Or, why do we, as classical educators, care so little about what modern foreign language research has to say?
The classical languages are generally taught very differently than modern foreign languages. Most people in classical education see this as a distinct plus: in addition to learning Latin, we also reap a variety of ancillary benefits inherent to deductive language study. Win-win, right?
But in the midst of all this brain-training grammar-translation study, do we actually learn Latin?
Reading about Latin education in the 19th century - the halcyon days of proper grammar-translation according to Cheryl Lowe of Memoria Press - one finds many noted classicists discussing how, after years of study, most students still couldn't just open a book and read a Latin sentence in Latin word order. They had to laboriously translate, word by word, to have any clue as to the meaning; even then, with each individual word correctly parsed and translated, they still often got it completely wrong. There was quite a lot of discussion about how to change Latin (and Greek) instruction to get better results, so that the majority of students attained some degree of actual reading ability in exchange for all their hard work.
I lurk on a list where the focus is applying modern language acquisition research and techniques to teaching Latin. List members cite firsthand experience with the low level of Latin skill achieved by the majority of their students under traditional g/t methods. In fact, they don't believe that learning Latin via a pure g/t approach trains the brain so much as it requires a brain with high cognitive ability in the first place. In other words, they say, learning Latin deductively doesn't make you smart - you had to *be* smart to successfully learn Latin via a grammar-translation approach in the first place. In their experience, only 5-10% of students - usually those gifted in math - can actually succeed in learning Latin by so-called traditional methods (they argue that teaching Latin almost entirely in the vernacular is actually an aberration compared to how it was taught for most of its history).
While I'm not sure if I fully agree with their assertion that grammar-translation study doesn't confer any brain-training benefits (though they seem to have cognitive science on their side with regard to it conferring - or more accurately NOT conferring - transferable skills; I'm not sure precisely where I stand on all that. Clearly, yes, critical thinking in a field requires significant domain knowledge; however, most people agree that "how to learn" skills, like note-taking, once learned, are applicable across many fields of knowledge. Which side of the line do the Latin claims fall? I'm not yet sure.), I've never seen anyone talk about a successful g/t program (one that conferred actual reading ability) that didn't involve a lot of Latin exposure and working in Latin (as opposed to the vernacular).
This observation dovetails nicely with modern language acquisition research (much of which is based on Krashen's five hypotheses). Some concepts pertinent to teaching Latin:
*The Input Hypothesis - A language is acquired through sufficient input that is one step beyond a learner's current level (input at i+1 for a learner at i); often referred to as comprehensible input, it should be self-selected and ideally be on a topic the learner finds intrinsically interesting. Thus they develop proficiency in the L2 almost incidentally while they are consciously learning about topics of interest. Krashen believes that generating output in the L2 has no effect on acquiring the language - input is all that matters - but other researchers disagree, variously considering output anywhere from helpful to necessary to acquire a language.
*The Acquisition-Learning Hypothesis - There is a difference between acquiring a language and learning about the language, with formal grammar study belonging to the latter and having no effect on the former ("language appreciation"), unless the grammar instruction is in the target language and the student is genuinely interested in learning the grammar. Then the grammar instruction functions as comprehensible input. (Traditional Jesuit Latin instruction would fall under this category; they utilized the direct method - striving to avoid, as far as possible, the use of the vernacular as the means by which Latin is learned - thus giving students large amounts of exposure to Latin.)
In reading threads (on the Well-Trained Mind boards) about learning modern languages, I find a lot of support for Krashen's ideas. There is a lot of emphasis about getting sufficient exposure to the L2 - at least an hour a day, if you want to achieve fluency. People talk approvingly about European schools that teach multiple languages by teaching other subjects (math, history, etc.) in the various target languages. Yet, when it comes to the classical languages, we're content to see maybe 15 lines of Latin a lesson?
We seem to forget that the classical languages are, in the end, still languages. Whether acquiring Latin confers greater benefits than acquiring another language, you still have to manage to acquire it. Research that is valid and applicable to learning modern languages is equally applicable to learning classical languages.