This is the first in a series of blog posts exploring what research says about the most effective ways of learning a language: "What does the research say? Part 1"
I recently came across a discussion about the most effective methods for learning a language. You know, the kind of discussion that is bound to happen when more than two language-learning enthusiasts enter the same room. One of the main claims in this specific debate was that the language-learning methods commonly used in language-teaching institutions - a mix of mainly grammar, rote memorization of vocabulary, and translation - must be the most effective, because otherwise they wouldn't be used in these institutions. As my reply became quite long (and rambling), I decided to make it a blog post instead.
TLDR: Judging the effectiveness of a language learning method based on its uptake by universities and language schools is ultimately a logical fallacy (known as appeal to authority), because one assumes that the behavior of renowned institutions is an adequate proxy for judging a teaching method's effectiveness. This, however, is a big assumption, with little data to back it up.
A variety of factors unrelated to the effectiveness of a specific language-learning method could be motivating language-teaching institutions to go with the traditional grammar + vocabulary + translation. Some of the main ones are inertia, misaligned incentive systems, and difficulties to measure effectiveness, and we'll be looking at those here.
Inertia effect: "this is how it's done"
This is basically resistance to change: second language acquisition has always been taught this way, that's how the teachers were taught, and that's what everyone else does.
There are many cases of such inertia leading to aberrations, throughout history and across sectors. Medicine is full of particularly famous (and graphic) examples: in keeping with the Zeitgeist, have a look at plague doctor costumes.
Or take education in general: the Finnish model of education is often presented as being close to perfect, despite diverging from how most other countries educate their children (no standardised grading, short days, no perverse quotas, etc.), but the political will to apply this has failed to manifest as of yet outside of Finland.
The inertia holds even if the new model is well-studied and proven to be superior: in the homelessness sector, the effectiveness of the approach that consists in giving homeless people accommodation before working on things like issues with drugs - Housing First - is beyond questioning, but it still remains a rare occurrence in any given country (apart from Finland maybe).
Basically, once something exists, it takes a lot of effort to change it.
Incentive systems are like names: everyone has one
In other words, "follow the money".
Incentive systems, and especially market forces, are another factor that needs to be taken into account when thinking about our quandary: language-learning is a huge industry, and a lot of powerful and well-established stakeholders (textbook publishers, universities, language schools) have strong incentives in maintaining the status quo.
This does not necessarily stem from malice. A language teacher might think grammar + vocabulary + this specific textbook might indeed be the one true way. The textbook's author might think that they're doing a service to language learners worldwide by promoting their work. The language school might think that their method works sufficiently well (more on that below), so why experiment with new ones? A language-learner who just went through a language school's program is likely to convince themselves that they chose the best option.
Of course, there can be less-than-altruistic factors at play as well. For example, the textbook publishing is an oligopoly known for its predatory practices. Language-teaching institutions could be said to have an incentive to make their study program last longer (so it costs more).
But, however good the stakeholders' intentions, the result is the status quo being maintained.
Measuring is not straightforward
A paradigm shift is especially hard to implement if the effectiveness of the different methods is difficult to measure. Mind you, even those methods that are easily measurable will not dominate a field easily (bloodletting was the most common medical practice for two millennia, even though it was harmful to the vast majority of patients). But the lack of evaluation definitely doesn't help.
Learning a language is a perfect example of that: most people are not that motivated to learn, most of the motivated ones will rarely use their target language, most of those who do will take years to get proficient and will be severely overestimating their proficiency up to that point. Add to that an unholy amount of confounding factors that are difficult to fully take into account.
Imagine method 1. It takes 6 months, it's cheap, takes place online, and 5% of those who start this method (including those who drop out) have a B2 level one year after.
Now imagine method 2. It takes two years (so learners have ample time to expose themselves to their target language outside of classes and they do, through their own volition, because that's what they're enthusiastic about), it's taught in a school that is located in the target-language-speaking country (read: free immersion), it's very expensive (so most everyone who starts it is already more motivated than the average learner and has a strong incentive to not drop out, so as to avoid wasting money). 40% of those who start method 2 are B2-level one year after.
Is method 2 better than method 1? There is no way to tell, because method 2 starts with much more motivated learners and reaps benefits - for free - from its location in the target country and the learners' free time. Maybe simply going to the target country for a year would yield the same result. Is method 2 worse than method 1? Method 1 sure seems much more accessible to the average person, but apart from that it's difficult to say. One thing you can tell for sure is that more research is needed, so that we can compare apples to apples.
Again, this blog post is not interested in claiming that the standard language-learning methods are wrong (we do not push any specific method at OPLingo), it simply aims to encourage taking a more evidence-based view of this space.
Is it possible that the most common language-learning methods are indeed the best? Yes, it is possible. Do they work? They most certainly do, at least to some extent. But has their superiority been examined in a critical, scientific way? Well, not really. And let me tell you that the research can be quite surprising (take this with a handful of salt before I have the chance to write more about it, but, as a foretaste, there is some evidence that learning grammar can be counterproductive to learning a language, for example).
In the blog posts in this series to come, we'll look at the existing research on the most effective second language acquisition methods. They might have their own biases or research design flaws, which we'll try to identify. In any case, now is a good time to be looking at this research, as the pandemic has given a second wind to new developments in education, at the same time as an opportunity for change.