Language Learning and AI: 7 lessons from 70 years (#2)

2. Communication in context

Oxford (1993) desired that “communicative competence must be the cornerstone of ICALL” (p. 174), noting that many ICALL projects of her time did not meet that goal, although communication and by extension communicative language teaching have been central ideas in applied linguistics for decades. Canale and Swain (1980) transferred the concept of communicative competence by Dell Hymes – developed in opposition to the Chomskyan linguistic competence – from sociolinguistics to language learning. Hymes (1974) had introduced communicative competence with the mnemonic SPEAKING = Setting and Scene (time and place), Participants (speaker and audience), Ends (purpose and outcome), Act Sequence (progression of speech acts), Key (tone, manner), Instrumentalities (language modalities), Norms (social rules), and Genre (kind of speech act or text) (pp. 53–62). These different facets and components of communication go well beyond the idea of a generative grammar (Chomsky, 1957) and that of linguistic competence. In ICALL as in NLP generally, however, generative grammar and other formal grammars, which are sets of rules that rewrite strings using mathematical operations, are the backbone of a system. The partial disconnect between formal grammars, such as Head-driven Phrase Structure Grammar and Categorial Grammar, and communicative competence with its focus on meaning, situation, and context, as clearly illustrated also by Hymes’ mnemonic, meant that ICALL systems did hardly play a role in communicative language teaching.

3 computer laptops talking to one another — Image generated in WordPress.com

This is the third part of a short series. All parts are based on a manuscript that I wrote recently. Part 0 gives a historical introduction. Lesson 1 focuses on the necessary exposure to authentic language and whether this can be done with GenAI. And I mean exposure and not so-called comprehensible input.

The GenAI chatbots, however, have been said to be a suitable conversation partner (Baidoo-anu & Owusu Ansah, 2023) and learning buddy (https://www.khanmigo.ai/). The GenAI output in a number of languages is certainly well-formed and plausible; GenAI’s natural language understanding is fast and precise. But is a conversation with a chatbot the same as a human conversation? Is it a negotiation of meaning as understood in communicative language teaching? The NLP researcher Emily Bender and her colleagues compared GenAI chatbots to stochastic parrots, which skillfully aim but proceed by guesswork (see Merriam-Webster, n.d.), and argue that “coherence is in fact in the eye of the beholder. Our human understanding of coherence derives from our ability to recognize interlocutors’ beliefs … and intentions … within context … That is, human language use takes place between individuals who share common ground and are mutually aware of that sharing (and its extent), who have communicative intents which they use language to convey, and who model each others’ mental states as they communicate” (Bender et al., 2021, p. 616). In other words, the chatbot spits out forms that are plausible but that do not mean anything; the (student or teacher) reader imbues these hollow forms with meaning and thus anthropomorphizes the GenAI tool, by then reasoning about its ‘intention’ and basing their response on the result of the reasoning, as humans do in conversation. Computers, however, do not have or formulate intentions. Something was clicked, data was input, and a condition was met. This triggered a digital operation, and forms that are numbers to the computer and look like words to the human user of the device became visible or audible. We add meaning after the form of the text has been generated.

My inspiration for this title came from the book
Snyder, T. (2017). On tyranny: Twenty lessons from the twentieth century. Tim Duggan Books.

I am sharing these early drafts of a book chapter I published in
Yijen Wang, Antonie Alm, & Gilbert Dizon (Eds.) (2025),
Insights into AI and language teaching and learning. Castledown Publishers.

https://doi.org/10.29140/9781763711600-02.

References

Baidoo-anu, D., & Owusu Ansah, L. (2023). Education in the Era of Generative Artificial Intelligence (AI): Understanding the Potential Benefits of ChatGPT in Promoting Teaching and Learning. Journal of AI, 7(1), 52-62.

Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021). On the Dangers of Stochastic Parrots. In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (pp. 610-623). https://doi.org/10.1145/3442188.3445922

Canale, M., & Swain, M. (1980). Theoretical bases of communicative approaches to second language teaching and testing. Applied Linguistics, 1(1), 1-47.

Chomsky, N. (1957). Syntactic Structures. Mouton.

Hymes, D. H. (1974). Foundations in Sociolinguistics: An Ethnographic Approach. University of Pennsylvania Press.

Merriam-Webster. (n.d.). Stochastic. In Merriam-Webster’s unabridged dictionary. Retrieved December 30 from https://unabridged.merriam-webster.com/unabridged/stochastic

Oxford, R. L. (1993). Intelligent computers for learning languages: The view for Language Acquisition and Instructional Methodology. Computer Assisted Language Learning, 6(2), 173-188.

… to be continued …

Sign here

On the complexity of change

Hey, Friend,

Yes, I have never signed these blog posts. I did not think it necessary. I write a lot more than a signature about myself in two other places: the short version and the long version. No signature on these pages either. I guess, one just does not do that. People put their sign in other places. I sign emails. And so do my colleagues. The other day, one colleague told me that he and a few others did an informal study of these automatically added bits of text, and they all agreed they liked mine. I really appreciated the comment. I had had comments on my emails on occasion, but never on the signature. This is what it looks like during the – work – day.

=================
Mathias Schulze, PhD

https://texterium.org/ and @mat_schulze and https://pantarhei.press/
he works at SDSU-LARC and the Dept. of European Studies
and now lives on the land of the Kumeyaay
[SDSU logo]

The full name, three short lines of text, and a small picture of four letters. Others have shorter signatures or no added auto-text at all. And yet others, have auto-text that is longer than the text body. So, what does it say? I sign – from Latin signare “to set a mark upon, mark out, designate; mark with a stamp; distinguish, adorn.” There often is a lot in a word, and I believe it is useful to play with words.

A tiny piece of canned text. Where is the complexity? What does this have to do with change? The change makes it complex.

In the beginning, I had no signature. Well, before that I had no email. In the early 1990s, I started with Pine and then Elm. I subscribed to some distribution lists to get at least an email or two a day. The list I remember fondly got me one or two essayistic and narrative texts from Mark Warschauer each week. In the late 90s – I still had no signature – the IT staff told me that I should delete some read and saved emails in my folder structure, because I was getting very close to my quota of 10 Megabytes on the mailserver. [Exponential growth in storage space, the staff could not predict!] And then the signatures came, giving complete mail addresses and telephone and fax numbers. Was I hoping to get a letter or phone call instead? And then, titles. Others added accolades. The fax number, I deleted soon from the “bottom lines”. Address and phone number were longer than the URL of a home page and my sig was shorter again. I like it that way.

Colleagues began to add quotes, statements, and stories. I added links. I like traffic on the blogs, on a website, on Twitter, Facebook, and Instagram, still hoping there will be days on which I which I am using more time writing and reading for these media than for email. And colleagues added gender identification, job titles, and land acknowledgements.

I have lived in many different places; does it still matter what I am a native of? As a boy, I watched East German Westerns, in which the Sioux, the Delaware, the Mohicans, the Apache, and the Seminoles were the main characters. Courageous, strong, smart, and true. I wanted to be like them. And in 2001, I got much closer. For sixteen years, I lived on the Haldimand Tract, six miles on each side of the Grand River, the land of the Haudenosaunee. I like to acknowledge that I am now beginning to learn about the Kumeyaay.

With many first names, I am unsure about the person. Should I use he or she when I email about Pat? Or Sasha? It’s a new custom to list pronouns after the name. They are short and clear. As a linguist, I did not want to list them; these little words live from their context, and I did not want to deprive them unnecessarily. So, I wrote a little sentence in the signature. The subject pronoun got an antecedent (a word to which it refers back) – the signer of the email by name – and it hooks up with the verb work, which makes it happy, grammatically speaking. And that’s what I do during the day – work. Like many, many others. I try to do my work as best I understand, as they do theirs. And then I write and read emails. With a short signature. On the weekend, I carve out some time to write other texts. No signature. Yet.

I am getting back to writing these short – or longer – texts on the complexity of change, exploring different examples and grappling with necessary concepts. Using my theoretical lens to better understand practical things, stuff I stumbled over, processes that proceed or peter out, … You can look at them in the order in which I wrote them, and once in a while, I make an attempt to sort them in a linear order.

What has language got to do with it?

Hey, Friend,

Mind your language.
The language of COVID-19.
The language of instruction.
Language and cognition.

Hope I did not startle you with the imperative: mind your language. I do mean you personally, but I do not mean to imply that you have said something inappropriate. It’s simple: all of us should be mindful of the language we use. All the time. I know this is hard to do. I have failed miserably. Too often. It is more something to strive for, to be aware of. We are focused on What we want to say. We inform, explain, promise, declare, question, list, pray, or baptize. All by using language. We rely on our utterances when we teach, train, coach, or mentor. And it is this area I want to focus on in this blog post.

Personal confession: I have been studying language, teaching it, and teaching about it for many years. I enjoy thinking about language and texts. Taking them apart and putting them together (again). Language or the way we communicate is one of our characteristics that makes us who we are—human, so I believe. Language and thought – or more technically: cognition – are inextricably linked in many different and complex ways. And yet, we are all able to use a language we grew up with without ever having to necessarily learn about it, think about it, or reflect on it. What a luxurious gift!

So, we all got a gift. Would it not be better to take good care of that gift? One polishes it to make it shine for the joy of others. Another monitors it to make it a precise and useful instrument. And a third ensures that no injury or misunderstanding results.

And if each of us strives to do all three at least most of the time, then I would call that being mindful about how we use language. This can improve all of our social interaction, and I will focus on the role of language and how we use it when we teach, train, coach, or mentor in this series of blog posts.