Learning a new language makes it harder to remember previously learned foreign languages

[Saim]

I think that being highly multilingual and competent it at it is a kind of meta-skill independent of your knowledge in any specific language. For this reason I still do deliberate learning — translation exercises and continuing to learn
vocabulary even as words get rarer and rarer — even in my stronger languages. People say that at a certain point you can rely on naturalistic exposure and speaking occasionally once you’re sufficiently advanced, but in my experience things start to fall apart the more languages you add if you stick to this paradigm.

[Iversen]

I don't agree, and I should know from previous experience. The one thing that can cause old languages to get rusty is lack of use, but the more you have used those old languages and the better you know them the less upkeep they need. The only role a new language can have in this game would be if it totally monopolizes your study hours.

[Le Baron]

I don't agree, and I should know from previous experience. The one thing that can cause old languages to get rusty is lack of use, but the more you have used those old languages and the better you know them the less upkeep they need. The only role a new language can have in this game would be if it totally monopolizes your study hours.

This topic stuck in my head after rdearman posted it. I think it has more implications generally than the dedicated language-learning fraternity (and sonority) might want to allow.

If, as the article claimed, and which I think has a grain of truth, there is a depreciatory effect on previously learned languages, every attempt at an extra language could mean a situation of 'more languages', but lesser overall benefit or net gain. Let's say someone got to B1 or B2 in Spanish - which is often cited here as a reasonable maintenance level - and then shifted onto German, then experienced the effect described by the article. If the only real 'maintenance' of the Spanish is passive and less focused, since it now isn't main language being actively learned, who is to say how extensive the depreciatory effect will be? Assuming this differs in magnitude between learners.

It could be the cause of being in possession of a succession of langages which are generally being weakened by the combined pressure of each additional language. Of course more learning over time might strengthen them, but genuine maintenance is time consuming and the time is spread increasingly thinner by every addition.

I'd be interested to know if perhaps there is a time element in the sense that e.g. five or more years of operating in an L2 with no interference makes it easier to retain vocabulary and facility in that language, should a new language be pursued. I can imagine it is much worse when the languages are stacked together in a smaller time frame.

[Cainntear]

I don't agree, and I should know from previous experience. The one thing that can cause old languages to get rusty is lack of use, but the more you have used those old languages and the better you know them the less upkeep they need. The only role a new language can have in this game would be if it totally monopolizes your study hours.

I think the thing ignored in the headline (thread title is a word-for-word copy of the linked article) is that the test showed that learning a new language makes it harder to remember words in previously learned foreign languages, not harder to remember the language.

To be fair, the paper itself clickbaits this, as although they're only measuring vocab, they present the results as something about language more generally: they're abstract opens with:

Anecdotal evidence suggests that learning a new foreign language (FL) makes you forget previously learned FLs. To seek empirical evidence for this claim, we tested whether learning words in a previously unknown L3 hampers subsequent retrieval of their L2 translation equivalents.

and ends with

Thus, learning a new language indeed comes at the cost of subsequent retrieval ability in other FLs. Such interference effects set in immediately after learning and do not need time to emerge, even when the other FL has been known for a long time.

The "general conclusions" section opens with this:

In this study, we asked whether learning a new language can make you forget previously learned foreign languages, and whether such detrimental effects set in immediately after learning, or only later in the learning process.

It then goes on to talk only about the experiments using vocab, but without clearly stating that it doesn't at any point measure language as a whole. I see no clear indication in the entire paper that the authors warn not to assume the same is true of other elements of language, so it is understandable that the finding might be taken as indicative of something more fundamental. The authors should know that something proven for one part of language is not proof of an analogous thing for other parts of language. They should also know that there is a historical problem of vocabulary studies being the easiest thing to carry out, and so people are predisposed to do more vocab studies than anything else, so there have been quite a few conclusions about language reached based on vocab stuff that only applies to vocab, and not the rest of language.
They may have mentioned the limitations in passing, but there should be a section titled "limitations" or similar to make any such limitations immediately clear and unmissable.

Here's an alternative hypothesis (NB, I state this grammatically as a fact not because I have any certainty of its truth, but because that's how hypotheses are generally written... which is probably a problem, but anyway):
The opinion that learning a foreign language leads to problems in other foreign languages is prevalent because there is interference between languages in terms of vocabulary, and vocabulary is the most immediately noticeable feature of language; hence people will notice vocabulary interference more easily.

Of course, I really suspect that there's some degree of interference in grammar too, so this is not my own personal hypothesis -- it's just an alternative hypothesis that the study does not investigate... hence why they should make it clear that they're only talking about vocab.

[Iversen]

I can't see that limiting the claim to vocabulary changes much - I still don't accept the claim, based on my own experiences which say that learning more languages makes me remember the older one better, not the opposite. If learning a new language really accelerated your forgetting of the old ones then the reason is that you neglect them and spend all your time on your new darling. And interference is just the other side of mutual support betwen languages: if two languages are close enough to cause interference they are also close enough to give useful hints on the vocabulary level.

[Cainntear]

I can't see that limiting the claim to vocabulary changes much - I still don't accept the claim, based on my own experiences which say that learning more languages makes me remember the older one better, not the opposite. If learning a new language really accelerated your forgetting of the old ones then the reason is that you neglect them and spend all your time on your new darling. And interference is just the other side of mutual support betwen languages: if two languages are close enough to cause interference they are also close enough to give useful hints on the vocabulary level.

Well that leads to another subtlety that is lost in the article because it wasn't properly stressed in the paper:
It's not about hampering memory, but interfering with recall. I've certainly had times where I've had to make an effort to remember a word and it's popped into my head in the "wrong" language.

Now I've looked slightly closer at the paper and I can see even more reasons it's not a great piece of work -- most notably that there isn't any description of what the actual test protocol was. I was looking for that because I was thinking that the task was probably so little based on language skills that the latency in responses was more about task-focus than language skill. Eg I was playing around with Duolingo and switched from Ukrainian to Portuguese, and despite already having good command of several Romance languages, I'd been playing the Ukrainian so much that I was dropping articles. Given that my default strategy with Portuguese was to think about Spanish, the dropping of articles is probably down to association with the expected response to the Duolingo task.

[emk]

Well that leads to another subtlety that is lost in the article because it wasn't properly stressed in the paper:
It's not about hampering memory, but interfering with recall. I've certainly had times where I've had to make an effort to remember a word and it's popped into my head in the "wrong" language.

When I'm speaking French, I often have a surprising amount of difficulty remembering words in English. Particularly the proper names of things. Apparently my brain suppresses English vocabulary in general, and proper names get caught up, too. If I consciously switch my inner thoughts back to English for 30 seconds or so, I can often get the proper names back.

And honestly, my intensive efforts on Spanish are beginning to mess with my less-used French vocabulary. Stuff like octroyer "to grant" gets harder to recall. Anything outside the top 5000 words or so may get temporarily misplaced from my active vocabulary. Probably I will need to go read a few books in French at some point to restablize things like octroyer.

And as the authors admit, this is a really limited paper. They're doing that thing where they test a toy version of the problem under easily-managed laboratory conditions:

In the first experiment, 31 native Dutch speakers, all of whom had English as their second language and no prior knowledge of Spanish, participated. Initially, they completed a vocabulary test to confirm their familiarity with certain English words. They then learned Spanish translations for half of these English words, potentially creating interference with their memory of the English words. Subsequently, the researchers evaluated how quickly and accurately the participants could recall the English words.

This is a very different question from what happens if you already speak two languages regularly, and you devote 500-1,000 hours learning a third. The authors do seem to recognize this, but of course the reporters want to turn this limited research into something much more sweeping. Nearly all language learning research in the popular press turns out to be some tiny, limited experiment involving toy languages and less than 10 hours of each students' time. Which is fine as far as it goes, but generalizing that to the massive project of actually learning a language is fraught.

[iguanamon]

And honestly, my intensive efforts on Spanish are beginning to mess with my less-used French vocabulary. Stuff like octroyer "to grant" gets harder to recall. Anything outside the top 5000 words or so may get temporarily misplaced from my active vocabulary. Probably I will need to go read a few books in French at some point to restablize things like octroyer.

Yes, when learning a new language, especially a similar one- because I am concentrating so much on the one I am learning, I find it can push aside my old languages a bit. I try to read or listen to something in the others every day. The process is akin to the old "plate spinners" on old US variety shows. Got to keep the other language spinning along in your mind.

It took me years to avoid mixing up Spanish and Portuguese after reaching a high level in Portuguese. It is much, much less of an issue now. I also have to consciously remember not to mix up Haitian Creole and Lesser Antilles Creole French which can be hard to do. My HC is more advanced than down-island Kwéyòl.

Sometimes, when I am speaking English, I forget how to say something in English and have to ask for the word. So, you are not alone. There is a price to be paid for being multilingual as an adult. Yes, read some engaging French books; listen to French and speak/write it... especially while you are learning a new language. Use your French.

[emk]

There is a price to be paid for being multilingual as an adult. Yes, read some engaging French books; listen to French and speak/write it... especially while you are learning a new language. Use your French.

French has been my primary language at home for something like 10 years now, at least when talking to my wife. My "kitchen French" is very much part of me, certainly after all that time and constant usage.

My "academic" spoken French? Eh, it comes and it goes. Realistically, I don't need it that much on a day to day basis. And it comes back quickly enough—a few podcasts, several chapters of a book, and a few days in full immersion, and I can spruce it up quite a bit. But I've accepted that college-educated native speakers who spend every day speaking French at work and at home have advantages that I don't need to knock myself out to match. If it comes to that, I know I can take a college-level course in French, or work with someone to debug a computer program in French (and have done both). But trying to keep that level of linguistic performance active all the time would require living significant portions of my life in French. Which is isolating in an English-speaking world.

So sure, maybe I could have easily recalled octroyer two months ago, and it took a certain effort to dig for it now. But I'm not sure I've ever even heard a French speaker use octroyer in casual conversation! It's very much the sort of word I'd expect in Le Monde or maybe in a medieval reenactment. If I need to accept some short-term interference with that sort of vocabulary while working intensively on Spanish, no big deal.

So I guess my feeling is that it's OK to focus on something specific sometimes, and to have faith that even if other skills are affected a bit, they'll be easy enough to recover when needed.

[Cainntear]

And as the authors admit, this is a really limited paper.

Not directly and explicitly enough. My TESOL masters team were so dead set on having a "limitations" section in every dissertation that I'm not sure whether you could even pass without one.