This question continues from Literacy in the classical world
It seems that literacy rates for populations prior to mass education becoming prevalent were estimated to be 20% or lower. However, it can clearly be observed (when learning new languages) that literacy (as defined as being able to read and write in a given language's script) is often far easier than fluency.
For example, the Hangeul script was derided by Chinese-educated Korean scholars as being trivially easy to learn, being designed for ease of study. http://en.wikipedia.org/wiki/Hangul#Other_names It was given names such as "script you can learn in a day" . Despite this, censuses taken during and after the Japanese colonialisation era showed that prior to aggressive governmental efforts to push literacy by implementing mass education, literacy rates were extremely low. http://fightforjustice.info/?page_id=3174&lang=en
The percentage of South Koreans unable to read decreased sharply from 77.8% in 1945 to 41.3% in 1948, down to and 13.9% in 1954.
Similar arguments can probably be made for European alphabet-based scripts such as Latin, Greek and Cyrillic, which were not significantly harder to read and write than Hangeul (which was only easy to achieve literacy compared to Chinese)
Since we can assume that people living in communities would need to be fluent in a common spoken language to communicate with each other, the fluency rate would then be very high. Why then, was literacy so low prior to the advent of mass education, when literacy was clearly highly beneficial? Were there any artificial barriers to entry that caused this low literacy rate?