Sunday, 5 May 2013

A Strange Form of Speaking; (notes on language storage, retrieval, and processing)

I'm reading a novel called ‘The Seamstress’ by Maria Dueñas (original Spanish title 'El Tiempo entre Costuras').  It's a story about love, war, and espionage set in Spain and Morocco during the Spanish civil war and second world war.  One of the characters, a young British woman named Rosalinda Fox, is described as having ‘a strange form of speaking in which words from different languages leapt about chaotically in an extravagant and sometimes incomprehensible torrent’  (page 193).  I felt an immediate bond with Rosalinda because this is a perfect description of what happens to me when I 'mix up' foreign languages.  


Building the Tower of Babel,
Bedford Master, c.1410-1430
I first experienced this phenomenon more than 30 years ago, around 1979 or 80.  Living in Bavaria, I’d become pretty fluent in German.  However, I still considered my strongest foreign language to be French, at which I had excelled in school.  At the hotel where I worked, nobody else spoke English, so I was often called upon  to translate when there were English or American guests.  I never mixed German and English in the same sentence; I could switch between them with ease.  However, one day, a call went out in the hotel, ‘can anyone speak French?’  Of course I said I could, and hurried down to the reception desk to perform my usual interpreting service.  But then it all went wrong. Although I was able to understand everything the French couple were saying, when I tried to reply, I found to my embarrassment that the words coming out of my mouth were not French, but German!  Sentences began in French but changed, somewhere in the middle, to German. My brain was telling me to say one thing, but something completely different was coming out of my mouth.  What was going on? My colleagues were looking at me as if I must be crazy; why was I speaking German when I had claimed to be able to speak French?  In the end the only way to stop it was to stop talking; I had to write out each sentence in French and then read it.  Apparently, although I could still understand French, I could no longer produce it; at least, not spontaneously.  

Because of this, I developed a theory that foreign languages must be stored in the same part of the brain (a different place from our mother tongue) and that they are layered one on top of another, with the most recently used on top.  When the brain goes into 'foreign language mode', the neural channels that send messages from brain to mouth pass through that area, and whatever language is on the top is the one that's going to come out.  

Language confusion:

So back to 2013:  After spending January and February in Spain, my Spanish, albeit limited, felt comfortable and easily accessible.  

view of Alps on flight from Bilbao, Spain, to 
Munich, Germany, March 2013
Then I went to Germany for a week.  Even though I know much more German than Spanish, it was harder to access.  German sentences kept switching to Spanish halfway through.  I was producing horrendous hybrid constructions like this one (when apologizing for not filling out the ski rental form properly):  Lo siento, ich habe mein gafas nicht mitgebracht.  I’m sorry, I didn’t bring my glasses with me.  (Spanish words in italics):

My language mixing theory, updated for the computer age, now goes as follows:  If we think of the brain as a computer, and languages as computer programs, then the program(s) for the language(s) we use every day are always running.  They reside in RAM and there is no need to reload them every time we need them. However, languages that we don’t use are stored in long term memory, somewhere on the hard drive, and before we can use them, we have to retrieve them and load them into RAM.  It wasn't that I didn't know the German words for ‘I’m sorry’ and ‘glasses.’  It was that I was trying to load a foreign language program into a place where another one was already running.

So how long does the loading process take?  I guess it’s different for everyone, just as a computer program will load faster in a computer with more power.  Age probably counts, as well as how much you've used the 'foreign language' part of your brain. Does it load in any particular order? Interestingly, what happened in my case was that the overall structure of the language loaded first, followed by specific words. As the hybrid sentence above illustrates, I experienced no overlap of grammatical structures, such as word order, between the two languages.  It was only individual words that would 'jump' language boundaries.  This suggests that the 'big picture' (the grammar) loads first, and that the 'details' (the words) come in more gradually, erratically, sometimes requiring great concentration, and sometimes falling easily into place.  In an earlier blog post, Garmisch then and nowI described an occasion on the ski slopes when I saw a sign (for a ski run I used to go down), the sign triggered memories, and the memories activated language.  I could almost feel brain neurons rearranging themselves as complete German sentences appeared in my head, like a blurry picture suddenly coming into focus. 

There's an explanation for this in the literature.  Michael Swan, in his article The influence of the mother tongue on second language vocabulary acquisition and use notes that we remember ‘fixed and semi-fixed expressions which are conventionally associated with recurrent situations and meanings.’ Therefore, specific places and situations trigger the memory of the language that goes along with those places and situations. Even, apparently, after more than 30 years.  

That night, in the hotel bar, 'it' happened again when I asked for ‘ein apfelstrudel, ein Glühwein, und ein wasser tambien’.  (It's the words you don't think about, I told myself).  

In the article mentioned above, Swan examines this phenomenon, which he calls  ‘unintentional code-switching', and how it often affects the words you don't think about, the ones linguists call 'function' words.

'Certain kinds of word may be more closely associated crosslinguistically than others in bilingual storage or processing.  In some second language learners, for instance, function words such as conjunctions are particularly liable to importation from the mother tongue and other languages.'  

Swan gives examples of Finnish students learning English (their third language) who mistakenly used Swedish (their second language) for words like ‘and’, ‘but’, and ‘though.’  Like me, the students knew the correct words, but, as Swan points out, 'knowledge and control are not the same thing.'  

where did that French word come from?
On day 3, just as Spanish and German were sorting themselves out, there was a further layer of complication, due to the word mal. It's a word in German (meaning time) and Spanish (meaning wrong).  When I used the word (correctly) in a German sentence, I immediately recognized it as ‘a Spanish word’, jumped to the conclusion that I’d made another mistake, and kept frantically trying to come up with the correct word  (which, of course, I’d already used).  In the midst of this confusion, what should pop up to join in the fun but fois, the French word for time.  Where on earth did that come from?  It was as if a new program, the 'emergency foreign language override program’ had gone into action, searching the depths of the language storage area for more options.  

A few days later I was back in Spain, and as expected, the German program lingered a couple of days in RAM before going back into long-term storage.  My second day back, Corin started laughing at me in the supermarket; I had just asked the cashier for ein bolsa, bitte.  One bag, please. (German in italics). For the first couple of days back, I had to stop and think about every single Spanish word, resulting in a strange, staccato form of speaking.  

And so I was left to reflect on the mysteries of how we store, retrieve, and process languages.  As a student and teacher of languages and language teaching, I wanted to know the science --  the cutting edge research from the field of neurolinguistics!  But I didn't want to slog through pages of intimidating academic articles.  I wanted to read something that gave funny examples and explained the key points in layman's terms.  In the end, when I couldn't find an article like that,  I decided to write one myself.  

Anecdotal research:

First, I wanted to find other people who had experienced this type of erroneous language processing, and where better to look for corroborating evidence than the blogosphere?  It turns out that there are plenty of examples.  Below I've shared some of the best stories from my fellow bloggers, with links to their respective blogs.  

Life in Ljubljana:  Kristina Reardon, in her blog post Third language learning, relates her experience mixing up Spanish and Slovenian.  She tells funny stories about 'my brain’s inability to keep my Spanish and Slovenian separated in my head.’  She relates putting Spanish words into Slovenian sentences even though she knew perfectly well the correct Slovenian word in each case.  She describes how, when switching back to Spanish after learning Slovenian, there seemed to be 'a weight on my brain' and how she felt that Spanish was 'buried somewhere in my brain.'   

Phrasemix:   Then there’s Aaron, the guy who switches from French to Japanese in mid-sentence.  In his post Mixing up two foreign languages he describes his theory about why this happens:  It’s odd that I would think of the Japanese word instead of the English, but I think that’s a product of how I mentally file my languages.  There’s a separate file for foreign languages that I search in, and when I can’t find what I’m looking for in the French pile, I pick up the closest thing I can find, which ends up being Japanese’.  

Taken by the Wind:  My next example is travel writer Reannon, who writes a hilarious account of mixing up Spanish and German called Do you ever mix up your second and third languages?  She writes:  'My German hung awkwardly between me and the confused Mexican woman as I tried to think of a way to explain why a Germanic language had suddenly taken my Spanish language skills hostage'. She also describes having to resort to 'staccato' Spanish.

Kalinago English:  My final example is Karenne Joy Sylvester, who describes her experience (also with Spanish and German) in Brains = filing cabinets or QuadPro hard drives?   It's a short post, but the comments section has many more examples of others' experiences.  

Academic research:

Next I looked for some 'proper' academic articles about what happens in the brain when we process language.  My first question was where/how languages are stored, and my second was why/how they get mixed up.  

A.  Where are languages stored in the brain?  

I started with an article in the science section of the New York Times called When an adult adds a language it’s one brain, two systems.  Sandra Blakeslee describes a study in which MRI (magnetic resonance imaging) was used to discover where in the brain languages were stored.  The subjects were all bilingual; however, half of them had acquired two languages at the same time in childhood, while the other half had learned their second language between the ages of 11 and 19.  The question at hand was whether there were differences between the two groups regarding where they stored the different languages.  Below is an excerpt, or you can click the link above to read the full article. 

Aspects of language ability are distributed all over the brain, Dr. Hirsch said. But there are some high-level, executive regions that are usually localized in a certain neighborhood on the left side of the brain, but are sometimes found in the same neighborhood on the right side, or on both sides. One is Wernicke's area, a region devoted to understanding the meaning of words and the subject matter of spoken language, or semantics. Another is Broca's area, a region dedicated to the execution of speech as well as some deep grammatical aspects of language. The regions are identified by observing brain function.
None of the 12 bilinguals had two separate Wernicke's areas, Dr. Hirsch said. In an English and Spanish speaker, for instance, Spanish semantics blended with English semantics in the same area. But there were dramatic differences in Broca's areas, Dr. Hirsch said. 
In people who had learned both languages in infancy, there was only one uniform Broca's region for both languages, a dot of tissue containing about 30,000 neurons. Among those who had learned a second language in adolescence, however, Broca's area seemed to be divided into two distinct areas. Only one area was activated for each language. These two areas lay close to each other but were always separate, Dr. Hirsch said, and the second language area was always about the same size as the first language area.
This implies that the brain uses different strategies for learning languages, depending on age, Dr. Hirsch said. A baby learns to talk using all faculties -- hearing, vision, touch and movement -- which may feed into hardwired circuits like Broca's area. Once cells in this region become tuned to one or more languages, they become fixed. If two languages are acquired at this time, they become intermingled.
But people who learn a second language in high school have to acquire new skills for generating the complex speech sounds of the new tongue, which may explain why a second language is harder to learn. Broca's area is already dedicated to the native tongue and so an ancillary Broca's region is created. But Wernicke's area, which handles the simpler semantic aspects of language, can overlap.

This is far from being the whole story, of course.  Jennifer Wagner, Ph.D. student in Languages and Linguistics at the University of South Australia, points to newer research such as this 2010 report suggesting 'nouns and verbs are stored in different parts of the brain, similar to how our mental lexicon is further divided into phonological, semantic and grammatical areas (for example, function words are only stored in Broca's area.)'
Noam Cbomsky
Other studies have shown that there's more going on in Broca's than just speech processing and function word storage.  Two studies using MRI, Tettamanti et. al, 2002, and Musso et. al, 2003, found that Broca's area activated when learning new grammatical rules.  More interestingly, it did not activate when subjects tried to apply non-grammatical (i.e. false) rules they had been taught (for example, 'negatives are formed by putting no before the fourth word in a sentence'.)  This suggests that Broca's somehow knows the difference between possible and impossible grammar rules; it knows how grammar works on a fundamental level.  Maybe Chomsky's 'universal grammar' is in Broca's?  (More here on Chomsky's theory of universal grammar).  

For more fascinating facts about Wernicke's, Broca's, and other areas of the brain that are used in language acquisition, understanding, and production, check out Language in the Brain.

So far, then, the science seems to back up the anecdotal findings.  Broca's area, where speech is processed, creates a separate space for languages learned after the age of first language acquisition.  The 'language mixing' phenomenon seems to occur only when speaking, suggesting that it is processing and production, rather than knowledge, that is affected. Furthermore, Broca's area is where function words are stored, and it's the function words -- the ones you don't think about -- that are most affected.

B.  Why do we mix our foreign languages?

Now back to the original question:  The next two articles examine theories about why this phenomenon occurs.

(1)  In Second language transfer during third language acquisition, Shirin Murphy describes L2 (second language) interference in L3 (third language) 'often without conscious awareness’, and how 'short L2 function words appear unintentionally in an L3 utterance.'

Murphy points out that interference from L1 (first language) is rare, suggesting that the brain recognizes whether we are using our 'base' (native) language or what Murphy calls a 'guest' language.  When in 'guest language' mode, we may experience  'failure to inhibit a previously learned second language adequately,' resulting in two of the 'guest' languages getting mixed up.  Level of activation is another factor.  At any given time, according to Murphy, different languages are at various stages of activation depending on proficiency and recent usage. A highly activated L2 (one that’s been recently used) is more likely to show up uninvited in L3 sentences.

(2)  The last article I read was Faulty language selection in polyglots (no link; not available electronically).  Benny Shannon first explains how what he calls ‘faulty interlingual selection’ is different from 'code-switching', as follows:

Whereas code-switching may be likened to the (planned or spontaneous) joint employment of two (or even more) instruments (e.g. musical instruments) that are at one’s disposal, faulty interlingual selection may be likened to the faulty, unintended employment of one instrument instead of another.’  

Whereas Murphy described languages as being 'base' (native) or 'guest' (foreign), Shannon, divides language status into three categories:  dominant (native), foreign (fluent but not native), and weak (in the process of being acquired).  He presents the hypothesis of 'the last language effect,' where the ‘intruding language’ is more likely to be the last language acquired and/or used.  Shannon suggests that this happens because the weak languages are stored in a ‘push button stack’ in our brains.  

Of special significance in this stack is the terminal position, that occupied by the last language the speaker studied or acquired (typically, this language is also the polyglot’s weakest language).  This terminal position is constituted by slots assigned to lexical items which have not yet been assimilated into one’s general conceptual database.  Whenever a new language is acquired, new values are assigned to these slots.  As long as the language has not been mastered, or as long as another language has not been acquired, the lexical items of this language occupy the slots in question.’  

If this is the case, and if there is more than one weak language (i.e. two or more languages still in the process of being acquired) then there are two or more competing lexical items in each slot.  Attempting spontaneous production of language (i.e. speaking) in this situation must be like a computer struggling to run too many programs at the same time and making that dreadful grinding noise -- or worse, locking up....


Tower of Babel; Pieter Bruegel the elder, c. 1563

I've described my experiences and theories, as well as those of others.  I've read some articles from the field of linguistics.  I have a much better understanding now of the phenomenon of language mixing; what I still don't have, however, is a good name for it. The linguists call it 'crosslinguistic transference', 'inadvertent code-switching, or 'faulty language selection.'  The language teachers called it 'interference' or 'transfer.' The bloggers call it 'mixing up languages.'  I prefer the quote I used at the beginning:  'A strange form of speaking where words from different languages leapt about chaotically in an extravagant and sometimes incomprehensible torrent.'  Vivid and eloquent, but not very academic sounding, and too long!  If you can think of a good name, please post it in the comments section below. 

To conclude, then, the science seems to support what those of us who experience language mixing feel instinctively: that our foreign languages are indeed stored in a separate area of the brain from our native language.  Although there's nothing in the science to suggest they are stored on top of each other, that's still how I imagine it (note my feeling 30 years ago that German had 'layered' itself on top of French; note Kristina's feeling that Spanish was 'buried' in her brain, and note Shannon's hypothesis of a 'push button' system in which words from different languages compete for the same slot).  How well we keep these languages differentiated depends on many factors including proficiency (how well we know the languages) and activation (how recently we've used them).  

Which brings me to those people who are able to switch between several languages simultaneously and easily.  How do they do it?  It's probably a combination of brain structure and constant practice. An integrated Broca's region means they are not running all their different languages as separate programs. Without that anatomical advantage, the next best option is to keep all languages activated by speaking (not just reading and writing) in them regularly.  

As it happens, even computers and their complex advertising algorithms are not immune to language confusion.  It's been two months since I came back to Spain from Germany, but Facebook is still sending me ads for 'Spanisch, ganz einfach!' 


The Seamstress, Maria Dueñas, Penguin, London, (2009)

Life in Ljubljana, Third language learning

Blakeslee, Sandra: When an adult adds a language, it's one brain, two systemsNew York Times (1997)
Ekiert, Monica:  The bilingual brain, Working papers in TESOL and applied linguistics, Teacher's College, Columbia University; Volume 3, No 2, (2003)

Nouns and verbs are learned in different parts of the brainScience Daily, (2010)

Tettamanti, M., Alkadhi, H., Moro, A., Perani, D., Kollias, S., and Weniger, D. (2002); Neural Correlates for the Acquisition of Natural Language Syndicates, NeuroImage 17, 700-709

Musso, M., Moro, A., Glauche, V., Rijntjes, M., Reichenbach, J., Büchel, C., and Weiller, C., (2003); Broca's Area and the Language Instinct, Nature Neuroscience 6, 774-781

Examining Chomsky's inborn universal grammar theory,

Language in the brain, 
Murphy, Shirin; Second language transfer during third language acquisition, Working papers in TESOL and applied linguistics, Teacher's College, Columbia University; Volume 3, No 1, (2003)
Shannon, Benny; Faulty language selection in polyglots, Language and Cognitive Processes, Volume 6, Issue 4, (1991)


  1. interesting blog. It would be great if you can provide more details about it. Thanks you

    iMarque - Form Processing

  2. Thank you for reading and commenting! Please send me an e-mail at and let me know what specific aspects of this article/blog you would like more details about. Thanks :)

  3. Hi Sarah, thank you for this blog entry! This explains why I always want to use Italian words in Polish class...Have fun in Spain! Best wishes, Itta

  4. Thank you Itta! I looked at your blog as well and I think it's really great! :)