I'm an American citizen living in Montreal, Quebec. I'm also an activist for information rights, including privacy and free expression as well as Free Software and Free Content. As such, I've been particularly interested in the community-oriented Web sites of Canadian information rights organisations, and particularly how they reach out to French-speaking people around me in Montreal. I think that most community Web sites address French-speaking audiences poorly, for no fault of their own.
The problem
Two such community Web sites that I've seen recently are:
Both sites are monolingual, in English. However, about 1/4 of the constituency that they're reaching out to -- Canadians -- are native French speakers. In Quebec, where I live, it's about 85% of people who speak French as their first language.
Now, many francophones who are interested in the Web and in information rights have by necessity learned to accommodate and read and write English, usually quite fluently. But there are a lot of French-speaking Canadians who are functionally monolingual. Many of these people could be sympathetic to our ideas if they were able to access and understand them.
Even for bilingual readers, English-only content has a very clear implicit message. It sends a very clear signal that they are not the target audience -- that they are at best an afterthought. English-only content says that it's up to French-speaking readers and not Web site creators to overcome the linguistic mismatch between content and audience.
If you are trying to convince someone that you care about them and their rights, failing to put that message in their native language puts you at a grave disadvantage.
Further complicating the problem is the issue of Quebec sovereignty. Of French speakers in Canada, more than 80% live in Quebec, and somewhere around half of Quebecois are sympathetic enough with Quebec sovereignty to vote for a sovereignist political party or to vote "yes" in a referendum. Many of those who support federalism still see Quebec as distinct and functionally separate from the rest of Canada. In other words, many Quebecois readers are not Canadian-identified.
If your Web site has the .ca TLD, or uses "Canada" in the title, there's already a disadvantage for reaching people who don't think that "Canada" means them. Having English-only content will seal the deal. For Quebecois readers, anything English-only is obviously for eux autres in the "ROC" (Rest of Canada).
The main point here is that English-only Canadian community Web sites miss a lot of opportunities to reach a sizable minority of potentially sympathetic Canadians.
A sub-optimal solution
So, what can community Web sites do to reach French-speaking Canadians? One possible solution is to provide parallel French and English sites. This is typically what we see with federal government sites like the CIC or with online retailers like FutureShop. There's a link that says "English speakers go this way" and another that says "French speakers go that way". Usually there are links between equivalent pages, such that an English-language page on water safety has a link to the French-language equivalent.
I think there are some serious problems with this strategy for community and grass-roots Web sites.
- It's extremely labor-intensive. Having a translator translating all of your content from English into French is very expensive if you pay, and it burns out volunteers very quickly. Most people find translating text to be a meticulous, tedious, mind-numbing task, and those people who enjoy it and are good at it come at a premium cost.
- It makes second class citizens. I've had a number of francophone friends complain about the obligatory translation into French on food packaging or for other products. Usually the translations are quite bad, so as far as they can see, they're not actual participants but a market that's being mollified. For people who do the translating, it's even more stratifying; you receive your orders from your English-speaking superiors and translate the text for your lowly francophone brethren. In either case the lack of empowerment isn't likely to excite francophone readers to action.
- It's hard to maintain. This rolls out of the labor intensity, but it's worth pointing out separately. Bit rot is a serious problem for Web sites, and it doesn't take long for a page that was once fresh and new to get out of date. If your intention is to let French speakers know that you care about them and that they're not an afterthought, pages that say Watch here for our upcoming 2002 news and events don't really convey that.
- It's separating. Starting off with an online retail or government Web site, there's a distinct feeling of being shunted into a particular cattle stall away from some other breed or species. Raising a wall between two communities keeps either from gathering enough critical mass to effect change.
- It's top-down. This is OK for government and corporate sites, of course, but sites that depend on user-generated content or user comments have a real problem if they're bifurcated into separate linguistic halves. Do you allow conversations to grow naturally on the francophone side? Or do you just translate the English comments into French?
I think that all of these factors combine to make the government and business solution to the language problem on the Web distinctly sub-optimal for community and grass-roots sites.
I think the immense costs of maintaining a bifurcated site is one of the main reasons most non-profit and community sites don't bother with bilingualism at all. But I don't think that this kind of site is the only solution.
My ideas about bilingual Web sites
All of which is a prelude to a description of what I think could be a very useful strategy for building bilingual community Web sites for Canadians.
- Keep bilingual content on the same site and if possible on the same page. A mixed-mode Web site gives a consistent reminder to both English-speaking and French-speaking readers that both languages are supported. It supports the idea that all Canadians can participate in the same conversation, whichever language they are speaking. No other mechanism can present a unified community like mixing content on the same page.
- Register domain names in both languages. Having a domain name in both languages makes it easy for French-speakers to pass along your URL by word of mouth.
- Make site navigation available in either language. Most dynamic content management systems allow building menus and other navigation tools in multiple languages. Just letting people navigate in their own language, even if content is in one or the other language, can be a big usability win.
- Generate original content in both languages. This gives both languages a first-class status on the site. It can be kind of hard to find people who are willing to contribute in the other language, though; it's a chicken-and-egg problem. However, bringing French speakers onto your core team can be a real boon for other reasons.
- Use wiki techniques for translating content. I think this is the key innovative idea here. Wiki editing -- letting readers edit pages -- distributes a boring and tedious task among lots of participants. It provides a way for people to alleviate a problem quickly and painlessly, and for later participants to refine and improve a sketchy first draft. It allows microcontribution of effort to get a useful task done; it doesn't cost you anything, and it doesn't burn out volunteers.
I think these simple techniques can potentially make for extremely effective bilingual community Web sites. There may be some technical difficulty with identifying different parts of a site with different language tags but I don't think it's impossible (see Content-Language in HTTP for one technique).
Most difficult, I think, is exposing a wiki interface for translation without making your entire site a wiki. Most content management systems either are entirely wiki-based, or they have little or no wiki features. Putting a wiki block into a page that could be edited by any reader is possible, I think, with Drupal, but not easy. I'd like to work with some CMSes to see if it's possible to add this pattern to them.
What do I know?
But, then again, what do I know? I'm just a carpetbagger spending my time in a foreign land; I should probably avoid telling those people how to fix their problems. But I hope that some of the ideas I've given above can be helpful for people who are advocating for freedom online in reaching some hard-to-hit populations.