HARNESSING
THE WISDOM OF CROWDS:
The
New Contours of Intellectual Authority
Remarks to The Fields
Institute for Research in Mathematical Sciences
Annual General Meeting
by
Dr. Peter J. Nicholson
President, Council of Canadian Academies
613-567-5000; peter.nicholson@scienceadvice.ca
|
OUTLINE ·
The Decline of Deference ·
Role of the Media ·
The Dilemma of the Information Age ·
Massively Distributed
Collaboration ·
Harnessing the Wisdom of Crowds ·
Implications for Mathematics |
HARNESSING
THE WISDOM OF CROWDS:
The
New Contours of Intellectual Authority
I will argue that what
qualifies as intellectual authority
in contemporary societies – who and what to believe – is changing fundamentally. I will speculate as to the reasons, and draw
out some of the implications for intellectual work in the future. I will even venture to speculate on possible
implications for the conduct of mathematics research, though the main object of
my message this evening concerns the broad sweep of intellectual endeavour and
not any specific domain.
The thesis in a
nutshell is this. People today are much
less prepared to defer to the experts.
But at the same time, we are being swamped with data and information – a
glut that cries out for analysis and summary.
So there’s a dilemma. Who to turn
to? Increasingly the answer is – Well,
to ourselves of course, as individuals empowered by a world wide web that has
rapidly evolved into a social medium. More specifically, it is a medium that today supports
massively distributed collaboration on
a global scale that – we can only hope – will help us make sense of it all.
This phenomenon of
massively distributed collaboration is perhaps best exemplified by Wikipedia –
the wildly popular, on-line, user-created encyclopedia. I will have a lot more to say about Wikipedia
later in these remarks. But the underlying
notion has more prosaic origins, many of which are described in James
Surowiecki’s fascinating book – The
Wisdom of Crowds, [1] the title of which is an ironic inversion of Charles
MacKay’s 1841 classic, Extraordinary
Popular Delusions and the Madness of Crowds.
The wisdom of crowds, as Surowiecki describes,
was encountered – perhaps to his dismay – by the great statistician and elitist,
Francis Galton, during a visit to a country fair in
So here at least we
have a counterexample to the elite convictions of such great intellects of that
time as Freidrich Nietzsche who wrote that “madness is the exception in
individuals, but the rule in groups.”
Or, as Thomas Carlyle haughtily put it: “I do not believe in the
collective wisdom of individual ignorance.”
A lot has changed in
the hundred and twenty-five years since Carlyle’s death. My purpose in these remarks is to offer a perspective
on what seem to me to be the deepest and most pervasive changes that are now shaping
the contours of virtually all forms of intellectual authority, transforming its
landscape in ways that stand Carlyle’s elitist conviction on its head.
Let me say at the
outset that I am not particularly comfortable with the future I foresee. I am, after all, a charter member of the ‘old
guard’ and will never really belong to the new.
But I am also an optimist and a realist.
The world has changed – and so must we.
The Decline of Deference
Let me begin with some very general remarks on contemporary
attitudes toward hierarchical authority generally – of which intellectual
authority is but one instance.
As President of the new
Council of Canadian Academies – an organization that will oversee expert
studies of the science underlying important public questions – I am in the
business of brokering intellectual authority.
I admit to being a traditionalist in the sense that I believe
intellectual authority should have a close correlation with expertise. And it should flow from the tried and true,
though never infallible, processes of peer review and other forms of elite
consensus building.
More than that, I am comfortable with
hierarchies that are based on merit. And
I am quite willing to defer to the well-established institutions in today’s
society since, on balance, I believe that their power is adequately constrained
by the legal, economic and political structures of modern democracy.
But I am also convinced
that the values that have shaped my world view – and that of my demographic
peers throughout the industrialized world – are being eclipsed by a new
paradigm. This new framework is shaped
by technology – primarily information and communications technology; by
globalization; by post-industrial affluence; and by a culture which, as never
before, celebrates and empowers the individual.
One of the most
significant symptoms of this pervasive shift is the decline of deference to
virtually all forms of traditional authority – whether the church, the school
teacher, the family doctor, the business executive, the union leader, the
politician, and not least, the intellectual.
In short – out there on main street, mistrust and scepticism reign.
While all this is
widely recognized, the truly fundamental reasons for the decline of deference seem
not to be generally understood in broad sociological terms. The explanations we do see typically cite the
public revulsion that stems from specific
cases – for example, scandals in the Catholic Church; or in businesses like Enron;
or in politics – Watergate; our own “sponsorship affair”; the failure to find
WMDs in Iraq; or to warn the British public of BSE. Take your pick.
The key point is this. The decline in trust of – and therefore deference
to – traditional sources of authority is a nearly universal feature of advanced
societies. It transcends every specific,
local instance. And it didn’t just
happen yesterday. Deference to
hierarchical authority has been declining for at least the past 50 years. Clearly, therefore, we are witnessing a socio-cultural
change whose roots run deep in the character of economically advanced
societies.
From whence does it spring? The best account I have read is by U of T
political scientist, Neil Nevitte. His
1996 masterpiece, The Decline of Deference,
draws on a rich vein of multi-country time-series data – the World Values
Survey – to establish convincingly that “the new citizens are less likely than
their predecessors to be satisfied with any form of authoritarianism. . .
Citizens, cut from the newer cloth, are more attracted to formations that are
bottom-up.” [2]
Thus societies formerly based on deference to
authority, community loyalty, and the struggle for the material basics of life
have given way to societies, the affluence of which has engendered a generational
shift toward the so-called “post-materialist” values of self-esteem, quality of
life, and the search for personal fulfillment.
When these objectives are
combined with the empowering tools of universal education, a rights-oriented
political culture, and the Google search engine, we should not be surprised
that people – and particularly younger people – regard ex cathedra expert authority with scepticism, if not outright
hostility.
The paradox is that
expert opinion is being sought and cited more than ever. But increasingly, it is individuals
themselves who weigh the various authorities and come to their own conclusion. Just ask doctors about their web-savvy
patients.
Role of the Media
Let me open an important parenthesis here on
the role played by the media in shaping broader public attitudes toward
intellectual authority. The prevailing
ethic in journalism is that “fairness” requires that all views on an issue be
presented, often without regard for the relative weight of authority of various
sources being quoted. The objective is
simply to report point, and counterpoint, with an emphasis increasingly on
sensationalism, official screw-ups, and conflict – i.e. those things that can
attract at least fleeting attention, and advertising dollars, in an information
environment that has become super-saturated.
The net effect is to create in the public mind
an impression that experts can never agree; and expert authority is thereby
diluted. Fortunately, I don’t see this as
representing much of a threat to the authority of expert mathematicians – but
then again, math controversies don’t get a lot of front page ink anyway!
The same certainly cannot be said for medical
journalism where the daily reported
advice does make the front page. And the
advice in the mass media on how to stay healthy keeps flip-flopping, whereas
the full text of the journal articles would reveal the provisional nature of
findings, statistical caveats, and so forth.
The bottom line is that superficial media treatment of scientific and
technical issues reinforces the prevailing scepticism as to the consistency and
trustworthiness of expert authority.
The Dilemma of the Information Age
Coming back to the main line of my argument, we
find that while expert-based authority is being challenged, the volume of
information and the economic significance of knowledge are exploding. Information technology itself – whose capacity
continues its four-decade exponential improvement – is clearly a key part of the
reason. But so too is the huge global
expansion of knowledge-generating capacity, the more so as
So we are confronted
with a dilemma.
On the one hand, the whole world is struggling to
cope with an information explosion that shows no sign of letting up – quite the
contrary. We need somehow to transform a
data torrent into useful information and knowledge that can power economic
progress and human fulfillment.
But on the other hand,
the agents we have relied upon traditionally to filter and manage information,
and to broker formal knowledge – agents like research universities, the serious
media, and highly trained experts of all kinds – are less trusted as
intermediaries than they once were. And
even if that were not the case, we might doubt that these expert resources are
really up to the task of managing the information glut anyway. Just ask journal editors and referees, or
researchers in any dynamic field, how well they are keeping up. Ask yourselves.
Part of the response,
of course, has been to deploy the same computer technology that is facilitating
the information explosion in the first place, to help cope with its
management. In other words, the offence
is also the defence. That’s why Google Inc. today has a total stock market
value of more than US $115 billion – over four times the combined worth of Ford
and GM. And it’s also why “Google” has
become a verb. (Who remembers when it
was merely ‘one’ followed by a hundred zeros!)
But Google and its ilk notwithstanding, the
sheer volume of information, its global origins, and especially the dynamic,
real-time nature of information today is simply overwhelming our traditional, centralized
institutions of information screening and management – whether research
libraries, book and journal publishers, or newspapers and other mass
media.
The infosphere,
if I could use that term, therefore needs new and decentralized mechanisms of
self-regulation and self-organization, much like a complex economy which, as
Adam Smith realized, needs the guidance of an invisible hand.
Massively Distributed Collaboration
I believe that the outlines of just such a
mechanism are already emerging in the multifaceted development of what
cyber-prophet, Mitch Kapor, recently dubbed “massively distributed collaboration.”
[3] Probably the single best example, as
I mentioned at the outset, is Wikipedia,
the user-edited encyclopaedia that in just over five years has become one of
the most-visited sites on the web.
In fact, something much
broader is going on. The world wide web
has already morphed into a social
medium – what some are calling Web 2.0 – a global many-to-many meeting place,
very unlike the one-to-many connections of radio, TV, books and
newspapers. The latter media are
inherently hierarchical – a communicator of one to an audience of many. The social web, on the other hand – like
Thomas Friedman’s new world – is flat. It is in tune with today’s ethos. Just consider some of the manifestations:
·
20
million blogs – and counting;
·
Self-expression
portals like My Space and Facebook, growing explosively – indeed a
new cultural phenomenon, tellingly dubbed “Me Media;”
·
Massive
multiplayer games like Ever Quest and
Second Life where the players
themselves shape the dynamic environment;
·
The
Linux operating system, flagship of the open source software movement, and
maintained by a worldwide network of volunteers;
·
eBay
– the many-to-many model implemented as a phenomenally successful digital
marketplace;
·
Amazon,
and countless other “collaborative filtering” sites that tally and report user
satisfaction;
·
And
Google itself, which indirectly exploits massively distributed collaboration
via its page-rank technology to aggregate the behaviour of millions of users
into an index of relevance.
In summary – and this is my key message – we
are witnessing in these examples the convergence
and mutual reinforcement of two
of the great defining movements of the past half-century – one cultural, the
other technological – i.e. the ascendancy of the ordinary individual together
with the empowering technology of the computer, now enormously amplified by
global networking – creating essentially a “cyber nervous system” for the
entire planet.
This is an epochal development that will not be
reversed. The job for all of us beyond a
certain age, but still hankering to be part of the action, is to figure out how
to be a constructive part of it.
Harnessing the Wisdom of Crowds
In the remainder of these remarks I want to
take a closer look at one important example of massively distributed
collaboration (MDC) – specifically, the on-line encyclopaedia movement, since
this illustrates most directly how MDC is already harnessing the wisdom of
crowds and thereby reshaping the contours of intellectual authority.
The flagship example is Wikipedia, founded only
in January 2001, but already the site of nearly four million entries in almost 200 languages. There are more than 1.1 million articles in
English, growing by about 1,500 a day. [4] The Encyclopaedia Britannica, by
contrast, has merely 65,000 articles in the print edition and 75,000 on-line. [5]
What is most amazing is that Wikipedia is doing
all this on an annual budget of just $1.3 million, 60 per cent of which goes
for the cost of computer hardware, leaving only about $500,000 to cover
everything else! [6] How can that possibly be?
With apologies to those who are already very familiar, here’s what’s
going on.
For starters, the articles are written, and re-written,
by volunteers. The website is equipped with so-called “wiki”
software that allows anyone with a browser to edit virtually any article at the
push of a button. In the flat culture of
Wikipedians, experts and dunderheads are equally welcome. The main editorial principle is that articles
should reflect a neutral point of view. This is not a site for cranks and
propagandists. Acts of deliberate
vandalism are not tolerated and are usually corrected very quickly. On the other hand, decisions as to what is
deemed to be unjustified bias are taken consensually, and this can be
excruciatingly drawn-out in contentious areas.
At first blush, it admittedly sounds a lot like
“monkeys with typewriters.” But in fact
it’s not. In a widely publicized and
controversial head-to-head test with Britannica, reported last December in the
journal, Nature, expert reviewers
determined that Wikipedia articles, on average, contained “only” a third more
inaccuracies than their Britannica counterparts. More to the point, only eight
serious errors were reported in the sample of 42 topics with an equal number, four,
attributed to each source. [7]
Having read the full text of the debate between
Nature and Britannica over the methodology
of the comparison, I would grant many of Britannica’s objections, but would
still conclude that the essence of Nature’s
findings remains intact. Wikipedia is
surprisingly good, especially for a five-year old; and even a source as
well-researched as Britannica still contains a significant number of
inaccuracies. So much for any
presumption of expert infallibility.
The real bottom line, of course, is that notwithstanding
doubts about its reliability, Wikipedia has taken off like a rocket. We need to understand why.
Obviously, being instantly accessible and free – no ads at all – is a big
plus. But the real power of Wikipedia is
that it’s in perfect synch with web culture – which mirrors today’s attitudes,
and even more so tomorrow’s. Wikipedia
is also in synch with globalization – 200 languages represented with much of
the content original to each language, not simply translated. And Wikipedia – like Google, and blogs, and
open source software – operates in synch with the rhythm of the web, incorporating new information continuously in
real time, 24/7.
This last point is important, and is part of a
much larger story. I can only summarize. The “half-life” of active information has been getting shorter and shorter due
primarily to the sheer rate of information generation. There is more and more to process, but not
more hours in the day, and not more raw individual brain power to apply. So we graze, or we gulp, and then we move
on.
The half-life is also shrinking due to the very
nature of electronic technology which makes overwrite so easy and natural. We are all becoming addicted to the “refresh”
button. Documents of every kind –
certainly in my experience in business and government – are being revised
continuously until the moment they become virtually obsolete. And as the shelf-life of any particular
information product gets shorter – whether it’s an e-mail or a position paper –
basic principles of economics dictate that fewer resources of time and money
can be put into its creation. The
ubiquitous deck of bullet points is the iconic example.
The result is a dumbing down of written
communication. We can decry it – and I
do – but it reflects a probably necessary trade-off in favour of easier and
quicker absorption, unfortunately at the expense of nuance and rigor.
This has profound implications for how good is “good
enough” when it comes to authoritative information. Of course, complete accuracy still matters as
much as ever where lives or fortunes (or mathematical theorems) depend on it. But for most everything else, the tradeoff
point is moving toward faster, not deeper.
This is a context in which massively
distributed collaboration systems like Wikipedia excel. But the advocates of MDC claim much more, and
believe that it can be both faster and deeper. They may have a point based on the old adage
that two heads are better than one – and thousands or millions of heads are
incomparably better. And so we come full
circle to the wisdom of crowds and to the validating belief of the open source
software movement, summed up in the motto – Given
enough eyeballs, all bugs are shallow.
[8]
Maybe. But in the case of specialized subjects where
quality criteria are more judgemental (unlike software bugs), or where relevant
expertise is spread very thinly – and perhaps nowhere so thinly as at the
frontier of mathematics research – the “crowd” is unlikely to be sufficiently
wise. So there will always be a secure
niche for expertise in the traditional sense.
Indeed, that conviction led Wikipedia’s co-founder, Larry Sanger, to
leave what he had created out of despair over the hostility toward expert
authority that dominates Wikipedian corporate culture. Sanger is now creating a new on-line
authority, Digital Universe, that seeks to provide both expertly-created as
well as collaboratively-developed content. [9]
We should stay tuned,
because the puzzle that the Larry Sangers of this world are trying to solve is
to evaluate and integrate very different methods of ascertaining intellectual
authority – ranging from the continuously-flowing, collaboratively-determined
“truth” of Wikipedia and its ilk, to the timeless records of solitary
genius.
Indeed we should be
thinking of the infosphere as an ecosystem where different “species” are
adapted to specific niches. Google, for
example, delivers fantastic volume but the measure of relevance is still pretty
crude. Blogs give you an
up-to-the-minute read on what’s hot. Wikipedia
provides a great first cut at coherently organized material plus a good set of
relevant links. And if reliability is a critical
objective, then refereed journals and original documents become progressively
more important. But the contemporary
niches in the information ecosystem are neither stable nor secure. Instead they are shifting continuously in
response to technological and cultural changes, which, as I have argued this
evening, are reshaping fundamentally the contours of intellectual authority.
Implications for Mathematics
What are we to make of
all this? As always, the one key
question for each of us is – What does it mean for me? For most of the audience in this room that
question translates to – “Does any of this relate to the world of mathematics
research?”
One might reasonably
think not, arguing that mathematical creativity of the highest rank has always
been a solitary and esoteric activity, the talent for which is extremely rare,
much like artistic and literary genius. Proving
theorems about non-Abelian groups is not a bit like averaging guesses about the
weight of an ox.
True – but on the other
hand, writing the software to implement the Linux operating system is – in terms
of sheer intellectual complexity – comparable to resolving some of the
knottiest problems in pure mathematics.
Yet the Linux challenge – and a growing number like it – are being met
through a massively decentralized collaborative effort by thousands of
volunteer hackers motivated both by the sheer challenge, and by the psychic and
social rewards of belonging to a very
special club of peers.
Might not similar conditions
apply to important areas of mathematics research?
When I googled
“collaborative mathematics,” one of the first entries was something called the Flyspeck Project which is dedicated to
the collaborative development of a purely formal proof of the ancient Kepler
Conjecture on the maximum packing density of spheres. This famously difficult problem was
apparently solved by traditional methods by Hales and
The Flyspeck Project
hopes to use specialized computer software and an army of volunteers to develop
a “formal” proof of Kepler, the correctness of which will be virtually assured
by the method of its construction. The Flyspeck
web site includes the following line:
“We are looking for mathematicians,
from the advanced undergraduate level up, who are computer literate and who are
interested in transforming the way that mathematics is done.”
An isolated case, you
say. Perhaps. And certainly the traditional skills of mathematicians
will continue to be relevant in collaborative projects just as world-class software
skills are table stakes for the open source hackers who keep improving the
Linux operating system.
But I would wager that
mathematics will not remain isolated from the deep and pervasive changes we
have been discussing this evening. And
fundamentally, that’s simply because the coming generations of mathematicians
will be children of the web culture, globally-networked, equipped with
unimaginable information-processing power, and devoid of deference to
hierarchical authority, intellectual or otherwise.
Can we believe that
they won’t do things very differently?
Source Notes
1.
James
Surowiecki; The Wisdom of Crowds; 2004.
2.
Neil
Nevitte; The Decline of Deference;
1996.
3.
Presentation
by Mitch Kapor at UC Berkely;
4.
Internet Encyclopedias go head to
head: Nature 438;
5.
Who Knows? The Guardian;
6.
Cited
in article on Wikipedia in
Wikipedia. It is reported that 4th
Quarter, 2005 costs were $321,000 ($1.3 million at annual rate) with hardware
making up almost 60%.
7.
Nature
438; op.cit.
8.
Eric
Raymond; The Cathedral and the Bazaar;
First Monday;
9.
Larry
Sanger; Why Wikipedia Must Jettison its
Anti-elitism; Kuro5hin.org;
10.
http://www.math.pitt.edu/%Ethales/flyspeck/