You may need to be helped in a newcomer program for basics, but once you're interacting with native speakers that provides motivation and opportunity. Eventually you parse below the lexicalized sentences and sentence fragments and construct the grammar. Pretty much everybody has strong oral and aural conversational skills in their L1s, at age appropriate levels, by the time they hit 1st grade.
Teachers have to focus on academic language, and they have to focus on that for L1 English speakers as well as LEPs. It's where a lot of kids appear to fall behind in late elementary and in middle school. Some kids get home support for academic language in a variety of areas, some don't. That makes a big difference, since if you can't read and take in information you're going to fall further and further behind. (No, it's not possible to say the requisite number of times in class everything that needs to be learned, even assuming that every student pays attention and is focused each time you say it.)
Every state implements its ESL programs differently. But every year in Texas you are subjected to the TELPAS if you're a LEP. It's a standardized reading test, writing samples are collected and evaluated by teachers, and their listening and speaking skills are also evaluated by teachers. Why every year, when you're not going to be age-level fluent in 1 or 2 or even 3 years? (And if you're 15 when you arrive, you're always at best going to be "near native fluent"?) Because they need to track progress. You're expected to advance one level per year.
As for "confusing language," that's a tough nut to crack because what I consider confusing somebody else may not. What I think is confusing today I may not consider confusing tomorrow. Stress increases confusion, whether it's high-stakes stress or time-limitation stress. Sometimes page layout can be the problem. Or the text is just too advanced or requires too much working memory for a kid otherwise occupied. Or the text assumes prior knowledge that's lacking. Some of the most confusing snippets of English I've ever seen were headlines. And some examples from English syntax where it's claimed there are "clear" grammaticality judgments. (Really?)
Take this as a potentially confusing example: "A typical eighth grade English language learner is subjected to one standardized test per week from early April through mid-June, all of them with multiple parts meted out across multiple days." You must know what "typical" means, what an ELL is, get the coherence relation with "them" right and know that each of the multiple parts is given on a separate day, not that each "multiple part" straddles several days. You also need to know that many of the tests the ELLs get are given to everybody.
It helps to know, also, that you really can't study for one of the ESL tests. They're too cumulative. It's like saying you're going to study for your French IV final in which it's not specific points of grammar and vocabulary that are up for grabs, but anything. So I had to take a kind of Russian test to gauge fluency. Every year it was different. We were given a simple conversation--topic unknown, situation unknown--to translate. We were given a passage to translate into English. One year it was from mid-19th century literature, another year from 20th century lit, a third time from a cultural journal, a fourth year from a political speech decrying the lapse of norms in public speech. There's a point at which saying "you must study everything" is very similar to "there's nothing in particular to study."