Coregraterold Authathlike Licile Spokshariation Powelceously

February 10, 2023 12:04 PM

This page generates nonsense words based on a frequency list of phonemes as they occur in legitimate English words. Generated words are checked against a list of approximately 500,000 English words. An actual word may slip through occasionally but it should mostly generate pronounceable gibberish.
posted by Etrigan (66 comments total)

How much for an oz. of becoweed?
posted by Liquidwolf at 12:09 PM on February 10, 2023 [6 favorites]

This scratches an interesting itch for me. My son has recently been coaching his peers in English proficiency tests, and one of the main exercises is "select the fake word" from a list of five.

All of them sounded so plausible! "How did they make them up?" Must have been some similar methodology. I wonder what fake Italian, Japanese, etc., words would sound like.

"How does this actually test vocabulary?" Still a little fuzzy for me.
posted by Meatbomb at 12:10 PM on February 10, 2023 [1 favorite]

Inbreathiate?

This is great, especially because I'm working on a sci-fi parody story in which I'd love to use a ton of made up but realistic words!
posted by MiraK at 12:17 PM on February 10, 2023 [4 favorites]

Fake English words previously on the Blue. "Prisencolinensinainciusol" is very catchy!
posted by obfuscation at 12:20 PM on February 10, 2023 [9 favorites]

Back in the 90s, when phone calls were expensive and so email was still the best way for people who were apart to keep in touch, I had a long-distance girlfriend who typed as though the world was ending in three minutes and never touched a backspace key. These lists read like her emails did.
posted by The Card Cheat at 12:21 PM on February 10, 2023 [7 favorites]

Metafilter: should mostly generate pronounceable gibberish
posted by genpfault at 12:21 PM on February 10, 2023 [15 favorites]

Thankyou for this soporkworkly post. I found it confoused my ratiatorum, while not being diputeous.
posted by inflatablekiwi at 12:22 PM on February 10, 2023 [3 favorites]

This reminds me of the Swamp Thing issue "Pog" in which Alan Moore paid tribute to Walt Kelly's Pogo by having the suspiciously-familiar-looking aliens speaking in neologisms that were mostly portmanteaus.
posted by Halloween Jack at 12:24 PM on February 10, 2023 [1 favorite]

I don't believe in fake words; they are just words that haven't found their bgoriegisnash, yet.
posted by NoThisIsPatrick at 12:24 PM on February 10, 2023 [14 favorites]

I don't know. Most of these words seem pretty cromulent.
posted by brundlefly at 12:25 PM on February 10, 2023 [8 favorites]

Also, Blackadder congratulates Dr Johnson for finally completing his dictionary.
posted by inflatablekiwi at 12:32 PM on February 10, 2023 [7 favorites]

Remember, when you're applying disalves to your crompts, always keep your plomers fluttacted. That's just common applanism.
posted by mittens at 12:34 PM on February 10, 2023 [3 favorites]

"caterrupt"

I'm pretty sure I've seen this happen when our cats unexpectedly ran into each other under the coffee table.
posted by Hairy Lobster at 12:37 PM on February 10, 2023 [13 favorites]

Also functions as an Amazon brand name generator.
posted by jocelmeow at 12:53 PM on February 10, 2023 [12 favorites]

In the proficiency testing context, in my experience the point of the fake words is typically to provide a reality check for self-assessments of receptive vocabulary size. (The question of what it means to "know" a word is immeasurably complex, but "can recognize it as a word" is a prerequisite to most other levels of knowledge, so it can give you a general idea of where you stand.)

Ideally you could just go down a list of words from different frequency bands and check off the ones you don't know, but even in the absence of cheating you need a way to be sure that you aren't fooling yourself, so that's where the fake words come in. (But normally you would want a fairly small number of distractors for this purpose, not like 80%, so maybe those tests are measuring something else?)
posted by Not A Thing at 12:53 PM on February 10, 2023 [1 favorite]

Back in the 90’s, a friend and I were going to a lot of open mic poetry readings. After a number of these, I wrote a little program that used English phonemes to generate nonsense words of various lengths. Not similar to English but pronounceable. I generated about a third of a page worth, printed it out and went to a reading and got on the list. I went up and gave a dramatic reading of this nonsense, cold by the way, and got a rousing bit of applause. San Francisco in the 90’s was far more tolerant than Zurich and Paris in the teens, when the Dadaists did something similar and either got pelted with rotten food or caused riots.
posted by njohnson23 at 12:55 PM on February 10, 2023 [5 favorites]

I'd have loved these these in one of my college papers on modern art.

Painted in the casailes style, but drawing heavily on Cubillo Ulates tribal motifs, the widearchermit presentation clearly is a critique on Ediblersaked philosophy and it's imsymmon followers.
posted by cccorlew at 12:56 PM on February 10, 2023 [5 favorites]

/- Also functions as an Amazon brand name generator.

I was immediately thinking similar, but instead that this could be a great way to generate a new plausible short company name that could buy the respective unregistered .com domains for a new business.
posted by many-things at 12:59 PM on February 10, 2023 [3 favorites]

Or new pharmaceuticals:

Don't take Climplashlest if you are allergic to Climplashlest or its ingredients, ask your doctor if Climplashlest is right for you! If you can't afford Climplashlest, Warbia-Ankriss may be able to help.
posted by AzraelBrown at 1:01 PM on February 10, 2023 [7 favorites]

Mertafilter: poining mentrate cremakenes since 1999
posted by salishsea at 1:02 PM on February 10, 2023 [6 favorites]

AzraelBrown, exactly. This is the ideal pharmaceuticals name generator.
posted by bz at 1:02 PM on February 10, 2023 [2 favorites]

Also, I asked ChatGPT to make up some definitions for some these words and here is what I got:

Here are definitions for the imaginary words you provided:

Poining - a feeling of intense longing for something that is out of reach.
Mentrate - the act of thoughtfully considering and evaluating one's mental state.
Cremakenes - the state of being pleasantly surprised and taken aback.
Routinkle - a repeated pattern or habit in behavior.
Fertift - the quality of being full of life and energy.
Penancied - the act of regretting past actions and making amends for them.
Incollinates - the process of bringing disparate elements together into a cohesive whole.
Weltonale - a harmonious and well-balanced emotional state.
Baskeps - the feeling of contentment and satisfaction derived from achieving one's goals.
Odluckworr - a powerful and mysterious force that determines one's fate or destiny.
posted by salishsea at 1:06 PM on February 10, 2023 [16 favorites]

"Prisencolinensinainciusol" is very catchy!

All right!
posted by tclark at 1:07 PM on February 10, 2023 [6 favorites]

Doonn't nooo y theees ess pursably kneedid?
posted by sammyo at 1:24 PM on February 10, 2023

Also, I asked ChatGPT to make up some definitions for some these words and here is what I got:

As much as I dislike the concept of ChatGPT, these definitions are cromulent in a way I wasn't expecting. It must have some knowledge of word roots, either through direct exposure or through inference through language exposure.
posted by hippybear at 1:27 PM on February 10, 2023 [5 favorites]

I feel at home
posted by clavdivs at 1:28 PM on February 10, 2023 [2 favorites]

Ok now make an ipsum lorem generator with this.
posted by oddman at 1:34 PM on February 10, 2023 [1 favorite]

“Each of these could be improved by adding the suffixes -onius, ainous, or anous.” -Dr Ronald Chevalier
posted by Abehammerb Lincoln at 1:39 PM on February 10, 2023

weird when i try, it just generates random Scots words
posted by logicpunk at 1:40 PM on February 10, 2023 [1 favorite]

It must have some knowledge of word roots, either through direct exposure or through inference through language exposure.

The tokenizer for the original transformer used subwords that probably helped the system learn root words and conjugations. But if you try that output on the GPT tokenizer, you can see it tokenizes real words as a single entity.
posted by pwnguin at 1:46 PM on February 10, 2023

a very revectuative post, assuming you're not feeling byrious; I hope it doesn't get munisticized, or turned into poemenace...
posted by terretu at 1:46 PM on February 10, 2023

I feel like I could use this to great advantage in a tabletop RPG,
when the players are trying to communicate with someone but don't understand their language.
posted by chromecow at 1:49 PM on February 10, 2023 [3 favorites]

Twas spristly, and the darleyed snowne,
Did bridate and conson in the fannes;
All rupted were the ummendered,
And the daunct reights orassed

Beware the fibnesses, my proons!
The flaxle that bite, the biangs that catch!
Beware the Auslaught bird, and shun
the semiguist boundablaniles

Ok, I don't have the rhymes down yet...maybe tonight...
posted by foonly at 1:49 PM on February 10, 2023 [7 favorites]

GPT is quite good at creating novel words and definitions, although it sometimes repurposes an existing word. Example: "Generate a nonsense glossary of Scottish-sounding words about nihilism."

• Gloamshrift: A belief that nothing matters.
• Skirlin: The feeling of emptiness when one realizes the futility of all things.
• Maunsman: A person who believes that all life is pointless.
• Glaikit: An attitude of apathy towards life and its events.
• Mootie: A person who has no faith in anything and has no hope for the future.

posted by credulous at 1:50 PM on February 10, 2023 [6 favorites]

turned into poemenace

Was that the fate of the family living in the House Of Leaves?
posted by hippybear at 1:50 PM on February 10, 2023

The gostak distims the doshes.
posted by dywypi at 1:52 PM on February 10, 2023 [2 favorites]

I put a random word generator online a couple decades ago. It works by analyzing supplied text, rather than a standard corpus. There's a database of word lists that people can try (or add to). I don't think I've looked at the site for ten years or so, but it's still running.
posted by seanmpuckett at 1:53 PM on February 10, 2023

I don't think I've looked at the site for ten years or so, but it's still running.

Now I'm wondering how many more words it has in its vocabulary (if that's how it works) than it had when you first launched it. A couple of decades of further input by users should yield something significant.
posted by hippybear at 1:55 PM on February 10, 2023

I like the idea, but as someone who's tried to explain English spelling, I feel like the program is missing a lot— mostly the fact that it mixes English and Latinate bits. E.g.:

facateeighes - starts out Latinate, but gh only appears in native words
dolliked - dollicked would be OK, or doll-liked
lenckeept - generally can't have a long vowel before two final consonants; and ck doesn't work like that
oundipitent - this one mixes a Frenchy beginning with a Latinate ending

But this is par for the course for random word generators... you can just use the good ones and ignore the rest.
posted by zompist at 1:55 PM on February 10, 2023 [2 favorites]

note: it's furking hilarious how fast my site is, compared to most modern garbage. It's all static html, perl cgi scripts, and a mysql backend. mein gott in himmel it rips.
posted by seanmpuckett at 1:56 PM on February 10, 2023 [2 favorites]

I have a PhD in expermalanties and I still can't find a job.
posted by Splunge at 1:57 PM on February 10, 2023

I have a PhD in expermalanties and I still can't find a job.

The field often requires extensive experience in permalanties before they allow you to move above that.
posted by hippybear at 1:58 PM on February 10, 2023 [2 favorites]

Smetchory!
posted by crazy_yeti at 2:00 PM on February 10, 2023

Sockpuppet names abound! My libraggloble runneth over!
posted by Greg_Ace at 2:11 PM on February 10, 2023

Immediately thought of the Library of Babel website.
posted by msbrauer at 2:12 PM on February 10, 2023

Poining - a feeling of intense longing for something that is out of reach.

Example: He's poining for the fjords!

Weltonale - a harmonious and well-balanced emotional state.

"Have a Welton Ale, it'll leave you feeling weltonale!"
posted by Greg_Ace at 2:14 PM on February 10, 2023

For those who enjoy fake-word-discrimination quizzes, there's one available at vocabulary.ugent.be.
posted by What is E. T. short for? at 2:18 PM on February 10, 2023 [1 favorite]

Weltonale - a harmonious and well-balanced emotional state.

"Have a Welton Ale, it'll leave you feeling weltonale!"

It was probably 5 years after I came home from living in Germany for a year that I realized that the ice cream brand Langnese was the same as the Good Humor brand in the US. That's a weird language blind spot.
posted by hippybear at 2:22 PM on February 10, 2023

Perrotchril
Gishorte
Decianchar
Surasping

These are all Culture citizens
posted by GCU Sweet and Full of Grace at 2:22 PM on February 10, 2023 [2 favorites]

This reminds me of the Dadaism writing trick.

Get a chair to nap in with a small table next to the chair. On the table you put a pen and an open notebook. In your hand you place a steel spoon.

And before you finally settle back into your nap you place a pie tin beneath the hand holding the spoon so that when you finally drift off for your nap eventually you will loosen your grasp on that spoon and it will fall and hit the pie tin and wake you up.

Then you immediately grab the pen and write as fast as you can everything that you remember from your dream into the notebook.

It is like we are trying to teach computers that trick.
posted by NoThisIsPatrick at 2:40 PM on February 10, 2023 [3 favorites]

For those who enjoy fake-word-discrimination quizzes, there's one available at vocabulary.ugent.be.

Ooooh thank you, I did enjoy that!
posted by an octopus IRL at 2:49 PM on February 10, 2023 [1 favorite]

This reminds me of the Dadaism writing trick.

Alternately, a hand hanging over the arm of a chair holding a marble.

I guess this method has the hand closer to the pen and paper, but the thing is the waking up.
posted by hippybear at 2:50 PM on February 10, 2023 [1 favorite]

Hasn’t happened in years, but sometimes as I fell asleep, instead of the usual hypnic image, I would see a scrolling MetaFilter or Askme page, but I couldn’t quite make out the words, and when I tried it would wake me back up.

One night I discovered I could zoom in or out by willing it, and the words all turned out to be like these, and it was still hard to catch them because they changed while I was looking at them. Infralucubrationism' is one that I remember.
posted by jamjam at 2:50 PM on February 10, 2023 [2 favorites]

Reading text in dream or trip states (text that is manifested and not real that you're trying to read) is always an interesting thing. It's a standard thing (trope?) that it shifts as you read it.
posted by hippybear at 2:52 PM on February 10, 2023 [3 favorites]

"caterrupt"

I agree that this needs to become an actual word and I'm going to start looking for opportunities to use it
posted by Jacqueline at 3:00 PM on February 10, 2023 [2 favorites]

I agree that this needs to become an actual word and I'm going to start looking for opportunities to use it

I experienced a caterrupt earlier today, but fortunately I got newspaper down in front before it happened so I didn't have to soak feline vomit out of the carpet.
posted by hippybear at 3:18 PM on February 10, 2023 [1 favorite]

Poining - a feeling of intense longing for something that is out of reach.

Example: He's poining for the fjords!

I noticed that about a number of the ChatGPT explanations, that it was essentially interpreting them as misspellings of ordinary English words.
posted by atoxyl at 3:31 PM on February 10, 2023 [1 favorite]

Don't take Climplashlest if you are allergic to Climplashlest or its ingredients, ask your doctor if Climplashlest is right for you! If you can't afford Climplashlest, Warbia-Ankriss may be able to help.

Stop taking Climplashlest immediately if you experience a downplembaffer lasting longer than four hours.
posted by dephlogisticated at 3:39 PM on February 10, 2023 [2 favorites]

My sleep has often been caterrupted of late, when Maxwell decides we should wake up when he does.
posted by mollweide at 4:40 PM on February 10, 2023 [1 favorite]

This thread absolutely requires the input of Stanley Unwin.
posted by How much is that froggie in the window at 4:59 PM on February 10, 2023 [1 favorite]

Old and busted: Correct Horse Battery Staple

New hotness: loudenting brairbivord ausent bunize
posted by ChurchHatesTucker at 6:25 PM on February 10, 2023 [2 favorites]

I feel like I could use this to great advantage in a tabletop RPG,
when the players are trying to communicate with someone but don't understand their language.

I have done this in TTRPGs in the past, without such a list or website. Knowing more languages than any of the players (even if none of them fluently) is handy for evoking this sort of situation. I recall the party crossing a mountain range and encountering a civilization of which they had no experience, so I confronted them with lots of ungrammatical Finnish with a smattering of Arabic.

Routinkle - a repeated pattern or habit in behavior.

Surely a routinkle should be the pro forma visit one makes to the washroom before a car voyage, even if a brief internal assessment suggests there is no pressing need (also referred to as a “safety pee.”)
posted by ricochet biscuit at 7:09 PM on February 10, 2023 [3 favorites]

How English sounds to non-English speakers
posted by ricochet biscuit at 7:12 PM on February 10, 2023

MetaFilter: lots of ungrammatical Finnish with a smattering of Arabic.
posted by hippybear at 7:14 PM on February 10, 2023 [3 favorites]

Love this; tickles the same bit of my brain as the nonsense-adjacent text generated by madlibbing 'AI', but without the obvious suspooperies that they imply.
posted by sarble at 4:53 AM on February 11, 2023

"Do you want ants? Because that's how you get ants"
posted by sebastienbailard at 8:37 PM on February 11, 2023