Oxford University Press's
Academic Insights for the Thinking World

The Seasons, part 2. From three to four, summer.

By Anatoly Liberman

The ancient Indo-Europeans lived in the northern hemisphere (see the previous post), but, although this conclusion is certain, it does not follow that they divided the year into four seasons. Our perception of climate is colored too strongly by Vivaldi, the French impressionists, and popular restaurants. At some time, the Indo-Europeans dominated the territory from India to Scandinavia (hence the name scholars gave them). They lived and traveled in many climate zones, and no word for “winter,” “spring,” “summer,” and “autumn” is common to the entire family; yet some cover several language groups.

It is rather probable that the worldview of the earliest Indo-Europeans was in part determined by a tripartite model of the universe. Julius Caesar must have divided Gaul into three parts almost instinctively. He grew up knowing three main gods: Jupiter, Mars, and Quirinus. The terrible hell dog Cerberus had three heads. An echo of the old beliefs is still distinct in epic poetry and fairy tales. The story usually revolves around three brothers or three sisters. The protagonist performs three difficult tasks. The Scandinavian gods often travel in three’s company, and so do Russian warriors. Heaven, our earthly habitat, and the underground kingdom make up another familiar triad. At some time, the “Indo-European year” may have consisted of (1) spring and summer, (2) summer and autumn, and (3) winter.

But the speakers of Greek, Latin, Slavic, and Germanic already knew four seasons. At least some of their names are still clear to us. Gothic asans (a word recorded in the fourth century, like the rest of Gothic) glossed Greek théros and meant “harvest” and “summer; heat” (English has many therm– words from this root). Its Slavic cognate (for example, Russian osen’) means “autumn, fall.” Thanks to Gothic, the Slavic word becomes transparent: not “autumn,” but “harvest.” German Ernte “harvest” and Engl. earn are related to asans ~ osen’. Perhaps gathering a crop was called simply “work”; then from “work (in the field)” to “harvest” and “autumn.” German Herbst “autumn, fall” corresponds to Engl. harvest (see again last week’s post); here English sheds light on German. It is of course more natural to associate harvest with the fall than with summer, but all depends on when summer ends and autumn begins. According to the conventional division of the year, spring consists of March, April, and May. However, “real” spring comes to us on March 23, which pushes summer to June 23, and so on.

Été by Alphonse Mucha, 1896.

Summer and its cognates dominate the Germanic languages: compare German Sommer, Dutch zomer, Old Icelandic sumar ~ sumarr, and so forth. If Armenian amarn “summer” (the transliteration has been simplified) is related to it, we get a glimpse of another sense of summer, because the Armenian word is a derivative of am “year”; summer emerges as “(time of) year.” The identification of one season with the whole year is not uncommon. In Germanic, people counted years by winters (more about it will be said in the forthcoming post on winter), but perhaps not always. In Slavic the count is by summers: leto means both “summer” and “year.” Such things are unpredictable. Our unit of time is day: she works all day, the store is open seven days a week, etc., but compare fortnight (a contraction of the original form of fourteen nights) “two weeks” and the obsolete sennight “week” (seven nights). Both words preserve an archaic and common system of reckoning by nights.

The original form of summer remains a matter of dispute. From Old English we have sumor, but, outside English this word also existed with the suffix –ar. All modern dictionaries identify the root (sum-) with Sanskrit sáma “half-year; year (!); season.” Despite the fact that this root is familiar from Latin semi– “half-,” questions remain. We are supposed to conclude that the year was divided not into three or four seasons but into two (rainy and dry? warm and cold?), that the name of the warm season became synonymous with “season as such, season in general,” and that those who used the word in question reckoned years by summers (like the Slavs). None of it is improbable. The trouble lies elsewhere. Many languages have words with the root sam– ~ sem– “half,” but the people among whom sáma was current lived on the Indian peninsula. So it is unclear why the Germanic speakers, who never inhabited those regions and had limited contacts with its population, adopted their word for “summer” and used it instead of some other Indo-European name of the season or of a more appropriate neologism. Their year was not divided into two seasons, as happens in south Asia. Despite the consensus on the etymology of summer, all is not clear, and this is the reason it might be useful to cite another etymology lost in the jungle of conjectures and hypotheses.

In the past I have had chance to refer to Francis A. Wood, who at one time was an influential scholar and the most active representative of the Chicago school of historical linguistics. His etymologies are imaginative but not always reliable, so that cautious philologists tend to reject many of them. Yet everything he said was ingenious and interesting. He compared summer and soft. The German for “soft” is sanft, in which f is secondary; the original root was san– or sam-. English soft is a modification of the form with n in the middle. According to one suggestion, sanft ~ soft has the root present in same: allegedly, soft meant “unified, with all the parts beaten into one ‘same’ mass.” But perhaps sanft ~ soft is a cognate of seem (compare Engl. beseem and seemly); in this case, soft meant not “unified” but “proper; pleasant; agreeable.” This is what Wood appears to have thought. If summer is a congener of sanft, the name of the season meant “pleasant time.”

Here we have to return to Slavic leto. Its etymology is unknown. Numerous hypotheses cancel one another out (even Swedish dialectal låding “spring” has been compared). I will cite only one of them, which has greater potential than the others. Leto may be related to a Lithuanian adjective for “mild, quiet, agreeable.” Engl. let, whose original meaning was “ease up, let up” has the same root. If this etymology of leto is right, the Slavic word turns out to be an ancient adjective derived from a verb. With Wood’s etymology as our support, we end up with the Germanic and the Slavic name for “summer” referring either to the joy people feel when this season arrives or (more prosaically and realistically) to the time when frost yields to warmth and our discontent melts like ice.

Next Wednesday will be devoted to the monthly gleanings. An etymological winter will arrive on April 4.

Anatoly Liberman is the author of Word Origins…And How We Know Them as well as An Analytic Dictionary of English Etymology: An Introduction. His column on word origins, The Oxford Etymologist, appears here, each Wednesday. Send your etymology question to him care of [email protected]; he’ll do his best to avoid responding with “origin unknown.”

Subscribe to Anatoly Liberman’s weekly etymology posts via email or RSS.
Subscribe to the OUPblog via email or RSS.
View more about this book on the

Recent Comments

  1. Marc Leavitt

    Professor Lieberman:
    I know this is a bit off the track,but I’ve always been fascinated by the traditional form of the joke: Three parts to the buildup, followed by the resolution. This protocol is by no means all-inclusive, but it predominates. You’re comments about three gods and three seasons brought this to mind. Any thoughts?

  2. […] The etymology of summer. […]

  3. John Cowan

    Marc: Jokes based on threes are Indo-European. In many Native American cultures, the “magic number” is four, not three.

  4. […] just as in the discussion of summer, one clever hypothesis has been ignored, so it is with winter. In 1960 Fritz Mezger derived winter […]

Comments are closed.