Post written by François Grosjean.

Any bilingual will tell you that there are concepts that are best articulated in a particular language. In fact, when bilingual friends or acquaintances fumble for the right word or expression, how many times have we not heard, or proposed ourselves, "Try your other language". But words are just a small part of our knowledge. What about other forms of knowledge that we have stored in our memory?

Northwestern University researcher Viorica Marian has spent many years studying the link between language and memory. In one of her earlier studies, conducted with renowned cognitive scientist Ulrich Neisser, she interviewed a number of Russian-English bilinguals, in English and in Russian. They were given English prompt words in the English part of the study (e.g. "summer", "neighbors", birthday", etc.), and Russian prompt words (translation equivalents) in the Russian part. The task of the bilinguals was to describe an event from their own life that the prompt word brought to mind. They were also asked, after the study, to indicate the language they had spoken, had been spoken to, or were surrounded by, at the time that each recalled event took place.

What the researchers found was that the bilinguals accessed more Russian memories when interviewed in Russian than when interviewed in English, and more English memories when interviewed in English than in Russian. They concluded that the accessibility of autobiographical memories was improved when the language used at the time of remembering corresponded to the language in which the memories were initially formed.

But can the accessibility of general knowledge, and not just autobiographical knowledge, also be guided by language? In other words, is factual knowledge acquired in a particular language more likely to be recovered when the same language is used at the time of recall?

To study this, Viorica Marian and Margarita Kaushanskaya examined the retrieval of general knowledge in Mandarin-English bilinguals by means of three tasks: a multivalent task in which they tested the retrieval of multiple items in a category within each language, a bivalent task in which the questions asked had two possible correct answers, one in each language, and a univalent task where there was a single correct answer, in just one language.

For example, in the bivalent task, the bilingual participants were asked, either in Mandarin or in English, "In a famous love story, what were the names of two lovers who died because of family disapproval?". Another example, this time a request, again expressed either in Mandarin or English, was, "Name a statue of someone standing with a raised arm while looking into the distance". What the researchers found was that the participants were more likely to access information that had been encoded (learned) in Mandarin when interviewed in Mandarin (the encoding information was obtained by questioning the participants at the end of the experiment), and more likely to access information encoded in English when interviewed in English.

Thus, for the first question above, the participants were more likely to say, Liang Shanbo and Zhu Yingtai when asked in Mandarin, and Romeo and Juliet when asked in English. Likewise, they answered more readily Chairman Mao when asked the statue question in Mandarin, and the Statue of Liberty when asked in English.

When the bilingual participants were given the multivalent task (they were prompted with "lakes", for example), the researchers found similar results. Mandarin responses such as Qinghai Lake, Lake Poyang or Lake Tianchi were more likely to emerge during the Mandarin interviews and English responses such as Lake Michigan, Lake Ontario or Lake Erie were more likely to be given during English interviews. As for the univalent task (where there was only one answer, encoded in only one language), even though language did not influence what particular memory was accessed, it did influence how quickly it was accessed.

The authors concluded that different types of knowledge are differentially sensitive to language-dependent memory, with language dependent effects more likely when multiple alternatives are available for retrieval. Thus the bivalent and multivalent results could be due to the fact that whenever more than one correct answer is available, a selection mechanism relies on additional markings, such as the encoding language at the time of learning, to choose the answer.

So the next time you try to remember something, and you can't seem to do it in one of your languages, try changing language ..... IF you are bilingual, of course. It might just work!

 Photo of a woman thinking from Shutterstock.


Marian, Viorica & Neisser, Ulrich (2000). Language-dependent recall of autobiographical memories. Journal of Experimental Psychology: General, 129(3), 361-368.

Marian, Viorica & Kaushanskaya, Margarita (2007). Language context guides memory content. Psychonomic Bulletin and Review, 14(5), 925-933.

"Life as a bilingual" posts by content area.

François Grosjean's website.

Most Recent Posts from Life as a Bilingual

Poetry in a Second Language

A language learning secret from the past

What Does It Mean to Be Dominant in a Language?

Domains of language use and dominance in bilinguals