Invisible Influence

The chaps over at Anecdote have taken to thanking the people who link to them or comment on their posts, once a month or so. It’s a nice thing to do, though the cynical might say they are showing off their network visibility and the insecure might say “hey I’d better comment, all these other people are”. But even if there is a bit of theatre about this, where’s the harm? They are out to make a living, they do great work, and we all display our feathers once in a while in our own different ways.

But a coincidental comment by Steve Denning on the com-prac forum yesterday made me think. 

I’ve never found Steve to be particularly collegial so I don’t follow him closely in communities or on his blog, but this particular comment caught me. Shawn Callahan had posted about the fascinating work that Anecdote is doing with the Most Significant Change technique, a process for impact evaluation using stories about changes brought about by any given initiative, and choices by stakeholders in regard to those stories.

Steve’s question was, what about non-changes? What about an initiative that was, say, intended to stop something happening? He gave the example of taking vitamin C to ward off colds or flu. Now there’s more to unpack in Steve’s question, but this reminded me of an on-off conversation I’ve been having with Maish for some weeks now, about the importance of mapping “white space” – ie the things that are not currently there – when we do information and knowledge mapping. If we don’t try to map the stuff that’s not there, we’re only mapping our assets and resources for history, not for current and future needs. It’s an aspect of ignorance management.

But I digress. So there I am, looking at Anecdote’s thank you list, with Steve’s question ringing in my mind, and I realise – in a flash – that I hardly ever link to or comment on the blogs that influence me most. There are four blogs that I read religiously, all the others on our blogroll are good for an occasional though often illuminating visit (no, I haven’t entered the RSS generation yet). My four cannot-miss blogs are Dave Snowden (for truth, humanity and intellect), Anecdote (for their reading and their practice), Maish Nichani (for his serendipitous treasures) and James Robertson (for his pragmatic no-nonsense say it simply approach). And I hardly ever link to them. Or comment.

I’m not entirely sure why. It may be that I have unconsciously turned them into extensions of my own cognitive apparatus. I don’t see them very often because I see through them. The cynical might think I want to keep them all to myself, so I can appear clever without attribution whenever I feel like it. The insecure might think I don’t want to appear derivative by acknowledging the extent to which they influence me.

But now there are all sorts of bells going off in my head (metaphorical ones, to those of you who might be worried). The last few days the on-fac community has been energetically discussing the ownership of words in forums, and attribution of ideas to their originators (for overlooking which Steve Denning has been scolded in the past). I have been simultaneously reading Jorge Luis Borges, who if you’ve read him, is a terrible literary tease. His stories are full of (and sometimes composed entirely of) direct, indirect and heavily manipulated quotations, some of which are attributed, not attributed, or misleadingly attributed to their original authors. He plays with plagiarism as a fire eater plays with fire, and some of his intent is surely to point out that no author authors alone. Anything we compose is (almost) always composed in the midst of a social conversation space.

There are other forms of invisible influence, and they have particular relevance in KM. Tonight, Graham Durant-Law is presenting a talk to the ACTKM folks in Canberra on his experiences in using social network mapping tools and measures. His data is taken from the discussion threads on the ACTKM listserve as it reconstited itself between January and June this year. The measures suggest certain individuals play key brokering roles in the community, and suggest certain combinations of personalities who form sub-groups or cliques.

But as anyone who participates in an online community knows, the visible explicit messages are only the tip of an iceberg, the bulk of which is back channel email chat between participants, between participants and moderators, face to face meetings and conversations, blog posts and comments, and so on. Not to mention the likes, dislikes, gossip and petty politics that goes on among friends. That is to say, the visible conversation space can only take you so far in mapping apparent influence networks (as Graham recognises in his talk). Even interview based techniques only give you the influencers that the participants want to display.

And then there is the invisible influence of withheld participation. I don’t reply to people in public forums if I don’t trust them. I don’t participate when I’m not confident that my contributions will be treated with respect (or rather, am confident that they will not). I don’t continue to comment on blogs if I’m not acknowledged (eg if my comment seeks a response). And while this influence is minute at the individual level, if this kind of influence is reflected across a population, then this can affect shifts in participation and activity and popularity and visibility and – visible influence – of the communities, forums, individuals that benefit from the empty spaces we have created by our avoidances.

To my mind, it’s the invisible influences that tell us much more about important influences. The visible ones tell us about conscious display, and to a degree about explicitly acknowledged influences, but they don’t tell us the whole tale by any measure of means. The fact that they are easier to track and map is tempting, however. We’d be happy to settle with them, particularly if they give us pictures that we like. Social network maps, hyperlinks and commenting behaviours are all informative, but give us clues that need to be interpreted at much deeper levels before we can be confident we have mapped the stuff that counts.

9 Comments so far

Shawn Callahan

A fellow Borges reader. I should have guessed Patrick. I was reading some of his stories today and of course this is easy when they can be as short as 2 pages in length. Thanks for the hat tip.

Posted on September 05, 2006 at 05:59 PM | Comment permalink

Patrick Lambe

And so should I! What hat?

Posted on September 05, 2006 at 06:26 PM | Comment permalink

Tonight’s presentation on the network analysis of actKM was terrific.  What really struck me was the gender imbalance in the network participation.  Yet on further analysis the one key female has a pivtal role to play.  Great stuff Graham.  Looking forward to learning more about this.

Posted on September 05, 2006 at 06:44 PM | Comment permalink

Patrick Lambe

I could tell it was going to be good from Graham’s slides, I’m only sorry I couldn’t be there for it. The gender imbalance also came through from Greg Timbrell’s paper for last year’s actKM conference,
and from the study of conflict participation I did a few months ago.


Posted on September 05, 2006 at 06:54 PM | Comment permalink

Graham Durant-Law

Hi Patrick,
What a great blog and site! Thanks for mentioning my presentation and including a link to my site.  The presentation can be found on my site at the bottom of the “My Papers” page.  I’ll send you the PowerPoint as well.  I’m writing a PhD at the moment so I did this exercise to understand the tools and any potential data problems using real data, and for no other reason - but none the less the results were fascinating.  Some of the results I excluded were just too meaningless in my opinion and some that I have included border on the meaningless. This was the real point of my presentation – real caution needs to be applied when doing an analysis and the user must understand the tools, the assumptions, and the mathematic algorithms underlying them.  I’d be happy to expand on this for anyone who is interested. 
Best Regards

Posted on September 06, 2006 at 06:11 PM | Comment permalink

Its a great post Patrick - thinking about it cost me an hour of time that should have gone on the book

Posted on September 09, 2006 at 06:28 PM | Comment permalink

Matt Moore

Borges is a great writer - The Library of Babel and Funes The Memorious might be descriptions of disastrously successful knowledge management projects.

Regarding SNA, I actually see it a tool of limited analytical use. When you start mapping a network, you change it. And if you are mapping a network, then you are normally doing so because you want to change it.

Posted on September 10, 2006 at 12:01 AM | Comment permalink

Shawn Callahan

I agree with Matt that SNA is of limited *analytical* use and it is better employed as a sensemaking device that triggers new conversations in an organisation.

Posted on September 10, 2006 at 09:36 AM | Comment permalink

Graham Durant-Law

Dear Matt and Shawn

I don’t think SNA/ONA is of limited analytical use, instead one must know the limitations of the methods.  It entirely depends on what it is being used for, and how it is being employed. 

I think ego SNA has some limitations.  For example the actKM network analysis has 140 actors, but only 106 are linked.  So what you say.  My post to the actKM List elicited just over 60 private responses, most of which are from people I have never heard of, and I assume (perhaps wrongly) are actKM members.  Obviously there is a strong and vibrant underground network which an ego SNA may or may not reveal.

On the other hand I have also mapped project relationships in Defence – specifically this project is dependent on that project to deliver part of the outcome.  I then correlated this with a SNA of project officers.  The method is analytical and the results are most useful from a corporate viewpoint.


Posted on September 10, 2006 at 11:04 AM | Comment permalink

Page 1 of 1 pages

Commenting is not available in this weblog entry.

Comment Guidelines: Basic XHTML is allowed (<strong>, <em>, <a>) Line breaks and paragraphs are automatically generated. URLs are automatically converted into links.