Software Use in Bible Translation

[John of Japan]

There are a variety of programs that can help translators, but none that do the actual translating. Actual translation software is like Google Translate: you have to physically type in or copy and paste what you want translated. There are no actual AI programs that do the translating themselves after simply being commanded to. So, this thread will be about the software packages and other electronic helps that can be of use to a translator. By all means, join in.

When I first began translating in about 1995, I joined a team which was working on the Gospel of John as the Greek nerd. We got through John 4 before things fell apart. The only software we used was a dedicated Japanese word processor. I'll attach a photo. I'm the guy on the left in a blue shirt.

That old term "dedicated" means that typing in Japanese was absolutely all that it did. It word-processed, period. However, that was a huge jump from the traditional Japanese typewriter, which was about 3' x 2', and had an arm that you used to pick up the individual type piece that had the Japanese letter or kanji (Chinese character) you needed. There were about 3000 of those type pieces.

 

[John of Japan]

In 2002 (I think it was), I began translating on my own, and foolishly did not use a computer, but drew all of the characters by hand. Check this one out, which takes 20 separate strokes to draw: 議.

I finally figured out I should use my Japanese digital resources. The first Japanese word processor I had bought in the mid 1980's was awesome! It had about a half page of memory, could print your half page, but had no way to save the file. Later after the technology evolved a kind supporter sent $400 bucks so I could buy a word processor that could save onto a 3.5" disk, and it was a huge help. That's the first digital machine I used to translate with.

Then I discovered a computer program called Seedmaster, and I thought it was incredible. It actually accessed Strong's glosses (short definition). Later I found a JW website, believe it or not, that allowed me to add Thayer's and another resource or two. Also, I could directly copy the Greek text into my WordPerfect word processor. What will they think of next?!!

 

[kyredneck]

What will they think of next?!!

A.I. of course.

 

[John of Japan]

A.I. of course.

How come I didn't think of that!

 

[John of Japan]

Here is a website talking about those early Bible software programs that were available in the 1990's: 1992 Survey of Software for Preachers

I used Seedmaster for many years after it came out. Somewhere I got a file for it of the original Byzantine Textform Greek NT, edited by Maurice Robinson and William Pierpont. While translating, I compared almost all of it with the Scrivener TR that I was translating from, and have those notes in the first draft files of the Lifeline Japanese NT.

So, software has made both translation and textual criticism easier, but it has by no means replaced the human in such disciplines. As I said on that *other* thread, a computer chess program operates by brute force. It calculates many moves ahead tactically, but struggles with making strategic decisions such as: Shall I move h3 (that king's rook pawn up one space) to prevent his possible knight position there three moves from now? (For a chess game in which this occurs, see this one by former world champion Magnus Carlson playing Grushnuk in the Candidates Tournament of 2013, his move #7: https://players.chessbase.com/ecogamesreplayer/Carlsen_Magnus_266?isWhite=True).

 

[Piper]

Here is a website talking about those early Bible software programs that were available in the 1990's: 1992 Survey of Software for Preachers

I used Seedmaster for many years after it came out. Somewhere I got a file for it of the original Byzantine Textform Greek NT, edited by Maurice Robinson and William Pierpont. While translating, I compared almost all of it with the Scrivener TR that I was translating from, and have those notes in the first draft files of the Lifeline Japanese NT.

So, software has made both translation and textual criticism easier, but it has by no means replaced the human in such disciplines. As I said on that *other* thread, a computer chess program operates by brute force. It calculates many moves ahead tactically, but struggles with making strategic decisions such as: Shall I move h3 (that king's rook pawn up one space) to prevent his possible knight position there three moves from now? (For a chess game in which this occurs, see this one by former world champion Magnus Carlson playing Grushnuk in the Candidates Tournament of 2013, his move #7: https://players.chessbase.com/ecogamesreplayer/Carlsen_Magnus_266?isWhite=True).

I totally respect you for your years in the dead soil of Japan and the work you did there.

 

[John of Japan]

I totally respect you for your years in the dead soil of Japan and the work you did there.

Thank you for the kind words.

 

[John of Japan]

Another software Bible program I have used probably 1000s of times is Power Bible 3.3. It's a low end product which you can find here: Power Bible CD – Power Bible. You can get the DVD for just $7!

It's very valuable for Bible translation, and Bible study as well, because of it's very fast and easy search capability. I used it again today while working on a sermon to search the word "householder." Making one of the passages "interlinear" and clicking on the word, I learned the Greek word occurs in 12 verses with several different renderings.

That capability is very valuable to a Bible translator for making sure the renderings are consistent across the translation. I am not advocating translating by concordance (making every rendering the same), but am just touting the usefulness of this tool.

I also copied and pasted into my translation documents for a long time the TR as the version to compare in the software. The font for the TR is just the old Windows "symbol" font with no breathing marks or accents or iota subscripts, but it worked fine for our translation purposes.

 

[John of Japan]

Now a word about word processing. That has advanced tremendously over the years, though I think the "docx" version of MS Word was a needless upgrade.

As I said, for many years I used WordPerfect, which I felt was a superior platform to MS Word, but I eventually had to surrender and use Word, though I did not like it at first.

For most of the translation, however, I used a dedicated Japanese word processor called Ichitaro, which kind of means "number one son." It is an excellent program, and I loved it. The sole disadvantage was that it could save as a Word file, but was not backward compatible. That is, Word does not read an Ichitaro file, which has a jtd extension. So when we left Japan nine years ago, I had to save all of my Ichitaro files in the Word format, which was a lot of work. I previously had saved all of my dedicated Japanese word processor files into the text format which could be read by Word, so I guess you could say I saw a lot of conversions in Japan.

Wikipedia has something about Ichitaro here: Ichitaro (word processor) - Wikipedia

 

[John of Japan]

In the meantime, I'm not sure when, Microsoft did something great: they invented the IME, which is a way to input non-English alphabets. So with the Greek and Japanese IMEs I regularly input Japanese and Greek directly into my documents.

This differs from the previous font methods which I used for Greek. Oftentimes a font written Greek document would lose its formatting and become English gibberish when files were transferred, but IME files keep their formatting in the language they were written in.

As mentioned above, I originally had just the symbol font for Greek, which does not do accents, breathing marks, etc. Then along came some pretty good Greek fonts which I was happy to use, but IME is much better.

 

[John of Japan]

Machine translation is nothing new. Eugene Nida had a whole chapter about it in Toward a Science of Translating in 1964. He wrote, "There is no doubt that eventually machines will be able to take over some of the humdrum tasks of 'low-grade' translating of certain types of material; for example, translating technical documents of a highly specialized nature, in which the multiple meaning of words are at a minimum and literary quality is not required" (pp. 263-264).

Early on, about 15 years ago, I bought a cheap translation program which could do Japanese to English and tried it out. The results were very poor. There were terms the software did not have in its dictionary, and it failed to capture the nuances. Even then you could buy a top drawer program for hundreds of dollars, but I had no confidence in the results.

Now we have Google Translate, which started out very poorly but does much better now. I just ran a few verses through it from the ESV to Japanese, and it handled them, but did not completely handle the nuances. I then ran the idiom "raining cats and dogs" through it and it got the Japanese right, with 土砂降り (doshaburi), instead of being strictly literal.

To my knowledge, there is still not any programs which can handle koine Greek into other languages. Until then, machine translation will be poor.

 

[John of Japan]

I just tried the ESV Psalm 95:2 to Japanese through Google Translate, and it did pretty well with: 感謝を持って神の御前に出ましょう。詩篇をもって主に向かって喜び叫びましょう。It even got an honorific correct.

There is one nuance it missed, though. It gave the name of the Bible book of Psalms (詩編, Shihen) when it should have had simply a word for a song or poem.

Then I tried the same verse in Google from Hebrew, and it could not handle it at all.

 

[John of Japan]

I tried the NKJV Psalm 95:3 through Google translate: "For the LORD is the great God, And the great King above all gods."

It came up with: 主は偉大な神であり、すべての神々の上にある偉大な王だからです. Not bad, except I would use an honorific verb there for the second clause. That's a cultural thing, which illustrates that an AI translation program would have to fully know the culture of the target language. Also, Google doesn't fully handle literary Japanese, Google made an effort with the verb in the first clause, but then missed something. Literary Japanese would have used the verb なる instead of the adjectival particle な.

 

[John of Japan]

I have to say here that the Japanese written language is very complicated, especially compared to English, with its paltry 26 letter alphabet and capital letters. Japanese has two alphabets of 46 characters apiece, with several combinations to make a combined sounds. Then, it uses thousands of Chinese characters called kanji. The government's list of "toyo kanji" (kanji for general use) is 1,850, but many more are used for place names. Besides that, it uses the English alphabet also in documents, calling it Romaji (ローマ字，"Roman Characters").

You can input with two different systems in a Japanese word processor or IME, either with the Japanese alphabet, but then you can't input numbers because the alphabet is so big. I use the Romaji input system, with which you input the sounds with the English alphabet, then hit the space bar to change what you've written into the characters you want.

Example: To get "faith" I type in shinkou. I then get this in the Japanese alphabet: しんじゃ. I then have to hit the space bar and choose which word I mean of the 54 homonyms:
信仰
親交
振興
神鋼
and so forth. The one I want is the first one listed here. Then I hit the return button, and it locks the word in place.

One reason that there are so many homonyms in Japanese is that the Japanese alphabet cannot handle all of the possible phonemes (sounds with no meaning) in Chinese, so it simplifies them for Japanese pronunciation.

 

[RighteousnessTemperance&]

How come I didn't think of that!

Because you fear it may displace you?

 

[RighteousnessTemperance&]

… I then ran the idiom "raining cats and dogs" through it and it got the Japanese right, with 土砂降り (doshaburi), instead of being strictly literal….

This is a good case in point. The AI is able to translate that idiom “correctly,” not because of its “intelligence,” but because a human hardcoded that information into the system.

I just tried it in another language and it did fine with just the idiom ("It's raining cats and dogs.") However, in a simple sentence ("It's raining cats and dogs today.") the program lost its way, translating it literally.

To be really useful, the program should be able to translate idioms either way according to the user's choice.

 

[John of Japan]

Because you fear it may displace you?

Oh, no, let's not go there. (Whistling, looking around anywhere but here. )

 

[John of Japan]

This is a good case in point. The AI is able to translate that idiom “correctly,” not because of its “intelligence,” but because a human hardcoded that information into the system.

I just tried it in another language and it did fine with just the idiom ("It's raining cats and dogs.") However, in a simple sentence ("It's raining cats and dogs today.") the program lost its way, translating it literally.

To be really useful, the program should be able to translate idioms either way according to the user's choice.

Good points. Therefore, a NT translation program must have a comprehensive database of idioms in both Greek and the target language. And there are many, many of them!

 

[RighteousnessTemperance&]

I tried the NKJV Psalm 95:3 through Google translate: "For the LORD is the great God, And the great King above all gods."

It came up with: 主は偉大な神であり、すべての神々の上にある偉大な王だからです. Not bad, except I would use an honorific verb there for the second clause. That's a cultural thing, which illustrates that an AI translation program would have to fully know the culture of the target language. Also, Google doesn't fully handle literary Japanese, Google made an effort with the verb in the first clause, but then missed something. Literary Japanese would have used the verb なる instead of the adjectival particle な.

JoJ, thanks for all the information in this thread.

Having to know a culture is a difficult thing. Culture is a major reason why understanding language requires much more than understanding vocabulary, spelling, grammar, punctuation.

We often don't think about our own language and culture compared to another, and many times don't even know how. Many things are so ingrained that we don't bother to analyze them.

 

[John of Japan]

JoJ, thanks for all the information in this thread.

Having to know a culture is a difficult thing. Culture is a major reason why understanding language requires much more than understanding vocabulary, spelling, grammar, punctuation.

We often don't think about our own language and culture compared to another, and many times don't even know how. Many things are so ingrained that we don't bother to analyze them.

You actually have to immerse yourself in the culture and live there. One cannot understand a culture without interacting with its people.

I have read that it takes about ten years to learn an Asian culture. After we had been in Japan for about ten years, one day I said to my wife, "I think I understand Japan now." And she said the same thing to me. That was after speaking with Japanese for ten years!