There Are Many Ways to Commit Suicide, But Post-Processing of Machine Translations Would Be a Horrible Way To Go

There are many ways to commit suicide, but all of them (shooting or hanging yourself, cutting your veins with a razor, or even taking a lot of pills) are pretty painful. That is why assisted suicide is now available for a price to ease the pain for those would prefer to check out from this valley of tears in some countries, (for example in Switzerland, or here in US in Oregon), because sometime living with an excruciating pain or a disease that robs us of our memory or cognitive abilities may be worse than death.

Suicide also exists in nature as a part of life under certain circumstances: salmon must return from the ocean, which is rich in nutrients, and swim upstream in a river to the spawning grounds only to die there in an environment that is poor in nutrients.

Perhaps this is how nature prevents overabundance of certain species of fish.

Societies also sometime commit suicide at a certain point in history. The suicide of democratic society is happening right in front of our eyes: when everybody is being secretly watched 24/7 in violation of existing laws, most importantly the Constitution, democracy is committing suicide and nobody seems to give it much thought, least of all the politicians who took a solemn oath to protect the Constitution.

Perhaps this is how human societies eventually die so that they could be regenerated and replaced by something that works better for most humans, rather than just for a few kings or billionaires.

It is also possible for an entire occupation to commit suicide.

In the classic American movie from 1946 “It’s a Wonderful Life”, a complicated, troubled but basically honest and likeable banker played by Jimmy Stewart is thinking of committing suicide and in one version of his life on the screen he actually does it by jumping into a freezing, cold river … only to be saved by an elderly, chubby guardian angel for another version of his on-screen life in black and white. Because this suitably uplifting and honesty celebrating movie is played every year around Thanksgiving and Christmas here on TV, I saw parts of it maybe four or five times already. I never finish it because I know how it ends, but some scenes I watch over and over again.

Unfortunately, honest bankers in this and many other countries have committed suicide en masse when they were bailed out by the taxpayers after their dishonest schemes unraveled in 2008 as a result of the mortgage crisis, a crisis that was caused by the greed and rampant illegality of rich bankers.

Although the great majority of the voters were very strongly against the bailout of the financial industry, that was irrelevant and the voters were as usual completely impotent. The big banks were given money at interest rates close to zero, money that was taken from the taxpayers, and the banks then turned around and lent the same money to cash-hungry people who were trying to survive in an economy that was ruined by the same bankers at rates that were at least 20 times higher.

Not a bad business model when it comes to profitability. But when it comes to morality …. one can see why Jesus decided to overturn moneychangers’ tables in the temple (Mathew, 21:12).

And that is how our great, capitalist financial system was saved. While socialism is a dirty word in this country, the bailout of the Western financial system in 2008 is a textbook example of how socialism was practiced for decades in East block countries – before the economies of these countries collapsed as a result of too many bailouts.

So as far as Joe the citizen is concerned, an honest banker of the type played by Jimmy Stewart in the film from 1946 no longer exists, it is only a distant memory of a dream from a bygone era, a dream that is revived on the TV screens in an old movie once or twice a year. In real life, if the bankers create another crisis by stealing too much money again, they know that the politicians will bail them out again.

*******************

Another profession that may be about to commit suicide is the translating profession. Not because translators are as dishonest and powerful as the bankers – quite the opposite.

Most translators generally do not make a lot of money, and they have very little power over anything. But if they fall for the latest scheme of some of our “translation industry leaders”, they will have no power whatsoever, not even the power to determine how much they should be paid for their work, and they will make even less money, probably much less money, than they are making now because they will no longer even exist as translators.

Translators already fell for the first hoax that was perpetrated on the translating community by “the translation industry” in the last decade. Many translators, although by no means all of them, believed the promises that Computer Assisted Translation (CAT) tools would make it possible for them to increase their productivity by a factor of 2, or 3, or more, and spent hundreds of dollars and Euros (I am told that Trados costs 800 dollars) for these tools. Once they started using them, they were told by many (although by no means all) translation agencies that from now on, they would only be paid a fraction of their usual rate, or nothing in some cases, for what in the industry parlance is referred to as “repeat words”, “fuzzy matches” and “full matches” determined by the CAT tools.

When I wrote a post in July of 2009 titled “Friends Don’t Let Friends Use Trados or Any Other Memory Tools”, only 103 people saw it, and by the end of 2009 it still had only 347 views. But after 2012, this became one of my more popular posts, several thousand people read it now every year and the total number of views of this old but not forgotten post will probably go well over 10,000 this year.

So maybe, just maybe, people are finally beginning to understand that all of this talk about higher productivity levels that would lead to higher incomes for translators was really just a smoke screen for an ingenious attempt by the “translation industry” to achieve greater profitability by slashing the fees paid to translators – without necessarily passing to the clients of translation agencies any of the savings achieved at the expense of the reimbursement for the work of human workers called translators.

But when some translators agreed to be controlled by the CAT tools instead of just using these tools (that they themselves have to buy!) for legitimate purposes, and only if they felt like it, this was not really a suicide for these translators – only the equivalent of shooting yourself in the foot. When you shoot yourself in the foot, out of carelessness and stupidity, you may be limping around for a while, but you will probably live if you learn that it is not a good idea to play with a loaded gun.

The second scheme, called “post-processing of machine translations”, is likely to be much more deadly to the translating community, because this tool was designed by our beloved “translation industry leaders” to simply kill off the translating profession and replace it by a new type of workers called post-processors of machine translations.

The word “post-processor” evokes in me the image of a processing line in cold halls of meat processing plants where workers in dirty white coats and nets on their hair, usually illegal immigrants, slash and cut to smaller pieces carcasses of slaughtered animals.

The nature of the work of post-processors who work on the detritus created by algorithms stored in machines that is generally referred to as machine translation would not require the use of sharp knives, nor would it require the use of a sharp brain, based on the design of this new profession.

Anybody who has or claims to have some knowledge of a foreign language would qualify as a post-processor because as far as the designers of this new profession are concerned, most of the work has already been done by the machine. The compensation for this activity, which is comparable to basic computer data entry, would thus be commensurate with the low skill set level required, and much of this post processing work would be done in third world countries to maximize the profitability of the economic model for the owners of these enterprises.

Should real translators agree to cooperate en masse with this interpretation of what their profession is or should be in the future, an entire occupation would be wiped out and replaced by new, post-processing human robots.

What a horrible way to go.

The post-processing design of human-assisted post-translation of what machines armed with software regurgitate when they are done analyzing communication between humans will inevitably generate very poor quality of translation, if we can still call it that. But “translation industry leaders” know that, and they are ready to attack this problem by offering lower prices for this product, significantly lower than what customers have to pay for human translations. There will still be plenty of profit left for them given how little the post-processing human robots will be paid.

Can this scheme work?

Yes, I believe so. In some translation fields, where “the translation industry” is already using or trying to use “cloud workers” (also called “clown workers”), a similar concept is already being implemented. The clown workers are not really translators, they don’t need to have attained a certain level of education and specialization, as long as they say that they “know a foreign language”, and as long as they are ready and willing to work for peanuts.

There should be plenty of clown workers like that on this planet! There are over seven billion people living on planet Earth and about half of them are starving. Let’s send in the clowns and get rid of translators!

But I do not believe that the post-processing design can possibly work in specialized translation fields, such as my field of patent translation.

Trying to apply this simplified model of what used to be called human translation to patent translation would result in a lot of garbage that can no longer be called translation.

When I translate a patent application, which is something that I have been doing for almost three decades, I generally always print out a machine translation if it is available. But I basically use it only as a dictionary, not a translation. A real translation must take into account a number of things that machines don’t understand.

For example, since most patents already have an English summary prepared by humans who sometime knew what they were doing, and sometime did not, I have to try to use the terms already existing in these summaries because these are the terms that my clients use to communicate with patent lawyers who may live in different countries and speak different languages.

But the most important thing to understand is that machine translation is not really a translation; just a suggestion made by a dumb machine. Google Translate sometime makes very good suggestions because it picks a similar translation, originally produced by a human translator, if it is available in its huge database. But because even Google Translate has absolutely no understanding of the meaning of the text, it will often generate hilarious mistranslation. These are easy to spot, but more subtle mistranslations can be very difficult to detect.

The concept of post-processing of machine translation is thus based on a faulty premise because it is almost always faster to translate anything, or at least complicated and highly specialized texts, by human translators from scratch. Post-processing is a just a scheme designed to get around the problem of high cost of human translation.

The scheme may work, to a limited extent, on very simple, repetitive, relatively unimportant texts that still need to be translated, regardless of how “unimportant” they are.

But it is not going to work for example in the field of patent translation where companies may be fighting out awesome battles in court for years to come, battles that may be in the end determined by a slightly different nuance of a technical term used in a patent application.

The key to surviving the coming wave of professional suicides, figuratively speaking, among translators who will decide to switch professions, ditch their old profession and become “post-processors” of output that was generated by machines is thus clear – specialization in highly complex texts, which even our “translation industry leaders” are likely to leave to highly educated, highly experienced and highly qualified human translators – unless they are much dumber than I thought.