[LH646] Learn Spelling Ignores Words With Apostrophes

Windows 10 Home Version 1803 OS Build 17134.112
Scrivener Version: 2.9.0.8 Beta (273511) 32-bit - 01 Aug 2018

I am encountering this behavior when attempting to teach Scrivener words which contain apostrophes.

For example, John’s.: If I have John and the spellcheck doesn’t recognize it, I’ll right-click and then click “Learn Spelling” and the red underline is cleared.

However, when I do the same with John’s, the red underline does not clear, leading me to suspect there’s something about the presence of the ’ which is causing the learn-spell routine to fail.

I’ve tested this with several “apostrophied” words and have gotten the same result.

Oddly, “Ignore Spelling” does learn the apostrophes.

Thanks. This has been filed.

Just an update - same issue is occurring in Beta 10. The rest of learn spelling is working perfectly, thanks for tuning it up!

Windows 10 Home Version 1803 OS Build 17134.228
Version: 2.9.0.10 Beta (331052) 32-bit - 28 Sep 2018

The problem is that Scrivener doesn’t recognise the curly quote (i.e. the typographic apostrophe) as an single inverted comma. You can fix this by closing Scrivener, opening the word list, then searching and replacing (in Notepad or similar). Changing the code for a single comma works, but you have to do it each time you add a word, so it’s a temporary solution (at best).

Hmmm… does Scrivener have the option of turning off “smart quotes” as Word does? That could be a workaround as well, and when I do final formatting &c in Word, it’s very simple to put 'em all back.

Does the “ignore spelling” actually add the apostrophied word to the dictionary, or does it just stop alerting for that word?

If you add a word that includes a typographic apostrophe, e.g. Darwin’s, it will be saved in your custom dictionary (wordlists.ini – you will usually find it in \Users[yourid]\AppData\Local\LiteratureAndLatte\Scrivener) as Darwin\xe8re’s, and the spelling check then doesn’t recognise it as equivalent to Darwin’s (straight quote). But, if you have Darwin’s (straight quote) in wordlists.ini, the spelling check recognises it as being the same as Darwin’s (tyopgraphic apostrophe).

So, one fix is to close Scrivener, open wordlists.ini using Notepad or a similar plain text editor. and search and replace, changing \xe8re into ’ – save and reopen Scrivener, and all is well, but that’s a pain (obviously) since you have to keep closing and reopening Scrivener.

So, a slightly simpler alternative is that when you type the quote, Scrivener immediately replaces it with a typographic quote. but if you hit Ctrl-Z (undo) immediately, it will change back to a straight quote. Then type the s, then add the word to your dictionary. That saves you from having to switch back and forth from autoformatting straight and curly quotes as you go.

But, obviously, this is just a temporary fix to what is likely to be a temporary problem; I assume it will be changed in a future beta.

A small update, as I’ve just observed something I hadn’t seen until now on this issue. I’ve stitched a couple screen shots together to illustrate - now, when opening the menu for learning a word, I’m seeing a ? in place of the ’ character. Not sure if this has anything to do with anything, but wanted to make sure I added it to the thread.

Apostrophe-S-Issue.png

… sure, but the point he is making is that it shouldn’t have to, and this is unnecessary extra work for the writer, because the system should simply detect words in certain categories, for which certain affixes ( prefixes and suffixes ) are validly applicable, and simply accept all such words and automatically add them to your custom dictionary ( or appending the default dictionary ).

This is a common problem with a great deal of word processing software, where it fails to understand the lexicographic rules for word construction, and so countless perfectly valid words are not included in its dictionary, which is nothing more than a simplistic and incomplete list, instead of a list of root words, morphemes, and affixes, along with the rules required to combine them << which would be vastly more sensible and sophisticated.

It’s possible such an issue is font substitution related ( where the character set mapping between fonts is not the same ), or keyboard layout related, where the keyboard mapping is not what you expected ( this happens when you want UK English dictionary, but it gives you the UK English keyboard layout along with it ) … but I cannot be sure without seeing how it all happened. Maybe someone else can clarify ( or I may be wrong and it’s something else ).

That’s a possibility I hadn’t considered. I run my keyboard in US-International mode because I regularly do things in German and Gaelige, which have all those fun accents… when I get a moment, I’ll have to reset my keyboard to stock US English and see what difference that makes. Thanks!