cognitive artifacts.mp3
[00:00:00] Happy 2019 everybody. This is Mark. And today I'd like to talk about cognitive artifacts. As people are tool builders, we've made many, many kinds of artifacts ranging from hammers to crowbars to bicycling to cars and cognitive artifacts are one subcategory of those. Cognitive artifacts are physical objects that we've created for improving or enhancing or in some way changing our cognition.
[00:00:30] An example would be a todo list. It's a physical thing that you create that helps you remember the things you need to do and track which ones you've already done. A string, tied around your finger as a reminder, would also be the example. A calendar helps you organize your days and make longer term plans. There are some interesting characteristics of cognitive artifacts and there's is a pretty sharp divide between two kinds—some that are complementary, and some that are competitive with human cognition.
[00:01:04] Let's go back to a couple of the examples I listed at the beginning: the bicycle and the car. Both of them help you get from point A to point B more quickly than you otherwise could. But in the case of a bicycle, you're training your body as you do it and becoming stronger and more capable of moving from point A to point B, even if the bicycle were taken away. With a car on the other hand, the more you use it the less fit your body becomes and the less capable you become of traveling without the car. The same kind of divide exists with cognitive artifacts.
[00:01:41] Let's look at a few examples of complementary cognitive artifacts. One of the very best is the written number. Whether it be Arabic numerals, Roman numerals or Chinese numerals, written numbers gave ancient people greatly increased abilities to talk about, to represent and even to think about large numbers... instead of just say the numbers between 1 and 10.
[00:02:03] However they were not equal. Roman numerals for example were just fine for representing numbers but not very good for computation. Mathematicians in ancient Rome—and even centuries later in Europe—who were still using Roman numerals, actually struggled to multiply large numbers by each other. And long division was pretty much right out.
[00:02:27] In contrast, mathematicians in the Arabic world, armed with a slightly different and in this case better cognitive artifact, had no problems dividing numbers. And it wasn't just because they had the numbers in front of them. If you've ever learned the Arabic numeral system—as most people in the modern world have—and you've learned how to do division, then there's a very good chance that with no books, no pencil, no paper and no writing implements, you can do more complex division in your head than the best mathematicians in ancient Rome could do with all of the resources at their disposal. That's the power of a cognitive artifact. And in this case even when it's taken away you can do amazing things.
[00:03:13] A second example is the sextant, which was a navigation tool sailors use to use. Its primary use was measuring the angular distance between the horizon and any given star. And, this let them do things like calculate what their longitude was, so they wouldn't get lost when making a long voyage. As with written numbers, experience using the sextant made it less and less necessary. In other words, a sailor who had many years of experience could simply look at the horizon and stars and get a reasonably good estimate of their longitude. Maps themselves have this property, to a degree. Someone who has used and studied a map that represented the sum efforts of thousands of people will still be able to navigate when the map is taken away from them. In fact, someone with that knowledge would have an incredible advantage in ancient times, when the map simply hadn't been made yet.
[00:04:12] And finally, perhaps the most impressive example of a complementary cognitive artifact, except for language itself, is the abacus. That is an old tool which was used for doing calculations by sliding beads along a wire frame in different directions, with upper and lower decks used to identify orders of magnitude. And with an abacus, people could add, subtract, multiply and divide large numbers very quickly. Even better, after becoming an expert and using an abacus, the item itself is no longer necessary. If you go on to YouTube, you can find videos of Japanese children who are multiplying and dividing huge numbers remarkably quickly while moving their fingers as if they're manipulating a mental abacus in front of them... even when there is no device.
[00:05:06] They've essentially learned how to use a part of their brain, which is normally used for visual-spatial processing, to process numbers. And, it's remarkably more efficient than what non-abacus users do. Even more amazingly, benefits of learning to use an abacus have a spillover effect. Not only is an expert user's ability to do math transformed more or less permanently, but there are also benefits to language abilities which is very interesting and somewhat surprising.
[00:05:40] Now in contrast, let's look at the digital calculator. A digital calculator is also a very fast, very efficient way to multiply and divide large numbers. And unlike the abacus, it doesn't take many months or years of training to become an expert in using it. It makes it easy to perform complex calculations pretty much right away.
[00:06:05] However, there is a cost. When and if that calculator is ever taken away from a user, regardless of how much experience they had using it, they retain none of the abilities that it gave them. Worse than that, the abilities they previously had have probably atrophied because when using the calculator they're not bringing to bear any of their own cognitive resources in performing the calculations. This makes a digital calculator a competitive cognitive artifact, because it competes with your cognition.
[00:06:37] The modern world is full of competitive cognitive artifacts... whether if it's online driving directions where—instead of looking at a map and navigating—you're simply told when and where to turn, or if it's auto completions on a text or Netflix telling you which movie to watch, instead of figuring out how to decide on your own. There are a lot of cases with machine learning or pretty much any kind of narrow A.I. where something that used to require the user's effort is now being done by a machine in a completely opaque way. And in many cases, it's even opaque to the designers of the system. So, if one were to use traditional programming methods to make a book recommender for example, that programmer would at least need to be thinking about what factors people are considering when they choose a book. But, if it were say, a deep learning set up, then the programmers themselves might have no idea why or how the recommendations worked but they would simply know that the system was very adept at recommending books that people would go on to read or buy or whatever the goal was.
[00:07:48] So why do people use so many competitive cognitive artifacts? Well, I think it's it's pretty straightforward. They're extremely useful, and in the case of the calculator versus the abacus, not only are they useful, but they're easy. They don't take that much work to use and I wouldn't want to abandon all competitive cognitive artifacts, either. In many cases, there's just some tasks that I want done, I don't really care whether or not I am skillful or become skillful in doing it, and the competitive cognitive artifact is just very, very convenient and effective.
[00:08:23] But, I do think it's worth it to at least be aware of the distinction and think about whether if the tool you're using is more like a bicycle for the mind or a self-driving car for the mind.
[00:08:36] If you're interested in this topic and would like to dig a bit deeper, then I would recommend checking out David Krakauer's work. He's the president of the Santa Fe Institute, in New Mexico, which is a very interesting Center for Interdisciplinary Studies. He's a complexity scientist, and he's written a lot about this topic. He also touched on it in an excellent podcast interview I can recommend from the Making Sense podcast, formerly known as the Waking Up Podcast, episode number 40. And, you could also check out Donald Norman's 1991 paper, "Cognitive Artifacts". Thanks for listening to The Metacast.