Discrete combinatorial system

Language is the quintessentially human activity. It is quite amazing yet so natural that we don’t notice how fantastic it is. It is built on two simple yet profound facts.

The first is that we randomly assign meaning to sequences of sounds. The word duck does not look like a duck, walk like a duck, or quack like a duck, but it means “duck” all the same. It means duck because we all think that it does. Other people think that canard (French) means duck; or idada (Zulu), or oli (Korean), or thousands of other words, but English speakers know that duck means duck. Anything could mean duck, but only one word actually does.

The second profound fact is that we combine words together in an almost infinite number of ways. Language is a kind of “discrete combinatorial system.” It is discrete because it is made of a limited number of discrete pieces. English has only 26 letters in its alphabet and only a few thousand words. (The Oxford dictionary lists about 200,000 words, but the average person knows between 20,000 and 40,000.)

Language is combinatorial because we combine the letters and words together and different combinations give different results. The meaning of “man bites dog” is different from the meaning of “dog bites man” in spite of the fact that the two sentences contain exactly the same words and letters. Our limited set of letters and words can be combined into so many sentences that you regularly say sentences you have never said before. With some regularity, you even say things that have never been said before by anyone.

In his book The Language Instinct, Steven Pinker wrote:

Go into the Library of Congress and pick a sentence at random from any volume, and chances are you would fail to find an exact repetition no matter how long you continued to search.

The number of potential 20-word sentences is so great that if you said a different 20-word sentence every five seconds it would take you 100 trillion years to say them all. When you finished you could start on 19-word sentences, or 21, or…

God gave us complexe language so that we could communicate with each other and with Him. It would be a shame to neglect cand tell God about something that bothers you, or that you are thankful for. You never know, maybe you’ll say something to God no one has said to him before.

Language is impossible

Noam Chomsky is arguably the world’s greatest linguist. He has done more to advance our understanding of human language than anyone alive today. He contends that the greatest challenge to understanding language is understanding its infinite creativity.

Knowing a language, says Chomsky, means being able to produce an infinite number of sentences never spoken before and to understand sentences never heard before.

It is a fact that you say sentences you have never said before with regularity. A surprising number of them have never been said before by anyone. Yet the people around you mostly understand when you say something never said before. That’s rather astounding, if you think about it.

I use this feature of language when I read to my granddaughter. To spice things up, I stop and ask her crazy questions. If there’s a bear in the story, I might ask: “Is the bear going to come in the house and climb in the freezer and go to sleep?” Even though she’s only 4, she immediately understands this question she’s never been asked before. I’ll bet no one has ever asked you that question. Yet you understand it. My granddaughter will simply answer no, or say I’m being silly, or tell me no more questions, just read the story. The never-asked-before question doesn’t throw off her 4-year-old brain, not even a teeny, tiny bit. (That last sentence has probably never been written before.)

Knowing a language means being able to produce an infinite number of sentences never spoken before and to understand sentences never heard before.

We take the God-given and breathtaking creativity of language in stride. We don’t even notice it. But those working on artificial intelligence (AI) notice because it’s a huge problem for them. To be realistic, a robot would have to regularly produce never-said-before sentences and understand them when others said them.

I’m going to go out on a limb and make a prediction. AI won’t reach that level, ever. Well, at least not for a very long time and certainly not without a revolution in computer technology. I’ll make a shorter term prediction. No publisher will use computer translation for its best selling novels in any of our lifetimes. The human translators they hire will use computers to help them. AI isn’t even close to good enough to translate artistic or elegant prose, let alone poetry.

God is the ultimate creator, but he gave us some of his ability to create. Everyday, you exercise that creativity effortlessly when you speak. Every person does, even the uneducated and illiterate. In fact, we can’t stop being creative with language. That’s one part of what it means to be made in the image of God.

Source: unknown

Missionary technology

The first commercially-available computers could only display or accept English. That was a problem. By the time I first started using computers, we had a few more characters available because of something called “extended ASCII”. This allowed the computer to display and accept keyboard input for most European languages. But it still didn’t work all that well. Specialized technicians had to fiddle with the computer to get it to accept and display the characters in the alphabets of African languages we were working on, But every language had its own system, making archiving and computer support a mess.

Technical details for unicode for one specific language

Fortunately, the tech companies wanted to sell their products everywhere, so they were interested in solving this problem. Missionary-linguists got involved in a worldwide consortium working on the issue. We jumped in so that the smaller languages wouldn’t be left out. Besides, we were often the only ones who had thought about what they needed. In the end, we got unicode; a world-wide standard for accepting, displaying and printing all the characters of all the languages of the world, even Tai Lue, wherever that is.

Your smart phone has unicode, your computer has it, your TV has it, maybe even your car and your refrigerator. Someday your doorbell might have it. Actually, I think some already do. You use missionary technology every day. So do atheists.

Now, anyone who wants to read the Bible in his language, no matter how strangely it is written, can see it displayed on a smart phone, tablet or computer. Because of unicode, the Bible in any language can be sent across the internet or put on small chips and carried anywhere. Whatever electronic device receives it will display the strangest characters correctly. Unicode, hidden the background, makes it happen automatically.

Cubes

My sisters refrigerator makes ice. Two kinds of ice, no less. The display panel offers cubed or crushed. I chose cubed. This is what I got. I don’t know if this three-dimensional shape has a name, but if it does, it’s not a cube. Yet this is an ice cube. And I was not at all surprised, disappointed or indignate when I selected “cubed” and got this. I didn’t feel mislead or duped. I’m perfectly happy with my ice cubes that aren’t cubes at all. The ice cube trays (There’s that word again!) of my youth produced ice in the shape of cubes, roughly speaking. I suppose that’s how chunks of ice made for human use came to be called cubes.

How did it happen that we are happy to call something a cube that’s not at all a cube? Well, “ice cube” ceased to be a combination of two words each with its own meaning and instead developed one meaning. We still write it like two words, and we can take it apart into two words, but the two words together have one meaning. We recognize ice cubes that are not even remotely shaped like cubes proving that “cube” has lost its independent meaning when found in the combination “ice cube”. There are lots of such cases, like “the White House”. It’s still white, but it’s quite unlike most houses.

The correctness of considering ice cube as one word is shown by the fact that in other languages it is one word. In German it’s a compound word (like icecube), but in French it is one word (glaçon) that is not composed of two parts. Glaçon is derived from the word for ice but it has nothing to do with cubes.

When translating or interpreting the Bible, we can’t pull phrases like “ice cube” apart to determine their meaning. In fact, we should be a bit sceptical when preachers make a big deal of subjecting words and phrases to great scutiny especially when the result is a novel interpretation. God spoke to us in ordinary language which is subject to ordinary understanding. So don’t think you need deep scrutiny of words to understand. Or you might be like someone convinced by fancy reasoning that my sister’s refrigerator doesn’t make ice cubes. Or that a quarantine is not a real quarantine unless it last 40 days.

Language and Ideas

A couple years back I attended a public lecture in Accra. It was given by a Ghanaian academic. The thing is, he is also a village chief. One of the things that is different about Ghana is the fact that a number of chiefs are highly educated. They might be academics as this man, or high level civil servants, or business people. Apparently, some villages in Ghana want their chief to be able to interface with the outside world.

At one point, the lecturer talked about language. He said that some of his subjects come to him speaking English and say things like “Your Highness”. He responds by asking them to say that phrase in their language, but they can’t. He said that if one of this subjects cannot translate things they say in English into their language, this is a good sign that they really don’t understand the thing that they are saying.

I have actually seen this first hand. I was helping a young man study for an exam he had to pass to get a diploma. As is the case for many Africans, he was studying and learning in the official language which was not his mother tongue and he was learning that language as he learned the other subjects. He would read and study in the official language, then come to me with things he didn’t understand. One day he came to me with a question in a sample exam. He knew the answer, but there was still a problem. He said that he knew that the answer was “The earth turns on it’s axis every 24 hours.” The thing is, he said that he had no idea what that meant. He knew that he could go to the exam an get questions like that right, but he also knew that he really wouldn’t understand what he was writing. As I started explaining, I found that I had to back all the way up to the fact that the earth is a sphere. I think, or at least hope, that he went away knowing what his answers meant.

Because many people go to school and learn in the official language which they don’t speak at home or in their community, and they are learning that language as they learn other subjects, then can learn to say something that they don’t understand, misunderstand or only partially understand. So it is possible to carry on an intelligible, coherent conversation with someone and find out later that they didn’t understand it, or understood it differently than you did. People can learn religious jargon without understanding it or without understanding it well. If you ask then to say the same thing in their language, they can’t. If someone doesn’t know how to say “Jesus is Lord” or “human rights” in their language, that’s a sign that they might not really understand the phrase when they say it in the official language.

Mind you, there is a layer of educated people who understand perfectly what they say in the official language. But that layer can be very thin in some places.

The lecturer-chief is on to something. He saw clearly the limitations of the official language. The lecture was in the official language, so obviously he thinks that it has it uses. But he is not naive enough to think that “Everyone speaks English”, because he knows that some who appear to speak English and who think they do, really don’t. We can’t build solid Christians and churches on a language people only seem to master.

Bibleless Peoples part 2, language myths

I recently read “The Narrative of the Life of Frederick Douglass, An American Slave”, Frederick Douglass’ autobiography. Douglass wrote it in 1845. It contained a number of words I did not know. Take this passage, for example.

His carriage-house contained three splendid coaches, three or four gigs, besides dearborns and barouches of the most fashionable style.

I had to look up gigs, dearborns, and barouches, having never had occasion to be around or talk about these types of horse-drawn wagons and coaches. I knew of covered wagons and stage coaches because I have seen those and heard them called by their names, but everything else was just a wagon to me.

Choir in Congo singing AIDS prevention song in their language

That brings me to bibleless peoples. They are often accused of having impoverished, substandard or unscientific languages. These accusations are bolstered by pointing out that the languages do not have words for modern things. But in Frederick Douglass’s day, even English did not have the word “microbe”. Was his language therefore impoverished? Was mine impoverished because I didn’t know barouches? Not having words for something is a silly way to judge someone’s language. Many common words today were unknown to Shakespeare and it is ridiculous to call the English he spoke and wrote impoverished or substandard. In fact, such accusations are misplaced for any language because languages adopt or invent new words for the things they encounter. Vocabulary is a function of context rather than a sign of adequacy or inadequacy. All languages can and do develop – acquiring new words as needed. Did you know that “vegan” wasn’t invented until 1944 and remained obscure for decades after that? When we were involved in translating AIDS information into Congo languages, we had to find ways to say HIV, AIDS, seropositive, virus and many other things. Those doing the translations always found good translations of all the terms and even had them approved by medical professionals.

Unfortunately, sometimes bibleless peoples believe the negative statements made about their languages. They can even believe that the lack of certain words will make it impossible to translate the Bible. Sometimes we have to convince church leaders who fear that the Bible will be degraded by translating it into “substandard” languages. It’s all part of mobilizing churches in Africa for Bible translation.

But when the translation produces the joy of salvation and the fruit of godly living, the language myths are often dispelled. Besides, dare we call any language “impoverished” in which God speaks to people?

Publications in many topics in Ghanaian languages

Long tail phenomenon

The long tail phenomenon was first described by Chris Anderson who has written a book about it. It’s gets its name from the shape of the graph that is created when you chart certain kinds of data. Mr. Anderson uses it to describe and predict how products sell online. But it turns out that lots of things have long tails, including the languages of the world. If we chart languages by the number of people who speak them, we get a long tail.

This is because there are a very few languages which are the mother tongue of lots of people. They form the spike on the left of the chart. Then there are lots of smaller languages. With just 10 languages, you can reach over 75 percent of internet users. The remaining 25 percent use the Internet in thousands of languages.

There are 7,097 languages spoken in the world. Just 23 of them are the heart language (mother tongue) of half the world’s population! They form the spike on the left. The other half of the population speak 7,074 languages. At the end of the tail are 467 languages spoken by a total of 12,758 people. That’s 27 people per language on average. These are mostly dying languages spoken only by a few older people. In fact, the tail gets so thin that I had to exaggerate its thickness in the graph to get it to show at all.

It’s the languages in the middle that come into focus for Bible translation. The biggest languages already have a translation and dying languages don’t need one. But about 1 billion people speak languages that are not dying and which do not have the Bible. The average size of their languages is 220, 000. A very few are large but many are smaller precisely because of the long tail phenomenon.

Several passages in the Bible, including the parable of the lost sheep, tell us that God cares for those in the long tail.

The Lord did not set his heart on you and choose you because you were more numerous than other nations, for you were the smallest of all nations! (Deuteronomy 7:7 NLT)

Let the red reduce

I was with some Ghanaian colleagues presenting Bible translation at a Christian College in Accra. After the presentation, the dean of the college was taking an offering. He told the students “Let the red reduce”. This sentence is an example of implicit information. When something someone says or writes relies on information that is not directly expressed in what they said or wrote, then the meaning of their words depends on implicit information. In this case, the implicit information is behind the word “red”. Here’s how. The money of many countries has different colored bills for different denominations. Ghana is one of them.

Here are the bills, so that you can interpret the dean’s comments for yourself.

Just like all language, the Bible also contains passages that imply information that is not found in the words themselves like this one:

But when the fourteenth night was come, as we were driven up and down in Adria, about midnight the shipmen deemed that they drew near to some country; And sounded, and found it twenty fathoms: and when they had gone a little further, they sounded again, and found it fifteen fathoms. Then fearing lest we should have fallen upon rocks, they cast four anchors out of the stern, and wished for the day. (Acts 27 ESV)

The fact that the water was getting rapidly shallower confirmed that they were approaching land, but the text does not state this obvious fact. Approaching the land at night is dangerous, but the passage doesn’t say that either, although it does say that the sailors are concerned about shipwreck and take action – they drop anchors. The anchors are designed to stop or slow the ship, another piece of obvious and therefore implied information.

What if you were translating for a landlocked people who had never seen the ocean and had no lakes? Would they know that the decreasing depth of the water meant land was approaching, that this was dangerous and that dropping anchors would slow the ship and help prevent disaster? For them, this passage might be as obscure as “Let the red reduce” was for most of you.

The Dean also said, “Give me brown, I will smile.”The Dean’s comments are analogous to an American preacher encouraging people to put fewer Georges in the offering and to even throw in some Bens.

By the way, we used to use the Acts 27 shipwreck passage in seminars on advanced translation principles in Burkina Faso – a landlocked country where most people have not seen an ocean or a lake large enough to navigate with a ship.

IPA

IPA is the International Phonetic Alphabet. It contains a symbol for every sound made in human speech. If you master the IPA, you can literally write down what anyone says in any language whether you understand the language or not. Not only is there a symbol in the IPA for every sound made in any language, the IPA contains the definition of how that sound is made by the human articulatory apparatus otherwise known as your mouth – well actually a bit more than your mouth. Every sound in human speech can be defined by the position of the parts of the articulatory apparatus. Does the tip of the tongue touch the alveolar ridge producing t or d, or does the back of the tongue touch the palate producing k or g? Do the parts touch and briefly stop the flow of air completely, as happens when you pronounce t, d, k or g? Or do they merely restrict the air flow as happens for s, f, sh and th. Are the lips rounded or not? Do the vocal cords vibrate or not? Does the air come out of the mouth or out of the nose as it does with n, m and ng?

All of this is taught in courses on articulatory phonetics and it is described in detail in textbooks. We have abundant and widely-available knowledge of the way the sounds in human speech are made. One of the foundational books on the subject was written by a Bible translator, Kenneth Pike. The process of translating the Bible starts with someone who knows the IPA sitting down with someone who speaks the language to write down words and phrases in the language using the IPA. It sounds like magic – writing down a language that has never been written – but its all described in the IPA and the books about it.

Congolese man telling of unsuccessful attempts to write his language

Without this bit of human knowledge, writing a language for the first time may prove impossible. I remember this older man in Congo saying that every since he was a child his people had been trying to write down their language without success. Oh, those who had been to school in French (the official language and the one taught in school) could write, but they couldn’t really figure out how to write some words. Also no one could read what they wrote. After a few weeks, not even the person who wrote the words could make sense of them. Decades went by. Then a few months work by a missionary trained in descriptive linguist and the problems were fixed. The old man said he was thrilled, and that he was finally confident that the translation could now move ahead. In fact, with the training given to the local translators, he said that the translation could succeed even if we missionaries left.

This was not the first time I have heard Africans tell of their repeated failed efforts to write their language. Some have even mistakenly concluded that their language could not be written. Then, decades later, a missionary linguist solved the problem in relatively short order. This happened so often in one area that I heard leaders of the biggest church in the area tell churches in other areas to be sure and ask for missionary linguists.

Bible translation is a spiritual ministry, but the science of linguistics sure helps, especially when it empowers local people.

Small languages: Part 1

The August 9th is the International Day of Indigenous Peoples. So I’m going to post some blogs about small languages.

Sometimes, people ask me how big a language has to be for us to translate the Bible into it. You may be surprised at my response – how many people speak a language is not an important criteria for whether we translate into it. Don’t get me wrong, it is a valid criteria, just not a very important one.

Language Vitality in Africa

That is because other criteria are more useful, especially the criteria of language vitality. Language vitality asks the question whether the language is being passed to the next generation, in other words whether there are signs that it is dying. To understand this, let’s imagine a situation that is and has been quite common in the USA. Say Swedish immigrants, a married couple, arrived in Minnesota in the early 1900s. In their home, they speak Swedish, but they learn English through contact with their neighbors and others in the community. When they have children, they continue to speak Swedish in their home, but the children quickly learn English through their friends and at school. In fact, the children speak English better than their parents. As the children graduate from high school and move out of the home they speak Swedish less and less, perhaps only when they visit home. Then the children get married. One or two may marry the children of other Swedish immigrants in the community, but others have spouses who do not speak Swedish. In any case, the couples speak English together, not Swedish. So when they have children, they speak to them in English. So the grandchildren of the immigrants no longer speak their language.

This imaginary story shows a typical case of the interruption of transmission of a language from one generation to the next. This process typically takes three generations. While my imaginary story concerns one migrant family, the same thing can happen to a whole community without migration being a factor. The same process can be found in communities of Native Americans where one generation speaks the language at home, the next learns the language at home but has as much or more contact with English and starts using English as its preferred language, then the next generation does not learn the language from those parents. Or they may learn only very limited parts of the language.

Language Vitality in North and South America

So, a crucial criteria for translating the Bible into a language is the language’s vitality – whether the language is being transmitted to the next generation. A simple survey can determine if the language is being passed to children in the home. When we know that, we can project the number of people who will speak it in 30, 50 or 70 years. If that projected number is increasing because of population growth and children learning the language in the home, then a translation might be warranted even if a smaller number of people speak it today. On the other hand, a language with more speakers but low vitality and hence a projection of decreasing numbers of speakers, might not get a translation. Vitality is more important than number.

While I was in Côte d’Ivoire, language surveys were being done to assess language vitality and other relevant factors, so that resources for translation can be allocated wisely. While some languages in Côte d’Ivoire have low vitality, most of them them are alive, well and growing.