What are the various possible sources of knowledge and learning? What are some of the most important aspects to consider if we want to be able to measure all learning? [Essay]
Humans are learners, problem-solvers. We learn from one another; by observing nature, making assumptions, by testing those assumptions more or less systematically; we learn by trying, then failing, then figuring out what went wrong, then trying again — trial and error. We sometimes conduct thought experiments, sometimes using logic and philosophy to reach conclusions by delving into abstract aspects of reality. We use our imagination to look beyond the obvious, to conjure up fictional worlds with fictional elements, to think out of the box, and somehow come across a solution that remarkably fits the real world use-case. Mostly, again, we learn from one another. We store our knowledge in clay tablets, papyrus scrolls, books, computers. We pass it on. This collective learning helps us cover vast landscapes of knowledge in a remarkably short time, making the global pool of human knowledge too large and ever-growing to be comprehended by any single human brain. The solutions to our problems lie in solving them together, though not necessarily at the same time and place. Collaboration can take place across huge gaps in space-time. That’s the power of languages and artifacts we created to contain them: books and modern digital memory devices. We may even store the entire growing body of knowledge found across the Internet and private databases in DNA someday.
So when we sit at our desks and decide to learn something, what are the possible sources of learning? And how atomic can they be? Do you need to read the entire book or finish an entire online course to learn what you need to? More often than not, we do not have that sort of time. We only need to find the right information at the right time. So, we rely on search engines, and also computational knowledge engines, to zero in on the bits of knowledge that contain whatever it is that we need to solve our problems at the time. So, instead of locking ourselves in a cave with a personal library of thousand great books for a decade and read every book, we might search online libraries of books such as Google Books or Amazon bookstore to zero in on relevant paragraphs from a few books among millions of books ever written. We reduce the time required to reach the relevant knowledge and thus reduce the time required to solve our problems. Having a vast knowledge in your brain can be helpful to get creative while connecting the dots, but we also forget a lot of what we learnt if we don’t use it, or somehow find it useful enough to retain. This ‘decay’ of knowledge might have a half-life. Keep watering the plants at appropriate intervals or they will dry up and die - this is from Memrise I think. Duolingo has a similar idea of ensuring that we strengthen our recollection of words and grammar at regular intervals, lest we lose our grasp on that foreign language we decided to learn. Imagine you read a thousand great books, and most of that knowledge was not directly relevant to your day job. Wouldn’t you tend to forget or mix up that knowledge over time? When our memories are fading, do our brains consolidate knowledge into wisdom, ingraining this compressed set of ideas into our feelings and instincts, so that specific facts might go but the essence will stay? Who knows... In any case, exposing yourself to a wide range of ideas can be a joy and have intrinsic value. But to solve our everyday problems, we also need means of getting just-in-time knowledge. The right information at the right time. Perhaps our scholarly reading of a thousand books can create a framework of wisdom in our brains, helping us be vigilant while judging the authenticity and applicability of the just-in-time knowledge that we seek from our online searches.
Safari Books Online is a useful service for this reason — digital versions of tens of thousands of books on technology and business are made searchable, so that you can zero in on those paragraphs from the books that are most relevant to you right now. Expert-curated tutorials are also being made by stringing together relevant, thematically coherent paragraphs or chapters from many different books to create meaningful guided learning experiences with a comfortable curve. You don’t have to buy all those books, stack them up on your desk, and read them all cover to cover first, and then come back to solve the problem you are facing. Chances are, given the vast landscape of knowledge out there, that your rigorous reading of useful technology books will still require you to make those just-in-time online searches to zero-in on the information you need right now. Perhaps you will be better equipped with what terms to put in, and which sources to query, with your scholarly knowledge base. Perhaps you will automatically avoid dead ends, or already have a partially formed solution. But you still need to do that last-mile online search and then actually try using the info you find to know if you’ve really solved the current problem.
Which is why, learning by doing and searching smartly are two important aspects of learning and problem-solving in the 21st century. Not that they weren’t before. But this brings me to the first point: to recognize all learning, we need to recognize that books are not atomic units of knowledge, they are divided into chapters and paragraphs - locatable by co-ordinates like word numbers, page numbers, edition number; online courses are divided into video segments, locatable by lecture number and timestamp. Hence, we should make it possible and convenient to reach these smaller areas of information within books and courses through online searches. And to recognize that within the given context, we might learn a lot from just the right paragraph of 100 words than from 10s of books, or similarly, from just the right 1 minute of video than from entire library of online courses. Which brings me to the second point: the context of learning needs to be defined: why are you learning something = what problem are you trying to solve? what do you already know? what have you already tried? This makes it easier (whether for the learners or the system which is designed to help them) to zero-in on the right information bits. Third point is to recognize what the learner did after gaining that information, and the result- to what degree was the problem solved? If problem is solved, cool! If not, iterate. To summarize: recognizing tiny sources of knowledge, specifying the context for learning, and capturing the activities done by the learner— these are the three important requirements if you want to recognize and measure all learning.
Summary
I) Two important aspects of learning are:
Just-in-time knowledge acquisition by perform searches on online databases to solve a problem at hand. (search engines, computational knowledge engines, libraries etc.)
Learning comes from 'doing' (as opposed to just 'consuming info'), or in other words, from problem-solving.
II) To measure all learning:
Tiny sources of learning need to be recognized - more often we'll just read a paragraph or chapter from some book than the entire book, or similarly watch only one video or part of it, in order to get the required information.
The context of learning needs to be specified - meaningful learning experiences occur in a purpose-driven framework.
Steps taken by the learner after gaining information from the sources needs to be accounted - we can't just settle for the sources from which the learner consumed info, we also need to know the learner's response if we want to capture the complete learning experience.
