I don't know... that all sort of discounts things like countenance, inflection, and gesture. Reading forces us to infer an awful lot of context, and text forces us to translate those things into a singular form of expression. I'd say reading appears faster because we're processing 1 language at a time instead of 2 or 3 or more. Can I parse your sentiment faster than you can convey it? Maybe our thinking is limited by language only because the process of thinking is actually distilling so many languages into a single language that becomes actionable, of which only part is verbal? The operating system is limited by the capacity of the hardware. Just other angles to consider.