Bertrand Russell, in his essay “On Being Modern-Minded,” writes, “We imagine ourselves at the apex of intelligence, and cannot believe that the quaint clothes and cumbrous phrases of former times can have invested people and thoughts that are still worthy of our attention.”
Consider his use of cumbrous. Perhaps you knew the word’s meaning before encountering this passage. I did not. But from the rest of the sentence—its overall sentiment, as well as the parallel use of another adjective, quaint—I was able to deduce that the phrases Russell speaks of are unsophisticated. And what sorts of phrases might a witty scholar such as Russell deem unsophisticated? Awkward, cumbersome ones. Which, wouldn’t you know it, sounds an awful lot like cumbrous.
This feat of Holmesian reasoning was brought to you by redundancy. Discourse is redundant, which is why we so often understand the gist of a message even when our cell phone cuts out on us or a train blazes by or we can’t remember for the life of us what c’est la vie or crème de la crème means. It’s redundancy that writers like Junot Diaz and Anthony Burgess exploit when they integrate unknown foreign (or made-up) words into their stories, adding color local without bamboozling a readership conditioned to less exotic colors.
Discourse redundancy is a powerful tool for word learning. Most of the studies that have looked at this topic have done so within the context of reading. This is in part because we learn the majority of our vocabulary—not the majority of the words we use on a daily basis, but the majority of all of the words we know—from reading. Most words are rare, and rare words occur almost entirely in print. According to a 1988 study, even children’s books and comic books contain a higher proportion of rare words than the casual speech of college graduates does.
So how good are we at using discourse context to learn new words while reading? In 1999, Dutch researchers M. S. L. Swanborn and Kees de Glopper analyzed the results from 20 separate experiments on incidental word learning (that is, learning when readers did not expect to be tested on vocabulary afterward). The researchers concluded that for every 100 new words a schoolchild encounters, he’ll “learn”—derive and remember enough about that word to answer a question about it later on—about 15 of them.
But this number isn’t magical. Who the reader is, what he’s reading, and how much he cares about what he’s reading all matter, and are all highly variable. Indeed, the average percentage of new words learned across the 20 studies ranged from three to 54.
I’d like to think that my personal rate of word learning falls on the higher end of this range, given that I’m an avid, experienced reader. Because readers must use their existing vocabulary to correctly infer new word meanings, those with already-rich vocabularies are better equipped to get richer, a phenomenon scholars have termed “the Matthew effect.” I, who already possessed relevant meanings for apex, quaint, and invested, was a prime candidate for acquiring cumbrous. (Someone with an even richer vocabulary could probably have learned cumbrous from a sentence like The anonymuncule’s cumbrous lexiphanicism led to severe redactophobia, but nobody likes a smarty-pants.)
A confession: this was not my first time reading Russell’s essay. This is fairly damning evidence that I did not learn cumbrous the first time around, which brings me to my last point. Though discourse redundancy makes learning new words possible—even easy!—for readers, it can also make learning new words unnecessary. C’est la vie; crème de la crème.
