Using Diffit - Creating Differentiated, Tiered Readings

Earlier, I wrote up a blog post about learning about AI and how it can be used in the classroom, especially for creating readings in the target language. Here is another web app resource which Stefanie Gigante demonstrated: Diffit.

According to its website: What does Diffit Do?

• Adapt any reading, excerpt, article or video for any reading level - to help all students access the content you’re teaching. 
• Generate short informational or narrative texts on any topic, for any reading level.
• Cherry on top? Diffit creates an automatically generated summary, comprehension checks, vocabulary words, and more to go with whatever text you’re working with! 

Then export to our ready to use Google Slide, Google Doc, and Google Form Templates.

Essentially, Diffit can take any text and create a differentiated reading of it based on the level parameters which you set.

In her presentation, Stefanie gave an example of creating a differentiated text using a Cambridge Latin Course reading. I decided to do the same with a stage 27 reading, where Modestus and Strythio hide out beneath the granary. I entered the text with the following parameters to create the following readings (the unique word count is my own):

Original (202 words/152 unique words)

Modestus et Strythio, e carcere egressī, ad horreum fugerunt. per aditum angustum repserunt et sub horreō celati manebant. centurio Valerius, cum portās cellārum apertās carcerem desertam vidisset, iratissimuserat.

imperavit militibus ut Modestum Strythionemque caperent. militēs tamen, quamquam per tota castra quaerebant, eōs invenīre non poterant.

illī duōs dies manebant celati. tertiō die Modestus erat tam miser ut rem diutius ferre non posset.

Modestus: “quam infelix sum! malim in illō carcere esse potius quam sub hoc horreō latēre! qualis est haec vita? Adest Strythio, comes exiliī, sed me non adiuvat. nam Strythio est vir maximī silentiī, minimī ioci! eheu! me taedet huius vitae!”

Strythio: “mi Modeste, difficile est nobīs hic diutius manēre. Nunc advesperascit. visne me, ex horreō egressum, cibum quaerere?

Modestus: id est consilium optimum! i primum ad coquum! eum iube cenam splendidam coquere et huc portāre! deinde quaere Aulum et Publicum, amicōs nostrōs! invitā eōs ad cenam! iube Aulum amphoram vinī ferre, Publicum lucernam. tum curre ad vicum; quaere Nigrinam! optima est saltatrix! saltatricēs mihi solacium afferunt.”

Strythio, ut mandata efficeret, invitus discessit. persuasit coquō ut cenam splendidam parāret. persuasit Aulō et Publicō ut vinum et lucernam et aleas ferrent; oravit Nigrinam ut ad horreum venīret, sed ei persuadēre non poterat.

4th grade  (144 words/105 unique words)

Modestus et Strythio, duo homines qui ex carcere evaserunt, ad horreum fugerunt. Ibi se celaverunt. Valerius, qui erat centurio, erat iratus cum vidit portas cellae apertas et carcerem desertum. Ille imperavit militibus ut Modestum et Strythionem caperent, sed milites eos non poterant invenire per tota castra.

Illis duobus diebus, Modestus et Strythio manebant celati sub horreo. Sed Modestus erat tristis et miser. Vita sub horreo non erat bona. Strythio, amicus Modesti, non iuvabat eum. Modestus dixit, "Vita haec qualis est? Me taedet huius vitae!" Strythio respondit, "Difficile est nobis manere hic diutius. Nunc sol occidit. Visne me ex horreo egredi et cibum quaerere?"

Modestus acceptavit consilium Strythionis. Strythio discessit ut mandata efficeret. Ille rogavit coquum ut pararet cenam splendidam. Ille rogavit Aulum et Publicum ut ferrent vinum, lucernam, et aleas. Ille quaesivit Nigrinam, saltatricem optima, ut veniret ad horreum, sed ei persuadere non potuit.

2^nd grade  (106 words/77 unique words)

Modestus et Strythio erant duo amici qui evaserunt ex carcere. Ibi erat periculosum, itaque fugerunt ad horreum. Sub horreo se celaverunt et ibi manebant. Centurio Valerius, miles iratus, eos non poterat invenire. Modestus tamen erat tristis et misellus. Strythio dicit, "Nobis hic diu manere difficile est." Modestus respondet, "Vade ad coquum et iube eum cenam splendidam parare. Postea quaere Aulum et Publicum, amicos nostrum, et invita eos ad cenam. Iube Aulum amphoram vini ferre et Publicum lucernam. Tum curre ad vicum et quaere Nigrinam, saltatricem." Strythio discessit ut mandata efficeret. Coquo persuasit ut cenam splendidam pararet. Aulo et Publico persuasit ut vinum, lucernam, et aleas ferrent.

Observations

1. While it did create differentiated readings of the original passage, the various versions were still very vocabulary driven. I do wish there were a way to explain to AI how to create a reading which "sheltered vocabulary, not grammar." In other words, while the length of the passage was shortened, for a differentiated reading, being an "easier" read for some students is debatable.
2. I am still impressed at the grammatically-correct Latin (for the most part) which AI can produce!
3. Diffit can also produce assessment questions - this is a tool which I wish to explore further!

What is Behind the Door?

Here is a great 5-10 minute activity to use in class to reinforce vocabulary and structures (and even to preview vocabulary). It is a simple but very engaging prediction activity. I have seen Haiyun Lu demo this many times as part of her live online Mandarin session for ABC Camp, and then I saw Elicia Cardenas do this in-person at this summer/s Acquisition Academy in Dallas. Last week at the Fluency Matters Conference in Denver, I saw Haiyun demo this live again in one of her presentations. Although I had never done this activity before as a teacher in my own classroom, I decided to implement it in the adult participant Beginning Latin language class which I was teaching there, and it went really well! When I told Haiyun that I was using her activity, she told me that she had gotten it from Laurie Clarcq!

The activity is called "What is Behind the Door?" 

1. It does involve creating a slide presentation with images of vocabulary and animation, of which one of the images must be a door. There is an image "behind" the door. You need to layer it so that the vocabulary image is set as "back" behind the door.
   
2. Ask students to predict one of two choices behind the door by holding up a number using their fingers- either #1 or #2.
3. Using the slide animation to make the door disappear, you reveal what is "behind" the door.
4. Proceed to the next slide, and repeat.

   

Observations

1. Oh my gosh, adults can do this activity forever, so I can only imagine how much students would enjoy this! Janet Holzer told me, "People like it, because it is like betting!"
2. Although "What is Behind the Door?" is highly engaging, I would only do it for around 10 minutes at most to preserve the novelty and to keep students wanting to do it in the future.
3. I like that I can use it with single vocabulary words, phrases, and full sentences depending on what I want to target.
4. Extension activities - You can easily extend this into becoming a more communicative activity. In her demo, Haiyun says that it is a good way to get in exposure of numbers, because she can count aloud in Mandarin how many students say "#1" vs "#2. You can also use the images then as a springboard for a Picture Talk.

"Research Talks: Motivating Acquisition-Centered Classrooms" by Eric Herman - Book Review

While I was at CI Iowa, I attended two sessions presented by Eric Herman regarding academic SLA research. I had never met Eric previous to CI Iowa, but I was very familiar with his name due to his reputation for being the "research guy" in the CI/ADI community. In meeting him, I can tell you that although Eric is the "research guy" (a title to which Eric probably does not want to be referred lol), he is so much more than that, is incredibly down to earth, and presents it all in a very relatable, understandable manner. During one of Eric's sessions, Jim Tripp promoted a book written by Eric called "Research Talks: Motivating Acquisition-Centered Classrooms," and I am currently reading it now.

Here on this blog, I have stated numerous times that I absolutely hate reading academic research. Reading academic articles makes me feel stupid, because I have such a difficult time wrapping my head around everything being stated. Latinists will understand this reference: For me, reading academic research is like translating Cicero - while I can translate it into English, I still have NO IDEA what he is trying to communicate due to his complex sentences and structures.

This book is SO different though. For some reason, I want to call this book "Research Bites," much like the mid 90's film "Reality Bites" (sidenote : that is such a Gen X movie. I saw it a month before I was to graduate from UCLA, and I was so depressed for a few weeks after that - much like Winona Ryder and her friends in the movie, I had NO idea what I was going to do with my life post-college outside of graduate school, and that was just delaying the inevitable for a few years!). However, in some ways "Research Bites" could be an apt title, because Eric presents this research in bite-sized chunks. Published in 2019, this book is organized into 30-week chapters, with each chapter divided into 5 "days," with a single SLA research quote and citation for each day related to a theme: input, development, textbook grammar, communication, and beliefs. I believe that this book is based on his Acquisition Memos, which were daily "memos' sent out to his subscribers? In his introduction, Eric writes:

This book can be read straight through or spaced out over time. My recommendation is to read and reflect on your quote per school day, providing you with daily inspiration to align language teaching with language science. The quotes have been organized for this purposes (page vi).

I like the ability to reflect on and to interact with a single academic quote. And while it is only a single citation, that citation packs a punch in what it conveys about SLA research! While I do possess a great deal of second-hand academic knowledge related to CI/ADI which I have learned from others, what I like about Eric's book is that it is both providing and exposing me to the actual academic research sources from where this is coming, thus helping to fill in some gaps. Already, some of my own personal annotated notes for quotes include:

• Even in 1977, researchers were wary of trendy educational reforms!
• Wow, this research predates Krashen!
• This!
• Latinists were questioning this (grammar translation method) back in 1657??!! Preach it, Comenius!
• There were "ADI" advocates in 1973??

Even more so, although for me Eric is preaching to the choir, I do appreciate his attitude towards those who may read his book and experience some objection to some of the quotes and ideas presented. He writes:

There is a significant gap between SLA and mainstream language teaching, so some readers will undoubtedly experience some dissonance...I hope that this book will motivate future learning, including exposure to alternative ways of teaching (page vi).  

That quote typifies Eric so much! He will not try to hit you over the head with academic research, but if you ask Eric (or even disagree with him) about something related to SLA research, he will enter in a friendly dialogue with you but will back it up a wealth of knowledge. I highly recommend this book - it is a definite must-read!

P.S. Remember how I hate reading academic research? Due to the compelling nature of many of these quotes, I have actually gone to Google Scholar/JSTOR to find the original articles to learn more, and I have read a number of them! I will admit that many still feel like a "forced march through Gaul" due to their high-level academic nature, but I am actually enjoying reading that which I am able to understand. Progress!

Playing Around with ChatGPT - Creating a Story

This past weekend, I attended the American Classical League Summer Institute in Tucson at the University of Arizona. It had been 5 years since I had attended a Summer Institute fully in-person since the pandemic, and at first, I was not sure what to expect. I am so glad that I attended, because 1) there were a lot of very good presentations there and 2) I had forgotten how much I missed my Latin teacher community. (I was also there as one of the recipients of this year's Merens Award).

One of the presentations which I attended was "Artificial Intelligence (AI) for Latin Teachers" by Stefanie Gigante. I have had little to no training on AI in the classroom, and quite honestly, I have been very leery of how it can be used properly. Stefanie gave an example of how ChatGPT can be used to create a lesson plan, and the key is to be as specific as possible with your parameters in terms of what you want. She gave us a bit of time to play around with it during the presentation, so I ended up asking ChatGPT to write me a little story in Latin using some specific vocabulary and grammar forms. It produced a very short paragraph in Latin for me, and I was amazed! 

Cut now to being home from the Summer Institute, and today I decided to play around a bit more with ChatGPT in asking it to write me a story with some VERY SPECIFIC parameters:

This is the story which ChatGPT produced:

However, I wanted to see if I were to do it again and to give it the same parameters, would ChatGPT produce the same story? The answer is NO!

Observations

1. Wow, ChatGPT actually produced a story in Latin which makes sense, i.e., it did not create a bunch of gibberish using Latin words. In both versions, there is an actual plot with a beginning, middle, end, conflict, and resolution!
2. It took less than 30 seconds for ChatGPT to create a story - I would venture to say around 15-20 seconds!
3. For the most part, the grammar is spot on (with a few minor errors in each version).
4. ChatGPT basically fulfilled all of my parameters except for the imperfect tense, perfect tense, and gerundive of purpose. I may have to be more specific with my parameters next time or better learn how to get ChatGPT to include that in a story.
5. For the purpose of creating a story, ChatGPT did add a lot of words in addition to those which I had requested. 
6. I do not know to what degree ChatGPT will ever understand the concept of "sheltering vocabulary, not grammar," but I did notice some different grammatical uses of the verbs - maybe it needs more input regarding the concept? Maybe I need to say that each vocabulary word which I list has to be used at least X times in the story?
7. I made a spelling mistake in my parameters in asking for "neecsse est" to be used, but ChatGPT "knew" that I was actually asking for "necesse est."

I am absolutely blown away by this tool! Essentially, AI could help write a framework for a story which I can then edit and "CI/ADI" it up! There are tons of other uses which AI can be used in lesson planning, so using some other AI web app resources, I will investigate writing up assessment questions, creating tiered readings for differentiation, and creating images. What success have you had with CI creating a reading in the target language?