Pattern matching is an easy way to check if a thing looks right. Serial processing is a hard way to check if it is right. A big difference. hydeandrugg.wordpress.com
There are two computational mechanisms for solving a problem, regardless of whether you’re a human or a computer. One of these mechanisms is parallel processing, where you carry out lots of tasks at the same time; this mechanism is very good for pattern matching, where you identify patterns (whether physical patterns, or underlying regularities in events, etc). The other mechanism is serial processing, where you do one task at a time; slow, but steady, and much better for catching errors in reasoning.
Humans are very good at pattern matching, which we find swift and easy, and very bad at serial processing, which most of us find slow and painful. So what? So this is why we appear to be an illogical species, and why demagogue politicians can get so far despite having policies that are little more than word salad.
Pattern matching is fast, easy for us to do, and lets us solve a lot of everyday problems (like crossing the road) that are extremely difficult to handle via serial processing.
That’s fine if you’re crossing the road, or recognising the person that you’ve just met. However, it’s not fine if you’re trying to follow a chain of reasoning, whether it’s a legal argument or the assembly instructions for something you’ve just bought. For the chain of reasoning, you need to use serial processing, which human beings find difficult, slow, and hard work.
So there’s a strong temptation to use pattern matching for tasks like assessing what a politician has just said, as a quick, easy approach – even if it gives the wrong result sometimes. Politicians learned this a long time ago, and populist politicians are very good at producing word salad, which doesn’t actually make any sense when you examine it via serial processing, but which has all the right ingredients to look good if you only examine it via pattern matching.
The same mechanisms occur across a huge swathe of human activity. The distinction between them is important. Once you start doing that right, then you can start dividing tasks up efficiently between humans and computers, since we and they have complementary strengths and weaknesses. The Search Visualizer software is one example of how that can work in practice – humans are slow readers in comparison to how fast they can scan patterns, so if you want to assess something quickly within a text, then it’s a good idea to translate that text into a format that you can handle via pattern matching.
There’s a more in-depth discussion of this in my book with Joe D’Agnese, Blind Spot: http://www.amazon.com/Blind-Spot-Solution-Right-Front/dp/0062097903
Some articles and links:
This is one of our series of Tweet sized thoughts. These are intended as a quick, effective way to publicise concepts that aren’t as widely known as they should be. We’re using this format rather than simply tweeting because it gives us a way to unpack the concepts in more detail, with examples, links and references if needed. We’re taking our usual approach of indicating useful related concepts in bold face where it’s easy to track down further information, and of only giving references where they’re not easy to find.