't know whether our brains can be called digital- neural networks seem more analog than digital, but there might be quantization going on at some levels to manage the complexity and the noise of handling analog information.) Maybe symbolic reasoning emerges despite being really hard to master, because it actually manages to grasp some fundamental aspect of reality?