Socially Intelligent
Computing
·
By: Daniel Goleman; Clay Shirky
Goleman: The way I understand that
is that when we're talking to someone face to face
when we're with someone in real time, and in real life, the social brain is in
its natural ecological niche. It's picking up information that it wants, moment
to moment. It's reading porosity invoice,
it's reading emotions, it's reading non-verbal cues. And it's doing it invisibly
and constantly out of awareness, and then telling us what to do next to keep
things smooth, to keep things in sync, to keep things on track, so that we can
get to where we want to go. The problem from my point of view with the design
and maybe the hardware itself of the web
is that is it has no channel for the social brain to attend to. You have no
emotional signal in real time. You have nothing for the orbital frontal cortex,
which is dying to get this information to latch onto, and one of the main
functions of that part of the social brain is to inhibit impulse. To say, 'no,
don't do that, do this. And it has nothing to go on so we're flying blind and
as far as I can see, the web is actually designed to optimize flaming.
Shirky: That's exactly right. In
fact I've often joked that if you looked at the software for supporting, say, a
mailing list, you would never find anywhere in the code something that said
'this next bit of code ensures that people will get into huge vituperative
arguments that last weeks and weeks and involve lots of name calling'. And yet
when the software is deployed in the field, reliably that's the effect. And so
plainly there is something social going on. Flaming, interestingly, is, in fact, one of the most social effects and
one of the earliest ones to be commented on, because it's much more severe in
groups than it is in two-person exchanges. Famously, one of one of the
antidotes to flaming is to contact the flamer personally by email and very
often people find that you have a much more reasonable conversation. When you
take them out of the social part of the conversation where they're performing
in front of an audience, and address them as an individual, they become much
less prone to the kind of name-calling and vituperation. But in the group
setting, designing software that actually inhibits flaming has proved to be
fantastically difficult. And as a design challenge, something I keep looking
for is what principles would you try to engage if you wanted to create an
environment that was more convivial to group agreement and conversation, rather
than just altercation and disconnect.
Goleman: Well I remember a
conversation years ago with Mitch Capo about this very problem, and what he was
suggesting was re-engineering the web itself so that there was real time,
basically so that it's a conference call with the picture. If you have a video conference, you at least know what the
person who's not saying anything is feeling about what you're saying. You can
see how the group moves, by posture even, whether there's consensus or dissent,
who wants to speak next or not. You don't have it on the web. Frankly, I don't see that happening anytime
soon, so we're left with the dilemma of how to reverse-engineer what we have
now to make it work better. Do you have any ideas?
Shirky: I think
that sense of reverse engineering is absolutely right. The Internet has had
truly major upgrades only two or three times in its entire history. It's a very
difficult proposition to change the basic structure. But it's had lots and lots
of changes, which involve what David Weinberger calls 'small pieces loosely
joined': some additional piece of software. And I think that essentially
opening up the channel for communal awareness, somehow adding that channel, is
a big is a big design challenge, and I think a huge opportunity. You alluded
earlier to our inability to get the sense of the room, the read of a room. If
you're in a meeting and someone is saying something you can immediately tell
whether or not there's basic agreement, basic disagreement, whether people are
rapt or bored. On the net, for instance if you're on a mailing list, not only
is that not possible but if someone posts something that captures the mood of
the crowd and someone else follows up that message by saying, 'oh yes, me too'
or 'I agree', that's actually considered rude. So it's actually both social and
technological pressures that are suppressing a lot of that kind of emotional
awareness. One of the things that instant messaging has shown us is that awareness of someone else in real time, if I'm
instant messaging with you, I have a different sense than if we're emailing,
simply because of the acceleration of time. There's been a lot of looking at
whether or not presence can be made more
general. And presence can start to convey things more than just 'I am here and
breathing and attending to the computer' but 'I am here and I agree'. 'I am
here and I disagree'. So in a number of pieces of relatively novel chat
software, there's one wonderful piece of chat software called a really simple
chat. It's just an open source tool but it works in pretty much any web
environment. In addition to type what you would like to say here, there's also red, yellow, and green buttons. If
you click a red button after someone says something, it just shows up with your
name meaning I disagree. Its not 'I'm going to construct a whole written
disagreement' it's just 'I disagree'. The green button is 'yes, absolutely I
agree with that'. And there are ways that people can convey those emotional
senses that make those chats in that environment emotionally rich, or socially
richer, than in the more tenuated text only environments.
Goleman: Do you think that kind of
patch, so to speak, might help move away from what the current myth is - that
social software as it exists now actually supports group decision-making,
actually supports business function, and actually supports meetings? People
tend to believe that what they get is what they need, but what you're saying is
that there are implicit omissions in the software that we use to, for example,
try to come to a collaborative decision. Or if you're the WHO or Accenture and
you have a global team trying to work on how to eradicate HIV in Africa or
you're trying to roll out a new product in Shanghai and Berlin and you've got
people who never see each other, but who operate together via the web, what
they don't realize is that the very mechanism they're using has blind spots.
And the kind of thing you're suggesting is a way to correct that.
Shirky: That's right, I often say
that social software is not better than face-to-face contact, it's only better
than nothing. There was a period that people wanted to believe that this as a
channel was essentially better than face-to-face contact, or somehow superior.
In a handful or circumstances its virtues
work well enough, for instance if you're trading source code. If you're trading
source code, you'd actually like a kind of asynchronous and logic-heavy
environment. But if you're really trying to get a group to come to a decision
that they don't just have to agree to rationally, but they also have to
emotionally buy into - a business that's starting, a risky new enterprise, or a
group of people that's going to set out in some direction as explorers, whether
physical or intellectual in any medium, what internet tools now will help you
do is come to some rational listing of pros and cons. But it won't help the
group attain that emotional core of agreement that keeps people together.
Goleman: Well there's a real flaw
there, and it has to do with how the brain makes decisions. Antonio Damasio,
who is an expert on what I'm calling the social brain and the emotional
centers, tells this story and this is a very important case in point of a
brilliant corporate lawyer who had a pre frontal brain tumor that was
discovered early and operated on successfully. During the operation they
disconnected the links between the brain's pre-frontal cortex, the brain's
executive center, and the emotional centers. And after surgery it was quite
intriguing, because this guy who had been brilliant was still brilliant. He had
no loss of memory, no deficit of attention, but he couldn't do his job anymore.
And Damasio was completely puzzled because on every neuropsychological test he
was perfectly fine. And then one day he gets a clue as to what was wrong with
this guy, he says 'when should we have our next appointment?' And he realizes
the guy can give him every rational pro and con for the next two weeks, but he
doesn't know which is best. Damasio argues that in order to make a good
decision individually, we need to attune to our emotional centers, because a
vast amount of information processing goes on out of awareness. And our entire
life wisdom on the current topic, the current decision, you know, 'should I
adopt this business strategy?' 'Should I leave my job?' Who should I marry?'
Whatever the decision, it isn't made purely rationally, it's made because the
emotional centers valence information for us, and give us a gut sense of what's
right and what's wrong. Lacking that, he says, we have no moorings, and what
you're describing in the collective situation is software that will help people
make the emotional decision without any emotional valencing. When you're
face-to-face and you're in a room, you can see people's expressions, you can
tell their tone of voice, you pick up a multitude of non-verbal signals that
tell you in every moment how we collectively feel about what's being said. So
here's the question, it seems that there is huge power in information
processing collectively to be had in web discussions, but if you're functioning
as a group, when do you need to get together? When do you need to have that
conference call? When do you need to be together for two days to get to know
each other as people and establish trust? Do we have any good rules of thumb?
Rules of Thumb for Online Collaboration
Shirky: That’s a really
interesting question, and in fact you've just covered one of the big 'aha'
moments for me in social intelligence. I've often told my students that emotion
is your best tool for traversing large decision spaces. They will get very
worked up about breaking down the variables exactly like the guy in the Damasio
story. And the fact that it ultimately has to come down to a very narrow set of
decisions means that they have to find some way to compress that huge number of
variables. And we greatly overlook emotion as a decision-making tool in favor
of doing things like decomposing, ranking, and adding, as has often been taught
in business schools. Just before I came to this meeting I was having a
conference with a former student who now taken his thesis research and is
launching a business out of it. And as happens the collaboration with the
department once it's gotten out into the real world is under stress because one
of his business partners is also working on a different idea at the same time.
And so he's absolutely torn about how to structure this because this person is
both a key collaborator in the business and also a long time friend. And he was
going back and forth and back and forth and clearly not coming to much of a
decision and I finally said look just take out a coin. Heads you're going to
try and work with the guy, tails your going to not try and work with him. Flip
the coin in the air, and the minute it's in the air, figure out which side you
hope it comes down on, and do that. Throw the coin away.
In terms of when you should meet face-to-face, there are two
different models I call them the umbrella model and the banyan model, although
since those are such different objects I need better words for it. But the
umbrella model is you periodically close together and then spread out. Close
together and then spread out. I've been the chair of two large-scale
distributed network design efforts over the last five years. One was for the
Library of Congress's Digital Preservation Network, which involved a consortium
of research libraries talking about how we're going to share this material. The
other, more recently, which I still have, is the chair of the technical
sub-committee of a clinical data exchange network, a medical network. And my
rule of thumb is your first face-to-face meeting should be as soon as you can
possibly make it. In practice we've ended up having that meeting be after an
introductory phone call which sets the themes, but the earlier you can have
face-to-face contact, the sooner you can take advantage of the kind of
remembered valence of personality and so forth. I think everyone has had the
experience of emailing back and forth with someone, whether it's a new friend
or someone you meet online, or someone in your business. Then meeting in
person. Afterwards, the email takes on a very different tone. And this is
really bad news for much of the business communications world, who have from
the mid 1960's with the AT&T videophone been pushing the idea that
communication will be a substitute for travel. In fact communication's a spur
to travel. When people meet and talk in mediated environments like the
Internet, eventually they want to meet and talk face to face.
Goleman: Let me give a perspective
from the social brain because what you're saying really is that social brains
need to attune before they can before they can use a shorthand. Because what
happens face-to-face is you come to know the person in a very deep way which
actually builds representation in your own brain of that person - of who they
are, of their style, of their feelings, and your feelings about them. Lacking
that, you don't know how to take things that are said. I got a very interesting
email from a women that works for my publisher and she and I had been in a
meeting once but not spoken and then we had an email exchange and at one point
she sent me the following email. She said: "It's difficult to have this
conversation by email. I sound strident and you sound exasperated." Now
what's intriguing to me is that I had no idea that I sounded exasperated, nor
that she was strident. But once she named it, I realized that there was something
really off. And the problem with email alone is that you can be off and not
know it, and a small seed of off grows into a big misunderstanding. I was just
in Europe and I was talking to someone who is consulting to two big European
telecoms who have a business alliance, and he said the whole thing was a great
business idea, great business plan, great business projections, but it is
stalled from the get go because two sets of engineers are flaming at each
other. And his solution was to get them together for two days and let them know
each other as people, and on top of that to work out norms for how they're
going to communicate online together going into the future.
Shirky: The engineering model is
another way to approach the problem, and that's the thing I'm calling the
banyan model, which is not the whole group gets together all at once
periodically. Very often in large distributed businesses you can't do that. But
you put down little roots of face-to-face contact everywhere. I had someone who
was the head of a global IT department for a bank. Talking about the design of
social software he said 'what can we do to replace face-to-face?' and I said,
'you can't. What you should be doing is using face-to-face to sort of
strategically augment electronic communications. The bank is global, the
security people in Singapore are going to stay in Singapore the security people
in London are going to stay in London, and you've got to have the distributed
24/7 communication of the internet as a piece of that, but you should
periodically eat the expense of flying those people someplace where they can
get together from time to time.
Goleman: Let me add to that
because it's not just that you want people to have a business meeting - you
want people to get to know each other.
Goleman: The fact that Wejung in
Shanghai knows Carlos in Barcelona has four daughters and his wife is an opera
singer. And Carlos knows that Wejung's wife teaches English in a college and he
has one son and that bit of knowing you as a person is background, because it
builds a foundation of trust and understanding that is a safe container for
when things might go awry otherwise.
Shirky: One of the design
principles that I think anybody managing groups has to deal with is to realize
that the group is a unit that's separate as a unit from the individual. And so
much focus has gone on individual employees and career paths and so forth that
regarding the group as anything other than an accidental conglomeration too
often gets cast by the wayside. One of the things you'd really like to have
happen, particularly in the business setting with distributed groups, is not
for everyone to know everyone on a large scale, that is simply impossible even
in companies that are physically in the same city. But when two groups interact
you'd like at least one person in that group to know one person in the other
group, because that prevents enemy culture from appearing. There's a wonderful
book by Wilfred Bion, the group psychologist of the middle 20th century, who
outlined the ways in which the group would silently collaborate with one
another and he outlined the things that produced group cohesion. One of the
things that he said which has always struck me as indicative for some kinds of
businesses is that nothing creates group cohesion faster than an external
enemy. And that in cases where a group doesn't really face an external enemy,
they will tend to rally around their most paranoid member because that person
is best at locating enemies where there are none. And so to defeat enemy
culture you really want some strand, that kind of deep knowing, to go across
whatever bit of geography or hierarchical separation the company has. And that
network works differently than, but strengthens the ability of, hierarchically
organized groups to also get things done.
Goleman: You
know, since Bion there's been an updating of our analysis of group dynamics and
it's a guy who started in the same psychoanalytic T-group tradition that Bion
was in, his name was Freed Bales, I had a course with him when I was a graduate
student at Harvard years ago. The course was fascinating, because it was two
classes at once. I was in the graduate course and there were eight of us behind
a one-way mirror and on the other side of the mirror there were thirty Harvard
undergraduates who met twice a week in a class that was this group where they
were told nothing about what to do or how to do it. And the class was to simply
observe what happens in free space when people get together, which is a
beautiful analogue of the web. And Bales developed a system to analyze it
called Symlog. Symlog its an acronym for systematic, multiple level observation
of groups. And the fascinating thing about Symlog is that there are now decades
of findings of groups. And the one place that this has not been applied so far
as I know is to social dynamics on the web. But let me give you some of the
next generation, after Bion, thinking about groups.
Bales had
a typology of groups. One is the unified group. This is what I guess Bion
called the cohesive group, the well-functioning group, this is the star
performing team. This is the group that gets it done on time with everybody in
good spirits and celebrates. Then there's a group where people are simply
fragmented, where people are off, they're just not working together. And then
there's a group that's polarized and that's the enemy group that Bion so
brilliantly described. Now Bales went beyond this to talk about typical roles
that people tend to fill in a group.
One of the roles, and this is the role you need in every
well-functioning group, is a leader who's able to keep people on focus and on
target and make sure we're having a good time, and are in harmony and unified.
So this leader pays attention both to task and emotion.
Both to task and emotion. Then there's a type which is someone
who's very popular - who gets in touch with people, who's outgoing, who's open,
who's sociable. But that’s a different function. There is a type who is kind of
a nurse, the protector, the one who comes to the defense of people who are
attacked, who's very encouraging for example, and who keeps morale up. Then
there's the clown. We all remember the classroom clown, well that's one of the
types. Someone who just keeps it funny and by the way in Bales' research it
suggests that every high-functioning group needs to have fun. Needs to clown
around, because it keeps it light, it keeps it playful, it means that you can
say things and take risks, put it out there, without it being deadly serious
and maybe disastrous. And then, interestingly enough, there's the flamer,
there's the person who's insulting, who's exhibitionistic. It's intriguing to
me that you say that flaming is a phenomenon of groups. It's the exhibitionist
of the flamer that's keeping them going and Bales saw this. There's the rebel,
there's the person who's hostile and threatening and obstinate, the problem
person of the group. There's someone who's very dictatorial and bossy, and
there's someone who is very bureaucratic and business-like and impersonal, and
is only task-focused - tuned out of the emotions. What Bales was able to do was
to establish not just a typology but a scoring system based on what people say,
which is where this becomes possibly applicable to the web. Because if you can
computerize a scoring system of every exchange in a group perhaps you could
post what's going on in the group at this deep level so that people in the
group or designated people in the group perhaps the designated leaders could
then take the responsibility to go off line and speak to someone who's starting
to flame. Talk to the the bossy person in other words I think you've pointed
out that problems in groups on the web are best dealt with one to one.
Shirky: One of
the really interesting things about problems in groups on the web is that
almost all of our political tradition has us facing in the wrong direction. We
have a tradition, from Hobbes, of trying to figure out the relationship of
majorities to group as a whole, and we've been enormously concerned about
majoritarian tyranny. When can and when can't the group enforce its will on an
individual? The tradition of civil liberties, of the bill of rights, all of
this comes out of a focus on majoritarian tyranny. On the Internet generally in
all group tools the risk is almost the opposite. It's autonomy risk. The risk
comes from one individual who wants to slow down or stop or redirect things.
There are almost no mechanisms for the majority, even if it is the entirety of
the group minus one, to redirect that individual's efforts - to either ask them
to remove themselves or to moderate their behavior. And while moral suasion has
proved to be an effective tool, more effective than anyone ever thought it
would be, its effectiveness is not unlimited. And its reach is also not
unlimited. It's really designing environments in which autonomy risk is dealt
with on the web that becomes the new problem. And listening to that typology
and thinking about things like the rebel and so forth, one of the really
interesting things about face-to-face contact is that the warm blanket of
consensus will settle over a group. Even if that consensus is relatively thin
or relatively false. Just to stay together the group needs to create some wave
of living together out of a need for social comfort as much as anything. None
of that happens online. And so where a rebel might be brought into the fold as
the kind of 'oh he always complains' or 'she's always going on about this
particular problem'...
Goleman: By the way, that
counterbalancing can be done well by the clown. Rebel speaks, clown jokes,
done.
No comments:
Post a Comment