Google translation is coherent not

A bad translation by Google sparked some hot-stove baseball rumors. The story, which seemed to say a highly sought Japanese pitcher will go the the Arizona Diamondbacks, may have been the product of repeated bad translations. One thing is for sure: Translation services that don't provide information about sources and how the language may be interpreted differently are not market-ready.

Being a close observer of baseball, I was intrigued to see a scoop that appeared to have come from Google's translation of a Japanese baseball site. Baseball Musings pointed to a Google translation of a page on, an MLB-licensed site, discussing the posting of Seibu Lions pitcher Daisuke Matsuzaka.

Posting is a bid process that Japanese teams use to auction the rights to negotiate with a player, allowing the team losing the player to earn a windfall from the transaction, even before the player begins to talk salary. For baseball fans, whoCalling this "translation" is like believing Borat is a serious journalist. spend the winter talking about the business of the game, this kind of information is a rich meal--when it makes sense, that is. Where Matsuzaka will pitch in 2007 is one of the hottest topics this winter.

The translation, which, if it can be said to be readable, reads:

(the bid system) with the Matuzaka Daisuke pitcher who aims toward measure transfer (Seibu) proposal of the measure club for was closed up on the 8th. It is seen that at present time New York [yankisu] and New York [metsutsu] etc are powerful, but coming here, also the jackpot the team which can be called it surfaced as ahead Matuzaka's transferring. It is Arizonan [daiyamondobatsukusu].

The local paper [isutobare] tribune of the D that backs (electronic edition), with 7 dates the team proposal will do in Matuzaka, it reports. Scot [borasu] it is Matuzaka's representative, also the representative of the young promising stocks [suteibun] [doru] shortstop who this season you measure debuts having served, is wanted bearing the future of the team with 2 main columns of the ace right arm [burandon] web pitcher of 27 years old, conveyed the fact that among other things the Far Eastern charge scout has extolled Matuzaka.

However as for the same paper, there is a limit in entire annual salary amount of the D backs, as for the possibility [yankisu] and [metsutsu] and others exceeding the D backs with bid amount it points out that it is high. The question mark is attached to actualization characteristic of Matuzaka acquisition. 

Who knows what's going to become of Matsuzaka, the MVP of Japan's World Baseball Classic-winning team, but the point is this is a far cry from adequate translation of information. If we're really going to weave a global society with real transparency on the foundation of the Web, we have a long, long way to go.

Mundane topics, like a baseball fan's interest in where a player will pitch next year, are the building blocks of a global community. With the question mark still being attached to the actualization characteristic of the acquisition, the Web leaves most folks gesturing dumbly, like a Star Fleet crew whose universal translator is broken.

With garbled information, we're more likely to become confused and misunderstand one another. Google translation plays the role not of interpreter but comic. Google's playing the role of a kind of Borat making fun of how little we want to understand others. Calling this "translation" is like believing Borat is a serious journalist.

Google should consider taking this translation service, which it describes as "state-of-the-art," out of beta until it works, because it doesn't help people find or understand information, the company's much-vaunted mission.

In fact, the mistranslation may be a compounded engarblement, as Baseball Musings points out. The page might be a bad translation of an article on Rotowire, a U.S. fantasy baseball site. In fact, it appears the scoop that the Arizona Diamondbacks had won the Matsuzake bidding is only a story about the fact the team bid. The apparent importance of the story is a product of the mangled language and nothing more.

We need better sourcing information in the metadata associated with pages so that loops of nonsense do not become news or accepted "fact." Translations should be annotated with information about the source and links to examine the meaning of phrases. Then, perhaps, we will all be on that jackpot team which can be called it surfaced.