How many minimal pairs are there in English?
Total number of Minimal pair in English is 92,253 pairs That is the total number of minimal pairs identified in this dictionary.
I cannot guarantee the accuracy of this figure. Many editorial decisions about including or excluding certain words were made arbitrarily and may be modified or reversed, and I know there were a few mistakes in transcription. However this is a fair working approximation that should not change by more than 0.2% either way.
32,722 different words appear somewhere in the list of minimal pairs out of the 70,646 in the whole of the Mitton dictionary. The commonest words to enter pairs are the homographs READ (75), BOW (71) and LEAD (70). The highest placed non-homograph is BOARS/BORES with 67 , followed by LAW/LORE (58). The high placement of LAW is partly accounted for by its entry into compounds such as brother-in-law . At the other end of the scale about 12,200 words enter into just one pair. As one would expect the majority of the paired words are short. Of the 8,472 monosyllables in the dictionary, all but 47 appear in pairs. The 47 "unmarried" monosyllables are
albs | months | smith | strength |
blasts | ninths | smiths | strengths |
blimp | opts | smooth | swifts |
blitzed | plebs | smooths | tempts |
chefs | plunge | smoothed | texts |
clefts | pranced | smudged | twelve |
cusp | quizzed | spleen | twelfth |
Delft | scripts | squelch | twelfths |
depth | scrunch | squelched | twixt |
depths | scrunched | squint | |
eighths | sixth | squints | |
Gwent | sixths | squire |
In addition there is the word ask which was paired with the River Usk . The dictionary did not find a pair for it outside proper names.
Enlarging the size of the vocabulary might increase the total of minimal pairs by a few thousand, but a law of diminishing returns applies. Obscurer and more specialised words are likely to be polysyllables which are less likely to enter minimal pairs. This dictionary contains 26,166 two-syllable words, 21,286 three-syllable words and 14,721 longer words. I am not proposing to count their distribution among the lists, but assume that most of the three-syllable and longer words are unpaired, although a number of polysyllabic verbs are included in the list which contrasts /z/ with /ŋ/ in inflections ( buzzbung.html ), such as vulcanizes/vulcanizing . The longest pairs my wife and I have noticed are nationalization/rationalization and undramatically/ungrammatically with six syllables each.