The Growing Connectedness of Things: How English Words Have Evolved Since 1953

In this article, TOEFL ‘s John Clark interviews Jasmine Nackash, author of a brilliant visual essay in The Pudding, How the Words We Teach English Language Learners Changed, and What That Says About Us. (We recommend exploring Jasmine’s piece before you read this discussion!)

Jasmine, I sent you a cold email after reading your fantastic essay and was surprised to learn you’d previously taken TOEFL. Forgive the shameless plug – why’d you take the world’s English test?

Of course! I took it when I applied to NYU to pursue my master’s, almost four years ago. I didn’t give it much thought at the time; it was just another step in the application process.

For a lot of people, it’s the standard for getting into higher education, so it was a nice surprise, listening to the TOEFL Talks podcast recently, to learn how much work and thought go into making it. It’s not as clear cut and dried as it looks from the other side of the exam!

Well, you get 100 bonus points for referencing the podcast. And I’m delighted to hear that preparing for the exam didn’t dampen your interest in the English language

I was struck by the depth of your research on the evolution of English vocabulary as it’s taught to English learners, but the design of the article also elevated the message. How’d you develop this unique craft of visual storytelling?

Thank you for the kind words, John! My background is in graphic design. I studied Visual Communication for my bachelor’s, and at some point I learned to code and started making computational, data-driven work, a lot of it about language.

When I started looking into the topic of this project, I had no idea it would become a visual essay. Then once I did, it took a lot of trial and error. To name one example, the concreteness chart went through more than ten different versions.

For a lot of what I found, the interesting part is the words themselves, more than the counts or proportions. You can’t see that unless you can see the words, so the guiding principle was to show them whenever possible. The tricky part was doing that without putting so much on screen that it turns into noise.

Well, you pulled that off – the final design is a real work of art. On to your research: You dove into the General Service List archives to compare two sets of commonly taught English terms from 1953 (“GSL”) and 2023 (“NGSL”). Before discussing your findings, why did you have interest in this topic?

I’ve always had an interest in language that I could never quite articulate; how it encodes ideas, how it communicates information, how the same word may land differently in different contexts. This inquiry started last December, when I came across the 2015 controversy of the Oxford Junior Dictionary removing several nature-related words. It made a lot of people upset.

I started looking into all the Oxford dictionaries, and then other “curated” word lists, which is how I stumbled onto the GSL and the NGSL. What ultimately drew me in is that, unlike the Oxford Junior Dictionary, these lists aren’t that curated; they are grounded in actual language use. 

They’re largely built from the most frequently used words in the corpora used to make them, with some learner-focused curation. So they’re closer to a record of what people were reading and saying than to a set of editorial decisions made by their authors.

The fact that they’ve already been studied and shown to cover most of everyday English, along with the fact that the NGSL is a modernized version of the GSL, made comparing them unusually compelling.

_{Words removed from the GSL (left) vs. Words added to the NGSL (right), sorted by their concreteness rating. 1 = highly abstract; 5 = highly concrete.}

Over the course of 70 years, you found that this list shed about 600 words and gained more than 1,100. Then you used a linguistics tool to assign each word to 21 categories. This is where things got really interesting.

The largest category across both eras was General & Abstract Terms. But this group grew by about 28% over 70 years, adding words like concept, justify, and perspective. Can you explain why these abstract words grew in prominence?

There are probably many contributing factors, but I think at its core it has to do with the growing connectedness of things. People’s lives became more globalized, more grounded in institutions, more far-reaching. They needed language to coordinate across borders, navigate larger systems, manage bigger responsibilities.

Formal education expanded over this period, and people were exposed to far more written language through mass media and later the internet. Words such as concept, perspective, and justify are part of the vocabulary of explicitly framing, evaluating, and communicating ideas at scale.

That shift from the immediate and the local to the systemic and structural runs through the whole piece, and is the main insight that kept showing up in all the different angles through which I inspected the lists.

Elsewhere, you found that the sharpest drops were in the most tangible categories: Substances & Objects, the Body & the Individual, Food & Farming, Life & Living Things. What social changes might have prompted this shift in language use?

I think it’s the other side of the same shift. Life became less self-contained. Supply chains took over, services replaced tasks people used to do themselves, and the physical substrate of daily life got more remote even for people who had been already living in cities for decades.

You may not need the word wheat if you just buy bread at the store. It’s not that wheat became irrelevant, it’s that our relationship to it changed. As the language usage shifted to accommodate that, some of those highly concrete terms fell out of the core vocabulary while more abstract terms became common enough to enter it.

_{Changes in semantic categories between the GSL (left), and the NGSL (right)}

Abstract words from 1953 like courage, wisdom, and honesty – emotions “grounded in experience” – have ceded ground to words like theoretical, evaluate, and justify, which describe “how things are structured and how processes unfold.”

You also found that “the vocabulary of inner life dropped words for feelings such as patience, pity, despair, and sorrow, and picked up more words for thinking: focus, judgment, investigate, and logic.”

Could it be that these terms are just more immediately useful for those who learn English as a second language?

That's a great question, and one I asked myself, too. My first instinct was the same as yours: maybe this shift is at least partly by design, because words like evaluate and justify are just more useful in professional or academic settings, where a lot of ESL learners would need to operate.

Both the GSL and the NGSL are primarily built from frequency counts, but I've also come across external research* that independently corroborates a part of this pattern. There's a 2012 study that tracked 50 virtue words (honesty, patience, compassion, courage, wisdom...) across the full 20th century in a large corpus of American books and found that 74% of them declined in frequency over that period.

So the disappearance of words such as courage and wisdom can’t be explained only as an editorial judgement call. They were also appearing less often in all American writing. But I don't think it means we stopped experiencing courage or feeling sorrow.

It might mean that as the world we needed to describe got bigger and more complex, those bigger ideas took up more of the conversational space. It became more important to talk about systems, institutions, processes, things happening at a scale beyond our immediate surroundings. The internal, more personal vocabulary just became less central to what people needed to communicate to one another – whether they are native speakers or not.

This shift from experience-based words to language-based abstractions strikes me as a little unnerving. You also mention that concrete and abstract words are processed differently by the brain. Does that have any implications for people learning the language?

I totally get why this feels a bit unnerving. I felt the same way, and it's one of the things I find most interesting about this in the context of language learning. There's this psychological theory called Dual Coding – the idea that concrete words are easier to learn and remember because they get encoded twice in our minds: once as a verbal term, and once as a sensory trace, be it an image, a sound, a tactile experience, etc. This is why concrete words tend to "stick" better.

But lacking a clear sensorial or emotional anchor, abstract words tend to rely more heavily on verbal context and associations. This arguably makes them harder to acquire for learners. And if you buy into that theory, this may also explain something that a lot of people, not just ESL learners, have been feeling: the sense that just getting through the week now involves quite a lot of reasoning, weighing, and deciding.

For a non-native speaker, that added linguistic load could make an already complex environment harder to navigate.

Indeed. Elsewhere, you note that the 2023 list hints at a social world that has “shifted from something you navigate through relationships to something with which you identify through categories” and offers “fewer words for the people directly around you, but more for belonging at a distance.”

Can you provide some examples of this shift from local to global terms?

This might be one of my favorite findings, if I'm allowed to have favorites. The "social" category barely moved, which made it seem like not much had changed, but if you look at the words that got switched, the difference is quite striking.

The 1953 list had more words for navigating direct, one-on-one relationships: companion, confess, humble, generous, polite, loyal... They assume you're dealing with a specific person. The 2023 additions are of a different kind: community, identity, ethnic, gender, narrative, organization... They're more about how we understand people at scale, as groups, or as categories.

I can't say for certain why, but my read is that as more people become connected across larger distances and across different cultures, the language also shifts from navigating individual relationships to thinking about collective identity and belonging.

In a way, you're less often describing one person and more often describing what kind of person they are, or what group, or structure they belong to. This is my own interpretation of the pattern; it's not something that comes directly from the semantic analysis, but I think it's one of those cases where reading the words themselves tells you more than any category count would.

You also found that the volume of adverbs taught to English learners has nearly doubled. In school, I was taught that adverbs should be used very, very sparingly. (See what I did there?) Any thoughts on the Rise of the Adverb?

There's a relatively clean connection between adverb growth and everything else we've been talking about. Many abstract words are inherently open-ended in a way that concrete words aren't. They tend to need another word alongside them to specify degree, or qualify certainty, or mark how strongly you mean it. Almost all of the adverbs that were added are doing that kind of work; they're calibration words.

I’d also speculate this has to do with the broader cultural shift we’ve seen elsewhere. It feels like the world now asks us to be more exact about what we're saying, more aware of uncertainty, more careful about how strongly we're asserting something. When we're navigating institutions, reading the news, writing an email to someone we've never met (like the one you wrote to me), we don't just say things are "good" or "bad". We say they're "really good," or "probably fine," or "significantly worse."

There's a kind of epistemological self-consciousness baked into the new adverbs. We're constantly flagging how certain we are, in which direction, and to what extent.

_{All the adverbs that were added to the 2023 NGSL (116 in total)}

That’s a great point. And, again, perhaps another subtle point of stress for non-native English speakers, who often have to navigate unfamiliar norms related to these “certainty” cues.

Absolutely. (See what I did there?) As a non-native speaker myself, I’ve often had to map those cues out explicitly rather than just absorb them, which might be why I noticed this pattern in the data at all.

I have to confess, I left your article feeling ambivalent about the evolution of the world that the modern vocabulary list represents. On one hand, the 2023 list evokes a world that is “not so much about what’s within arm’s reach, but more about the larger world we navigate.” This seems great!

But you also note: “Bread survived both lists. Flour, wheat, harvest and bake didn’t. The word for what sustains us remained essential, while the words for how we’d make it weren’t.” This strikes me as a real loss.

Am I being too sentimental? Or did you also feel a sense of cultural loss when you looked at your final results?

I feel it, too. There's something a little heartbreaking seeing stove and apple and wool go, and thinking about what their disappearance reflects. There's a kind of world encoded in these words; small enough to feel with our hands, close enough to name from where we're standing. But I tried not to write this piece as a lament, because I don't think that's the whole story.

We didn't just get computer and server and blog, we also got democracy, and independence, and perspective - they are not trivial replacements. They are the vocabulary of a world where our lives are entangled with more. 

We eat food we didn't grow, governed by regulations we didn't write, processed through supply chains we can't fully trace. That's dependency, yes. But it's also civilization. It's the whole project of people building things together no one person could build alone.

And I find some comfort in the fact that we notice the loss at all. That we look at a word like tide leaving the list and feel something. I think it means we haven't stopped caring, even if other things occupy more of the language now.

I like this framing: It’s a culture in flux, not one suffering a loss. One final question. And it’s deeply unfair. How do you anticipate this list changing when the Jasmine Nackash AI Bot looks at it again in 2093?

It's only fair! I've been wondering about this a lot. In 2093, when our consciousnesses are all fed into sophisticated Artificial Intelligences running in the ether, I suspect language will have changed even more than it did over the past 70 years, simply because change itself seems to be accelerating. The meaning of words will shift just as much as the words themselves.

The word cloud comes to mind. Awful used to mean worthy of awe, the kind you feel in a cathedral. Manufacture comes from the Latin for made by hand. And artificial used to describe things made by human skill, rather than by nature. On that note, AI language models are, in a way, statistical portraits of language, not unlike the frequency-based logic behind the word lists. But they also feed back into the language itself.

If we increasingly let models write for us, and those models output statistically likely language, and that language gets used more often and then scraped back into the next generation of training data – I do wonder if we'll be seeing a kind of flattening. A reduction in variance. Everything drifting toward a statistical center of gravity.

Or perhaps we'll see the opposite happening in pockets: specialized vocabularies spread to more people, terms that used to take years of training to use correctly get picked up and normalized. Hard to say. 

I think it will be a combination of sorts. The mid-20th century wasn't all farms and homemade bread; we also had TV dinners and assembly lines. Nowadays there's a whole counter-movement around local produce, handmade goods, and caring about how products are being made and by whom.  

We're biological creatures, and as long as that's true, the physical world still matters. But it makes me wonder whether the vocabulary of the self will become less experiential and more analytical. We increasingly measure, track, diagnose, and optimize our bodies and minds, using increasingly specialized language to describe them.

I'd guess we'll know far more about our own inner workings in 70 years, so some of that specialized vocabulary, words that right now live in medical or scientific journals, might become everyday language. And climate change will almost certainly force a new kind of relationship with the physical environment, less as backdrop, more as something we actively manage and negotiate with.

We’ll probably be seeing words such as displacement, adaptation, resilience, and others we might have to invent or borrow from other languages (or eras). The 2023 list captured a world entangled more than before. I think the 2093 list might capture a world reckoning with the cost of that.

Jasmine, this has been a profoundly enlightening walk through 70 years of linguistic – and social – change. This deserves a companion piece that analyzes how the language of TOEFL has evolved since its inception in 1964. But we’ll save that for another day!

Note for the reader: If you’d like to dive into Jasmine’s full dataset, you can access it via this publicly available spreadsheet. And again, if you haven’t read the piece itself, you really need to see the full design. It’s a work of art.

*Kesebir, P. & Kesebir, S. (2012). The cultural salience of moral character and virtue declined in twentieth century America. Journal of Positive Psychology.