Rosetta Stone: Not Quite There Yet

Recently I downloaded the Rosetta Stone app for Android in great anticipation. I think the Rosetta Stone concept is fantastic (putting an immersion type approach to language learning into a CALL environment), and having a mobile version of the Rosetta Stone system just makes sense in today’s mobile world. However, all that glitters is not gold and to my great disappointment I found this to be the case with Rosetta Stone’s current offering.

At first glance, the mobile app seems to be all that it claims. Rosetta Stone on a mobile. However after a bit of testing, I found certain key functions did not promote language learning as expected.

For my first test, I excitedly dove right in to a relatively new language for me, Arabic. I opened the Arabic language learning option and found the typical (to anyone familiar with Rosetta Stone) beginners course. It starts by teaching the target language’s words for boy, girl, man, and woman. The Rosetta Stone system first provides pictures and pronunciation teaching the target language, and then a picture matching exercise where the learner matches the correct picture to the audio spoken, followed by a pronunciation practice stage. It was in the pronunciation practice stage that things began to go awry.

At first it seemed I was an expert in Arabic pronunciation… perhaps even a little too much of an expert… My hackles of suspicion rose. So, I tried an experiment; I purposely began pronouncing differently from what I was being taught.

RosettaStoneAdAt first I made slight changes, then gradually larger and larger mispronunciations. It seemed I could not go wrong! It was virtually impossible for me to mispronounce a word unless it was entirely off in syllable count. Then I tried purposely pronouncing the Arabic word for girl, when I was prompted for boy, and vice versa. To my great chagrin, I was rewarded with approval for my completely and contextually wrong pronunciation. This, with only my first two Arabic words!

Because of my determination to solve the riddle (for I was sure it was a riddle), I proceeded to test other languages which I was already familiar with. I tried the course in Japanese (which I have a fairly strong command of), and my native language of English (American form). Again, I was met with similar results. It was nearly impossible to mispronounce words.

Then I had an idea; perhaps there was a setting I was unaware of.  So, I searched the apps menu to find the culprit setting. I was somewhat relieved when I discovered a setting for accuracy relating to pronunciation checking. It was preset to easy. Though I am not sure that entirely wrong pronunciation constitutes “easy,”  I adjusted the setting to the highest difficulty level and tried again. Much to my disappointment, there was little if any change to the programs operation. I finally retired the app.

It seems to me that from a programming perspective simply making a call to Android’s build in speech-to-text feature, would enable the Rosetta Stone app to function better than it does now.  When simply using the phones built in speech-to-text software, I have to pronounce things very clearly to get it to recognize my intended phrase. In fact, here is a very simple Android app which attempts to teach pronunciation doing just that.

RSad2So what was the programming decision, or perhaps marketing decision, which brought about the offering currently available?  And what is the purpose of a piece of software that fools someone into thinking they are learning a language, when in fact they are learning to speak gibberish? Well, that will have to remain the domain of our imagination. But, whatever Rosetta Stone’s reason’s are, one thing seems clear. Our jobs as language teachers are still secure.