Well. This does somewhat miss the point of why she wanted to bring this Latin speaker back. Beard is a classicist, which is not the same thing as a linguist, and she would have her own reasons which would be very different from the reasons I would give for bringing back a Latin speaker, I'm sure. But I think we would overlap in one opinion, which is that it really isn't the same thing at all!
Linguists do lots of different types of research, but almost all of it involves finding out how people speak, recording that and using the resulting data to test a hypothesis. Some linguists use their own intuitions as data, some use other people's intuitions, and some use other people's natural speech.
Let's say I wanted to find out about the use of Isn't he not? and Isn't he? in the Geordie dialect, and what the differences are in the use of each. I might try to record some people's natural speech, and I might do some kind of survey where I asked people what sounded most natural to them. I might use my own intuitions to start me off, but I wouldn't rely on them, because I wasn't born and brought up in Newcastle so my intuitions might not be reliable.
When I wrote my PhD thesis, I needed to know facts about a lot of different languages. I used books for this, grammars that describe how the language works. That meant that I didn't have to spend a lot of time and money travelling around the world finding people to record. But these books can only take you so far. There are two problems with them: the first is that they might not be reliable. Modern descriptive grammars are good, thorough and accurate, written by linguists with a lot of training. The grammars written by missionaries from the SIL are generally very good, too. But an old grammar might be written by any unqualified person with little training, and with who-knows-what purpose. And grammars are almost never written by a native speaker of the language, so even with the best intentions they don't always know for sure that what they're being told is accurate: perhaps their consultants are subconsciously telling them how they should speak - like if you told a linguist that every sentence has to have a subject and then went off and said 'Don't know what he wants all this information for', without using a subject.
The second problem is that they can never tell you absolutely everything about a language. Hardly any of the grammars I looked at for my PhD could tell me if the question particle could be used in embedded questions. Therefore, I had to supplement this knowledge with intuitions about some of the languages. Obviously I don't speak these languages, so I couldn't give the intuitions, so I asked native speakers of those languages. And it is important that they are native speakers. Just because someone has learnt a language doesn't make them capable of making subtle judgements about what is and isn't grammatical, especially when you get into non-standard forms.
There are some linguists who can't ask people about their language because there aren't any people to ask. These are the historical linguists. Someone studying Old English can't ask a speaker of Old English what the language is like, but fortunately there are some texts written in Old English that still survive. Without those, we'd have no idea beyond the reconstructive work that can be done based on regular changes over time. There is a massive amount that can be learnt from historical texts, and historical linguists do some truly amazing things. But it is limited to what exists already, whereas a living language is infinite and can constantly provide new data.
Latin, similarly, has no living native speakers. It has evolved and become the modern Romance languages, but these are not the same as Latin any more than present-day English is the same as Old English (the language of Beowulf). There are living people who speak it, but they are not native speakers, and more than I'm a native speaker of French because I learnt it at school. In fact, these people are even further removed from being native speakers of Latin because they didn't even learn it from a native speaker (in fact, I didn't learn French from a native French speaker, but that's beside the point). There are lots of things we don't know about Latin, and those are the things linguists might be interested to find out, and there's no way we can find them out from modern Latin speakers because their knowledge of Latin is based only on what we already know, not their native competence. We aren't going to get very far trying to find out what we don't already know if we ask people who only know what we already know.
So whatever reason Mary Beard has for bringing back a real live Latin speaker, perhaps for the sheer pleasure of hearing them speak, I too would like to bring one (or ideally, several) back, so that I could do research on their language without having to rely on written sources.
