Still tinkering with the troubles of the aggregation chasm (as in this here previous post) and the hardly visible but still probably most fundamentally related concept of emergent properties (as in this, on Information), when some dawn of insight crept in.
[Somewhere IL, IN, OH; anyone has definitive bearings? JustASearchAway found it. WI]
Because I’ve dabbled with Ortega y Gasset’s stupidity of the masses for a long time. Whether they constitute Mob Rule, or are (mentally to action) captives of the (or other!) 0.1%, or what. My ‘solution’ had been to seek the societal equivalent of the Common Denominator – No! That may be in common parlance but what is meant here, much more precise, is the Greatest Common Divisor.
Since that is at work when ‘adding’ people into groups: Through stripping differences (as individuals have an urge to join groups and be recognized as members, they’ll shed those) and Anchoring around what some Evil Minds may have (consciously or not) set as GCD-equivalent idea, the GCD will reinforce itself ever more (immoral spiral of self-reinforcement), mathematically inherent through adding more elements to the group for GCD establishment and (not ‘strictly’) lowering. [The only difference being the possibility of a pre-set GCD to center around; just make it attractive enough so the mass will assemble, then shift it to need ..!] Where the still-conscious may not want to give up too much of their individuality but may have to dive under in their compliance coping cabanas just to survive (!?).
So, aggregation leads to the stripping of ground noise which may lead to patterns having been pervasively present but covered by that noise, to emerge. Like statistically, a high R2 but with a low β – but still with this β being larger than any of the others if at all present. This may be behind the ‘pattern recognition’ capabilities of Big Data: Throw in enough data and use some sophisticated methods to ensure that major subclasses will be stratified into clusters and be noise to the equation. [That GMDH, by the way, was the ground breaking method by which I showed anomalous patterns in leader/follower stock price behavior (Shipping index significantly 2-day leading one specific chemicals company; right…) in my thesis research/write-up back in 1994, on a, mind you, all hard-core coding in C on a virtual 16-core chip from mathematics down to load distribution. Eat that, recent-fancy-dancy-big-data-tool-using n00bs..!]
By which all the patterns that were under the radar will suddely appear as patterns in Extremistan DisruptiveLand would; staying under the radar until exploding out of control through that barrier (but note this). As emergent.
But just as metadata is not Information but still only Data, the Emergent isn’t, really. Darn! Close, but no Cuban.
As the pattern is floundering on the research bed when the noise around it dries up, it is not necessarily part of every element in the data pool and potentially can only exist (be visible) at aggregate level. But can and ex-ante very much more probable be part in one or some or many elements of the pool, which would be methodologically excluded from the definition of Emergence / Emergent Characteristics (is it?). And, if the noise is quiet enough, would already be visible in the murky pool in the first place as characteristic not ‘only’ as emergent as the definition of that would have it.
So, concluding… a worthwhile thought experiment, sandblasting some unclarity, but still, little progress on understanding, felling-through-and-through, how Emergence works; what brings it about. But we should! It is that Holy Grail of jumping from mere Data to Information ..!
Joe Cocker just died a couple of weeks ago. Fulfill his request, and little help this friend here, with your additional thoughts, please…