I presume everybody knows about the absurdly addictive Wordle by now (and its many offshoots like Dordle, which you can play as often as you want); I’ve been playing them without feeling any need to post about them, language-based though they are. But I’m frustrated enough by Semantle that I’m passing it along to frustrate you as well:
Each guess must be a word (of any length) or short phrase. The game will tell you how semantically similar it thinks your word is to the secret word. Unlike that other word game, it’s not about the spelling; it’s about the meaning. The similarity value comes from Word2vec. The highest possible similarity is 100 (indicating that the words are identical and you have won). By “semantically similar”, I mean, roughly “used in the context of similar words, in a database of news articles.” […]
The “Getting close” indicator tells you how close you are –if your word is one of the 1,000 nearest normal words to the target word, the rank will be given (1000 is the target word itself). If your word is not one of the nearest 1000, you’re “cold”. (By “normal” words”, I mean non-capitalized words that appears in a very large English word list; there are lots of capitalized, misspelled, or obscure words that might be close but that won’t get a ranking. Those get marked with “????”).
You will need more than six guesses. You will probably need dozens of guesses. There’s a new word every day, where a day starts at midnight UTC or 19:00 your time. Yesterday’s word was “consume”.
I’ve already made 48 guesses on today’s word and have gotten no closer than 20.23 (today the nearest word has a similarity of 74.10, the tenth-nearest has a similarity of 45.08 and the one thousandth nearest word has a similarity of 20.28, so my best guess is damnably close to the top 1000); I don’t seem to have a good instinct for how to approach the semantic center. May you fare better!
I don’t seem to have a good instinct for how to approach the semantic center.
I have the same experience with so-called ‘quick’ crosswords. (I suspect my vocab is too abstruse for them.) I can’t think of a similar-meaning word that fits. Then I see the answer, and it’s not semantically close at all IMO.
Give me a cryptic crossword any day — usually I can do them quicker.
https://xkcd.com/2576/
Give me a cryptic crossword any day
Me too. I can’t do “quick” crosswords, because you don’t know if you’ve got the right answer until you look at the crossers and realise that you haven’t.
After a strangely tortuous journey I got the secret word on guess number 84. I started finding some words for which I could see a broad connection in meaning and/or context, but many guesses that I thought would get me closer sent me way off track instead. Getting the right word was pretty much a lucky guess.
Alerted by LH’s post I tried it today (19 February, in Victoria, Australia). Annoying. Took me 20 guesses, but that’s without reading the introductory material.
As for Wordle, the rules are not stated at the site. The three “examples” are not complete examples, and the glosses on them are misleading and indeed literally incorrect. As an exercise I compressed all the actual rules (which took some research) into a mere 29 words; but in fact about 75 words would be needed in real-world use to induct novices. Even sophisticated online analysts of the game found quite late that they had misunderstood aspects of it.
Given that irritation, compounded by the fact that different players were working with different information about the possible daily solutions, I gave up Wordle itself as uninteresting. A limited set of more common words but with de facto exclusion of plurals in -S (rendering useless any guess like the permitted try MOUES); an unknown mechanism by which the game selects daily solutions from that list (in fact they are pre-set, and easily unearthed); an unknown policy regarding the re-use of a correct solution. Such vaguenesses, and such levels of meta-game, interested me more: as a philosopher and a specialist in technical communication. But I gave up exploring those too.
Crosswords? Tediously meta-gamic also. Give me the writing of bilingual palindromes, any day.
Twenty?! You are truly a Master! I’m up to 68 and ready to pack it in.
Until that xkcd, I had not bothered to learn how Wordle even worked. Apparently, it’s a lot like Mastermind.
In the last couple weeks, I have actually made another effort to learn to do British-style cryptic crosswords, and I finally got the knack of it. The key was finding an online puzzle site that would let me know as I got each word correct. Ironically, I now seem to have the hardest time with the puzzles that include the most American-style clues, possibly humorous, but without the explicit division of each clue into a straight definition and a clue based entirely on wordplay.
A mere 130 guesses!
(This may not be my thing …)
I have, while remaining in the Wordle-avoidant control group, recently taken to making a daily visit to Worldle,* which has nothing in common with Wordle or most of its imitators except for some wordplay in the name. Possible distinguishing advantages include:
1. Has nothing to do with words or spelling.
2. I am better at it than my wife is.
*https://worldle.teuteuf.fr/
My take:
At 15 guesses I finally got into the top 1000 (963), with a similarity of 37.76, then on guess #16 I got the word (which I’d have thought would be closer than 963, though it is a different part of speech). I doubt that’s going to be the way it usually goes.
You are truly a Master!
Nah. Here in Australia it is already tomorrow, and so far I have not solved the latest Semantle. I’m up to guess 131, and have got no closer to word 1000 (the answer) than word 996 (similarity = 45.56). Too arbitrary, too vaguely prescribed, too many variables, too dependent on blind luck – too short a life.
Give me the writing of rhymed double acrostic sonnets, any day.
Australia too here. Here’s my result:
I solved Semantle #21 in 160 guesses. My first guess had a similarity of 5.61. My first word in the top 1000 was at guess #101. My penultimate guess had a similarity of 21.72. https://semantle.novalis.org/
I gave up after 48 guesses, best 978. Like too many tests, one spends too much time trying to understand the tester. I may give it another chance or two.
“rendering useless any guess like the permitted try MOUES” — while it helps a player to know that the set of answer words is smaller than the set of guess words, it can be useful to guess a non-answer word; it won’t be your winning guess but it might pinpoint the answer.
By “semantically similar”, I mean, roughly “used in the context of similar words, in a database of news articles.”
Most of the news articles I run across are a dog’s breakfast of Begrifflichkeiten that don’t cohere. Our dog wouldn’t touch them. He prefers expensive roquefort.
My best guess on the one I started last night was 984/1000, so 16th closest, I guess? My biggest breakthrough was guessing a capitalized word that turned out to be close, which reoriented my thinking, but I’ve run out of interest.
I just hit “Give up”, and wow, no, I probably would not have gotten there, as I don’t have a supernatural mindset. My closest guesses were things that I don’t believe in. I guess that’s the path I should have followed.
Of course, my sixth guess was extremely good at 704/1000, but I don’t think of it in concert with these other things.
I just hit “Give up”, and wow, no, I probably would not have gotten there
Same here, and I don’t think I’ll try again. Not my thing.
Tried again today. My fifth totally random guess was heresy, which scored 27.85 and 575/1000. But then my next three guesses — belief, religion, lie — all scored in the mid teens and were officially cold. It’s very hard to make useful inferences from these scores. So now I am officially bored.
I think the key is that ‘context’ is always relative to the secret word. ‘Heresy’ and ‘religion’ might be very close to each other, but they’re not equally close to the secret word, especially since ‘context’ here is the sentence.
Word2vec really only models semantic similarity in a way the approximates human intuition when it’s used to solve a specific problem. When it’s used in an open-ended way like this, it’s us humans who have to think like machines.
it can be useful to guess a non-answer word
Quite right, MM. I have used MOUES myself in Wordle as a rather good diagnostic (three vowels, two common consonants) rather than what I somewhat woollily called a guess: a genuine stab at the right answer.
So now I am officially bored.
Me too, David. Woolly tasks. I have more rewarding ways of wool-gathering, of my own devising.
Something in the vvay she moues …
Took me 100 tries to get to within one word of the word. Then I thew in the trowel. Immensely frustrating, it was, but also mind-expanding. But expanding how, and to what end?
Today 30 guesses only. Was this an easier word or just luck?
Holy crap, I got it in 17. Must be an easier word. Quite a feeling to jump from “cold” to 981/1000 — “Come, Watson, come. The game is afoot.”
50. Which is an improvement on 130, at any rate …
It would be (moderately) interesting to see if there actually is some way of ranking words as intrinsically more or less difficult. (Maybe that is actually the secret plan behind the website …)
It would be (moderately) interesting to see if there actually is some way of ranking words as intrinsically more or less difficult
You could probably come up with a score that weights proximity against absolute frequency, such that words with lots of high-frequency neighbours score higher than words with low-frequency ones.
Dordle (linked in the initial post) doesn’t seem to work. There’s no puzzle (game), nor a link to one, just information.
Just click on “Free Dordle.”
Today 47. Another “easy” word?
I prefer Quordle to Dordle. Octordle is a bit much though.
On today’s Semantle I got to 1 away from the secret word (999/1000) on guess 134, but now I’m stuck.
Hmph.
The current Semantle word — that is, as of 0:01 UTC 2022-02-24 — was a word that means a role that a person can have. I actually came within one word of the actual one (999/1000), then stalled out, because all of the reasonable synonyms were further away, sometimes much further away. I eventually switched to reading a novel, noticed in the novel the use of a vaguely similar word I hadn’t tried yet — and that turned out to be the exact word. But I feel lucky, not smart.
One of the very early words I tried was “dinosaur” — and all of the following (sorted alphabetically to obscure the order of guessing and actual ranking) were “colder” than “dinosaur”.
boss
council
general
governor
guard
head
lord
oligarch
pope
ruler
student
I call bullshit.
SPOILER for 2022-02-24
GUESS WORD METRIC ORDINAL
----- ------------- ------ --------
104 historian 100.00 FOUND!
100 archeologist 53.60 980/1000
56 archeology 46.52 928/1000
40 history 37.94 722/1000
Harrumph
Yes, I gave up at about 80, having previously got “history” and “writer” (46.67.)
It’s semantics, Jim, but not as we know it.
@Mollymooly: The spoiler warning didn’t help, since I scrolled to the bottom to catch new comments. I saw the answer before the spoiler warning. 🙁 (Actually, I wanted to see Owlmirror’s comment again.)
I got the closest word (999/1000) on my 51st guess. And, given that word, I’m particularly surprised at “student” being, supposedly, semantically farther away than dinosaur, given that (semi-spoiler) the closest word and student are semi-synonyms.
Normal people don’t know there’s a difference between archeo- and paleontology, and correspondingly classify dinosaurs as history. Maybe that’s what we’re looking at.
The ranking for “dinosaur” is 17.44, the ranking for “student” is 15.31.
The ranking for the plural of the actual word is 998/1000, 65.48.
Colossal bullshit, I say!
Other words in the top 1000 which are, I insist, more bullshit:
943/1000 woodturner 47.90
886/1000 buff 44.07
885/1000 landscapist 44.00
877/1000 watercolourist 43.65
837/1000 birder 41.65
836/1000 raconteur 41.63
829/1000 suffragist 41.34
828/1000 typographer 41.31
823/1000 realist 41.10
819/1000 plantsman 41.02
817/1000 nurseryman 40.98
815/1000 pamphleteer 40.88
812/1000 watercolorist 40.76 (yes, the same word as 877 with American spelling)
769/1000 ironmaster 39.20
767/1000 potter 39.18
747/1000 filmmaker 38.53
(probably more, but those are some that caught my eye in the high rankings)
Semantle is a weird game indeed but feeling it out is interesting. I got today’s word, spoiled above, in 45.
Other words in the top 1000 which are, I insist, more bullshit:
Hear hear.
I’ve been dipping into it again at idle moments. Mixed success. Today:13 guesses (Semantle 27). I’ll stop while I’m ahead, and resume life.
Today only 25 guesses, a PB.
I got today’s in 382. It took 314 guesses to get into the closest 1000. I don’t know why I continued playing.
I just got an 0.54 on guess #40 and am officially giving up.
There weren’t negative numbers before today, were there?
The ranking for the plural of the actual word is 998/1000, 65.48.
That’s probably because plural nouns don’t appear in the exact same contexts as singular nouns. If I’m understanding Word2vec correctly, it’s a vector derived from the words that appear near (in the same sentence as? I think there’s some weighting for how proximate the context is, but there also seem to be a couple different options for how to implement the algorithm) the target word. So if your corpus had three sentences, “The archaeologist was applying for a grant,” “The historian was applying for a grant,” “Archaeologists were applying for a grant,” then “archaeologist” would be closer to “historian” than “archaeologists,” then the model would (correctly) capture that archaeologist(s) and historian(s) both do the same kinds of things (applying for grants), but it also assumes (incorrectly) that the differences between the “archaeologist” and “archaeologists” sentences are just as fundamental to their semantics as the similarities with the “historian” sentence. A lot of this smooths out with larger data sets, but asking the model to triangulate the exact semantics of a single word throws up a lot of noise.
For February 27, I reached the secret word after 221 guesses. The 183rd guess had me at choreographer. Earlier after reaching mathematics on guess #18, I needed 169 more to reach “salsero.”
On today’s Spanish semantle, I surrendered after a couple hundred guesses; the closest 1000 were a string of 1st-person plurals.
I am truly addicted.
If the word2vec system doesn’t group alternate spellings together, singular/plural very close to each other, and synonyms at least somewhat near each other, what the hell use is it?
How was “violin” the closest word to “piano”? I can see that “organ” and “keyboard” have various meanings, but what about “harpsichord”?
Of course I still haven’t been able to stop, despite my frustration with the arbitrariness. At least I got tonight’s in 58. Part of the problem is that the other daily puzzles all start in the morning, so if I’m finished with all of them then Semantle offers something to do in the evening.
What is todays answer?
Heardle: a phonemic version of wordle
https://heardle.glitch.me/
It’s much harder to think of words with 5 phonemes than with 5 letters. And, note, it’s got an American English bias.
Heardle Daily 2022-03-05 3/6
⬜⬜🟩⬜⬜
⬜🟨⬜⬜🟩
🟩🟩🟩🟩🟩
https://heardle.glitch.me
Note, there’s another game called Heardle that involves music. It was a via a Google search for that that I found this.
Wow, that’s tough. I didn’t have a hard time coming up with words, but once I discovered there was a vowel that was there but couldn’t be in second place where it was most at home, I was in a morass. I couldn’t believe it when my final Hail Mary pass worked!
Heardle Daily 2022-03-05 6/6
⬜⬜⬜⬜⬜
⬜⬜⬜⬜⬜
⬜🟨🟨⬜⬜
⬜🟨⬜🟨⬜
🟩🟨⬜⬜⬜
🟩🟩🟩🟩🟩
https://heardle.glitch.me
“This project has received too many requests, please try again later.”
Such is the awesome power of LanguageHat. Slashdot, nothing.
Heardle Daily 2022-03-05 6/6
⬜⬜⬜🟨⬜
🟨⬜⬜⬜⬜
⬜⬜🟨🟨⬜
🟨🟨⬜🟨🟨
🟩🟨🟩⬜🟩
🟩🟩🟩🟩🟩
https://heardle.glitch.me
Definitely harder than Wordle.
Our host might want to test his proficiency in the profane arts with Lewdle.
L + ratio + no damsels???
L(ack of victory) + ratio (relative to the amount of supercilious attention your comment has accrued in relation to mine) + no damsels (for those who may be unfamiliar in the profane arts, damsels are women)
Yesterday’s word was “manager” and the closest word was “manger”? I think this so-called game is designed to cause as much frustration as possible in a certain subset of the population that includes most of the commenters here.
Yesterday’s word was “manager” and the closest word was “manger”?
OK, that’s stupid beyond belief, and makes me even gladder I gave up on this.
@ki
That looks like some kind of bug. I got there via assistant, associate, partner. What I found more curious was having “poet” rank higher than “narrator” and “storyteller” when the word was “speaker”.
I don’t think “manger” as the closest word to “manager” is a bug, at least not in the technical sense. It’s a result of the dataset, and typos, and how words are judged to be semantically similar. Info on Semantle says:
By “semantically similar”, I mean, roughly “used in the context of similar words, in a database of news articles.”
“Manger”, in the dataset, is usually a typo for “manager”, thus is very high semantic association with “manager”. I mentioned that to a friend, and later that day they texted back a screenshot showing that typo of “manger” for “manager”.
Today (14.03) 24 guesses, a new PB.
Yesterday was circle, and the closest I got was goalpost (899). The closest words were circles, clockwise, corner, timer, semicircle, semicircles, deking, slapshot, coterie, circled. The top 1,000 are full of hockey terms and other random sports terms. The game is an interesting idea, but the implementation is just too flawed.
Today (16.03) only needed 6 guesses. But my helper provided the initial and winning guess ..
Looking at it again:
I’m so glad I managed to break away.
What are some good ‘category’ words to try initially? I’ve tried topics such as philosophy,math, geography, physics history etc but it’s not helping. Food housing health another.
I guessed semantle#77 at 11 guesses and #80 at 12 guesses…I was lucky! (I guess)
Semantle now has hints!
Today’s was definitely heavily influenced by “semantically similar” meaning roughly “used in the context of similar words, in a database of news articles”.
The word was “spill”. Words related to household spills, not closely related. High words did include “gusher”, “disaster”, and “catastrophe”. Makes sense if you realize that in the data set used “spill” usually means an oil spill.
I didn’t get the word, even with 3 hints (after which point I gave up and had it show me the word).
Man, I hate to do this to anyone who’s successfully weaned themselves off, but:
Daily Duotrigordle #74
Guesses: 37/37
1️⃣7️⃣ 0️⃣4️⃣ 3️⃣7️⃣ 0️⃣9️⃣
0️⃣7️⃣ 3️⃣4️⃣ 1️⃣5️⃣ 0️⃣8️⃣
1️⃣8️⃣ 1️⃣6️⃣ 1️⃣9️⃣ 1️⃣4️⃣
1️⃣0️⃣ 0️⃣6️⃣ 2️⃣0️⃣ 2️⃣1️⃣
3️⃣2️⃣ 1️⃣1️⃣ 3️⃣5️⃣ 3️⃣3️⃣
3️⃣6️⃣ 3️⃣1️⃣ 3️⃣0️⃣ 2️⃣9️⃣
2️⃣5️⃣ 2️⃣8️⃣ 2️⃣7️⃣ 2️⃣6️⃣
2️⃣4️⃣ 2️⃣2️⃣ 2️⃣3️⃣ 1️⃣3️⃣
https://duotrigordle.com/
The thing with Semantle is not to try and grasp connections in some true semantic network, but to remember that the database is based on what journalists and editors put in some unknown (to me) sample of publications. It’s a bit like the game show Family Feud. What is correct doesn’t matter. What you’re going for is the responses of a random sample of Americans with time to kill.
I believe that people with extensive vocabularies may actually be worse at it, because while it tells you straight out that the answers have been chosen from a list of the top 5000 most common English words, I have no idea which those are, and the vocabulary in my head is not sorted by difficulty. This was underscored by the fact that I am absolutely no better at Semantle Junior, because it’s the same word net, just guaranteed that the actual word chosen will be a word every child knows.
I play it just enjoying seeing how many words I can find that are green, and imagining the context that puts giblet next to chocolate (really!) rather than trying to make sense of it.
I decided sometime in my 20s that I would only ever play one solitary game henceforth, and that was the card game known as Solitaire to some (it used to ship with Windows), Klondike to others, and Canfield to a third group. It seemed to me that I needed no additional addictions to those I already had. And I have given up on Solitaire.
Me too. I can’t do “quick” crosswords, because you don’t know if you’ve got the right answer until you look at the crossers and realise that you haven’t.
I do not count crossword puzzles as solitary games, because I do them only as part of a group (most recently just Gale and me). I once hung out with a group of people at Google who did cryptics in a group during lunch, though I was more of a kibitzer than a participant.
I had never heard of quick crosswords before today, but I gather they have the same topology as cryptics (only a limited number of crossers) but use straight dictionary clues. In American-style crosswords, every letter participates in both a horizontal and a vertical clue, which helps a great deal: you ask yourself (or your partner) “What is a word of the form A-O-E that means ‘Southwestern building material’?” Then after cudgeling your brains for a while you either come up with ADOBE or you don’t, in which case you move on to another word and eventually return to it in hopes that you will now have ADO-E or A-OBE. If I still can’t get it, I mutter to myself in alphabetical order: abobe, acobe, adobe … aha! Thus each answer initially stands on its own, as in cryptics, but as other answers are found the remaining ones become easier: what was once impossible becomes feasible.
Not all clues are dictionary definitions, and these provide entry points to the initially faceless squares. Some are encyclopedic lookups like “1975 World’s Series losers”, and for these I ask Dr. Google for help if I don’t already know. On the other hand, a clue ending in “?” indicates a pun or other wordplay, and when we get these we groan and imagine the well-deserved defenestration of the constructor.
And so it goes until the puzzle is complete. Sometimes we can’t fill in a few squares, or even just one, because both the horizontal and the vertical words remain impenetrable (e.g. a person’s name we just don’t know), and then we heave the puzzle overboard: this can also happen in mid-solution if there are too many stupid clues like “Award-winning actress in romantic comedies”.
I had never heard of quick crosswords before today, but I gather they have the same topology as cryptics (only a limited number of crossers) but use straight dictionary clues. In American-style crosswords, every letter participates in both a horizontal and a vertical clue, which helps a great deal
Russia seems to distinguish crosswords (кроссворды), which (at least prototypically) have limited crossers and descriptive non-cryptic clues, from scanwords (сканворды*), where most letters participate in both directions (the clues are typically written in the spaces that don’t contain letters, with direction indicated by arrows).
I very much prefer the latter, for exactly the reasons described by John Cowan; in a regular crossword I can know all the crossers and still have very little idea what the intended word was, while in a scanword that hardly ever happens even when a letter is not part of a crosser (usually because it ended up between two clue squares).
…Exceptions do occur, of course; one time in the early ’00s I was stuck on a “Capital of Western Samoa: AP_A” where the last letter wasn’t a crosser. Apoa? Apua? Apea? I thought it was probably a vowel in there but had no idea which (they all sounded about equally implausible) and ultimately decided that I didn’t need to care. (It’s Apia, by the way.)
And sometimes a long corner word only has crossers at the ends and is separated from everything else by a large clue area; those I tend to just ignore unless I can immediately guess the answer.
I’ve never done a cryptic crossword, and didn’t even find out they existed until stumbling into vague references in online blogs; it sounds like an implausibly hard puzzle, especially for the first few clues where I couldn’t just work it out by the crossers. I suspect that most American crosswords would also be too hard for me, but for other reasons.
*) TIL that the “scan” part is actually short for “Scandinavian” and has nothing to do with any kind of scanning
What J1M calls сканворды are indeed the type you find in your daily paper in Denmark, and I think in the rest of Scandinavia — the UK cryptic type looks totally impossible to a Dane. I can do the “quick” ones but as soon as you get into multiple word answers about stuff you only learn at Eton, I’m lost.
Same in Germany – what you usually get as Kreuzworträtsel is the сканворд type, you get a sprinkling of american crosswords only in specialist Rätselhefte, and I have only seen a few attempts to reproduce the UK cryptic type in German publications.
it sounds like an implausibly hard puzzle
They can be, but they don’t have to be: the main reason why they’re much more tractable than you might think on first principles is that there is a whole enormous body of familiar conventions underlying the way most cryptic clues work. Good setters make it interesting yet still doable by throwing in occasional clues that deviate from the conventions, or phrasing the clues in such a way as to lead you up the garden path by looking as if it follows one set of conventions when in fact it follows another. A really good clue makes you laugh when you finally see it. There is also an art in making the ostensible, non-cryptic, reading of the clue (the “surface”) plausible and funny in itself: I have an impression that female setters tend to be particularly good at this, but that may be the result of overgeneralising from a very small sample.
The fact that convention and experience (rather than sheer intelligence) are so important means that if you’re good at cryptic crosswords you can bask in a more or less wholly undeserved reputation for brilliance (or at least, you could, if anybody was impressed by crossword-solving skills apart from other crossword fans.)
…and they’re actually called amerikanisches Kreuzworträtsel.
The Наука и Жизнь (‘Science et Vie’) magazine has crosswords where the questions are set in the form of a picture, a text fragment, a map, etc, like this one,
I have a distinct feeling that sometimes the setters poke sly fun at someone or something, eg in one crossword, the answer to a question across was (Евгений) Матвеев (an actor) and the answer to the question down starting with the same м was макака ‘macaque’.
I put openly on record what is tacitly available above: AmE constructor vs. BrE setter for the authors of these things.
I would think of a setter as a dog breed, often to be found lounging on a settee.
On the internet, nobody knows that your crossword was constructed by a dog.
The greatest setter of all time. Much loved, much expostulated against, much missed.
I remember solving the crossword in the solutions of which he announced his fatal illness (with details.)
From this weekend’s cryptic:
Composition written by a setter one organised (8)
Lights (swipeover)
The greatest setter of all time. Much loved, much expostulated against, much missed.
I swear to god, until I clicked through I thought that was going to be about a dog.
So love, much expostulate, very missed, wow
Araucaria araucana, the monkey-puzzle tree after which Araucaria named himself, looks like it got its name from a minim-misreading of Araucana araucana. It was originally Pinus araucania, after the region of Chile where the Mapuche live; when it was moved to its own genus, it became Araucaria araucana. Go figure.
I like how once you get into a domain in Semantle, you think you’re on the right track, but there can be multiple distinct tracks to to answer. Like if the answer were head, you might get into the green with president and spend the whole time with chief, and leader and officer, while someone else might start with body. Also, I think that Semantle Junior and possibly Semantle itself are actually easier if your personal vocabulary is smaller, because you can’t get stuck listing the forty near synonyms you know for your best word so far. I find it helps a little to think of the way a word might be used in a newspaper.
That’s not allowed in botany: genus and species name are not allowed to be the same word. When the fir was separated from the spruce, it changed from Picea abies to Abies abietina.
There’s no such constraint in zoological nomenclature: add subgenera and subspecies, and you get Bison (Bison) bison bison sometimes. Oh, and if you see Salamandra maculata in 90-year-old literature, kill it with fire.
David Marjanović: Bison bison bison… appears to be an exact parallel to the use of any n ≥ 2 instances of ‘buffalo” to form a syntactically valid sentence.
Well, the theoretical maximum is Bison (Bison) (bison) bison (bison) bison. Unlike higher ones, the genus- and species-group ranks are fully enumerated in the Code: genus, subgenus, “group of species”, species, “group of subspecies”, subspecies.
(I’ve never seen “groups of subspecies” used, and “groups of species” are very rare. Even subgenera are far from common.)