Consider the Luddites

by Jost Zetzsche

ve been reading up on the Luddites lately, those anti-industrialism protesters of the early 19th century. It makes for some great - and painful - reading. Like the story where three Luddites ambushed the mill-owning Luddite hater William Horsfall with a shot to the groin to avenge his threat that he would "ride up to his saddle girths in Luddite blood." He died soon after. As a matter of fact, so did the three Luddites (and many others), especially after the destruction of machines—their main target—was made a capital offense. As we know, the short-lived Luddite uprising came to an end and technology continued to develop. And while the Luddites were proven right that some jobs were lost, many more and varied jobs were created in the process.

The attack on Horsfall

In the mid-nineties of the 20th century, however, Luddism experienced a renewal of sorts. This time it was not directed against mechanized looms but against new technologies, in particular the personal computer and the Internet in general. Here is an interesting quote that I found from an NPR program in 2001 at the height of Neo-Luddism:

Some 85 percent of us are uneasy around machines. About a third of us are what's called "resisters." That anxiety might explain why the use of the word "Luddite" has grown so much over the past decade.

Hold that thought.

This week the New York Times published a much-commented-upon article on Google's machine translation efforts. It was posted on their digital homepage for almost 24 hours and was the fourth most popular article according to the Times' own record keeping:

Most popular articles in the New York Times on March 9, 2010

Clearly machine translation is not only something that people are using but also something they're interested in. The article describes the quick advancements of the quality of Google Translate due to the large amount of data that Google has access to (and also correlates it in a separate but related blog entry to the data that comes—or will potentially come—via the Google Translator Toolkit that I discussed in an earlier issue of this journal.)

Coincidentally, I just spent this past week working on a project within a translation environment tool (TEnT) with a strong link to Google Translate. Years ago I tried to use machine translation for a project and quickly stopped after the first fruitless minutes. Now I had a chance to give it another shot. Long story short: I still didn't like it—when it came to complex text. When I had to translate short segments such as software commands, Google Translate was helpful in eight out of ten cases and right on target in maybe six or seven out of ten cases. But again, as soon as I had to deal with more complex segments, the machine translation proved to be unhelpful at best and distracting at worst.

Was it better than when I first tried to use this technology ten or so years ago? Absolutely. Will I use it again? No, because I have excellent translation memories and terminology databases that I can usually use with much better results. But if I have to use a tool like the one last week where I can't access my own data and the text lends itself to machine pre-translation? Maybe. Either way, it began to make some kind of sense to me that virtually all translation environment tools now have a direct and very unashamed link to machine translation. But like I said, it made some kind of sense.

What I also realized again last week is that we have many translators who secretly (or not so secretly) side with the Luddites—not with the crotch-shooting kind, mind you, but with the technology skeptics, with the above-mentioned "resisters."

And in fact, most of us can probably sympathize with this notion by a thoughtful Neo-Luddite:

Luddism favors a thoughtful use of appropriate technologies that does not damage those relationships we hold dear.

Here's what I think: Technology is fine if it remains a tool, and it's fabulous if it's a useful tool. But developments in technology are moving so very fast these days that it seems we often don't have the time or the space to evaluate the usefulness of a tool in creating a better product. That's the part where we need to be cautious.

One thing that we can learn from the mistakes of the Ur- Luddites is to remember that jobs will not go away because of technology. Ben Sargent makes a convincing point on machine translation and its positive effect on the future of the human translator. And consider this: The only way that tools like Google Translate can be "successful" is by feeding them good, high-quality data, data that has been and will always need to be translated by you and me. We can confidently use the future tense in this claim because of the quickly changing nature of language. The language of today will be outdated tomorrow, so we will need to continue to feed the monster . . . and continue . . . and continue . . . and continue. . . providing plenty of jobs in perpetuity for all our neo-Luddites!