（评论）

（评论）
(comments)

原始链接: https://news.ycombinator.com/item?id=41022653

荷兰语字母“IJ”最初代表长“e”音，就像英语“ee”一样。随着时间的推移，某些单词的发音逐渐类似于英语中的“eye”。尽管“ie”如今在荷兰语中常用来表示长“e”音，但“IJ”在荷兰语字母中仍然是独一无二的，因为它通常被认为是一个独立的单位并且是大写的。从历史上看，学校中“IJ”的教学时间早于“y”，因为荷兰语本土单词中没有后者。然而，现在在计算机系统中它被排序在“i”下，然后是“j”。一些以“Y”结尾的地名实际上发音为“IJ”。在某些情况下，“iy”（长“i”）的使用也被“ie”取代。在荷兰语中存在三种表达相同声音的方法：“y”、“ie”和“ei”。这些变化是由于拉丁语、希腊语和法语的历史影响而产生的，导致语言中出现复杂的语音模式。与多次尝试改革英语拼写规则（导致美国和英国版本之间存在差异）不同，荷兰语拼写在各个地区保持了相对一致性。尽管随着时间的推移，语言和地缘政治发生了许多变化，包括罗马人、法兰克人、维京人和西班牙人等各种入侵者的影响，但荷兰人民和语言的统一在历史上始终保持不变。这种一致性的一个值得注意的例外是来自其他语言的借用。由于其作为全球主导力量的作用，英语吸收了拉丁语、希腊语和其他各种语言的大量词汇。因此，数百年前的英文文本仍然难以理解。相反，法国努力保护自己的语言并压制相互竞争的地区方言，从而形成了相对统一的国家语言，除了细微的语言差异。虽然法语拒绝采用英语借词，但英语则拥抱来自不同来源的术语的折衷组合。当然，不同语言和口音之间的音素表示不一致的情况比比皆是。例如，即使荷兰人能够完美地发出荷兰语“ij”的发音，说英语的人也可能很难正确发音。同样，说日语的人可能会面临理解用古英语创作的作品的挑战，例如《贝奥武夫》，而

The Dutch digraph ij originated from the same custom. Originally written ii, later ij, and pronounced as a long i (English "ee"), the sound later shifted to be similar to English "eye", see the Dutch names for the cities Berlijn and Parijs. Meanwhile, the long i sound is now written as ie. IJ is the only Dutch digraph that tends to be treated as a single letter and is capitalized as such. In education in the Netherlands it is taught as the 25th letter instead of y, which does not occur in native Dutch words, and in the phone book it used to be sorted together (mixed) with y. But since the advent of computers it is generally sorted as i followed by j. There are a few place names like Ysselsteyn where Y is pronounced as IJ.

I'm pretty sure I was taught the letter y (I-grec) as part of the alfabet and not the ij (I'm Dutch as well). The other day I was surprised to see that an alfabet song that my kid was watching used the ij instead of the y. And I thought you were talking about IJsselstein, but saw there's indeed also Ysselsteyn, and they're not the same town (though pronounced the same). Interestingly we've then 3 ways to write the same sound (though y can also be pronounced as 'ee', e.g. Yvonne); y, ij, ei, oh, and I suppose 'ey' should be counted as well though that's not used in modern spelling. Cool

My Mum is an Australian Yvonne and married my Dutch father Jan. When we spent a whole year in the Netherlands back in 1971, Dutch acquaintances would always pronounce her name something like Yuh-vonna, quite different to how in English it is said like Ee-vonn. (My Dad's name is pronounced like Yun of course)

And yes as a 7 year old Aussie it was interesting learning that the digram "ij" was the 25th letter on the Dutch alphabet

One of the things that blew my mind when I took a few semesters of Dutch in college is that the language occasionally gets amended (or more precisely, the way it's taught in schools is amended, although after a generation or two of delay the effects are the same) to fix inconsistencies. I can't remember if clarifying 'ij' and 'y' was one of those, but I remember the professor telling us about how a while back people were upset when they updated the spelling of the word "pannenkoek" (which I think had previously been spelled "pannekoek"), which some people apparently still haven't gotten over after almost three decades[1].

I wish there were enough willpower for something similar for English, but it's probably too late to reach any sort of compromise on whether to use a "u" in words like "color"/"colour".

[1]: https://dutchreview.com/news/dutch-government-argue-over-spe...

Well, I think the difference is that for Dutch, people are mostly colocated, i.e. share the same space and have a national curriculum to guide the use.

For English, it's quite different because many countries list it as their official language but may have diverged spellings and meanings and there is no single body to direct the curriculum. The most notable is the US vs. British English and the u in colo(u)r is a mere spelling example. Consider surgery or elevator, which are bigger discrepancies. I remember from my school days, as a non-native speaker, these were much more troublesome and we had a special test just to check we could tell which is which in US and GB English.

> these were much more troublesome and we had a special test just to check we could tell which is which in US and GB English.

I'm American and sometimes the only way I know that I'm using the GB spelling is when autocorrect flags it. Grey and defence are two just from the top of my head.

> For English, it's quite different because many countries list it as their official language but may have diverged spellings and meanings and there is no single body to direct the curriculum. The most notable is the US

The US doesn't have an official language, though most institutions operate primarily in English, and all US states that have one or more official languages include English on that list.

Nor does the US have an official curriculum. In the 80s, there was a wide variation both between and within states as to what a student with a high school diploma might have been expected to know. When the state of Illinois increased the graduation requirement for high school students from 1 to 2 years of mathematics, my high school added a second-year general math class because too many students couldn’t pass pre-algebra because fractions were beyond their ken. (The graduation requirement was subsequently amended such that no math below algebra could be counted for graduation and then later that students needed to have taken at least math up to the level of algebra and geometry because of districts stretching algebra I into a two-year sequence).

> Nor does the US have an official curriculum.

The US (i.e. the federal government) has no legal authority over education at all -- schools are run at the state level, and individual states often do have standard curricula.

Most (maybe all now?) states have standard state-wide curricula. Some are painfully vague, some are painfully detailed. When I was looking at this when I got my credential, it seemed like many small-to-medium sized states copy-pasted their curricula from other states.

I think the reason is different. Some languages have official bodies that decree how a language works. For example the French have the French Academy.

English has no "boss" in charge of how the language works and who decides what the correct anything for the language is. The closest are style guides, but they come from multiple organisations and each often different to each other. So, it's harder to just decide that something will be spelt differently.

A language is a dialect with an army and a navy.

English had multiple attempted spelling reforms, but that mostly resulted in differences between American and British spellings of the same words. There was no central power to enforce a single English spelling because America had successfully broken off from the British Empire whose coercive force backed the English language.

In contrast, France was extremely good at putting down revolutionaries in their colonies. Even when they gained nominal independence France would invent debts to impoverish their former colonies. France was also extremely good at putting down local languages, even within France itself. France waged a century-long campaign to hunt its own minority languages to extinction[0] and the language regulators are themselves part of the same infrastructure.

In other words, "regular spelling" is a colonial atrocity and English's incomprehensible spelling rules are a demonstration of freedom.

This can even be demonstrated in each language's attitude towards loanwords. English stole huge swaths of Latin, Greek, and French vocabulary. Text even a few centuries old is difficult to read and anything older than Shakespeare is completely inscrutable[1] as there's too many Germanic root words for this ostensibly Germanic language. Meanwhile in France, their language regulators fight tirelessly to invent new calques for English loanwords to stop the "Anglicization of French", as if that were a thing that needed to be stopped.

Well, I suppose there is one reason why you'd want to stop it: loanwords make it far harder to regularize spelling. Think about, say, all the Japanese loanwords in English. Hepburn romanization[2] tried to make them readable to English speakers, but didn't do a great job, so now English has to accommodate spellings like karaoke, tsundere, or sakura alongside spellings like pleonasm, beef, or Olympic. Good luck inventing a spelling system that can both faithfully represent the "correct" sounds for all of those words, while also remaining legible to present English readers, and being adopted in every country that speaks English to some degree - off the top of my head that includes America, Canada, the UK, Australia, India, Nigeria, and South Africa.

[0] Which I suspect may have also inspired analogous atrocities in Canada against indigenous Canadians

[1] As I put it to a native Japanese speaker: "You can read Genji, I can't read Beowolf."

[2] Which itself isn't used by native speakers since you can't round-trip Hepburn to Romaji. Japan itself uses Kunrei-shiki romanization, it's the one that spells つ as tu instead of tsu.

> Meanwhile in France, their language regulators fight tirelessly to invent new calques for English loanwords to stop the "Anglicization of French", as if that were a thing that needed to be stopped.

That's something one hears quite often from English speakers but I don't really know where this myth comes from.

French is a quite decentralised language with each country having its own "regulator" but none of them have legal power. They are all just advisory organisations. Many the one with the most power is the OQLF in Québec. The French ministry of education decides what is taught in France and some of its reforms are sometimes followed by the other countries, but they don't deal with vocabulary itself, mostly orthography.

The Académie Française does invent words, but they have no official value or power. Their main occupation is documenting actual usage.

Most other European languages do have international, language wide regulators though, often with actual legal value, but French doesn't have anything like this.

I think it all comes down to English speakers knowing even less about the other languages than they know about French.

This looked as incomprehensible to me (a native English speaker) as a foreign language, albeit one with a bunch of scattered words I recognized…until I figured out that þ meant th and all the sudden I could mostly read it fine.

The French do something similar, as do the Germans. They've had relatively little luck with substantial changes recently, though. The 1990 French reforms were rather trivial. For example. the placement of the diaeresis was shifted to the first vowel in some digraphs. So "aiguë" (acute, sharp) is now "aigüe" since it is two syllables, but the former spelling misleadingly hints at three syllables. That one seems to have been well-accepted. Many of the other reforms have been only partially accepted, or explicitly rejected by many speakers. (I don't think I am ever going to get used to "ognon" for onion.)

> I wish there were enough willpower for something similar for English

English tends to develop through evolution, not design. Whose willpower would we be talking about anyway?

Dutch has an amazing variety of dialects. Strong dialect speakers from the southwest and northeast corners of the language area will not be able to understand each others dialects.

> there's only one prestige Dutch dialect though

Which one would that be? AN which nobody actually uses (at least here in Flanders)?

I don't know much about how it works within the Netherlands but within Belgium at least the Dutch accent gives you no prestige at all. I would expect the reverse to be true in the Netherlands (whichever accent one would choose to represent the Flemish) since the Dutch sometimes don't even recognise Flemish speakers as native Dutch speakers and answer in English instead.

Afrikaans is a related but separate language. Suriname, as the third Dutch speaking country, does use the same spellings as Belgium and the Netherlands.

It’s more like dye-kstra. But that would still be wrong. It’s a sound that doesn’t exist in English.

Daikstra - you can pronounce this

Dijkstra - a sound in between you cannot pronounce coming from English

Diekstra - you can pronounce this too

In my region and accent (Boston area), I don't know how I would pronounce the first and third one differently. Maybe this is a Mary/Merry/Marry thing based on accent.

Just guessing here, because my pronunciation is the same as yours...

Dai = Day

Die = Die

If you make an "ay" sound, then transition to an "i" sound, you can feel that the sounds are continuous. So maybe that inbetween sound OP is talking about literally the sound you make with your tongue between "ay" and "i".

In IPA, the name is /dɛikstra/. Americans (English vowels vary quite extremely by dialect/accent, so I'm sticking with Standard American English) can easily pronounce ɛ, we say the word 'bed' as /bɛd/.

But the 'ɛi' is a dipthong, and we don't have that dipthong. We have aɪ, as in 'bride' /bɹaɪd/, which is what usually comes out, at least for those of us who know that /dʒɪkstɹə/ is not even close to right. What you'll hear is generally something like /daɪkstɹə/, which make sense, as we pronounce 'dike' /daɪk/ as well.

So yes, completely correct, I just wanted to add the IPA for clarity.

To my understanding, most American and British English dialects[0] have /eɪ/, e.g. in "maze", which I imagine is quite close to /ɛi/.

[0] The exceptions being dialects where it monophthongizes to /eː/

Dutch /e/ has the tendency to become /eɪ/ as well (not in Flemish, I think), so /eɪ/ just sounds mostly like "e" to us (Dutch speakers in NL). The "ij" is really /ɛɪ/, but because we don't really have /aɪ/, it turns out that /aɪ/ works fine as a fairly non-ambiguous substitute.

True, but to my ear at least, /eɪ/ is further from /ɛi/ than /aɪ/ is. Maybe more to the point, English speakers (American ones at least) are going to "realize" the name Dijkstra with an /aɪ/ dipthong. "Dayksra" would not occur to us to say.

Phonemes are a funny thing: if an American were to pronounce that last name perfectly (or a Dutch person speaking English, for that matter), other Americans would still hear /daɪkstɹə/, or at most, "/daɪkstɹə/ but it sounds a bit foreign".

Some sounds which we don't produce are audibly not-English-phonemes, like German ü or Scandanavian ø, but others are not. Another example of this is that native English speakers aren't usually sensitive to the difference between the r sound in our word "pear" and the r sound in Spanish "pero", but recognize that "perro" is a different rhotic sound.

> True, but to my ear at least, /eɪ/ is further from /ɛi/ than /aɪ/ is.

If I may ask purely out of curiosity, are you a native English speaker? I only say this because, to my American English ear, [eɪ] and [ɛi] are virtually indistinguishable. Maybe I'm just misunderstanding what you're trying to say....

A local theme park as a "dutch quarter" and I as a kid was always fascinated by one booth having a letter "ÿ" in the name. I later learned that it was some kind of alternativ spelling for ij, and if you spell ij in black letters its really easy to see why that might have occured.

It's also quite common in handwriting; ĳ in cursive looks exactly like ÿ, and even when the letters aren't all joined, that is a common ligature. When I studied Flemish as part of the integration course, all of the teachers wrote ÿ, and so I presumed that it was just the done thing.

In Zeeuws (a Dutch dialect) many words in standard Dutch with an "ij" are pronounced as "ee" and written as "ie". So not everything has changed everywhere. (e.g. Dijk is Diek)

My favorite part of ĳ is the folk transliteration to the character "ÿ".

In cursive script, it's easy to see how "ĳ" becomes "ÿ". However, "ÿ"s Unicode codepoint of U+FF means it often shows up as a placeholder for "unrecognized character", doubling down the misapprehension that "ÿ" is the correct character in dutch text.

My favourite parts are (1) how you can see a capital U with a small cutout at the bottom of the left straight line used as IJ, and (2) how some foreign words are mangled by Dutch people who are not so good at spelling, like bijoux spelled as byoux, which can in turn confuse foreign visitors.

this makes so much sense - speaking German and English fluently, my brain always defaulted to interpret "ij" in dutch as "ee’" - but it took your comment to make me understand that it's indeed correct, and actually why

Well, it used to be correct until about 1600 or so. There are a couple of words like bijzonder where it is still pronounced as English "ee", and in the suffix -lijk it is reduced to a schwa.

30 years ago I knew a very old Swiss woman who told me they used the ‘letter’ ij (“yotee” if I remember correctly) instead of just a j (j in Swiss/German makes a y sound, like in jawohl). Hoping a Swiss person can fill in the details.

There's also an Amsterdam movie theatre across the IJ named "Eye" Film. I(expat)'ve always heard dutch "ij" as an over-annunciated "eye" where the sides of your mouth are a more widely open.

It's a sound that does not exactly exist in English, somewhere between the diphtongized a in maze and eye. Quite many dialects kept the old ee pronunciation.

As far as preventing forgery/tampering with the number, it's funny because this applies to merely the smallest part.

The "j" instead of "i" prevents you from adding another 1 or 2 to the number, at most.

Meanwhile, you could still change "ii" to "vii" or "xii" or "lxii" or whatever on the left side.

And still often append "i" on the right side when it doesn't end in "j" -- "x" can still become "xj", "xij", "xvij", etc.

It's just funny that you'd bother at all with "j" when it achieves so little. I see that it was also a practice to draw a line over the number -- why not just draw a whole box around it?

My theory #1: It's not protection against forgery, it's protection against accidental misreads. It's less likely a random smear looks like an "x" of "v" than an "i", so they don't really benefit from the terminal marker. And "i" can never start a large roman numeral, so just a terminal marker is fine.

Theory #2: It didn't start as such protection. It started as a typographical convention -- remember, there was a time when "i" and "j" weren't separate letters, "u" and "v" were not separate letters, and so on -- which then people started using as a reading guide.

It's weird nothing in the Stackexchange answer makes the point about the letters not being separate. In Latin (at least during the Middle Ages) J and I were used interchangeably. It was simply a matter of style or, at times, the random whim of the scribe. A person's name might start with a J and then in the same document be an I. As for numerals, tacking a J (or really what often just looked like an I with a longer straight descender) was common in late medieval Gothic scripts.

Forgery doesn't make sense, yeah. You gave some reasons, but also western medieval and early modern conceptions of forgery, for instance, were very different than our modern one. Ease of reading and clarity makes sense.

Nothing in the answer (or anywhere in this thread) distinguishes "when". It don't doubt that in the nineteenth and early twentieth centuries the "j" was used in this way, but that might not have always been the case.

I've been skimming a bunch of 16th–17th century English books with blackletter type recently, and the relation of ꝛ to r, ſ to s, v to u, j to i, ʒ vs z, etc. seem generally similar, just typographic variants of a letter to be used interchangeably or with one form in the middle of a word and another at the end or beginning depending on the letter. (Disclaimer: I am not an expert on this topic.)

There is an old German idiom for deceiving or defrauding somebody that roughly translates to "faking an X for a U". It's origin is people turning "v" into "x" on bar tabs or records. Given that this made it into an idiom and far outlived the use of roman numerals I assume this used to be a very relatable experience once upon a time.

On cheques, you write the dollar part of the amount in words, with no margin to add anything on the left, and then immediately followed by a long horizontal dash to strike out anything that may be appended.

In Japanese and Chinese, there are alternate numerals used on bank notes that have more strokes and therefore can't be amended. The normal number 1 is 一 and the fancy bank version is 壹.

The source from the article doesn't sound like it's a measure against tampering, but instead a way to stop the patient receiving the wrong dose by accident. Prior to the early 1900s I'm pretty sure you could buy cocaine and opium over the counter in most of the English-speaking world, so there wouldn't have been any reason to alter your prescription.

Edit: the source given by the article was from 1919, so it would have been within a couple of decades of the earlier laissez-faire drug policy.

It was actually ſ+z, which is easier to see if you look at it in Fraktur, and is literally the name of the character. I have long been puzzled by how its romanization became „ss“. In older writing ‘s’ was just the terminal form of ‘ſ’ — if you needed two you just wrote ſſ, same as in English into the early 19th century.

Other ligatures like tz and ch did not survive the jump into Roman letterforms, only ß for some reason.

By the way the now obsolete dual letterforms of s (ſ/s) was inherited from Greek. We’re lucky it was still in use in Leibniz’s time or calculus notation would be more confusing!

Not only the romanization, also the capital form in German is SS.

The ß sticks around in German because it prevents some ambiguities. For example Busse and Buße are different words with different pronunciations.

Basically German orthography backed itself into a corner because of two rules: a vowel before a doubled consonant is short, and an s before a vowel is pronounced /z/. This leaves you with no clear way to write a long vowel followed by an /s/ sound followed by a vowel: VssV would make the first vowel short, and VsV would keep the first vowel long but the consonant would be pronounced as /z/. The ß solves this problem.

sz is officially still an accepted substitution for ß though. It's basically never used, except if you capitalize a word with ß and substituting ss would lead to confusion (if both words exist and would make sense). Of course since 2017 the upper case ẞ is also an option, but nobody wants to use that either.

No, it's because the letter J came about in the 4th century AD to cover the consonant version of I in Latin.

J and I are thus equivalent.

There is a similar history between U and V. The original latin alphabet just had V. U covers the case where V is a vowel. That is why in old carvings and statues they sometimes write U as V.

I'm talking about historical use of these letters in Latin.

But if you've seen middle and early modern English text that hasn't been sanitized for modern readers, you will see a lot of stuff. You'll see W written as uu. You'll see lots of Vs and Us trading places, in places we no longer put them.

And eg. I kind of suspect the UK pronunciation of lieutenant, where u is /f/, has roots in it being /v/.

I read it as supporting my theory.

It says that in French of the time as heard by English people at the time, the phonetic value of that u was ambiguously heard as f or v.

My position is v and u were phonetically ambiguous or sometimes heard as the same in Latin and romance languages. Maybe also in older forms of English.

In roman times June was Iunius (pronounced You-nee-ew-s, or something like that, I am not a native English speaker).

The change in the pronunciation is just centuries' worth of language evolution.

In majuscule forms (i.e., all upper case) V was the normal shape for u/v while in minuscule forms (i.e., all lower case), u was the normal shape. This is why you’ll see V in engravings rather than U. Albrecht Dürer‘s Of the Just Shaping of Letters which shows only upper-case forms provides a V shape but not a U shape. His textura, at the end of the book, has a single letter in the upper-case, but both ü and v in the lowercase with a distinct base character beneath the umlaut for ü.

>Why isn’t “june” then “iune”?

I'm not certain about the etymology, but many languages do use an "i" sound for the names of June and July. I wouldn't be surprised if it started with an "i", then people switched to writing it with a "j" while still using the same sound, and then later people began pronouncing it differently.

> Why isn’t “june” then “iune”?

Because it’s not Latin? Things change quite a lot as centuries pass. But it could have been different, like in German, in which Juni (=June) is pronounced /iuni/.

J is one of those letters that gets treated wildly different depending on the language. Compare Germanic and Slavic languages where j = /y/, French and Portuguese where j = /zh/, Spanish where j = /h/ (but with a bit more edge than English /h/, kind of halfway between /h/ and /kh/).

The International Phonetic Alphabet is helpful, though it may add confusion to talk about this with people unfamiliar with IPA.

What you describe as y is /j/ in IPA.

What you describe as zh is /ʒ/.

Spanish j is usually /x/. Some languages transcribe /x/ as kh or ch. There are accents of Spanish that pronounce it as /h/ or something slightly different however.

I don't understand why it's clear that xx is terminated, but not clear that iii is terminated, requiring the last i to be a j.

(Furthermore, if you have additional syntax on it like a superscript th, why is the j still needed then. That's a pure onion in the varnish at that point; someone just using the j without thinking about its purpose and necessity.)

It’s easier to squeeze in another i than another x in modifying a written note. Sort of like why amounts are written out in words as well as numbers on checks (young readers, ask your parents what checks are).

Interesting stuff! The 'j' at the end of Roman numerals used to be a flourish, indicating the end of the number, especially in older texts and prescriptions. It's a neat example of how writing practices evolve over time. Anyone else come across quirky historical notations like this?

Those come from latin lex, rex, so possibly more related to sloppy handwriting getting standardized?

Edit: on the other hand, x used to be read /j/ (e.g. Quixote, Mexico), so also possibly lex —> lej —> ley

> Those come from latin lex, rex

No, in fact they descend from the accusative forms, lēgem and rēgem. I don’t know much about the precise evolution, but g → y strikes me as a perfectly normal sound change, especially before /e/. Looks like it happened in the other Romance languages too: compare French roi, Portuguese rei, Italian re, etc.

In that case, German switched to "t", not English from. High German is the main odd one out of the Germanic languages in this case.

Compare "Dag" (low German)"; Saterland Frisian "Dai"; West Frisian "dei"; "dag" in Old Saxon, Dutch, Norwegian, Swedish, Danish); Faroese and Icelandic "dagur"; Gothic "dags"

d -> t is one of the major High German consonant shifts. Other examples of the d -> t shift is Vater vs. vader, fader, father (from Middle English "fader", so there the "th" in modern English is also not from the German)

Others prominent changes around the same time included e.g. p -> ff (Schiff vs. ship, schip, ship, skip, skib in other Germanic languages) and /t/ -> /t:s/ etc. - compare Zwei with two, to, två in other Germanic languages.

Very few of those changes made it into English, either Old English at the time, or later - English in that respect tends to be closer to Low German than to current standard German.

(as an English speaker - or a speaker of basically any other Germanic language -, to learn German having a rough idea about the consonant shift is a major help in figuring out German vocabulary - you'll find a lot more German words are "close enough" to be understandable once you recognise the changed sounds)

correction: spanish is heavily ablative case instead of accusative. original statement is correct in principle since lex, rex are nominative and first in a nominative-genetive form in latin dictionaries. eg. lex, legis. rex, regis.

> spanish is heavily ablative case instead of accusative

Is it? Aren't the descendants of the ablative mostly -amante adverbs and e.g. conmigo? The wiki article of Ablative Case doesn't even mention Spanish.

Got any other examples of ablative in Spanish?

X in Mexico or Quixote used to be pronounced as "sh". I think X is still used for that sound in Portuguese or Basque or Catalan if I'm not mistaken.

In the case of Mexico they were transcribing a word from indigenous people.

In a bunch of Spanish words coming from Latin, it was often from Latin double S. Eg. páxaro, from Vulgar Latin passarum. I can also think of cases of this sound coming from Arabic, eg. modern ojalá was once oxala which is a cognate with "inshallah", note the sh sound.

Sometime around the 1500s this sound changed to /x/ and they changed spellings of a few different sounds whose values had shifted over the centuries. X had merged with a sound they spelled with J, derived from Latin words with J in them.

Old Spanish was probably closer to the real pronunciation and had also more variety, depending on the whim of the scribe. Most of the greatest works ever written in Spanish are prior to the standardisation by the RAE at the end of the 18th century, indeed

I don't know. An /i/ near a vowel in Spanish tends to become /j/. I usually see the consonant y transcribed as [ʝ].

Wikipedia on Spanish phonology notes the difference in pronunciation between viuda and ayuda which is an interesting contrast in two very similar sounding words.

Old Spanish had a lot of phonological differences so I'm not sure you can do a straight comparison.

I know; but since Don Quixote and Lazarillo de Tormes the modern phonetic changes were already there.

I own an anotated Lazarillo book for modern Spanish readers, btw. It will describe you every ancient jargon and old sayings as footnotes. And, of course, some of the context.

> IJ probably developed out of ii, representing a long [iː] sound [...]. In the Middle Ages, the i was written without a dot in handwriting, and the combination ıı was often confused with u. Therefore, the second i was elongated: ıȷ. Later, the dots were added, albeit not in Afrikaans, a language that has its roots in Dutch. In this language y is used instead.

> Alternatively, the letter J may have developed as a swash form of i. In other European languages it was first used for the final i in Roman numerals when there was more than one i in a row, such as iij for "three", to prevent the fraudulent addition of an extra i to change the number. In Dutch, which had a native ii, the "final i in a row elongated" rule was applied as well, leading to ij.

> Another theory is that IJ might have arisen from the lowercase y being split into two strokes in handwriting. At some time in the 15th or 16th century, this combination began to be spelled as a ligature ij. An argument against this theory is that even in handwriting which does not join letters, ij is often written as a single sign.

So indeed maybe the same as the Roman numerals. Interesting, I didn't know this. (Also interesting that no one really knows where this letter came from.)

It doesn’t undergo the phonetic process of umlaut, Don, and as it’s a vowel the two dots are more like a simple diaeresis.

It is used in that way in French too.

It is neither umlaut nor diaeresis, but just an i and a j with each their own dot smushed together into a ligature, which can take either the form of a y with both those dots on top, or a more rounded form that more resembles an i and j joined together.

I guess in other languages than Dutch it could be a y with a diaeresis or umlaut.

The most interesting thing is that that post sat for months without the answer "why" it was done, but chatgpt knew, from the comment added 2h ago.

The j was to prevent forgery, or altering the document. ii could be altered later to iii.. but if it was ij its obvious its been tampered with if it later appears as iji

The question was asked over 10 years ago and now it's getting clicks from HN someone's copypasted a claim from a LLM, possibly without verifying it. That is kind of interesting, but not in a good way.

The question was "asked Oct 10, 2013 at 1:54" and the answer (which explains why it was done) was "answered Oct 10, 2013 at 2:58". A more likely take on chatgpt's answer is that their anti-fraud explanation is not even correct; the 2013 answer provides citations from the time explaining why, and anti-fraud is not mentioned.

> if all the letters i and J are not dotted, the pharmacist may be in doubt as to the number intended.

Oh, is that where "dot the i's and cross the t's" comes from? I can imagine that saying being hammered into many medical practitioners back then.

Pretty sure that's just a derivation from everyday handwriting. The i-dots and t-crosses are two typical elements that can't be written without lifting the pen, so it's quite a frequent thing to put pen to paper, write out a word that includes i and t but leave them unfinished so you can continue the rest of the letters in one fluid motion, then go back and finish them off by dotting and crossing them.

Hence dotting is and crossing ts is going back to do final touches that you've neglected or may have missed while working quickly or casually.

Really interesting thanks for sharing. It's great to learn these little things which would have been well known at one point but start to be forgotten as time and society progresses.

They're more difficult to fake.

If the doctor prescribes 10 tablets, the user could be tempted to add a zero and thus get 100 tablets instead of 10. It's more difficult to turn an X into a C.

But you can add a C just as easily as adding a zero, and now you've turned 10 into 110 or 90 depending on whether there was space before or after the X. I guess adding a D or M would be a bit more suspicious.

Why? If I'm prescribed a course of something for 90 days, I will get 90 pills, even if the packs have 100.

It might differ by location, but here (I'm in the UK), it's not unusual to e.g. get 14 pills of antibiotics if prescribed a two week course even if the package is in units of 10, for example.

I never claimed it was impossible, just more difficult. Prevention of misreading is more important thar prevention of forgeries, I assume, and any misread og forged number would also have to match existing packaging sizes.

Adding a c is hardly more difficult than adding a 0.

Matching packaging sizes is not necessary - I have had prescriptions that do not match packaging sizes many times - they will cut up sheets or transfer to other containers as needed.

That's not universal. In my country, prescription medicine can only be sold in sealed boxes. Fractioning or tranfering to a different container is expressly forbidwn by law, so doctors always prescribe a number of boxes, not of pills. This makes a little more difficult to forge prescriptions.

That must be recent, as it certainly was not the norm in the UK even prior to Brexit. For some medicines, UK pharmacists are required to split packs as needed to give exactly the number of pills prescribed.

I can't find anything to suggest EU rules prevents split packs - I do see there was a tightening of the rules on how to treat packs if dispensing split backs a few years ago, but nothing since. Do you have a source?

EDIT: At least as late as per the False Medicines Directive going into force in 2019, split pack dispensing was explicitly allowed - the FMD set out rules for the controls required to do so. E.g. the original container can not leave the dispensary until its entire content is used, and the pharmacy can't re-sell or supply other pharmacies with it.

The cited text has an extra detail:

  «… and it is customary to draw a line above all the letters making up the number, the dots of i and j being put above this line; for example, xviij»

So, XVIII would have been written likes this:

  _____
  XVIII
  -----

Lines would clearly delineate the beginning and the end of the number. This is how I have written the Roman numerals since school.

So, if the script had X written that way, adding a C, would have also required both lines to be extended. Not impossible to forge, but an extra effort nevertheless.

Extending a straight line sounds like the easiest part of the process compared to reproducing a letter so that it matches the handwriting. As an indicator of error this seems to work fine, but to prevent fraud it seems marginal at best.

And as per the instruction of 1919 "it is customary to draw a line above all letters making up the number, the dots of i and j being put above the line" (in other words, the horizontal line is directly above v, x, c, and so on but between the dot of i and j and the lower part of the letter).

I think this way it is difficult to add more Roman digits like x to fake a higher amount than prescribed.

The letter j seems to be an additional security because i as a letter is narrow and one could perhaps squash in one or two more i's. The instruction also mentions that the number of dots is important, if the number of i and j is not the same as the number of dots above the line, the apothecary should assume an error.

Checks are quite foolproof too - spelling out the number (a hundred eighty eight) and writing the number (188), like on checks.

The paragraph you mention seems to worry more about mistakes more than tempering attempts though (taking 10x the dose prescribed could indeed be problematic)

The important part there is the termination characters. "Dollars" / "Pounds only" etc. and the decimal point (though that's easier to turn into a comma and make into thousands... Funnily enough the German way of using comma as the decimal separator avoids this)

Polish still regularly does this for contracts and all kinds of legal document.

I regularly see verbiage like "person A promises to pay person B 100 (in words, one hundred) zloty upon termination of this contract."

Someone forging a number is already going to the trouble to find seclusion so that they are unobserved, and to try to match the writing instrument (color, stroke and texture) and the handwriting of the existing digits.

A horizontal square bracket would have been a more effective deterrent than just a line.

My point is more that the people of the past weren't stupid; they didn't believe that a horizontal line couldn't be easily extended by a forger to hang over newly added digits. Rather, our explanation today is lacking; there must have been some other reason for the practice. Or no reason other than that someone stated it and it become something that is always done that way.

They say that you should never fix a curfty system by rewriting it. The transition from Roman numerals to Arabic numbers is a great example of a situation in which you should ignore this advice.

Here I was hoping it was indicating a complex number, since i was already taken.

...on that note, how would one write complex numbers in Roman numerals, now that both i and j are taken?

> A j was used for the final i, to make it clear the number had ended.

Yeah I imagined this was the case. Same as other common ways to avoid misunderstanding (and possible forgeries)

（评论） (comments)

（评论）
(comments)