b) KM Frontiers: Self signifying intelligence

[First published in Ark Group’s Inside Knowledge Magazine in November 2009]

By Jan Wyllie

INPUT Intelligence


In the previous article, I recounted my understanding (sense) of what Dave Snowden of Cognitive Intelligence told members of the International Society of Knowledge Organisation in May 2009. For me, his term, “self signifying”, pointed not just towards the kinds of intelligence analysis techniques which I have used for years, but it also began to bring into focus what had hitherto been a rather inchoate sense of something emerging in the world of Web 2.x, social networking and collaborative intelligence.

So I added, the term ‘self signifying’ itself to the list of flows being monitored, just to see whether the concept would yield a useful and interesting information flow for analysis. Even though the time frame and the sample are very small rendering the self signifying counts themselves pretty insignificant, the question – is this narrative about self signifying knowledge? — is already creating information flows which suggest perspectives that may never have been seen before, which this article aims to share.

But before going any further, more reflection on what is meant by ‘self signifying’ in this context would be useful. Self signifying knowledge consists of inferences derived from the patterns in metadata as they change over time. The focus is not interpreting the content, but on what analysis of the patterns in the content might signify.

In the world of self signifying knowledge, like in the world of science, there should be no ideology, no searching for right and wrong. It is the patterns in the metadata which are the object of study. Information flows are like rivers. People think that they command and control them with their notions of truth and falsehood. But the self signifying perspective sees the arguments merely as indicators supporting (or not) the inferences drawn from the flow. The object of the inquiry is not whether the content is true or false, but how notions and narratives change. Being wedded to beliefs about what is true and false will blind the observer to next zeitgeist of a collective intelligence.

What the patterns might indicate is always an open question, although over time some inferences can gain credence (but can never be proven), other inferences though can certainly be disproved.

If the object of the exercise is neither to interpret content, nor decide whether it is true or false, what is the relationship between the researcher and the content of his sources? It is still possible to report what is being said. The difference between reporting and interpretation is that in the former the voice of the reporter is eliminated as much as humanly possible. The job is to record who said what to whom and when. No more, no less.

There are academics for whom it is fashionable to say that reporting is an impossibility, and that to pretend that reporting is anything other than an individual’s interpretation is intellectually dishonest. The problem inherent in this way of thinking is that it would, if applied consistently, render the process of scientific research impossible. If scientists could not report the results of their investigations without letting premature interpretation skew the results, the process would be unable to function.

Good reporting is not easy. However, there are ways and means of keeping bias to a minimum, for example, direct quotation from sources, pre-defined questions, summaries etc. The most important factor is a professional disinterest in the content, other than as an indicator from which some future unknown inference can be drawn.

So what is an inference? It is thinking about what the self-signifying data might be indicating. An example might be: ‘As reported cases of swine flu increase, the number of swine flu Twitter Tweets continue to fall. The inference is people’s concern about swine flu is declining, as they become more experienced. It seems obvious, except that next month, swine flu Tweets might suddenly spike indicating the need for a different inference. When information is treated as a flow, rather than as static sets of documents, the data is always changing. Therefore the work of thinking and inferring is never done, if the assumption is that changing information flows are somehow related to events in the ‘real world’. If no such relationship exists, then all information would be useless anyway.

As more information flows are monitored, the opportunities to make inferences become richer because links between developments in different sectors can be made which otherwise would be impossible to see.


In this case, the collection of sources under ‘self signifying’ is very small which is hardly surprising since the concept is so new and only one person is monitoring. The real power self-signifying knowledge happens when groups of people monitor the information flows. Humanity then enters into the domain of collective intelligence.

Creating meaningful information flows is a process akin to asking questions in order to see what the sources are addressing. I found that the material could be meaningfully analysed according to the questions how? and why? Then, somewhat unsurprisingly, under the ‘How?’ category I found that sources broke down nicely into Techniques and Tools. I analysed the ‘Why?’ category into Purposes (obvious) and Reports where I have put inferences that are being made by the early practitioners.

Please note the open nature of the questions which create the information flows. The judgements needed are not about agreeing or disagreeing, right or wrong etc. They are about the application of neutral concepts. At this point the analyst (me) is literally disinterested since I could not care less whether a source is classified as a Technique or a Tool. My only interest is to be as accurate as possible using the best of my native intelligence. Remember the process works much better when groups use the process.

Once the articles are classified, then it becomes possible to make counts of how the numbers under each heading change over time. In this case, the time is too short and the numbers are too small to be meaningful. Nevertheless, note that in the sources that I was monitoring, there were twice as many articles about Techniques, than there were about Tools. Already, the question can be posed whether this statistic is significant and whether any inferences can be drawn. Perhaps it is too early to tell, but I would certainly want to know if the percentage of articles about Tools began to increase compared with Techniques. I could then know when an emerging idea becomes a real business opportunity, for example.

Since because of the small sample, it is not possible to draw any inferences from the counts, all we can do is report on the characteristics of the content in the different question flows. I cannot emphasis to much that the narrative indicators collected under each heading have not been collected to prove a point, rather they appear in juxtaposition as a result of a disinterested process. The job is to report on what is found there, that is, a form of knowledge which could not otherwise have been known in advance.

The next section results from the application of the method outlined above on an initial collection of articles pertinent to the term, self signifying. The findings are, I hope, of some value despite the very small sample used.

OUTPUT Reports




Report: Two approaches to generating self-signifying knowledge are identified: 1) automated output and 2) socially constructed output. The former approach is much more cited than the latter.

Inference: A battle of intellectual paradigms could be brewing.


The Social Media Marketing Book spent nine months analyzing roughly 5 million tweets and 40 million retweets. He noted when they were posted, which words they used, whether or not they included links, and more.


The more citations, the more powerful the concept appears to be. The recommendation economy that has defined academia for centuries has been brought to the Web was started in the same fashion; the more citations of a page, the higher the ranking). [See] Dan Zarella’s brilliant quantitative analysis


Real-time web search – which scours only the latest updates to services like Twitter – is currently generating quite a buzz because it can provide a glimpse of what people around the world are thinking or doing at any given moment.


The counts are a measure of the attention given to a topic by the sources, and how they change over time. … After time passes, it becomes possible to compare “what is being said” during different periods, then we are in a position to “sense” turning points and trends in the data, and it all becomes much more interesting.



Report: Volume and distortion of information flows are portrayed as unresolved concerns.

Inference: Methods enabling degrees of reflection and ethical standards will be at a premium.


But with more than 10,000 tweets per hour about the [swine flue] virus, keeping up with the news has become challenging – setting up a Twitter search or following a hashtag is pretty much worthless at this point, unless you want to dedicate your entire day to monitoring it.


When we observe something we change it. It is a recognized effect in physics. By allowing us to observe trending topics in real time, Twitter gives us the opportunity to change them. Contrast that with post-hoc analysis like Google. It is way past the event, and as such, harder to influence. This is the danger [and opportunity Ed.] with real-time statistics. When we find out about something in real time, we don’t just observe, we also participate and change the phenomenon itself.


About 50,000 activists are reported to have downloaded a programme called Megaphone that sends an alert to their computers when an article critical of Israel is published. They are then supposed to bombard the site with comments supporting Israel.




Report: Trendrr is presented as a favourite among a growing number of self signifying knowledge tools.

Inference: Look out for more tools, as well as any indication of a revenue stream.

Trendrr is a new favorite among analysts looking to keep track of trends and compare information. It tracks statistics on multiple social media platforms, but for Twitter, its Twitter Search graphs are invaluable – they provide graphing of keyword mentions on an hourly basis. There’s also Twitter user stats available, and all of this information can be compared to things like blog mentions. Also check out TweetStats Trends. For more in-depth information about trend tracking on Twitter, take a look at last month’s article, 15 Fascinating Ways to Track Twitter Trends


Twitter Search will also get a reputation ranking system soon. When you do a search on a “trending” topic (a topic that is so big it gets its own link in the Twitter.com sidebar), Twitter will take into account the reputation of the person who wrote each tweet and rank search results in part based on that.



Report: Google dominates yet another field by providing automated analysis of the massive amounts of data it controls.

Inference: As self signifying data becomes valuable, copyright and access conflicts are liable to escalate.


Why is “Ethylene Oxide” #3 on Google? ATSDR – ToxFAQs™: Ethylene Oxide Exposure to ethylene oxide can … cause cancer? Is there a medical test to show whether I’ve been exposed to ethylene oxide? …


Google researchers Hyunyoung Choi and Hal Varian combined data from Google Trends on the popularity of different search terms with models used by economists to predict trends in areas such as travel and home sales. The result? Better forecasts in almost every case


Google is the biggest participant in this new market with products such as its much publicised Google Flu Trends.



Report: Socially constructed self signifying knowledge practice is a comparatively rare and unknown phenomenon despite many years of experience.

Inference: Growth potential is high.


Both Cognitive Edge and our own company, Open Intelligence, are both early instances of using self-signifying data created by human intelligence to make real time inferences about group awareness.




New Consciousness

Report: A new form of “global meta-consciousness” is envisaged in which the human mind changes from being ego- and nation- centred to a sense of “oneness” with others.

Inference: If something simplistic and raw as Twitter can invoke ideas of a new level of consciousness, it becomes important to think about the implications opportunities, and dangers as people become more adept at working with it. H. G. Wells concept of World Mind is re-emerging more than 70 years after he conceived it.


[Twitter’s success] corresponds to the accelerating spread of a global consciousness, one in which our sense of boundaries no longer end at national boundaries and we are increasingly in touch with our sense of “oneness” with others. http://openintelligence.amplify.com/2009/06/19/the-philosophical-significance-of-twitter-consciousness-outfolding/

[It is] one more technological innovation enabling the outfolding of consciousness — the collective turning-outward of human thought. A single tweet  [could be] the butterfly’s wings that eventually leads to a big bang of global meta-consciousness. interconnected now.


Collective Intelligence

Report: In the view of its advocates, fluid conversation is seen as replacing static documents, as the principal repository of human knowledge. Writing and editing functions are reported becoming less centralised and hierarchical, and more distributed and community based.

Inference: Although the Collective Intelligence ‘clip stream’ is not quite as all encompassing as a new level of consciousness, it is developments here that are likely to be the most significant because this is where the hard work will need to be done before anything like a World Mind is to emerge.


The value of the content itself … is nothing but fodder for sense-making conversations.


Twitter is less about defining identity than it is about managing the creation of communities and the flow of information. Like the production of news, the ability to decide what is newsworthy is fast becoming a bottom-up process. Users can influence content by shaping trends.


Documents are no longer the focus of communication. Rather, documents become just one element in a conversation. And a conversation, one might note, in which any kind of editor function has been eliminated. It remains to be seen [whether] disintermediation helps or hinders effective information sharing.


The question, then, is whether there is more in the tag cloud than mere words? [5] Does something like a “collective intelligence” leave its traces in a tag cloud?



Virtual World

Report: The list of top trending topics suggests that the ‘twitterverse’ is dominated by a “twitparade” of fashion, celebrity and commercial topics, rather than significance. Reported measurements rank on single words or phrases, severely limiting the kinds of questions which can be addressed to Twitter communities. Research is cited suggesting that the mainstream media is still overwhelmingly the source of trending topics.

Inference: The content of Twitter has an even higher noise to knowledge ratio than the traditional document. The imbalance increases the need for new ways of filtering out noise.


Bing Tweets Trend Topics: Twitter: Popular Now; People; Places; Products


Today’s Twitter TrendCloud – AIDS Apple Better Designer Big Screen Kindle Boston Globe #comedyheresy Cyber-FM #ecomonday #edaust09 Fourth Goodnight H1N1 Happy Star Wars Day #hoppusday I Can’t Believe How Kindle Larger Kindle Masterchef May The Fourth Be More Sex Than One #musicmonday Newspapers Coming As NIN App Gets #StarwarsDay Steve-O Swine Flu #swineflu Terrific Twitter Trent Reznor Wolfram Alpha Wolverine X-Men Origins


For the most part, the traditional news outlets lead and the blogs follow, typically by 2.5 hours, according to a new computer analysis of news articles and commentary on the Web. 3.5 percent of story lines originated in the blogs.


‘Real’ World

Report: Two types of application are reported: forecasting and intelligent human sensing systems.

Inference: The lamentable performance in current practice relating to both forecasting and real time sensing systems make the need for methods of improving them all the more important. Self signifying inferences are only as good as they are useful. They need constant testing.


Twitter engineers noticed that the word “earthquake” had suddenly started trending up. They didn’t know where the earthquake was. Several seconds later, their building started to shake.


The bots crawl through relevant web pages, noting keywords and examining the text around them. The theory is that this gives an insight into the “wisdom of crowds”, as the thoughts of thousands of people are aggregated. Its study of “web chatter” is said to give advance warnings of terrorist attacks, and proponents claim that it successfully did so ahead of 11 September 2001. Its latest and most sweeping prediction is that 21 December 2012 signals the end of the world http://openintelligence.amplify.com/2009/09/28/web-bot-project-makes-prophecy-of-2012-apocalypse/

By trawling scientific list-serves, Chinese fish market websites, and local news sources, ecologists think they can use human beings as sensors by mining their communications. Much of the pioneering work in this type of Internet surveillance has come in the public health field, tracking disease



The analysis suggests that a new phenomenon has been identified which merits the continued use of the term ‘self signifying’.

It is too early to say whether this new perspective will be the seed from which a new level of real time global consciousness will sprout.

In the meantime, opportunities exist, especially in the areas of marketing, as well as socio-technical, cultural, political and economic trends, to develop self signifying intelligence systems. Knowledge managers and taxonomists should be key players in realising those opportunities.

It’s all about designing forms of questions to be cast like nets into the knowledge flow.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: