I'd been curious about the Rosetta Stone software for a while before I decided to try it with Spanish. Their ads make some bold claims about "dynamic immersion" but don't give much detail about how it works.

It turns out that what they've done is quite clever. By using only pictures and words in the target language, the learner's native language is irrelevant. Instead of having to make separate versions of the software for an English speaker learning Spanish and an Italian speaker learning Spanish, as is the case with most any book out there, anyone can use their software for any language. Think about how much money that saves them!

So how does it work? The software's courses have a multi-level hierarchy. For each language, there are levels 1, 2, and 3 in some cases, which are all separate products. Within a level there are about 8 units, and within each unit, 10 chapters. A chapter has a set of words or phrases and pictures that match them. The software essentially gives you many different ways of learning these phrases, which correspond to the language skills of reading, listening, writing and speaking. First you can see the phrase written out while hearing it spoken, and match it to a picture. Next you can see the phrase without hearing it spoken, or hear it without seeing it, and again match it to the picture. Next you speak the phrase into a microphone after hearing it, and the software will actually grade your pronunciation. Finally you hear the phrase and then type it out. Within this sequence the software offers several variations, so that you have nearly a dozen different exercises you can do in the course of learning the chapter. It's entirely up to you which ones you actually do, and there is a gradual curve of difficulty within each chapter.

The material starts with basic nouns for physical objects, like girl and boy and ball and car. Soon adjectives are introduced, and then whole phrases and sentences that describe something going on in the picture, like "The man is putting on a shirt" or "The girl catches the ball." There are chapters on numerals and all the basic segments of vocabulary that you would expect. And that's really all there is to the structure of the Rosetta Stone's courses.

At first I really liked the software for two reasons. First, for learning pronunciation, I think it is unparalleled. Listening to a native speaker over and over and then having your own pronunciation graded is the best way I've ever experienced to learn to truly pronounce the language correctly. In a class setting, I realized, there is often the way the students speak and the way the teacher speaks, and the students tend not to speak like the teacher due to an ironic embarrassment at appearing too eager. The good students get to a certain level and stop there. Even if you know a native speaker, they may not have the patience or the desire to help you perfect your pronunciation. With the Rosetta Stone you're in an isolated environment where you can really concentrate on the sounds with no embarrassment factor. (The grading of your pronunciation by the software isn't perfect, and is quite sensitive to background noise, but it's still pretty darn good.) Second, the multitude of exercises gives you many ways to experience the words, and I found this extremely helpful for retention.

After a while though, I started to see the downsides. First, the repetitiveness. You're just clicking the same images over and over, and eventually I found it unbearable to pay much attention. This is clearly not so good for retention.

Second: you might be asking yourself how the software handles complex grammatical issues such as the difference between ser and estar. Well the answer is, it doesn't, at least not as far as I went with it. There are no grammatical explanations of any kind; only a limited set of examples, from which the student is expected to pick up the pattern. If I hadn't already studied many topics in other places, I would have had no idea what was going on with anything but the simplest elements of the grammar.

Third, again as far as I went, there were absolutely zero examples using the first or second person. If the Rosetta Stone were my only source, then months into my study I would have had no idea how to say my name, or say that I or you or we were doing anything. The words "yo" or "tu" simply did not appear. This seems insane to me.

Fourth, the software gives you no opportunity to ever construct your own sentences. I don't think any more really needs to be said about that.

One of the tenets of the philosophy behind this software is that you're learning a language the same way you learned your first language as a child. You simply absorb by seeing and hearing. But there's a few flaws to this. For one, our brains as infants and toddlers are very different from our adult brains. They are not as developed, but they absorb like sponges. Second, it takes us many years to learn our first language! Most people in the Rosetta Stone's target market are probably not willing to wait that long. It's just really not an appropriate analogy to make, though it seems attractive before you try it.