This is efficient in a digital medium, where scrolling down is easy and unbounded: not a waste of pixels the way it might be a waste of paper. Linebreaks come for free: they don’t take up any more bytes than a period and a space, and they add a lot in readability. Both “new line” and “send message” take a single keystroke, often the same enter key, so the muscle memory is easy. Plus, it helps the conversation flow better if you hit “send” after every utterance rather than waiting and sending a whole essay: the reader can start thinking of a reply sooner. Even in more formal genres online, such as news articles, paragraphs have gotten shorter and are separated by a blank line rather than a space-saving indent as they are on paper.
For people whose linguistic norms are oriented towards the offline world, the most neutral way of separating one utterance from the next is with a dash or a string of dots. After all, you definitely wouldn’t want to send each of these phrases as a separate email, let alone as a separate text in the days when we were billed per message. You’d take up four times the space on a postcard if you started a new line every time! Here’s the same example in the punctuation style of the offline-oriented:
hey...how’s it going.....just wondered if you wanted to chat sometime this week......maybe tuesday....?
This, too, has a logic to it: while some kinds of punctuation are traditionally reserved for joining full clauses (periods) and others for dependent clauses (commas), ellipses and dashes are deemed acceptable for joining both sorts, even in the most conservative styles. So if you’re writing informally and you don’t want to bother deciding whether your string of words is a full sentence or merely a clausal fragment, one way to split the difference is to punctuate ambiguously—to use an ellipsis or dash. Sure, classically speaking, the ellipsis indicates omitted text or a trailing off, but that’s fine: in speech we sometimes trail our sentences off for casual effect. And sure, classically speaking, the ellipsis gets three dots in the middle of a sentence and four dots at the end and gets a slightly different spacing from simply three periods, but that’s the kind of rule that copyeditors care about, not composers of casual emails who have no dedicated ellipsis character on their keyboards. Informal writers who are oriented towards offline norms, like the 1970s Beatles postcards we saw in the previous chapter, sprinkle in dots and dashes to show they’re not standing on ceremony by committing to formal, clause-typing punctuation. It’s exactly the same motivation younger folks have for separating utterances by linebreaks or message breaks. The same reason, in fact, that Jane Austen sprinkled her original manuscripts with what seems to the modern reader to be an absurdly high number of commas, or that Emily Dickinson’s poetry contains a metric ton of dashes, if you can get ahold of an edition where they haven’t been edited out. Pause marking is really intuitive, and it always has been.
The problems start when you combine multiple sets of norms. A message like this, say, from an older relative to a teenager, or a boomer boss to a millennial employee, reads quite differently depending on what you think of as neutral.
hey.
how’s it going....
just wondered if you wanted to chat sometime this week......maybe tuesday....?
For some, it reads as a compromise between the new text messaging linebreak style and the older dot dot dot. But if you’re solidly in the linebreak camp, you see those extra dots or even just a single period where a linebreak or message break would have sufficed, and assume that anything that takes more effort than necessary is a potential message. The dots must be indicating something left unsaid: “how’s it going [there’s something I’m not telling you].” From a peer, something left unsaid might indicate flirtation. But from an older relative, that would be weird. What other kinds of hidden messages are left? The most common assumptions are either passive aggression or sheer confusion.
The passive-aggressive potential of the single period started being reported in thinkpieces in 2013, in a list at New York magazine and then later the same year as a full article in the New Republic, before popping up in a handful of other publications in subsequent years. The string of dots got a thinkpiece in 2018, though it has been popping up in comment threads since at least 2006, while its cousins, the hyphen and string of commas, have been less extensively reported but have occasioned long comment threads on blogs and internet forums. Despite the fears mongered by headlines, it’s not the case that the passive-aggressive meaning has completely killed all other uses of the period. The linguist Tyler Schnoebelen, who’s definitely younger than the peak dot-dot-dot generation, did a study of periods in his own 157,305 sent and received text messages. He found that, true, periods were rare in short, informal messages—ones less than seventeen characters or containing lol, u, haha, yup, ok, or gonna. But they were still often found in messages longer than seventy-two characters or containing words like told, feels, feel, felt, feelings, date, sad, seems, and talk. The added weight of the period is a natural way to talk about weighty matters.
So how is a person to tell whether a given period is supposed to be passive-aggressive, sad, or merely formal? The jumble of meanings associated with the period became clear to me when I started interpreting it as a marker of typographical tone of voice. Just as a question mark can indicate a rising intonation even without a question (Like so?), the period can indicate a falling intonation even when it’s not serving to end a statement (Like. So.). When I put on a newscaster voice, I deliver every sentence with falling intonation. Solemnly. Portentously. But in an ordinary conversation, we don’t speak in full sentences, and we especially don’t round them all off with a distinct fall. (“And now over to: The Weather.”) Instead, we speak in utterances, and our intonation is neither rising nor falling: by default, it’s flat or trailing off, like a dot dot dot or an unpunctuated linebreak.
Both the dot-dot-dotters and the linebreakers have instinctively brought us back to an idea from the very beginning of punctuation. The first kind of punctuation marks indicated breaks between utterances, and medieval scribes were the ones who first used them. One important medieval punctuation mark was the punctus, a dot which was placed in the location of the modern comma for a short breath, midway through the line for a medium breath, and up around the position of the apostrophe for a long breath. Before that, ancient Greek and Roman writing had all the formatting of a wordsearch puzzle grid, with no punctuation, paragraphs, or even spaces between words, in all caps (carvings) or all lowercase (ink). Like a wordsearch, the reader had to figure out where one word stopped and another began—and also like a wordsearch, they often did this by muttering under their breath. (Mercifully, unlike a wordsearch, the words didn’t appear diagonally and there weren’t additional distractor letters.)
With the rise of the printing press and dictionaries in the sixteenth and seventeenth centuries, spelling and punctuation became more complicated and standardized. Scribes had spelled and punctuated idiosyncratically, but printers could—and did—change things while typesetting to match everything else they were printing. People may never have wholly followed these elaborate guidelines in their personal correspondence—Austen, Dickinson, and the Beatles certainly didn’t, and the handwritten notes of famous writers are among the most analyzed. But when a printing press was the easiest means of reaching a large audience, edited and formal punctuation became the main kind that people saw. The internet made our personal punctuation preferences public, and brought with it a different set of priorities: writing needs to be intuitive, easy to create, and practically as fast as thinking or speaking. We drew these requirements together to create a system of typographical tone of voice.
Strong Feeling
WHEN YOU WRITE IN ALL CAPS IT SOUNDS LIKE YOU’RE SHOUTING.
All caps to indicate strong feeling may be the most famous example of typographical tone of voice. But there are several kinds of strong feelings. Linguist Maria Heath asked a cross-section of internet users to rate the difference in emotion between a message in
all caps and the same message in standard capitalization. She found that all caps made people judge happy messages as even happier (IT’S MY BIRTHDAY!!! feels happier than “It’s my birthday!!!”) but didn’t make sad messages any sadder (“i miss u” is just as sad as I MISS U). When it came to anger, the results were mixed: sometimes caps increased the anger rating and sometimes it didn’t, a result which Heath attributed to the difference between “hot” anger (FIGHT ME) and “cold” anger (“fight me”). A single capped word, on the other hand, is simply EMPHATIC. Looking at examples of all-capped words on Twitter, Heath found that the most common single ones included NOT, ALL, YOU, and SO, as well as advertising words like WIN and FREE: the same kinds of words that are often emphasized in spoken conversations (or commercials). When we want to emphasize something in speech, we often pronounce it louder, faster, or higher in pitch—or all three at once. All caps is a typographic way of conveying the same set of cues.
Emphatic caps feel like the quintessential example of internet tone of voice, and sure enough, they’ve been around since the very early days online: linguist Ben Zimmer found people in old Usenet groups explaining that all caps meant yelling as far back as 1984. What’s more intriguing is that capitals were available for emphasis long before the internet as well. The linguist John McWhorter dates shouty caps back to pianist and writer Philippa Schuyler in the 1940s, while author L. M. Montgomery has a character use both capitals and italics for emphasis in her fictional diary entries of the 1920s, which another character criticizes as “Early Victorian”—meaning old-fashionedly melodramatic, even back then. Going yet further back, a newspaper in 1856 described a line of dialogue with the phrase “This time he shouted it out in capital letters.”
Back in the heyday of personal letter-writing, all caps were just one part of a broader emotional ecosystem for expressing strong feeling, along with italics, underlining, larger letters, red ink, and other decorative formatting options. The emotional use wasn’t even the most prominent option: all capitals were widely used to avoid the idiosyncrasy of joined handwriting, such as in comic strips, on forms (“Please fill out your name in block capitals”), or in official documents by lawyers, architects, and engineers. Similarly, several of the postcards that I looked at for the previous chapter were written in block capitals, especially in the address field. Typewriters and early computer terminals made illegible handwriting less of a problem, but they also introduced a new one: they wouldn’t let you type italics and underlines and font sizes (for that matter, many social media sites still don’t). This created a vacuum into which the preexisting but relatively uncommon shouty caps expanded.
This brings us to a puzzle. Early internet guides like the Jargon File, Wired Style, and website FAQs mentioned all caps, but not to facilitate shouting, the way that *bolding asterisks* or _italicizing underscores_ were recommended to compensate for the lack of other formatting that can indicate emphasis, or a smiley face was recommended to facilitate sarcasm and joking around. No, they were generally trying to discourage it, meaning that a fair percentage of eighties and nineties computer users were writing their routine correspondence in all caps.* Where did the idea that it was ever okay to type a full message in block capitals come from? After all, people have been handwriting in lowercase for over a thousand years, and even the melodramatic early Victorians didn’t capitalize everything. Why would anyone suddenly switch to all caps on a computer?
Part of the blame may go to Morse code, that dashingly dotty system used for sending telegrams. Morse code represents every letter as a combination of dots and dashes, suitable for transmitting as long or short taps along an electrical line: A is dot dash, B is dash dot dot dot, and the rest of the twenty-six letters can all be represented as combinations of up to four dots and/or dashes. But if we wanted to include lowercase letters, we’d need a fifth and a sixth dot or dash, because we’d be representing fifty-two symbols, and telegraph operators would have to memorize twice as many codes. Unsurprisingly, people decided it wasn’t worth it—if all caps was good enough for the Romans, it would be good enough for telegrams.
Early computers were very similar. Some used teletype machines—the mechanical descendants of telegraph operators—as a way to transmit or print out information. The classic first command that you learn when you start to code is something like PRINT(“HELLO WORLD”), which causes the computer to display HELLO WORLD onscreen. It doesn’t make the computer print out on paper HELLO WORLD, but at one point it did—back before screens, when we commanded computers by keying words into a teletype machine and received their replies printed out onto rolls of paper. Even once computers had screens, storage space was still expensive, as precious as the brain cells of a telegraph operator, so many of them, such as the Apple II, displayed everything in just one case—all caps. Relicts of this setup are still in place on some commercial computer systems: teletypes are uncommon, but your grocery store receipt, bank statement, or airplane ticket might very well appear from a roll of shiny paper, printed in all caps.
By the time computers did start supporting lowercase characters, we were faced with two competing standards: one group of people assumed that all caps is just how you write on a computer, while another group insisted that it stood for yelling. Ultimately, the emotional meaning won out. The shift in function happened in parallel with a shift in name: according to the millions of books scanned in Google Books, the terms “all caps” and “all uppercase” started rising sharply in the early 1990s. By contrast, in the earlier part of the century, the preferred terms were “block letters” or “block capitals.” People tended to use “all caps” to talk about the loud kind, while block capitals more often referred to the official kind, on signs and on forms. But the addition of all caps for tone of voice didn’t eliminate the official kind of capitals, which remain common on EXIT signs and CAUTION tape and CHAPTER ONE headings: they may be emphatic, but they aren’t interpreted as especially loud. Rather, our interpretation seems to flip depending on whether we read the text as formal or informal: HOME in a website’s menu bar is a mere graphic design choice, while HOME in a message like “ugh I want to go HOME” is typographical tone of voice.
* * *
—
Another way to do emphasis online is by repeatingggggg letterrrrssss, especially for emotive words like “yayyyy” or “nooo.” Just like shouty capitals, the origins of this practice predate the internet by maanyyy years. I searched the Corpus of Historical American English for sequences of at least three of the same letter (to eliminate common English words like “book” and “keep”). The corpus contains texts from 1810 to 2009, but to my surprise, there were hardly any results in the first half of the corpus. The few earlier examples were mostly just typos, like “commmittee,” or numerals, like “XXXIII.” Here’s the oldest real example I could find, a character pretending to be a candy-seller in a novel published in 1848:
“Confectionary, confectionary,” he cried, bursting into a louder tone of voice, which rang forth clear and deep-toned, as a bell. “Confectionary!” and then he added with grotesque modulations of his voice, “Confecctunarrry!”
“By Jove, how this reminds me of the little fellow in London. I’ll go the complete candy-seller. I might as well.”
“Ladies and gentlemen! Here’s your fine candy, lozenges, apples, oranges, cakes and tarts! Heeeere’s your chance!”
The “grotesque modulations” of this 1840s faux confectioner were an anomaly, ahead of their time. The author respells the elongated “confectionary” with a “u,” rather than preserve the component letters like a modern writer would do. Even now-commonplace elongations of sounds like “ahhh,” “oooh,” “hmmm,” “ssshh,” and “brrr” don’t start showing up in this historical corpus until the decade before and after 1900, whereupon they increase steadily for the next hundred years, displacing word-like versions such as “ahem” and “hush.” Rare, one-off elongations of full words like “confecctunarrry,�
�� “evvveryone,” and “damnnn” follow in substantial numbers a few decades behind, starting to rise in the 1950s and 60s and really getting popular in the 1990s and 2000s. The period when lengthening became popular lines up with the rise of recorded speech, such as phonographs, records, cassettes, and CDs. It might be coincidence, but it might also be that when we started being able to play and replay recorded speech, we started paying more attention to representing it precisely. At any rate, it’s clear that the goal of repeated letters is to represent speech in writing because the early examples show up in fictional dialogue, especially in play scripts and novels.
Repeating letters is an expressive tool that’s been growing for over a century in informal writing, not just on the internet. And it isn’t haphazard. One study looked at the most commonly lengthened words on Twitter and found that they still tend to be sentiment words. The top twenty most lengthened words are a cornucopia of emotions: nice, ugh, lmao, lmfao, ah, love, crazy, yeah, sheesh, damn, shit, really, oh, yay, wow, good, ow, mad, hey, and please. Several studies have found that this expressive lengthening, as linguist Tyler Schnoebelen named it, is sensitive to social context: people lengthen more in private texts or chat messages than in public posts.
People are also sensitive to linguistic cues. In a study I did with the linguist Jeffrey Lamontagne, we found that while people generally lengthen the rightmost letter in a word, they’ll also lengthen the rightmost letter in a smaller unit of sound. For example, in the word “dream” the “ea” together indicates the vowel sound, so people will lengthen this word as either “dreaaam” or “dreammm.” But in the word “both,” the two middle letters “ot” are not a unit (the “t” belongs with the “h” instead), so people lengthen it as “bothhhh” or maybe “boooth,” but never “botttth.” But people aren’t completely tied to phonological feasibility. They often write things like “stahppp” or “omgggg,” but it’s not physically possible to hold ppppp or ggggg for more than an instant. Even more improbably, people sometimes “lengthen” silent letters, writing “dumbbb” or “sameee.” What’s cool about expressive lengthening is that, although it started as a very literal representation of longer sounds, it’s ended up creating a form of emotional expression that now has no possible spoken equivalent, making it more akin to its typographical cousins, all caps and italics.
Because Internet Page 12