The Slow Death of non-English Languages by Navigation Software


Ok, that title is a little off. What I really want to do with this post is help to make it a priority for companies that make navigation and or text to speech software a little less English specific (or whatever language it's built for). What am I talking about?

Lots of people use GPS navigation services either built into their car or on their phone. Those navigation services almost always read out the directions. "In 300 feet, turn left at main street" for example.

The problem is that text to speech software doesn't understand pronunciation and so arguably it's slowly killing off the correct pronunciation by teaching everyone, especially children, the wrong way to pronounce things.


In Hawaii, "Turn left on Kalakaua". "Kalakaua" should be pronouced Kah−lah−kah−ou−ah where the "ou" rhymes with "you". But the navigation system says Kah−lah−cow−ah 😞

I'll probably get this wrong myself but in Daly City, "Turn right on Junipero Serra Blvd". "Junipero" I believe should be pronounced "Juan−ip−eh−row" but the navigation system says "June−eh−pear−oh".

In Temecula, "Turn right on Via Princessa" the "Via" should be pronouced "Vee−ah" but the navigation system says "Veye−ah". You could even get pickier as if I understand it correctly "Princessa" should have the 'r' slightly rolled but the navigation system doesn't do that.

I'm a little bit surprised some people haven't made this more of an issue. I know at some level you can argue it's not important but at another it's arguably destroying culture. Millions of children will grow up with GPS navigation systems and Siri and Google Voice search etc. They'll learn from those systems. As long as these systems get the pronunciations wrong they'll be teaching those kids the wrong things which will over time arguably lessen the number of people who know the correct pronunciations.

I'm confident it's not intentionally this way. I'm hopeful it's on a TODO list somewhere. But, how many years do we give it? At what point does it go from being a low−priority TODO item to systemic marginalization of minority cultures?

I'm sure it's relatively hard problem. It seems like every word/name would need to be tagged for a language it comes from. I'm sure there are even words that have the same spelling but different pronunciations depending on the region, language or country. But still it seems like it should be done. Here's hoping the various navigation and text to speech system engineers can take this seriously. Who knows, maybe it would also help speech recognition to have all this info.

Comments Needs to Update!!!
Helsinki, Assembly 2013