How well can you communicate over email? (or blog posts? how about in person?)

Justin Krueger, Nicholas Epley, Jason Parker, and Zhi-Wen Ng write (pdf):

Without the benefit of paralinguistic cues such as gesture, emphasis, and intonation, it can be difficult to convey emotion and tone over electronic mail (e-mail). Five experiments suggest that this limitation is often underappreciated, such that people tend to believe that they can communicate over e-mail more effectively than they actually can. Studies 4 and 5 further suggest that this overconfidence is born of egocentrism, the inherent difficulty of detaching oneself from one’s own perspective when evaluating the perspective of someone else. Because e-mail communicators “hear” a statement differently depending on whether they  intend to be, say, sarcastic or funny, it can be difficult to appreciate that their electronic audience may not.

The pointer is from Sendhil Mullainathan on Twitter.


I don't think anyone here appreciates the condescending tone with which you posted that.

Gota admit, you make me look.


Race baiter

First comment will not be topped for brevity or wit.

As there are at least two points in the communication attempt, there are also two points for failure. Not only might emailers be certain they are getting across their intended messages, emailees likewise may be (erroneously) certain they fully understand the intended messages.

And with the exception of intonation, this is also true for voice only communication. And if either the caller or callee is in a loud or very stimulating environment, intonation cues may also be lost.

The paper is from 2005.

In Internet years that paper is really old.... Reminds me of a recent comment here about "middle aged" and "deep bucolic nostalgia".

That's a brilliant point. People can totally tell when you're being sarcastic today.

You know who invented sarcasm? *I* did.

How did that joke come through?

Al Gore invented the internet because he was tired of people like you.

Joe made me laugh. Over time I predict that youth will increasingly find ways to bridge this gap. "Emoticons", bold, underlined, and abbreviations (LOL) will develop a tone all in their own. The next generation will understand intended gestures more easily as they pioneer virtual lives. But that only accounts for the emotion involved with communication. When speaking of effective email communication in a business sense, emotion is not as important as urgency, and expectations. This will eventually force people to be increasingly particular and articulate. Do you think tone will ever be clarified using electronic communication?

"This will eventually force people to be increasingly particular and articulate."

In another words, people will have to write emails in a format closer to the recent historical standard for letters. Somebody better inform the US school system, then. They've been phasing out articulate writing for over 30 years at this point.

Grumpy reaction: We're all just talking to ourselves, why put our thoughts in words or interact with people.
Pragmatic reaction: Real communication takes real effort..way more than pushing some words out. (Hand-waving and tone doesn't add that much either...context from prior conversations and actions is the biggie.)

Bad news for proponents of online education

If it's a course taught by Brad DeLong, you wouldn't know the difference.

In related news, people with crippling social problems, e.g., high-functioning autism or disfiguring injuries, can conduct meaningful social lives via the internet.

When I first started using email at a large high-tech firm in 1985 the colleague who was informally training me mentioned the issue with tone-deafness and resulting tendency for what was then called flaming in response. She said it erupted almost immediately after the world's first email system was created. It was bad enough that one of the three big computer companies at the time (I'm thinking IBM though it might have been AT&T) commissioned a study at Carnage Mellon to figure out WTF.

They tested with email vs. hand-written notes vs. phone conversations and, yeah, determined basically that people consistently gave ambiguous statements in text the most negative interpretation possible.

My colleague said that emoticons and tags like and LOL were proposed specifically to address the misinterpretation issue. (In other words they were basically engineered rather than emergent.)

The paper cited here might not be derivative. At Carnage Mellon they carefully studied the effect but only speculated about the cause.


Anecdote: While I was still at the high-tech firm one of my subordinates came roaring into my office saying she wanted a (senior!) developer from another work group fired. The cause? Even though they were on very good terms in person, he'd concluded an email with the following text: "or I could just reformat your hard drive. ;*)"

My direct report was a whiz at workstation computing but absolutely brand new to email (as were most people in 1986.) I repeated the story about Carnage Mellon and pointed out that the "line noise" at the end of the email was actually a "smiley." Problem indicated, problem solved.


Emoticons and abbreviations such as LOL existed existed well before 1985. We were using them at UCB in 1982 when I was there, using "msgs" (the ancestor to netnews) and email. I used "talk" (a two-way talk program) frequently to talk to people at MIT, and emoticons and such were already widely used (and understood).

Well smiley faces obviously predate the computer age. Not to mention teenage girls who dot their i's with a heart.

Can girls still do the latter with computers?

Yes, yes they can.

BF Best friend
BFF Best friends forever
BFFL Best friends for life
BFFLNMW Best friends for life, no matter what
BFFN Best friend for now

And 4,000 messages per week seems a bit on the high side, but 5,000 to 10,000 per month was fairly common when my youngest daughter was a hs sophomore and junior a few years ago. As a college student, she and her buddies all now use WhatsApp, which allows pictures and video quite easily. (When you meet a male 45 to 60 who texts, it's usually because he has a daughter in high school or college. And yes, now I use WhatsApp. To send her pictures of her dog at home.)

At the rate she's going, my daughter may well be in the 4000/week range by the time she's a HS freshman; I've had people scoff at some of the numbers thrown around, until I pull out a bill and show them the stats for her phone. Thank God for unlimited messaging plans. It'll be interesting to see if Facetime keeps her numbers down.

I am dubious that people at MIT understood the emotions.

The neuroscientists would. They have electrodes.

They'd have to explain it to everyone else, though (but again, electrodes).

What were the results from a handwritten note?

Kissing under a tree at recess is my guess.

I thought anonymous flaming was what the internet was originally designed for.

How else would we get such brutal honesty?

Obviously not from you, you weaselly... oh wait. That joke's been done better already upthread. Never mind. Carry on.

Web communication about text positioning, it is about having the text spotted by a bot that will deliver it to the intended audience. If you are good at it, you can get the web lag down to a month.
I do it, when I know some scientist or economist has good data, a model, and I want to flesh it out, make him/her put it up on the web, see what's under he covers.

Web bots are doing the reading, the commenter is doing the taunting, the intended audience is updating the information.

I forget. The blog posters often do a great job, they get e mails and counters from researchers to comments that show up in the searches, and that leads to subsequent article. It is a closed loop system optimized to flow information in units of posts!.

With e-mail or blog comments, the main problem isn't the lack of non-verbal cues. It's the fact that people very often don't read what you write.

I don't mean that they speed-read, or fail to take the time to contemplate the nuances and implications. I mean they literally don't read it: they completely skip over entire sentences and paragraphs, miss the point entirely, and then take several minutes to compose a reply that is, at best, a reply to something other than what you actually wrote.

These are shortcut habits born of the sheer verbiage most of us need to plow through every single day. But in the end, it simply wastes the time you spent writing and the time they spent replying.

Didn't read the rest of your post. Don't read Anonymous's post. He is generally cowardly and troll-like.

If that's happening to you a lot; maybe it is hinting at something?

I guess in a way this proves the original point of this blog post, about paralinguistic cues. Lacking those, you might interpret my comment as some kind of peevish rant against all the inconsiderate dolts who fail to appreciate my own carefully-crafted prose masterpieces.

But seriously, I'm just pointing out what seems to be a real and even rather common phenomenon. Has no one else noticed this? I'm not talking about subtleties or ambiguous wording either, I mean things that can only be explained by someone not reading an entire sentence at all, as if it had been blanked out entirely. On the order of: "The sky is blue." "I disagree, the sky is blue, not green, and here's four paragraphs about why".

In conversation this doesn't happen because the feedback loops are shorter. You can interrupt people as soon as it's apparent they missed a key point, and in any case people don't talk in paragraphs-long soliloquies.

I do see your point but I'd put some more blame on the senders. Some people have never heard the word "brevity".

Some online media can be amazingly precise though: the unix / linux forums on uSENET for example are gems of precision and brevity.

I am not referring to people ignoring what you write because they think it's stupid, and not bothering to reply.

I mean: people who ignore what you write, and then reply in detail anyway.

Sometimes it's a paragraphs-long reply that clearly took them several minutes at least, which would have been obviated entirely if they had spent ten seconds longer reading.

If it's a blog comment, that's sometimes OK, because the reply could still be an interesting riff in its own right for a wider audience. The wonder is when it's an e-mail reply and there is no other audience but yourself.

I know what you mean. I have a co-worker who conducts long email exchanges with friends, arguing about sports and politics. The other day he admitted he rarely actually reads their emails.

Won't this only be true for a generation or so? Eventually the majority of communication will be electronic, not face-to-face, and the generations after us will have realistic expectations of what that means. Or doesn't. They won't be projecting their context of direct communication and its meaning onto indirect communication.

I doubt this will happen.

I say we bring back the percontation point؟ (

Let me see how forceful I can make this statement without para linguistic ques:


The emotion is strong in this one.

If it was in bold caps, it would work even better.


What is more *effective* communication? Sometimes you don't want to convey your emotion--say, if angry, nervous, bored, etc.

That said, movies are better than books because they better convey the emotion in the dialogue.

No, the lesson here is that 95% of human communication is pretty much garbage.

If you need to say when you are will be home tonight, or what the Bruins' score was, or whether the cat needs a shot, or what time Jeopardy is on, or what the corporate tax rate is in Nebraska.... yeah, the internet or texting is just fine.

But if you need to opine on something -- which is really about 95% of human communication -- then you are really just satisfying yourself. This comes across in "old school" definitions of interaction by the simple fact that you hang out with people who largely agree with you. But today, you just do it on the net with a bunch of morons who cannot understand your genius. This leads to sarcasm, and strawmen, and all the rest.

Please reply below about how well you understand my genius. Please! PLEASE!!!!

Online writing has improved significantly since that paper was written. We err. We fail. We get slapped around. And we learn.

Or at least, most people do. :)

The three biggest problems with online discourse are:
a) that it's very difficult to negotiate a contract on the meaning of terms, and as such, most debates are eristic or autistic.
b) that the medium does not tolerate the level of exposition needed to convey vast differences in the categories and judgements that are under discussion.
c) almost no one, even the very best people, are able to articulate their positions by other than allegorical means, or without relying on the assumption that the methodology underlying their reasoning, is merely a convenience, not a representation or means of identifying true statements. (My glossary is fifty pages long. and it's not anywhere near complete.)

Conservatives are the worst offenders because they rely on sentimental, historical and allegorical concepts, which if fully articulated as human actions, are demonstrably true. But since they're so poorly expressed, usually as post-religous moral statements, they are all but useless in debate.

FWIW: I am absolutely nothing like my online persona, and everyone who meets me in person comments on it. Interpersonal relations are, well, personal. Debate online is political - purposeful. If I learned anything from the 20th century its that Friedman's and Rothbard's antagonistic relentlessness was more successful than Hayek's modest civility.

Tying together today's topics, maybe this is yet another explanation for the Fermi paradox.

The aliens aren't talking to us because they're mad at us. Was it something we said?

They're right. Sarcasm is NEVER understood in electronic communications.

Somewhere in all this, we need to formulate a hypothesis for why teens in particular prefer to use their phones for texting rather than speech. Not only do they rarely make phone calls, sometimes they text each other even when both are sitting in the same room.

Maybe they want wiggle room for plausible deniability after the fact... for instance, if you flirt in an ambiguous way, you save face if you give the other person room to construe it as merely a playful joke and reply accordingly. That's harder to do when your face or voice betrays you with paralinguistic clues galore.

Sometimes failure to communicate is a feature, not a bug.

I've seen them text each other in the adjacent chair.

Simpler to say that people are thinking conversationally when they dash off an email, but are generating a permanent or potentially permanent written document. Many misunderstandings and hurt feelings can arise from failure to appreciate this fact. I like the high culture language though ("paralinguistinc cues"!)

OTOH, you can't edit in real time. That's why some of us prefer email.

You betcha! If I say something and five miliseconds later I wish I hadn't, I can't take it back. (Politicians still seem to struggle with that concept.) OTOH, if I write something I merely have to discipline myself to look at it again before I hit Send. And if I take the trouble to edit what I've written I have a better chance to use just the right words, instead of simply the first ones that came to mind.

Isn't anybody in this entire thread old enough to remember when there were NO computers, NO word processing, and ONLY the handwritten letter through the U.S. Mail to send your grandma a thank you note for your birthday present? I seem to remember that back then it was worth the trouble to think first, then write a draft, then edit it, then copy it neatly onto nice stationary. Now I can do all those steps up to the last, substituting the Internet for the U.S. mail, and the thank you note arrives in a few seconds instead of a few days. But it's still right to do all those other steps.

I am from these kind of people who prefer to write rather than talk. For me communicating via e-mail is more effective than talking.

I am genuinely astonished that one can write and publish a paper on a fact that diplomats have known for centuries.

Comments for this post are closed