How about enabling multilingual search in the main search engines

There are those of us who are proficient in two or more languages or are wanting to become so. This is due to countries like Canada, Belgium or Switzerland or even parts of countries like the South West USA that are inherently or officially multilingual.

It also extends to societies that maintain a multicultural character; as well as people who are setting themselves about to learn languages in addition to their native one. In some societies, a desire to work across multiple languages has been enhanced through activities like the increased viewership of subtitled foreign-language film and TV content like European thrillers or Nordic Noir; or particular cultures bestowing attention to particular countries such as the gamer culture’s obsession with Japan being known for manga / anime and fast cars.

The current problem

But using Google or similar search engines may become awkward for those of us who are or want to be multilingual. Typically, you have to know a concept in a particular language if you want to see the results in resources based in that language and you only see those resources. But if you are multilingual, you may want to see the resources in the languages you are familiar with, even if you type the search terms in one of the languages you are familiar with.

What needs to happen is for a search engine to implement “on-the-fly” translation of search phrases from one source language to a multiplicity of user-chosen target languages. Then the search engine would show the resources that are natively written in the target languages.

At the moment, most search engines can work across dialects of the same language such as to understand American or British English, showing resources in either dialect.

Questions that can be raised concerning this idea would be to assure a grammatically-accurate translation of the source search terms, including where there are multiple equivalents specific to that language.

Handling language peculiarities

There are situations where source and target languages maintain particular peculiarities when referring to some concepts or objects.

An example of this would be a reference to the lightweight commercial vehicles which are described as a van if they are enclosed or a pickup truck in most of the English-speaking world or a “ute” in Australia and New Zealand in the case of those with an open tray. But the French refer to these vehicles as either a “camionnette” or a “fourgonnette” while the Germans would use “Lieferwagen”, “Kastenwagen” or “Transporter” for a van for example.

Similarly, there are loanwords that are used across multiple languages to mean the same thing although some languages like French cut back on the use of these loanwords to maintain language purity. It may be preferred to use the loanwords or the language-specific equivalents or both as search terms for searching within a particular language. The same issue can also apply to proper nouns where there isn’t a language-specific equivalent such as place names, trademarks or business names.

There is also the issue with some Asian languages like Chinese and Japanese which use different writing styles. This can cause problems if search terms are provided in one writing style but you are confident in using the other writing styles offered by that language and want to see resources offered in those styles.

Handling multilingual resources

As for showing results, some Web resources, typically resources written by organisations in or targeting multi-lingual areas, tend to provide resources in multiple languages. This practice has been encouraged in Europe since the adoption of the Maastricht Treaty which underscores the Single European Market under the banner of the “Are you ready for 1992”. This approach may be through a translation process that the author implements as part of their editorial workflow or some end-users simply “pipe” the resource through a site-wide machine-translation resource when they view the site.

A situation that can come up with some multilingual Websites is that the site carries more comprehensive information in the site’s native language or a few other languages than in the other languages. Or if the site is targeted to multiple countries like all of the European Union’s resources, the translations may be deeply localised such as to refer to governmental workflows specific to that country.

A search engine could allow the user to set preferences for multilingual searches such as preferred languages and / or language priority. This would mean that the user would see the results from resources written in the languages they specify; along with the ability to have certain languages appear first. The language priority could be fixed by the user or be determined by the search engine if the user supplies the search expression in a language-specific form. But if a resource carries translations, the user could see results from that resource in the highest-priority translation first plus a reference to their other chosen translations.

Similarly, a search engine could compare the amount of information that is available in multilingual versions of the same resource to identify language peculiarity or content richness.

User preferences concerning multilingual search

A search engine that implements individual user preferences such as being linked to a user account could implement a set of preferences for multilingual search.

This could be through a list of languages that the user knows so as to prioritise resources in those languages. Similarly, a user could determine whether to place a multilingual resource’s native language as a higher priority over the translations.

Providing a multilingual results list

A multilingual results list could have each native language as a sorting or grouping factor when ordering the results. It may also allow results that come from a multilingual resource to be identified as appearing in the chosen languages.

To cater for multilingual resources where there is a differing level of comprehensiveness amongst the languages. the user interface could identify which languages have more comprehensive results. It can also be used to call out translations that underscore area-specific terminology or colloquialisms.

Catering to language learners

Some users who are learning a language may want a multilingual search interface to provide features conducive to this task.

This may include the ability to show their “home” language under foreign-language headlines in a search list using a different typeface so they can build their vocabulary up for example. Some user interfaces like the traditional mouse-based interface or a touch-based interface that allows the user to dwell for more options may allow for a “pop-up” or similar translation. This can also apply to languages that implement an intermediary phonetic script along with one or more different written scripts.

An augmentation that can work with text-to-speech setups may allow for the user to have all or part of a foreign language read aloud. This could permit them to hear how the word is pronounced in the context of the sentence.

Conclusion

What needs to be provided with a multilingual search option is to accept searches in multiple languages and to show resources that are native to different languages in a search-results page.

It also includes dealing with multilingual resources including resources that are focused towards a few languages along with supporting a multilingual user’s preferences.

Send to Kindle

Leave a Reply

*