In older variations of English in history, how much evidence of written language samples are needed to accurately define the grammar and usage of that period?
For example, if we want to define how English was in the year 500 CE, how much evidence in the form of parchments and samples must we have and analyze? Is there an official figure or standard?
The short answer is that there are so little available sources that you would need all you can find. In addition, the corpus has already been gathered and is readily available.
Indeed, one of the most complete corpus of Old English sources is the one gathered by the University of Toronto to elaborate their Dictionary of Old English. This dictionary is still a work in progress and they have only reached the letter 'G'. It is online and the corpus is also available online or on a CD.
From their own web site, they claim to have gathered a comprehensive corpus:
The DOE is based on a computerized Corpus comprising at least one copy of each text surviving in Old English. The total size is not quite five times the collected works of Shakespeare.
Considering the scarcity of sources, this is a remarkable achievement (consider for instance that the story of "Beowulf" is only known from a single, half burnt and incomplete manuscript).
As for year 500 AD, there are simply no sources (except some runic carvings) and the words are reconstructed. The oldest sources are dated ca 650.
Another problem is that before widespread printing the language, grammar and spelling changed from region to region and person to person. Except perhaps in copying Gospels spelling doesn't seem to have mattered much to anyone
