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.

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.

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.)

Q. Is it grammatically accurate to say something like, “I’m going to dress warmly”? My hunch is no, because “dress warmly” means that I’ll be smiling and emotionally warm as I’m dressing, given that “warmly” modifies the verb “dressing.” If all that is true, then what I’m unsure about is how to fix the sentence. Can you suggest any good alternatives besides writing around it like so: “I’m going to dress in warm clothing”?

A. If you say you’re going to dress warmly, that means you’re going to put on warm clothes, whereas addressing someone warmly would mean greeting that person with affection or kindness. Words often have more than one sense depending on how they’re used; according to Merriam-Webster (among other dictionaries), warmly is no exception.

Or maybe you’re thinking of the feel badly versus feel bad principle. That’s different, though, because unlike dress, the word feel can be a linking verb. Linking verbs reflect the predicate back onto the subject—as when you feel bad, where the adjective bad modifies the pronoun you. But when you feel badly, you are literally not good at feeling something (either physically or emotionally). See also CMOS 5.175.

Q. How should one style the title of a work in a discussion not of that work, but of its title. As an example, consider the following sentence:

The novel’s title, “Pride and Prejudice,” refers to a pair of traits seen in all of its characters.

Should the title be set in roman and within quotation marks because it is a phrase being mentioned (rather than used)? Or does the fact that it IS a title prevail, so that it should be italicized and without quotation marks? Or perhaps some tertium—or even quartum—quid? My sense is that because in that sentence its referent is not Austen’s book itself but the character flaws that recur in its plot, the italics would be inappropriate. Do I have that right?

A. You might be overthinking this. The novel’s title is Pride and Prejudice, a three-word italic phrase that names a pair of traits exhibited by many people, including the characters in that book. Chicago-style italics for book titles doesn’t prevent you from discussing what the words mean.

But if you really want to get your readers to home in on the title words as words, try something like this: “The nouns in the novel’s title, pride and prejudice, refer to a pair of traits . . .” or “The nouns in the novel’s title, ‘pride’ and ‘prejudice,’ refer to a pair of traits . . .”

For the use of italics or quotation marks to refer to words as words (either treatment is correct), see CMOS 7.66.

Q. Does CMOS have guidance for the White House’s recent changes to the names of the Gulf of Mexico and Denali?

A. Relative to matters of style, yes: Spell a generic word like gulf with an initial capital when it’s used as part of a proper name but not otherwise (“the Gulf of Mexico,” but “the gulf”; see CMOS 8.54); don’t add “Mount” before the name Denali (see 8.56); and spell out “Mount” in names like Mount McKinley rather than abbreviating it as “Mt.” (see 10.35).

Other than that, it’s easy enough to confirm that the names of those two geographic entities were officially changed (at least in the United States) to the Gulf of America and Mount McKinley, respectively, in accordance with an executive order issued on January 20, 2025, by the White House. What an author does with that information will then depend on various factors, including a publisher’s house style (if any) and considerations related to historical accuracy.

For additional guidance, see the AP Stylebook, which issues regular updates geared toward those who cover the news. In a pair of updates added on January 30, 2025, that guide says to use the original name for the gulf “while acknowledging the new name,” but, for the mountain, it says that “the Associated Press will use the new official name of Mount McKinley.” For the gulf, AP points to the long history of the older name together with the fact that the gulf shares its borders with Mexico; for the mountain, AP cites the fact that it lies entirely within the United States, which lends broader authority to that name change.

Q. With coordinate adjectives separated by a conjunction, there’s no comma: “A stable and sensible approach.” I assume it would be the same for contrasting adjectives: “A sensible yet volatile approach.” Though if you wanted to emphasize the volatility, you might set it off with commas: “A sensible, yet volatile, approach.” Does this all sound right?

A. That sounds right to us, though whenever you interrupt an adjective-plus-noun construction with an intervening phrase set off by commas, the result tends to be a little awkward. If you want to smooth things out while keeping the emphasis on volatility, try rephrasing. For example:

an approach that’s sensible yet also volatile

a sensible approach, albeit a volatile one

Or you could embrace the interruption by applying something stronger than commas:

a sensible—yet volatile—approach

a sensible (yet volatile) approach

See also CMOS 6.51.

Q. I work in curriculum. I need to be able to spell out large numbers so as to model how to read numerals correctly. I can find rules for when to hyphenate whole numbers, but I can’t find any for hyphenating decimals. Specifically, I need to know when to hyphenate the words to the right of the decimal (tenths/hundredths, etc.). Please advise. Thanks so much!

A. That’s a challenging question! Let’s start with a few numbers and how we would suggest spelling them out—on both sides of the decimal point:

1,357,201.5: one million, three hundred fifty-seven thousand, two hundred one and five tenths

1,357,201.58: one million, three hundred fifty-seven thousand, two hundred one and fifty-eight hundredths

1,357,201.580: one million, three hundred fifty-seven thousand, two hundred one and five hundred eighty thousandths

1,357,201.5803: one million, three hundred fifty-seven thousand, two hundred one and five thousand eight hundred three ten-thousandths

Note that we’ve used commas between groups of numbers to the left of the decimal point but not to the right, which reflects how the numbers are grouped (and punctuated) as digits. But we’ve treated the numbers to the left of the decimal point as a single value rather than as a series (which might require a serial comma before and).

The rules for hyphenation are the same on both sides of the decimal point (see CMOS 7.96, section 1, “numbers, spelled out”). But note that Chicago’s preference for hyphenating simple fractions doesn’t apply to five tenths in the first example above, which simply names the number in the tenths place. It does, however, apply to an ordinal fraction like ten-thousandths:

five-tenths (a simple fraction)

five tenths (the number in the tenths place; see first example above)

ten-thousandths (the ten-thousandths place; see last example above)

See also CMOS 7.96, section 1, “fractions, simple.” Finally, note that some writers add and when spelling out certain numbers that include hundred (three hundred and fifty-seven thousand; two hundred and one); Chicago omits this and (see CMOS 9.5).

Q. I am wondering how best to cite, within one chapter of a multiauthor book, other chapters from the same volume. I am accustomed to simply adding “(see chapter X)” to the text, but one author is pushing back and wants to see them in the reference list. We are using author-date style.

A. Chapters in multiauthor books are likely to be consulted separately (and are sometimes offered for individual download or sale), so listing other chapters from that same book in the reference list at the end of your own chapter wouldn’t be the worst idea. But you should still let readers know that the chapter you’re citing is in the same book.

We’d suggest adding this information to the author-date reference for that chapter in your text—for example, like this: “(Smith 2025, in this volume).” The corresponding entry in your reference list would include the book’s title (per CMOS 13.109), so a similar comment shouldn’t be needed there:

Smith, Jane. 2025. “Chapter Title.” In Title of Book, edited by Joe Anyone. Publisher details.

If you cite more than a few chapters from your book, you may want to use a shortened form for the book: “In Anyone, Title of Book.” But even if you do that, you shouldn’t need to add a separate entry in the reference list for the book as a whole, the identity of which should be obvious to anyone who is already reading something from that book. For some additional considerations, see CMOS 14.10, under “author-date.”

Q. CMOS 13.128 shows how to use author-date citations in a footnote, but what about an informational footnote like, “The history of exclusion of Chinese people in the United States has been highly researched. To begin, see . . .”? Should the parentheses around the citations be removed, as in “To begin, see Chan 1991, Lee 2003, and Kurashige 2016”? Otherwise, it might seem as though the citations are substantiating the statement, rather than being offered as suggested reading.

A. In Chicago style, the “see” in “see Chan” means that you’re referring to a work rather than a person, and the year would retain parentheses whether in the text or in a note: “To begin, see Chan (1991), Lee (2003), and Kurashige (2016).” If you’re instead referring to the author in terms of the work, the wording would need to make that clear: “See the earlier efforts by Chan (1991) to digitize the archival records.”

Parentheses for the year are omitted only when the citation is itself in parentheses, in which case semicolons rather than commas separate the sources, as in “(to begin, see Chan 1991; Lee 2003; Kurashige 2016).” But if your parenthetical reference is to an author rather than a work, the year would get square brackets: “(See the earlier efforts by Chan [1991] to digitize the archival records.)” See CMOS 13.122 and 13.124.

Q: Am I showing my age? Once upon a time, a couple of decades ago, when you were ill or had an emergency, you would “call in sick.” Now it’s “calling out.” When did that happen?

A: People have been calling “in,” “out,” and “off” sick for dozens of years, but “call in sick” is the oldest and by far the most common expression for reporting one’s absence from work because of illness.

The Oxford English Dictionary defines “call in sick” as “to contact one’s employer, school, etc., typically by telephone, to report one’s absence that day, esp. due to illness.” The OED’s earliest example is from the 1940s:

“This being a holiday weekend, employees in Treasury’s loans and currency section … were warned yesterday not to call in sick either today or Monday under any circumstances” (The Washington Post, July 3, 1943).

The dictionary’s first citation for “call off sick” is from a West Virginia newspaper: “Personnel who frequently call off sick … should be checked at their homes to ascertain the legitimacy of their absence” (Charleston Daily Mail, May 24, 1958).

The earliest OED example for “call out sick” is from a Massachusetts newspaper: “Bray said no one called out sick in the DPW at all this week … [due] to his demands that anyone out sick must have a doctor to certify illness” (Sentinel & Enterprise, Fitchburg, April 16, 1976).

A similar expression, “call off work,” has been around since the the mid-1960s and means “to report one’s absence from (work, school, etc.), typically by making a telephone call,” Oxford says.

The dictionary’s first citation is from a 1965 labor arbitration decision: “He would receive a final warning if he didn’t improve on the tardiness, absenteeism, and calling off work without notice” (from Labor Arbitration Awards, Vol. 65-2).

A search with Google’s Ngram Viewer, which compares words and phrases in digitized books, indicates that “call in sick” is clearly the most common of the four expressions.

We’ve seen suggestions online that “call out sick” may be especially popular in New York, New Jersey, and Connecticut, but we’ve seen no linguistic evidence to support this. The Dictionary of American Regional English doesn’t mention the usage.

Scientists have made new improvements to a “brain decoder” that uses artificial intelligence (AI) to convert thoughts into text.

Their new converter algorithm can quickly train an existing decoder on another person’s brain, the team reported in a new study. The findings could one day support people with aphasia, a brain disorder that affects a person’s ability to communicate, the scientists said.

A brain decoder uses machine learning to translate a person’s thoughts into text, based on their brain’s responses to stories they’ve listened to. However, past iterations of the decoder required participants to listen to stories inside an MRI machine for many hours, and these decoders worked only for the individuals they were trained on.

“People with aphasia oftentimes have some trouble understanding language as well as producing language,” said study co-author Alexander Huth, a computational neuroscientist at the University of Texas at Austin (UT Austin). “So if that’s the case, then we might not be able to build models for their brain at all by watching how their brain responds to stories they listen to.”

In the new research, published Feb. 6 in the journal Current Biology, Huth and co-author Jerry Tang, a graduate student at UT Austin investigated how they might overcome this limitation. “In this study, we were asking, can we do things differently?” he said. “Can we essentially transfer a decoder that we built for one person’s brain to another person’s brain?”

The researchers first trained the brain decoder on a few reference participants the long way — by collecting functional MRI data while the participants listened to 10 hours of radio stories.

Then, they trained two converter algorithms on the reference participants and on a different set of “goal” participants: one using data collected while the participants spent 70 minutes listening to radio stories, and the other while they spent 70 minutes watching silent Pixar short films unrelated to the radio stories.

Using a technique called functional alignment, the team mapped out how the reference and goal participants’ brains responded to the same audio or film stories. They used that information to train the decoder to work with the goal participants’ brains, without needing to collect multiple hours of training data.

Next, the team tested the decoders using a short story that none of the participants had heard before. Although the decoder’s predictions were slightly more accurate for the original reference participants than for the ones who used the converters, the words it predicted from each participant’s brain scans were still semantically related to those used in the test story.

For example, a section of the test story included someone discussing a job they didn’t enjoy, saying “I’m a waitress at an ice cream parlor. So, um, that’s not … I don’t know where I want to be but I know it’s not that.” The decoder using the converter algorithm trained on film data predicted: “I was at a job I thought was boring. I had to take orders and I did not like them so I worked on them every day.” Not an exact match — the decoder doesn’t read out the exact sounds people heard, Huth said — but the ideas are related.

“The really surprising and cool thing was that we can do this even not using language data,” Huth told Live Science. “So we can have data that we collect just while somebody’s watching silent videos, and then we can use that to build this language decoder for their brain.”

Using the video-based converters to transfer existing decoders to people with aphasia may help them express their thoughts, the researchers said. It also reveals some overlap between the ways humans represent ideas from language and from visual narratives in the brain.

“This study suggests that there’s some semantic representation which does not care from which modality it comes,” Yukiyasu Kamitani, a computational neuroscientist at Kyoto University who was not involved in the study, told Live Science. In other words, it helps reveal how the brain represents certain concepts in the same way, even when they’re presented in different formats.

The team’s next steps are to test the converter on participants with aphasia and “build an interface that would help them generate language that they want to generate,” Huth said.