In this Sunday's New York Times Magazine, I take over the "On Language" spot to pay tribute to the man who originated the column, William Safire. (You can already read the online version here.) It's not quite as personal as the remembrance I posted here after learning of Safire's death, but it's no less heartfelt. As preparation, I took a stroll through some of the thousands of columns that Safire produced over three decades, focusing especially on his first year of language punditry, 1979. Though many of his early columns stand the test of time, one example where he was less than on-target had to do with a popular peeve: "could care less."

On September 30, 1979, Safire devoted a few paragraphs of his column to consider how in common parlance "I could care less" means quite the opposite of what it says, namely, "I could not care less." Right out of the box, he asserts that this usage "seems to be petering out." He goes on to write:

The phrase first popped up around 1960, appeared in a letter to columnist Ann Landers in 1966 and in a few years began to cause concern as a barbaric attack on meaning. ... Usage seems to have peaked in 1973, when The Wall Street Journal headlined: "More and More Girls Flip for Gymnastics: Boys Could Care Less." A healthy derision set in. ... Maybe the attacks on the antimeaning helped; eventually, like most vogue phrases, it wore out its welcome.
Farewell, "could care less"! You symbolized the exaltation of slovenliness, the demeaning of meaning, and were used by those who couldn't care less about confusing those who care about the use of words to make sense.

With thirty years of hindsight, we can see that Safire was far too quick to dismiss "could care less" as a mere "vogue phrase" that was already on its way out. If anything, this turn of phrase has grown more entrenched over the years, and it's showing no signs of abatement in American English.

Hindsight also allows us to pin down the approximate age of the expression a bit more accurately, thanks to searchable newspaper databases. The year 1960 was a decent ballpark estimate (that was, in fact, when Ann Landers first fielded a reader complaint about "could care less" in her advice column), but I've been able to take it back another five years. Here is sportswriter Shirley Povich using it in 1955:

"The National League clubs have always shied from pitching left-handers against the Dodgers, but Casey Stengel could care less about the Dodgers' reputation for beating southpaws."
—"This Morning . . . With Shirley Povich," Washington Post, Sep. 25, 1955, p. C1

Notably, this is a mere 11 years after the first known usage of "couldn't care less," which I found in the archives of the Chicago Tribune:

"I couldn't care less, darling," said Frederica who, being on duty in the ward, could not go to the party.
—"Danger List" by Christianna Brand, Chicago Tribune, May 15, 1944, p. 18

When I brought these citations to the attention of University of Pennsylvania linguist Mark Liberman, who wrote about the expression on Language Log in 2005, he commented: "Given the expected resistance of editors to 'could care less', the fact that it appears in print 11 years after the first citation for 'couldn't care less' suggests to me that the two expressions probably arose at essentially the same time, like quark-antiquark pairs in a high-energy collision."

It's true that newspaper editors of the past were resistant to printing "could care less" in their pages, at least in the early decades of usage. But they have by and large abandoned that struggle, despite Safire's wishful thinking that the expression was breathing its last gasp in 1979. Safire's own newspaper, the New York Times, is typically very conservative when it comes to usage matters, but even the Times has thrown in the towel. I compared results for "couldn't care less" versus "could care less" in the Times archive over the past four decades, and this is what I found:

"couldn't care less" "could care less"
ratio
1960-1969 245    5    2.04%
1970-1979 369    42    11.38%
1980-1989 274    77    28.10%
1990-1999 248    147    59.27%
2000-2009 259    194    74.90%

These numbers should be taken with a grain of salt, especially in the earlier decades, since the character recognition of scanned pages isn't entirely accurate. Also, there could be a number of false matches: an appearance of "no one could care less," for instance, should not count towards the "modern" sense of the expression, since it has the same negative polarity as the original, more logical, "couldn't care less." Despite such noise in the data, there's a very clear upward curve in the proportion of usage. While "couldn't care less" has held steady, the frequency of "could care less" has nearly tripled in the pages of the Times since Safire hastily declared it extinct.

Other databases of U.S. news outlets confirm this trend. (Elsewhere the historical rise has been more gradual, but the 75% ratio of "could" to "couldn't" currently found in the Times accords with other 21st-century journalistic sources.) Editors no longer resist printing "could care less," despite its seeming illogicality. The fact is, this is simply how people talk, and contemporary newspapers seek to reflect that reality. Mark Liberman reports that "could care less" has in fact been much more common in American speech for at least a decade or so. Based on a survey of sources compiling spoken American English, he estimates that "could care less" now outpaces "couldn't care less" by a ratio of about 5 to 1.

There are plenty of theories for how this state of affairs came to pass. (See, for instance, this post by Arnold Zwicky elsewhere on Language Log for a discussion of competing linguistic analyses, with further discussion linked here.) Regardless of how we got here, there's no denying that "could care less" is here to stay — even if you agree with Safire's initial assessment of it as a "barbaric attack on meaning."

One important lesson to learn from all of this: as the linguist David Crystal warned us in an interview last year, "Never predict the future with language!" Over time, I think Safire learned this lesson and contented himself in the role of language observer, not prognosticator.