Why PDF copy-paste splits words with hyphens

The specific damage

Copy a paragraph out of a PDF and you frequently get fragments like this:

The depart-
ment reviewed the require-
ments before publi-
cation.

Three words broken, each with a stray hyphen and a line break in the middle. Search will not find "department", spell check flags every fragment, and a word count is wrong.

Why the hyphen is there at all

Justified text needs even margins on both sides. When a long word will not fit at the end of a line, the typesetter splits it at a syllable boundary and marks the split with a hyphen. This is centuries old and entirely normal in print.

That hyphen is a rendering artefact. It means "this word continues on the next line" and should vanish the moment the text reflows to a different width.

In a word processor it does exactly that, because the file records the word and calculates line breaks on the fly. A PDF records the opposite — the finished appearance — so the hyphen is stored as a literal character sitting between two literal line fragments. There is no flag distinguishing it from a hyphen someone typed on purpose.

Why "remove all hyphens" is the wrong fix

The tempting fix is to delete every hyphen-plus-newline. It works until it meets a compound word that happened to land at the end of a line:

• well-being → wellbeing
• part-time → parttime
• e-mail → email
• twenty-five → twentyfive
• X-ray → Xray

These are silent corruptions. Nothing is flagged, the text reads almost normally, and the error is only caught by someone who notices the specific word.

The rule that actually works

One heuristic separates the two cases most of the time: **what follows the line break**.

A hyphenated split leaves the remainder of the word on the next line, so it begins with a lowercase letter. A genuine compound at a line ending is normally followed by a complete word, which in running prose also begins lowercase — but the crucial difference is that a typesetter only ever splits *within* a word, never after a real hyphen it could keep intact.

In practice the reliable rule is: join when a hyphen is immediately followed by a newline and then a lowercase letter, because that is the signature of a mid-word split. Leave the hyphen alone in every other position.

This is why a general-purpose find-and-replace does poorly here and a purpose-built cleaner does well. The rule depends on position and on what character follows the break — neither of which a plain search for "-" can see.

It is a heuristic, not a proof. A compound word split across lines exactly at its own hyphen will still be joined incorrectly. That case is rare, and no automatic method can resolve it without a dictionary, because the two are genuinely indistinguishable in the file.

Soft hyphens, the invisible variant

Some PDFs use a soft hyphen (U+00AD) instead of a regular one. A soft hyphen displays only when a word actually wraps, and is invisible otherwise.

Copied out, it may arrive as an invisible character sitting in the middle of a word. The text looks perfect and still fails a search, because "depart\u00ADment" is not "department". If a word looks right but cannot be found, this is a likely cause.