As a science teacher, I teach very vocabulary-heavy content, and this is a particular challenge for EAL students. A solution I’ve started working on to support EAL students with their vocabulary is using artificial intelligence to build bilingual keyword glossaries like this one for each new topic and giving these to students as a printout.
It makes learning more efficient, has the potential to minimise cognitive load, and supports their language acquisition.
As an international school teacher I work with students from all over the world, most of whom do not speak English as their first language, and this is a barrier to their learning. What’s more, the content I cover often has a LOT of complex vocabulary (just take a quick look at the first few lessons in IB ESS Topic 1.1 and you’ll see the amount of new language my students have to digest). Direct translation of new vocabulary (i.e. via Google Translate) does not always support understanding and, at times, it can sometimes create or reinforce misconceptions.
I’m a science teacher but I’ve also worked as an EFL (English as a Foreign Language) teacher in the past. I’ve taught in 5 different countries outside of my own and so, for the entirety of my teaching career, I’ve had classes with a large portion of English Additional Language students. We all know the importance of differentiated instruction as educators and, in my classroom, catering for a wide range of English language abilities and finding ways to support EAL students has always been a factor in my lesson planning.
AI is now opening doorways, allowing me to provide support materials in my students’ first languages – something I couldn’t have done just a few years ago.
What My Bilingual Glossaries Look Like
I produce the glossaries in French, Japanese, Mandarin, Russian and Thai (as well as an English-only version), and plan to expand this list over time.
My initial plan was to include the English key term, its translation, its definition in English, and its translated definition but after I got ChatGPT to create a draft version I asked it, “Is there anything else I might include that would support my EAL learners?” It gave me several suggestions, but the one that seemed most impactful was example sentences showing the key terms in context. So I added an example sentence (in English) along with its translation for each key term.
This is great because it aids students’ understanding of the key term (by seeing it used in their own language), but also supports their English language acquisition by seeing the way it can be used in English. This is far better than a simple one-word translation.
You can see an example of one here: IB ESS Topic 2.5 Bilingual Glossary (English, French, Japanese, Mandarin, Russian, Thai).
How I Make Glossaries With ChatGPT (and how you can too)
I recently got access to the paid version of ChatGPT. Most of what I’ve been doing can be achieved in some way with the free version, but the paid version allows you to create your own GPTs and give it instructions to use for every query, which makes it a lot more efficient and specialised towards a given task.
Training the Model
First I had to train the model to look in the right place for course information. This GPT is intended specifically for IB Environmental Systems and Societies (and I plan to create a new one for my IGCSE Biology classes), but the ESS course got a significant update this academic year and ChatGPT tended to find information on the previous version of the course, so it needed instructions on where to look.
Before even thinking about the word list and definitions, I focused on formatting. I wanted everything laid out in a table, with the translations in the same cell as the English content and had a few other specific requirements regarding formatting. I had a few conversations with the model before the result looked how I wanted, then I asked the model, “What do I need put into your instructions so you format my glossaries like this every time?” It gave me the instructions and I pasted them into its own instruction section.
Side note: it gave me instructions in bullet points, but the instructions have to be entered into the GPTs instructions section in prose, so I asked it to re-write in prose before I copy/pasted.
Making the English Version First
For each glossary I make, I start by asking for a word list: “Give 20-30 keywords for topic 1.1”. I take the list, remove the terms I don’t think matter, and add in any I think are missing.
Next I paste the list back in and say, “Give me an english glossary for topic 1.1 using these terms.” Keep in mind it knows how to format a glossary now. I also add, “Do it alphabetically” because the model generates the keywords list in the order it “thinks” them up, not alphabetically as a good glossary should be presented.
When I’ve got the definitions and example sentences (English only) I copy/paste them into a Google Doc and I go through them with a fine-toothed comb, making tweaks and edits so the definitions are appropriate to the syllabus. I find I don’t usually need to edit the examples sentences; the model does a good job of using the key terms in the context of the chapter which is ideal.
Generating the Bilingual Version
I then download that Google Doc and upload it to the GPT’s knowledge bank. Now all I need to tell the model is, “bilingual glossary, English-Thai topic 1.1”. It then uses the English version in its knowledge bank and produces the bilingual one, ready for me to copy/paste into a new Google Doc and print.
One side note: ChatGPT is a little limited in how much work it will do at once, so it generally only does the first third or quarter of the glossary table, then asks if I’d like to continue. Unfortunately this means copying and pasting chunks of tables and joining them together in Google Docs, which makes the process a little more laborious and time consuming. I had initially hoped I could say, “bilingual glossary in French, Japanese, Mandarin… [and others]…” and come back 5 minutes later when it was all done for all languages at once. In practice, it won’t even complete the glossary for one language in a single instance, which is a shame.
The Benefits of a Bilingual Glossary
Learning Efficiency
The conventional process when students see a a new word or phrase goes like this: Read, see unknown word, pause, open laptop, navigate to Google Translate, re-read the new English word, type it into the translator, read the translation, note down the translation.
That takes time! What’s more, I don’t like my students having to flick from a paper task to a laptop and back again, or between tabs in their browser, when they really need to just dedicate themselves to the task at hand. A bilingual glossary does not remove the time it takes to interpret new, unknown vocabulary, but it takes some of the inconvenience away having the information at hand, making the process smoother and more efficient. This will certainly support EAL students in their science learning.
Add to that, the keen students will take the glossaries home and read through them before next lesson, meaning much of the “new” vocabulary isn’t so new by the time they need to use it in class.
Reducing Cognitive Load
I’ve never been in a situation where I’ve had something new to learn and only been given the information on it in a foreign language, but this is the situation my students are in every day in school.
I often see them translating words in the worksheets and notes I give them meaning that, in addition to interpreting the meaning of the question (in the way a native speaker also would), they also have the added strain on their working memory of doing the translations as well.
Having a bilingual glossary at hand probably won’t change this – they still have new terminology to learn, putting strain on their working memory. However, it gives students the chance to see the vocabulary in advance. As I’ve said, those that are keen will review the glossary between lessons and so, by the time they come to class, the new vocabulary might not be so new. Anything that reduces this cognitive load in class can only be helpful in the learning process.
Language Acquisition (in both languages)
The benefits regarding language acquisition are two-fold: it benefits both their English acquisition and their native language acquisition.
If students rely only on translators, they may successfully learn how the English keyword translates to their own language, but they’ll have limited understanding of its meaning (unless the context makes it very obvious) or some of the ways the vocabulary can be used. The definition in the glossary, as well as the example sentence, gives students the information they need to effectively build this new vocabulary into the mental schema they have for the subject content and language.
The glossaries also help with first language acquisition. Many of the terms the learners come across are new to them in any language, including their native tongue. Think about it… let’s say you’re studying in Japan, in Japanese, and you’re presented with a new word that you don’t know. You translate it, and find the English translation is phototropism. Do you now understand this Japanese term and how to use it? If you were a student early in your science education, the word phototropism would probably have no meaning to you in any language.
What I’m saying is, it is useful to students if I help them understand new vocabulary in both English and their own language. And it’s also important to keep in mind that translation alone does not alway equal understanding.
Some of my students will study in English in my school, but go on to university or into a job in their home country. If they’ve spent their whole high school education learning subject-specific vocabulary in English, they’ll be behind their peers in Higher Education who know this terminology in their own (non-English) language of study. Supporting first language acquisition can only help students in the long run.
Conclusion
The barriers to learning for EAL are many, and there’s no silver bullet. However, there are plenty of strategies we can implement to make the process a little smother, and make our lesson content a little more accessible.
AI has opened a doorway where we can now start creating learning resources that use a student’s own first language and that is significant!
Give them a try yourself (the free version of ChatGPT will do a pretty good job if you don’t want to pay for it yet). And if you’re an IB Environmental Systems and Societies teacher have a look at my IB ESS glossaries like this one for topic 2.5, and keep an eye out for more on the way!