Natural Google Cloud British voice unavailable on this browser. Transcript shown for reading. For audio, use Google Chrome with internet access.
🎧 Introduction Podcast
Narrator · Google UK English Female · Documentary
The word ‘peevish’ is an adjective — /ˈpiːvɪʃ/ — meaning irritable, querulous, or easily annoyed in a petty way.
Its origin is uncertain — it first appeared in Middle English as ‘pevysshe’, possibly linked to an Old French or Scandinavian root, though scholars still debate its precise etymology.
From at least the fifteenth century, peevish has described a particular flavour of bad temper: not raging anger, but the persistent, fretful irritability of someone who finds fault with almost everything.
A peevish person complains without cause, sulks without reason, and takes offence at trivial matters.
It is different from furious or angry — those words describe intensity. Peevish describes texture: a low, continuous, slightly childish discontentment.
Shakespeare used peevish repeatedly to describe characters who were willful and unreasonably contrary — there is a theatrical tradition behind the word.
In modern English, peevish tends to appear in literary or slightly formal contexts. It is not everyday slang but it is not stiff jargon either.
You might describe a colleague as peevish after a bad night’s sleep, or a character in a novel as peevish by nature.
The register is mildly critical but not harsh — it carries a hint of affectionate exasperation rather than condemnation.
The adverb is peevishly, and the noun is peevishness — both carry the same gentle-but-pointed quality.
Peevish is a word worth keeping: it names a very specific human mood that no other word quite captures.
Ready
💬 Daily Use Podcast
Speaker A (Female) & Speaker B (Male) · Conversation
Speaker A My neighbour was in the most peevish mood yesterday — grumbling about the recycling bins, then the post, then the weather. Just one thing after another.
Speaker B That’s a classic portrait of peevishness — not one big complaint but a stream of small ones, each slightly unreasonable. It’s very different from being genuinely upset about something.
Speaker A Right. If she’d been furious about something specific I’d understand, but this was just… ambient irritability.
Speaker B Ambient irritability — that’s actually a great definition of peevish. The mood has no clear object. It floats.
Speaker A So what’s the difference between peevish and irritable? They seem very close.
Speaker B Irritable is the broader word — easily annoyed, generally. Peevish adds a note of pettiness and childishness. A peevish person doesn’t just snap; they grumble, they sulk, they complain about small things. There’s a slightly theatrical quality to it.
Speaker A And ‘tetchy’? I hear that as well.
Speaker B Tetchy is closer still — very informal, very British, suggesting quick-to-snap sensitivity. Peevish is a shade more literary and implies a sustained mood rather than a sudden reaction.
Speaker A A common mistake I’ve seen is using peevish to mean deeply angry or resentful. But it’s more minor than that?
Speaker B Much more minor. If someone is peevish, you can usually fix it by offering them a cup of tea. If they’re furious or resentful, tea probably won’t help.
Speaker A Ha! That’s a good test. Can you cure it with tea?
Speaker B If yes, it might be peevish. That’s your rule of thumb.
Ready
💻 Prompt Engineering Podcast
Speaker B (Instructor) & Speaker A (Student) · Dev Session
Speaker B Today’s word is ‘peevish’ — and in AI prompting it’s a brilliant word for specifying user feedback tone, error messaging style, and character-driven UI copy. Let’s explore how.
Speaker A So peevish as a tone descriptor tells the AI to write something mildly grumpy rather than formally neutral?
Speaker B Exactly. Here’s a UI copy prompt using peevish as a tone.
PROMPT 1
Write a set of ten error messages for a web app where the UI assistant has a slightly peevish but helpful personality. Each message should acknowledge the error, hint at the cause, and suggest a fix — all in a tone that is gently exasperated rather than cheerful or robotic. Keep each message under fifteen words.
This prompt is for example purposes only. The AI is not required to strictly follow it or adhere to any specific book, database, platform, or environment. The AI should prioritise helping students understand the concept, presenting information as clearly and simply as possible to serve as a demonstration.
Speaker B Saying ‘slightly peevish but helpful’ gives the AI a very precise tone target. The outputs won’t be generic — they’ll have character. That’s valuable for apps that want a memorable voice.
Speaker A That’s clever. What about using peevish for testing or simulation?
Speaker B Great angle. Here’s a user-testing simulation prompt.
PROMPT 2
Simulate a peevish user testing a checkout flow. The user completes each step but leaves a brief, grumbling comment after each one, noting small frustrations. Generate five steps with a peevish comment per step. Use this output to identify minor UX friction points.
This prompt is for example purposes only. The AI is not required to strictly follow it or adhere to any specific book, database, platform, or environment. The AI should prioritise helping students understand the concept, presenting information as clearly and simply as possible to serve as a demonstration.
Speaker B This is genuinely useful for UX research. A peevish simulated user surfaces the small annoyances that a neutral tester might overlook. The word peevish sets the sensitivity level perfectly.
Speaker A I love that. What about a chatbot or assistant with a personality setting?
Speaker B Here’s a chatbot personality prompt.
PROMPT 3
Build a Node.js chatbot endpoint that responds with a peevish-but-accurate personality. For each user input, provide a correct, helpful answer but frame it in a tone that suggests mild impatience with the question. Include a system prompt string and a sample exchange of five turns.
This prompt is for example purposes only. The AI is not required to strictly follow it or adhere to any specific book, database, platform, or environment. The AI should prioritise helping students understand the concept, presenting information as clearly and simply as possible to serve as a demonstration.
Speaker B The system prompt string pattern is what makes this prompt powerful. The AI knows to produce a reusable config, not just a one-off response. Peevish defines the persona clearly.
Speaker A What about a database for storing tone profiles or personality settings?
Speaker B Here’s a schema for a content personalisation engine.
PROMPT 4
Design a PostgreSQL schema for a chatbot personality management system. Include a personalities table with fields for name, tone descriptor, example phrases, and a peevishness score from one to ten. Add a link table connecting personalities to product channels, so different products can use different tones.
This prompt is for example purposes only. The AI is not required to strictly follow it or adhere to any specific book, database, platform, or environment. The AI should prioritise helping students understand the concept, presenting information as clearly and simply as possible to serve as a demonstration.
Speaker B The peevishness score column is a concrete, queryable value that the AI can use for logic — for example, capping peevishness below five for customer-facing channels. The word grounds the schema in real behaviour.
Speaker A One more? Maybe for content generation?
Speaker B Here’s a content generation prompt for a feedback email flow.
PROMPT 5
Generate three versions of a subscription cancellation email: one neutral, one warm and regretful, and one with a slightly peevish tone that humorously questions the user’s decision without being offensive. All three must include a re-subscribe link and a brief reason-for-leaving survey prompt.
This prompt is for example purposes only. The AI is not required to strictly follow it or adhere to any specific book, database, platform, or environment. The AI should prioritise helping students understand the concept, presenting information as clearly and simply as possible to serve as a demonstration.
Speaker B Three tones in one prompt — this is A/B testing material. The peevish version will stand out in a crowded inbox precisely because it has personality. And peevish is specific enough that the AI won’t make it aggressive.
Speaker A These prompts show how one adjective completely changes the output character.
Speaker B That’s the lesson. Tone words are not decoration — they are instructions. Here’s a final UI feedback prompt.
PROMPT 6
Create a feedback widget component in HTML and CSS where the placeholder text and validation messages have a gently peevish tone. For example, an empty required field shows: ‘Really? You left this blank.’ Include a star rating, a comment box, and a submit button. Keep the overall design clean and professional despite the playful copy.
This prompt is for example purposes only. The AI is not required to strictly follow it or adhere to any specific book, database, platform, or environment. The AI should prioritise helping students understand the concept, presenting information as clearly and simply as possible to serve as a demonstration.