Cainntear wrote:There's a lot of doubt about the effectiveness of minimal pair practice. I've seen research showing that leaners got better at distinguishing between minimal pairs in the task, with no improvement in distinguishing them outside of minimal pair tasks.
I'd love to see what the study design looked like. Do you still have a reference?
My guess is that if minimal pairs help at all, they work as part of a much larger, longer-term effort. My experience is that making phonemic changes can be difficult and may require a sustained effort. For example, I used to tap my French R ([ɾ] instead of [ʁ]). I knew it was wrong, I could hear the difference in isolation, and I could make the correct sound if I tried. But as soon as I started speaking rapidly, I immediately reverted to a tap.
I fixed this because my tutor kept getting on my case about it. But the process of fixing it took something like 3 months. I practiced it daily in the shower, I sang along with songs that lovely, sustained [ʁ] sounds, I corrected myself when speaking, and I paid a lot of attention. And after several months, it did get better, and [ʁ] now feels entirely natural.
Now, the French R doesn't have a minimal pair. But for sounds which do have a minimal pair, I would expect that any attempt to "fix" a learner's speaking and listening could easily take months of work using a variety of techniques. Explaining the minimal pair and learning to produce and distinguish it is just a first step.
Still, I find it useful to remember that there are some very embarrassing minimal pairs in most languages. For example:
- The English "sheet" and "shit", which occasionally causes ESL students to say things like, "I need to put a shit on the bed."
- The French cou /ku/ "neck", cul /ky/ "ass" and queue /kø/ "tail, dick".
I find that most people don't mix up these minimal pairs more than once or twice. I admit these minimal pairs are not suitable for the typical classroom, but public embarrassment is notoriously memorable.