Does memorising songs tell us much about language learning?

For over 25 years I have sung in choirs, mostly barbershop choruses and quartets. This has made me spend many, many hours memorising melodies and lyrics. I'm in that mode at the moment as our community choir in East Dulwich, Note-Orious, prepares for our 'disco extravaganza'. Yay! 

So what does this process of song learning have to tell us about memory and, perhaps by extension, language learning?

1. Repetition + Little and Often is best. This is in line with spacing theory from cognitive science. Hermann Ebbinghaus famously demonstrated how quickly we forget items from our short-term memory. The antidote to forgetting is to make yourself retrieve from memory, at spaced intervals, the information you first tried to remember. No one knows quite what the spacing should be, but one hypothesis is that after the first encounter with information, the intervals between retrieval should be shorter, then gradually increase over time. They call this 'expanding spacing'. I wrote about it here.

So with song learning I try to keep my learning sessions not too long and spaced out. if you spend too long on a session you lose concentration and interest. Similarly, in language learning, it is very likely that shorter, more frequent lessons will lead to better retention. Alas, many schools fail to offer this diet. With respect to vocab learning, the same principle applies. Keep sessions short, spaced and as frequent as possible.

2. Consolidation. In cognitive science the concept of consolidation of memory is well-known. Even when we sleep, the brain is continuing to work and expend energy, as networks and memory traces are strengthened. In song learning you can get frustrated when words don't seem to stick short-term, but later you can find that you've actually remembered them better than you thought. In language learning, consolidation of memory is happening all the time. It's even been shown that memorising before sleep is a good idea, since the following day, memories have been consolidated and strengthened. Research on this goes back to Jenkins and Dallenbach (1924)

3. Implicit learning. This is where I sense the strongest association between song learning and language acquisition. Through repetition and almost 'not trying too hard', tunes and words eventually stick. You can't easily force the process too quickly. Sheer repetition and unconscious acquisition (with consolidation) lead to eventual mastery where you don't even have to think about what you are singing. Just as an L1 speaker rarely has to stop and monitor what they are saying, once you have memorised a song it is just there and stays there for quite a long time. I can come back to a song, say, three months later, and it is still pretty much in memory.

In language learning we know that learning is largely of this unconscious, implicit nature. Explicit focus on form and setting vocabulary to memory may help the process along, but these are not the main drivers of growing proficiency. Classroom language learners will take time to become proficient. The amount of time depends on aptitude (including working memory) and motivation.

4. Noticing

When learning parts for a song I sometimes have those moments where I notice something in particular and pay extra attention, perhaps repeating them more often. It might be a tricky combination of words, a flattened note, an awkward change of pitch. Giving that extra bit of attention seems to come naturally when I want something to be more 'salient'. By doing this I think it will help me remember. I am sure it does.

In language learning noticing is sometimes claimed to be a prerequisite for acquisition. (See Schmidt's Noticing Hypothesis.). This is actually a contentious area, partly because it's hard to know precisely what 'noticing' means. Sometimes a piece of language will come sharply into focus as we 'notice' it. This probably helps us remember and reuse it later. Other times we seem to acquire new words and constructions without ever having been aware of noticing them. Perhaps we did notice, but at a subliminal level. Geoff Jordan wrote an interesting blog about this, examining a range of evidence and how Schmidt developed his hypothesis over time.

As language teachers we often force students to notice and use vocabulary and grammatical constructions. Sometimes we have to make an extra effort for students who may not notice things which seem redundant, like, for English L1 users, adjective agreement. Most teachers believe that getting students to notice items is useful and necessary, as well as getting students to rehearse specific language items in working memory, for example through delayed dictation or delayed choral repetition.

5. TAP. TAP stands for Transfer-Appropriate Processing. This is the well-established finding that we recall items better when we recreate the conditions in which we first learned them. This is why it can make sense to test language using the same type of tasks you used when teaching it. So, if you did gap-fills in the lesson, use gap-fills in the test, unless you want to make it deliberately harder.

In song learning, when you learn the words together with the melody, it is easier to recall the words later together with the melody. If you try to say the words aloud without the tune, you'll probably find it harder. In language lessons, the same phenomenon occurs, for example, if you teach the letters of the alphabet with a tune, students will be able to sing back the alphabet but will find it harder to just say the letters. And spelling out words using the letters in a  different order will be much harder. At a more general level, if you want students to become good listeners, do lots of listening. If you want them to be good at conversation, do lots of conversation. I would call these general examples transfer, perhaps, rather than TAP which is more specific.

To sum up, I wouldn't want to push the analogy between song learning and language learning too far. It works when we are talking about memorising language, for instance for a an oral test or a learned presentation. general language proficiency is developed, largely implicitly, through input, repetition and interaction. As researchers like to say, the linguistic system is built by hearing and reading comprehensible input, including new language just abve the learner's level (Stephen Krashen's famous i +1). But when it comes to helping students memorise language (not a useless practice at all), the song learning comparison has some things to tell us.

References

Ebbinghaus, H. (1885). Über das Gedächtnis. Untersuchungen zur experimentellen Psychologie (On Memory: A Contribution to Experimental Psychology). Leipzig: Duncker & Humblot

1. Jenkins J, & 
2. Dallenbach K. 

(1924). Obliviscence during sleep and waking. Journal of Psychology, 35, p. 605–612.

The term obliviscence is now old-fashioend but refers to the process of forgetting or the gradual loss of memory over time. It is derived from the Latin word oblivisci, meaning "to forget." In psychology and neuroscience, obliviscence is studied as part of memory processes, where information that is not reinforced or retrieved may fade or become inaccessible. It is a natural and common aspect of human memory, though the rate and extent of forgetting can vary depending on factors like the importance of the information, how it was learned, and whether it is periodically reviewed or used (Mistral AI).