I will give my answer to my question, the way I see it, but bear in mind that I am not a linguist, or even a mathematician or a physicist, I am just a computer programmer who likes to learn about futurism.
In my opinion, we can definitely conduct any scientific knowledge we wish, as long as it is not related to our psychology and culture.
As showed by the voyager spacecraft, and the Arecibo message, which I didn't thought of at first. But yeah, we start by a calibration circle, then with some universal constant, then we define all the math and physics, we show chemistry of life, pictures of blueprints and stuff. Sure, if it is aliens that find the info cache, they would have a hard time figuring out how we used some of our inventions, but they will have the inventions none the less, and with a good system of organization for all human knowledge (witch we still lack by the way), we can describe it all, we can even describe the syntax of a language by using the Backus–Naur forms.
But we will not be able to conduct the meaning of most of the words of a language, especially of verbs and adjectives, as those tend to be mostly related to our inner motivation and our perception of the world. We can try, and with some success probably, but their understanding of our culture and the way our minds work will be, at best, very rudimentary and misguided. That is, until we figure out a way to perfectly map human brains into neural networks they might simulate, in order to have the next best thing to living humans, that can be thought their language and teach them about our culture, and while at that, why not save genomes, that can be used to recreate the actual human beings?
Bottom line, I believe that most of our scientific knowledge can be preserved in a readable manner for other species to decode, but our culture can't, nor can languages be thought well without translation to beings that are not very similar to us - that is, human languages that involve more than just pure logic - programming languages and other descriptive languages can.
On the other hand, I don't know how feasible it is to teach humans a language by examples, but I think it is very difficult, even if somewhat feasible.