Futility in Innovation

Digital workersAlthough there is no compelling evidence of Henry Ford ever saying, “If I had asked my customers what they wanted, they would have said a faster horse”, in a 1923 interview to The Christian Science Monitor, he spoke about the replacement of horses on farms by machinery.

The intriguing element in Henry Ford’s alleged quote is the reference to speed. Customers would not have been looking for stronger or long-lived horses, but for faster ones, as if it there were a way to be sure of breeding an animal with such a feature. Faster horses to eventually get automobiles is the perfect way to depict futility in innovation.

The World’s Fastest Lorry

John Steed, the central protagonist on the popular 1960s British spy series The Avengers, drove a Bentley Speed Six, while 007 drove a Blower Bentley in the first three novels and in From Russia with Love.

Bentleys were bulky cars. After the defeat of his lightweight racers at Le Mans in 1929 and 1930, Ettore Bugatti referred to the Bentley as le camion plus vite du monde (the world’s fastest lorry).

Opulence and Shoddiness

What do vintage cars have to do with the translation industry? Bentley and Bugatti are now both part of the Volkswagen Group. They outlived their glorious past by manufacturing luxury cars, and speed races are now the prerogative of other car manufacturers.

This might also be the near future of language services. To date, translation boutiques co-exist alongside with (relatively) large language service providers, but none of them offers luxury services in the same distinction as Bentley or Bugatti, despite the claims of the delusional advocates of the mythical premium market.

Why? Because it does not make any sense to show off a translation.


In many respects, when it comes to questioning the cornerstones of translation practice, the translation community finds itself instinctively compact, sturdy and resolute, despite any evidence against sustainability, as if hearing testudo!

Although its shielding barrier enabled both assault and defense, the greatest strength of a testudo was also its major drawback. Its complexity required perfect coordination to execute orders in the unison, while its density affected agility, at the expense of speed.

The Trilemma

The triple constraint is an archetypal management trilemma of trying to get something of high quality quickly and cheaply. The solution to the trilemma is in balancing turnaround time, number of defects, and cost.

Indeed, expecting customers to grant you all the time in the world for a standard product while paying the fee you demand without vetting your credentials and your work is an illusion, just like looking for something cheap, fast and good.

The balance to solve the trilemma is very hard to reach, and even to achieve two goals, planning and preparation are crucial to gather as much information as possible and clearly pinpoint, prioritize and present milestones to vendors and negotiate terms and conditions. The clearer and unambiguous these are, the better the odds.

The hardest part is always deciding how much to waive of each element of the trilemma to achieve a suitable outcome.

Whether you are on the customer’s or the vendor’s side, an open mind is necessary to be able to see things in your counterpart’s perspective and foresee their moves whenever possible.

In any production environment, the trilemma finds its synthesis in sustainability.

In this respect, when commenting Henry Ford’s alleged quote in The Ten Faces of Innovation, Tom Kelley was quite easily right to suggest not expecting to receive any help from customers to envision the future. In fact, customers might think they know what they need, but they hardly can tell what they want, and rarely do they know how to achieve it. However, this does not imply “educating the client” is a solution. On the contrary, with the marginal utility of translation getting lower and lower for at least thirty years, this is senseless.

Two Types of Innovation

Innovation is of two types: product innovation and process innovation.

Product innovation involves the development of a new product, the improvements to existing products or the introduction of new features into existing products, and it is generally the effect of technological advancements, less often of outdated product design, seldom of changes in customer requirements

Process innovation is the application or introduction of new methods to run business in a radically different way. It is not continuous improvement and it generally aims at reducing costs more often than driving an increase in revenue. Also, it is not less hard than product innovation for involving all the facilities, skills, and technologies of a business, and it can include changes in equipment and technology as well as in techniques and methods.

While product innovation is generally visible, process innovation is mostly visible only internally and it is typically less risky, unless it involves a change in business model, which may almost always result transformative.

In translation practice, process innovation has touched ancillary—albeit far from marginal—aspects of professional practice due to the introduction of supporting technologies such as TMs and TMSs and, more recently, MT. The typical workflow has remained virtually unchanged together with all the corollary of a century-old legacy. In contrast, there has never been any product innovation.

As a matter of fact, social media and the associated marketing evolution have been the greatest innovation of all, making smokescreens way much easier to raise and make tiny steps look like giant leaps.

The Educational Smokescreen

In recent years, periodically, academic institutions in the old and the new continent have been competing in announcing new educational programs to make future translators be able to succeed in an ever-changing scenario. Curiously, or maybe not, the names or profiles of heralds are no anticipation of innovation; indeed, these announcements seem rather a veil, which is more and more transparent, torn and jagged.

The emphasis on high-value skills and thematic specialization does not mean any student at any university is going to develop any of them any time soon. It might also be a “wise move”, in a marketing perspective, but it won’t lead to anything real and significant, and no praise from any prominent figures in the industry will make it such. By the way, sooner or later, these ‘figures’ should tell what are the more valuable tasks that linguists will perform instead of “doing words” and, since methods are nothing without tools, which tools these linguists should use to curate linguistic assets, build multilingual knowledge and engineer workflows. Especially when the figures above have never “done words” themselves. Ever. And most academics too. On the other hand, “He who can, does. He who cannot, teaches”. And he who cannot teach, leads.

MT has been around for over 70 years. The importance of language data come out nearly 20 years ago with SMT, when TMs had been already around for a decade. Now, every new day sees a new Solon voicing the importance of data in translation training.

Also, the last two and a half decades have seen plenty of books on internationalization and localization, even though only one or two of them are now part of the program of many university courses in translation and localization. This is a a non-exhaustive list:

  • Carter D. R., Writing Localizable Software for the MacIntosh, Addison-Wesley, 1991, ISBN 0201570130
  • Digital Equipment Corporation, Corporate User Publications, Digital Guide to Developing International Software, Digital Press, 1991, ISBN 1555580637
  • Apple Computer Inc., Guide to MacIntosh Software Localization, Addison-Wesley Publishing Company, 1992, ISBN 0201608561
  • Taylor D., Global Software: Developing Applications for the International Market, Springer Verlag, 1992, ISBN 0387977066
  • Uren E., Howard R., Perinotti T., Software Internationalization and Localization, TGP Consulting, 1993. ISBN 0442014988
  • Madell T., Parsons C., Abegg J., Developing and Localizing International Software, Prentice Hall, 1994, ISBN 0133006743
  • O’Donnell S. M., Programming for the World: A Guide to Internationalization, Prentice Hall, 1994, ISBN 0137221908
  • Kano N., Developing International Software for Windows 95 and Windows NT, Microsoft Press, 1995, ISBN 1556158408
  • Luong T. V., Lok J. S. H., Driscoll K., Internationalization: Developing Software for Global Markets, John Wiley & Sons, 1995, ISBN 0471076619
  • Bishop M., How to Build a Successful International Web Site: Designing Web Pages for Multilingual Markets at the National and International Level, The Coriolis Group, 1997, ISBN 1576101584
  • Hall P., Hudson R., Software without Frontiers: A Multi-Platform, Multi-Cultural, Multi-National Approach, John Wiley & Sons, 1997, ISBN 0471969745
  • Tuthill B., Smallberg D., Creating Worldwide Software: Solaris Int’l Developer’s Guide, Prentice Hall, 1997, ISBN 0134944933
  • Ott C., Global Solutions for Multilingual Applications: Real-World Techniques for Developers and Designers, John Wiley & Sons, 1999, ISBN 0471348279
  • Esselink B., A Practical Guide to Localization, John Benjamins Publishing Co., 2000, ISBN 9027219567
  • Schmitt D. A., International Programming for Microsoft Windows, Microsoft Press, 2000, ISBN 1572319569
  • Savourel Y., XML Internationalization and Localization, Sams, 2001, ISBN 0672320967
  • Symmonds N., Internationalization and Localization Using Microsoft .Net, 2002, APress; ISBN 1590590023
  • Maxwell Chandler H., The Game Localization Handbook, Charles River Media, 2004, ISBN 1584503432
  • Dr. International, Developing International Software, Microsoft Press, 2002, ISBN 0735615837

And yet, every day, still today, despite the many brilliant, smart, sharp, knowledgeable lecturers, teachers and professors, too many localization projects still suffer of many of the issues that all those books above deal with. Any more bullshit?

Now it’s the time of AI and its many deceptions. For example, without labeling, no AI program can tell a skyscraper from a spider and autonomous technology companies employ thousands of low-wage human workers to painstakingly inventory data and images and enable AI programs to learn. These labelers are the new sharecroppers in one of the most dynamic parts of the economy, as the authors of Ghost Work argue.

Similarly, the highly valuable tasks that linguists should perform instead of “doing words” (i.e., again, curating linguistic assets, building multilingual knowledge and engineering workflows) all involve a set of skills requiring much effort and time to develop, and should therefore be adequately remunerated, but they are just another form of digital sharecropping.

In The General Theory of Employment, Interest and Money, John Maynard Keynes formulated his principle of effective demand, according to which, “demand creates its own supply”. In this view, producers’ expectations on demand and consumption determine their investments and, therefore, the employment and production that derive from them. Can you see how this principle applies to the translation industry?

The Nastiest Taboo

Not surprisingly, none of the books in the list above came from academics, who are only interested in the liturgies and erudition they have got acquainted with over the years and how to perpetuate them, however pointless or obsolete and despite being peddled as science. On the contrary, the mere idea of doing anything different or considering anything different scares them to death. Their creed is “adapt and blend” rather than “challenge and innovate”: throwing a pebble in the pond could make a storm break loose, so forget about cultivating doubts and investigating them.

Not surprisingly, to paraphrase Orson Welles and Graham Greene, ‘harmony’ and stillness have produced inextricable, perverse, impractical metrics and wooden standards, and challenging them—i.e. crying out that the Emperor has no clothes—makes you a contrarian, if not an outcaste. And no one is willing to lend their ears to outcastes, especially in pompous events that are “places to discuss solutions”, not questions. After all, when those who should institutionally offer solutions cannot find any, stealing some is the only path to survival.

Therefore, the best advice is still to avoid the nastiest taboos such as quality, metrics and standards. Bullshitting and appeasement are best companions for everyone looking for a career, even though they do not lead to any innovation, any change, any real improvement and, possibly, any solution. And do not hesitate to talk about anything, especially if you know nothing about it. After all, everybody talks about the weather, but nobody does anything about it.

Again, innovation is a perfect example: the internal combustion engine was unquestionably a disruptive innovation; who cares if it was invented more than two centuries ago? The same goes for AI, as well as for MT, TMs, and so on and so forth.

The Challenges Ahead

If you think the most serious threat to translation practice comes from AI, you are terribly wrong. It comes from platform economy.

By offering a common technology framework for building applications and services, innovation platforms are those growing faster. Translation businesses might benefit from platform economy rather than being harmed by exploiting innovation platforms to create their own platforms or making use of existing third-party ones.

Unfortunately, the translation industry is behindhand also in this respect. The major risk in this case is that the drawbacks of the innovation economy go public, while rewards go private and that the translation industry may become one of the next source of ‘rent’ in the knowledge economy.

Innovation is cumulative and collective by nature; it is the result of long-term investments over the years, thus taking a long time: what today may appear as a radical innovation is the result of years of hard work by many researchers. It is also very tentative, as many attempts fail and many results are unexpected. And yet, as Robert Solow showed, improvements in the use of technology explained more than 80% of growth.

Can you tell of any innovation coming from translation-related research in the last two decades?