top of page

From 'Meh' to 'Magniloquent': Building Vocabulary with Gen AI & Prompt Engineering

An open dictionary page.
Photo by Romain Vignes on Unsplash

I often encounter advice around the theme of how to communicate better that equates to “simplifying” language use to make it easier for readers or listeners to understand.

I do believe that we should take our audience into account as we craft our communication, and I do believe that certain kinds of texts – for example, legal documents – should absolutely be written for maximal accessibility. But I also believe that there are times when “simplifying” our language use not only makes our meanings less precise, but also robs our audiences – both intended and otherwise – of the opportunity to learn and deepen their own literacy. In short, I believe that some of the onus is on readers and listeners in comprehending authors and speakers.

And the way we accomplish this, I believe, is through focused lexical upskilling.

Focused lexical upskilling?

Yup. As in, studying vocabulary.

A Case for Studying Vocabulary

You may be thinking, Studying vocabulary? Didn’t we leave that sort of thing behind in high school English?

No. No, we did not.

Or, rather, we shouldn’t have. Consider how often you encounter unfamiliar words, whether in speech or writing. Or words that you feel like you kinda know, but in reality, would struggle to define, or use in an actual sentence. How often do you stop and look up at that word in the dictionary before moving on? How often do you record that definition somewhere and revisit (study!) it so that you don’t promptly forget it?

Anecdotally, I would guess that very few people do the latter, and again, I would argue that we absolutely should.

Think of it this way: Everybody is a learner of their native language. Your native language is the one you learned as a young child (some people have multiple native languages). We don’t stop learning our native language once we reach adulthood. Every day, we read and hear words that we’re either completely unfamiliar with, or are unsure how to deploy in the contexts in which we regularly communicate.

And my strongly held belief is that when it’s possible, we should look those words up, attempt to understand them deeply, and work toward incorporating them into our working vocabularies. When I read an article from The New Yorker, or The Atlantic, I encounter at least one word every paragraph for which I lack either comprehension, or facility, or both. And until I look up that word, my understanding of the writer’s ideas is compromised. So I look up the word in the moment to enable my understanding of the writer’s ideas in that moment, and I record the word and continue to study it so that when I encounter it in the wild again, that meaning is unlocked for me without my having to look it up again. Do this again and again and again, and your literacy – your fluency in your native language – continues to grow, unlocking new meanings for you, and increasing your ease and facility in interacting with the world of ideas.

A Method for Acquiring New Vocabulary

So, yeah – we’re studying vocabulary. Now let’s talk about how we might bolster our lexical knowledge using modern tools.

There’s a slew of research that supports the idea that we need to repeatedly engage with new information in order to incorporate it into our long-term memory. When it comes to vocabulary learning, we typically start with declarative knowledge - knowing what the word is and what it means. With time and practice - using the word in various contexts, seeing it used in texts, hearing it in conversation - we develop procedural knowledge on how and when to use the word appropriately. Both forms of memory are vital for comprehensive language learning and usage.

In order to move new words into procedural knowledge – in other words, to actually incorporate new words into your working vocabulary, we must consciously and deliberately engage with them. We must practice using these new words.

After experimenting with the OpenAI Playground, I believe that this is an excellent tool for making vocabulary study effective, efficient and even fun.

Here’s a prompt that I engineered to get you started. You would place this in the OpenAI Playground SYSTEM field, or as a ChatGPT prompt:

You are an expert on the English lexicon. Hell, maybe you're even a lexicographer.

For the provided word, please tell me its:
- part of speech
- meaning
- lexical derivatives

Please also give me two example sentences for each derivative. These example sentences should relate to the context provided.

Example input:

WORD: sycophant

Example output:

WORD: Sycophant
POS: Noun
MEANING: A person who acts obsequiously toward someone important in order to gain advantage.
DERIVATIVES: sycophantic (adjective), sycophantically (adverb)

For 'sycophant':

1. In the work environment, it's critical not to be perceived as a sycophant lest your coworkers lose respect for you.
2. Marie is such a sycophant at work; she's always flattering the boss in hopes of getting a promotion.

For 'sycophantic':

1. His sycophantic behavior at work towards his superiors raised many eyebrows amongst his colleagues.
2. The meeting was dominated by John's sycophantic praises of our boss which didn't go down well with the team.

For 'sycophantically':

1. At work, Charles behaves very differently when management is around—always agreeing and nodding almost too energetically; he acts quite sycophantically.
2. Gina has been acting more and more sycophantically every day now that annual performance reviews are nearing—it doesn't go unnoticed among her peers though.enjoys near ubiquity in certain industry sectors.

Then, in the USER prompt field, input the word and context for which you’re currently seeking a definition and example sentences.

In the example below, I’m looking for the meaning of “vivify” with example sentences relating to an educational context.

WORD: vivify

CONTEXT: education

Press “Submit”, and you’ll get the definition, derivatives, and example sentences for your new word:

A view of the OpenAI playground system prompt and output.

And it’s as easy as that! You can continue to engage with the AI if you’re looking for additional example sentences in different contexts, if you’re interested in learning more about the etymology of the word, or if you’re curious about synonyms and collocations (words that are frequently encountered alongside this word).

These exploratory steps are just the beginning in any earnest attempt at vocabulary acquisition, of course. As discussed earlier, in order to make these words stick, you should record them somewhere where you can practice them (I use Brainscape digital flashcards), and when you’re ready, come up with example sentences that reflect things that you might actually say in your actual life. Coming up with authentic example sentences will increase the likelihood that you’ll find occasion to use them, which will help you add them to your procedural knowledge.

Warning & Limitations

There is, of course, a limitation with Generative AI that we must address: At this moment in time, many Generative AIs (and the Large Language Models that power them) are notorious for hallucinating, or in other words, making stuff up. So, how can we be sure that Chat GPT isn’t contributing to the degradation of our working vocabulary and literacy by coming up with fake definitions or word derivations that we don’t actually use?

Truth be told, in my experience using Gen-AI for lexical upskilling, sometimes stuff gets made up. Primarily around suggesting word derivatives that aren’t used conventionally.

But here’s something fun: English is known as a productive language, meaning that our robust affixation system allows for the constant creation of new or unconventional forms. So while converse is the conventional verb form of the noun conversation, folks will understand what you mean if you say conversate. That, and when it comes to language use, I’m a descriptivist – not a prescriptivist – meaning I don’t think language use must abide by rules. In a descriptivist approach to language, anything we utter that results in successful communication is valid! That said, just know that ChatGPT – just like us humans – will sometimes provide you with a derivative or an example sentence that is not “acceptable” in Standard American English. If you’re questioning whether the provided derivatives are valid, you can quickly and easily look them up in any online dictionary.

Alright, then. Go forth, learn some new words, increase your literacy, and unlock your access to meaning and ideas!


bottom of page