University Of Oxford Wants Volunteers To Transcribe 'Lost Gospels'

[urroner]

Found this little article on the internet where they, researchers at the University of Oxford, are asking for volunteers to "help transcribe two million segments of ancient Greek writings in a bid to find the next 'lost gospel.'"

Volunteers Wanted to Transcribe 'Lost Gospels'

OXFORD, England – Thousands of volunteers are needed to help transcribe two million segments of ancient Greek writings in a bid to find the next "lost gospel," researchers at the University of Oxford said.

Some 200,000 segments of the ancient papyri, found in Egypt, were uploaded to the website www.ancientlives.org -- and now volunteers are needed to transcribe the Greek letters using character-recognition tools.

Researchers from the project already discovered previously-unknown "gospels," including stories about Jesus casting out demons that were written in the third century.

But now the project is looking for people across the world -- who are not required to read classical Greek -- to help translate the ancient documents because the task is too big for the research team.

The software allows people to match Greek letters with the symbols on the papyri, then translates the words and stores them on a searchable database.

The scripts include a range of bills, receipts, letters and other everyday documents, but researchers believe they also could contain more stories about Jesus.

The papyri were part of a haul discovered in the Egyptian city of Oxyrhynchus early last century and date from between 500 BC to 1,000 AD. They were written in Greek when Egypt was under the control of the Greeks and Romans.

Previous interesting finds among the scripts included plays by Sophocles.

[ldsfaqs]

That's just dumb. Using software to match Greek with Symbols and then it translates?

They are completely ripping the texts out of their context by doing that, not to mention software can't detect nuances, exceptions to the rule, etc. For example, we know related to the BOA that standard/common Egyptian translations were not always the ones actually used, because there is always variations in dialect, culture, and language usage. What is further sad about this, is that this shows the actual state of the educational systems of the world, that there are just not enough people who can do Greek and Egyptian to make this their own couple of years project, so the job is done right. Just sad.

[volgadon]

For example, we know related to the BOA that standard/common Egyptian translations were not always the ones actually used, because there is always variations in dialect, culture, and language usage.

First of all, scholars are aware of that. Second, the BoA shows howa different group could adopt and adapt symbols, not words. Third, once scholars have a database then they will be able to scrutinise the original much quicker. Computer databases are a boon. It can take decades otherwise, such is the case with the Cairo Genizah.

[David T]

That's just dumb. Using software to match Greek with Symbols and then it translates?

They are completely ripping the texts out of their context by doing that, not to mention software can't detect nuances, exceptions to the rule, etc. For example, we know related to the BOA that standard/common Egyptian translations were not always the ones actually used, because there is always variations in dialect, culture, and language usage. What is further sad about this, is that this shows the actual state of the educational systems of the world, that there are just not enough people who can do Greek and Egyptian to make this their own couple of years project, so the job is done right. Just sad.

I think you misunderstand what's going on. The database is for ease in transcription, and searching/indexing keywords in a database - not doing the definitive translation itself. It would be a key resource in determining the priority certain documents are given for full attention and scholastic translation.

[mfbukowski]

Sounds to me that it is just "letter matching"- "Oh- this symbol looks like this "gamma" over here in the reference- the next one looks like a "beta"" etc.

I can see how that would be highly valuable, but holy cow, we are talking major tedium here, at least for me.

[David T]

It's actually not much different than the process used for Family Search Indexing.

[jo1952]

It's actually not much different than the process used for Family Search Indexing.

You beat me to this very same observation. All I can say is thank heavens people go over our work and review it! Whew.......

Regards,

jo

[jo1952]

Found this little article on the internet where they, researchers at the University of Oxford, are asking for volunteers to "help transcribe two million segments of ancient Greek writings in a bid to find the next 'lost gospel.'"

Volunteers Wanted to Transcribe 'Lost Gospels'

Thank you for the heads up, Urroner! In these latter days more and more will be revealed. Who knows what methods Father will use to accomplish His purposes??

Love,

jo

[Joseph Antley]

I'm pretty skeptical about the software's accuracy in dealing with the Greek on papyri. At least from my experience reading Greek handwritings is more of an art (and one that I'm very, very bad at) than something you can just program, especially since in most Greek handwritings the same single stroke or double stroke is almost identical for multiple letters.

Of course I'm sure they've thought of this, so I think I'll ask around. If there is actually is character recognition software for transcribing Greek papyri this would be a huge boon to papyrologists all over the world.

[maklelan]

I'm pretty skeptical about the software's accuracy in dealing with the Greek on papyri. At least from my experience reading Greek handwritings is more of an art (and one that I'm very, very bad at) than something you can just program, especially since in most Greek handwritings the same single stroke or double stroke is almost identical for multiple letters.

Of course I'm sure they've thought of this, so I think I'll ask around. If there is actually is character recognition software for transcribing Greek papyri this would be a huge boon to papyrologists all over the world.

There was a PhD student in computer science at BYU when I was there who was developing character recognition software for ancient scripts. I don't know where he is now.

[Doctor Steuss]

If they had to end up creating Captcha to help with printed English, I can't even imagine what will be involved in creating this software (hopefully I don't get some crazy Captcha in Greek next time I try to sign up for a site ).

[Joseph Antley]

There was a PhD student in computer science at BYU when I was there who was developing character recognition software for ancient scripts. I don't know where he is now.

I can understand it working for uncial or any spaced scripts, but for any less-literary handwriting I'm not sure how it could work. The only thing I can think of is if the recognition program has a massive database of Greek words from which it can deduce what letter sequences are probable. Other wise I'm not sure how it would always be able to tell the difference between letters like a pi and a nu or an alpha and an omicron, which are virtually identical in a lot of ancient hands.

The other problem is that there are nearly hundreds of different styles of Greek handwriting on Egyptian papyri where different scribes can write the same letter completely differently. The software would have to be able to recognize both a delta from Tebtynis in 100 A.D. and a delta from Oxyrhynchus in 500 A.D., which look significantly different -- without confusing the 100 A.D. Tebt. delta with an almost identical-looking character from a different papyrus but which might actually be a different Greek letter.

But again if they're employing this program they must be aware of this and have thought of some solution for it. I'm just curious as to what it is.

Edited July 27, 2011 by Joseph Antley

[ldsfaqs]

Indeed, there are countless problems with such a "program", I'm surprised that so many people think it's just so good. Obviously, in certain circumstances such can be helpful, but I see the potential for so much bad translating in this KIND of case, that someone should have done the work manually in the first place.

And who-ever said it's not going to be used to "translate" the texts, read the article again. It specifically states it.

match Greek letters with the symbols on the papyri, then translates the words and stores them

[mfbukowski]

Indeed, there are countless problems with such a "program", I'm surprised that so many people think it's just so good. Obviously, in certain circumstances such can be helpful, but I see the potential for so much bad translating in this KIND of case, that someone should have done the work manually in the first place.

And who-ever said it's not going to be used to "translate" the texts, read the article again. It specifically states it.

match Greek letters with the symbols on the papyri, then translates the words and stores them

Nothing says its OCR software that does the matching- people do the matching. Of course the people will have to be trained in recognizing Greek letters- it is the software which translates the words once they are input.

I dunno- I guess I am having trouble seeing why that is so complicated. You see a hand-written letter, you type a Greek letter, those letters make words. Software translates the words. What am I missing here?

[Joseph Antley]

Indeed, there are countless problems with such a "program", I'm surprised that so many people think it's just so good. Obviously, in certain circumstances such can be helpful, but I see the potential for so much bad translating in this KIND of case, that someone should have done the work manually in the first place.

And who-ever said it's not going to be used to "translate" the texts, read the article again. It specifically states it.

match Greek letters with the symbols on the papyri, then translates the words and stores them

I think it means it translates them into modern, standard Greek characters, not into English words.

Edited July 28, 2011 by Joseph Antley

[David T]

Indeed, there are countless problems with such a "program", I'm surprised that so many people think it's just so good. Obviously, in certain circumstances such can be helpful, but I see the potential for so much bad translating in this KIND of case, that someone should have done the work manually in the first place.

Again, this isn't done to make definitve translations of the pieces. It's to determine which documents are worth prioritized attention and interest. It's a huge stash of documents, not of equal interest to everyone. If a scholar could do keyword searches to determine what is on the records, then they could dig up the actual record, and give it an actual translation and study. It's an identification and indexing tool. Not intended to be a work of scholastic literature.

Kind of like 'Google Books' for the ancient document. Google Books the software won't tell you what the book means, but you can use it to find a book that does talk about what you want by searching for keywords. Then you can either download the actual book to read yourself, or buy it. I've found several fascinating books I didn't know existed this way by searching through the Google books index.