Monday, December 17, 2012

Media 'Mistakes' as Digital Dye


In particular circumstances, medical doctors may "[deliver] into the body" a substance called "contrast" in order to "[help] certain areas show up better on...x-rays" (Source). Intuitively, the idea is for a competent medical professional to inject, cause the patient to ingest, or otherwise put, a sort of dye (contrast media) into the patient's body in order to facilitate investigation of various parts of the body that would otherwise be less visible (or entirely invisible).

Is there an analog to this "contrast media" process in the study of information systems? It is conceivable (or so it seems to me) that "high profile" events, such as the latest, tragic massacre at Sandy Hook elementary school provide systems analysts with a relevant data set.

One may imagine that there are (at least) two pertinent considerations.

First consider an admittedly vague category of news report that I may, for convenience, term "outlandish claims". Understand, please, that this label is not intended to convey - covertly or overtly - the notion of falsity. To be more specific, I wish to allow that particular Outlandish Claims (OCs) may very well be true. Although, surely, they might not be.

As an illustration of an OC, consider the case of the "$43 Trillion Dollar lawsuit." Under the ponderous heading "Major Banks, Governmental Officials and Their Comrade Capitalists Targets of Spire Law Group, LLP's Racketeering and Money Laundering Lawsuit Seeking Return of $43 Trillion to the United States Treasury", PR Newswire reported, on October 25, 2012, the filing of what is alleged to be the "largest money laundering and racketeering lawsuit in United States History". (I register awareness of the fact that this September report had relevant antecedent, e.g., a Marketwire article of April 22, 2012.)

Now various language employed in this press release, not least of which is the identification of the exorbitant trillion dollar amount and the labeling of the alleged defendants as "banksters," might lead skeptical readers to question the veracity of the report. Herein I wish to make no claim as to the report's truthfulness. Rather, my interest is in suggesting that such an OC might serve, for interested analysts, as a sort of informational equivalent of a high-contrast dye. After all, the phrase "$43 trillion" is, prima facie, fairly likely to pick out this alleged news story quickly and uniquely. Anecdotally, as I type "$43..." into Google, the Google engine automatically recommends to me (words like) "Trillion" and "Banker lawsuit" as "auto-complete" suggestions.

If one thinks of the internet as a vast "body," one might suspect that "$43 trillion" key phrase might serve as a useful "dye" which would indicate, to interested observers, the precise informational "nodes" that have picked up, and repeated, the relevant story. Notice that this schema arguably holds up whether the story is true or not.

Second, one might also consider stories where reporting "confusions" serve to generate different permutations of a story. In this vein, one has an example ready at hand. As Loren Coleman has summarized:
Some are seeing the media making massive mistakes. The miscues piled up: two shooters were said to be in the school, when there was only one; the wrong name of the suspect was identified; the incorrect Ryan Lanza was named; the Lanza mother was not a kindergarten teacher at the Sandy Hook school; the number of dead was given out prematurely and was wrong; a kindergarten class was said to have been attacked, when it was a first grade class; the Bushmaster semi-automatic rifle was not found in Lanza's car, but instead was next to his body and the primary weapon of the school shooting; and a growing list of mistakes were aired. The media began writing about their own media missteps (see herehere, and here).
Supposing that we focus only on the elements explicitly named by Coleman (and, for all I know, there could be more besides) - viz., number of alleged shooters, name of alleged shooter, affiliation of alleged shooter's mom with the school, victim count, grade level of targeted classroom, and variety of weapon used - we have a model with the following positions: <shooters, name, affiliation, count, grade, weapon>. Now, each of the first three variables have two possible values, the fourth has numerous possible values, the fifth again has two possible values, and the sixth has (by my quick survey of the news articles) either three or four possible values.

Think of a blogger, call her "Ninmah." Ninmah posts on the Newtown, CT tragedy as soon as she hears a news report (or a cluster of them). Let us suppose that her blog post ends up being some instance of the <s,n,a,c,g,w> schema. In effect, the values embodied in Ninmah's post mark it with something like a sort of ID-tag. The tag, one might think, may disclose two separate but related pieces of information.

Number one, the specific values that Ninmah selected (e.g., the name of the shooter and what grade level he attacked) may reveal her source or sources. This is particularly relevant for information that may turn out later to be false. (One of the information spreading investigations cited elsewhere [infra.] concluded that there were "three phases: Expansion, Front-page, and Saturation.") An illustration of this is also provided via Loren Coleman, who quotes the following:
The father of Newtown Connecticut school shooter Adam Lanza is Peter Lanza who is a VP and Tax Director at GE Financial. The father of [the alleged] Aurora, Colorado movie theater shooter James Holmes is Robert Holmes, the lead scientist for the credit score company FICO. Both men were to testify before the US Sentate in the ongoing LIBOR scandal.
Coleman subsequently remarks: "One informant tells me that Sorcha Faal is the dispenser of Holmes' father Libor info. But Faal is allegedly a disinformation agent who makes claims about several people."

One may see, then, that the permutation-tagging route can converge with the Outlandish Claim route in the following way. If a tag includes information that may be justifiably termed an OC, it may be possible to trace the belt of transmission very cleanly all the way from the initial source through various intermediates to various terminal nodes.

Hence, number two, persons that "repost," "reblog," quote, or otherwise reproduce Ninmah's text will be traceable to Ninmah via the repetition of Ninmah's tag.

All of this information, in theory, could be diagrammed as a sort of branching tree.

Is there any good reason to think that this sort of analysis is actually being carried out?

For one thing, one can detect evidence of analyses that are at least in the vicinity of what is being suggested here. Again, a simple, non-scientific Google search seems to show that a handful of pertinent studies have been conducted over the last decade or so. Here is a sample listing:

Philip Anderson, "Complexity Theory and Organization Science," Organization Science: Special Issue: Application of Complexity Theory to Organization Science, Vol. 10, No. 3 (May - Jun., 1999), pp. 216-232, <URL>.

A.C. Salvania and J.P. Pabico, "Information Spread Over an Internet-mediated Social Network: Phases, Speed, Width, and Effects of Promotion," Philippine Information Technology Journal, 3(2):15-25, <URL>.

Dashun Wang, et. al., "Information Spreading in Context,"  <URL>.

"Mapping Cyberspace: Tracking The Spread of Ideas on The Internet," <URL>.

Lars-Erik Cederman, "Computational Models of Social Forms: Advancing Generative Process Theory," American Journal of Sociology, Vol. 110, No. 4 (January 2005), pp. 864-893, <URL>.

Although I have nothing like a rigorous proof of the contention, the following seems likely to me. Given that it has occurred to complexity theorists publishing in the "Philippine Information Technology Journal" to analyze patterns of information spreading over the internet, probably, it has occurred to an organization such as DARPA, as well. And the appeal is, apparently, obvious. The presentation titled "Mapping Cyberspace" contains the comment that "statistical analyses" can be used "to understand the reasons for particular trajectories along which an idea spreads" (op. cit.).

It is evident that this talk of "trajectories" has clear application to the study of "social networking" websites such as, without limitation, Facebook, Twitter, LinkedIn, Google Plus, Myspace, and the so-called "blogosphere." In conversation, I have previously remarked that the now somewhat dated criminological techniques related to the area once called "social network analysis" have perhaps become largely passé. It is apparently true that it is just as important today as it once was to "not only identify key participants" in various "conspiracy investigation[s]" but also to "grasp the connections among" conspirators in order "to determine the scope of an illicit operation" (Charles Swanson, Neil C. Chamelin, and Leonard Territo, Criminal Investigation [NY: McGraw-Hill, 1996], p. 294). However, it may well be that the amount of "legwork" required in 2012 has decreased, as compared to that required in 1996 or before, on account of the fact that people now enthusiastically map (if not flaunt) their own connections. To be sure, I have no real idea the extent to which serious criminals "Friend" their associates on Facebook. (But see here.) I think that my point is clear enough, however.

(Cf. "Just about any kind of update can be used against you, it seems. One deejay posted on his MySpace page that he'd be working at an upcoming party. Agents decided to crash it to collect their cash." [Source: "Tax collectors now hitting social networks to track deadbeats," Mon Aug 31, 2009 11:39AM EDT, Original url:] "U.S. law enforcement agents are following the rest of the Internet world into popular social-networking services, going undercover with false online profiles to communicate with suspects and gather private information... Think you know who's behind that 'friend' request? Think again. Your new 'friend' just might be the FBI. [Source: "Richard Lardner, "Break the law and your new 'friend' may be the FBI," Associated Press, Mar. 16, 2010, Original URL:] & here & here. Examples can be multiplied easily.)

For another thing, what is now broadly termed "complexity theory" has been around for quite some time. Consider just a few examples. Note that there is a "...diversity of thinking within the complexity sciences field, which includes the study of complex adaptive systems, chaos theory, catastrophe theory and a number of other sub-fields..." (Valerie J. Lindsay, "The development of international industry clusters: a complexity theory approach," Journal of International Entrepreneurship, Vol. 3 No. 1 [March 2005], p. 71).

In his 1980 book, An Introduction to Catastrophe Theory (Cambridge UP), P.T. Saunders relates that "catastrophe theory" is well suited "for the study of systems whose inner workings are not known, and for situations in which the only reliable observations are of...discontinuities" (p. 1). Whatever else might be said of the "massive mistakes" and "miscues" noted by Loren Coleman (and rehearsed above) it certainly seems plausible that they qualify as "observations ...of...discontinuities" and, thus, fall within the purview of the catastrophe theorist.

Similarly, according to James Gleick in his popular level explication Chaos: Making a New Science (Penguin, 1987), the interdisciplinary "chaos theory" has been around since at least the mid-1970s, with tantalizing hints preceding it by decades. Such an area deals with systems that conjoin "chaos and order together" and that display a "sensitive dependence on initial conditions" (p. 8). "Scientists...define chaos as 'stochastic behavior occurring in a deterministic system'" (Stacy Shapiro, "Combating risk management woes: Chaos theory may hold some clues," Business Insurance, May 4, 1992, International, p. 41). Chaos theory therefore promises to unlock heretofore mysterious systems from weather forecasting to virus distributions. "[O]nce mathematicians and scientists come to terms with chaos theory -- which holds that order may be found in apparent disorder -- risk managers may learn to use it as a tool to predict the unpredictable" (Ibid.). Not only this, but when we observe "that chaos theory helps to explain, and possibly control, systems that are highly non-linear in nature", we can understand why governments might have an interest (Jason Makansi, "Chaos theory," Power, December, 1993, Vol. 137, No. 12; p. 65).

Again, it seems intuitive and plausible that questions like "which bloggers are picking up and transmitting which bits and pieces of this or that multifaceted story?" could be glossed by the descriptors quoted previously in relation to chaos theory. Such questions could be represented by so-called "stochastic models." And the goal might not simply be to analyze information transfer but, as one of the above quotations mentions, actually to "control" (or thwart?) said transfer. If it now seems that I am making my own Outlandish Claim, consider a recent article sketching the use of stochastic models with the aim of trying to control the spread of biological pathogens.

Sometimes you want to prevent extinction. In other cases, you want to hurry extinction along. Take, for instance, a cold virus hacking its way through an elementary school. Extinction of the virus within the student body is something everyone cheers for. ... Leah Shaw says there are many paths to extinction no matter if you want to hasten extinction or try to stave it off but, every system has one route to oblivion that's most sure and direct. Shaw, an assistant professor in William and Mary's Department of Applied Science, works on the mathematical modeling of extinction. ... Understanding the mechanics of extinction is important in many regards, she says. First of all, it would be informative just to understand the system better. If you can understand how a part of a system goes extinct, you know something about the dynamics, you know what factors are important in the extinction, she explained. But the major reason for interest, of course, has to do with how you might control the extinction. (Joseph McClain, "Plotting the Best Route to Oblivion," States News Service, Press Release: College of William and Mary, Monday, July 18, 2011.)
("I control your extinction..."?; Source)

Additionally, note that many documented military experiments concern investigations that are arguably close cousins of the sort that are being considered herein. For example, " 1957 and '58, an Army plane released zinc cadmium sulfide over northern and western Minnesota..." (Greg Gordon, "Sabo and expert to urge ban on secret civilian chemical tests," Star Tribune, Minneapolis, MN, September 28, 1994, Metro Edition, p. 5B.) The purported aim of this "release" (itself "...part of a large-scale spraying program...") of chemicals was given as follows: " monitor chemical dispersion..." (Ibid.). Incidentally: "One of the sites sprayed in Minneapolis was a public elementary school where former students have reported an unusual number of stillbirths and miscarriages...One former student of the sprayed school told the station that of her three children, one had Down syndrome, another was profoundly retarded and a third had a learning disability" ("Minneapolis Called Toxic Test Site in '53," The New York Times, June 11, 1994, <URL>.).

Moreover: "The Army used serratia to test whether enemy agents could launch a biological warfare attack on a port city such as San Francisco from a location miles offshore. For six days in late September 1950, a small military vessel near San Francisco sprayed a huge cloud of serratia particles into the air while the weather favored dispersal. ... Army tests showed that the bacterial cloud had exposed hundreds of thousands of people in a broad swath of Bay Area communities including Sausalito, Albany, Berkeley, Oakland, San Leandro, San Francisco, Daly City and Colma, according to reports that later were declassified. Soon after the spraying, 11 people came down with hard-to-treat infections at the old Stanford University Hospital in San Francisco. By November, one man had died" (Bernadette Tansey, "Serratia has dark history in region / Army test in 1950 may have changed microbial ecology," San Francisco Chronicle, 4:00 am, Sunday, October 31, 2004, <URL>.).

Relatedly, just this past September 25, St. Louis based news station KSDK reported a story on "...details of the Army's ultra-secret military experiments carried out in St. Louis and other cities [e.g., in Corpus Christi, Texas] during the 1950s and 60s..." (Leisa Zigman, "The Army's secret Cold War experiments on St. Louisans," I-Team, <URL>). The report states that "The I-Team independently verified that the spraying of zinc cadmium sulfide [possibly with radioactive particles] did take place in St. Louis on thousands of unsuspecting citizens. ... The Cold War cover story was that the Army was testing smoke screens to protect cities from a Russian attack." Researcher Lisa Martino-Taylor claims that "There is a lot of evidence that shows people in St. Louis and the city, in particular minority communities [centered on the Pruitt-Igoe housing complex], were subjected to military testing that was connected to a larger radiological weapons testing project" (Ibid.). Digital Journal of September 28, 2012 adds the familiar refrain: "The spraying was meant to simulate the airborne dispersal of biological warfare agents" ("US Army sprayed zinc cadmium sulfide on poor St. Louis residents," <URL>). Again: "In the 1950s and 1960s, the Army used motorized blowers atop a low-income housing high-rise, on schools and in other locations to spray zinc cadmium sulfide, a fine fluorescent powder, into the air" (Jim Salter, "Army: No risk from secret St. Louis tests," The Associated Press/Army Times, Friday Nov 2, 2012 17:52:51 EDT, <URL>.).

"Joker, I've told you, we run two basic stories here. ...Winning of Hearts and Minds...And combat action that results in a kill..."
~ Lt. Lockhart, Full Metal Jacket

In other words, the Army injected a compound into the air simply in order, at least partly, to track which way the wind would blow it. Are pieces of information released similarly in our digital era?

Finally, in this connection, reflect on the fact that it is a matter of public record that the Central Intelligence Agency has been in the business of media finagling since the 1950s. See here. It's enough to make one wonder at the origin of various pieces of information, like perhaps the number and name of alleged shooters at a particular elementary school, or even something as "historical" as the initial description of an alleged Presidential assassin.
District Attorney Wade ...[held]...a lengthy formal press conference ... in which he attempted to list all of the evidence that had been accumulated at that point tending to establish Oswald as the assassin of President Kennedy. ... [Wade] told reporters that Oswald's description and name 'went out by the police to look for him.'... The police never mentioned Oswald's name in their broadcast descriptions before his arrest. ... (Source)
If the police didn't release that information, then who did? I don't know.

 "The Pentagon...has begun deploying forces to mount psychological operations, or 'psy-ops'"
Source: Carla Anne Robbins, "Spin Control: U.S. Has Early Priority: Managing Its Message? Stifle News on Military? 'Psy-Ops'," Wall Street Journal, Eastern edition, New York, N.Y.: Oct 4, 2001, p. A.1 
(I owe my knowledge of the existence of this quotation to Michael Hoffman).