Help With The Sounds of English Words

Reading the columns

Rank

Position by frequency: rank 1 is the rime readers meet most often.

Rime

The spelling, from the vowel to the end of a one-syllable word.

IPA

The rime's usual sound, in the International Phonetic Alphabet (American English).

ARPAbet

The same sound in ARPAbet, the code used by the CMU Pronouncing Dictionary and many speech tools.

Frequency

How often the rime occurs in running English text, per million rime occurrences.

Consistency by use

Of all the times readers see this rime, the share that has its usual sound. Common words count more.

Consistency by word

The same share, counting each different word once. Use this one to judge a word family for teaching.

Family size

How many common one-syllable words share the rime and its usual sound.

Exceptions

How many common words spell the rime the same way but say it differently.

Example words

Members of the family, most common first.

Exceptions by sound

The words that break the pattern, grouped by the sound they make instead.

Who it's for

ESL learners and teachers. Rimes high on the list with consistency near 100% are safe patterns: learn -ake once and make, take, lake, shake follow. Rimes with low consistency show where English spelling can't be trusted, along with the exact words to learn by heart.

Reading tutors and phonics teachers. Tick Reliable families only, sort by family size, and print a ready-made list of word families for a lesson.

Developers and researchers. Download the whole table as CSV or JSON. The ARPAbet column matches the CMU Pronouncing Dictionary, so it lines up with many speech and text-to-speech tools.

Sorting, copying and printing

Sort

1. Choose a column in Sort by and a direction.
2. Choose a second column in Then by to order rows that tie on the first. For example, sort by family size, then by consistency by word.
3. You can also click the arrow in any column heading. Hold Shift while clicking a second heading's arrow to add it as the next sort.

Hide columns and narrow the rows

Use Columns to hide anything you don't need, or to show the ARPAbet column, which starts hidden. Hidden columns are left out of copies, downloads and prints. Type in Find a rime or word to filter: ight finds every rime containing it, and a whole word such as cake finds the rime it belongs to.

Select a block of cells

Click a cell and drag, or Shift-click, to select a rectangle of cells. Click a column heading to select the whole column, or a row number to select the whole row. Then set Take away to Selected cells only.

Copy, download or print

Copy puts the cells on your clipboard as tab-separated text, which pastes cleanly into Excel, LibreOffice or Google Sheets; Ctrl+C (⌘+C on a Mac) copies a selection too. CSV, Excel and JSON download what is shown, in its current order. Print prints it. The complete dataset is also available as rimes.csv and rimes.json.

How it was made

The table was built by a program, not by hand. It starts from the 30,000 most common English words, keeps the real everyday words, about 21,000 of them, and drops names, abbreviations and fragments.

1. Each word's pronunciation comes from the CMU Pronouncing Dictionary. Where it lists a reduced form (and said as "n"), the careful form is used.
2. A program lines up each word's letters with its sounds. Silent letters are counted with their neighbours (kn in knight), and a silent final e is linked to its vowel (a…e in cake).
3. For one-syllable words, everything from the vowel onward is the rime. Words are grouped by rime spelling, and the sound most of them share becomes the rime's usual sound.
4. Frequency is weighted by how often each word appears in English text, measured across books, news, subtitles and the web.

Limits

Pronunciations are American: -ar and -our keep their r, and some vowels differ from British, Australian and Canadian speech. Frequencies reflect adult text, not children's books. The alignment is automatic, so a small number of rows will be wrong; if you find one, the row number and rime are all that's needed to fix it.

Sources and licence

Pronunciations: the CMU Pronouncing Dictionary (BSD-style licence). Word frequencies: wordfreq by Robyn Speer, data licensed CC BY-SA 4.0. Name filtering used the Brown Corpus and the NLTK names list. Table software: Tabulator (MIT) and SheetJS (Apache 2.0).

The rime table itself is released under Creative Commons Attribution-ShareAlike 4.0. You may copy, adapt and share it, including commercially, if you credit phonics4u.org and share your version under the same licence.

Suggestion box

Found a row that looks wrong, or have an idea for the site? Tell us here. Your suggestion goes over an encrypted connection straight to this site's own server. It doesn't pass through email or any outside service.