Affichage des articles dont le libellé est translation. Afficher tous les articles
Affichage des articles dont le libellé est translation. Afficher tous les articles

vendredi 21 janvier 2011

Ads & Marketing Translator

Lire mon dossier complet sur le binôme Google - Traduction (PDF, 4Mo)

Ce blog est né il y a presque six ans (déjà !), à l'enseigne de la traduction publicitaire et marketing, Adscriptor signifiant à l'origine Ads & Marketing Translator!
Un néologisme avec une double trace signifiante, où l'on a une superposition d'Ad - ou Ads - (Advertisement en anglais, l'équivalent de notre "pub") et de Scriptor, mot latin qui a la même racine que Scriptum (comme dans P.S., l'écrit) ou Scriptura (écriture) et a plusieurs sens en latin : secrétaire, copiste, écrivain, auteur, rédacteur, et même législateur, ou encore historien (celui qui rédige...). Ceci dit, le nom « Adscriptor » a encore d'autres caractéristiques, puisque c'est également un mot qui existe en espagnol (où il qualifie la fonction spécifique d'un professeur, malgré mes recherches je n'ai jamais très bien compris de quoi il s'agit vraiment), formé à partir de deux autres morphèmes existants, un mot et un préfixe latins.
Mais 730 billets plus tard, que d'eau a coulé sous les ponts...

Pour autant l'envie m'a pris de dépoussiérer un peu le filon "traduction" (ne pas confondre avec le tag "laboratoire de traduction", qui recouvre des billets traduits de l'anglais ou de l'italien vers le français), soit une trentaine de billets publiés sur le sujet, dans lesquels je m'efforce de faire un peu le tour de la situation, en évolution permanente.

Or vu que je suis en train de préparer une journée de formation sur le marketing pour des traducteurs-interprètes italiens, j'avais besoin de me replonger dans le bain en relisant ce que j'ai écrit, et notamment sur le binôme Google-Traduction (PDF complet, 4Mo):
mais également sur la traduction en général, et sur la communication multilingue en particulier :
En parallèle j’ai traité de la « foule-traitance », (crowdsourcing) de Facebook (7 janvier 2008), en approfondissant sa localisation quelques mois plus tard dans le cadre de mes quelque 90 billets sur le phénomène Facebook, ou encore, dans le genre « boutade », relaté l’épisode sur le traducteur facétieux et l'ego de Sarko (30 avril 2007) (ego bien connu, par ailleurs…), et même les traductions du Petit Prince !

Question terminologie, je me suis essayé à la création d’un moteur de terminologie, Translation 2.0, mis en ligne le 28 mars 2007, et d’un glossaire comptable multilingue : XBRL.name.

Sans oublier une « Réflexion quasi-philosophique poétique sur la terminologie et son évolution souhaitable » (3 juillet 2007) qui vaut ce qu’elle vaut...

Idem pour mes conseils sur le déploiement de la traduction automatique en entreprise (août 2008), qui m’ont valu d’être contacté il y a quelques mois pour occuper un poste de Directeur en charge du département linguistique d’une grande banque tunisienne : « The Director, Language Services Department (CLSD) will perform under the general supervision of the Vice-President, Corporate Services (CSVP) and will be based in Tunis, Tunisia. »

J’ai décliné pour ne pas imposer un changement de pays à ma femme et mon fils, même si j’étais loin de penser à l’actualité de ces dernières semaines…

Terminons ce récapitulatif par un passage en revue des principales places de marché dédiées aux traducteurs et à la traduction sur le Web : Top 20 of main Translators & Translation Workplaces & Marketplaces (12 novembre 2008).

En conclusion, après avoir annoncé il y a presque 3 ans le compte à rebours (25 février 2008) de Translation 2.0 Open Project - TOP², basé sur une vieille idée d’ontologie de la traduction dans le monde, un projet qui s'est plutôt avéré être un triste poisson d'avril, et après avoir fêté les 20 ans de bons et loyaux services du Studio 92 Snc (27 février 2009), l’été dernier j’ai finalement changé la raison sociale de ma société, aujourd’hui : Translation 2.0 S.a.s.

L’aventure continue dans la traduction, belle
infidèleinconnue !

Partager sur Facebook


www.translation2.com

P.S. Cette année je fête 25 ans de métier...

, , , , , , , , , , , ,

mardi 9 mars 2010

Google, traducteur universel


Après avoir introduit le "speech-to-text" (reconnaissance vocale + transcription texte), Google a aussi ajouté la traduction automatique à la volée sur YouTube...

* * *

Ça commence à devenir problématique de trouver des titres que je n'ai pas déjà utilisés pour parler du binôme Google + traduction. Pour autant les progrès que fait la société dans ce secteur, aussi rapides que gigantesques, m'imposent de faire des mises à jour de temps en temps.

Cette fois l'occasion m'en est donnée par un article du New York Times intitulé Putting Google to the Test in Translation, qui compare la traduction humaine d'un extrait de texte avec celles des principaux traducteurs gratuits disponibles sur le Web : Google, Yahoo et Microsoft.

La comparaison du journal porte sur cinq langues source (français, espagnol, russe, allemand et arabe) vers une seule langue cible, l'anglais. Or je ne m'intéresserai ici qu'aux deux premières, tout simplement parce que le texte choisi pour l'extrait est littéraire : Le petit Prince de Saint-Exupéry pour le français, et Cent ans de solitude de García Márquez pour l'espagnol :


Donc, ce qui m'a énormément frappé dans ce texte, c'est que pour la première fois j'avais sous les yeux la preuve absolue de ce que j'énonçais il y a plus d'un an et demi dans Google et la traduction, à propos du concept de mémoire de traduction universelle :
Pour rappel, cette mémoire sert également à l'auto-apprentissage de Google, et par bitexte il faut comprendre que l'on a texte source (ou texte de départ) et texte cible (ou texte d'arrivée) en regard l'un de l'autre. Exemple.

(...)

Donc en nous livrant à un bref exercice de prospective, on peut très facilement imaginer que dans un avenir proche, non seulement Google pourra coupler par défaut l'opérateur à votre profil (en clair, sachant que vous êtes anglais, il vous proposera par défaut des termes traduits en anglais, sauf indications contraires de votre part), mais aussi, et surtout, qu'il pourra puiser pratiquement tous les termes du langage humain, dans toutes les langues, au fur et à mesure que sa notre mémoire de traduction universelle prendra forme.

Alimentée autant par les traducteurs humains qui utiliseront les outils de Google pour traduire, que par ses technos automatisées à grande échelle (à ne pas confondre avec le déploiement d'un système de traduction automatique en entreprise, par exemple), voire par la mise en parallèle des œuvres littéraires qui appartiennent au patrimoine de l'humanité et sont déjà traduites dans de nombreuses langues.

Pour les profanes, mettre en parallèle un texte c'est prendre Les Misérables de Hugo ou votre livre préféré, le segmenter et mettre en parallèle les segments du texte original avec les segments correspondants traduits dans la ou les langues de votre choix (à noter qu'un segment ne correspond pas forcément à une phrase, découpée en plusieurs segments si elle est trop longue, par exemple). Vous faites ça avec français-anglais, et vous avez la mémoire français-anglais des Misérables. Ensuite vous faites de même avec anglais-italien, espagnol-allemand, russe-chinois, etc., et vous obtenez autant de mémoires que de langues dans lesquelles l'ouvrage a été traduit.

La pierre de Rosette est un parfait exemple de textes mis en parallèle. Et pour me limiter à n'en mentionner qu'un seul autre, pensez aux milliers de traductions de la Bible qui existent déjà...

Donc ajoutez-y tous les grands classiques mondiaux déjà numérisés, construisez les mémoires de traduction correspondantes dans les couples de langues dont vous disposez, et vous comprendrez aisément qu'on n'est pas loin de pouvoir mettre en parallèle pratiquement l'ensemble du langage humain, à toutes les époques.

Depuis l'aube de l'humanité, nul n'a jamais été en mesure de faire ça. Jusqu'à Google...
Or les deux échantillons pris en exemple par le New York Times montrent que Google construit bien cette mémoire !

En effet, par rapport à la traduction humaine, celle de Google est identique à près de 65% pour Le petit Prince (38 mots sur 59 qui forment des séquences équivalentes) et carrément à 99% pour Cent ans de solitude !!!

D'où une distanciation abyssale, en termes qualitatifs, vis-à-vis des deux autres traducteurs, qui ne réussiront jamais à combler leur retard s'ils ne passent pas à des modèles de traduction automatique fondés sur le couple "linguistique de corpus" + "analyse statistique". Et bien au contraire, l'écart se creusera démesurément au profit de Google et aux dépens de ses concurrents.

C'est là tout l'enjeu des mémoires de traduction à très grande échelle implémentées par Google :
Comme l'explique fort bien Franz Josef Och, responsable recherche et traduction automatique, la clé des modèles de traduction selon Google, perpétuel apprenant, c'est de pouvoir disposer, d'une part, d'énormes quantités de données linguistiques (very large amounts of datas), et de l'autre, d'une phénoménale puissance de calcul qui traite à très grande vitesse des milliers de milliards de mots (aussi bien données textuelles que vocales...) dans pratiquement toutes les langues, puisque plus vous alimentez vos modèles statistiques en données, plus la qualité des modèles s'améliore !

Il précise d'ailleurs les deux principaux enjeux pour Google dans la traduction automatique :
  1. augmenter la qualité de sortie, grâce au binôme quantité de données / puissance de calcul, ce qui explique clairement pourquoi la qualité est meilleure pour les couples de langues plus représentées ;
  2. augmenter en conséquence le nombre de langues (et donc de couples de langues) et de fonctionnalités offertes (comme les recherches croisées : je saisis un terme en français pour une recherche sur le Web chinois, et en sortie j'obtiens une page avec les résultats chinois à gauche et en vis-à-vis leur traduction française à droite), etc.
Alors au bout du compte on finira par obtenir ce que j'expliquais dans Google et la traduction, à savoir la
mémoire de l'humanité...
[MàJ - 10 mars 2010] Grâce à ce tweet de Christophe Asselin, je découvre les chiffres suivants :
Pour modéliser un langage, Google a besoin d'environ un milliard de mots, employés dans des combinaisons différentes. Pour l'anglais, l'entreprise a injecté « plusieurs centaines de milliards » de termes, selon le New York Times.
Donc, là encore, les prodigieuses capacités de calcul de Google, couplées à une approche d' « intelligence hybride » associant l'homme et la machine, font la différence. Google l'explique ainsi :
If you have access to enormous datasets, it opens up whole new avenues for scientific discovery and for solving problems. For example, Google's machine translation tools take advantage of "parallel texts": documents that have been translated by humans from one language to another, with both forms available. By comparing the sentences from enormous numbers of parallel texts, machine translation tools can develop effective translation tools using simple probabilistic approaches. The results are better than any previous attempts at computerized translation, but only if there are billions of words available in parallel texts.
Mais ce n'est pas tout ! [Début]

* * *

Prenons l'exemple de mon dernier billet, dans lequel il y a une vidéo qui affiche un bandeau avec capture automatique du texte source, obtenue par reconnaissance vocale.

Et bien le texte correspondant ne réside que dans la mémoire de Google, il n'est pas disponible en ligne. Par conséquent pour le traduire en français, j'ai dû d'abord le retranscrire intégralement en italien afin d'avoir le texte, qui n'est maintenant disponible en ligne que sur mon blog italien et sur la plateforme de blogging.

Première constatation : la capture en italien est fiable à 100%.
Deuxième constatation : une fois que Google détecte la traduction du texte source, vous pouvez être sûr que le bitexte part dans la mémoire qui alimente son système de traduction automatique.

Voilà pourquoi j'annonçais cette nouveauté il y a déjà presque 4 ans :
... une fois que Google maîtrisera avec un degré de pertinence suffisant la traduction automatique du texte, et a fortiori de la voix, je vous laisse imaginer le reste... (et) je suis sûr de ne pas me tromper en affirmant que Google nous prépare quelque chose de révolutionnaire avec sa fonction de traduction automatique : texte-voix, Internet, vidéo, mobiles, Adsenses personnalisés, etc., qui pourra concurrencer Google ?
Car en fait, qu'il s'agisse des données images, vidéo ou voix, il suffit d'obtenir le texte source dans une première étape, soit par reconnaissance de caractères soit par reconnaissance vocale (pour la téléphonie mobile), et d'appliquer ensuite la traduction automatique au texte obtenu.

Si je garde mon exemple des vidéos sur Youtube, vous sélectionnez la langue d'arrivée de votre choix (mais comptez sur Google pour vous proposer par défaut votre langue maternelle) et vous obtenez dans le bandeau le texte traduit. Du sous-titrage automatique à la volée.

Et enfin, pour la téléphonie mobile, il suffira d'appliquer la synthèse vocale au texte traduit pour entendre dans son oreillette la version parlée. C'est ce que promet déjà Microsoft, et ce qu'a déjà annoncé Google...

Si vous voulez tester, la fonction est disponible sur Google Translate : cliquez sur l'icône du haut-parleur (que j'ai signalée par une flèche) pour entendre la version parlée du texte traduit !


[Début]


Partager sur Facebook

P.S. Quand on dit que le hasard fait bien les choses, je viens juste de découvrir par un tweet que la version bêta de la traduction automatique est disponible sur Youtube !!!

Donc, vous paramétrez la langue :


Et voilà le travail...


Pour finir, je découvre que la fonctionnalité "traduction" sur Youtube est en ligne depuis deux ans !

Que dire ? Ça se voit que je ne m'occupe plus du Web tellement je suis accaparé par le berlusconisme... [Début]

, , , , , , , , , , ,

mercredi 12 novembre 2008

Top 20 main translation workplaces

Top 20 of main Translators & Translation Workplaces & Marketplaces

Source : Google Trends - Daily Unique Visitors (average)
  1. Proz : 80,000
  2. ForeignWord.com (+ .BIZ) : 15,000
  3. Tanslators Café : 10,000
  4. Cucumis : 8,000
  5. Word2Word : 4,500
  6. Language123 : 4,000
  7. TRADUguide : 3,000
  8. Translatorsbase : 2,000
  9. GlobTra : 2,000
  10. Go Translators : 1,500

Top 10 of main Translators & Translation Workplaces & Marketplaces
From 11 to 20 (alphabetical order, no traffic data available) :
* * *

Here is what these sites are saying of themselves:

1. Welcome to ProZ.com, the world's leading enabling and sourcing platform for language professionals. ProZ.com's unique online workplace provides tools and opportunities for its members to network, collaborate on projects and terms, expand their businesses, do better work and have more fun. Plus ProZ.com enables companies and organizations to rapidly search, find, qualify, evaluate and interact with the translators, interpreters and linguists they need.
See some stats.

2. While at Foreignword.com you find all your favorite online translation tools, the Eurêka database and the hundreds of language links, Foreignword.biz is dedicated to our software products as well as the Foreignword Translators Directory.

3. Welcome to TranslatorsCafé.com — Directory of translators, interpreters and translation agencies!
4206 translation agencies
81033 registered members

4. Cucumis is a community of translators who share their linguistic knowledge and help each other online.
Members : We have 149277 registered users

5. Word2Word Language Resources is dedicated to breaking down of language barriers and assisting the users who have the desire to learn a language, a need to communicate between languages, and for those who work with languages as a profession.

6. Language123 : The translation marketplace

7. TRADUguide - Your Guide to Translators and Translation Agencies

8. Translatorsbase is the leading provider of translation services via network of global high quality service providers. Translatorsbase.com is a source of translation jobs and provides tools and services to help you bring your translation business to the next level.

9. GlobTra is a free and open international community of professional translators, which lets you find jobs, promote your services and help each other. Make your business grow, while you have fun. This is your site!
Need a translator? Here's 12000! Are you a translator? Find work, friends and help.

10. Go Translators : the top tool for bringing specialist translators and their commercial or institutional clients closer together.
Freelance translators and specialist agencies publicise their skills to the millions of companies and private individuals who use the Internet.

11. Aquarius is the longest standing online marketplace for translation and localization projects. Since February 1995 we help translation vendors find clients and assist businesses in locating reliable localization specialists. This site is a public beta. Click here to access the old Aquarius website. Facts and figures:
28568 freelance members, 4904 corporate members, 40134 translators, 7866 interpreters, 7743 subtitlers, 2058 multilingual copywriters, 949 localization engineers, 1358 multilingual DTP specialists from 136 countries.
5331 projects were outsourced through Aquarius in 2007.

12. Welcome to GLTJobs.com, the specialist job board for all the best job opportunities in Globalization, Localization and Translation across the World.
The site connects industry professionals with globalisation, localisation and translation jobs in Europe, The Far East and America. Registration is simple and takes only a few moments.

13. LinguistFinder has thousands of professional translators, interpreters, and tutors worldwide, offering any language translation and localization services at cost price.
No Obligation Language Translation Quotations From 1000's of Professional Translators, Interpreters, & Tutors

14. Trally : Are you in need of professional translation services?
We are here to help you with more than 53,000 registered Freelance Translators and Translation Agencies !

15. TranslationStaffFinder is now probably the fastest growing global translation employment website dedicated to the Translation sector. Translation Staff Finder is a specialist translation job site for those looking for jobs and employment in the Translation industry and for companies looking to resource and hire personnel with multilingual, linguistic and associated skills.

16. Translationzone : The official home for SDL TRADOS - The leading freelance translation software
Providing support and resources to a global community of over 170,000 users!

17. Translator Planet is a portal for Online Translator and online auction for language translator projects. Demo.

18. Translator.search-in.net (translatordatabase.net) - Worldwide Translator/Interpreter Database

19. TranslatorPub / Stats
Total Jobs : 12 226
Open Jobs : 2 904
Total Notifications : 1 741 021
Job Views : 405 880
Welcome!

20. TransRef - Directory of translators
There are currently 8549 translators in the Directory.


Partager sur Facebook

P.S. I took Translation Directory off the list because of this Google advisory:
Of the 19 pages we tested on the site over the past 90 days, 7 page(s) resulted in malicious software being downloaded and installed without user consent. The last time Google visited this site was on 2008-11-12, and the last time suspicious content was found on this site was on 2008-11-12.
Malicious software includes 12 exploit(s). Successful infection resulted in an average of 3 new processes on the target machine.



, , ,

mardi 9 septembre 2008

Google and the Universal Translation Memory

Version française

Blown away from the outset by Google’s speedy and significant break into machine translation, I have long been predicting its brilliant progress in the field of translation. Surely there are quite a few surprises left to come.

I’m at the point where instead of the operator define:, sometimes I test the operator translate: just to see if it has been implemented since I last checked. And it seems that the moment has arrived, with the translation onebox (Via Google Live).

Apparently, it currently only concerns common expressions, although it is likely it will cover all terms already included in Google’s dictionaries.


I tried with Italian but it doesn’t work yet. Even so, it is likely that as soon as the future Google Translation Center is up and working, this operator will also search for translations in the bitexts that will populate the universal translation memory an army of translators is developing daily.

Remember that this memory also helps Google self-teach, and “bitext” refers to a source text (or starting text) and a target text (or final text) set side by side in alignment. Example.

In fact, the operator does not seem to function the same way as define:, where you use the English term no matter what language you speak.

With "translate", it’s apparently the operator that defines the target language: when you type translate ordinateur Google directly interprets this to mean translation from French to English, whereas with traduire computer, it interprets English to French.

So, if we engage in a brief forward-looking exercise, we can easily imagine that in the very near future, not only will Google be able to match the operator to your profile by default (that is, knowing you are English, it will offer you terms translated into English by default, unless you indicate otherwise), but it will also (and most importantly) be able to draw from practically all the terms in human language, in all languages, as the universal translation memory gradually takes shape.

It will be filled as much by the human translators who use Google’s tools to translate, as by its large-scale automated technologies (not to be confused with the implementation of an automatic translation system in a company, for example), or even by the alignment of literary classics that make up our common heritage and which are already translated into countless languages.

For the layman, aligning a text is taking for example Victor Hugo's Les Misérables breaking it into segments and aligning the segments of the original text with the corresponding segments translated into the language(s) of your choice (noting that a segment does not necessarily correspond to a sentence, which will be broken into multiple segments if it is too long, for example). You do this with French-English, and you have the French-English memory of Les Misérables. Then you do the same thing with English-Italian, Spanish-German, Russian-Chinese, etc., and you get as many memories as there are languages into which a work has been translated.

The Rosetta Stone is a perfect example of aligned texts. And if I could only mention one more, think of the thousands of translations of the Bible that already exist...

Add to that the great classics from around the world that are already in digital format, build the corresponding translation memories in the language pairs you have access to, and you can easily understand that we are not far from being able to align practically the whole of human language, from every era.

Since the dawn of humanity, no one has ever been able to do that. Until Google...

The talk about Google and translation is not over yet. In fact, it’s only beginning!


Partager sur Facebook

, , , , , , , , , , ,

dimanche 31 août 2008

Google et la traduction

English version

Bluffé depuis le début par la percée significative et ultra-rapide de Google dans la traduction automatique, voici longtemps que je prédis ses avancées éclatantes dans le domaine de la traduction, de nature à nous réserver bien des surprises.

À tel point qu'à l'instar de l'opérateur define:, de temps en temps je teste l'opérateur translate: juste pour voir s'il n'aurait pas été mis en place entre-temps. Or il semble qu'avec la translation onebox, le moment est venu ! (Via Google Live).

Apparemment, ça ne concerne aujourd'hui que les expressions courantes, même s'il est probable que cela puisse porter sur tous les termes qui renseignent déjà les dictionnaires de Google.


J'ai testé avec l'italien mais ça ne marche pas encore. Pour autant, il est probable que dès la mise à régime du futur centre de traduction de Google, cet opérateur ira également rechercher les traductions dans les bitextes qui peupleront la mémoire de traduction universelle quotidiennement élaborée par une armée de traductrices et de traducteurs.

Pour rappel, cette mémoire sert également à l'auto-apprentissage de Google, et par bitexte il faut comprendre que l'on a texte source (ou texte de départ) et texte cible (ou texte d'arrivée) en regard l'un de l'autre. Exemple.

Par ailleurs, l'opérateur semble ne pas fonctionner de la même manière que define:, où vous utilisez le terme anglais quelle que soit votre langue.

Alors qu'avec "translate", c'est apparemment l'opérateur qui définit la langue d'arrivée, puisqu'en saisissant translate ordinateur Google interprète directement le sens de traduction français vers anglais, tandis qu'avec traduire computer l'interprétation se fait de l'anglais vers le français.

Donc en nous livrant à un bref exercice de prospective, on peut très facilement imaginer que dans un avenir proche, non seulement Google pourra coupler par défaut l'opérateur à votre profil (en clair, sachant que vous êtes anglais, il vous proposera par défaut des termes traduits en anglais, sauf indications contraires de votre part), mais aussi, et surtout, qu'il pourra puiser pratiquement tous les termes du langage humain, dans toutes les langues, au fur et à mesure que sa notre mémoire de traduction universelle prendra forme.

Alimentée autant par les traducteurs humains qui utiliseront les outils de Google pour traduire, que par ses technos automatisées à grande échelle (à ne pas confondre avec le déploiement d'un système de traduction automatique en entreprise, par exemple), voire par la mise en parallèle des œuvres littéraires qui appartiennent au patrimoine de l'humanité et sont déjà traduites dans de nombreuses langues.

Pour les profanes, mettre en parallèle un texte c'est prendre Les Misérables de Hugo ou votre livre préféré, le segmenter et mettre en parallèle les segments du texte original avec les segments correspondants traduits dans la ou les langues de votre choix (à noter qu'un segment ne correspond pas forcément à une phrase, découpée en plusieurs segments si elle est trop longue, par exemple). Vous faites ça avec français-anglais, et vous avez la mémoire français-anglais des Misérables. Ensuite vous faites de même avec anglais-italien, espagnol-allemand, russe-chinois, etc., et vous obtenez autant de mémoires que de langues dans lesquelles l'ouvrage a été traduit.

La pierre de Rosette est un parfait exemple de textes mis en parallèle. Et pour me limiter à n'en mentionner qu'un seul autre, pensez aux milliers de traductions de la Bible qui existent déjà...

Donc ajoutez-y tous les grands classiques mondiaux déjà numérisés, construisez les mémoires de traduction correspondantes dans les couples de langues dont vous disposez, et vous comprendrez aisément qu'on n'est pas loin de pouvoir mettre en parallèle pratiquement l'ensemble du langage humain, à toutes les époques.

Depuis l'aube de l'humanité, nul n'a jamais été en mesure de faire ça. Jusqu'à Google...

Google et la traduction, on n'a pas fini d'en parler. En fait, on commence juste !


Partager sur Facebook

, , , , , , , , , , ,

vendredi 5 octobre 2007

The Communicational Circuit of Translation


French version / Version française

Translation = Communication
The Communicational Circuit of Translation
Conclusion: Choosing a service provider worthy of the name

* * *

Translation = Communication

I’ll start this post by quoting myself on the future of translation:
And especially, thanks to the Internet, the time has come for translation to finally become what it has been since the dawn of time: pure communication...
- Question to businesses: Do you leave your communication to just anyone?
- Answer: No, never.

- Related question: Do you leave your translations to just anyone?
- Answer: Yes, often.

- Why the disconnect?
- Because a large majority of clients still think - against all logic, except the money argument - that if you can babble a few words in a language, you can translate it.

So without batting an eye they entrust their corporate documents and even their commercial correspondence to so-called translators whose only merit is being the lowest bidder, and basing the decision to award the contract on but one criterion: the lowest possible price...

Overlooking skill and quality, and completely oblivious of the fact that translation is a complex operation which follows its own communicational circuit. [Top]

* * *

The Communicational Circuit of Translation

I am going to try to clarify things with an explanation of the real communicational circuit of translation as I see it, drawing inspiration from various communications theories and detailing each stage. In September 2004, I created the following illustration:


In short, combining several communication models, you have:

1. An information source (who)
2. A message to communicate in the form of a signal (says what)
3. A transmitter which will encode the signal (how)
4. A transmission channel which will carry the signal (by what channel)
5. A receiver which will decode the signal to recreate the message and restore the information it contains (to whom)
6. Feedback (with what return effect)

When we speak of interpersonal communication, it is generally accepted that the transmitter and the receiver understand each other because they speak the same language. But this assumption is skewed from the outset when they speak two mutually incomprehensible languages. Such is the daily lot of about 7 billion Earthlings in an infinite number of places and circumstances!

So, if we put aside the technical aspects to concentrate on the path a message takes from emission in Language A by its author to reception in Language B by its receiver - which is a definition of translation - we see the following communication correspondences:

1. An information source (who) (author)
2. A message to communicate in the form of a signal (says what) (message in Language A)
3. A transmitter which will encode the signal (how) (encoding into source message)
4. A transmission channel which will carry the signal (by what channel) (translation)
5. A receiver which will decode the signal to recreate the message and restore the information it contains (to whom) (recipient of the target message in Language B)
6. Feedback (with what return effect) (is there loss - or addition - of information, satisfaction or dissatisfaction with the quality of the translated message)

First comment: Some may be surprised that I have placed language between the author and the source message, or between the target message and the recipient.

The explanation is simple: For the creator, or for the translator, there is never ONE SINGLE way to encode a message, since every language allows for many encoding variants, which are called registers of the language. Between "Would you kindly leave" and "Get out of here", the message is the same but the presentation is a little different...

The encoding of the source - and target - message changes with the purpose of the communication, the recipient, the location, the circumstances, and so on; there are many criteria. Also coming into play are noise (what I call linguistic interference, such as poor mastery of the language, style, syntax, grammar, spelling) and filters (be they cultural, religious, familial, social, etc.), which can potentially modify or disturb the clarity of the information. In this way, every message is influenced by an infinite number of variables, whether conscious or not.

It is not uncommon for an engineer, a lawyer, a doctor, or any expert in a field, not to be able to express himself properly in his own language. A translator who has been practicing his craft for a long time is sure to have hundreds of examples on hand where his final text is noticeably more intelligible than the original document.

As for the TRANSLATION process itself, which is based on the intrinsic skills of the translator AND the technical aspects of localization and/or computer-assisted translation (from automatic translation on search engine to translation environments, from terminology management to translation memories, from bitext or multilingual corpuses to online glossaries and specialized search engines), it is the channel which will carry the message from one language to the other.

It is an operation of creating an equivalent message - not equivalent terms or words (a subtle distinction of vital importance which I have detailed in an analysis of professional technical translation) - which for translators consists in “increasing the elasticity of language” and “creating bridges between peoples”, quoth Victor Hugo.

Finally, you will note two breaks in my graphic which can potentially interrupt the message, either upstream if the integrity of the information is not transferred from the source message to the target message because of incomprehension or poor translation, or downstream when the recipient of the message is incapable of judging the correspondence between source and target, or if he couldn’t care less, which is not uncommon either. :-)

That being the case, on the (dis)satisfaction scale, there can be a thousand different situations depending on the (in)competence of either party. [Top]

* * *

Conclusion: Choosing a service provider worthy of the name

So if your SME is among the European companies which are "experiencing significant commercial losses for lack of adequate linguistic skills", and if by some lucky chance you are responsible for ordering translations and you have had the patience and fortitude to read this far, I dare to hope that you will entrust your next translation project or the localization of your corporate website to a service provider worthy of the name.

Like TRANSLATION 2.0, for example (but that’s just an example :-).

For in the end, as a general rule, the quality of a translation is always proportional to the price you pay - or are prepared to pay - for it. Multilingual communication is a craft, and like any craft, it can’t just be improvised! [Top]


Nota Bene: The French original post has been brilliantly translated by Theresa Shepherd. Many thanks, Theresa.  :-)

, , , , , , , , ,

vendredi 15 juin 2007

Challenges and nature of a formal language strategy

Challenges and nature of a formal language strategy

French version / version française

on ProZ.com


Introduction
I. The Challenges
II. The Languages
Conclusion

* * *

Introduction

In a document presented in November 2005 to the Council, the European Parliament, the Economic and Social Committee and the Committee of the Regions and entitled “A New Framework Strategy for Multilingualism”, the European Commission announced a “study on the impact on the European economy of shortages of language skills” to be published in 2006.

It’s the ELAN survey (Effects on the European Economy of Shortages of Foreign Language Skills in Enterprise, December 2006), which gives us “the most comprehensive survey of the use of, and need for, languages and cultural knowledge ever carried out” and specifically assesses:
• lost business, or underperformance, due to deficient language skills and cultural knowledge;
• barriers to trade due to deficient language skills and cultural knowledge,
studying 30 large companies (criterion for selection was that they operate internationally with sales in at least 10 countries of the world, including Total, BP, Canal+, Unilever, Saatchi & Saatchi, Pernod Ricard, Accor, Air France, General Electric, GlaxoSmithKline, etc.) and almost 2000 small and medium enterprises (SMEs) from 29 European countries; 89.6% of the firms have more than 5 years in the market. The sample was selected in each country to be as close an approximation of the country's export profile as possible, and representative of the target country's export profile (Manufactures of food products and beverages, of machinery and equipment, of fabricated metal products, of chemicals, furniture, textiles, rubber and plastic products, radio, television and communication equipment, wholesale trade and commission trade and other business activities, etc.), offering a cross-section of company sizes (micro-companies up to medium-sized with 250 employees) and aiming to reflect a national rather than a regional picture. In aggregate (taking all SMEs as a whole), almost ¼ of the respondents were Export Managers, ¼ General Managers, ¼ Managing Directors and ¼ Administrators.

I don’t know if the study will be available with Reportlinker, anyway it’s freely downloadable on the Net, in English and French, so I hope you’ll be content with this solution. :-)

Even if summarizing the 85 pages of the survey in one post seems to be a tricky task, I quote the main elements in my opinion, that is the one of a senor translator who fight almost every day since two decades and longer to make companies more aware of language and cultural challenges, trying to explain to my customers the benefits of a quality translation, vs. the damage of their do-it-yourself solutions while attempting in vain to square the triangle[Top]

I. First point: the CHALLENGES

The survey of nearly 2000 small and medium enterprises found that a significant amount of business is being lost for European SMEs as a result of lack of language skills. Across the sample, 11% of respondents had lost an actual or potential export contract for this reason; 11%: the figure seems low, but related to 20 million businesses in EU19, this “significant amount of lost business” (estimated in a range between €16,400,000 and €25,300,000 of potential loss for respondent SMEs) due “to deficient language skills” would concern about … 2,200,000 companies!

Moreover, this figure of 20 million businesses is a prudent estimation, since it’s far exceeded in EU25. Anyway, by default it means that on a percentage basis, one point = 200,000 businesses at the European Community scale. A magnitude to keep in mind throughout the reading of this post...

Nevertheless, large companies would account for 0.2% of the total, that is to say approximately 50,000 global corporates!

So, for saying it with the survey, “Taken together, this constitutes potentially an enormous loss to the economy of the European Union” since “we estimate the total losses to the EU economy through lack of language skills in the SME sector are in the region of €100 billion per year”, with a €325,000 average loss per business over a three year period.

And if on average 48% of the firms (mainly from non-Anglophone countries as well from more recent applicants or entrants to the EU) “acknowledge having a formal communication (or language) strategy”, what would be the situation if they didn’t?

But what kind of strategies are we talking about? Four elements of “language management” seem to be associated with improved export performance:

1. possession of a formal language strategy (including frequent Website adaptation, with over 50% of businesses in 22 countries showing businesses providing websites in languages other than the national language, most frequently in English for 62% of the firms in the European sample, but also company/sales literature, merchandising information about goods or services and so on),
2. employment of nationals, with 22% of businesses drawing on this resource,
3. hiring staff with specific language skills,
4. using external translators/interpreters.

According to the survey, an SME investing in these four elements was calculated to achieve an export sales proportion 44.5% higher than one without these investments.

In other words, “the ratio of exports to total sales would rise 44.5 percentage points for those SMEs, now without any foreign language investment, that undertook all four types”. SMEs account for between 30 and 60 percent of output, depending on the European country. Assuming, for purposes of illustration, an average across the EU of 45 percent of output, and supposing that SMEs accounting for half of this output adopt the four language skill investments, then exports would rise by 0.45*0.5*0.445=0.10 or 10 percent of GDP (about €1.1 trillion for EU25 in 2005).

To be pointed out that big companies get a larger set of “language strategies” to adopt, for which the survey identifies nine “language management techniques”:

1. Language training
2. In-house languages department
3. Appointment of external translators or interpreters
4. Selective recruitment
5. Intercultural training for the internal workforce
6. Expatriation
7. Inpatriation (employment of local agents)
8. “Buddy schemes” (use of a language-skilled worker or a native speaker to support one with lower levels of language skills)
9. Use of machine translation and web-based tools


About selective recruitments, “there would seem to be an expectation on the part of large companies that appropriate levels of language skills should be available from the employment market rather than through capacity-building within the business.
Given the expectations of further increased demand for language skills both by large companies and by SMEs in future and current dissatisfaction in business with skills levels and output volumes in languages, it is conceivable that language skills, coupled with business awareness, will command a premium on the employment market in future” (causing a drawback for monoglot employees), “with a concomitant increase in demand for native speakers to close skills gaps which education systems cannot fill”.

Last but not least, about using of machine translation and web-based tools, “It is perhaps not surprising that this was the area where the highest proportion of respondents recorded that their company had tried the approach and had abandoned it (37%)”, but really I can’t understand how MT could be qualified as “language management technique”. In fact it should be worth to talk about “interactive semi-automatic translation systems for fast, high-quality translation of texts in specific fields”, following the orientations recommended by the EC in “A New Framework Strategy for Multilingualism”. Even if this field of study probably has to evolve quickly with Google & Co or the GALE Program, inter alia

Only common denominator between large companies and SMEs, the appointment of external translators or interpreters. Here 80% of the former companies had engaged external language specialists, with 47% doing so on a regular basis. Interestingly, two respondents of large companies indicated that the policy had been abandoned owing to unsatisfactory experiences with external providers.

These figures are again significantly higher than those for SMEs, where an average across the EU sample of 45% was recorded. “They suggest that some SMEs may be unwilling or unable to bear the cost of translating what might be essential sales and merchandising information and may be leaving the intermediary function to local agents.”

For sure, larger companies have many advantages on SMEs on this matter and are more likely to acquire language skills because of the threshold cost of this investment. Anyway, even if it is true that reason for SMEs to undertake such investments -which represent one of the fixed costs of exporting to certain countries- is improved access to foreign markets, then SMEs will acquire language skills only if the market is working and when their value to the firm exceeds their cost. It means a balance rather delicate to find, as well as an individual answer. Therefore each SME needs to analyze and better understand upstream where to position oneself in relation to the four principal dimensions of language management:

1. Language preparedness: the level of language competence possessed by the company expressed against current and anticipated needs.
2. Language responsiveness: the willingness and ability of the company to accommodate to the language needs of their international partners.
3. Language awareness: the extent to which language issues are embedded into the strategies and policies of the company.
4. Language management: the extent to which the company is able to satisfy its language needs through prudent deployment of a variety of language management tools including for example language training and expatriation. [Top]

II. All the above brings us to the principal matter of this study: LANGUAGES

Before asking us which languages and situations should be regarded as more problematical for companies, it is worth to focus on a first evidence.

Talking about LANGUAGES means talking about language AND cultural skills.

For instance, a translator doesn’t translate ONLY with words, but ALSO with words (that is what we call “terminology”: all the words I don’t know and need to find out). Each text hides an infinity of situations which go beyond the words: contextual, implicit, drawn from a culture, from a country, from unvoiced language and cultural matters, etc., so many reasons for which a translator having all the words essential to his work (thanks to the dictionaries, of course) not necessarily will provide a good translation if he/she is lacking in other essential capacities, as well technical as cultural, as much in the source language as in the target language, and yet so much in the latter (which should correspond to its mother tongue in a normality assumption).



To be noticed that the very low rate of Spanish or the absence of Portuguese is certainly ascribable to the fact that the sample belongs to Europe, so SMEs are “more likely to need European languages”, since the enormous natural markets of these two languages (Latin America and Brazil) may probably be of greater interest to large companies, who “are likely to be operating on a more global scale” and “have a need well more pronounced for great world languages” (Italian and German do not appear at all in their list of the needs).

These findings may reflect the fact that “SMEs may be looking to export to markets which are near at hand”, given that sometimes proximity could have a determinant impact, as well as political and free trade unions, or level of political and colonial association which are potentially correlated with sharing a common language. Nevertheless 30% of the companies are identifying “Spanish as a key language for improvement” and “26% of respondents seeing Chinese as a language that should receive investment”.

The need for English seems to be at a similar level in both types of enterprise, who “recognized that their use of English needed to be improved”, and that “levels of competence in business language and awareness of business practices fell short of requirements”.


So English is the first result, not surprisingly, even if the survey notices “It is surprising that English is not more widely spread”. This is probably “due to the tendency for companies to try to use the local language of the market if possible, and if not, then one of the major European languages, such as German or French”. However, “the survey results, as well as comments from individual company respondents, suggest that the picture is far more complex than the much-quoted view that English is the world language”, “almost as if English is no longer perceived as a ‘foreign’ language, but assumed to be the lingua franca for trade in many countries”. In fact, “demand for skills in languages other than English (is) greater than the demand for English itself”.

Evidently, a “global marketplace suggests a need for skills in a multiplicity of languages”, and while “English isn't becoming any less important on the Internet, other languages, such as Chinese, Russian, Spanish, and Portuguese, are becoming comparatively more important”. Since “as the balance of economic power shifts away from domination by the West, with the rise of the so-called BRICs economies (Brazil, Russia, India and China) the relative status and power of global languages such as Chinese, Hindi/Urdu, Portuguese and Russian will increase”.

Reasons mentioned by firms that have experienced actual/or estimated losses in trading opportunities due to lack of language skills and/or staff speaking languages (63 % of respondents) mainly concern:

• negotiations
• meetings, social relationships
• Etiquette (especially in Asia)
• correspondence
• enquiries / requests for information or quotations not followed up
• lack of confidence
• phone/switchboard problems
• problems with agents/distributors
• errors in translating/interpreting
• inability to capitalize on opportunities
• exhibitions/trade fairs, travels
• advertising, marketing
• lack of cultural affinity
• better understanding of mindset of other cultures, etc.

The survey, which adds a languages/situations breakdown, recalls that Usunier (2000) “suggests a number of ways to minimize cultural impact in negotiations in order to build effective transcultural relationships”:

• Being willing to adapt
• Being aware that interpreters influence meaning
• Being aware of cultural blocks to translation
• Avoiding negative stereotyping
• Good prior preparation in inter-cultural understanding

Indeed, negotiation is “one of the key skill areas for working effectively across cultures”. Moreover, “Intercultural problems arising from the use of IT have been identified relatively recently”, so the survey suggest “ways in which programme interface designers might develop their products for optimal use by people from different cultures”.

That is what we call localization in our job, a sector of activity who gaves origin to the recent GILT industry.

Although many alternative ways are not yet explored, Internet surely could offer a range of solutions to SMEs willing to expand in export markets but slow down by language and cultural barriers, even if some of them, rightly or wrongly, are unwilling to bear the related cost. To set an example, how many SMEs bear all costs of renting a stand in an international exhibition/trade fair, without having a good commercial communication, bilingual or multilingual, up to the situation? As for so many “localized” Websites where quality is nothing but an optional, even if by far it is not an economical problem, as I showed it in my study on localized global corporates Websites.

However, in the Internet Era, we can consider that all companies on the Web are “born global firms” having a vocation to export, therefore there is evidence that those which will make it possible to each one of their customer to “decide in which language he wants to communicate”, will have an obvious competitive advantage on those which will use a neutral English where they could not work in their partners' language, since we do know “that successful exporters recognize the value of using the customer's language as a basis for long-term relationship management” and that “the personal contact with foreign customer is necessary for every method of trade”.

Moreover, to communicate with foreign customers in their native language is the first step if you want to better understand them for overcoming communication obstacles due to cultural differences. Even if in most companies of Anglophone countries (United Kingdom, Ireland, United States, etc.), “there is clearly either an issue of complacency based on the lack of implementation of language strategies or simply a belief that English is adequate for all trading purposes”, since having English as a native language, or at least as a widely spoken second language, “opens up significant markets which do not have English as a mother tongue”, “used to trade in over 20 different markets” (vs. about 15 markets for German and 8 markets for French, according to the survey).

In fact 96% of Irish companies never hire external translators/interpreters (even if we can observe that, paradoxically, all main players of the GILT industry have their premises in Ireland…), vs. 85% of UK companies, to compare with 26% of German SMEs who rely upon their services, even if “we see a very strong positioning of German as a major second lingua franca of European business”. [Top]

Conclusion

“In each country five ‘influencers’ were asked to comment on the findings of the survey within their area. They represented a cross-section of entrepreneurs, academics working in the business or language training disciplines, political and civil servants working in the business support field and representatives of business organizations such as the Chambers of Commerce”.

“There was a high degree of agreement with the survey's findings, with 79% of those interviewed finding them totally accurate, 21% finding them partially accurate and none disagreeing”.

However, overall, 60% of the influencers thought that “there was too much emphasis upon English and there needed to be expansion of other language knowledge within their country”, and their feedback have “confirmed the outcomes from the analysis of the SME survey”. And given the “diversity of experience and starting positions across the 27 states reviewed”, “it is remarkable that, repeatedly, the issues associated with language skills and economic performance strike harmonious chords across the whole sample”.

Wrap-up:

Across the sample, “the most effective performers amongst export SMEs tend to have a language, or communication, strategy, in place”, an average of 13% of firms claim that the language competence of their staff has influenced the company's choice of export market”, 18% confirm that “they have experienced difficulties with foreign customers due to cultural differences”, and 42 % forecast “an increased demand for language skills” (both qualitative and quantitative) “in the next three years”.

Moreover, exports were important to the sample of SMEs, since “40% of the companies exported more than half of their total sales” (with Italian SMEs being distinctive in their higher relative productivity), and most of them have lost contracts as a result of lack of language and intercultural skills. In particular, smaller firms (with lower turnover) “lack the resource to make forward investment in language skills”.

Related to the European Community scale, these figures result in millions of SMEs and thousands of global corporates, but many of them “are not aware of what skills they possess”. Where this happens, the ‘right person for the job’ can be slotted in at the right time and may indicate “how dependent an SME is on external providers of language support”.

However, it is well known that one’s man joy is another man’s sorrow, so it is the duty of skilled worker of the language-related industries (translation, editing, proofreading, précis-writing, interpreting, terminology, language technologies [speech processing, voice recognition and synthesis], language training, language teaching, language certification and testing and research, etc.) to increase the customers and clients awareness on implications of a good (or bad) provision of language/cultural service, which is all about gained (or lost) markets, valuable (or deteriorated) image and improved (or definitively missed) customer loyalty…

Clearly our path is all laid out to put forward our professions, and our professionalism, which too often lack of recognition and of a fair income.

Lastly, the survey stresses in its general conclusions that “these investments are an essential factor in enabling the EU to compete on the basis of skills and knowledge rather than on the basis of low costs”.

Really, it seems to hear the very words of our professions and here is what is at the bottom of all of that: UE and translators fight for obtaining same results!

For once that a consensus has been reached on what is at stake when we’re talking about challenges and nature of a formal language strategy… [Top]



, , , , , ,