To me all it taught was just objects, nothing along the lines of how/what to refer to objects/people/etc.
For example, it showes a picture of a ball, book, and automobile then tells you what they are. Nothing like "Can you please tell me what the book you are reading is about?"
I actually found some video clips from the 80's that taught me more then what Rosetta stone did
Huh... Did you have an older version? I have ver. 3.xx. Even in the first lesson of the first unit they are forming at least simple sentences.
Yeah you have to go through a bunch of object identification drills first, but that is so you will have the components to form sentence.
Example (from memory so my Romanji spelling might be off):
So first they do bunch of drills to teach you man, boy, girl, woman..
i.e. man = Otoko no histo
Then they teach you plural by showing pictures of groups men, women, etc...
i.e. men = Otoko no histotachi (ahh so adding tachi to the end of a noun denotes plural. Got it.)
Then they teach you water...
i.e. water = Mizu
Then they show you pictures of people drinking various beverages....
i.e. drinking = Nonde
Then they show you pictures of people running....
i.e. running = (language filter) H a meatball t e
Then they start putting it all together in simple sentences along with contextual pictures:
i.e. "Otoko no histo wa, mizu o nonde imas." or "Otoko no histotachi wa, mizu o nonde imas." or "Otoko no histo wa, (language filter) h a meatball t e imasu."
I can infer that "imasu" is maybe like "is" or "are".
So now I can infere the simple gramtical structure from usage. It looks like they talk like Yoda.
I'm seeing "Subject wa, [object] verb is/are." At least for this simple sentence pattern.
Why don't they just passively feed you all the information on a platter? I believe it is because the act of forcing you to think and decipher the meaning is from partial information causes you to form a stronger mental association than if you were just being passively fed or just reading lists of grammer rules out of a book.
Think about how a young child learns to speak its native language. Its shown objects and told the names. For a long time it has no grammer to string them together but just builds a vocabulary. Later it learns to string together simple sentences out the vocabulary components it has collected. In time the vocabulary and grammatical complexity increases. By the time they go to school they already have a fairly strong grasp of the language through inference and intuition long before they start to learn any formal rules. I think that is a good approach to copy.
But like I said, use what ever works for you.
Regards,
Wab