What is the difference between a widow and an orphan?

A widow is the last line of a paragraph pushed alone onto the top of the next page. An orphan is the first line of a paragraph left behind alone at the foot of a page. The names are swapped constantly, including in software documentation. The mnemonic compositors use is that an orphan has no past, so it is left behind at the bottom, and a widow has no future, so she goes on alone.

Does Word's widow and orphan control fix the problem?

Not on its own. It is on by default and it does prevent the single stranded line, but avoiding a widow means moving a line, and moving a line leaves the page one line short. A word processor has no instruments to absorb that, because it is filling a column rather than composing a spread. A typesetting engine can tighten the leading fractionally, adjust the word spacing, or let the page run short while keeping its facing page in step.

How many lines should fall on each side of a page break?

At least two. The conventional rule is that no paragraph is broken so that fewer than two lines fall on either side of the break, and it is the default a typeset book follows.

What else besides widows and orphans decides where a page breaks?

A heading is never the last thing on a page, so if the paragraph it introduces cannot follow it, the heading moves. Facing pages end on the same line, because a spread where the left page runs one line longer than the right reads as crooked. And a chapter's last page carries more than one or two lines. Each of these is a constraint on where a page may break, and applying them all at once across hundreds of pages is what makes typesetting a craft.

What widows and orphans are, and why your book has them

Open any book published by a house you recognise and look at the top of a left-hand page. You will not find a single line of a paragraph sitting alone up there, with the rest of its paragraph on the page before. Open a manuscript printed from a word processor and you will find one every few pages.

That stranded line has a name. It has had one since at least the nineteenth century, when the person setting the page was the person who noticed it.

The two of them, and which is which

A widow is the last line of a paragraph, pushed onto the top of the next page alone. An orphan is the first line of a paragraph, left behind at the foot of a page while the rest of the paragraph moves on.

The two names get swapped constantly, including in software documentation. The mnemonic compositors use: an orphan has no past, so it is left behind at the bottom; a widow has no future, so she goes on alone.

Both are the same underlying event: a paragraph was broken across a page boundary at a place that leaves one line by itself. And both look like a mistake to a reader who could not tell you why.

A page is a unit of attention. A line that arrives without the paragraph it belongs to asks the reader to hold something in mind across a physical gesture, and gets a fraction of a second of confusion in return.

Why word processors produce them

Word does have a widow and orphan control, and it is on by default. It is also, on its own, not enough. Avoiding a widow means moving a line, moving a line means the page is now one line short, and something has to absorb that. A typesetting engine has choices about where: it can tighten the leading fractionally across the spread, adjust the word spacing, or let the page run one line short and keep its facing page in step.

A word processor has none of those instruments, because it is not composing a spread. It is filling a column and starting another one when that column is full.

What a compositor actually does

The rules a book follows are not complicated, and they are why professionally set pages feel different before you can say why:

No paragraph is broken so that fewer than two lines fall on either side of the break.
A heading is never the last thing on a page. If the paragraph it introduces cannot follow it, the heading moves.
Facing pages end on the same line. A spread where the left page is one line longer than the right reads as crooked, even to someone who never notices type.
A chapter’s last page carries more than one or two lines. A single line of prose under a running head is not a page; it is a mistake that got printed.

Each of these is a constraint on where a page may break. Applying them all at once, across hundreds of pages where fixing page 40 changes page 41, is the work that makes typesetting a craft rather than a setting.

What this looks like in Nueel

Nueel’s engine treats page breaking as a constraint problem rather than as a column fill. Widow and orphan thresholds, keep-with-next on headings, and recto and verso progression are all part of the same pass, so a fix on one page cannot silently create a defect on the next. The engine reconsiders what follows.

You can set the thresholds yourself, and the defaults are the conventional ones: at least two lines on each side of a break, headings kept with their first paragraph, chapters opening on a right-hand page.

A reader will notice the absence of widows, not the widows themselves. It arrives as the sense that this is a real book, made by someone who was paying attention.

Common questions

What is the difference between a widow and an orphan?: A widow is the last line of a paragraph pushed alone onto the top of the next page. An orphan is the first line of a paragraph left behind alone at the foot of a page. The names are swapped constantly, including in software documentation. The mnemonic compositors use is that an orphan has no past, so it is left behind at the bottom, and a widow has no future, so she goes on alone.
Does Word's widow and orphan control fix the problem?: Not on its own. It is on by default and it does prevent the single stranded line, but avoiding a widow means moving a line, and moving a line leaves the page one line short. A word processor has no instruments to absorb that, because it is filling a column rather than composing a spread. A typesetting engine can tighten the leading fractionally, adjust the word spacing, or let the page run short while keeping its facing page in step.
How many lines should fall on each side of a page break?: At least two. The conventional rule is that no paragraph is broken so that fewer than two lines fall on either side of the break, and it is the default a typeset book follows.
Why do widows and orphans matter if most readers cannot name them?: Because a reader notices their absence rather than their presence. A page is a unit of attention, and a line arriving without the paragraph it belongs to asks the reader to hold something in mind across a physical gesture. What they perceive is not the defect itself but the sense that this is a real book, made by someone who was paying attention.
What else besides widows and orphans decides where a page breaks?: A heading is never the last thing on a page, so if the paragraph it introduces cannot follow it, the heading moves. Facing pages end on the same line, because a spread where the left page runs one line longer than the right reads as crooked. And a chapter's last page carries more than one or two lines. Each of these is a constraint on where a page may break, and applying them all at once across hundreds of pages is what makes typesetting a craft.