"Lept" vs. "leapt" vs. "leaped"

I found this quote here by Pam Peters (The Cambridge Guide to English Usage), and she says:

leaped or leapt: Both spellings are used in American and British English, but the relativities work in opposite directions. In the US the more regular spelling leaped is commoner by a factor of 5:1, according to CCAE¹ data. In British data from the BNC², leapt is ahead by similar ratio.

Based on this, I think that you can use 'leaped' or 'leapt'; both are acceptable past-tense and past-participle forms for the verb leap. As Bryan A. Garner says³: “Because leapt is pronounced /lept/, the mistaken form lept is frequently encountered.”

¹_{CCAE: The Cambridge International Corpus of American English}

²_{BNC: British National Corpus}

³_{A Dictionary of Modern American Usage}

Almost ten years ago, I checked and recorded the relative popularity of leaped and leapt in Google results, because the following assertion in Webster's Dictionary of English Usage (1989) intrigued me:

In American English, leaped and leapt seem to be used with about equal frequency.

Recognizing that leapt is somewhat more common in British (and much of British Commonwealth) English than in U.S. English, I decided to run Google searches for the two terms, reasoning that if leaped came out significantly ahead in undifferentiated results, it was probably considerably more popular than leapt in U.S. usage.

My first round of searches for leaped and leapt (undated, but probably in 2004 or 2005) generated 6.2 million matches for leaped and 4.7 million matches for leapt. A second run, on 11 August 2010, produced these unexpected numbers: 8.73 million for leaped, 8.53 million for leapt, and 401,000 for lept. Today (22 January 2013), I ran the searches again and got these numbers: 7.06 million for leaped, 10.5 million for leapt, and 1.11 million for lept.

With regard to absolute numbers, Google searches are utterly unreliable because they generate so many repetitious matches for the same content. But for indicating shifts in relative popularity, I think that they can be useful. And in this case, a rather impressive shift in such popularity is evident—away from leaped and toward leapt. The numbers also suggest that lept is very far from vanishing as a variant spelling, though Merriam-Webster Online as yet declines to accept it as one.

An Ngram Viewer search for leaped, leapt, and lept confirms my initial impression/assumption that leaped is more common in published print content than leapt—and indeed indicates that it has been consistently so for more than 200 years. However, the rise of leapt in Google results and the absence of copyediting in much published work these days suggest that leapt may soon vindicate Webster's Dictionary of English Usage's prematurely enthusiastic assessment of its popularity in the United States.

---

Followup (April 4, 2021)

This was one of my first answers on English Language & Usage—and I didn't yet know how to reproduce Ngram charts in answers on this site. I also was unaware that Ngram lets users break out results for British English and American English. With this knowledge now in hand, I am belatedly revisiting this answer to provide the relevant Ngram results.

First, here is the Ngram chart for all published occurrences of leaped (blue line) versus leapt (red line) versus lept (green line) for the period 1750–2019:

Next, here is the Ngram chart for the frequencies of the same three words over the same period in books and periodicals from British publishers:

And finally, here is the Ngram chart for the frequencies of the same three words over the same period in books and periodicals from North American publishers:

The most striking thing in these results is the evidence they provide that leaped was significantly more common than leaped in British publications from 1750 until the early twentieth century (the crossover in British preferences from leaped to leapt technically occurred between 1916 and 1917, according to Ngram). What caused this change? Writing in 1926, Henry Fowler, A Dictionary of Modern English Usage attributes it to a tendency (which he commends) to match spelling to pronunciation:

-T & -ED. Typical words are bereaved & bereft, burned & burnt, dreamed & dreamt, kneeled & knelt, leaned & leant, leaped & leapt, learned & learnt, smelled & smelt, spelled & spelt, spilled & spilt, spoiled & spoilt, tossed & tost.

In the last of these the point is purely one of spelling, & the sound is the same either way ; there are many other verbs of which that is true (husht, kist, whipt, curst, cookt, &c.), & individual writers make a practice of using the short form as a piece of spelling reform, as time-saver, or an eccentricity ; whichever the motive, the effect is with most words eccentric ; but tost, esp. in p. p. compounds such as storm-tost, is current, by the side of tossed.

Of the rest the spelling may affect the sound in some, & does affect it in others. Thus, burned may be sounded with d, but perhaps most even of those who spell it so sound it with t, whereas leaped & leapt are pronounced by everyone with different vowels—lēpt & lĕpt. The advice here offered is to use the -t spelling in both classes, & that in the face of the surely surprising figures to be given below ; it will hardly be denied that most people say bernt & lĕpt, not bernd & lēpt, & conformity between the written & the spoken word is worth securing where, since both spellings are already in use, it costs nothing. At present, however, the -ed forms still prevail in print over those in -t in most of our list ; and it should be added that, if the past tense were distinguished from the p. p., the preponderance of -ed for it would be slightly greater. The figures are arrived at by counting the occurrences in all OED quotations of the 19th and 20th cc. ...

Fowler then goes on to note (among other comparisons) seven relevant OED matches for leaped and five for leapt.

Writing almost forty years later, Ernest Gowers, A Dictionary of Modern English, revised edition (1965) updates us on the progress of relevant British English spelling preferences:

-t and -ed Some verbs that once had an alternative -t form for past tense and participle have now lost it, either wholly or to an extent that makes it archaic, e.g. curst, dropt, husht, kist, stopt, tost, and whipt. For many others -t is now the only formation, e.g. crept, dealt, felt, kept, left, meant, slept, swept. Typical of the words that have preserved both alternatives are bereave, burn, dream, kneel, lean, leap, learn, smell, spell, spill, and spoil.

In the first edition of this dictionary an attempt was made to assess the comparative popularity of the two endings when used in print by counting the number of times they occur in all OED quotations of the 19th and 20th cc. ...

So far as we can accept thee figures as a guide, it would seem that the -ed forms then still prevailed in print; ... But there has since been a movement (advocated in the original edition) towards -t. Lean is the only word in the first group [which includes leap] for which the C[oncise] O[xford] D[ictionary] still puts -ed before -t; for kneel it admits no alternative to knelt.

In the North American English, meanwhile, the -ed forms have long been viewed as preferable (or at least dominant). From Margaret Nicholson, A Dictionary of American-English Usage (1957), based on the first edition of Fowler:

-t & -ed. ... In US the -ed ending is much more general, esp. in writing; in Brit., the -t.

Bryan Garner, Garner's Modern American Usage, second edition (2003) has a somewhat surprising take on leaped versus leapt:

leapt; leaped. Both are acceptable past-tense and past-participial forms for the verb leap. Because leapt is pronounced /lept/, the mistaken form lept is frequently encountered ...

Leapt, which used to be the more common form, is steadily being displaced by leaped: in frequency of use, the two forms are neck-and-neck in modern print sources. Traditionalists prefer leapt.

The Ngram chart for American English shows something quite different from what Garner describes. Although it does show a considerable narrowing of the distance in frequency between the two forms from roughly 1900 until about 1975, that narrowing involves a reduction in the advantage of leaped over leapt—not the reverse. And since 1976, despite the fact that both forms have seen impressive increases in frequency of use, the difference between the two has remained remarkably steady.

One might also ask what tradition traditionalists are basing their preference for leapt on. As Fowler noted, and as the Ngram charts tend to confirm, traditional spelling in both Britain and the United States has generally favored leaped by a very substantial margin. It would perhaps be more accurate to say, as Fowler does, that spelling reformers who favor the pronunciation lĕpt over the pronunciation lēpt clearly prefer the spelling leapt.

Several cautions about the Ngram results may be in order. First, the fact that a book has a British or North American publisher doesn't mean that its spelling has been altered to suit the publisher's preferences. If the book is by a British author—especially if it is a scientific or literary work, it may retain the author's original British spelling choices, and likewise (in reverse) for a British publisher's version of a U.S. author's work.

Second, the "American English" evidently doesn't distinguish between U.S. and Canadian publishers, although U.S. and Canadian spelling preferences differ on a number of points, including (potentially) some difference in degree of preference for leaped versus leapt. Margery Fee & Janice McAlpine, [Oxford] Guide to Canadian English Usage, second edition (2007) note that Canadian usage on -ed versus -t verb endings aligns more closely with U.S. usage than with British usage on many words, although with some degrees of difference on others:

-t, -ed Leant, learnt, smelt, spelt, spilt, and spoilt are chiefly British forms of the past tense and past participle forms. Canadians, like Americans, are far more likely to use the -ed forms of these verbs. However, Canadians do use burnt, dreamt, knelt, and leapt quite often.

From this account, it is hard to tell whether leapt is more common in Canadian usage than in U.S. usage—but it would be unreasonable to rule out that possibility.

And third, Ngram results are subject to various optical character recognition errors, including misreading the spelling of the word in the original text and misreading the publication date of the cited book or periodical. Consequently, the more precisely one attempts to assert the accuracy of the percentage data that provide the basis for the Ngram line graphs, the more dubious the legitimacy of one's assertions becomes.

Still, in broad outline, the lines plotted for leaped and leapt over the period from 1750 to 2019—and especially since 1800 or so, appear to provide a reasonably reliable picture of the relative frequency of the two spellings in books and periodicals published by British, American, and combined publishing houses.