Should be an interesting discussion, Jonathan. I look forward to following it.
I am quite familiar with Duolinguo, Living Languages, Pimsleur, and bits and pieces of others in learning modern Spanish and Greek. I think they all can be effective if used intelligently, and I think that some of the strictures such as not speaking any English at all during a learning session can be overdone. It depends on the specific application, and sometimes on the leader/instructor. TPRS, for example, uses English sparingly and to good effect, while other methods which will not allow any English seem quite strained and unnatural at times.
I have found Randall Buth’s take on this topic to be very helpful - Surprise! Surprise! I have also found two of the references he recommended to be very helpful:
Randall Buth wrote: One of the most prolific and astute writers on Reading is Frank Smith. For example,
Smith, Frank (2004). Understanding Reading: A Psycholinguistic Analysis of Reading and Learning to Read. 6th ed.
A nice article on the phonological loop and reading that deals with the experimental data is:
"Phonology in Second Language Reading: Not an Optional Extra"
Author(s): Catherine Walter
Source: TESOL Quarterly, Vol. 42, No. 3, Psycholinguistics for TESOL (Sep., 2008), pp. 455-474 (Available thru JSTOR)
If you think about – and explore – real facility with a language itself, as opposed to competence in the associated meta-language, you begin to gain some appreciation for why Randall talks about such concepts as the 2 second (or less) ‘phonological loop’. Catherine Walter explains the phenomenon itself, but for me Frank Smith really brings it alive when he talks about how people actually learn a language – which is how they learn everything else.
This is not just a lot of techy banter, but
it really does inform you concerning how to create the right learning environment – and I choose that wording advisedly. Real language learning – that is, gaining a fluency and facility of use with a language – HAPPENS in the right setting. That setting must include a
context of interest – leaning happens when there is a matter of interest before the ‘learner’. For languages, when we are speaking about fluency and facility of use, real learning also happens
WHEN THE ENGAGEMENT IS DIRECT, as opposed to analytical. That is, if someone says to me “
Ὄνομα σοι τί ἐστιν;”, real language response is a ‘thoughtless’ “
Ὄνομα μοι Θωμᾶς” – without intervening time, without reflection, without trying to remember which case
Θωμᾶς should be in, without reference to another language parallel – without, even, recalling a memorized language piece. Of course, I don’t really mean “thoughtless”, and sometimes there is “reflection” of a different nature, but only to say that a real language response is immediate, and without intervening mental processes – to use a different metaphor.
It all becomes much clearer when one chats with a five year old, who knows very little of the meta-language, but who will respond
without any hesitation to a very broad range of language stimulus. He or she is an ‘expert’ language learner! I know L2 acquisition is different, but much of the underlying basics still apply. It is a direct, immediate, and learner-initiated activity. It cares not one whit for the complexities of grammatical logic, nor does it care about other-language parallels.
This informs one how one must LEARN the language if fluency and facility of use are the goal. I see a picture – I hear a word. I follow a simple narrative while I watch a depiction of the narrative in pictures, cartoons, video, play acting, etc. I make a connection without intervening analysis. That connection is now mine, and mine in an ongoing fashion if I go on to EXPRESS this language in a meaningful way. This is very different than the way meta-language is learned, and very different than the way translation proper is learned. This is what you are after if fluency is what you want.This is also a rather humbling experience for those who are proficient in the meta-language but who are less than babes in real language fluency.
The shortcoming I see with what you’ve posted from Duolingua is the strong bias towards “translation” as opposed to a direct and immediate connection with the thing itself. A picture of a boy combined with target language narrative
(“ὁ παίς”) allows me to make direct connection between
ὁ παίς and the visual/audio representation rather than connecting
ὁ παίς with the English parrallel “boy”. This is where I think use of English definitely
IS a distraction to the actual learning process. Having made the direct connection between the sound and a visual depiction,
ὁ παίς is now a mental entity (one struggles to find language!) associated with the
audio "ὁ παίς", and later on with the
written "ὁ παίς". It is there! It is mine! Given the right stimulus, it is immediately available without need for 'reflection' or 'analysis'.
Of course, once one begins to attain some fluency, the context of the language itself more and more instructs one about new entities, as opposed to having to depend on arranged representations with pictures etc.