The stay-at-home orders of recent months have meant a lot more time to indulge one of my least harmful vices — deep-diving into etymologies — and this has led me to reflect on a thing I hadn't noticed as much before: English has a surfeit of expressions for the idea 1 + 1. On the surface they appear oddly unrelated, but if we dig into their storied pasts, we find that some of the various words we use for indicating 2 are more related than you would expect.
First, let's consider the family of words that give themselves away as being tied in with twoness by the way they begin: with tw-. Here we have two, twelve, twenty, twin, twain, twice, twill (originally, "having double thread"), twine (which was originally two strands of fiber, twisted around each other). Twig is another member of this family, perhaps from the fact that a twig, when it emerges from a limb, represents an addition of one; it makes one become two. These words all walk around with their own handy mnemonic attached, and aside from two, whose pronunciation is anomalous, all of these words pass the phonics test as well: a child who has mastered the basics of English spelling and phonetics is likely to pronounce them correctly on first sight or spell them correctly on first hearing.
This amount of consistency is somewhat contrary to the spirit of English, and so of course we have alternative words to designate things that come in twos. We have another group of words that deal in doubleness, and these are the ones that begin with d, usually have round vowel, and may have other consonant sounds after that: duo, dual, deuce, duel, double, dozen, duplicity, duple, duplex, and dyad represent the happy campers in this group. The main source is Latin duo and Greek δύο (dyo), "two". The words that have an additional bl or pl consonant pair pick it up either from Latin plus, "more"; or plex, "fold".
Those who prefer to minimize the mnemonics in their minds may be relieved to know that the two foregoing groups are not in fact two families; they are cousins. Their granddaddy is the Into-European root that we represent as dwo-. You can read all about it in the excellent American Heritage Dictionary of Indo-European Roots. It's not unusual in the careers of old words for Ds and Ts to be interchanged as they pass through languages and dialects: these two letters represent the voiced and unvoiced versions of the same tongue-behind-teeth maneuver. A rounded vowel is mostly consistent across both groups of words. The two groups I've outlined above result from dwo- flowing into the Germanic languages to become the tw- group, and into Greek and the Romance languages to become the dyo/duo group before arriving in English.
As I noted in a column last year, "English is not alone in having apparently unrelated words for the low-value cardinals and ordinals. There's French un and premier, German ein and zuerst, Arabic واحد (wahad) and أول (ool), all of which are translations of one and first. Many languages have this odd disconnect.” Most languages, however, fall into line after the cardinal/ordinal split for 1. For next after the first, French has deuxieme, based on deux, German has zweite, based on zwei, and Arabic has ثانيا (thania) based on اثنان (athnan). But English does not have twooth, it has second, based not on two or on duo or anything like that, but ultimately on Latin secundus, "following, next, second". Unlike other Germanic languages, early English lacked an ordinal based on two; it was handy to just borrow something (as English often does) more or less ready-made from another language.
Second is not the only Latinate borrowing about twoness in English. We also have couple, whose etymology is related to 2 indirectly. The etymon of couple is copula, "bond", which English has borrowed for different purposes. The existence of a bond suggests two things joined that were originally separate, and of course it does take two to copulate. Yet another Latinate borrowing is pair, from Latin paria, "equal things," which also gives us par.
Three other related words in English are not so much about 2 as they are about implying 2: they are either, its negation neither, and whether, which is most closely related to the common etymon of all three, Old English hwӕther, "which of two". As soon as you use whether you're cluing up your listener that there's an alternative; it's either the one thing or the other, and together of course they make two.
The remaining oddball in the English lexicon of twoness is both, a very handy pronoun and adjective that throws a yoke over whatever two things you have been talking about and makes them a pair. Of all the words talked about here it is the second oldest, after two itself, with roots in Old English and cognates in other Germanic languages. It suggests that the earliest speakers of the language had a need to designate things that come in twos. And why not, since our bodies have two of so many things, and but for the coupling of two people, our bodies would not exist?
Both, either, neither, and whether are vestiges of a feature of Old English that we have now discarded: dual grammatical number. It's a rare bird in modern European languages generally, but not so unusual in world languages. There was a time when English speakers had at their disposal not only singular and plural, but specialized forms for designating two things (as opposed to one, or more than two). Modern Semitic languages (Arabic and Hebrew, for example) maintain dual forms robustly and have individual words to designate natural pairs such as the eyes and limbs, as well as dual forms for things that often come in twos. For example, Arabic يوم yawm (day), يومان yawman (two days).
Does 2 enjoy such diverse representation in other languages? I would be interested to know from speakers of languages far from English whether most of their words for twoness are based on the word for 2, or whether they also derive from multiple sources.