Google edges toward Rosetta Stone status

Google is making a new move to lower language barriers, offering the ability to translate search results from one language to another.

The search giant is in the process of adding the feature to the "show options" button that shows at the top of search results page. "We've offered this feature in Google Translate for a while, but now we're integrating it fully into Google search, making it easier for you to find and read results from pages across the web, even if they weren't written in a language you speak," said Maureen Heymans, the project's technical leader, and Jeff Chin, its product manager, in a blog post.

Clicking the option can dramatically change the results you see. For example, my ordinary search for "Taipei Museum of Fine Art" produced mostly English-language results. The translated results, though, featured Chinese Web sites with a different perspective (see the result below). Among other things, there was a Chinese Wikipedia entry--also conveniently translated by Google when I clicked the link--where there is none written in English.

Translation is an obvious tool in Google's mission to organize the world's information and make it universally accessible, but it's not technologically easy. The company applies its large-scale data-processing methods to the task, with machine-learning algorithms that compare the same passages of text in different languages.

Language is at the heart of information transfer among humans. And because so much of Google's effort to understand the Web is based on what people are saying, it's got to be a central focus at the company. With automated captioning of speech in YouTube videos, transcribed voice mail in Google Voice, and the combination of speech recognition and machine speech in the phone-based Google 411 service, Google is understanding speech as well as text, too.

The end result of all this speech and language work is that Google, a company dominated by massive data centers, numeric analysis, and engineering, is one of the best at offering a human touch.

It's now got the ability to translate among dozens of pairs of languages in an ever-growing matrix of combinations. For example, "automobile" in Croatian translates to "karozzi" in Maltese. Or at least Google Translate assures me.

Google already has been adding translated results to search pages in several languages through a program called Cross Language Information Retrieval (CLIR). For example, a Google search in Russian for Tony Blair's biography will present an option, in Russian and located at the bottom of the search results page, to search pages written in English. Clicking on a link then translates the English page into Russian.

Google has greater ambitions yet in translation.

"We will eventually do 100 by 100 languages, to take this set of languages and convert to another," Google Chief Executive Eric Schmidt said in a 2008 talk. "That alone will have a phenomenal impact on an open society."

Google's dictionary
When it comes to word comprehension, Google has long offered a feature I use two or three times a week to find word definitions. In the Google search bar, you can type "define:disinterested" and "define:uninterested" to compare the two words' meanings.

That feature only provides links to outside sources, though. Now Google is offering its own dictionary, a move noted by the Los Angeles Times. For a comparison of the dictionary and "define:" results, check the illustrations below.

No doubt this will cause some indigestion at sites such as Merriam-Webster.com, Ask.com's Dictionary.com, and perhaps Wiktionary. Unlike those first two, Google doesn't--at least yet--show ads on its dictionary results.

The Google dictionary offers synonyms and in some cases images drawn from around the Web. That can produce some indifferent results. One of the two pictures I got for a Google dictionary search for "pismire" was a black-and-white cartoon drawing of three ants.

If the dictionary doesn't have any of its own results, it shows what the "define:" operator produces. If there is a Google entry, those results are shown below. Because that can often veer away from just the basics of definitions and pronunciation, the dictionary feature sometimes functions somewhat like an encyclopedia, or at least as a referrer to encyclopedia-like content elsewhere on the Web. That's been done in traditional dictionaries, too, that have entries to describe things like "Bergen-Belsen" specifically and not just define "concentration camp."

Google's logophilia
A close understanding of words is related to many parts of Google operations.

Spell checking, a useful if imperfect service computers offer, also is beginning to arrive in the world of Web-based applications. For example, Google has touted the spell-check abilities of its Google Wave service.

Understanding synonyms is relevant to both translation and dictionary results, but perhaps more significantly for Google, it's part of the underlying processing that interprets what people mean in their search queries.

Note also that Google can translate Gmail subject lines. The world abounds with language barriers, but Google has significantly lowered many of them.