Konstantin Tovstiadi

In a famous piece titled On viewing rhetoric as epistemic, Robert Scott writes:

The attractiveness of the analytic ideal, ordinarily only dimly grasped but nonetheless active in the rhetoric of those who deem truth as prior and enabling, lies in the smuggling of the sense of certainty into human affairs

This is one of my favorite papers ever written on communication or on rhetoric; but that's beside the point. The point is about smuggling certainty into human affairs.

A lot of times we drag certainty in in the metaphors we use. What do we liken our lives to? A road. A path. Our lives do indeed have a beginning and end; but that's where the likeness ends. Our lives are not linear; nor are they marked clearly so one know where to proceed next. Maybe an old 2D arcade game is a more realistic metaphor - dodging enemies all the time, not knowing what's coming next, things getting nastier over time, getting slaughtered every now and then. Except of course you can't reload the game in real life (that's where this metaphor ends, too - they all do).

Some scientists somewhere once asked people if they see their lives as similar to a game of cards or a game of chess. Most picked chess, quite mistakenly. When was the last time you knew the location of all your opponents pieces? Even the game of cards analogy is overly optimistic: at least in cards a jack is a jack and is always below a queen. Real life is even more random.

Which finally brings me to my point: we have bad metaphors for education, too. We think of it as a building. Getting bigger and prettier over time. Growing from a foundation. Orderly with a top to bottom hierarchy - simple facts at the bottom, differential calculus at the top.

I don't feel like my education is like that. The foundation has crumbled away. What I learned when I was ten I could not recall now - at least no more than 10% of it. I used to be fluent in Spanish. I used to know trigonometry. I used to be able to write in hiragana. No more.

And yet I am surprisingly comfortable with my current level of knowledge. I think of myself as an icebreaker - ploughing through the unknown, with enough power to crunch through any problem, one at a time. Loose ice stays loose in my path for a while, then freezes over again. Over time, I gain skill in navigation, a knowledge of the terrain, and maybe even a more powerful engine to run faster and through thicker ice. But I know I will never crush all the ice. The goal is to never be locked in it, not to crush it all.

The icebreaker metaphor lets you forget without feeling guilty. Just keep moving and don't let the water freeze around you, and the rest will take care of itself.

Today I am going to revisit some of my basic How to be successful on the web mantras:

form is nothing. content is nothing. audience is everything
Meaning you could have the best graphic design layout, but with sucky content it doesn't matter. And then you can have spectacular content, but if readers can't access it nor contribute to it, it is a waste of time, too.

for every good idea, there is already an implementation
You want a photoblog? There are tons of open source PHP / MySQL (or Python / Ruby / Perl if you prefer) applications available. Want Flickr exports? There are already plugins that do that. Want to change the layout? Find a theme that would do it. Do not write new code unless you absolutely have to.

if your solution requires manual data entry, find another solution
This is really a variation on the previous rule of no custom code development. Unless you have some really unique data, the stuff that you need is already somewhere, typed up and ready to go. Find a way to acquire it.

the wider the user base, the better the product
Works this way with trails - come to a fork on a trail, not sure which one to take - always take the more trodden one (metaphorically of course the opposite is true, but that's a different story). If choosing between two platforms, pick the one with the larger user community - more people to report bugs, fix the code, write plugins, answer your questions on forums, and develop custom themes.

you gain by giving away
Putting your work online - not all of it, not in full resolution - but enough to give people a flavor of what you do and make them want to eat more - is the quickest, cheapest and the most flexible way to advertise your talents. Plus it makes you look cutting edge - and actually does make you cutting edge. Even if you do a little bit of it, you are still ahead of the game - even in 2009. An amazing number people still consider it unnecessary, too complicated, or worst of all, below them. Getty Images sneered at the microstock model when it appeared. Then they re-thought it and bought the most successful microstock agency. Now that business is pulling them through the crisis as the traditional macro stock market is dwindling and tightening.

I hear it all the time around me in the academic world. "I am not going to tell you what I am working on". "We won't share the custom code we have developed". "My bibliography took me ten years to compile, like hell will I put it online". Or better still, "I put my favorite photo online and so-and-so took it and used it on their site without giving me credit". "How do I put my papers on my site but prevent people from cutting and pasting my stuff?"

There are several false assumptions here. The truths are,

1. Most likely nobody gives a fuck about your ideas. Just because you saw them re-used somewhere doesn't mean they are unique or valuable. You just happened to be on the first page of Google search for a few days.

2. The best way to capitalize on your ideas is to make them public, not to guard them like a dragon with a pile of gold. You gain by giving away (more on that later).

3. Once you have put something online, consider it public property. If it is online and if it is any good, somebody will re-use it without your permission eventually. You are better off accepting it outright and smiling every time you see it happen. Put an image up as a CSS background, and somebody would look in the style sheet and copy it from there. Hide pictures in a Flash animation, and somebody will take a screen shot and steal it that way. If you don't want to share it, don't put it online.

Why is it pointless to protect your data? Well, there is an obvious reason - sharing means feedback, feedback means improvement. The less obvious reason - dummies won't understand it, smart people will steal it anyway. Even if you don't explain how you did something, clever people with similar knowledge will look at the result and reverse engineer the process. Those who are not capable of doing that, are usually not capable of understanding it even if you give them the full detailed specs. If somebody gave me the full die specifications for the upcoming Intel Nehalem 8-core chip, I would gain nothing from it. It would be all Greek to me.

Your data - your photographs, your manuscripts, your ideas - serve as ads for your abilities if you put them online. Let as many people as possible see them. They are the manifestations of things no one can steal from you - your experience, your skills, and your networks.

It doesn't matter if you have access to data. Yes, you can get a million search results on Google, big deal. Do you have the experience to ask the right questions? Do you have enough experience to tell bad answers from good answers?

It doesn't matter if you have the right answers if you don't know what to do with them next. Full processor specs are useless if you don't know how to run a 45 nanometer production line. You may have a Canon EOS 1Ds Mark III with a 7 thousand dollar lens and sill take a God awful picture.

And you may have the best knowledge and skills, but nobody to hire you to do the job. Knowing everything has no value unless you know everyone. Putting ideas online, having an avant-garde, a first-strike battalion of your thoughts on the web, lets you spread the knowledge about what you can do. The benefits of that far outweigh the risks of seeing that cartoon you drew on somebody else's blog.

I used to waste hours of my life playing The Batte of Wesnoth, a strategy conquest game. Some of the maps in the game are hidden: you start with a tiny part of the terrain open and visible, the rest covered in dense fog. Enemies jump out of the fog where you least expect them. As you move your troops, the explored terrain grows, more of the map becomes visible. Yet there are still too many unknowns to predict the flow and the position of new enemies. And then at some point - while a sizable part of the map is still covered in fog - you just know it all, and the enemy does not stand a chance. For practical purposes, a clear view of maybe three fourths of the terrain is as good as the complete view.

And that reminds me of the famous complexity theory thought experiment - imagine a thousand beads on the floor. They are disconnected. You take a piece of string, randomly select any two and connect them. Then you repeat the procedure again and again, until small clusters start forming - 2 beads here, 3 here, 4 over there... And then you reach a point when almost every bead is in a cluster, big or small; but there are still dozens of clusters. And finally with just a few more iterations everything coagulates into one giant cluster. You have not at all exhausted all the possible connections between the beads, far from it; yet everything is connected, albeit not necessarily directly, to everything else. That transition is not unlike the Eureka moment of playing the game and all of a sudden "knowing" the whole map.

So what kind of "knowing" are we talking about? What does it mean to say that somebody knows their stuff? Well, we only make knowledge claims about patterned phenomena. Nobody in their right mind would claim knowing a sandpile, a primarily chaotic system. Up in the Bezengi area of the Caucusus mountains is the 18 kilometer long Bezengi glacier, the longest in the Caucasus. To get to the base camps for many of the climbs, you have to walk up the glacier for about ten kilometers. It is very monotonous; everything looks the same the whole time, especially to a new comer. There is no discernible pattern. Part of the glacier may collapse overnight, shifting the optimal path a hundred yards to the left. It is hard to "know" the glacier.

But even there you can get to a point that you run up and down the glacier in dense fog without a compass and know exactly where you are going. You have learned to recognize the subtle patterns of the system, same as you would "know" a big city. You don't really "know" every street, every house, every station; but with enough knowledge about some of them (like the beads connections reaching a transition) you can predict the rest of them and find them with an almost 100% accuracy. It is almost like you have latent knowledge of every node in the system, based on the positions of some of the nodes and the overall structural principles of the system. This is what Confucius was talking about it when he said, only those who can imagine the three other corners of a quadrant when shown one corner are worthy of being a student.

For years and years and years I have been pushing myself to learn more about information technology. It always seemed like a giant map covered in fog, with me standing in some corner of it blabbering about PHP and CSS. Yet yesterday I realized how far I have come already. I had my "beads-in-a-cluster" / "this glacier has a pattern" moment.

I don't need to see the whole map. I feel (=know) the rest of it already, without having ever seen every element of it.

In a recent meeting, I heard John Snow, the dean of Meteorology at OU, use a mutilated proverb: One man's noise is another man's signal. I am a big fan of mixed metaphors and corrupted sayings - An apple a day makes 365 a year; Wake up and smell the early bird; Beauty is in the eyes of the beer holder.

The source of entertainment in mixed metaphors comes from associations, patterns, predictions that are violated - a mismatch between your internal semantic map and what you hear. And the joy of science and education, on the other hand, comes from discovering patterns and connections and associations - after all, it seems like our brains are capable of no other data processing mechanism than comparing patterns.

So in science, the same rule applies - different strokes for different folks - you pick the type of patterns that amuse you and call it a degree or a career. I personally find two types of patterns particularly engaging. The first type is about the natural structures that exhibit emergent qualities - water atoms creating the pattern of a snowflake, an ecosystem of cheetahs and gazelles balancing at an optimal size, a chain of chemical reactions enabling DNA replication. The second type of patterns that I enjoy are symbolic in nature. In that respect they are different from the first type (not created by humans and not dependent on them). These patterns generate a mind-boggling array of challenges - translation, culture, identity, linguistic change, language acquisition are just some of the problems on that list.

And yet they have a lot in common with the patterns of the first type. Just like snow flakes emerge from the interactions between individual atoms, language changes occur thanks to the actions of thousands or sometimes millions of language users; the changes are dynamic, bottom-up, spontaneous. The spontaneity and the lack of top-down control is what makes them fascinating. Introduce top-down control and the adaptive capacity of the system diminishes or dies.

To me, there are several disciplines too fascinated with the top-down control to be interesting - political science, law, and to some extent history. Rather than start with complexity and emergent qualities of individual agents, they tend to focus on upper level top-down phenomena - social institutions (whose structure is predetermined by a finite set of codified instructions); laws (another rigid symbolic structure) and attempts at post-mortem re-evaluations of social institutions and laws to fit the extant power structure (I am talking about history now).

On April 17, 2009 I am giving the Jesse H. Jones Centennial Chair Lecture at the School of Journalism at the University of Texas at Austin. This page contains links to the resources mentioned in the lecture. You can access this page directly by following a shorter URL.

utexas.mikt.net

Software links

Zotero – a Firefox plugin for personal bibliography management
EndNote – commercial package for personal bibliography management with catalog linking functionality
Connnotea – online portal for personal and shared bibliographies
Cite U Like – a portal similar to Connotea; owned by a commercial publisher
WIKINDX – open source software package for personal and shared bibliographies; needs to be installed on a server
Aigaion – another open source package similar to WIKINDX
CWIS - a digital library software package from the National Science Digital Library

Other resources

Online bibliographies 1 - an early blog post from me about the subject
Online bibliographies 2 - a more recent update
Monster searches with EBSCO host - a blog post describing how to build complex Boolean searches with multiple blocks using EBSCO databases
Creating large scale online bibliographies - a work document used by us at CRCM for harvesting citations. Tailored for the OU IT environment, but could be useful elsewhere - especially if you are an EndNote user.
A blog about bibliographic tools by Kjell Fauske - an extensive summary of local and shared solutions


gapingvoid.com - Hugh McLeod's site, the source of the cartoons used in lecture slides
How much data is that? - a set of examples of different amounts of data, from the smallest to the largest
The End of Theory - a thought provoking article about science in the petabyte age

Lecture slides

A PDF file from the lecture

At work I spend a lot of time working with bibliographies. There are two issues here - finding the stuff and formatting it. That's pretty much all we do. Every step of the way in both processes I follow one mantra: minimize manual work. Automate / batch process everything you can.

At the search stage, that means doing a search in such a way that the results have a higher keeper rate. It means spending more time building a more intricate search string rather than doing a simple sloppy search and then looking through thousands of "results".

Most of the time that means using Boolean search operators - AND, OR, NOT. There are some pretty robust ways to do complex searches in the newest release of EBSCO databases (to which OU and many other research libraries subscribe).

The simplest searches can be done from the basic search interface and don't require a rocket scientist. You can look for

polar bear

or for

polar bear OR brown bear

or even for

polar bear OR brown bear NOT panda bear

But what if you want to deal with a more complex issue - like brown bear or grizzly bear conservation on the American continent, but related to national parks and not related to zoos, and not about efforts made in Europe and Asia? What we have here is three semantic blocks: types of bears, location, and type of conservation efforts. It would result in a complex search string that might look like this:

(brown bear OR grizzly bear NOT panda bear NOT polar bear)
AND
(national park NOT zoo)
AND
(America NOT Asia NOT Europe)

How do you run this monster in EBSCO? The easy way is to run the three blocks as three separate searches. Then, to do a monster search you switch to "Search history" - it would show you the three searches that you have just performed. Check the boxes next to the searches and click "search with AND". (Note: before you do that - make sure you clear the current basic search fields, otherwise EBSCO would add those to your search - bad idea).