Wednesday, June 30, 2010

Dictionaries versus Lexicons

I’ve been kept busy this past week doing some urgent professional translating. I came out of retirement to help some Spaniards going to work in Canada, because Canada and Spain don’t recognise the accreditations of one another’s officially ‘certified’ translators, so documents to be filed in Canada can‘t be translated by a Spanish traductor jurado. That’s not so surprising: the two countries don’t accredit one another’s driving licences either. Protectionism? Mistrust? Administrative lethargy?

To come back to this blog, the June 9 post ended with an undertaking to say something about the dictionaries inside our own heads. I was reminded of it this week by an article listed among the Most Popular Stories Now on BBC News: “Caffeine found in two strong cups [of coffee] impaired word recall.” A reminder that our mental lexicons are subject to the failings and vagaries of memory; printed or computer-stored dictionaries have the advantage that they don’t forget. On the other hand, we have to feed the dictionaries entry by entry, whereas lexicons can learn by themselves.

(Psychologists call our mental word storages lexicons, so I’ll follow their example and keep dictionaries for the external kind.)

There are numerous fascinating mysteries about lexicons, and more particularly the lexicons of bilinguals and hence of translators. For instance, does a bilingual have two separate lexicons or a single lexicon with the ‘entries’ marked for each language? We tend to think in terms of separate word listings, because that’s how the dictionaries are usually organised. But the other way is conceivable, and in fact I used to own an English/French dictionary, the Bellows, that had a single alphabetical listing. As the New York Times reviewer of the first American edition wrote:
This enabled him [Bellows] to save a good deal of space by giving only once… words that are identical in both languages. A further advantage is that in the case of words that are similar in spelling but not in meaning, the reader can glance at both, and get knowledge of distinctions and comparative value, which often he would forego had he to turn to another part of the book.
Psycholinguists, seeking to make discoveries with their new tool of magnetic resonance imaging, are very interested in where a bilingual’s two languages are located physically in the brain. If there are two lexicons, one for each language, do their neuronal representations overlap or are they served by separate areas? An open question.

Another Bellows innovation was to print all the masculine French headwords in Roman font and the feminine ones in italics. Thus gender information was conveyed instantly, without need to look for an m. or f. label. But how is grammatical gender coded in the lexicon? I notice, for instance, that some French speakers, when asked whether a word is masculine or feminine, reply by associating it with the definite article: la dent, le fleuve. In other cases, words themselves have a ‘flavour’ that informs implicitly: for instance, that a word ending in the sound -sion is feminine.

That brings us to two of the most obvious characteristics of the lexicon: it’s sound based and it’s random access. Even illiterate people have lexicons. Yes, there are some special dictionaries that list by sound, notably rhyming dictionaries. But we can ask our lexicon to find us words by sounds that occur anywhere in the word and to do so quickly without having to hunt through a list: for instance, words that include the f sound, and in a twinkling mine retrieved photo, atrophy and cough as well as fool.

Words, however, are two sided. One aspect is their form, their sound or spelling. The other is their ‘meaning’, the concepts they trigger or are triggered by. They wouldn’t be of any use if they were only forms, and they wouldn’t serve for communication if they were only concepts. Words that sound very different are translations of one another in different languages if they share, at least partially, the same concept. But do our lexicons in different languages have each its own store of concepts or do they share a single store? More to come.

A coffee can make you forgetful. BBC News, June 27, 2010.

John Earnshaw Bellows. John Bellows and his French dictionary: issued to mark the 75th anniversary of the publication in April, 1873, of John Bellow's French and English pocket dictionary. Gloucester: Bellows, 1948. The dictionary went through many editions, right up until the mid-20th century.

A new French dictionary. The New York Times, December 17, 1873.

Michel Paradis. A Neurolinguistic Theory of Bilingualism. Amsterdam: Benjamins, 2004.

Photo: BBC

Tuesday, June 22, 2010

A Dubious Involvement

My post last September 12 was about dubious translation for a terrorist trial in India.

Tomorrow is the 25th anniversary of terrorists blowing up an Air India flight from Canada to Londonn in midair. 329 people were killed.

The report has just been published in Canada of a public inquiry into the bombing and the two decades of Canadian botched investigations and attempted trials that followed there. For various reasons, only one of the cabal of perpetrators was convicted.

The one man who was successfully tried pleaded guilty. He was a Sikh immigrant engineer in Vancouver named Inderjit Singh Reyat and he made the bomb. He had also made the bomb that killed baggage handlers at Narita airport in Japan. It came out during his trial that his son Didar and daughter Prit were on his defence team's payroll and did various jobs that were billed to the government at $25 an hour.

And what was the work that they did? Anything from carrying lawyers' bags to court to translating documents.

Timeline: Air India Flight 182, the 25-year search for answers. The Globe and Mail (Toronto), June 22, 2010.

Friday, June 18, 2010

A Song Lyric Native Translation

All translators sometimes need help. It may be for a complete translation or it may be at any one of the stages in the translation process from understanding the original to assessing a translation already made.

Professional Translators have their own forums for ‘crowdsourcing’ their community at times of need. I have recourse, and occasionally contribute, to one called Apart from its forums, has a remarkably useful term and phrase bank that preserves the advice given.

This week a call for help has came through a forum for non-professionals called Lang-8. It’s all the more interesting for being about the translation of a song lyric. Previous posts (April 30, 2010 and September 28, 2009) have mentioned that Sebnem Susam-Sarajeva of the University of Edinburgh is doing research on crowdsourcing and the translation of song lyrics.

The song is sung by a very popular Spanish singer named Alex Ubago (see photo) and it’s available with the lyric on YouTube. The poster of the appeal, a young woman named Asmaa, asks:
I need help with the song's lyrics and translation. I got them from Yahoo Answers [another forum], apparently the poster was a native Spanish speaker. I just want to make sure that he/she got it right.
Here’s the beginning.
E: (With shouts of hope) or something like that lol.

S: A pesar, que la luna, no brille, mañana,
Dará igual, sólo verte reír,
Es lo que me hace feliz, mi alma.

E: If the moon doesn’t shine tomorrow,
It will be the same, only seeing you laugh,
It’s what makes me happy, my soul.

”my soul” is her, like saying, you are my life, he is saying you are my soul in that part.
You can read the rest at

She goes on,
Honestly, I don't feel like I fully understand it even with the translation. Here is the song… Could you please review the Lyrics and translation and tell me if there is something wrong with it?
Which brings us up against one of the elementary norms of all translation, from Native to Professional: it should be understandable. Not necessarily understandable to everybody, but understandable to the readers/listeners to whom it's addressed.

That the translator is not an Expert Translator is obvious from the first line, with its “or something like that”. Expert translators are not supposed to say this, even though they often think it.

Certainly the English could be improved in small ways. But interestingly, even English is a second language for Asmaa. She lives in Egypt and her first language is Arabic. Yet you’d never guess it from her forum post. She explains:
I'm studying English on my own. I already studied it in school but I'm trying to improve my fluency and to think in English exactly like a native speaker does. But I won't pull it off without your help, guys. So please help me with that and I promise that I'll try to be a hard working student.
“¡Muchas gracias, amigos!” is Asmaa’s ending. Maybe you can help.



Alex Ubago. A gritos de esperanza., search for Alex Ubago.

Photo: tucocheesnuestro

Monday, June 14, 2010

Why No Academy of the English Language?

This is a bit off topic, but it’s fresh from the press and it does have to do indirectly with dictionaries. In fact it was one of the most read items in yesterday’s The Observer, so it must have struck a chord in public feeling. In it the author, David Mitchell, fumes against a proposal for an Academy of the English Language with a vehemence reminiscent of the British opposition to identity cards.

The article is over-stuffed with exaggerated imagery. Stripped of all that, what it boils down to is that academies of the language are a kind of sinister ‘language police’ – as the title says – and as such are to be vigorously resisted. We British, it claims superciliously, don’t need them like “less successful languages such as French and Italian.”

Well, Spain has one as well. (Spain also has identity cards.) It’s called the Real Academia Española, and most of the Spanish-speaking Latin American countries have a similar academy. It’s regarded here not as a language police, for it has no coercive powers, but rather as an arbiter in cases of doubt or differences of usage. And like its illustrious predecessor on which it was modelled, the Académie Française, it publishes a dictionary (see image), a normative dictionary, i.e., one that purports to specify what is admissible as 'good Spanish' (and by implication, what is banned or not yet generally accepted). It works hard these days to keep the dictionary fairly up to date and to embrace regionalisms.

That’s not to say that British English doesn’t have language standards and authoritative dictionaries. And there was a time when its most authoritative dictionary, Samuel Johnson’s, was highly personal and prescriptive. But today its source of authority is diffuse, more like crowd sourcing of the well-educated members of its language community, especially its professional writers. And with this, I believe, goes a profound difference in attitude to language.

The countries with the academies belong to a culture in which language is regarded as manipulable by human intervention – at the limit, even by decree. Like a classic French garden, it should be be constantly planned, ordered and trimmed. Whereas native English speakers regard their language as an organic growth that should be left as much as possible to its own tendencies and to natural selection – like an English garden. Of course, now I am exaggerating, but there’s something of that in it.

The situation in Canada provides a neat illustration. In English Canada there’s no institutionalised language authority. Whereas in the main French-speaking region, Quebec, there’s a government agency for language, the Office Québecois de la Langue Française (Quebec French Language Bureau or OLF), and the OLF maintains a large online bilingual technical dictionary. The OLF really does have some ‘police’ powers to regulate the use of French and English in the province.

There’s an old joke that echoes the different cultures.
A Frenchman and an Englishman are each looking for a word in the dictionaries of their respective languages. Neither of them finds it.

“Mon Dieu, what a pity,” laments the Frenchman. “I can’t use it. It’s not in the dictionary.”

“But I’ve heard people say it,” insists the Englishman, “So I’ll use it. There must be something wrong with the bloody dictionary.”

David Mitchell. Snakes are evil, but save your venom for the self-appointed language police. The News Review section, The Observer, 13 June 2110, p. 7.

Real Academia Española. Diccionario de la Lengua Espanola. 22nd edition. Madrid: RAE, 2001.

Académie Française. Dictionnaire de l’Académie Française. 9th edition. Paris, 1992-.

Office Québécois de la Langue Française. Le Grand Dictionnaire Terminologique.

Dr. Samuel Johnson. A Dictionary of the English Language. London: Richard Bentley, 1755.

Photo: RAE

Wednesday, June 9, 2010

Dictionaries and Alphabets

The May 14 post ended with a promise of more about bilingual dictionaries, and an affirmation that learning to use or write dictionaries can’t be learnt naturally like translating. Here’s why.

Pardon my stating the obvious, but dictionaries must remain closed books to people of any age who can’t read. Children learn to speak instinctively, and if they’re bilingual they start to translate at the same time; but they have to be taught to read and many of them are never taught. In developed societies, they typically learn to read around the age of five or six; but bilingual children as young as three can translate orally.

However, there’s much more to it than just reading. Traditionally, the words described in dictionaries of languages with alphabetical writing systems are listed in alphabetical order. The invention of alphabets was of course crucial in the development of Western culture; but the subsidiary invention of conventional alphabetical orderings was also very important. It assumes mastery of the order for each language, which may vary even between languages using the Roman alphabet. Traditional Spanish dictionaries, for instance, use an extra letter (ñ) and count ll and ch as digraphs (combined letters); so a Spanish/English bilingual must learn one and a half alphabets.

Even alphabetic writing systems can be very complicated for dictionary use. Arabic dictionaries, for example, are ordered according to the three, sometimes four, radicals (basic consonants) of a word. Thus, to find the meaning or translation of sytkllmw (they will speak) (the vowels are usually not written in Arabic), one has to be able to recognise and abstract the radicals k-l-m, which, as it happens, are in alphabetical order but in other words they may not be. Recognising the radicals in a word form that is declined (like sytkllmw) requires a good knowledge of Arabic morphology. It took me a year to get to an efficient level of speed and accuracy .

And then there are the non-alphabetic languages, most notably the character based ones like Chinese, for which, in traditional dictionaries, we must know how the characters are built up from their component pen or brush strokes because that’s how the characters are ordered.

Equipped with the above, a primary school bilingual child may be able to find a word in a simple bilingual dictionary. But finding and decoding the information within each entry takes us to another level of complexity.

The highly developed structure of hard-copy dictionary entries for translators hasn’t changed much since the turn of the 20th century. To decode it, one also needs to know part at least of a long list of special abbreviations. However, the first bilingual dictionaries were simple lists of equivalences, and they were very hard copy indeed: baked clay. The writing system was cuneiform and the languages were Sumerian and Eblaite (a Semitic language forming a subgroup with Akkadian). A large library containing such tablets was excavated at Ebla (modern Tell Mardikh), in Syria, from 1974 onwards (see photo).

The tradition of alphabetical listing runs deep. Giovanni Pettinato, the epigrapher to the Tell Mardikh archaeological expedition, wrote (my emphasis):
Among the various lexical texts, a prominent position is occupied by the syllabaries for learning Sumerian, grammatical texts with verbal paradigms, and finally bilingual vocabularies properly speaking in Sumerian and Eblaite, the first such bilingual ones in history. This is perhaps the finest expression of the cultural maturity of Ebla, which has presented us with some works that are the equal of modern culture. Who would ever have dreamed that back in 2500BC Syrian teachers and students passed their time in classrooms compiling vocabularies that Italians would find 4500 years later? I still remember the fateful moment on that sunny afternoon of 4 October 1975 when with vivid emotion I could announce to my archeological colleagues that tablet TM.75[= Tell Mardikh 1975]G.2000 was a bilingual vocabulary. Thus Ebla was not only an economic power but a center of culture as well!

What impressed us in these bilingual vocabularies was the modernity of their layout. Today vocabularies follow the alphabetical principle; at Ebla a similar principle was adopted. Above mention was made of lists whose words are arranged according to acrographic criteria [i.e., according to their initial cuneiform sign]; the vocabularies show the same structure inasmuch as they contain... different sections set off from one another by their initial elements.

There are at least 32 bilingual vocabularies having each Sumerian word translated into Eblaite. One superb example (with 18 duplicate copies!) contains 1,000 words in both languages ― “an inestimable treasure for scholars today, as it was handy for scribes in antiquity,” says Kitchen.

Moreover, there is another basic principle of dictionary and terminology classification already evidenced at Ebla. It's thematic grouping, i.e. grouping by subject field. There are “long classified lists of the Sumerian words for animals, birds, fishes, terms for professions, types of personal names, geographical names (‘gazetteers’), and all manner of objects.” (Kitchen)

When traditions go back so far, we have to wonder about the human predispositions that engendered them.

So much for now about hard-copy dictionaries, the vade mecum of Native and Expert Translators. For Natural Translators, as for all language speakers, there’s another kind of dictionary: the ones we carry round encoded neuronically in our heads. More about them when I have time.

Giovanni Pettinato. The Archives of Ebla, An Empire Inscribed in Clay. Translated from Italian Ebla un impero inciso nell'argilla. Garden City, N.Y. : Doubleday. 1981. Available from Amazon.

Kenneth A. Kitchen. Ebla – Queen of Ancient Syria, chapter 3 of The Bible in its World: The Bible and Archaeology Today. Exeter: Paternoster Press, 1977.


Thursday, June 3, 2010

Panmunjom at First Hand

This is an unexpected appendix to the post of May 24, which should be read first.

Through the good offices of Mariela Fernández (see May 24 post), I’ve been put in touch with Richard Underwood, one of the United States Army interpreters at the Panmunjom negotiations that ended the fighting in Korea in 1953. He’s now in his eighties. His brother Horace, who also interpreted there and was older, died some time ago. In the photograph, Lt. Underwood is in the middle, fourth from left. To his right is CWO Kenneth Wu, his interpreter colleague for Chinese.

This contact is an extraordinary piece of luck. He confirms that most of the Panmunjom interpreters on both sides had not been trained to professional standards, if indeed they were trained at all beforehand. They were therefore Native Translators. He’s given me permission to quote from his very interesting email, so here’s the part that’s most relevant. The emphasis is mine.
I have never been a professional translator. Rather I was a bilingual officer in the US Army and was assigned to interpret for the initial liaison team of UNC [United Nations Command] officers who flew into Kaesong [a town 10 km west of Panmunjom] to set up the particulars for the Truce talks that followed. I always felt inadequate for this task in view of my total lack of training in the Korean language. I had just learned it naturally growing up and playing with Korean children, and surrounded by Korean adults who were literate and sophisticated in their language. My extreme weakness in reading and writing were a terrible handicap, costing me untold hours of sleep as I struggled with translations, aided by a Korean who patiently explained obscure words and phrases.

Soon I was joined by my brother U.S. Navy Lt. Horace G. Underwood. He was 10 years my senior, had studied Korean and the Chinese Characters and was miles ahead of me in all aspects except for fluency and pronunciation, where I was a little better.

The situation improved markedly when Admiral Joy [U.S. Navy, UNC Senior Delegate] invited us to sit in on their nightly staff meetings so that we would have advance knowledge of the gist of speeches to be made.
Richard also elucidates the kind of mistake that apparently cost one untrained Chinese interpreter his life (see previous post).
He continually tried to ‘gild the lily’ of [Chinese] Liaison Officer (Col. Chang)'s remarks. For example: Col. Chang said quite calmly on one inspection trip ‘Here are three of the Chinese People's Volunteers [i.e. Chinese soldiers] killed by your soldiers,’ but Sul [the interpreter] said words to the effect, ‘Here you see, in pools of their own blood these brave volunteers who left home and family to come to this foreign land in sacrifice for the noble cause of our side in this war.’

After he did this several times I turned to Col. Chang (against all protocol, for interpreters exist only to speak for their masters) and asked him if he indeed meant what Sul said, or what he himself had said. Chang glanced at a third officer (who had shown no evidence of speaking English) who gave him a quick nod, meaning I was telling the truth. At that Col. Chang blew up at Sul and ordered him to go – get out of my sight. My analysis of the situation is that Sul, who had been a ‘mole’ HS [high school] teacher in Seoul before the war, was simply trying to be super patriotic to ‘prove’ his loyalty to the North.

Photo: Walter G. Hermes. Truce Tent and the Fighting Front: The Last Two Years. Washington, D.C.: Center of Military History, United States Army, 1990. Fascinating details and richly illustrated. It's online at