by

Proven Winners

Does it matter if things have been proved or proven? I ask this as a grammatical question, not a philosophical one about the nature of evidence. Does it matter if one uses proved or proven as the past participle of the verb prove?

If you’re in the it-doesn’t-matter camp, you’re not alone. But the it-does-matter camp is not deserted (yet).

Bryan Garner is one of the folks in the latter camp. In Garner’s Modern American Usage (third edition), he writes, “Proved has long been the preferred past participle of prove. But proven often ill-advisedly appears.” The Associated Press Stylebook 2015 implies this ill-advisedness in its guidance: “Use proven only as an adjective: a proven remedy.” (Garner does allow an exception for legal phrasing such as innocent until proven guilty.)

These opinions on proven seem tame compared with Richard Grant White, writing in 1876 in Words and Their Uses, Past and Present:

Proven, which is frequently used now by lawyers and journalists, should, perhaps, be ranked among words that are not words. Those who use it seem to think that it means something more, or other, than the word for which it is a mere Lowland Scotch and North of England provincialism. Proved is the past participle of the verb to prove, and should be used by all who wish to speak English.

The variant proven is one of those words that is not a word? This now seems far-fetched.

The verb prove was borrowed from French in the Middle English period and created the past tense and past participle like other regular verbs, by adding -ed: e.g., I proved, I have proved. The past participle proven was originally Scots (as White refers to above); the editors of the Oxford English Dictionary and the American Heritage Dictionary of the English Language speculate that it was probably created through analogy with verbs like cleave (past participle cloven) and weave (past participle woven).

Proven started to gain momentum beyond Scottish English in the late 19th century and has become significantly more popular over the course of the 20th century, as captured by the Google Books Ngram Viewer graph below.

proven copy

Just how popular? Almost as popular as proved as the past participle. A search of the present perfect construction “has/have proved/proven” in Google Books shows the construction with proved declining and with proven rising to the point where they are strikingly close in frequency (see below). A search of the Corpus of Contemporary American English (COCA) corroborates this finding for the present perfect (proven: 2,616 instances; proved: 3,001 instances).

proven_has have copy

Both constructions occur regularly in academic journals, newspapers (no matter what the AP Stylebook says), and magazines. Here’s just one example from the Journal of International Affairs: “The above mentioned measures have been implemented half-heartedly and for the most part have proven to be ineffective.”

So here we see a verb that has added an irregular past participle, and the evidence suggests that both proved and proven could be considered standard. It is a nice example of the fact that standard edited English can allow some variation without significant if any stigma (at least for many of us).

What about other verbs that have both -ed and -en past participles, such as shave (shaved and shaven) and mow (mowed and mown)? Both of these were “strong” verbs in Old English, which meant that they formed the past tense through a vowel change (the way weave/wove still does) and formed the past participle through a vowel change and final -n (the way weave/woven still does). Over time, both verbs became regular in the formation of the past tense (shaved, mowed) and are on their way to becoming regular in the past participle. Shaven still shows up regularly modifying nouns (e.g., shaven head, clean-shaven face). Mown? Well, many undergraduate students in my classes question whether this is even a word. They laugh as they say it, marveling at its quaintness. There are 122 instances in COCA, with fewer and fewer over the course of the past 20 years.

It could be tempting to think the same historical trajectory must apply to the verb sew, which similarly has two past participles right now: sewed and sewn. But sew was a “weak” verb in Old English, which meant that it formed the past tense and past participle with what we now think of as -ed.  In the Middle English period the verb developed the irregular past participle sewn, probably through analogy with sow/sown. And while both sewed and sewn are standard, the irregular sewn has overtaken sewed in the present perfect construction, as captured in the Google Books Ngram Viewer chart below.

sewn_have copy

Taken together, these verbs provide a good reminder that in the verb system, like elsewhere in the language, there can be pressure to regularize, but new idiosyncrasies are introduced into the system all the time, some of which prove to be winners.

Return to Top