Irish names marked as typos: Tech companies' diversity failure

There are so many “correct” variations of most names that if you flag them all as valid, you’re also flagging most misspellings of them as valid as well. So instead of a glut of false positives, you have a glut of false negatives.

3 Likes

one interesting approach would be to flag, but not mark as “incorrect”, unknown words that look like proper nouns.

even in general, the idea that an unrecognized word is spelled wrong is a fair bit of chutzpah for our not very bright computing machines

that’s happened with some german names too. ( ask my ancestors how i know! ) and i think it’s interesting how the same original last name is now sometimes said by its bearers based on how it got written, rather than how it originally sounded

maybe they all gave up the good fight somewhere down the line

3 Likes

I see “flagging” vs “marking as correct” as a distinction without a difference. The whole idea of indicating an error is that the computer is alerting you to something that might be spelled wrong.

A reasonable approach might be to simply not flag anything that is capitalized. Smith? Smythe? Smydasdfasdf? It’s all good!

I don’t agree, as there is a difference between “Might be” and “is”. Maybe a yellow line when it might be a name.

Your rule is one of those simple but wrong. Teh first word of a sentence is capitalized.

2 Likes

Every flagged word is a “might-be-wrong”, as the computer doesn’t know what’s ultimately right. At best it can come close to being right, but it will never be 100% right. Even your “Teh” might be right in some contexts, but wrong in others. So “is” vs “might be” is still wishful thinking.

But if these two modes of flagging ultimately help humans better distinguish between things that are more likely to be wrong and things that are less likely, then sure, it’s a good idea! Maybe use a gradation of colors, though: bright red for “I think this is 100% wrong!” or yellowish-red for “I think this is probably wrong” or yellow for “This might be wrong.”

Agreed on the first-word-of-a-sentence; maybe just don’t flag capitalized words that aren’t the first word of sentence.

1 Like

The Most Common Irish Names and Phrases (and How to Pronounce Them) - Shamrock Gift

Irish names you’re probably saying wrong and how to pronounce them

3 Likes

yeah, it’s just that it’s presented as “this is wrong”, which leaves a very different taste in the mouth for the user than the computer asking for clarification.

i do like the sound of that. would have to be shape based as well, or leave the details till the review ( which shouldn’t be called “correction” ), because of accessibility for color blindness

3 Likes

My wife has a Japanese surname because she’s Japanese.

Under Japanese law, when a woman marries she must change her surname to that of her husband. This is done by registration on the Koseki Tohon, a largely patriarchal family register system. The wife is transferred from her father’s koseki tohon to her husband’s, and gains his surname. The problem is that foreigners don’t have a koseki tohon for the wife to be transferred to, so a Japanese woman marrying a British man like me must change her name but she can’t change her name.

The solution invented by the Japanese Ministry of Foreign Affairs for the purpose of issuing passports is that my family name is bracketed and attached to my wife’s family name. The family name in her passport reads XXX(YYY)

This is a pretty good compromise except that we haven’t yet discovered a computer system where you can enter brackets as part of your name. Thus, when she flies anywhere, her ticket is technically invalid because the name on the ticket does not exactly match the name in the passport. Yet somehow this complies with ICAA rules, or perhaps its just that the check-in desk agents have the human intelligence to ignore the bad results of the system.

To the topic! Irish names are wonderful and magical but very hard to pronounce correctly from the spelling if you don’t know the rules.

I expect that’s true of lots of names in European languages. I recently discovered my own first name’s spelling in Polish, which looks very exotic to me. It turns out to be pronounced more like French than English or German.

10 Likes

I have a bible like that.

4 Likes

Actually, the law does not require a woman to change her name to that of her husband. The law only says that spouses have to have a koseki tohon under a single surname, so it is possible for a man to take on the surname of his wife. In actual practice, though, it is as you say…

When I became a Japanese citizen, I was given the option of creating a new koseki tohon or being registered under my wife’s.

6 Likes

From my experience in Scotland, in the past when people didn’t have to deal with computers, “Mc” and “Mac”, and even “M’” (that’s “emm-apostrophe”) used to be pretty much interchangeable, with no consistency between different documents even dealing with the same individual.

In theory there’s a rule that says that when sorting, “Mc” should be considered to be equivalent to “Mac”, so “McKenzie” goes between “MacInnes” and “MacLeod”. But my computer doesn’t seem to have heard of that (even when switching the locale to en_GB).

7 Likes

Ooh, an excuse to repeat one of my favourite limericks:

There wis a young lassie named Menzies,
That askit her aunt whit this thenzies.
Said her aunt wi a gasp,
“Ma dear, it’s a wasp,
An you’re haudin the end whaur the stenzies!”

8 Likes

You’re missing out, then.

7 Likes

That’s a fun one but it doesn’t have anything to do with Gaelic. The joke comes from the fact that the Scots language (not Gaelic) used to have a special letter called “yogh” (ȝ) which to early Scottish printers looked suspiciously like a “tailed” variant of the letter “z” (Ʒ), to a point where they used the “z” in fonts that didn’t have a yogh available. Eventually the yogh fell out of use completely in favour of the “z”, as far as printers (and later the population in general) were concerned.

To be able to appreciate the limerick, one needs to be aware that the surname “Menzies” used to be spelled “Menȝies” and pronounced “Ming-is” (the rest follows from there). There are loads of proper names and place names in Scotland with “nz” or “lz” combinations which in former times would have been spelled with a yogh, and in many cases the original, not-obvious-to-foreigners pronunciation is still the appropriate one. For example, Culzean Castle, the popular Ayrshire tourist attraction, is pronounced “kul-AYN”, not “KUL-zeen”.

In the Menzies case, many people with that name will go with the traditional pronuncation just because they can, but if at the end of the 20th century you were at a UK train station or airport and asked around for the nearest John “Mingis” bookshop, people – especially outside Scotland – would probably look at you funny. (The bookshop chain has since been absorbed by its competitor WHSmith but the “Mingis” company is still around in aviation services.)

6 Likes

When I worked in a school there were often kids with two last names who didn’t know the second one. If I had to look them up on the computer, I wouldn’t be able to find them. I since learned if I was given an Hispanic name to look up, I’d use a wildcard, so instead of searching for (let’s say) Juan Zamora, I would search for Juan Zamor*. In this way, I’d get a list of Juans Zamoras. Some families consistently used hyphens, but most didn’t use them at all. But again, the wild card would find both. If someone said the name was Juan Zamora Amaya, and it had a hypen in the computer, the record wouldn’t be found if I left it out.

Sometimes whoever first registered the kid in the system would put the first and middle name in the FIRSTNAME field and the first part of the surname in the MIDDLENAME field and then the second part of the surname in the LASTNAME field. It’s easy to believe the kid isn’t registered at your school when the surname most commonly given is in the middle name field.

This can also be a problem just with paper files, as people unfamiliar with Hispanic last names will aphabetize by the second part of the surname, rather than the first part.

6 Likes

I like that “Our Father” is in bright red. In my mind, when you mouse over it a little pop-up appears: “Did you mean ‘Our Mother’?”

3 Likes

You often see it stated as an absolute rule that “Scotch” refers only to whisky, and the people are “Scots” or “Scottish”. But John Kenneth Galbraith entitled his memoir The Scotch, because that’s what all his neighbours in rural Ontario called themselves.

949169

I was reading the transcript of the trial of the H.M.S. Bounty mutineers, and I noticed the names M’Coy and M’Intosh, which was the first time I had seen that. Did it originally represent a glottal stop or something like that?

3 Likes

That could clear up some debates. One colour for “absolutely must obey”, one for “do your best”, and one for “more of a guideline, really”. That’s the way most people read the Bible anyway.

1 Like

Mc and Mac come from the word mhic, meaning son-of. The prefix O’ just means ‘of’

Also lets not forget that the Scoti originally came from Ireland, just to mess with historians I think. /s

8 Likes

I know, but I was wondering how it was originally pronounced, given that it was sometimes replaced with M’. A glottal stop is a consonant in some languages (I don’t know about Scots Gaelic).

1 Like