Skip to main content

Caption Chaos: When “Stephen Colbert” Becomes “Steven Coal Bear”

Caption Chaos: When “Stephen Colbert” Becomes “Steven Coal Bear”

If there’s one name automatic captions should be able to nail, it’s the host of the show you’re watching. And yet, over multiple episodes of The Late Show with Stephen Colbert, I noticed a recurring error: captions getting Colbert’s name hilariously wrong.

Thanks for reading Capturing Voices! Subscribe for free to receive new posts and support my work.

Sometimes it’s “Steven Coal Bear,” other times it’s just a jumble of guesses that make you wonder if the AI was multitasking. It’s hard not to laugh, but it also makes you think—if it can’t get the host’s name right, what else is slipping through the cracks?

It’s not just Colbert’s show. I flipped over to Jimmy Kimmel Live! and spotted the same issue—captions tripping over guest names, fast-paced jokes, and even simple phrases. Interviewing Kathy Bates, who starred in “Misery”, James Caan became James Kahn’s, and Jimmy’s sidekick Guillermo became… well, not even an actual word!

Automatic captions are improving, sure. But when accuracy matters—like for accessibility, clarity, or just keeping up with fast comedy—there’s no substitute for human touch. A real person knows that “Coal Bear” isn’t about climate change, and that timing is everything when delivering a punchline.

Thanks for reading Capturing Voices! Subscribe for free to receive new posts and support my work.

Sex, gender, and sociology

Sex, gender, and sociology

Where do the words “sex” and “gender” come from, and where are they now in their semantic journey? A fellow reader of the Grammarphobia blog asked the question, and here is the answer!

Thanks for reading Capturing Voices! Subscribe for free to receive new posts and support my work.

Q: What explains the increasing use of “gender” in place of “sex” in sexual terminology? For me, prudishness doesn’t explain it.

A: The nouns “sex” and “gender” have been used interchangeably since the Middle Ages for either of the two primary biological forms of a species.

Although the two terms are still often used like that, they began to go their separate ways in the 20th century. Here’s the story.

English borrowed “sex” in the late 14th century from Middle French, but the ultimate source is classical Latin, where sexus referred to the “state of being male or female,” according to the Oxford English Dictionary.

When “sex” first appeared in English, the dictionary says, it meant “either of the two main categories (male and female) into which humans and many other living things are divided on the basis of their reproductive functions.”

The earliest citation in the OED is from Genesis 6:19 in the Wycliffe Bible of the early 1380s. In this passage, the term is used for the sex of the animals in Noah’s ark:

“Of all þingez hauyng soule of eny flesch: two þou schalt brynge in to þe ark, þat male sex & female: lyuen with þe” (“Of all things living of any flesh, two thou shall bring into the ark, that of the male sex & female, to live with thee”).

When “gender” appeared in the mid-14th century, it was a term for a grammatical subclass of nouns and pronouns distinguished by their different inflections.

The earliest OED citation, which we’ve expanded here, is from the gender-bending legend of St. Theodora of Alexandria, who is said to have betrayed her husband and then done penance by dressing as a man and entering a monastery:

“Hire name, þat was femynyn / Of gendre, heo turned in to masculyn. / Theodora hire name was, parde, / But Theodorus heo hiht, seide heo” (“Her name, which was feminine of gender, she turned into masculine. Theoroda, her name was, by God, but Theodorus she was called, she said”).

From “De S. Theododra,” circa 1350, in Sammlung Altenglischer Legenden (Collected Old English Legends), 1878, edited by Carl Horstmann.

In the 15th century, the OED says, “gender” came to mean “males or females viewed as a group,” which the dictionary describes as the same sense as the original meaning of “sex.”

The earliest Oxford example is from the 1474 will of Thomas Stonor: “His heyres of the masculine gender of his body lawfully begoten” (from The Stonor Letters and Papers, 1919, edited by Charles Lethbridge Kingsford).

At the end of the 19th century, the noun “sex” took on an additional meaning—the sexual act—a sense the OED defines as “physical contact between individuals involving sexual stimulation; sexual activity or behaviour, spec. sexual intercourse, copulation.”

The dictionary suggests that the association of the noun “sex” with sexual relations ultimately altered the old senses of “sex” and “gender” for the principal biological forms of humans and other creatures.

In the 20th century, Oxford says, “sex came increasingly to mean sexual intercourse” and “gender began to replace it (in early use euphemistically) as the usual word for the biological grouping of males and females.”

The dictionary adds that the noun “gender” “is now often merged with or coloured by a sense that developed in the mid-20th century in psychology and sociology:

“The state of being male or female as expressed by social or cultural distinctions and differences, rather than biological ones; the collective attributes or traits associated with a particular sex, or determined as a result of one’s sex. Also: a (male or female) group characterized in this way.”

The first Oxford citation for this new sense of “gender” is from “Sanity and Hazard in Childhood,” an article by Madison Bentley in The American Journal of Psychology, April 1945:

“In the grade-school years, too, gender (which is the socialized obverse of sex) is a fixed line of demarkation, the qualifying terms being ‘feminine’ and ‘masculine.’ ”

Despite the evolving meaning of “gender,” the entries for the term in some standard dictionaries include both the old biological and the new social senses.

Merriam-Webster, for example, has two definitions for “gender” used in the ways we’re discussing:

(1) “either of the two major forms of individuals that occur in many species and that are distinguished respectively as female or male especially on the basis of their reproductive organs and structures.”

(2) “the behavioral, cultural, or psychological traits typically associated with one sex.”

And a search with Google’s Ngram viewer, which compares terms in digitized books, suggests that “sex” and “gender” are still both being used in the old biological sense.

Merriam-Webster says in a usage guide that “among those who study gender and sexuality, a clear delineation between sex and gender is typically prescribed, with sex as the preferred term for biological forms, and gender limited to its meanings involving behavioral, cultural, and psychological traits.”

However, the dictionary adds, and we agree, that the “usage of sex and gender is by no means settled.”

Thanks for reading Capturing Voices! Subscribe for free to receive new posts and support my work.

Plummet’s journey

Plummet’s journey

Thanks for reading Capturing Voices! Subscribe for free to receive new posts and support my work.

Yann LeCun‘s evaluation of political versus linguistic errors:

His comment is no doubt meant as a joke, but it’s worth exploring the usage that bothers him.

To start with, the English word plummet has already been on a long morpho-syntactic and semantic journey (like nearly all other words). It started as a noun in Old French, plommet, the diminutive of plom “lead”, meaning “ball of lead, plumb bob“, referring to a (typically lead) weight attached to a “plumb line”. Wikipedia tells us that

The instrument has been used since at least the time of ancient Egypt to ensure that constructions are “plumb”, or vertical. It is also used in surveying, to establish the nadir (opposite of zenith) with respect to gravity of a point in space.

And also, used by sailors to estimate the depth of water.

The noun plummet, under various spellings, was borrowed into English as early the 14th century, and of course was also used metaphorically, as in this example from Shackerley Marmion’s 1632 play Hollands Leaguer:

And when I haue done, I’de faine see all your Artists,
Your Polititians with their Instruments
And Plummets of wit, sound the depth of mee.

And as always, the metaphorical extension got looser, to the point that the noun plummet came to be used to mean simply a “rapid fall” — though this usage seems to be relatively recent, with the OED’s first citation from 1957:

1957 After his plummet from fame, Keaton became a writer. Atchison (Kansas) Daily Globe

Also, like many other English nouns, plummet was soon used as a verb — though interestingly, the OED tells us that the first uses were transitive, connected to the depth-sounding sense of the noun:

1620 This ought to be the barre, cancell and limit of our too scrutinous nature, which often will assay to plummet the fathomlesse and bottomlesse sea of Gods most secret and hidden actions. T. Walkington, Rabboni

The (now more common) intransitive use, meaning (literally or figuratively) “To drop or fall rapidly or precipitously”, came a couple of hundred years later, with the OED’s first citation from 1845:

1845 Our capacity for delight plummeted. N. P. Willis, Dashes at Life with Free Pencil

OK, so what about the usage that bothered Yann LeCun:

In just 52 days, Trump has started a global trade war, plummeted the stock market, fired thousands of federal workers, slashed government funding, and sparked fears of a recession.

The author of that sentence has evolved plummet following the pattern of (the English version of) the causative-inchoative alternation:

The Causative/Inchoative alternation involves pairs of verbs, one of which is causative and the other non-causative syntactically and semantically (e.g., John broke the window vs. The window broke). In its causative use, an alternating verb is used transitively and understood as externally caused. When used non-causatively, the verb is intransitive and interpreted as spontaneous.

(Note that inchoative in this context means something like “change of state”, applied to the intransitive subject; and in the (transitive) causative version, the subject causes the object to the undergo the state change.)

There are many English verbs exhibiting this alternation — boil, melt, sink, open, bake, bounce, blacken, hang, close, cook, cool, dry, freeze, move, open, roll, rotate, spin, twist, shatter, thaw, thicken, whiten, widen, march, jump,

And it’s common in English to extend this pattern to create a causative transitive verb from an intransitive inchoative one, as I did with evolve in an earlier sentence (though others have done this before me…).

But attempts at such extension don’t always go smoothly, and plummet is not the only example of possible failure. Thus fall is an intransitive inchoative verb, but “*Those actions are going to fall the market” doesn’t work. Why?

This post is already too long, so for now I’ll just direct you to Beth Levin and Malka Rapaport Hovav, “A preliminary analysis of causative verbs in English“, Lingua 1994:

This paper investigates the phenomena that come under the label ‘causative alternation’ in English, as illustrated in the transitive and intransitive sentence pair Antonia broke the vase / The vase broke. Central to our analysis is a distinction between verbs which are inherently monadic and verbs which are inherently dyadic. Given this distinction, much of the relevant data is explained by distinguishing two processes that give rise to causative alternation verbs. The first, and by far more pervasive process, forms lexical detransitive verbs from certain transitive verbs with a causative meaning. The second process, which is more restricted in its scope, results in the existence of causative transitive verbs related to some intransitive verbs. Finally, this study provides further insight into the semantic underpinnings of the Unaccusativity Hypothesis (Perlmutter 1978).

Among other things, they note the difference between “verbs of manner of motion such as roll, run, jog, and bounce“, which have causative counterparts, and “verbs of directed motion such as come, go, rise, and fall“, which don’t. You can read the paper to learn their theory of why this matters — but we can note that the intransitive verb plummet is arguably in between those categories, interpretable either way.

None of the dictionaries that I’ve checked have a causative-transitive sense for plummet = “cause to fall rapidly”, and I don’t think I’ve ever heard or read it — but it wouldn’t be a shock to find other examples, and it’s also understandable that it would trigger someone’s “wrong!” reaction.

Thanks for reading Capturing Voices! Subscribe for free to receive new posts and support my work.

Mustn’t Have Done and Couldn’t Have Done

Mustn’t Have Done and Couldn’t Have Done

A reader has asked for a post on the difference between “mustn’t have + past participle” and “couldn’t have + past participle.” He gives these examples:

a) Ahmed failed the exam. He mustn’t have studied hard.
b) Ahmed failed the exam. He couldn’t have studied hard.

Thanks for reading Capturing Voices! Subscribe for free to receive new posts and support my work.

Before writing to me, the reader queried native English speakers of his acquaintance and received these answers.

• Some native speakers say that ONLY the first example is correct.
• Others say that both are correct.
• Some say that “mustn’t have + pp” indicates a conclusion based on evidence.
• Some say that “mustn’t have” suggests an 80% certainty, whereas “couldn’t have” provides 100% certainty.

Both a) and b) are correct.

The first statement is more likely to be spoken by a speaker of British English and the second by a speaker of US English. Either way, in this context, the speakers are merely speculating as to why Ahmed may have failed the exam. In this context, the constructions with mustn’t and couldn’t are interchangeable.

I have found numerous discussions of the mustn’t/couldn’t dichotomy in ESL forums. I don’t think I’d ever seen percentages of certainty applied to grammatical constructions before.

Degrees of certainty
Here is an illustration from an actual grammar book:

In answer to the question “Why didn’t Sam eat?”:

“Sam wasn’t hungry.” (The speaker is 100% sure that this is the reason.)

“Sam can’t have been hungry.” (The speaker believes – is 99% certain –that it is impossible for Sam to have been hungry.)

Sam must not have been hungry. (The speaker is making a logical conclusion. We can say he’s about 95% certain.)

“Sam might not have been hungry.” (The speaker is less than 50% certain, and is mentioning one possibility.)

Rather than assigning percentages of certainty to these constructions, it makes more sense to me to say that sometimes they convey certainty and sometimes they don’t. It all depends on context.

Here are examples in which mustn’t have and couldn’t have do indicate a conclusion based on evidence.

If the blood was still fresh that meant this murder mustn’t have been too long ago.

From the style of his writing he mustn’t be older than 30 years of age.

The car’s windows are darkly tinted, so Snell couldn’t have seen Johnson inside.

She couldn’t have understood the radio broadcast because she does not speak Dutch.

The evidence for the conclusion lies in the sentence itself.

the freshness of the blood.

the writing style.

the windows were too dark to see through.

the listener did not know the language.

Other contexts
Lacking internal evidence, the application of percentages to the “certainty” of the meaning of these two constructions is an exercise in futility.

The following examples can convey ideas other than certainty.

You mustn’t have spent much time in New York. (sarcasm?)

He mustn’t have finished his homework on time. (Maybe he didn’t do it at all)

She couldn’t have tried very hard. (Maybe she tried as hard as she could, but lacked the necessary ability.)

Thanks for reading Capturing Voices! Subscribe for free to receive new posts and support my work.

Mustn’t Have Done and Couldn’t Have Done

Mustn’t Have Done and Couldn’t Have Done

A reader has asked for a post on the difference between “mustn’t have + past participle” and “couldn’t have + past participle.” He gives these examples:

a) Ahmed failed the exam. He mustn’t have studied hard.
b) Ahmed failed the exam. He couldn’t have studied hard.

Thanks for reading Capturing Voices! Subscribe for free to receive new posts and support my work.

Before writing to me, the reader queried native English speakers of his acquaintance and received these answers.

• Some native speakers say that ONLY the first example is correct.
• Others say that both are correct.
• Some say that “mustn’t have + pp” indicates a conclusion based on evidence.
• Some say that “mustn’t have” suggests an 80% certainty, whereas “couldn’t have” provides 100% certainty.

Both a) and b) are correct.

The first statement is more likely to be spoken by a speaker of British English and the second by a speaker of US English. Either way, in this context, the speakers are merely speculating as to why Ahmed may have failed the exam. In this context, the constructions with mustn’t and couldn’t are interchangeable.

I have found numerous discussions of the mustn’t/couldn’t dichotomy in ESL forums. I don’t think I’d ever seen percentages of certainty applied to grammatical constructions before.

Degrees of certainty
Here is an illustration from an actual grammar book:

In answer to the question “Why didn’t Sam eat?”:

“Sam wasn’t hungry.” (The speaker is 100% sure that this is the reason.)

“Sam can’t have been hungry.” (The speaker believes – is 99% certain –that it is impossible for Sam to have been hungry.)

Sam must not have been hungry. (The speaker is making a logical conclusion. We can say he’s about 95% certain.)

“Sam might not have been hungry.” (The speaker is less than 50% certain, and is mentioning one possibility.)

Rather than assigning percentages of certainty to these constructions, it makes more sense to me to say that sometimes they convey certainty and sometimes they don’t. It all depends on context.

Here are examples in which mustn’t have and couldn’t have do indicate a conclusion based on evidence.

If the blood was still fresh that meant this murder mustn’t have been too long ago.

From the style of his writing he mustn’t be older than 30 years of age.

The car’s windows are darkly tinted, so Snell couldn’t have seen Johnson inside.

She couldn’t have understood the radio broadcast because she does not speak Dutch.

The evidence for the conclusion lies in the sentence itself.

the freshness of the blood.

the writing style.

the windows were too dark to see through.

the listener did not know the language.

Other contexts
Lacking internal evidence, the application of percentages to the “certainty” of the meaning of these two constructions is an exercise in futility.

The following examples can convey ideas other than certainty.

You mustn’t have spent much time in New York. (sarcasm?)

He mustn’t have finished his homework on time. (Maybe he didn’t do it at all)

She couldn’t have tried very hard. (Maybe she tried as hard as she could, but lacked the necessary ability.)

Thanks for reading Capturing Voices! Subscribe for free to receive new posts and support my work.

5 Words You Use All the Time That Were Once Totally Unacceptable

5 Words You Use All the Time That Were Once Totally Unacceptable

Think language rules are set in stone? Think again. This fun and insightful piece from Time Magazine highlights five everyday English words that were once considered unacceptable or even flat-out wrong. It’s a great reminder that language is always evolving—and what’s “incorrect” today might be standard tomorrow.

Thanks for reading Capturing Voices! Subscribe for free to receive new posts and support my work.

Attention language lubbers: the fourth edition of Garner’s Modern English Usage is out this month, and while it’s as full of pedantry as any guide on how to properly wield the English language, the new version is also funnier and more data-driven than any reference you’ve likely come across in school.

By using searchable databases like Google Ngrams, lexicographer Bryan Garner has developed ratios for how often words or phrases, one standard and one upstart, are used in published writing. And he has developed his own five-stage system for how widespread a “mistake” has become, from being something sloppy and offensive-to-the-educated to perfectly unobjectionable.

Garner describes himself as something of a linguistic “epidemiologist,” tracking mistakes that spread within a population, sometimes escalating to pandemic levels and then, perhaps, reaching a point at which human bodies have all adapted to the virus and it becomes the new normal.

Garner provides several charming analogies in his preface for how readers of the book should understand his five-stage system of acceptability, one of which is this:

Stage 1: Audible flatulence
Stage 2: Audible belching
Stage 3: Overloud talking
Stage 4: Elbows on the table
Stage 5: Refined

Take buck naked and butt naked.

While professional editors still tend to stick to the more traditional buck naked—which may have been inspired by “the skin of bare buttocks” being reminiscent of a male deer’s soft hide—butt naked is by far the favorite in popular usage these days, according to Garner’s analysis. He categorizes the use of butt naked as “Stage 4,” meaning everybody’s using it “except a few stalwart holdouts” who might be principled but should also be preparing for defeat.

As Garner says, “There’s almost a universal rule that any two words that resemble each other in some way will be confounded.” Here follow five examples of confounded words that have shifted from being abhorred to standard.

contact (as a verb instead of a noun): Stage 5

We might think of this verb as being as normal as to be or to not be today. But there was a time when using contact as a verb (“Contact me later”) rather than a noun (“Avoid contact with angry bees”) had grammarians snapping their monocles. This is a good example, Garner notes, of a usage that had appeal because it was efficient. “It was a huge bugbear in the 1920s and 1930s,” he says of people using it as a verb. “But people like brevity. They don’t want to say get in touch with, and they don’t want to say call, email, text or leave a note at my door.” Gradually, he says, “it’s lost all the bad odor.”

nauseous instead of nauseated : Stage 4

Have you ever said you felt nauseous? In the traditional sense that would mean you felt like you were capable of causing others to woof cookies, not that you were feeling sick to your own stomach—much along the lines of how poisonous and poisoned work. Garner, who is something of a stalwart holdout himself, views nauseous as a “skunked term,” a label he coined for a word that gives off “a bad order” to someone whether it’s used correctly or incorrectly. His advice: just don’t use it at all.

graduate from instead of to be graduated from: Stage 5

Your great-great-great grandparents, Garner says, might have insisted that they “were graduated from” college, because it was the school that was bestowing a level of achievement on graduates after all. By the 19th century, your grandparents might have “graduated from” college, and if you’re over 30, you probably did too. Young people today, he says, are now dropping the “from” and simply saying they “graduated college,” which he classifies as Stage 3. “I wish I didn’t have uncharitable thoughts every time I hear somebody say I graduated college,” Garner says, “but I do.”

spitting image instead of spit and image: Stage 5

As Garner explains, the original spit and image came “from the notion of God’s using spit and dust to form the clay to make Adam in his image.” In fact, people used to say things like, “Tom is the very spit of his uncle!” But the corruption has become so common that using the original today might not only stop a conversation in its tracks but cause unpleasant face-scrunching. Per Garner, spitting image is now 23 times more commonly used than its precursor.

self-deprecating instead of self-depreciating: Stage 5

The oldest meaning of deprecate is to “pray for deliverance from,” which makes the notion of being “self-deprecating” pretty hilarious, or, as Garner puts it, “a virtual impossibility, except perhaps for those suffering from extreme neuroses.” And yet over time, this word has taken on the meaning of that slightly longer word that looks so much like it: depreciate means to belittle, to lessen the value of. These days someone might even try to correct you if, in an attempt to note someone was being (overly) humble, you said they were self-depreciating.

Did this just happen because the shorter word is easier to say? Partly, Garner says. Then again that makes little sense when trying to account for why people use the less-standard preventative instead of preventive or irregardless instead of regardless, he notes. Some linguists would argue that there’s no point fighting against slips like that—that language is forever unfixed and deviations should simply be observed and even appreciated—or that it’s silliness to tell people to follow rules that are as arbitrary as the meaning assigned to a certain jumble of letters. But Garner is not one of them.

“There are a lot of people who mistakenly think intelligibility is the standard. ‘Oh, you knew what I was saying.’ Well, that’s not the standard. That’s a really bottom-of-the-barrel standard,” he says. “People who are concerned with English usage usually want to have their words taken seriously, either as writers or as speakers. And if you don’t use the language very well, then it hard to have people take your ideas seriously. That’s just the reality.”

Thanks for reading Capturing Voices! Subscribe for free to receive new posts and support my work.