Automatically inserted spaces

When setting Western text, TeX automatically puts a slightly wider space after a sentence-ending period than between ordinary words. It decides “is this period the end of a sentence, or an abbreviation?” by looking at the case of the preceding letter — a guess that sometimes goes wrong. This page covers how that works, the two commands that fix it when it misfires (\ and \@), and \frenchspacing, which turns the wider sentence space off entirely.

The extra space at sentence ends

In the English / American typesetting tradition, the space at the end of a sentence is set slightly wider than the space between words (informally called “double spacing”). LaTeX follows this by default, adding extra space after a sentence-ending ., ?, or !. In 10-point Computer Modern, for instance, the interword space is about 3.33 pt, while at a sentence end TeX adds roughly another 1.11 pt on top. The sentence space is also given more stretch and shrink, so when a line is justified it is pulled wider before the ordinary word spaces are.

Internally, every character carries a “space factor.” After an ordinary character it is 1000 — the normal interword value. After ., ?, or ! it jumps to 3000, and TeX’s rule is: if the space factor is 2000 or more, add the extra width to the following space. That is the mechanism behind the wider sentence space. Spacing between Japanese (CJK) characters is governed by a different system entirely, so everything on this page is really about Western text.

How TeX guesses sentence vs. abbreviation

But if every . ended a sentence, the periods in abbreviations like Mr. Smith or etc. would get the wide space too. So TeX guesses from the case of the preceding letter. If the letter just before the period is lowercase, it treats the period as a sentence end and inserts the wide space; if it is an uppercase letter, it assumes an abbreviation or initial and uses an ordinary interword space.

That uppercase exception is implemented through the same space factor. Capital letters are assigned a space factor of 999 (not 1000), so a period following a capital cannot raise the factor to 3000, and no extra space is added. The rule of thumb “capital + period = abbreviation” works well for most English prose.

Being a heuristic, though, it misfires in two directions. First, a lowercase-then-period that is *not* a sentence end still gets the wide space (Fig. 5, etc. and). Second, a sentence that genuinely *ends* on a capital is mistaken for an abbreviation and gets too little space (...in the USA.). The next section shows how to fix each one.

Fixing the misfires (\ and \@)

(1) After an abbreviation (lowercase + period, but not a sentence end). To remove the extra space, put a backslash-space \ after the period. This inserts an explicit, non-stretching ordinary interword space: writing Fig.\ 5 gives exactly a word space after Fig. Alternatively, ~ makes the space a normal interword space *and* forbids a line break there (a non-breaking space) — good for keeping an abbreviation next to its number or name, as in Mr.~Smith or Fig.~5.

(2) A sentence ending in a capital. Because it is mistaken for an abbreviation and under-spaced, place **\@ immediately before** the period to declare “this is a sentence end.” \@ acts essentially like an invisible lowercase letter, restoring the following period’s space factor to 3000 and so the proper sentence space. \@ carries through closing parentheses and quotes, so it also works in forms like (\dots etc.\@).

A mnemonic: the positions are mirror images. For a non-ending capital abbreviation, put \ *after* the period; for a capital that *does* end a sentence, put \@ *before* it. Compiling the example below sets the spacing correctly after both etc. and USA.

We cite Fig.\ 5 and Mr.~Smith here.
Use etc.\ and so on, mid-sentence.
The rocket was built in the USA\@. It then flew.

Note that all of this matters only under the **default \nonfrenchspacing** (wide sentence spacing). If you switch to \frenchspacing in the next section, sentence spaces equal word spaces and most of this fiddling becomes unnecessary.

Turning off the wide space (\frenchspacing)

\frenchspacing makes the space after all punctuation, including sentence ends, a single uniform width equal to the interword space.** Once declared, the period case-test no longer matters, and both Fig. 5 and USA. come out naturally spaced. The command that restores the default is **\nonfrenchspacing**. You normally write it once in the preamble to affect the whole document, but placing it inside a { ... } group limits it to that scope.

\documentclass{article}
\frenchspacing
\begin{document}
No extra space after this period. Or this one.
\end{document}

The name “French” comes from the convention — in France and generally outside the English-speaking world — of not widening sentence spaces. Documents that are primarily Japanese sometimes also choose \frenchspacing to avoid uneven sentence spacing in their Western passages. In modern practice many authors simply **set \frenchspacing from the start** rather than patch individual cases with \@, and that is a perfectly reasonable choice. Note too that a language package such as babel may adjust sentence spacing automatically to suit the selected language’s conventions.