does passive listening actually help you learn a language?

what happens when your attention is somewhere else

in 2005 juan toro, scott sinnett and salvador soto-faraco played people a stream of invented speech in which the only clue to where one word ended was how often certain syllables followed each other. half the listeners simply listened. the other half did a second task at the same time: tracking bursts of noise, watching a stream of pictures, or catching changes of pitch in the speech itself. afterwards the people who had only listened could pick out the words. the people with a second task did dramatically worse, and the authors concluded that when attention is used up, this kind of word-finding is seriously compromised.

later work complicates that. in 2019 laura batterink and ken paller recorded brain activity while one group listened with full attention and another did a demanding visual task. in both groups the brain tracked the hidden words as the stream played, and both groups rated the words as equally familiar afterwards. a memory test based on reaction times, though, showed that diverting attention had cost something. their reading is that the ear can group sounds into words outside the focus of attention, while storing those words in memory may need some of it. the same paper describes earlier results on attention as conflicting and the question as unresolved.

both studies used invented languages in a lab, and none of the words meant anything. they still point the same way: a busy mind does not stop the ear working, but it takes something away from what the ear keeps.

two things called passive

the word passive is doing two jobs, and it helps to pull them apart. in the toro study, the condition the researchers called passive listening was the one that worked: people sitting with the speech and nothing else to do. that is not what most people mean by the word. they mean a podcast running while they answer email, or a radio station in another language filling the kitchen while they talk to someone in their own.

those situations are a long way apart. one gives the sound most of your attention without asking you to work at it. the other gives it whatever is left over from a task that wants all of it. the research above is kind to the first and much less kind to the second. a fair way to judge a listening habit, then, is to ask what else your mind is doing while it runs. washing up, walking a familiar route or folding laundry leave a lot of attention free, while reading, writing and conversation leave very little.

sound with nothing to hold

attention is one condition. the other is whether the sound can mean anything to you.

hilde van zeeland and norbert schmitt tested how much of a spoken passage listeners need to know in order to follow it. most of their second-language listeners understood adequately when they knew about nine words in ten, though their scores varied widely at that point, and they understood more consistently at ninety-five in a hundred. the authors estimate that ninety-five in a hundred means knowing between two and three thousand word families. someone in the first months of a language knows a small fraction of that. a radio station in that language, however long it plays, is mostly sound to them, because almost nothing in it gives them a foothold.

the linguist richard schmidt argued in 1990 that learning without any awareness does not happen, and that input only becomes useful to a learner once it is noticed. later researchers have disputed that strong form of the claim, and the lab results above suggest the ear does some of its work unnoticed. what is hard to dispute is that you can only notice what you can partly recognise, and a language with no meaning attached gives you very little to recognise.

a few minutes, watched

the most encouraging result comes from a study where both conditions were met. marianne gullberg and her colleagues gave dutch adults between seven and fourteen minutes of continuous mandarin chinese on video, natural speech with the gestures that go with it, and no instruction or training. they then tested what the viewers had picked up about the language's sounds and words, and looked at the part played by how often a word came round and by the gestures around it. their conclusion was that the adult learning mechanism is more powerful than is normally assumed when it meets small amounts of continuous speech that can be seen as well as heard.

what that study offered was not effort. there was no teaching and no drilling. it was speech presented as the thing to watch, with a speaker's hands carrying some of what the words meant.

making background listening count

if you already listen in the background, you do not need to stop, only to change two things about it. the first is the job running alongside it. move the listening to the parts of the day when your hands are busy and your thoughts are not: the washing up, the walk, the queue. keep it away from the hours when you are reading or talking, when the other task will take everything.

the second is the meaning. choose listening where you can reach what is being said: a podcast made slow and simple for learners, a film you already know well, audio that gives you the translation. then check it the way you would check anything else. after ten minutes, can you say what one line meant? if you can, the listening is reaching you. if nothing comes back, turn down the other task or the difficulty, and try again.

a line, then its meaning

driftspeak is built for that middle kind of listening. it is subtitles for your ears: lines from albums and films, a line at a time, in their language and then in yours, so each one arrives with its meaning beside it. it is paced for half your attention, and it claims no more than that.

if you want to test the middle kind of listening, attentive but not strained, listen to a sample of words i kept while you wash up or walk, and see whether "salí de casa de madrugá, sola" ("i left the house at dawn, alone") is still with you when the clip ends.