How Accurate Is Linguistic Reconstruction?

[Zyobot]

Despite the fact that they've been out of everyday use for quite some time and were once spoken in eras without modern record-keeping, long-extant languages still remain with us in some form or another. Enough, in fact, so that dedicated learners can acquire the knowledge and tools to become fluent and immerse themselves in the works and culture of the long-gone peoples who spoken them on a daily basis.

However, since the original speakers are no longer with us and there remains a complete lack of audio and footage of them in action, that leaves me wondering: how is it that professionals decipher these ancient languages in the first place, and how accurate can we bet that their efforts at linguistic reconstruction are? Though official pronunciation may be imperfect and we're probably missing some words and minor features here and there, at least Latin and Greek have written records and centuries of use to point to, whereas further back, currently theoretical tongues--like Proto-Indo-European--have no such direct evidence for us to study and decode.

Thank you in advance,

Zyobot

[Zyobot]

Perhaps people here will find Dr. Jackson Crawford's take on how linguists rediscover the sounds and other mechanics of dead languages a useful explanation.

How do we know what a dead language sounded like?


He has plenty of other videos concerning both linguistics and various aspects of Norse history, considering that's apparently his specialty.

[kyng]

I recently discovered that there are efforts to reconstruct the Norn language of the Orkney and Shetland Islands, which died out some time around 1850. However, they acknowledge that they have no hope of making it exactly the same as the old Norn language - so, their effort will be known as 'Nynorn', or 'New Norn':



Still, they're not going in completely blind. In the late 19th century, a couple of linguists went there to speak to 'rememberers' of the language (i.e. people who couldn't speak it themselves, but did know some words and phrases, passed down from ancestors who did speak the language). Between them, they managed to record over 10,000 Norn words, which is a decent base to work from. However, there will still be gaps - so, they're filling some of those in by borrowing from Icelandic and Faroese, which are closely related to Norn. Other gaps are being filled in by forming compound words, based on the Norn words that have been recorded.

So far, it doesn't have any new native speakers, but it'll be interesting to see whether it gets anywhere.