On the pages Why No-One’s Noticed This Before, The Dreaded Jungian Backlash and Logical Effects I describe how the effects of spending too much time in focussed attention conspire to conceal themselves. This includes the simple truth that we don’t notice what we don’t notice, our culture’s accomodation of focussed attention by leading us to describe our work in reactive, prescriptive terms, and reduced expectations of - and reliance on - self-consistency. Most strangely I argue that groups of people trapped in focussed attention are liable to escape contradiction when things go wrong by deluding themselves. All of this fits with studies of groups under stress. We know that stress makes groups more reliant on heuristics, less analytic, less aware of context. There are elegant “choice blindness” and “conformity” experiments which reveal the surprising way that ordinary people will backfill their own memories with fantasy.

It is not my purpose to dwell on these issues. What matters is that it can happen, it’s a spiral of decline as each fantasy leads to another, and it always reduces efficiency and quality. We need to be aware this kind of thing can happen, and ensure we maintain conditions to prevent it. It’s like finding rotten timbers - the exact growth pattern doesn’t matter, we just need to get rid of it. Confusion can reach a point where we need to rule off, take a deep breath and start again. Over-analyzing each mess is of little value - unless we are studying messes!

To demonstrate how crazy things can get and provide a little entertainment, I’m going to analyze a story that’s been in the news this week. It makes no sense at all until we think about it in these terms, and then what is happening is crystal clear. Do remember that this is a good textbook example - and textbook examples are inherently misleading. Real life is rarely this clear cut. Anyway, on with the yarn…

It comes from the Chicago Sun-Times, a gem entitled Voters are told pen had ‘invisible ink’. Ahem…

Voters are told pen had ‘invisible ink’

BY ANNIE SWEENEY Staff Reporter/asweeney@suntimes.com

When it comes to election shenanigans, Chicago has been accused of just about everything.

But invisible ink?

Twenty voters at a Far North Side precinct who found their ink pens not working were told by election judges not to worry.

It’s invisible ink, officials said. The scanner will count it.

But their votes weren’t recorded after all.

“Part of me was thinking it does sound stupid enough to be true,” said Amy Carlton, who had serious doubts but went ahead and voted anyway.

As it turns out, Carlton was one of 20 voters at the precinct who were given the wrong pen to use. They were also then told, apparently by a misinformed judge, that the pens have invisible ink, elections officials said.

As a result, the votes were not counted. But officials insisted there were no dirty tricks involved.

“This one defies logic,” said Jim Allen, a spokesman for the Chicago Board of Elections. “You try to anticipate everything. But certain things just … they go beyond any kind of planning you can perform.”

By late afternoon, five voters had been contacted and told to come back to the polling place to vote again. And elections staff had left messages at the homes of the rest, Allen said.

Carlton and Angela Burkhardt, another voter who was told the same invisible ink story, spent a good part of the day calling and e-mailing the Board of Elections to get answers.

“I am furious and devastated and I just feel stupid,” Carlton said. “I feel so angry.”

Both women agreed that this election meant a lot. They had spent a good deal of time researching candidates.

“I have been voting since I was 18,” said Carlton, 38. “This is the most important election of my life so far.”

Burkhardt planned to go back to vote late Tuesday. She worried about those who might not be able to return.

“I worry about the other people who were there,” she said. “Maybe [they] can’t get off work. I am a person of privilege. I can go back. What if you couldn’t?”



OK. First let’s dismiss the conventional kind of explanation. The piece suggests election shenanigans, but that’s just not a good theory. For one thing the disenfranchised voters left the polling station aware that something funny was going on. We need to find another cause.

If Chicago is like most places, the election official will be a local bureaucrat, postal officer or other person renowned for behaving in a predictably focussed attention kind of a way. There will likely be a strong addiction to the kind of low level background stress that such organizations generate within themselves. This will lead to a deeply habituated tendency to defer to heuristics - a fixation on proceduralism. Now elections are important and are taken seriously. The official might even need to be sworn in, and the apparatus of the Voting Procedure will have been explained in great detail. There isn’t much to it, but the Sacred Pen, Sacred Ballots and Sacred Box will each have been considered with due solemnity. There will have been no doubt that the Sacred Pen is an integral part of The Procedure, which is given a strange, ineffable mystery because it gives its devotees a stress hit with chemistry similar to the effects of cocaine.

So with the system boundaries and heuristics thus arranged, we arrive at the moment when a human being deferred to the authority of a broken pen. This caused a broken narrative when others also noticed, which the overly “compliant” official, filled with background fear, stress and defensiveness, subconsciously filled in with confabulation. The pen uses invisible ink! We can forget our pride in our wonderful technological society. The fantasy is funny not because it is alien but because we all recognize the poetic resonance with the Brothers Grimm!

Confabulators have a convincing certainty about them - it’s probably something to do with having suddenly reduced their fear. That, the tendency to groupthink, and likely the deluded one’s Armband of Office would all have pressured all the other workers and the voters into a textbook conformity experiment. The rest, as they say, is history.

It’s important to realize that no white cats were stroked in the making of this mess. The key event - the invention of the invisible ink - occurred without conscious reflection, in a mind squeezed like a foot in a too tight shoe, at a peak of reactive fear. Unfortunately, no conspiracy is necessary for there to be damage. On this occasion the delusion has seen daylight and at least some voters have had their ballots collected, but what tends to happen is that the nonsense kicks around and causes knock-on evasive fear much worse than the threat posed by a broken pen.

I once saw a work in progress management system that was working very well, but for some reason scaled really, really badly when it produced reports. The thing was quite well structured, there were about a half dozen people on the team, and a myth that a database server was inevitably compute intensive. The business logic never directly used the database - instead there were higher level calls. There was something odd about those calls though. When I tried to understand them, at first I failed. I just wasn’t getting it. I asked around, and realized that the team members weren’t talking straight to me. There was something very funny going on.

Eventually I got to the bottom of it. The guy in charge of the database stuff had a blind spot. He obviously had no concrete mental model of what a C program does and/or he’d never done the thing of thinking things through, round and round, explaining everything to himself. He’d written some frighteningly good code by applying rote learned incantations, but there are limits to the possible. He didn’t realize that an SQL query could be assembled in a buffer with the things to search for filled in between the SELECT, WHERE, JOIN and so forth. All his queries were in hard coded literal strings. He’d got some quite clever literal strings, which worked well with the somewhat distorted database schema to implement the bizarre programming idiom. These strings would be issued to the database, and then he’d parse and filter the result sets to find the records he wanted, doing memory management like crazy while he was at it, but never cottoning on to the fact that his query strings were just char * pointers that could point to any text he wished to cruft up!

There must have been blood on the carpet at some point (I didn’t find out about that, and didn’t want to), because this whole business was a no-go area. When discussions approached it, otherwise rational people would start logic-chopping and weaseling, as they attempted to make a continuous transition from the truth to Mr. Database’s misunderstanding. The sheer amount of processing needed to support the wriggling must have detracted from their work. My job was supposed to be tweaking the database and UNIX platform to make it go quicker, but the Project Manager was a canny fellow. I eventually figured out the real reason he’d hired me was to walk in, speak the truth and then leave discreetly.

The funny thing was, Mr. Database wasn’t stupid, just fixated on a misunderstanding and without a rich enough model to get him out of it. I wrote a “test program” that prepared SQL with sprintf(3) to INSERT and SELECT some pattern data, and showed it to him along with some “timings” (which I didn’t really care about at all). As soon as he saw what the program was doing he became very excited, and immediately saw how make things go faster.

Still, I’m looking forward to the Chicago Council coming to order in their splendid new Robes of Office. Will there be webcams? ;-)