By Anatoly Liberman
Very long ago, one of our correspondents asked me how irregular forms like good—better and go—went originated. Not only was he aware of the linguistic side of the problem but he also knew the technical term for this phenomenon, namely “suppletion.” One cannot say the simplest sentence in English without running into suppletive forms. Consider the conjugation of the verb to be: am, is, are. Why is the list so diverse? And why is it mad—madder and rude—ruder, but bad—worse and good—better? Having received the question, I realized that, although I can produce an inventory of suppletive forms in a dozen languages and know the etymology of some of them, I am unable to give a general reason for their existence. I consulted numerous books on the history of the Indo-European languages and all kinds of “introductions” and discovered to my surprise that all of them enumerate the forms but never go to the beginning of time. I also turned to some of my colleagues for help and came home none the wiser. So I left the query on the proverbial back burner but did not forget it. One day, while feeding my insatiable bibliography and leafing through the entire set of a journal called Glotta (it is devoted to Greek and Latin philology), I found a useful article on suppletion in Classical Greek. Naturally, there were references to earlier works in it. I followed the thread and am now ready to say something about the subject.
This introduction might seem unnecessary to our readers, but I have written it to point out two things. First, it sometimes happens that finding an answer to what looks like an elementary question proves a difficult enterprise. Second, the episode has a sobering aspect. The main work on the origin of suppletion is a “famous” book written more than a hundred years ago, and it had important predecessors. “Everybody,” as various authors say, knows it. Well, apparently, the book’s fame is not universal, and one can devote long years to the study of historical linguistics and stay outside the group defined by the cover term “everybody.” Nothing like a question from a student, friend, or reader to prick one’ vanity! And now to business.
Regular forms exist in both grammar and word formation. For instance, many languages use a special suffix to derive the name of a feminine doer from its masculine counterpart. Thus, German Freund “(male) friend” ~ Freundin “(female) friend.” English borrowed from French the suffix –ess; hence actor ~ actress, lion ~ lioness, and many others. But in no language are the words for “girl” and “woman” derived from those for “boy” and “man.” German and Italian have resigned themselves to the existence of Professorin and Professoressa, whereas English does without professoress despite the fact that the number of women on our faculty is now considerable. Man and woman, boy and girl form natural pairs (and their referents form natural couples); yet language keeps them apart, and no one feels the inconvenience caused by the separation.
Grammar follows thought and generalizes disparate forms. It makes us feel that work, works, worked, and working belong together. English has almost no morphology left, but it is enough to look at a summary of Greek or Latin conjugations, to see how many forms ended up belonging together. We can only reason backward and keep begging the question. Why do we have separate forms for man and woman? Because each member of the tandem was felt to be unique, rather than “derived.” How do we know that? From the fact that the words are different. The vicious circle is unmistakable. We have no way of deciding why thought combines some entities but separates others. However, certain moves can be explained. For example, horses is the plural of horse (one horse/many horses), but I cannot be multiplied, even though grammar says that we is the plural of I. Therefore, it does not come as a surprise that I and we have different roots. Likewise they is not the plural of he, she, or it.
The speakers of early Indo-European who coined the words for “first” and “second” understood them as “the foremost one” and “the next one” and saw no intrinsic connection between what we call ordinal numerals and the cardinal numerals one and two. Suppletive forms in the pairs one/first and two/second turn up in various languages with rare consistency. We wonder why the comparative of good is better. We should ask ourselves what the positive degree of better is! It has never existed. From an etymological point of view, better means approximately “improved; remedied; compensated for.” Good needed a partner meaning “more than good” and better offered its services. We would have preferred “gooder,” but our indomitable ancestors chose to do their work the hard way. They did the same all over the Indo-European world (compare Latin bonus/melior/optimus, and be grateful for the similarity between better and best). Worse probably meant “entangled.” Yet the suffix –er in better (it once existed in worse too) indicates that the comparative force of both adjectives was not a secret.
Perhaps the hardest case is suppletion in verbs. We encounter cases like go/went everywhere. Moreover, the present is affected as often as the preterit. In Italian, the infinitive is andare, but “I go” is vado; the French pair is aller and vais. A look at the entire panorama of Indo-European shows that suppletive forms occur in the conjugation of the verbs for “come; go,” “eat,”, “give,” “take, bring, carry, lead” (those who studied even a bit of Latin had fero/tuli/latum beaten into them at the very beginning), “say, speak,” strike, hit,” “see, show,”, and of course “be, become.” In most cases the relevant forms are individual (like andare and aller), that is, each language invented, rather than inherited, suppletion. The example of English is especially dramatic. The past of Old Engl. gan “go” was eode, a word derived from a different root. In Middle English, went, the historical preterit of wend (as in wend one’s way), superseded eode. The language had a chance of producing a regular past of gan but chose to replace suppletion with suppletion. Even in the carefully edited text of the Gothic Bible (a fourth-century translation from Greek) the preterit gaggida (of gaggan; read gg as ng) occurred once. In Gothic, but not in English. Those who know German may think that gehen/ging “go/went” are related, but they are not. The source of the illusion is the initial consonant g-.
No fully convincing explanation of this phenomenon exists, but some facts can be considered with profit. Early Indo-European did not have some of the tenses we take for granted. A classic example is the lack of the future in Germanic. This statement need not cause surprise. Even today we sometimes do very well without the future: the context does everything for us. Compare: I am leaving tomorrow and If I leave tomorrow…. The difference between the preterit and the perfect can also be hazy: “Did you put the butter in the refrigerator?” or “Have you put the butter in the refrigerator?” The difference is insignificant. Nor does any English speaker bemoan the absence of the aorist. Centuries ago, verbs were often classified according to whether they designated a continuous (durative) or momentary (terminative) action, and occasionally verbs like see (durative) and look (momentary) were later merged within a single paradigm. One thing is “go, walk”; something quite different is “reach one’s destination.” Consider the difference between speak and say. This is probably how went made a union with go. Eode is a word of obscure origin and its inner form meant as little to speakers in the fifth century as it does to us.
The merger of synonyms within one paradigm may not have been the only source of suppletion, but it was an important one. Perhaps the most intriguing question is why languages choose the same verbs and adjectives for defying regular grammar. It appears that the usual target is the most common of them: “good; bad,” “be; come; go; take; eat; speak” and the like (see the list above). Frequency in language always tends to defy regularization. Not every irregular form is the product of suppletion: man/men, tooth/teeth, do/does also have to be learned individually, but none of them is “suppletive.”
We have thrown a quick look at this vexing problem and see that final clarity avoids us, but such is the fate of all things whose past has to be not simply recorded but reconstructed. In any case, I have answered an old question, and my conscience is clear.
Anatoly Liberman is the author of Word Origins…And How We Know Themas 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.
Image credit: Boris Kustodiev. Portrait of the author Yevgeny Zamyatin. 1923. Drawing. Public domain via Wikimedia Commons.