989. What is a word? Does "that's" count as one word or two? In this excerpt from Paul Anthony Jones' new book "Why Is This a Question," we look at how things like plurals, compound words, and contractions make defining the word "word" a tricky exercise.
989. What is a word? Does "that's" count as one word or two? In this excerpt from Paul Anthony Jones' new book "Why Is This a Question," we look at how things like plurals, compound words, and contractions make defining the word "word" a tricky exercise.
| Find Paul Anthony Jones at https://www.paulanthonyjones.com/
| Edited transcript with links: https://grammar-girl.simplecast.com/episodes/what-is-a-word/transcript
| Please take our advertising survey. It helps! https://podsurvey.com/GRAMMAR
| Grammarpalooza (Get texts from Mignon!): https://joinsubtext.com/grammar or text "hello" to (917) 540-0876.
| Subscribe to the newsletter for regular updates.
| Watch my LinkedIn Learning writing courses.
| HOST: Mignon Fogarty
| VOICEMAIL: 833-214-GIRL (833-214-4475).
| Grammar Girl is part of the Quick and Dirty Tips podcast network.
| Theme music by Catherine Rannus.
| Grammar Girl Social Media Links: YouTube. TikTok. Facebook. Instagram. LinkedIn. Mastodon.
Grammar Girl here. I’m Mignon Fogarty, your friendly guide to the English language. We talk about writing, history, rules, and other cool stuff. Today, I have a piece about a question both practical and profound: What is a word? This is actually a condensed excerpt from the book "Why Is This a Question" by Paul Anthony Jones, who will be a guest on the show this Thursday. And I asked him if I could excerpt it because I think it's the best treatment of the question that I've seen in all my years as Grammar Girl.
by Paul Anthony Jones
Have you ever been asked what a word means but found yourself utterly unable to explain it? I can still remember the look of panic on my schoolteacher’s face when a boy in my class asked her what "grace" was (and why Mary was so full of it). And then there was the friend of mine whose endlessly curious three-year-old asked him what "depth" meant while he was filling her paddling pool during a lazy summer barbecue and in doing so instantly bamboozled every adult in earshot. Some words, it seems, are just difficult to define. We know what they mean and can use them without a second thought – but trying to put that meaning into words and it’s hard not to resort to little more than a string of synonyms. "Depth? Well, it’s just 'depth' isn’t it? Like . . . 'deepness.'"
When it comes to defining our indefinables, one of the great ironies of language is that the word "word" is one of them. It might not seem as though it should prompt the same navel-gazing as something like "grace," and if someone were to ask you what a word was, you’d probably be able to give them a fair idea. ("A word? Well, it’s a word isn’t it? Like . . . a little bit of language.") But in practice words are surprisingly difficult to pin down and practically every test or definition devised to do so quickly comes unstuck.
One common explanation is that words are everything found between spaces in writing. That’s certainly how word-counting computer programs operate. But how would you count contractions, like "that's" for "that is." Are they one word or two?
Another way of defining it is that when extra material is added to a sentence, additional words will always fall between, not inside, those already there. So "The owl and the pussycat went to sea" could become "The wise owl fledglings and even the aloof pussycat quickly went back to the sea." We’d certainly never find ourselves talking about a puss-aloof-ycat, but any definition assuming words can never be infixed like this is abso-bloody-lutely flawed.
Broad rules of thumb such as these are clearly little use here. A much better starting point is that simple definition from earlier: a word is just a little bit of language. As throwaway as that might seem at first glance it makes sense. We recognise "at," "first" and "glance" as words, and can read them here as individual ‘bits’ of language. But it’s explaining precisely what these "bits" are that proves difficult, because as it stands, that definition could be easily misinterpreted. After all, individual letters are just bits of language too, as are individual sounds, punctuation marks, nonsense jumbles of characters, and even whole sentences and paragraphs. To exclude everything that isn’t a word, while including everything that is, we’re clearly going to need some firmer ground rules.
Some are certainly more difficult to explain than others, but all words have a meaning. Adding that requirement immediately cuts out a lot of this excess noise, as letters and sounds have no meaning at all on their own, and sentences and paragraphs go too far the other way – blending multiple smaller units into larger more meaningful wholes. Calling a word a single meaningful unit of language certainly feels like a more reliable definition, but there’s still a problem: in language not everything that has meaning is a word.
In English we typically add an "–s" onto the end of a noun to create its plural, for example – changing one word into many words, one dog into multiple dogs, and a detached house into a row of houses. We’d scarcely think of that "–s" as a word in its own right, yet to be capable of creating this kind of change, it must have some kind of meaning. Put another way, if a dog is a canine animal, and "dogs" means "more than one canine animal" then surely "–s" must be the part that means "more than one". So wouldn’t that make "–s" too a single, meaningful unit of language?
The problem is that "–s" is not a word but a morpheme. Morphemes are the smallest possible meaning-bearing components of a language; that meaning is called a sememe. By definition, morphemes can’t be broken down into anything smaller that likewise has any kind of meaningful content. So while the "–s" of "dogs" is a morpheme, the meaningless "d–" at the beginning is not.
Confusingly, that definition means many words count as morphemes too. "Dog" can be broken down only to its individual sounds "d"’ "o,"’ and "g," and because they have no meaning on their own, "dog" is a morpheme as well. "Dogs" on the other hand can be split apart – into its singular root "dog" (meaning ‘canine animal’) plus the plural tag "–s" (meaning ‘more than one’). So while one dog is a morpheme, multiple dogs are not. That overlap can make morphemes a tricky concept to grasp, but this distinction is an important one: whatever definition of "word" we end up with, it will have to include the likes of both "dog" and "dogs" while excluding the likes of "–s." Dig a little deeper, however, and we have a neat way of doing just that.
Morphemes play a hugely important role in how our language operates. As well as changing singular words into plurals (like changing "dog" to "dogs"), we can use the likes of "–ing" and "–ed" to change the tense of verbs (giving us "talking" and "talked" from the verb "talk"), and we can use "–er" and "–est" to expand on our adjectives (making "quicker" and "quickest" out of "quick"). These are known as inflectional morphemes, so they work solely to alter the grammar of whatever root they attach to. Conversely so-called derivational morphemes work to change the meaning of their roots, and thereby create entirely new words. We can use the suffix "–less" to form words implying an absence of something for example, like "faultless" or "timeless," or tag "anti–" onto a word to create its opposite, such as "antihero" or "anticlimax."
Unlike the words they connect to, however, on their own the likes of "–s" and "anti–" are lost. Attached to nothing they mean nothing. You could no more draw someone’s attention to a pack of dogs by shouting "–s!" than you could label yourself "anti–" without there being something to be anti– against. These are bound morphemes – fragments of language whose meanings come to the surface only when they are ‘bound’ to other things. The opposite, like "dog" and "house," are free morphemes which need no such support. And when it comes to defining a word, this freedom is crucial.
The linguist Leonard Bloomfield defined a word as a "minimum free form" – a single unit of language smaller than a phrase or a sentence that is capable of maintaining its meaning on its own. The likes of "dog" and "dogs," "owl" and "pussycat," "linguist" and "definition" all pass that test, but on their own "–s" and "anti–" fail it. Demanding a word not only have a meaning but be independently meaningful therefore excludes bound morphemes like "–s" while including everything we can use them to create. It’s an ingenious solution. But where does that leave compound words such as the one word "greenhouse" or the two word "ice cream"?
Despite their twofold structure, compound words represent single concepts; if they didn’t, there’d be no difference between a "loudspeaker" (one word) and a "loud speaker" (two words).
But while "loudspeaker" and other such close compounds are spelled without a space, many other compounds, such as "post office" and "golf ball," are divided by a space. And these too are single concepts: you can't drop the "golf" from "golf ball" any more than you could claim a "house" and a "doghouse" are the same thing. Open compounds are therefore single, independently meaningful units of language built from single independently meaningful units of language. We can’t split them up because their meaning relies on both their halves working together. So we're effectively back where we started, asking: is "coffee shop" one word or two?
[And listeners, this is where I'm extensively condensing the chapter.] The answer about compound words is incredibly complex, and ultimately comes down to how you stress the different parts and syllables.
Perhaps, then, we need to rethink our approach. As soon as we’re forced to consider the stress patterns of individual words, our focus shifts from written language to spoken language. Speech predates writing, of course, as our evolutionary ancestors were talking to one another long before they thought to put pen to paper (or stylus to clay, as the case may be). Rather than bogging ourselves down in the problems posed by letters and spaces on a page, why not take our language back to its roots and consider a word as primarily a set of sounds? The sounds of "d," "o," and "g" together form the word dog, /dɒg/, which carries the meaning "canine animal". String together the seven sounds in /kɒfɪ ʃɒp/, and you’ll have the single meaningful compound word "coffee shop," regardless of whether it is written open or closed.
Drawing this line in the linguistic sand gives us this definition: a word is a single independently meaningful unit of language, consisting of a sound or series of sounds, that can be represented as a series of written characters. Focusing on sound first and considering spelling only as an afterthought avoids the issues presented by compound words, while leaving bound morphemes and everything else to one side. Does that solve our problem? Well, in basic terms, yes – if you’re looking for an answer to the question "What is a word?", then something along these lines is probably the closest you will come to defining a word in any kind of watertight way. But in broader terms, no – even at this late stage, there are innumerable problems here.
Earlier, we touched on whether a contraction such as "that's" for "that is" constitutes one word or two. We still haven’t answered that question, for the very good reason it’s all but impossible to do so conclusively. The same goes for contractions such as "dontcha," "shoulda," "gotcha" and "imma," which operate as single word-like units in speech, despite being built from multiple individual words smashed together. And consider this too: now that we know how bound morphemes work, would "anti-dog" qualify as a word? How about if someone were to comment on how "coffee-shopless" their neighbourhood was? You’d doubtless know what they meant, but is that a word?
Here’s a sobering thought as well: at no point here have we stepped outside the cosy confines of English to consider how this definition might fare in other languages. German is well known for its capacity to string multiple elements together to form single word-like units with multifaceted meanings. In 1999, the official Association for the German Language nominated Rindfleischetikettierungsüberwachungsaufgabenübertragungsgesetz [honestly, I can't say it, but it's 63 letters long] — the association chose this beast as one of its Words of the Year. Is that a single word, or a conjoined collection of individual words? Even before we get to other languages, we’re in a quagmire here. There is, however, some good news: in the grand scheme of things, none of this really matters.
As much as we might think of the study of language as being the study of words, when it comes to examining and comparing languages, the word "word" just isn’t a particularly useful one. The reason we’ve ended up discussing the likes of morphemes and sememes here is because labels like these are of much greater value to language study than anything as vacuous and temperamental as "word." Sure, when we don’t need to be quite so academically rigorous, having a word like "word" to throw around is immensely useful. It lets us talk about the words on a page, jot down a few words, have a few words with someone, learn and translate words into a different language, and count the number of words in an assignment or essay. And as something with which to casually divide up our written world, "word" works just fine. When it comes to trying to pin down its meaning more precisely, however, perhaps we shouldn’t be too surprised that something we use so loosely defies all our attempts to define it.
Again, that was a condensed version of the "What Is a Word?" chapter from Paul Anthony Jones' new book, "Why Is This a Question." Check your feed in a couple of days for my interview with him where we talk about why English doesn't have gender anymore, which languages are the hardest to learn (and why), the strange story of how we got the letter Q, and more.
Finally, I have a familect story from Annette:
Hi, this is Annette Lyon calling from Utah with a familect. As a novelist, editor, and the child of a linguist and immigrants, I have many familects that I might share another time, but here's one. I have four children, and one day, when our second oldest was about 12, she got this "Aha" look on her face, and she said, "I just realized that the Target logo is a Target." We all laughed with understanding because who hasn't had those "Aha" moments that are to other people "no, duh" moments. So ever since, anytime one of us realizes something that was probably obvious to others, we say, "I just had a Target moment."
Thanks so much Annette! I'll never forget the first time I realized that the space in the middle of the FedEx logo makes an arrow, and it turns out a lot of other people have had the same experience. I shared your story on my social media, and other common Target moments like this are realizing the Amazon logo has an arrow going from A to Z, that the name of the food chain Arby's is phonetically spelling out the letters R and B for "roast beef," not realizing the letter K and the number 9 for police units with dogs create the word "canine," and the funniest one to me was an editor friend of mine named Rudy who said she was in grad school when they realized that none of the Doobie Brothers were named Doobie. If you want to read all the funny comments, you can find the posts on Threads where I'm "thegrammargirl" and on Facebook where the page is just Grammar Girl. And thanks again for the story Annette.
If you want to share the story of your familect, your family dialect, a word your family and only your family uses, call the voicemail line at 83-321-4-GIRL. It’s in the show notes, and be sure to tell me the story behind your familect because that’s always the best part.
And if you're a Grammarpalooza subscriber, you can also send a voice memo. I send text messages with fun facts a couple of times a week. It's a great way to support the show and the first two weeks are free, and I'm hearing that the people who are signed up are really enjoying it. So that's great! To sign up, visit https://joinsubtext.com/grammar or text "hello" to (917) 540-0876.
Grammar Girl is a Quick and Dirty Tips podcast. Thanks to Davina Tomlin and Kamryn Lacey in marketing; Morgan Christianson in advertising; Nathan Semes, in audio; Brannan Goetschius, director of podcasts, and Holly Hutchings in digital operations, who is re-adding Duolingo to her routine so she can be ready for the next eclipse … in Spain! Fantástica!
And I’m Mignon Fogarty, better known as Grammar Girl. That's all. Thanks for listening.