Speak up

This exercise is a part of Educator Guide: There’s Extra Time to Learn a Language / View Guide

Directions: After students have had a chance to review the article “There’s extra time to learn a language,” lead a classroom discussion based on the questions that follow.

For overviews of world languages and their histories, see:

Bernard Comrie (ed.), The World’s Major Languages, 3rd ed. (2018).

Asya Pereltsvaig, Languages of the World: An Introduction, 2nd ed. (2017).

Nicholas Ostler, Empires of the Word: A Language History of the World (2006).


Discussion question:

1. What is the Indo-European language family?

The Indo-European language family is a group of hundreds of languages that are spoken in most of Europe, areas of European settlement and in much of Southwest and South Asia. Based on how similar or different those languages are from one another, scientists have constructed a family tree showing how far back each language branched off from the others. Using archaeological information about human migrations and settlements, and estimates about how rapidly languages change, it has been estimated when the major language branching events occurred. The Indo-European language family appears to have originated in or near what is now Turkey, around 6,000 to 8,000 years ago, with a language that has been dubbed Proto-Indo-European (PIE). There are similar language family trees for Native American languages, African languages, and languages spoken in East Asia and the Pacific Islands. Languages from very different parts of the world, say East Asia and Europe, have so few similarities that working out their connections is very difficult and rather speculative.

Extension prompts:

2. What is neuroplasticity?

An adult human brain has roughly 100 billion nerve cells, or neurons, and several hundred trillion synapses — the connections between neurons. As you learn new things and make new memories, synapses in your brain are constantly forming and rearranging. The ability to form new synapses and to change existing ones in response to a change in the environment is called neuroplasticity. While the human brain is certainly capable of neuroplasticity and learning throughout life, neuroplasticity may be greatest during childhood, when the brain is best equipped to learn languages and other new information.

3. What are mnemonics? Give an example of a mnemonic device that you have used this year.

Mnemonics are systems designed to make it easier to remember new information. Mnemonics have been around for centuries — some ancient writing systems used mnemonic symbols. Today, people often use written and verbal mnemonic systems, or devices, to learn new languages. But mnemonic devices can be used for pretty much anything. For example, “cat-ions” are “pawsitive” is one way to remember that cation molecules are positively charged.


Discussion question:

1. How can computer programs support language learning? Give specific examples.

Computer programs can provide real-time feedback to language learners, both in written and verbal forms. A program might engage a language learner by providing images of an object or action and having the student type or speak in the chosen language what is shown. A program might also provide vocabulary words for the student to define as well as prompts that require a student to use the correct verb conjugations or article/adjective/noun declensions. (Declensions are when the article, adjective or noun change form to indicate a number, gender or grammatical case). A program could evaluate a student’s performance on grammar or vocabulary assessments and follow up with more questions in that student’s areas of weakness. A program could detect differences in a student’s pronunciation when compared with the pronunciation of a native speaker. Then, based on the results, the program could provide examples and exercises to improve the student’s pronunciation.

Extension prompts:

2. How well can computer programs translate languages? Compare Google Translate (translate.google.com) and DeepL (deepl.com/translator). In the left-hand boxes of each program, type in a sentence or a paragraph in proper English. Then have each program translate it into the language of your choice. The translation will appear on the right side of the screen. Now, cut and paste the translation back into the left-hand box to have the program translate it back to English. What sorts of errors did the computer introduce? Compare these two programs.

Student answers will vary. DeepL translates fewer languages than Google Translate does, but tends to give more precise translations. DeepL usually puts words into the order that is appropriate for each translated language, whereas Google Translate often leaves words in the grammatical order that was appropriate for the original language. (Sound like Yoda, Google does.) Google Translate can also make some strange vocabulary mistakes. DeepL makes fewer mistakes.

3. What are some potential difficulties in conducting and analyzing survey studies, such as the one in There’s extra time to learn a language”?

Any tests involving humans or animals, even a survey, require safety and ethics approvals. In order for the results to have maximal statistical power, the number of volunteers should be as large as possible. The more complex the variations you are looking for are, the more volunteers you need, which is why this particular study recruited hundreds of thousands of volunteers. It can be difficult to find enough people, or to motivate them to spend the time and energy answering your survey. You have to figure out the questions that will yield the most useful information with the least time of the volunteers. You need to consider how accurate the responses from the volunteers are and what biases might affect volunteers’ answers? If the survey is written, it may not accurately test things that are not, such as fluency with spoken language. If the volunteer group does not include people from key demographics (such as young children within certain age ranges), it might throw off your results.