I’ve referred to it before, but Frank Jacobs’s Strange Maps site really deserves a post of its own. My initial impetus was the latest entry, “It’s 10:15 in Germany. Do You Know Where Your Isoglosses Are?,” perfect for LH with its clear presentation of the areas of the German-speaking world that say viertel nach zehn, viertel elf, viertel ab zehn, and viertel über zehn for 10:15. (As lagniappe, it throws in a map of the many words for indoor slippers.) But scroll down and admire the maps of Clapham Common, Ground Zero of the Saints, Europe’s Many Midpoints (“Early 1900s: German geographers concluded that Europe’s midpoint was not located in the Austro-Hungarian Empire but in – what a surprise – Germany: near the Frauenkirche in Dresden”), Nazis Up the Mississippi And Other Axis Invasion Scenarios, and Fanlands: Football Supporter Map of London (the post includes as well the strikingly simple Barassi Line dividing Australia between followers of Australian rules football and rugby football). And there’s hundreds more where those came from. If you’re at all interested in geographical visualization, you need this site bookmarked.
I can remember saying “quarter of ten” as a boy, but I can’t now remember whether we meant “quarter to” or “quarter past”. I wonder when I forgot.
“At the back of ten” was nicely unambiguous and usefully imprecise: I still use it.
In my experience “quarter of” means the same as “quarter to”.
You can also say “ten of”, or “five of” or “twenty of” (in each case, with the option of “to” instead of “of”). But you can’t say “fifteen of [or to]”. And if it’s thirty minutes before 10 you can’t say “thirty of [or to] 10”, or “half of [or to] 10”; you have to say “half past 9”. Or “half 10” if you’re British.
I knew a guy in college, from Colorado, who said “quarter till”.
It seems to be that the English have taken to saying ‘half-eight’ when they mean ‘half-past eight’. Unfortunately, this conflicts with what I learnt in German long ago, where halb acht means 7:30… (I hope I’ve got it right)
Oh yes, I think I’ve got the English expression wrong, and probably by confusion with the German expression.
And then there is the unique Cantonese way of telling the time.
Instead of saying ‘a quarter past ten’ or ‘a quarter to ten’, the Cantonese apparently say ‘ten three’ and ‘ten nine’. The ‘three’ and ‘nine’ refer to the digits on the dial of your watch. That is how you get ‘three’ meaning ‘a quarter past’. I can’t remember the exact wording that they use, and here in China for some reason the results of Google searches frequently come up as ‘page inaccessible’, so it’s difficult to find the information that I want.
(Despite what the 愤青 (‘angry youth’, read ‘chauvinistic young nationalists’) might tell you, Baidu is not up to Google as a search engine, which makes my life on the Internet in China pretty frustrating. Luckily I go back to Mongolia tomorrow.)
Found a source!
http://www.cantonese.sheik.co.uk/time.htm
Look at the left column for examples of colloquial Cantonese usage. The right column gives standard Chinese usage.
E.g., 1:05 becomes (left column):
一點一 jat1 dim2 jat1 (literally ‘one o’clock one’).
Compare Standard Chinese (which is also used in telling the time in Cantonese) in the right column:
一點零五分 jat1 dim2 ling4 ng5 fan1 (literally ‘one o’clock zero five minutes’).
viertel über zehn for 10:15
??!! Now I see that it occurs only in Austria: “Another national Option is viertel über zehn (‘quarter over ten’), used only in central parts of Austria (5).” But I have to admit that it makes sense.
Ø: I knew a guy in college, from Colorado, who said “quarter till”.
I say that. Possibly I picked it up from my parents, both of whom hail from Mississippi.
I had a Norwegian friend in England who missed an appointment because of the hour’s difference between the Norwegian halv elve and the English “half eleven” (Norwegian being like the German and, I read, Catalan). In Norway I say to my family “halv elve, ten thirty”, just to make sure we’re in agreement.
I’d forgotten about the map site. I love it. Have there been such things as “outdoor slippers” since Shakespeare’s sixth age shifted into the lean & slippered pantaloon (which I’d always assumed were outdoor)?
I’ve had it bookmarked for ages. It’s one of my favourite sites — one that I go to immediately whenever Google Reader tells me it has something new. I particularly liked the one about the commune boundaries of Liechtenstein, which are seriously weird for such a small country. Unfortunately they’ve messed* up the search function since moving to a new host, so searching for “Liechtenstein” won’t work, but it’s at http://bigthink.com/ideas/21375
*Your own host wouldn’t allow the word I had here originally.
My Grandfather, born in Victorian London, used to say five and twenty past (or to) the hour. It didn’t survive his generation as far as I know.
I wish I could now remember more of the words and phrases ‘globe’ for light bulb for example that were relics of his age but they’ve all gone,along with his collar-studs and meat-safe.
“‘half-eight’ when they mean ‘half-past eight'”: yes, I still use that.
@ Ø: “I knew a guy in college, from Colorado, who said “quarter till”.
I didn’t go to college in Colorado, but that is what I say (Southern AmE).
Does “five-and-twenty past”, the four-and-twenty blackbirds method of counting, have any direct connection to the similar method used in German, Norwegian etc., or is it coincidence?
“four-and-twenty”: are you thinking of this old favourite?
http://www.youtube.com/watch?v=mXiwkz-t5H4
I say that. Possibly I picked it up from my parents, both of whom hail from Mississippi.
I say it too, possibly picked up from my father’s side of the family (Arkie/Okie).
I’d like to see a cartwheeling, mini kilt-wearing Kenneth MacKellar sing a disco version of that in this year’s Eurovision song contest.
My Grandfather, born in Victorian London, used to say five and twenty past (or to) the hour. It didn’t survive his generation as far as I know.
I think it did. It sounds a little old-fashioned but otherwise unexceptional to me. I am pretty sure that is because I have heard it used by family members (Birmingham, UK) born long after Victoria died.
I was taught “virtel nach zehn” in ninth-grade German back at the end of the 70’s. Is that also the theoretical uniform/prestige Hochdeutsch standard or simply one contending regional variant? But what I think of as the Urheimat of prestige Hochdeutsch (Upper Saxony) is shown as solid “viertel elf” territory on the map. The only part of Germany I ever spent meaningful time in (summer ’82 in Neu-Ulm, so Swabia right on the Bavarian bank of the Danube) is I think shown as “viertel elf” territory (although the more Bavarian part of Bavaria is back to “nach zehn”) but I actually don’t recall encountering that usage. How do those people say 10:45? Drei-viertel elf?
Crown: No coincidence at all; the “five and twenty” style is a direct Germanic inheritance. A little googling showed me a 71-year-old Englishman still using it in 2008.
J.W.: How do those people say 10:45? Drei-viertel elf?
Just so. And not only “those people”.
Is that also the theoretical uniform/prestige Hochdeutsch standard or simply one contending regional variant?
It’s a variant, but there’s no contention. It would take a bold silly-billy to publish an article in Germany with the aim of talking up certain ways that people have of telling the time, and talking down the others. No matter how often German pundits scuffle over dative-versus-genitive matters, they don’t take up arms against the time of day.
It seems to be that the English have taken to saying ‘half-eight’ when they mean ‘half-past eight’. Unfortunately, this conflicts with what I learnt in German long ago, where halb acht means 7:30…
Dutch follows the continental Chermanic convention, of course. But “half februari” does *not* in fact mean the middle of january, and they keep refusing to tell me why not.
UK “half four” = US “quarter of eight”
It’s interesting to see that the viertel elf, drei viertel elf region in Austria is along the border with Hungary. In Hungarian the quarter-hours are expressed in the same way (negyed 11, háromnegyed 11, where negyed = 1/4, három = 3).
Very interesting indeed!
Federal Chemistry Olympics of 2000. The teacher wanted us to meet again at 14:15. Being Viennese, he said viertel drei. Most others thought that must mean viertel nach drei (15:15). The Tyroleans thought they had misheard viertel vor drei (14:45), which is dreiviertel drei elsewhere. Several minutes of fun confusion ensued.
Viertel über is, in Austria at least, rather restricted; I think it’s limited to southern Upper Austria.
Correct.
There is no such thing. The dialects in Saxony are closer to the standard than most others in many respects, but… I’ll stop here before spending an hour on the complications. 🙂
Oh yeah, I didn’t know about Hungarian; that’s fascinating. I wonder what other Sprachbund features than this, curses, and a bit of phonology there are.
I am surprised, given the Russophones who frequent here, that no one has mentioned that time in Russian is similar to Eastern German and Hungarian – you say “pol vtorogo” (half of two) for 1:30 and you say “chetvert’ vtorogo” (a quarter of two) for 1:15. I have never heard “tri chetverti vtorogo” for 1:45 though, that would be “dva chasa bez chetverti” (Two o’clock without a quarter).
On reflection, I should have mentioned that Russian uses ordinal numbers for hours rather than cardinal. I.e. you say “half of the third” (hour being implied) or “quarter of the third”, which to English speakers may seem more logical than “half three” or “quarter three”, which is the literal translation of the German expression.
Oops. The map says viertel über is less restricted and more scattered than I had thought.
I should have known about Russian, but didn’t.
Mandarin counts back from 29 minutes to the full hour, using chá “lack”.
Mandarin counts back from 29 minutes to the full hour, using chá “lack”.
Actually 差 chà, I think. And the chà theoretically comes before the hour, e.g., 差一刻十点 chà yí-kè shí-diǎn ‘quarter to ten’. Except that I don’t think I’ve heard many people actually talk this way…
After the characters, the tones are the first thing I forget. 🙁
What’s even more fascinating is that I promptly forgot this!
Mandarin counts back from 29 minutes to the full hour, using chá “lack”.
Western Yugur (which is only distantly related to Uyghur and not at all related to Eastern Yugur, which is a Mongolic language spoken in the same area as Western Yugur despite the names) uses the next following -teen word in 2-digit compounds, thus per otus lit. ‘one thirty’ is 21, not 31.
One of the third ten. Perfectly logical.
“quarter of the third”, which to English speakers may seem more logical than “half three” or “quarter three”, which is the literal translation of the German expression.
Mandarin counts back from 29 minutes to the full hour, using chá “lack”.
But Mandarin for the half-hour is half past the hour/same as English/unlike German: shí-diǎn bàn — gloss as ‘ten-hour half’. Shí-diǎn yī-kè — ‘ten-hour quarter’.
Danish — you all wanted to know about Danish! — Danish has halv otte for half past seven, but everything else has explicit i ‘to ‘ and over ‘past’. So minutes up to 29, except kvarter (pronounced kvart over and kvart i unless you’re trying to reduce redundancy, viz. et kvarter over fem).
(This is reading out 12 hour cl times; there is no fixed am/pm convention, but of course you can disambiguate with om natten/morgenen/formiddagen/eftermiddagen/aftenen or i nat/morgen tidlig/morgen formiddag/eftermiddag/aften to specify the first such in the future. 24 hour clock times are read out with hours and minutes as numbers, nothing else. [morgen = ‘morning’ but i morgen means ‘tomorrow’ more generally, so you need to force the time-of-day interpretation]).
It turns out that there is no simple way to specify a midnight unambiguously. Writing fredag klokken 0:00 will make some people think your service window is the night between Friday and Saturday and then you get angry phone calls when you power down the firewall. The easiest fix is to start at torsdag klokken 23:00 instead or say midnat mellem torsdag og fredag. How does that work in the Anglosphere? (contrast theory and practice).
Is 24:00 unknown? Donnerstag um 24:00 = Freitag um 0:00.
I keep forgetting which of 12 am and 12 pm is noon and which is midnight because the etymology gets in the way (I mean ante/post meridiem).
I keep forgetting which of 12 am and 12 pm is noon and which is midnight
I don’t think anybody knows this. I hate those expressions. What on earth is wrong with “noon” and “midnight”?
Digital clocks will show 12:00am at midnight if set to am/pm display. Unless that’s just something a Chinese chip programmer decided themself, I suppose it’s “correct” in some way.
Conventionally 12pm = noon, 12am = moon. When I need to indicate time close to either of those (sometimes I have to do it in an electronic system that only has am/pm choice) it is 12:01 or 11:59. I would prefer 24h clock, but it would be un-American and a sure sign of coming communism or antichrist or both.
Conventionally 12pm = noon, 12am = moon
Moon a typo for “midnight”, presumably?
If compelled to guess, I’d actually have had them the other way round. Digital clocks know nothing and in any case are an Abomination. Abomination, I say! *
Etymologically, neither term makes any sense at all, as noon is neither post nor ante meridiem, and midnight is exactly equally both.
* The only acceptable way of marking the hours is by striking a piece of metai with a hammer.
(cf Hausa ƙarfe “metal”, ƙarfe uku “three o’clock”; Kusaal karifa atan’.)
Torsdag 24:00 never happens, the hour field of the relevant structure is defined as 00-23. (Your clock is supposed to show 23:59:60 during a leap second, if you believe such exist).
More to the point, I’ve never heard a Dane say klokken 24. 0.15 is kvart over tolv om natten old style, but the 24-hour cycle starts over from 0. It’s just that some people manage to think that Friday 0:00 follows Friday 23:59 while Friday 0:01 was the night before.
(Train and bus timetables may not be blameless here– the column for “Saturday” continues to around 05:00 on Sunday. Possibly connected to the fact that one-day tickets are valid until then, However, official communications about cancellations and the like are careful to specify such times as, e.g., mellem 1:00 og 5:00 natten efter lørdag).
There is, in fact, a perfectly coherent reason why 12 a.m. is midnight and 12 p.m. is noon. Since time measurements are always rounded by truncation, they have to be right continuous. Clearly, 12:01 is a.m., since it’s after midnight. Likewise, 12:00:01, or even 12:00:00.01 should be a.m. When you look at your clock, it may read just 12:00, but you know that it’s actually some number of seconds after, and hence a.m., not p.m.
@Lars Mathiesen: The last time I paid close attention (which was admittedly a few leap seconds ago), the actual rule for the leap second was that the last two seconds of the year would both be 23:59:59.
Since time measurements are always rounded by truncation, they have to be right continuous.
A dictum a day keeps the grey cells in play. But “have to be” ? What does that even mean ? Seems that the arrow of time is in there somehow.
Full disclosure: I know what truncation is, and what right/left continuous mean.
“Around the clock”, when not rund um die Uhr, is 0 – 24 h. I’ve even seen a digital clock or two that showed 24:00 (before moving on to 0:01).
OK, that’s bad.
WP.en sez:
Also 8601 actually allows T240000 to designate the end of a day; this option was removed in 2019 and reinstated in an amendment two months ago. (So WP is imprecise — ISO 8601-1:2019 unamended did not allow 24 in the hours field).
Also also “they” have now decided that there will be no leap seconds added after 2035. If you ask me, that’s a case of laziness winning over technical correctness. And I’ll always be on the latter side myself, of course.
>Moon a typo for “midnight”, presumably?
I think that’s a joking reference to Luna as the goddess of the nighttime. I would mention that it’s very difficult for my kids to recognize that the moon has no real solar-system association with nighttime. Every few months, one of them will notice the moon in the daytime with renewed surprise. We need a mythology in which the moon is the messenger between sun (daytime) and stars (nighttime) to bring the subconscious understanding of the moon into agreement with reality.
Yes it was. Also, Moon (the lesser light, to be more precise) is charged with governing the night in Genesis 1:16, even if sometimes it does it in absentia.
No, it’s a genuine case of the rotation of the Earth speeding up (or at least slowing down more slowly). That happens sometimes when subducted ocean floor slabs sink toward the core.
I thought it was a deliberate rhyming mnemonic…
@DM, yeah, I saw that graph that looked like the slowdown was slowing down a lot — and actually might go the other way for a while, but the Earth does as it pleases. And the decision (last month, according to WP) was to stop doing leap seconds no matter what the Earth does.
What I mean by laziness is programmers going:
1) integer timestamps not increasing at one SI second per tick would need us to actually code.
2) The UTC date not changing every 86400 ticks exactly would need us to actually code.
3) So what if the UTC date doesn’t change when the sun passes nadir at longitude 0 — which is more or less the definition of the date — the users can’t tell. We’ll deal with it when we’re a minute wrong. Maybe.
(The leap second system breaks 1) because the support library people decided not to deal with 2). If the Earth is actually returning to keeping SI time, that mitigates the error in 3) — but not forever).
TIL that since Linux kernel 3.1, the code needed to mitigate 1) is as simple as calling clock_gettime(CLOCK_TAI, &time) (*). I call stupid on the CGPM.
_______________
(*) If your system can reliably find the offset when it starts. I hope NTP is our friend.
Is Hausa ƙarfḕ ‘metal’ literally ‘what has strength, strong stuff’, cf. ƙarfī “strength’, ƙarfàfā ‘strengthen’, ƙàƙƙarfā ‘very strong’? (A bit like Arabic حديد ḥadīd, both ‘sharp’ and ‘iron’, beside حد ḥadd ‘edge’ and حد ḥadda ‘sharpen, put an edge on, delimit’, Proto-Semitic *ḥdd ‘be sharp’.) Or is ‘strength’ derived metaphorically from ‘metal’ (like English mettle from metal)?
It never actually struck me before, but, yes, I think ƙarfḕ must indeed be related to the “strength(en)” root. It would have struck me as odd semantically, but the Arabic analogy shows that it’s not unlikely at all. Bargery actually gives the “iron” word as ƙarfi, homophonous with “strength”, though, interestingly, he lists the “iron” meaning first. I don’t know enough about Chadic to know if it’s possible to say which meaning is primary; I would imagine on first principles that the “strength” meaning was, but ironworking is old enough in West Africa to figure in some presumably quite ancient protolanguages.
Kusaal kut “iron” (wrongly written with a nasal vowel in the Bible) is formally plural, but the original singular survives in the personal name Akudug “Akudugu.” The same stem appears in the verb kud “work iron.” This etymon is reconstructable to Proto-Oti-Volta, cf Moba kud “iron, metal in general”, where the tones conveniently confirm that the word is cognate and not borrowed from WOV. It doesn’t seem possible to relate it to any words meaning “strength.”
Kusaal also has the unanalysable word saen /sãj/ for “blacksmith”, and the etymon can be reconstructed to the protolanguages of Western-Oti-Volta and Buli/Konni; it turns up also in Waama: cáárō (all regularly derived from *cã:ʎ-.) Nawdm has the agent noun kudta, from the “work iron” root.The rest of Oti-Volta has forms cognate with e.g. Moba maal, which looks like a specialisation of an old agent noun “maker.”
Waama is not close to WOV, so this makes you wonder if the maal languages have just lost the original word. But things aren’t simple: Waama shares the tone-inversion change with WOV/Buli-Konni/Yom-Nawdm, and it may be that it was historically just the first branch from the tone-inversion subgroup (there is some other lexical support for this idea.)
Proto-Bantu *-beda “iron” is itself old, but obviously unconnected with Proto-Oti-Volta *kut- , which is hardly surprising.
@ Ryan, bravo:
We need a mythology in which the moon is the messenger between sun (daytime) and stars (nighttime) to bring the subconscious understanding of the moon into agreement with reality.
This is a brilliant idea. Why have I never heard of it before?
The moon is the router of the heaven’s wi-fi system. (Let us pray that the universe is not rebooted in our lifetimes!)
DE
Re no kut with meaning “strong”, could the Kusaal etymon have had initial s instead of k?
sũtãanaũ̀-ã́-àsũtãananamn Satan, diable, démonSatan, devil
sũtook1n+nrudesse, amertumeharshness
sũtook2ncolèreanger
Compare
kũsṹvpiquer, poignarder, vacciner, percer, injectersting, pierce, stab, give an injectionAba’ane mɔrɩ õ bii tɩŋ tɩ kũs ye õ gu’ul ʋ ne bã’as.
sũsũ̀vcouvrir, recouvrir, cacherto cover, shelterKa sagbãʋŋ ma’asɩm tɩna sũsɩ ba.FrnEt un nuage est venu les couvrir.
sũsã’aŋ1ʋ̃̀-ã́sũsã’amɩsntristesse (lit. cœur gâté), frustration, découragementsadness
sũsã’aŋ2sũsã’amɩsnchagrin, peine, deuilsorrow, mourning
If this is just a silly question, you don’t have to answer….
Kusaal kut “iron” (wrongly written with a nasal vowel in the Bible) is formally plural, but the original singular survives in the personal name Akudug “Akudugu.” The same stem appears in the verb kud “work iron.” This etymon is reconstructable to Proto-Oti-Volta, cf Moba kud “iron, metal in general”
Kusaal also has the unanalysable word saen /sãj/ for “blacksmith”, and the etymon can be reconstructed to the protolanguages of Western-Oti-Volta and Buli/Konni; it turns up also in Waama: cáárō (all regularly derived from *cã:ʎ-.)
Not at all sure the connection is real, but I can’t look at all that without immediately thinking of (proto-)Songhay gúurú “iron” and zàḿ “blacksmith”, neither of which have obvious language-internal etymologies…
could the Kusaal etymon have had initial s instead of k?
No.
The sũ- in the Toende words you cite is the combining form of “heart” (Agolle sunf.)
Sũtook, for example, is the same word as Agolle suntɔɔg /sũtɔ:g/, literally “heart-bitterness.”
Sutaana “Satan” is a loanword, of course (in the first instance, from Mooré, and further back, Arabic.)
(proto-)Songhay gúurú “iron” and zàḿ “blacksmith”
The voiced initials are a pity … and the Kusaal s- is definitely from an original palatal stop: the changes of Proto-OV *c *ɟ -> s z are regular throughout in WOV apart from Boulba, where both fall together as /c/.
It’s easy to imagine words for iron technology migrating with the technology itself, though.
Saen and its cognates are odd, too. The stems are unanalysable. It’s conceivable that the forms might have been influenced by sɔen /sɔ̃j/ “witch”, which is reconstructable to Proto-OV, though any such remodelling would have to be quite ancient.
Iron and smith aren’t trivial either. Iron, PGmc. *īsarną ~ īzarną, is a loan from Celtic, where it turned out to be derived from the PIE word for “blood” once Hittite was well enough understood. Smith was unanalyzable until a conference presentation in 2015 (handout, slides) that still doesn’t seem to have been published; it’s related to smite and comes from good old *temh₁- “cut”, apparently…
It never actually struck me before, but, yes, I think ƙarfḕ must indeed be related to the “strength(en)” root.
Thank you for you comments, DE.
PGmc. *īsarną ~ īzarną, is a loan from Celtic, where it turned out to be derived from the PIE word for “blood” once Hittite was well enough understood.
The etymology of the Celtic word (as here, p. 172, under *isarno-/*īsarno-) was one of my reasons for asking about ƙarfḕ.
People still quote the etymology that associates Celtic ‘iron’ with Sanskrit iṣirá- ‘vigorous, active, lively, quick’ and Greek ἱερός, meaning ‘holy’ already in Mycenean, although Homer also has something like ‘powerful, fast’. For example, here is Daniel Kölligan (2018) “The lexicon of Proto-Indo-European” in Klein et al., Handbook of Comparative and Historical Indo-European Linguistics, vol. 3, p. 2237, but without any explicit discussion of the phonology:
(The association of ‘strength’ and ‘holy’, note also Old Irish níab ‘vigor, vital spirit’ (< *nei̯bʰo-) beside noíb ‘holy; holy person, saint’ (< *noi̯bʰo-; cf. Old Persian naiba- ‘good’, Persian نیو nēv ‘brave, valiant, warlike’), or indeed English healthy and hale beside holy.)
For LH readers who are curious, here is Warren Cowgill’s original proposal deriving Celtic ‘iron’ from Proto-Indo-European *h₁ésh₂r̥ ‘blood’, in Indogermanische Grammatik, vol. I.1 (1986), where he discusses the semantics of the etymology explicitly:
Yeah, that completely fails to explain the long vowel. Though if a short-vowel form *isarno-, as cited, actually existed, then maybe the two words simply got confused.
That *h₁isHro- was present in Celtic is shown by the river Isar in Bavaria.
> adjective formed with vṛddhi > nominalized, “the bloody stuff”.
*ē &t; *ī is regular in Celtic in the first syllable.
“I think that Celtic *īsarno- is derived from an older *ēsr̥-no- ‘bloody’; the reason for this may lie in the reddish color of some iron ores and/or the propensity of iron to rust and/or the efficacy of iron weapons. Compare Finnish rauta ‘iron’ from an IE designation for ‘the red thing’.”
Wiktionary on rauta: “From Proto-Finnic *rauta, borrowed from either Proto-Germanic *raudą (compare Old Norse rauði (“hematite”)) or Proto-Balto-Slavic (compare Latvian rūda (“ore”)), but in either case ultimately from Proto-Indo-European *h₁rewdʰ- (“red”).”
Wiktionary on ἱερός: “From Proto-Hellenic *iherós, from Proto-Indo-European *ish₁ros. There are a number of candidate cognates with this word. Compare Sanskrit इषिर (iṣirá) and Oscan 𐌀𐌉𐌔𐌖𐌔𐌉𐌔 (aisusis).” Add Bozzone’s law, and “*ish₁ros” could easily be the same thing as *h₁isHros. The Oscan word can’t be related, though.
The U.S. and U.K. militaries use the 24-hour system, and eliminate the ambiguity (which would be truly dangerous with reference to when an attack takes place) by beginning events at 0001.
@JC, that’s where the right continuity comes in. My mathematical way of thinking can’t contain the existence of people who think that Friday 0000 is (or might be) 1439 minutes later than Friday 0001. Absent an official rule to the contrary (as in bus time tables) the minute 0000 has to be the first one on a given date, or civilization will presently end.
(But they do exist which proves that Maths Is Wrong!)
The version of ISO 8601 currently in force allows you to specify 24:00 as well, which I would probably only use as the endpoint of an interval. But it might actually confuse fewer people to say Thursday 2400.
I will now shoot down this amalgam of “truncation” and “right continuous” in connection with a function.
If you’re talking about a discrete function f(x) arbitrarily defined at a finite number of values x0, x1,,,, of a continuous variable x, being 0 elsewhere, f is not continuous in x (unless all values are zero). Apart from this special case, f is neither left nor right continuous. Truncation of an f(xi) doesn’t change that (unless f(xi) was the only non-zero value, and was truncated to zero). f does not suddenly become right continuous at an xi on truncation of f(xi).
If you’re talking about a continuous function f(x) of a continuous variable, alteration of f(x0) at any x0 destroys the continuity at f. Both left and right continuity go up in smoke there.
I just learned the word unconformity: “a buried erosional or non-depositional surface separating two rock masses or strata of different ages, indicating that sediment deposition was not continuous.”
“Short paraconformities are called diastems.”
German (and other) terms for quarter-hours later on LH. (I don’t often get to phrase it quite that way…)
This is reading out 12 hour cl times; there is no fixed am/pm convention, but of course you can disambiguate with om natten/morgenen/formiddagen/eftermiddagen/aftenen or i nat/morgen tidlig/morgen formiddag/eftermiddag/aften to specify the first such in the future. 24 hour clock times are read out with hours and minutes as numbers, nothing else.
Russian works kind of like that; you say три часа дня for 15:00 and три часа ночи for 03:00. There are really just four such disambiguators, ночи/утра/дня/вечера “night/morning/day/evening”, with the “day” word corresponding to English “afternoon”.
IIRC archaic (19th century?) Russian did have a fixed am/pm convention: пополудни for PM, пополуночи for AM. I don’t think I’ve ever heard anyone actually using those terms IRL, but I do occasionally encounter them in old books.
24-hour times, when they occur (which is not too uncommon), are indeed read out as numbers, with added zeroes as appropriate [at least in the minutes part]. А вчера в девятнадцать ноль восемь…
It turns out that there is no simple way to specify a midnight unambiguously.
Yup – which is why it’s so common for deadlines to be 23:59 rather than 0:00.
In my own personal timing reports (e.g. timestamped file names) I’ve been known to use 24:xx, 25:xx, and 26:xx [I don’t recall if I’ve ever gone further] for times late at night when it was nominally the next day. I’ve heard of other people doing this as well, but AFAIK it hadn’t caught on in any major contexts.
Your clock is supposed to show 23:59:60 during a leap second, if you believe such exist
That’s what I heard as well, complete with a screenshot of such a clock. I never (knowingly) caught a leap second myself, and now I probably never will.
Train and bus timetables may not be blameless here– the column for “Saturday” continues to around 05:00 on Sunday. Possibly connected to the fact that one-day tickets are valid until then
TV programs also tend to run the day until about 5 in the morning of what is nominally the next day, then start the next day’s column at 6 (IIRC the exact timing of the transition varies).
I’m not sure of the one-day ticket thing; we didn’t have those in Moscow. Reportedly the Israeli ones used to be valid until 04:00; no idea if they still are. Monthly tickets, both in Moscow and (again reportedly) in Israel, run out at midnight when the month ends even though public transport is still open.
There is, in fact, a perfectly coherent reason why 12 a.m. is midnight and 12 p.m. is noon. Since time measurements are always rounded by truncation, they have to be right continuous. Clearly, 12:01 is a.m., since it’s after midnight.
The incoherent part (as far as I’m concerned) is calling it 12 (and 12:01) in the first place. It should be 0:00 a.m., then 0:01, 0:02, and so on to 0:59, at which point it transfers to 1:00 a.m. even in the incoherent US system.
(Russian has no problem with this because двенадцать ночи “twelve of the night” is unambiguous [at least in respect to not being noon] and anything after that would be however many минут первого “minutes of the first”.)
What I mean by laziness is programmers going:
1) integer timestamps not increasing at one SI second per tick would need us to actually code.
2) The UTC date not changing every 86400 ticks exactly would need us to actually code.
3) So what if the UTC date doesn’t change when the sun passes nadir at longitude 0 — which is more or less the definition of the date — the users can’t tell. We’ll deal with it when we’re a minute wrong. Maybe.
AFAIK some of it is also “we have already coded a lot to deal with added leap seconds, but the way the Earth’s rotational speed is going these days we’re likely going to end up having to remove a second to keep sync with it, and that would be an even more complicated bunch of coding to deal with, so we’ll rather get rid of the problem entirely”.
@J1M, we’re likely going to end up having to remove a second: Yeah, from eyeballing some of the graphs I found at random places, we may* never get to the next leap second going up. But unless something drastic happens, it looks like it could take several decades before it gets to where it goes the other way. (The difference UT1-UTC looks like about 0.6s, and it would need to be -0.6s and trending down before they can schedule a negative leap). But why “they” didn’t code for negative leaps in the first place I don’t understand. There are well-established systems for dealing with DST changes which are sometimes announced less than 6 months in advance. Even the kanji for the current Japanese imperial epoch (announced 2 weeks before the coronation IIRC) were updated in time on the distribution I was using. It’s not rocket science. (And as I noted, current Linux kernels will give you TAI if you ask nice. No leap seconds there, so why mess with the definition of UTC?)
* EDITED so as to not aggravate mine host with might.
EDITED so as to not aggravate mine host with might.
No, no, I have nothing against that usage. You’re misremembering my animus against counterfactual “may have” (for Proper “might have”).
The difference UT1-UTC looks like about 0.6s
In fact the difference you’re talking about [actually UTC-UT1, judging by your description] is about +0.02 seconds at the moment, and expected to go negative sometime in 2023; it hadn’t been close to +0.6 since the leap second before last.
But yes, even if the trend continues, it will probably take a few years before we get far enough to warrant a negative leap. (Though it might be before 2035, which might make things tricky.)
But why “they” didn’t code for negative leaps in the first place I don’t understand.
Presumably because at the time (the 1970s) the TAI-UT1 difference was going up so fast (at over 2 ms/day, so nearly a full second per year) that a negative leap second was a theoretical construct – and since long-term the Earth’s rotation is slowing down (due to tidal forces), it was expected that none would ever be needed.
Because local civil time is tied to UTC, so a clock that ticks LCT has to deal with seconds like (here in NYC) 2016-12-31T19:59:60.
Oh, 70s? That means the tidal slowdown was known, but plate tectonics had only just been developed and many of its implications had not been explored at all.
From the OP:
With cruel irony, this now forwards to https://bigthink.com/articles/ideas-are-immortal/ …
@J1M, ah, of course, UTC is behind TAI and each leap second inserted makes it behinder. And UT1 leads UTC in behindness. I didn’t use enough napkins to get it right.
I was looking at this graph but interpreting it wrong. (Zero is centred vertically, not at the bottom, and the curve has to go _down_ to trigger an inserted leap second). I can see that the -0.6 rule wasn’t followed strictly in the 90s, and especially in June of 2016 it must have looked as if the discrepancy would reach -0.6 at year’s end, but then UT1 stopped slowing so fast and the difference was only -0.4 when the scheduled leap second happened. Being a chronographer is complicated.
(I thought the rule was that the difference has to be greater than 0.6s [absolute value] when the decision is taken [6 months before the leap second] and the goal is to stay inside 0.9s, but it looks like “they” have been trying to stay inside 0.6s).
@JC, now that’s an argument I can understand. But I suspect that just like various orthographic reforms trying to remove obstacles to computer processing by removing “difficult” rules, technology will overtake the change — I bet in 2035 the displays will be running Linux and using NTP over a G5 data modem, so nobody will care if leap seconds exist; they just call localtime(3) and display the result.
(Actually I don’t know how localtime(3) is able to distinguish between 19:59:59 and 19:59:60 since its input is exactly that CLOCK_UTC value that doesn’t tick on the leap second. But the manual says that it can return 60 in the tm_sec member. A mystery).
Actually I don’t know how localtime(3) is able to distinguish between 19:59:59 and 19:59:60
…Come to think of it, did any timezones with non-integer-minute offsets intersect with the period where leap seconds existed?
Because if they did, that’s going to be a real mess for time conversion, because you couldn’t just assume that the 60 is the leap and all the others are normal.
[Looking it up, the answer appears to be “just barely no”; the last time zone with an offset with seconds in it was eliminated in January 1972, and the first leap second occurred in June 1972. But I’m not very confident that there weren’t later edge cases.]