I. Houston, I Feel Like We Have a Problem

I feel like we say "I feel like." Like, a lot. The phrase is so ubiquitous that it's invisible, but then some nerd points it out and it's all you can hear.

When did we all start doing this? Why did we start doing this? Should we be doing this?

On one hand, it seems useful. Many people I've asked about it say they learned to say "I feel like" in various conflict-resolution lessons in middle school and high school. It's a distancing tactic. The thinking is that if you couch your opinion as a private emotion, other people won't feel as if you are directly attacking them. That makes sense to me.

On the other hand, using a big, wimpy hedge to obscure nearly every criticism, complaint, observation, or opinion seems annoying, insulting, and transparently deceptive. It's annoying because it gives me extra work to do: If we're having a conversation and you say, "I feel like..." suddenly it's my job to establish the exact nature of your commitment to your own opinion. It's insulting and transparently deceptive because, in those cases when you're using the phrase in a snarky, passive-aggressive way, I feel the need to glare at you and say something like "We're not in a middle-school health class. Please describe your grievance, and I will do my best to help." For the sake of efficiency and honesty, shouldn't we just state plainly our criticisms and desires?

II. I Made a Mistake That I Feel Like I Should Tell You About

For six years, I taught at the college level.

In the grad-school pedagogy courses I took to prepare me for teaching, I read several essays about gender differences among students. One of the foundational texts is called "Sexism in the Classroom," by Maya and David Sadker, which basically says that instructors and students often fall into stereotypical gender roles in the classroom. Women don't speak as often as men, and when they do speak, they often preface their answers with phrases like "I'm not sure if this is right, but..." "I could be wrong..." and, of course, "I feel like..." Instructors, Sadker and Sadker claimed, needed to be aware of these patterns and needed to make a concerted effort to prevent college classrooms from reflecting the sexism of mainstream American culture.

That essay was written back in the mid-1980s, but Sadker and Sadker's thesis held true in all the courses I taught. While I never had much trouble getting women to talk in class, I noticed that they seemed to begin their claims with big qualifiers, and I was always conflicted about what I should do.

The pedagogical texts seemed to suggest encouraging women not to hedge. However, I wasn't very comfortable perpetuating the patriarchy by telling a woman how she should talk. I did, though, want to prepare them for the boardroom or the surgery room, where people perceive heavily qualified statements as soft or weak.

So every once in a while, I'd stop a student after class and tell her that she didn't have to say "I feel like..." or "I'm not sure if this is right..." before every answer because (1) she was typically correct, and (2) she needed to be direct with her claims in order to compete with other people, typically men, who have been socially conditioned to make direct, unqualified claims all their lives.

Once, after I talked with a student about this, she responded with something like "No no no, Mr. Smith. I qualify like that because if I'm wrong then I'm right because I said I could be wrong, but if I'm right then I seem humbly right and don't come off like a snob."

I thought her answer was so clever that I never again held another student after class for that reason. But was I dooming this clever woman—and the rest of the women I taught in the following years—to a life devoid of leadership positions? What about the boardroom? What about the surgery room?

I have since consulted several experts—including Matthew Gordon, an associate professor of linguistics at the University of Missouri, and Marisa Brook, a fifth-year PhD student in sociolinguistics at the University of Toronto, who recently researched the origins of "I feel like"—and I learned many things. Two things are particularly important: (1) The history of "I feel like" goes back much further than one might think, and (2) my impulse to tell women to use more direct speech was misguided and based in the very sexism I was trying to combat.

III. First, Let's Talk About Different Kinds of "Feels"

The old-fashioned usage of "I feel" simply means that you feel something, as in, "I feel lonely." In this case, you're citing a specific feeling of loneliness. You can get metaphorical with it, too, as in: "I feel like a crab in a sack." Now you're equating your state with the state of a crab alone in a sack.

The newer usage means "my opinion is"—as in, "I feel that I'm lonely." You could also say, "I feel like I'm lonely." Here, you've become your own therapist. You guess you're lonely, but you're not really sure—maybe your problem is something else entirely. The same thing holds true for the metaphor "I feel like I'm a crab in a sack." Perhaps you do feel like a crab in a sack, but with this construction you leave room for argument, for someone to say, "That's impossible! You're not a crab, you're a person! Get ahold of yourself!"

But what's the cause of this change from "I am feeling" to "my opinion is"? To answer that question, we need to consult the graduate student I mentioned earlier.

Marisa Brook has been researching the change happening with "I feel," and also the rise of "I feel like" specifically. Citing the research of Sandra A. Thompson and Anthony Mulac, Brook notes that the phrase "I think" went through a similar transition in the 20th century. That is, "I think" used to just mean "I am thinking," before it started to mean "my opinion is." She also cites Minna Palander-Collin, who argues that the same transformation occurred with the Renaissance favorite "methinks" in the 15th and 16th centuries. All of this is to say that there's lots of precedent for a shift in meaning from "I am engaging in mental activity" to "my opinion is," both of which are intimately related notions anyway.

Brook's data shows that the use of "I feel" to mean "my opinion is" started gaining purchase with people who were born around 1910, but the first unambiguous example of that usage comes from a Canadian speaker born in 1891, who was recorded in 1975 saying "I feel that" and "I feel" before stating her own opinions.

There's another small but important change going on during this time, too. "I feel" isn't just being used to mean "my opinion is," it's also being used as a strategy to distance the speaker from his or her own statements. For instance, the phrase "I feel that you're not listening to me" could mean "I have the opinion that you're not listening to me," but it could also mean "I definitely know that you're not listening to me, but I want to be polite about pointing it out, so I'm being indirect."

IV. Now, Let's Talk About "Like"

I'm now going to describe why I was wrong to correct that woman in my class in the wonkiest possible (but most exact) way. First, we need to get "I feel" and "like" in the same room together.

Brook told me that the phrases "I feel" and "like" have been going through changes at approximately the same time. As I mentioned earlier, the phrase "I feel" used to just mean "I am feeling," e.g., "I feel dumb." Recently, though, it has started to mean "my opinion is," e.g., "I feel like I'm dumb." Well, "like" has been going through some changes of its own.

In English, "like" can be used as something called a "comparative complementizer," other examples of which include "as though," "as if," or "that."

This is about to get pointy-headed, but stay with me—I promise it'll be worth it.

The verb "feel"—when used in this way we're talking about, e.g., "I feel like I'm not understanding this"—announces that a finite subordinate clause is coming even before you get to that "like." (You do not need to know what a finite subordinate clause is, but I have to use that term in order for the rest of this to be true.) Only four other verbs can be used to announce this kind of clause: "look," "sound," "seem," and "appear." These are all sensory verbs that we use to describe our perceptions of the world. In order to link one of these verbs—"feel," "look," "sound," "seem," or "appear"—to the finite subordinate clause, you need to get yourself a "comparative complementizer," like "like."

Brook's research shows that from 1860 to 1930 in Canada, "like" rose above the other comparative complementizers because people used "like" to describe metaphorical or literal things, whereas people used "as if" and "as though" for constructing metaphors. They used "that" (or nothing at all) to describe literal things. For example, someone would be much more likely to say "I feel like a crab in a sack" or "I feel as if I'm a crab in a sack" than they would be to say "I feel that I'm a crab in a sack," because a crab in a sack is a metaphor.

But people use "like" with literal things just as readily as they use "like" with metaphorical things. Since "like" doesn't pick a side, "like" wins.

During the late 19th and early 20th century, "I feel" began to take on the meaning "my opinion is" and also "I don't wanna stir up any shit here, but..." At the same time, people committed to "like" as their go-to comparative complementizer because it was so versatile. Brook told me (and Google's Ngram Viewer confirmed) that the phrase "I feel like" started blowing up in the 1970s, and then it really blew up, usage-wise, around the millennium, and especially among young women.

(If you want to see this for yourself, go to books.google.com/ngrams and type "I feel like" in the search bar.)

"Most linguistic changes are led by young women," Brook points out. "But not everything young women say disproportionately is linguistic change." That said, the fact that young women are saying "I feel like" a lot supports the notion that the phrase is new and is changing.

V. It's Worth Repeating—Most Linguistic Changes Are Led by Young Women

Does a woman's use of "I feel like" align with stereotypes of feminine insecurity and passiveness? Should educators and friends and spouses follow Sheryl Sandberg's lead and instruct women to "lean in," to state their opinions without qualification as men supposedly do, so that they might snatch up more leadership positions?

Um, no. No they shouldn't. Also, that question is wrong.

Dr. Mark Liberman is a big-deal American linguist who specializes in phonetics and teaches at the University of Pennsylvania. Along with Geoffrey Pullum, he runs the blog Language Log, which is a digital temple of word nerdery. Liberman's post on the phrase "I feel like" confirms Brook's finding that women are ahead of the game on "I feel like," and shows me how wrong I was to pull my student aside and tell her to stop qualifying her answers.

Liberman analyzed a block of data from the Linguistic Data Consortium, which is essentially a humongous collection of conversations donated by the government and by private entities. He looked at a bunch of transcribed telephone conversations, most of which occurred around 2003. And in his "I feel like" post, he describes the ways in which men and women use different kinds of phrases to hedge their claims.

He found that women use "I feel like" and "I think" more than men do, but this fact doesn't support the notion that women hedge more often than men.

Men preface their claims with the phrases "I guess" and "I believe" more than women do, plus they use inter-clausal qualifiers such as "seems," "somewhat," and "probably" more often than women.

When you look at the numbers, Liberman concludes, men hedge just as often as women do, they just use slightly different words to do it. So the fact that women say "I feel like" more than men only suggests what everybody in linguistics already knows, which is that women tend to lead language change. In this way, their use of "I feel like" is a kind of leadership, not a sign of passivity.

So it appears as if I only felt like the women in my classes were hedging more often than the men. I wasn't listening closely enough to hear the sneakier qualifiers that the men were using within their sentences. When I think back, I don't remember ever saying, "What do you mean Eliot is 'probably' using the metaphor of the crab scuttling against the floors of silent seas to describe his loneliness? He's either doing it or he's not doing it, Eric! Be more direct or they'll eat you alive in boardroom!"

My urge to "help" women "lean in" was misguided, given that men lean back just as much.

When I confessed this to Dr. Matthew Gordon, he put me in my place like only a linguistics scholar can. He said, "The general effort to get women to avoid hedges is certainly well-intentioned but also part of a long tradition of advising women to reshape their linguistic behavior to be more like men's." recommended