Automatically curly


Quotation marks: single or double

Quotation marks can be used to call special attention to words and phrases. For example, when text is quoted from another speaker or writer, the quotes distinguish it. Whether single quotes are employed for this purpose, and double quotes for other uses, or the other way round, or the same type for both, is a matter dealt with in style guides.

This also varies by language (e.g. Dutch may be different from English), by country (e.g. English in the United Kingdom or the United States, Dutch in the Netherlands or Flanders) and even by newspaper or book editor.

Straight or curly

ASCII and more

The simplest quotation marks are the straight quotes in ASCII. ASCII is a plain and simple character set, consisting of only 128 characters including punctuation and control characters. These straight ASCII quotes (single ' or double " ) are simple and clear, and always available. But in this day and age of powerful computers, multiple fonts and Unicode, they are a bit, shall we say, primitive.

A notable feature of ASCII quotes is that there is no difference between quote and unquote. The same signs ( ' or " ) are used for both. In case of single quotes, the opening quote is sometimes the ASCII backquote ( ` ), but this is not universally adhered to. For double quotes, there is no such distinction at all.

Apart from the fact that using the same sign for quote and unquote doesn't look very good, it also can make a conversion to more advanced characters difficult. But there's a solution to that.

Languages and countries revisited

Curly quotes and other special characters look much better. They have been used in printed books and newspapers for ages, but they are also available for computer texts, on the screen and on paper.

Details on which kinds are used, again depend on country, language, and even newspaper or a book editor's corporate design.

This Swiss site gives a lot of information about this. Some remarkable points:

Word's Autocorrect function

As you type

Straight ASCII quotes are the same at the beginning as at the end of the quoted material. Smart, curly, typographical quotes usually have different starting and ending shapes. That means that any automatic conversion needs to have a certain intelligence, it must consider the context.

Microsoft® Word offers such an automatic conversion, and it works quite well. You simply type straight quotes ( " ), and Word automatically places the correct round quotes, taking the language and the situation into account. The same works when typing single quotes ( ' ).

The following is based on a Dutch language version of MS Word 97, but I expect it to work with newer and other editions of that software too.

(Note that I am often rather critical of Microsoft® software, including Word, although I think that is one of their best products. But when there is reason for praise, I openly express that too.)

For MS Word's automatic conversion to work, you need to activate the “Autocorrect” function. This is done as follows:

Note: I usually disable all the other Autocorrect options in Word. But of course this is a matter of personal preference.

Language setting

The language setting of the text influences how the Autocorrect function changes straight quotes into curly ones.

You can change the language code as follows:

Quotes are correctly converted for English, Dutch (but see Apostrophe), and German.

With the language set to French, type straight double quotes correctly become French guillemets, each including the non-breaking space which is required in this language, to separate certain punctuation from the text ( « Quoted text » ). Word then also automatically inserts a non-breaking space before a colon ( : ) and semicolon ( ; ).

Typed single quotes in a text marked as French are not converted to single guillemets ( ‹ › ) but to round single quotes ( ’ ), which is correct for things like « c’est ».



Word's Autocorrect can also be used to obtain apostrophes simply by typing a single straight quote ( ' ). This is useful in English (it’s, don’t, John’s shoes), Dutch (opa’s, tv’tje, A’dam, m’n) and French (the aforementioned c’est).

In Dutch words like ’s ochtends, ’s morgens, ’s middags, ’s avonds, ’s nachts (meaning in the morning (twice), in the afternoon, in the evening, at night) the ’s is a variant of an older word “des”, the genitive of the article “de”.

It also occurs in geographical names like ’s-Gravenhage (also Den Haag, meaning The Hague) and ’s-Hertogenbosch (also Den Bosch, French Bois-le-Duc).


Word 97 doesn't handle these Dutch initial apostrophes correctly: it makes them 6-shaped ( ‘ ) instead of the correct 9-shaped ( ’ ).

A way to overcome that difficulty is to type Ctrl-'' (i.e., hold the ctrl key and type the single quote ' twice). Or you can copy and paste a correct round quote from somewhere else in the text, if it occurs there.

You could also type a redundant letter before the quote (x's-Hertogenbosch) and remove it later. This will persuade Word to use the correct type of curly quote for this apostrophe.

In search and replace

Once Autocorrect has been set up to convert from straight to curly as you type, it also works when doing a search and replace. Strangely enough though, the language code of the text doesn't influence this, and the conversion is only done from straight to curly, and from curly to curly. French guillemets and German quotes are not properly handled in this case.

When searching for straight quotes ( " ), curly quotes are also found. So Word finds ", “ and ”. When replacing these by a straight quote (so in fact, you ask Word to convert " to ", and seemingly make no change at all), the conversion is done in the process. So what you get is curly quotes, i.e. opening and closing as appropriate, taking context into account.

This opens the following possibilities:

Functional symbols?

So far we discussed quote characters in Word. They can also be made in HTML using symbolic entities or references to Unicode characters.

Examples of such entities are:

“left double curly quote
”right double curly quote
‘left single curly quote
’right single curly quote
„bottom double curly quote
‚bottom single curly quote

See also the aforementioned Swiss document and under Special Characters.

In both cases (Word and HTML), what is symbolically encoded, is still the wanted character itself. I think a better solution would be to symbolically encode the function of the characters, not their shapes. So you should be able to indicate you want to start or end a quotation, a quotation within a quotation, or put special attention to a word or expression.

Then in CSS entries (for HTML) or styles (Word) you should be able to specify which style you want to employ, with defaults for various languages, and the possibility to fine-tune the details and deviate from the defaults.

Presentation software (browsers, Word display and print functions) would then make the characters appear as they should, without explicitly having to encode this in the source text.

Perhaps such solutions already exist, but I am not aware of them. I didn’t look very well, though.

Copyright © 2008 by R. Harmsen.

Colours: Neutral Weird No preference Reload screen