How we count words
A word is a run of letters or digits, and every word of a file belongs to exactly one part, so the parts always add up to the words in the file. Type a sentence below to see it counted word by word.
The first 5,000 characters are shown word by word.
25 words, of which 7 in citations
Smith (2019) found that well-known e-mail habits don’t change: 1,000 students read 3.5 pages a day (Lee & Park, 2020, p. 4). See Figure 2 [3].
Underlined: one word each. Highlighted: inside an in-text citation, left out when you switch citations off.
Cases where counters disagree
| Text | Words here |
|---|---|
don’t | 1 |
e-mail | 1 |
well-known | 1 |
3.5 | 1 |
1,000 | 1 |
10:30 | 1 |
U.S. | 1 |
word — word | 2 |
São Paulo | 2 |
日本語 | 3 |
#2 | 1 |
https://example.org/page | 3 |
What counts as a word?
A run of letters, digits and accents. An apostrophe, a hyphen or a dot between two letters keeps them in one word, so don’t, e-mail, U.S and 3.5 are one word each, and a comma or colon between two digits does the same for 1,000 and 10:30. A dash with spaces around it, a bullet or a lone symbol is not a word. Each Chinese, Japanese or Korean character counts as one word, because those scripts do not separate words with spaces.
Which parts of a document are there?
Main text, headings, tables, text boxes, captions, in-text citations, footnotes, endnotes, headers and footers, the reference list and appendices. A word is in exactly one of them: a citation inside a footnote is a citation, a table inside an appendix is part of the appendix. The panel shows each part found, with its words and a switch, and lists the parts the file does not have.
Why do word processors disagree?
Each one decides what to leave out. Google Docs counts everything except headers, footers and footnotes. Word lets you include text boxes and footnotes from its Word Count dialog. Tools also differ on hyphenated words, numbers and dashes. Here the default leaves out notes, headers and footers, as Google Docs does, and every other choice is a switch you can see.
How are citations told apart from other brackets?
By their shape. An author–date citation has a capitalised name and a year (or n.d.) in brackets, with an optional page: (Smith, 2019, p. 4). A narrative citation is a year in brackets right after a name: Smith (2019). An author–page citation is a name and a page: (Smith 45). A numbered one is a number in square brackets: [3]. Brackets with a year but no name, a decade, a quantity or a label such as (Figure 3) are left alone.
Questions about the counting rules
Are my files uploaded anywhere?
No. The file is opened by your browser with the File API and read on your device: Word and OpenDocument files are unzipped in the page, PDFs are read by PDF.js running in the page. Nothing is sent to a server, stored or put in the address bar, so closing the tab clears it. You can check this by opening a file with your network switched off.
Is the count the same as on WordCount.tools?
Yes. Both sites use the same counting code for words, characters and sentences, and the same reading and speaking rates. WordCount.tools is the live counter for text with a platform limit, such as posts, captions and text messages; this site counts files and essays.
How are sentences and paragraphs counted?
A sentence ends at a full stop, question mark, exclamation mark or ellipsis followed by a space or the end of the text, except after an abbreviation such as Dr. or e.g. or a single capital initial. A paragraph is a block of text in the file, or a line in pasted text.
How are characters counted?
As you see them: an accented letter or an emoji made of several code points counts once. Characters with spaces include spaces and line breaks; characters without spaces leave out all whitespace. Both are counted for the parts that are switched on.
Where do the reading and speaking rates come from?
From a 2019 review of reading-rate studies by Marc Brysbaert: 238 words per minute for silent reading of non-fiction and 183 for reading aloud, averages for adults reading English.
How are headings found?
From heading styles in Word and OpenDocument files, heading tags in HTML and Markdown, larger type in a PDF, and in pasted text a short line without a full stop that is followed by a paragraph. A heading that says References, Bibliography, Works Cited or Appendix starts the reference list or an appendix.
More ways to count
- Document word countDrop DOCX, PDF, ODT or text files: counts by part, section and page.
- Essay word counterLeave out references, notes and citations, and track your word limit.
- DOCX word countA Word file counted part by part, next to the count Word saved.
- PDF word countWords per page, without running headers, page numbers or split words.
- Words to pagesPages from a word count or a file, by font, size and spacing.
Sources and limits of the count
- See word count & page count in Google Docs
“Unless you select a section of the document, word count applies to everything except headers, footers, and footnotes.”
Google Docs Editors Help. Opened on .
- Word count
The Word Count dialog on the Review tab shows more detail and has the option to include text boxes and footnotes in the total.
Microsoft Support. Opened on .
- Is my list of references included in my word count?
The reference list or bibliography is not included in the word count; in-text citations are included; in footnote-based styles the footnotes are not included. There may be exceptions: check the module handbook.
University of Hull Library. Opened on .
- How do I exclude the in-text citations from my word count?
You usually will not have to exclude in-text citations; the Postgraduate Research Regulations exclude citations from the word length. Its Word method removes all text in round brackets.
Lancaster University Library. Opened on .
- ECMA-376, Office Open XML File Formats
ECMA-376, Office Open XML file formats, the standard behind .docx, including Part 2, Open Packaging Conventions: the zip package that holds the parts of a document.
Ecma International. Opened on .
- Open Document Format for Office Applications (OpenDocument) v1.3
The schema behind .odt, which defines headings (text:h), notes (text:note), comments (office:annotation) and the document statistics element (meta:document-statistic).
OASIS. Opened on .
- PDF.js
“A general-purpose, web standards-based platform for parsing and rendering PDFs.” Used here to take the text of a PDF, with the position of each piece, out in the browser.
Mozilla. Opened on .
- How many words do we read per minute? A review and meta-analysis of reading rate
Average adult rates in English: 238 words per minute for silent reading of non-fiction, 260 for fiction, 183 for reading aloud.
Brysbaert (2019), Journal of Memory and Language 109. Cited reference: we could not open the page on the day of our check, so it is listed as the published reference.
Word processors and course rules count some parts differently, so treat the total as a check and your course handbook or publisher as the rule. Files are read on your device: nothing you drop or paste is uploaded.