It all started because I have been teaching Greek to some dedicated friends. Now if you know anything about classical languages, you know that many people every year take these languages, such as Greek, Hebrew, and Latin, in graduate school as prerequisites toward further study. And you may also know that the dropout rate post-graduation for the enjoyment and retention of these languages is not good. Now I and my friends are putting a lot of effort into learning Greek. I don't want them to be a statistic--and, frankly, I don't want to become one either. So I ask the question, "What does it take to escape the attrition trap and break through into enjoying and so sustaining and even growing, in a language?"
I work a job, so I don't have a lot of time for deep reading and research about this. But here are a few things I've learned.
Immanuel Kant was a Linguist
According to Immanuel Kant, human beings construct the world along two planes: extension and change, in other words, along space and time.
- The noun system captures space
- The tense system captures time
Both employ the same method to do so.
Beginning with a root, prefixes and suffixes are added or removed to fix that root within a matrix that assigns it jobs--being a direct object or a subject, for example.
We memorize a lexical form (lemma) of a word, but understand that the lemma really exists as a root that can manifest anywhere along a matrix.
Understood in this way, there is a deep repetition, a strategic recurrence, between verbal and noun systems. Furthermore, taken together they answer Kant's qualifications for building the kind of world that human beings experience (an appropriately phenomenological world.)
With this in mind, one can begin to interrogate a language and ask why this matrix is chosen and not another one? Why five or seven cases and not thirteen or twenty-one?
Open- and Closed-Class Words
Linguists parse the words of a language into two categories: open-class and closed-class. Open-class words are the usual nouns, adjectives, adverbs, and verbs that we usually associate with language. Open-class words are also called content words or lexical words. These are the words that carry the meaning in sentences, and it is interesting that new words that come into a language are always open-class words (thus the "open.") Closed-class words, also called function or grammatical words, are things like determiners, qualifiers, prepositions, conjunctions, and intensifiers. They serve a variety of functions, as their names demonstrate. They do not, themselves, carry meaning in the way that content words do. Instead, they serve to grammatically connect the open-class words. There are far fewer closed-class words in a language than open-classed words by several standard deviations. And, here's another kicker, unlike open-class words, to which new terms may be coined or invented or taken whole from another language, closed-class words are stubbornly fixed. Languages have all the function words that they need, and it is near impossible to delete or add to their number, even when it would be useful to do so. Closed-class function words, then, are the inner skeleton upon which the open-class content of the language is attached. In short, if you are going to learn a new language, get to know function words well, and become adept at watching for them when reading. "Once the framework of grammar has been transferred to long-term memory," says author Tim Ferris, "acquiring vocabulary is a simple process of proper spaced repetition." I've read that an ESL teacher, for example, should aim at the perception of the structure of a text before the individual words. So by zeroing in on function words, one will learn more about a target language than would be done through memorizing hundreds of content words.
". . . the task facing the child is not to learn how language works, starting from scratch. Instead, since children are born with an implicit knowledge of languages in general, they have to figure out how the particular language (or languages) they hear functions. For example, all languages have something like prepositions, words that show relationships among things (The book is on the table). In languages like English, these words that show position come in front of the noun, so they are called prepositions. In other languages, these words follow the noun, so in those languages, a child would encounter sentences with this pattern: The book is the table on.In such languages,these words are called postpositions because they come after (post), not before (pre)." 
An Interesting Note from a Translator
"A translation is not just turning one language into another. It’s also about opening up a foreign mindset . . . to hear the text and experience it absolutely as intensely as I can, allowing myself to fall into its way of thinking about things. A good translator has to be an interested sponge when it comes to the idiom and cultural setting of the language he or she is translating from [--] fascinated by the picayune details of language. Every complex translation would be somewhat different if we had done it a month before, or a month later, or even an hour."
Benny the Irish Polyglot Says to Ditch the English
One of the biggest lessons Benny learned in his transformation to langauge polyglot was to give up his English as quickly as possible. Be stupid, he says. Make mistakes. But leave your English behind and deal with the frustrating, bewildering, but language-acquisition-fast method of leaving your comfort language behind.
Now, where dead languages are concerned, you can't just avoid your English-speaking expat friends and hang out with the locals. What you can do, however, is to up your exposure and go cold-turkey on sections of text. For example, one could go native on the Gospel of Mark or on a chapter of the Gallic Wars or something.
Tangentially, get away from code-thinking as soon as possible. A new language is not just a code for your old one. Stop occasionally and forget your native tongue while holding the new language in stasis like wine held in the tongue. Enjoy and parse out the sensation before swallowing.
A Large Bit on the Importance of Reading
"Studies of vocabulary development through reading give further support to the claim that most vocabulary is acquired. Anderson and Nagy carried out a series of studies on how children acquire words during reading. They found, for example, that there is about a one in twenty chance that a student will acquire a new word from seeing it in context. . . . If students see a word more often, they are more likely to acquire the word.Anderson and Nagy report that the average fifth grader reads for about twenty-five minutes a day. They comment "This number is certainly lower than would be desired, but it translates into about a million words of text covered in a year." If even 2 percent of the words were unfamiliar, students would encounter twenty thousand new words in a year. If they acquired one out of every twenty, they would acquire at least one thousand words a year.
"These authors go on to say, "An avid reader might spend an hour or two a day reading, and thus cover four or more times as much text. The rate of learning from context for self-selected text is likely to be closer to one unfamiliar word in ten than one in twenty. For children who do a fair amount of independent reading, then, natural learning could easily lead to the acquisition of five to ten thousand words a year, and thus account for the bulk of their annual vocabulary growth."
. . . .
"A study carried out with adult speakers of English and students learning ESL showed that both groups were able to define many new vocabulary words just from reading a novel.
. . . . .
"[Another researcher] found that picking up words from reading is ten times faster than learning words through intensive vocabulary instruction. However, they also suggest that some vocabulary study can be useful. They encourage teachers to develop a sense of what they call "word consciousness." "We believe that the goal of instruction should be to develop what one lexiphile has termed word consciousness."
. . . .
"One of the benefits of acquiring vocabulary through reading is that students develop a more complete understanding than the superficial knowledge gained by memorizing a definition. . . . When students see and hear a word in different contexts, they build a subconscious understanding of that word. Extensive reading is the best way for students to build a rich vocabulary." 
Note also the following quotation.
"It is perfectly possible to expand your vocabulary in a sustainable manner that is relevant to your current level, needs, and usage. Remember the “output” versus “input” distinction I’ve made earlier? Focusing on input (i.e. interesting resources to consult to immerse yourself in the language) rather than output (i.e. the number of words to memorize every day) is a good start. Rather than becoming obsessed with memorizing new words, simply accept the fact that you will end up forgetting a lot of the stuff you come across. But by being exposed to material that is at a suitable level to you and by trusting your common sense and learning as much as you can from context, you will inevitably end up assimilating sentence structures, collocations, and words in a more natural fashion through exposure. Sure, at this stage, if you come across a word several times and you still can’t get your head around it, it doesn’t hurt to write it down in a notebook or even save it in a flashcard app. But by changing the way you approach language learning, and by taking the focus away from memorizing a set amount of words every day/week/month to actually getting exposed to the language in a more holistic manner, you will find that slowly but surely, your skills will progress and your fluency in the language will follow a path in harmony with the vocabulary and sentence structures you are exposed to." (Lingholic)
Lend Language Your Ears
Human beings use air, their vocal chords, and the contraction and expansion of their oral and nasal cavities to express or suppress sounds. The Phonecians were the first to begin writing down sounds phonetically rather than resorting to picture language. Therefore writing for every other language that does this, which are most languages, is a kind of shorthand for the contraction or relaxation of organic sound production. Writing isn't where meaning lives. Writing tells you how to move your body to make the kind of sounds that produce meaning. Generally speaking: language is an oral thing. When you look at a page of English or French or Latin or German, you are seeing instructions for producing the sort of sounds that community of speakers agreed on. What this means is that when learning a new language, you need to keep your mouth and ears involved. Never read silently, and listen to as much as possible.
Set Short-Term Goals
One thing that Tim Ferris and Benny the Irish Polyglot talk about is not dying at the hands of language perfection. Language courses tend to teach against an ideal of perfection, and this, they say, kills motivation. Language is the way that minds connect with minds. It is the tool for making connections between people and cultures. So learn what you have to have to get to the point where you can start connecting and let the rest take care of itself. That's what they would say.
This is not as applicable where classical languages are concerned, but it is still helpful. If the point is to make connection, then read for meaning first before you read for parsed perfection.
Ferris and Benny talk about setting smaller short-term goals: to carry on a two-minute conversation, to order coffee, to read a weather report, to read a chapter of Plato without reference to a lexicon.
And finally, I'm thinking about the idea of micro-grammars within languages right now. It is one thing to learn how to say hello or to construct a simple sentence, but the conversation dies quickly thereafter because of a lack of micro-grammar. A micro-grammar is a word I use to talk about spheres of language: the weather, sports, the office, family life, religion, what's new in politics. Native speakers move from micro-grammar to micro-grammar as easily as moving from room to room, but even in native language there are times when one must acquire new words for a new environment (the micro-grammar of your city or neighborhood.) I'm not sure how yet to best incorporate this into language acquisition, but when I think of various children's books I can see that they recognize the existence of such grammars and take steps to teach words accordingly.
The Two-Face Technique
From the beginning, I have thought about language acquisition using the metaphor of climbing Mount Everest. There are a number of reasons for this. The effort and focus that climbers display even years before their attempt. The social, material, and physical expense and exertion, if not pain, required to successfully summit the mountain. The way that climbing Everest has become a well-understood and apportioned process of having such-and-such gear and moving up the mountain through established camps. The fact that hundreds if not thousands of people make an attempt every year (you aren't any different from them.)
Now one element of the metaphor that has become very useful is the difference between Everest's south and north faces. The south face is the (qualified) easier of the two. The south face was the way Hillary and Norgay made the summit in 1953. The north face, however, is a beast.
For my purposes, each face represents a technique of language acquisition. The north face represents what we usually think of when we think of learning a language: wrote memorization and paradigms. The raw violence of making our minds sink new synapses into new patterns unattached to any other familiar information. The south face represents the way native speakers learned their language. South face techniques are reading aloud and reading a lot. They are fun and easy, and in my experience they charge a session with energy and life. North face exercises feel like work. South face s just fun.
The trouble, of course, is that south face takes a good long while and a lot of exposure--far more than we'd achieve even through a course of immersion. (Non-native speakers tend to achieve a homeostasis of "good enough," which is why I say even immersion is not sufficient.) Few people have the time or patience for such an approach. On the other hand, north face is not so great either. I already mentioned the abysmal rates of attrition by graduates who have taken even years of a language in formal instruction. So what to do?
My hypothesis is that good language work needs both faces delivered in appropriate amounts. Overall, south face activities are best, but north face activities should be used to speed up the dial. You swallow a paradigm or construction quickly, via north face, and that reinforces and makes your south face work more capable. The resulting success pumps endorphines into the whole and keeps the arrow of acquisition moving forward.
Linearization or Discourse Up and DownDr. Steven Runge talks about something called the linerization problem. "Linearization describes the fact that we can only produce one word at a time, one sentence at a time," he says. That means that "the reader/hearer can only take in one word at a time, one sentence at a time." So the hearer or reader has to construct the architecture of meaning that's coming at her, and she gets one shot at it. How does she do it?
Runge says that she does it through two methods. First, she uses deictic markers or textual markers that help her structure the stream of information she is hearing or reading. These tell her what is more important or less, whether the subject has changed or is going in a new direction. Citing the work of linguist Walter Kintsch, Runge calls this method construction.
As it turns out, there is a lot of debate about construction. After all, if meaning is really found in the big structures of language, then why is it that so few of those handy markers are present at that level. And why are so many found at the level of sentences? Runge says that it is a matter of debate whether meaning is top-down or bottom up, but he thinks it is both.
What he's really fascinated by is how textual markers on the sentence level go on to build those big, discourse-level structures our reader uses for understanding. "Call me silly," he says, "but it would seem that if one has properly understood how a device operates in simplex context at the lower-levels, then one will be in a much better position to adequately describe its much more complex interaction with other features at the higher-levels of discourse processing, i.e. the integration stage." Discourse Analysis needs its more humble, lower-level cousin, discourse grammar. "Lower-level structures are they keystone to understanding higher-level structures."
The second method the hearer or reader uses to extract meaning is more contextual. Runge calls it integration.
The newly forming mental representation of a text doesn’t exist in an isolated silo of our brain. Instead Kintsch has demonstrated that we integrate the new one into our existing, larger mental representation. This integration is not simply with the earlier portion of what we’ve read or even other books we’ve read, but with the sum of our knowledge about the world and how it operates based on our prior learning and experiences. . . . Differences in background knowledge, goals, and presuppositions all play a role in how we process a text. We don’t just read a text, we also integrate it with what we already know.
Integration is why two people can read the same text and get completely different answers. They may both be doing the construction side at nearly the same level, but their integration is widely different, as are no-doubt their life experiences. "Our own mental representation of the world . . . plays a huge role in how we process new texts or communication."__________
 David E. Freeman and Yvonne S. Freeman, Essential Linguistics (Portsmouth, NH: Heinemann, 2004), 14.
 Dennis Abrams, "The Art of Translation: Something New, Something Old" Publishing Perspectives
https://publishingperspectives.com/2013/04/the-art-of-translation-something-new-something-old/ accessed April 2, 2013.
 Watch his interesting TEDx talk at http://youtu.be/HZqUeWshwMs
 Freeman. Linguistics. 2004.
 A deictic marker is a word or phrase (such as this, that, these, those, now, then) that points to the time, place, or situation in which the speaker is speaking. Deixis is often and best described as “verbal pointing,” that is to say pointing by means of language. Deictic expressions fall into three categories: person (you, us), spatial (here, there), and temporal (now, then) deixis. Deictic expressions are tied to the speaker's location/context (deictic center). The most basic distinction is between near/proximate and far/distal.