January 12, 2012
"If it bleeds, it leads," goes the
cynical saying with television and newspaper editors. In other
words, most news is bad news and the worst news gets the big story
on the front page.
So one might expect the New York Times to contain, on average, more
negative and unhappy types of words - like,
...than positive, happy ones - like,
Or take Twitter. A popular image of what
people tweet about may contain a lot of complaints about bad days,
worse coffee, busted relationships and lousy sitcoms.
Again, it might be reasonable to guess
that a giant bag containing all the words from the world's tweets -
on average - would be more negative and unhappy than positive and
But new research shows just the opposite.
"English, it turns out, is strongly
biased toward being positive," said Peter Dodds, an applied
mathematician at the University of Vermont.
The UVM team's study "Positivity
of the English Language," is presented in the Jan. 11
issue of the journal PLoS ONE.
This new study complements another study the same Vermont scientists
presented in the Dec. 7 issue of PLoS ONE, "Temporal
Patterns of Happiness and Information in a Global Social Network."
That work attracted wide media attention showing that average global
happiness, based on Twitter data, has been dropping for the past two
Combined, the two studies show that short-term average happiness has
dropped - against the backdrop of the long-term fundamental
positivity of the English language.
In the new study, Dodds and his colleagues gathered billions of
words from four sources:
twenty years of the New York
the Google Books Project (with
millions of titles going back to 1520)
a half-century of music lyrics
"The big surprise is that in each of
these four sources it's the same," says Dodds. "We looked at the
top 5,000 words in each, in terms of frequency, and in all of
those words you see a preponderance of happier words."
Or, as they write in their study, "a
positivity bias is universal," both for very common words and less
common ones and across sources as diverse as tweets, lyrics and
Why is this?
"It's not to say that everything is
fine and happy," Dodds says. "It's just that language is
In contrast to traditional economic
theory, which suggests people are inherently and rationally selfish,
a wave of new social science and neuroscience data shows something
quite different: that we are a pro-social storytelling species.
As language emerged and evolved over the
last million years, positive words, it seems, have been more widely
and deeply engrained into our communications than negative ones.
"If you want to remain in a social
contract with other people, you can't be a…," well, Dodds here
used a word that is rather too negative to be fit to print -
which makes the point.
This new work adds depth to the Twitter study that the Vermont
scientists published in December that
attracted attention (above video)
from NPR, Time magazine and other media outlets.
"After that mild downer story, we
can say, 'But wait - there's still happiness in the bank," Dodds
notes. "On average, there's always a net happiness to language."
Both studies drew on a service from
Amazon called Mechanical Turk.
On this website, the UVM researchers
paid a group of volunteers to rate, from one to nine, their sense of
the "happiness" - the emotional temperature - of the 10,222 most
common words gathered from the four sources.
Averaging their scores, the volunteers
rated, for example,
"laughter" at 8.50
The Vermont team - including Dodds,
Isabel Kloumann, Chris Danforth, Kameron Harris, and Catherine Bliss
- then took these scores and applied them to the huge pools of words
Unlike some other studies - with smaller
samples or that elicited strong emotional words from volunteers -
the new UVM study, based solely on frequency of use, found that,
"positive words strongly outnumber negative words overall."
This seems to lend support to the so-called
Pollyanna Principle, put
forth in 1969, that argues for a universal human tendency to use
positive words more often, easily and in more ways than negative
Of course, most people would rank some words, like "the," with the
same score: a neutral 5. Other words, like "pregnancy," have a wide
spread, with some people ranking it high and others low.
At the top of this list of words that
elicited strongly divergent feelings:
"profanities, alcohol and tobacco,
religion, both capitalism and socialism, sex, marriage, fast
foods, climate, and cultural phenomena such as the Beatles, the
iPhone, and zombies," the researchers write.
"A lot of these words - the neutral words or ones that have big
standard deviations - gets washed out when we use them as a
measure," Dodds notes.
Instead, the trends he and his team have
observed are driven by the bulk of English words tending to be
If we think of words as atoms and sentences as molecules that
combine to form a whole text,
"we're looking at atoms," says Dodds.
"A lot of news is bad," he says, and
short-term happiness may rise and and fall like the cycles of
the economy, "but the atoms of the story - of language - are,
overall, on the positive side."