You are here: Encoding > Unicode
Short URL: http://scripts.sil.org/Unicode5QuoteMirroring

Why are my quote marks backwards?

Bob Hallissy, 2009-03-02; 8330 reads

If you work with right-to-left text in Unicode and have certain quote marks in your text, then this article is for you. More specifically: if you suddenly see your smart quotes reverse direction, this article will explain why and what to do (or not do) about it.

What you may experience

You have right-to-left text that uses one or more of the characters from the list below. After upgrading some software on your computer, or viewing the text on another computer, the quote marks appear reversed.

U+2018 LEFT SINGLE QUOTATION MARK
U+2019 RIGHT SINGLE QUOTATION MARK
U+201A SINGLE LOW-9 QUOTATION MARK
U+201B SINGLE HIGH-REVERSED-9 QUOTATION MARK
U+201C LEFT DOUBLE QUOTATION MARK
U+201D RIGHT DOUBLE QUOTATION MARK
U+201E DOUBLE LOW-9 QUOTATION MARK
U+201F DOUBLE HIGH-REVERSED-9 QUOTATION MARK
U+301D REVERSED DOUBLE PRIME QUOTATION MARK
U+301E DOUBLE PRIME QUOTATION MARK
U+301F LOW DOUBLE PRIME QUOTATION MARK

Background

One of the properties of characters in the Unicode standard is the "mirrored" property: If a given character has this property, then the graphical image must be adjusted in right-to-left contexts. For example, the parenthesis characters U+0028 LEFT PARENTHESIS and U+0029 RIGHT PARENTHESIS have this property so that U+0028 is always used for opening parenthesis, and it is the responsibility of the software to turn its image around (to "mirror it") if the text flow is right-to-left.

In an attempt to tighten up the Unicode standard with regard to mirroring, a number of existing characters were given the mirrored property. (For more background, see Public Review Issue #80). Unfortunately the characters listed above were included in this change. That is, these characters went from not being mirrored (in all versions of the standard prior to 5.0), to being mirrored in version 5.0. This means that any software that is upgraded to Unicode 5.0 conformance will now start mirroring these characters in right-to-left contexts, which is why you may see your quote marks reversed!

However, before you go "fixing" your data, please read on. This will be a temporary blip on Unicode's historical lifeline: a corrigendum has already been issued to undo this change.

Software affected

The list of affected software is incomplete at this time — watch this space for updates.

For the Windows platform, a lot (but not all) programs obtain their Unicode processing via the system component called "Uniscribe" (see sample list). Some other applications license Uniscribe from Microsoft — even when running those applications in pre-Vista environments.

Microsoft implemented Unicode 5.0 conformance in the versions of Uniscribe that shipped with the initial releases of Windows Vista and also Office 2007.

At this date, I believe all the following are affected:

Microsoft Vista (pre SP1)
Microsoft Office 2007
Paratext 6.1

In particular, the versions of Uniscribe that we know implement the 5.0 behavior have version numbers starting with 1.626. But see below.

Is this being fixed?

The Unicode Technical Committee (UTC) has taken the unusual step of issuing an official corrigendum that, in effect, undoes the change. See Corrigendum #6.

With regard to the mirroring of these characters, then, Unicode 5.0 is a temporary anomaly: all prior versions did not mirror them, future versions will not mirror them, and Unicode 5.0 With Corrigendum 6 Applied does not mirror them. Only Unicode 5.0 (without Corrigendum 6) does it, and only software that implements Unicode 5.0 without Corrigendum 6 will be impacted. The number of products affected are limited and are likely to get updated for Corrigendum 6.

Microsoft incorporated Corrigendum #6 into Vista Service Pack 1 (SP1). The earliest verison of Uniscribe that has Corrigendum #6, that I am aware of, is 1.626.6001.18000.

What should I do? Should I change my data?

Based on the above discussion, it should be clear that you should not change your data just to make it look right when using Unicode 5.0-conformant software — doing so will only make things worse in the long term.

Basically the rule is: if your work involves right-to-left scripts, you should avoid using any version of Uniscribe from 1.626.0.0 up to, but not including, 1.626.6001.18000. Preferably upgrade to a later version. If that is not possible, downgrade to a previous version.

Other Uniscribe information

Uniscribe versions
How do I tell what version of Uniscribe I’m using?
http://www.babelstone.co.uk/Software/BabelPad.html also has useful information. Search for “Uniscribe Issues”.
Wikipedia: Uniscribe versions
Introduction to Uniscribe
Rendering technologies overview

Add a response to this article

Note: the opinions expressed in submitted contributions below do not necessarily reflect the opinions of our website.