One of the biggest problems facing many learners today is the problem of incidental vocabulary. One of the prevailing themes in education is the preference for so-called "authentic materials".
"Authentics", to use a bit of teacherese, is just another word for "native materials". Except you're allowed to doctor them a bit without making them less authentic. Once you hit upper-intermediate, you will be subjected to more and more authentic materials. Your lessons will be based around texts (that damn word again) drawn from newspapers, magazines and websites, or excerpted from novels or non-fiction books, and will be arranged in chapters based on themes like science and technology, arts, education and the like.
But there's suddenly a problem here. Throwing together a heap of articles that are related in topic is all well and good if your goal is to study the topic, but if you are aiming to study the language, then the relationship between the articles is entirely superficial. The structures that you are looking for just won't be there.
The simplest manifestation is in vocabulary. Key vocabulary in that would be incidental for a native sends many a learner scuttling for his dictionary. So a word appears once only in the article, and never again in the entire course -- that dictionary time is wasted as the learner does not learn the word. In a textbook chapter on technology, you can switch from satellites to biometrics to textiles. Each of these semantic domains has a vastly different stock of basic words. The learner is left trudging through a bog of heavy, unknown words, and looking back later those are lost in a fog, never to be recalled. A cursory look back may occur in a programmed "revision lesson", but it's a random sample of a fraction of the language covered, and usually limited to matching exercises (word+definition, sentence halves or question+answer) that tacitly admit that the student isn't capable of recalling the word without heavy prompting.
The same occurs with grammatical patterns, where a pattern may appear and be taught in one text, but then subsequent texts don't support, revise or otherwise consolidate that structure. On the other hand, it's in the grammar that a lot of doctoring takes place, with complexity thrown out of window to make the text easier to understand. But it is very difficult to program for increasing complexity when you're not writing the material yourself.
The end result is a lot of wasted time -- lots of material is presented to the learner, and little retained. The goal of exposing the student to native materials slows down the learning process, and in effect means that students may leave the course less well equipped to deal with native materials than they otherwise would be.
But wait... there's more!
There is a secondary consequence of this.
The reliance on themed material favours the well educated, because they are more likely to have knowledge of the topic under discussion, and in many language pairs, the terminology will be very similar. For example, "biometrics" in French is "la biométrie". This particular example came up in my last French course, and my classmates were confused by it, but as I already know a few things about biometrics, there was nothing in the articles themselves that challenged me.
The well educated and well read get better marks, and we justify this with by accepting that people who have done well in the past are "good students" and that those who haven't aren't. This is only marginally better than the old fallacy of equating education and intelligence.
But in reality, whether the historically "good" students are genuinely better than those without a good history of attainment, that's a side issue. The material presented favours those who really have the least need for favours. In my book that's a bad thing.
It's particularly worrying to me that the course I was studying is with an institute of higher education that takes pride in the fact that it makes learning available to everyone, regardless of educational background.