Oxford University Press's
Academic Insights for the Thinking World

A few more of our shortest words: “if,” “of,” and “both”

The post of 21 June 2017 on the “dwarfs of our vocabulary” was received so well that I decided to return to them in the hope that the continuation will not disappoint our readers. Those dwarfs have a long history and have been the object of several tall tales.


Like most subordinate conjunctions, if conveys a rather abstract meaning. Its bulky, bookish synonyms provided (that) and in case of have little virtue. Sometimes it is possible to do without if by inverting the word order, as in “were I (had I been) there…,” “should you ever meet him…” and so forth, but in most cases, when we want to introduce a conditional clause, we say if. The word has a respectable ancestry. Thus, in Old English, we find gif with the puzzling initial g-, pronounced as Modern Engl. y– in yes.  (Although this additional y- was not restricted to the conjunction, its appearance is always a riddle.) All “dwarfs” tended to interact with their likes as regards both meaning and pronunciation. For instance, in gif, the initial g- may have been “borrowed” from “yea.” German speakers, unless they are linguists, do not know that -r in oder “or” was added to it under the influence of aber “but” and other similar form words. However, Old Frisian also preserved both ef and jef, so that j– was not a mere freak of Old English phonetics. Indeed, Old Frisian had several other forms that will soon make the plot thicken.

How do people coin a conjunction like if? Why should a combination of short i and f refer to a possibility of something happening? Could if be a relic of some longer and less abstract word? The once famous politician Horne Tooke (1736-1812) was also a self-made philologist. His two-volume book The Diversions of Purley had numerous distinguished admirers (purley “purlieu”; the title must have referred to the joys of hunting, but the work’s Greek title Epea pteroenta means “winged words”). Tooke’s career and the reason for his popularity need not delay us here, though even today some people find virtue in his etymologies. According to one of his bizarre ideas, numerous words go back to the imperative of verbs. He suggested that if is an ossified relic of give! in the assumed sense “grant, suppose.” I am mentioning this derivation only because it has been repeated in so many sources.

The famous Horne Tooke, a great admirer of imperative forms.

Alongside Engl. if, we find a group of Old Icelandic nouns: if, ef, ifi, and efi, all of them meaning “doubt.” The corresponding verb ifa meant “to doubt”; it alternated with efa. The Icelandic cognate of Engl. if was ef. Old High German had the noun iba “condition, stipulation, doubt.” It would be almost ideal to trace the conjunction if to the dative or instrumental case of a word referring to hesitation and uncertainty. But despite some excellent suggestions the origin of those nouns remains unclear, and, as always, I resort to the law that one obscure (or, let us say, not fully transparent) word cannot shed light on another word of dubious origin. Also, the derivation of such a conjunction as if from a noun would be a most unusual process. Could all those nouns and verbs go back to the conjunction (in order to designate something iffy), rather than the other way around?

The iffy parts of our life are never behind us.

Finally, it is not improbable that if is a stub of some longer word. And indeed, in Gothic (recorded in the fourth century), the cognate of gif ~ if is ibai “whether.” The Gothic form is suggestive, but it too needs an explanation. Perhaps ibai is ib + ai! In this case, the short form is more important than the long one, and we are left with the enigmatic ai. It was the great Indo-European scholar Karl Brugmann who seems to have offered the best explanation of ibai. He isolated two elements in it: i and bai.

While reading the history of the Indo-European languages, one constantly runs into something called pronominal roots (or stems). Among them, the most often mentioned are *i and *e. The asterisks mean that they have not been attested but only reconstructed. In the remote past, we think, such roots were used as the basis of other words. Did they ever exist as independent units? Apparently, they did (for roots are not the flotsam and jetsam of language waiting to be attached to other words), though it would be hard to give details. Did our very remote ancestors go about and, while pointing to the objects around them, shout “i” or “e” (that is, the short i and e of Engl. in and hen)? Perhaps they did. In any case, i- ~ e is part of several well-known pronouns and conjunctions. Among the pronouns, the most obvious example is Modern German es “it,” from et. The second element of the conjunction ibai (i-bai) is then –bai, either an emphatic particle (in Gothic, ba “even though” occurred) or both. Gothic did have bai “both.”

Strangely, Gothic had an even closer cognate of gif than ibai, namely jabai, a compound made up of ja “yea, indeed” and the same bai.  One may assume that two such similar words as ibai and jabai interacted and were even occasionally confused in Gothic.  But alas, we have no evidence of everyday Gothic, because only parts of the Bible in that language have come down to us. Colonel Pickering, a colleague of Professor Higgins’s in G. B. Shaw’s Pygmalion, was the author of Conversational Sanskrit. As it happened, that worthy officer’s interests did not extend to Germanic.

German had the noun iba “doubt” (see above), but it also had an unquestionable cognate of English if, namely obe and ubi “whether” (Modern German ob; note that in English, whether and if can also alternate, as in “I don’t know whether/if she will come”). We can see that the vowels of if and obe ~ ubi do not match.  The German case is not an aberration. Old Saxon had both ef and of, while the Modern Dutch conjunction is simply of. In Old Frisian, we observe the whole panoply: jef, ef, jof, and of. (Do you remember the caveat about the thickening of the plot?) Those who have read the post, referred to at the beginning of the present essay, will recall that the conjunction and also had several variants with alternating vowels, especially in German: unta and inti. The cause of the alternation was ablaut. The same factor will account for the alternation here. Ablaut was an extremely flexible and widely used means of word formation and inflection. Consider our ticktock, dickery, dickery dock, chit-chat, fiddle-faddle, fee-fi-fo-fum, and the rest? So why not obe ~ ubi ~ (y)if ~ iba?  Does that mean that Engl. of is related to if? And what is the ultimate origin of if?

In doubt about the the etymology of the word if.

Let us pretend that we are characters in 1001 Nights. Although dawn has just broken, the fairy tale is not over. The king postpones the execution of the story teller, in order to hear what happened to the characters.  If ifs and buts were candies and nuts, what a beautiful etymological dictionary it would be! You already know the story of but (see again the post of June 21), and in the nearest future I’ll tell you all I know not only about if, of, and both but also about nut. For the moment, suffice it to realize that the shorter the word, the longer its etymology.

A scene from “1001 Nights”.

Featured image credit: “Children look” by Luisella Planeta Leoni, Public Domain via Pixabay.

Recent Comments

  1. Stephen Goranson

    Yes, Horne Tooke was indulging in word play with his book title, surely. Yet it is worth noting that he indeed lived in Purley. I notice this because my first years of life were spend in Purley, Surrey.

  2. Johann Blagojevich

    Thank you. I find the etymology of ‘structural words’ fascinating.

  3. Herma Beute

    Although I can go along with your thread of thinking, I feel obliged to mention that the Modern Dutch conjunction ‘of’ has a totally different meaning than the English ‘if’ or even the German ‘ob’. Modern Dutch ‘of’ mostly means ‘or’, not ‘if’ or ‘whether’. Granted, ‘or’ also indicates a choice, maybe even an uncertainty or doubt, and there is a limited usage of the word that can be translated to English ‘whether’, but the way you connect both words evidently lacks some nuance.

Comments are closed.