Is it possible to tell from a word’s form whether it describes something concrete or abstract? Is there something in the sounds tree or sun or giraffe that screams I can be experienced with the five senses! and something in independence that warns us not to even try?
Jamie Reilly, now a professor at Temple University, and his colleagues analyzed over 2,000 English nouns and found that, to some extent, the answer is yes: Abstract nouns tend to be longer and more morphologically complex (having more prefixes and suffixes) than concrete ones.
Moreover, we as English speakers have internalized these differences, and will happily use them to gauge the concreteness or abstractness of strings of letters like inhighosht that don’t describe anything at all. And nouns that don’t conform to our ideas about how they should sound—abstract nouns that are short and simple, concrete ones that are long and complex—are understood more slowly than those that meet our expectations.
On one level, this isn’t so surprising. If a pattern exists in the language, there’s a pretty good chance we’ve picked up on it. (Nouns that resemble verbs and verbs that look like nouns also present difficulties, for instance.) But why might this pattern exist in the first place?
One possibility is that the two classes of nouns differ etymologically. Old English has Germanic origins. But since the 11th century, Latin has also exerted a powerful influence on the language. And beginning in the Renaissance, when it became more acceptable to engage in religious and scientific discourse in one’s own mother tongue, the English language hungrily acquired (and anglicized) a swath of Latin terms wholesale. English writers have long played the distinctive properties of Latinate and Germanic words against each other. Consider this passage from Shakespeare’s Macbeth:
Will all great Neptune’s ocean wash this blood
Clean from my hand? No; this my hand will rather
The multitudinous seas incarnadine,
making the green one red.
As Robert McCrum puts it, “Having flattered the classically educated men of substance sitting at the side of the stage, or in the twopenny seats, with a scintillating Latinate phrase (‘The multitudinous seas incarnadine’), Shakespeare repeats it in good, plain old English (‘Making the green one red’) for the benefit of the groundlings crammed into the pit.”
In a 1984 essay, historian Jacques Barzun makes the explicit observation that Latinate words tend to be more abstract than even their nearest Germanic equivalents. The English language, he writes, “possesses two vocabularies, nearly parallel, which carry the respective suggestions of abstract and concrete, formal and vernacular. A writer can say concede or give in; assume or take up; deliver or hand over; insert or put in; retreat or fall back.” Indeed, in tracking down the etymology of those 2,000 English nouns, Reilly and his colleagues find that abstract ones tend to be Latinate, while concrete ones have a wider range of etymological origins—most commonly Germanic. (Barzun celebrated the two vocabularies; in “Politics and the English Language,” George Orwell griped at writers who preferred the Latinate over the Germanic because it led to vagueness and dishonest euphemism.)
On one level, then, the mystery looks to be solved. Latin loves its prefixes and suffixes; it has also had an undue influence on the words we use to describe ideas. Thus, our polysyllabic abstractions and stubby thing-words are the results of historical accidents. It could have been otherwise. Right?
But Reilly doesn’t think this is the whole story. In a very recent study, as yet unpublished, he looks at abstract and concrete nouns from eight additional languages and finds that, overall, abstract nouns also have a longer duration than concrete ones. In three of those languages, Dutch, Russian, and Hindi, the difference is significant, and in two others it trends in this direction. Only in the one sign language studied—American Sign Language—did the results pattern reliably in the opposite direction. Might there be an etymological explanation for these results as well? Or is there a deeper explanation, perhaps one that taps into something universal about how we humans label the world? As Reilly confessed to me in an email, “It’s a puzzle.”