Bug with title and page number in citations using RTL languages (Arabic and Hebrew)

scholarsgrotto

Suppose you cite a book written in Arabic or Hebrew by Jane Doe using Chicago style. Everything looks fine. You then cite a different book by Jane Doe. Still good. If you next want to cite the first book again, however, you run into a problem: this subsequent citation of the first book inverts the position of the book title and the page number in the citation. See attached image for what I mean.

This appears to be a Zotero issue, not a CSL issue, because it happens for both Chicago and MLA and for both Hebrew and Arabic. What seems to be happening is that the page number is somehow getting "glued" onto the title field, and although the page number is getting placed at the end of the title, this actually puts it to the left of the title since the title is in a RTL language. Does that make sense? Again, take a look at the image.

Note: MLA will actually exhibit this behavior on the first mention of a book if you have more than one work by the same author cited anywhere in the document.

Note: For some reason, the title and page number appear as font size 12 in a font-size-11 document (while the Author name remains properly in font size 11). This is a piece of evidence that leads me to believe book title and page number are getting glued together. In any case, I don't know why the font change is happening and seems to be an additional bug.

https://s3.amazonaws.com/zotero.org/images/forums/u3061023/lrmqnz4vhyppjoji9ie3.png

scholarsgrotto

Just checking in to see see if anyone has encountered this issue with Arabic and other RTL languages?

AbeJellinek

Arabic characters switch the text direction to RTL, and numbers and punctuation are neutral, so they appear to the left of the RTL text. Word doesn't give you any good way to fix the ordering in mixed LTR/RTL lines, but you may be able to improve it by forcing the citation footnotes to be either LTR or RTL (see Word documentation). That will likely prevent the citations from being updated by Zotero, so you should only do it on a copy of your document as you're finalizing your citations.

Font size issue: Apparently Word allows styles to set different default font sizes for different scripts, so try editing the style being used for footnotes in your document and see if "Complex scripts" have a different size than "Latin".

scholarsgrotto

Thanks, AbeJellinek. So, taking your cue, I found that if I insert a Latin letter in front of the page number in the page number field, Word correctly orders the elements of the citation, implying that numbers are indeed seen as "neutral" with respect to text directionality.

HOWEVER, if I manually insert spaces into the citation I find that I also get the correct ordering of elements, so there must be a structural solution to the problem. Looking into the underlying XML of the document, I see that this is indeed the case. In the picture below, compare the incorrect and correct appearance of the citation, as it is rendered in Word and as it appears in the underlying XML:
https://s3.amazonaws.com/zotero.org/images/forums/u3061023/elxtkefj90faeqjgxh6z.png

This clearly shows that the root of the problem is Zotero outputting the page number inside the w:r element and the w:t sub-element along with the title.

The fix would be simply to output the comma and page number in their own separate w:r elements, no?