Monday, August 31, 2009

Employment, History

He had been a precocious child. An intellectual. At twelve, he had translated the poems of T.S. Eliot into English, after some vandals had broken into the library and translated them into French.

(Woody Allen, Without Feathers)

Been born to Bologna 3 January 1952. Married, daughter has one. Graduated in psycology, she is journalist professional from 1979. National secretary of the Forehead of the Youth in 1977, is elect deputy for before turns the 26 June 1983.

(Excerpt from the biography of Gianfranco Fini, then deputy Prime Minister of Italy, briefly published in 2001 on the website of the Italian Government)

Some time in the next few months it will be twenty years since I completed my first paid translation. I picked up those initial jobs rather casually, as a time filler during the interminable hours spent in an office where I was supposed to answer a phone that never rang. Ironically enough, the material I was given to translate concerned early cell phone protocols and preluded to a technology that would make asking a school leaver to man the phones of an otherwise deserted office entirely redundant.

The mobile office had yet to be invented, and my place of employment hadn’t switched to the then primitive version of Windows either, so I worked in WordStar for DOS. Here’s a screenshot, but please be warned that if you are old enough to have used it, clicking to enlarge might be triggering.

Wordstar was a ‘what you see ain’t what you get’ type of word processor, meaning that you had to tag the portions of the text that were to appear in, say, a bold typeface and then, just like that, they wouldn’t look bold on the screen at all, but rather you’d have to keep track of the tags and make sure they were all open and closed in all the right places, or the printout would look all funny. Then at the end you got to save your efforts onto one of these bad boys.

This 5¼-inch floppy disk could easily pack as much as 1.2 MB of data.
It seems like such a primitive setup, doesn't it? Still, it was electronic writing, and you could revise whole paragraphs on the fly without leaving a trace of how poorly you might have expressed yourself in draft form. Outside of school assignments, I have never experienced the act of translation when words couldn’t be recombined at will in that manner. To me, it’s just part of how things have always worked, but for professional writers and translators the transition must have been quite dramatic at the time of its occurrence. I have in fact just learned from Wikipedia that William F. Buckley Jr. loved WordStar so much that he kept installing it on new machines until long after it had become obsolete, and once had this to say: “I’m told there are better programs, but I’m also told there are better alphabets.”

Did you know that banana bunches are sometimes wrapped in bags for protection? Image by Fairsing, from Wikipedia.

I’ve been a translator for nigh on twenty years, and besides the obvious aspects of the job - the writing discipline, what it can teach you about communication, culture and language - I’ve enjoyed it for the insight it has given me into other professions. This is something I have very much in common with my father, the interest in tools of various trades but also and primarily in how people organise in order to carry out certain tasks, the intricacies of which you’re unlikely to discover as an outsider. Translating a company’s internal documents, or the contracts that regulate certain arrangements, or the manuals for this or that working procedure, will go some way towards satisfying those particular curiosities.

Often these documents are written at a time of transition, when a new system is being put in place, and remind me of the history and trends of my own profession. In those early days of sitting in front of Wordstar in a lonely office, I researched my translations through technical encyclopaedias and dictionaries, or by asking experts in the relevant field who happened to be available to my employers, and I delivered my work in the form of printed documents. Then came better word processing, automatic spellcheckers, the delivery of electronic files, the World Wide Web and computer assisted translation - each of them an incremental step in what has been a radical transformation of the industry.

Translating a text about, say, the cultivation of bananas, may still entail the same linguistic skills and fundamental understanding of how to produce a text in the target language that is as close as possible in content, style and feel to the source; but the manner in which these texts are exchanged, and the ways in which the translations themselves are researched and produced, have changed enormously. The initial stages of this transformation, with the perfecting of electronic word processors, could be said (although not by me) to have done little more than make us faster and possibly a little less error prone. But the second phase, which began when the Web reached critical mass in the mid- to late nineties, has had effects that are rather more dramatic.

In my capacity of freelance translator, I work with clients in different parts of the world who sometimes work in turn for other clients in other parts of the world, all from the relative comfort of wherever I happen to be at the time. This means that I can take my job with me when I go to Italy, which is a significant advantage in terms of my ability to remain connected with my family back there; but even that is really not nearly as significant as the fact that I can be a full-time Italian translator living in New Zealand in the first place.

As a professional group we have been scattered around the globe, and are now competing with one another globally. At the same time, and contrary to some early predictions, the Web hasn’t accepted English as its lingua franca, quite the opposite: the credibility of many businesses and organisations has been staked on their ability to offer accomplished and appropriately localised multilingual translations, an aspect of globalisation that is sometimes overlooked. Even open source software and independent media organisations often rely on the efforts of volunteer translators, and the ability to operate in several languages is considered integral to their philosophy - it’s one of several ways in which people on the Web are working to counteract the tendency towards cultural homogeneity inherent in the medium. More and more, the token gestures that were common earlier in the history of the Web are either frowned upon or laughed at, as was the case when the second Berlusconi government published the biographies of its Ministers in a crude Babelfish translation that had to be taken down in a matter of hours (but not before several of us had saved them on our computers, thankfully - that stuff was pure gold).

The geographical scattering is aided by the fact that one no longer needs to reside in the country where a language is spoken in order to access the finest expertise in that language. I took with me three very large dictionaries and a small encyclopaedia when I moved to New Zealand ten years ago, but all of these exist now on disk; more importantly, technical vocabulary that used to require days of research can now in most cases be located by means of a well constructed Google search or, if you're still stumped, by asking other translators or experts in the relevant field who can be easily reached online and are invariably more than happy to assist. With all that help so readily available, having to go a bit further out of my way to nail down the exact Italian word for an obscure piece of machinery or procedure or naturally occurring phenomenon remains, I'll nerdily admit, the source of a professional buzz followed at times by embarrassing smackdown-style exclamations.

So far, so relatively obvious - translating is hardly the sole profession to have gone through these kinds of changes. But some of you might not be familiar with the concept of computer-assisted translation, and the impact it's had on our work. I'll refer in particular to an aspect that is of obvious relevance here, namely translation memory software.

Translation memories are files that store the text entered by a translator against the relevant segments of the source document. Here's what a step in that process might look like.

Created in Wordfast.
The next time that in the course of your work you encounter that exact sentence, or one similar to it, the software will automatically insert your previous translation in the target segment, with a percentage and a colour code indicating how closely it matches the old source. Then you can amend as needed or simply validate the proposed segment and move on to the next bit of text.

Translation memories have been around for a while, but their emergence as a key tool of the trade dates back to the late Nineties. It was at that time that I started using Wordfast, following the recommendation of a colleague, and between then and now I have entered something in the vicinity of three million words in my database. The Woody Allen example above is actually not a very good one in that literary translations are not an area where you're going to find translation memories very useful. But think of technical documents, legal contracts and in general any text that has a large degree of repetition and where consistency of terminology is key. Think also of the needs of an organisation working on a large project that has to be split amongst several translators. This is how we do it, and the benefits in terms of productivity are very significant.

And how do I feel about this?

If you know me or have been following this blog at all you'll know that I'm prone to criticise the semantic slippages that allow the word memory to be applied to something other than the human capacity to remember things, and in ways that obscure what constitutes knowledge and value in favour of the much more easily quantifiable and less problematic category of information. Saying for instance that the Library of Congress contains X terabytes of information is a sure way to get my facial muscles to twitch. But in the case of translation memories and, worse still, my own translation memories, I find it very hard not to concede the point.

Too many times I've approached a text that looked completely new to me only to discover via this other 'memory' of mine that I had translated a nearly identical one five years earlier; too many times I've seen those results pop up in rapid succession, and have yet been unable to recognise anything other than the writing style and idiosyncrasies that I'm not even to supposed to allow to seep into my work; too many times I've seen my computer translate entire chunks of documents thanks to its superior knowledge of my past work. Too many times I have experienced all of these things not to recognise the extent in which not just my past translations - after all, that would be no different from saving on file or committing to paper any old piece of writing - but the method, the style, the approach I bring to my work have been successfully externalised and replicated. It is a form of memory, in that from a collection of words and symbols you can make out meaningful connections, recognise a form of autonomous intelligence. Mine, and yet not mine.

Most soberingly of all, the entire database - the three million words of translation and the glossaries I've compiled over a decade - occupies a mere 160 megabytes. This means that, should I stick with translating until the day I retire, all of my precious words, the definitive memory of my working life, would still most likely fit onto one of these:

Serves me right, I reckon.


Deborah said...

I used Wordstar when I wrote my Honours dissertion at Otago (the Accounting one, not the later Philosophy one), before shifting to some program on the Vax (?) where we used some kind of mark-up language with opening and closing tags and the like. And stored it on those old floppy disks.

It hadn't occurred to me that professional translators would have software that gets better, or at least bigger, with use. It reminds me of The Practice Effect, by David Brin.

Giovanni said...

Heh! I had totally removed The Practice Effect. I should try to reread it to see if it's improved over time.

It's customary amongst translators to fret that all this help we're getting might prelude to a leap of technology that makes us redundant altogether. I don't worry so much about fully automated translators based on the principles of AI, but if you were to feed a lot of translation memories into an expert system you could be getting somewhere. And there already are systems in place to build collaborative super-TMs.

merc said...

At a place I worked as a Tech writer, I had to write for 12 localisations (translations)...every word stalked me, and i it, writing in the vernacular was verbotten; a word like "handy" could cost 1000's of dollars.
Whole Govts. write in this faceless language, especially the UN (a Govt?). I went mad, again.
The Poesy survived, intact but has taken some years to grow back to approaching the sensual.

wordverf; fixasm a steady gaze whilst excited.

artandmylife said...

I used WordStar, but then I used dBase software before it even had a number. Enough said on the subject of aging here I think

Paul said...

I worked at a language school for a while; the Asian students would use pocket translation devices, to find English words to express their thoughts. The results were quite peculiar, sometimes.

If only I could find software which would produce phrases like "constitutes the Alloy North," I would be able to write lyrics for Progressive Rock bands.

Giovanni said...

@Merc: Yes, the tiranny of the words you must use, and those that you must not - you'll either put it out of your mind or go mad. These days when I translate the phrase "in these tough economic times" my brain barely hurts.

@artnandmylife: *gasp*! I came in at db3, maximum respect. Gave myself flashbacks checking out the interface. I see they still make it, though, unlike WordStar which has become a piece of abandonware (new word I learned this week, I like it very much.)

@Paul: The Alloy of the North is a personal favourite, although the transformation of the fascist youth front in “Forehead of the Youth” remains a masterpiece of black humour. I solved a few formatting issues and pasted the whole text on this page. Enjoy.

rob said...

Wordstar- yep, that picture brings it back.
Wrote a master's thesis on it, on our first computer, a reject from the journalism programme my partner was on.
It ran an OS on one 5 1/4" leaving another 5 1/4" drive for Wordstar and the text. When I filled the disc, I had no idea what to do. Pulling either disc out wasn't an option.
Somehow I managed to copy wordstar onto the OS disc, and then onto another floppy.
Printed the wretched thing out on a dot-matrix printer, two copies, up all night, tearing the cheap paper at the perforations as the sheets slowly rolled out.
It was only just readable. Beautifully bound (a friend who was a bookbinder) by now both copies must have, mercifully, faded to illegibility. One day I'd like to retrieve the copy held in the Dept library, and write something fresh on the blank pages.
Wordstar. The name carries with it some of the panic as I tried to change discs, and the stress of that sleepless night.

Paul said...

I had a Word Processor, by which I mean an assistant - a human being - who typed all my letters (which I wrote with a fountain pen on a huge machine which I never touched; then she would return them to me for my signature.

Good times.

Giovanni said...

This is going to turn into Four Yorkshirmen any minute, isn't it?

Love the story about the two full disks, neither of which could be removed from the respective drive without undoing the work.

merc said...

Ok, ok, I found a Gannet primary feather on the sand, made a quill of it, ground some charcoal with stream water for ink and wrote a poem on a dessicated seal bone...true story.
Years ago though i did work in advertising pre-crash for a multi-national on a HUGE account, I can't even begin to tell you what the typists did with my garble...heh, I was 23 and i thought everyone drank Champagne for lunch...

wordverf; "turrun" or not to run, that is the question.

George said...

I'd be surprised if machine translation wasn't standard within ten years for major languages. With the use of huge parallel corpus, things are progressing very quickly. It won't be perfect, but for 95% it will be "good enough".

Captcha: ingism - a phrase that can only be captured with badly translated English.

Giovanni said...

If it happens, I'll console myself by recycling this post with a slightly amended title: Unemployment, History. Think of the savings.

merc said...

Don't worry Mr. Gio, from what I have seen of them, the machines have no Deus, and they don't want to translate for us anyway.
Bad machines.
wordverf; "phompi" acid prayer = circumference of square.

Daleaway said...

Oddly enough it's the photo of the bananas that brings back memories for me - I have some similar photos I took in Queensland 45 years ago. At the time we were told it kept spiders and snakes and weather out of the ripening bunch.

George said...

Giovanni, are you familiar with Ethan Zuckerman? He's written some interesting things, including on translation.

Giovanni said...

@Daleaway. Nice that it brought back memories, I was quite struck by that image when I researched that particular translation. Now apparently they have high tech banana bags, sun filtering and laced with insect repelling stuff.

@George. Thank you, great link. One of the aspects of the Google settlement that is getting less coverage is the potential that opening such a huge corpus would have for linguistic research and, yes, machine translation.

harvestbird said...

Upon this theme I recommend the recent Three Novellas for a Novel by Carl Shuker, which was released online in limited edition. It concerns a man who works for a machine-translating company in (near future?) Japan. The second novella of the three is a machine-narrated version of some horrible events with historic precedents. There's also quite a lot of amnesia and bafflement. It's a tremendous work.

If anyone would like access to a copy, let me know and I will share mine.

Giovanni said...

Oh, pick me please. The idea of a limited digital release is intriguing in and of itself.

rob said...

I'd love to have a go at it, too. Is it small enough to email?
Sounds a little like Mishima's 'Sea of
Fertility'- maybe it's just the idea of linked novellas.

katy said...

Every week for about three of the years I lived in Tokyo I met with a computational linguistics researcher for chat and occasional editing of his papers. His thing was HPSG parsing which is where they try and teach the computer to learn grammar, as opposed to programming the computer to apply grammatical rules. His first love was computers, rather than language, and he came to computational linguistics via the understanding that "teaching" a computer to use language was probably the greatest challenge they had. It was a fabulous period for me with the very real pleasure of trying to understand his papers (and spot any dodgy use of articles).

On Google as a corpus, in my job I only use English which I have a pretty good grasp of. However, 80% of what I do is writing and I find myself Googling expressions on a very regular basis to see whether I am using them in the accepted manner. What did anyone do before Google??

Giovanni said...

It can be a bit of a trap, though, no? Google "millenium" and you'll get over 14 million results. We could discuss whether the fact that so many people get it wrong means that actually it ought to be considered right, but you're likely to get pinged for it by a more traditionally minded editor if you choose to trust Google on that one.

Paul said...

Googling "going forward" got me 6,410,000 results. Conclusion: this phrase is so popular that it must be good English.

Giovanni said...

It is a well constructed adverbial phrase and it stands to reason that platitudes should be well represented in a corpus. But spare a thought for the translator, who has to come up with an equally trite phrase in the target language.

Some of my colleagues will tell you of the nightmare of having to deal with technical translations in obscure topics. The empty phraseology of marketing and PR is to my mind is far harder to work with, and not just because of the soul crushing aspects.

merc said...

Going forward = Veni vidi vici?
On having a soul...a Surgeon (cap S) once said...
"I have cut many bodies open and I have yet to see one."
Me personally, avoid the S word like a plague...(translate that in 12 languages why dontcha!).

wordverf; antioust, "yes" is definitely off the cards in Franglais.

Di Mackey said...

Thank you. I have always, but always, wondered about translators. I love when I find a good translation of a good book, one that either stays true to the author's original text or that lifts it with the translator's own voice and flavour.

Thanks for letting me know about Half-pie too, much appreciated.

Keri h said...

'Half-pie' here is mainly 'half-pai' - unenthusiastic/not well-done/half-hearted job:

"So, how did the working party go?"
"Half-pai really."
"O bugger."

harvestbird said...


In the beginning was the Word, and the Word was with God
and the Word was God

In die begin was die Woord, en die Woord was by God
en die Woord was God

Në fillim ishte Fjala dhe Fjala ishte me Perëndinë
dhe Fjala ishte Perëndi

في البدء كان الكلمة والكلمة كان عند الله
وكان الكلمة الله

У пачатку было Слова, і Слова было ў Бога,
І Слова было Бог

В началото бе Словото и Словото беше у Бога;
И Словото бе Бог

Al principi era el Verb i el Verb era amb Déu;
I el Verb era Déu


Cesta je u početku u Boga u istoj;
I Riječ bijaše Bog

Cesta je na počátku u Boha ve stejné;
A to Slovo byl Bůh

Stien er i begyndelsen hos Gud i det samme;
Og Ordet var Gud

Het pad is in het begin bij God in hetzelfde;
En het Woord was God

Tee alguses Jumala juures selles;
Ja Sõna oli Jumal

Gawin ito sa simula ng Diyos;
At ang Salita ay Diyos

Tee näin alussa Jumalan
Ja Sana oli Jumalan

Pour ce faire, au début de Dieu
Et le Verbe était Dieu

Para iso, a comezos de Deus
Ea Palabra era Deus

Aus diesem, dem Beginn der Gott
Und das Wort war Gott

Για αυτό, οι αρχές του Θεού
Και ο Λόγος ήταν ο Θεός

לכן, עקרונות של אלוהים
והמילה היתה אלוהים

इसलिए, परमेश्वर के सिद्धांतों
शब्द ईश्वर था

Ezért az elveket, az Isten
Isten szava

Því að meginreglur Guðs
Orð Guðs

Untuk prinsip-prinsip Allah
Firman Allah

Chun na prionsabail a bhaineann le Dia
An Focal Dé

I principi di Dio
La Chance parola


하나님의 원리
단어의 기회

Dieva principi
Word of iespējas

Dievo principai
Žodis galimybių

Божјите принципи
Збор на можности

Prinsip Allah
Word peluang

Il-prinċipju ta 'Alla
Opportunità Word

Prinsippet om Guds
Opportunity Word

اصل خدا
فرصت ورد

Bóg Artykuł
Szansa Verde

Deus artigo
Chance Verde

Articolul lui Dumnezeu
Chance verde

Статья Бога
Шанс зеленый

Члан Бог
Могуће Зелена

Boh členských
Prosím, Zelená

Boga država
Prosimo, Green

Dios miembros
Por favor, Verde

Mungu Member
Tafadhali Verde

Gud Medlem
Please Verde


Üye Tanrım.
Lütfen Cape.

Члени Богу.
Будь ласка, мис.

Các thành viên của Thiên Chúa.
Xin vui lòng mũi.

Aelodau Duw.
Os gwelwch yn dda trwyn.

מיטגלידער פון גאָט.
ביטע נאָז.

Members of God.
Please nose.

harvestbird said...

... with apologies for the lack of right justified-margins for the Arabic, Hebrew and Yiddish lines.

Giovanni said...

That is quite something... why had I never thought of doing that before? Very interesting how the phrase got shorter and shorter, and the Italian (God's principles / The word of Chance) is just brilliant.

harvestbird said...

Something seems to happen across the Indonesian - Irish - Italian translation, and then again coming out of Korean, Persian and Thai. I've enjoyed playing around with machine translations ever since Babelfish was the best option, so this seemed like a natural opportunity.