Tuesday, 30 September 2025

Emergent Language and GenAI: New Opportunities for Learner Autonomy

I wrote about some ideas I had about 'emergent language' after reading an article by Richard Chinn in The University Grapevine. I wasn't entirely happy with what I had written so I asked ChatGPT, "Can you challenge me on the points I make here with a view to me finding better ways of expressing them?"

I was offered two responses and from the one I preferred I selected some changes to make. I put these to ChatGPT and was offered some suitable titles, subtitles and an introductory paragraph. I chose the ones I liked best. There was another round of changes, but I didn't feel the resulting text was written in 'my voice'. So I planned to offer my website as an example of 'my voice' and ask for the article to be written that way. (I had to wait till one o'clock the next day to try that as I use a free account!)

I said, "If I add the URL of my blog, can you try to replicate my voice in a new version of this article? https://onewaytotefl.*****.com/"

Here's a link to the whole conversation showing how ChatGPT helped me.

The article below is the result. Do you agree with everything in it?


Emergent Language and GenAI: New Opportunities for Learner Autonomy

Exploring how spontaneous language, AI feedback, and task repetition can foster independent, lifelong learning.


Broadening the definition

Emergent language is often defined by what teachers notice and work with during a lesson. It’s usually tied to how a teacher reacts to something a student says in real time — for example, picking up on a phrase, drawing attention to it, and building an activity around it.

But in practice, emergent language isn’t owned by the teacher. It’s any spontaneous communicative language that learners produce, whether or not it’s noticed, corrected, or exploited. And that broader view is more useful now than ever, because learners have tools at their disposal that can help them exploit emergent language for themselves.


What emergent language is

Emergent language is the language learners produce spontaneously to express ideas. It can be spoken or written, in or out of class.

Teachers can choose to work with it — by highlighting an interesting word, unpacking a grammar point, or showing alternative ways of phrasing. But emergent language doesn’t depend on teacher action. It exists whenever learners communicate in real time. Even without feedback, using the language is part of the learning process.


From teacher-led to learner-driven

Traditionally, it’s been the teacher’s role to guide learners in noticing and refining emergent language — through scaffolding, feedback, and prioritising what’s worth focusing on.

But today learners have access to GenAI tools that can play part of this role. They can paste in something they’ve written or record themselves speaking, then ask the AI to respond in useful ways.

A teacher still has a role here: helping learners ask better questions, interpret responses, and decide what to take away. For instance, a learner might ask for a correction, but a more productive prompt might be:

  • “Give me three alternative ways to say this.”

  • “How could I make this sound more polite?”

These prompts don’t just fix mistakes; they show learners new choices.


Experimenting with prompts

Learners should experiment with prompts and see which kinds of feedback they find most helpful. The “best” response often depends on the type of text or task. Sometimes a short alternative is enough; at other times, it helps to see a reworked version of the whole text.

This experimentation is part of developing autonomy. Learners discover how they like to learn, and what makes feedback stick.


Task repetition and practice

Once learners have feedback, the next step is practice. Repeating a task immediately after feedback helps consolidate learning. Repeating it again later — as in spaced practice — strengthens retention.

There’s no single formula for timing or frequency, but a combination of immediate and delayed repetition works well. The important thing is that learners feel they’re progressing, not just repeating for the sake of it.


The teacher’s role

Can learners do all this on their own? Possibly. But teachers still add value. They can help learners become more efficient, avoid overwhelm, and focus on what’s most useful.

Our aim shouldn’t be to keep control of emergent language, but to help learners take ownership of it. Encouraging learners to notice, reflect, and act on their own emergent language — with or without AI — is part of preparing them to be lifelong learners.


Final thought

Emergent language happens whenever learners use the language to communicate. Teachers can work with it, but learners can also exploit it themselves — especially with the support of AI. If we, as teachers, help them develop the skills to use feedback effectively, we’re not just teaching language for the moment. We’re equipping learners with habits and strategies they can use for a lifetime.


Saturday, 27 September 2025

Turboscribe can be used (with ChatGPT) to create quizzes on an audio or video file


Subtitles

Turboscribe can be used to get a transcript of any audio or video file using a free account as long as it is less than 30 minutes long. There is a limit of three transcriptions every 24 hours.The transcript can be exported as an .srt file which means it can be used to add subtitles to a video on Youtube. As Turboscribe (and Clipchamp) are better at making transcripts than Youtube and in both cases any mistakes can be edited out quite easily, there is a lot to be said for uploading these .srt files to your videos on YouTube. See here for a comparison between subtitles from Turboscribe and Clipchamp.

Quizzes

Personally, I am not a fan of comprehension questions as I prefer getting students to do the much more demanding task of retelling the story, but many teachers DO want comprehension questions - I only used them a week later as a test of memory or a scaffold for one student to retell the story to someone who had missed the previous week's class.

But because Turboscribe has a way of linking the transcript of a recording to ChatGPT by selecting ChatGPT from the menu and then selecting Custom Prompt, it can be used to create multiple-choice comprehension questions.

The Custom Prompt I used was, "Can you make five multiple-choice questions about this story? Give the ansers at the end of the quiz"

Here they are:

1. What happened at the start of the story?
a) Hannah saw Jamie at the theatre
b) Hannah nearly hit a man with her car
c) Jamie called Hannah
d) Hannah stopped to buy coffee

2. What did Hannah do after stopping the car?
a) She called for help
b) She stayed in the car
c) She got out and spoke to the man
d) She drove away

3. Where did Hannah and Jamie go together?
a) To the theatre
b) To a park
c) To a coffee bar
d) To Hannah’s house

4. Why was Jamie in the street?
a) He was going home
b) He was buying theatre tickets
c) He was meeting friends
d) He was looking for Hannah

5. How did Hannah feel when she heard about the tickets?
a) She was angry
b) She was surprised and happy
c) She was bored
d) She was worried

To be able to answer the questions you could either listen to how Carla retold the story (top), or you could listen to the version corrected by ChatGPT (below).

 Carla

 Ada



Impact of Long-Term Extensive Reading Programs on Students

These three transcripts originate from separate plenary sessions at the ERWC7 conference, focusing on the topic of Extensive Reading (ER) and its role in language education. The first features David Beglar, who discusses the promises and limitations of ER for developing reading fluency, arguing that while extensive reading is crucial, it should be supplemented with other intentional fluency activities. The second presents Akio Furukawa and details the six-year extensive reading program at SEG, sharing student success stories and the methodological approach of starting with simple stories and gradually increasing complexity, alongside the importance of the Yomi Yasa Level (YIL) grading system. Finally, Yoko Yamashita provides an overview of an ER program in a Japanese school and the results of a graduate survey, highlighting the long-term positive impact of the programme on students' language skills, cultural awareness, and the development of a global mindset. 

I'm a fan of NotebookLM and this summary of the three plenaries at the ERWC7 conference gives a great introduction to what the three speakers covered.

The Video Overview adds to that and here it is at 7 minutes long with subtitles extracted by Turboscribe.




Here's the same video with subtitles generated by Clipchamp




Here are the three plenaries from the ERWC7 Conference:








Thursday, 25 September 2025

A 40-year Reading Experiment described in 8 minutes!



This is a Video Overview produced by NotebookLM based on just one source - the YouTube video of my presentation to the JALT ER Sig, which is below.

It is a bit over the top in its praise for me, but it encapsulates well what I said in only 8 minutes. There are a couple of places where it gets details wrong, but I've added in an * in the subtitles and a note in red at the top of the screen when these occurred.



Friday, 12 September 2025

ChatGPT simply doesn’t understand how to grade or create texts on the IELTS scale

I offered a week ago to adapt my 9-step prompt for studentsto use to get feedback on their speaking from, which uses the CEFR scale and  the Global Scale of English to work with the IELTS scale. In many countries the IELTS scale is used more often than the CEFR scale.

Here is a table showing how grades on the IELTS scale correspond to the other two scales:

It also shows the levels of the different versions ChatGPT produced from the original transcript of “Bargain Sweater”.

Unfortunately, after experimenting with seven different variations of my 9-step prompt with various differences between steps, I began to suspect that ChatGPT had no idea about how to rate texts or create them at different points on the IELTS scale. I wanted to test this.

Daniel Tena Calderon had experimented with using Gemini to do something verysimilar to what I’m trying to do with my 9-step prompt, so I decided to see how the two, ChatGPT and Gemini, compared using the same transcript and the same prompt.

I asked each of the LLMs to grade the five versions on the IELTS, CEFR and GSE scales and they both asserted that they had done what they had been asked to do: produce texts that go from the original to half a level up and then one level further and one level further again.

 

ChatGPT

Check

Gemini

Check

Original

4.5-5.0

3.5-4.0

4.0-4.5

3.5-4.0

Corrected

5.0-5.5

3.5-4.0

5.0

4.0

½ IELTS level up

5.5-6.0

4.0-4.5

5.5-6.0

4.5

1 more IELTS level up

6.5

3.5-4.0

6.5-7.0

6.5-7.0

1 further IELTS level up

7.0

4.5

7.5-8.0

6.5-7.0

Differences

 

9.5 too high

 

3.5 too high

 

Gemini appeared to grade the original more severely, but appeared to reach higher levels later in the sequence.

I haven’t got a tool to independently check if the LLMs were right, but I have used an average of the CEFR grades given by Text Inspector and Pearson’s Text Analyzer as a reliable way to grade texts. I then used the table above to predict the level on the IELTS scale and I’ve added a Check column with the results for each step for each GenAI tool.

Gemini was much closer to the checked levels and ChatGPT was so far out that I think it is fair to say that ChatGPT simply doesn’t understand how to grade or create texts on the IELTS scale.

Before writing this, I made a screen recording making the same point and illustrating it with the texts created by the two LLMs. As usual, I used Clipchamp to make a couple of edits and to generate the subtitles file, which I then used to create subtitles on YouTube. I experimented with various formats of the video and subtitles and uploaded the four formats to YouTube:

·         Wide screen with subtitles to the right

·         Portrait with subtitles at the bottom

·         With red subtitles on top of the screen recording

·         With subtitles you can turn on or off