R.Remember the tags? I was all over them around 2007 and I’m not ashamed about it. [David’s book, Everything Is Miscellaneous, was published in 2007. —Ed.] Allowing users to apply whatever tags and folksonomies they like to digital content has blown away the limitations of knowledge that we assumed were the strengths of knowledge for thousands of years. The idea that each object has only one “real” tag has been a cornerstone of knowledge for thousands of years. Tomatoes are vegetables, nothing else.
Ok geek, you are right. Tomatoes are actually berries. But you’re just proving what I mean. At least in some contexts.
Before tags, of course, apply multiple taxonomies to things. Books about tomatoes may be categorized into recipes, health foods, and genera. SolanumBut the tomato is also the classic red object: round, delicious, squishy, a source of juice, bad for learning to juggle, and thrown at bad actors and corrupt politicians. Etc.
Second, sites that allow user-based tagging allowed users to tag tomatoes with attributes that were important to them at the time. Where you’ve been, the apps you use, the sites you visit, and the music you listen to. Tags have become so common that awareness has faded since 2007, but clever hashtags can still pop up.
AI in the form of machine learning can apply tags automatically, potentially reducing the need for tags. Already, you can search for photos based on content, color, and even mood, without having to tag them.
Machine learning redefines Tagging
But more than that can be at stake. Can machine learning complete the conceptual job that tagging began, taking us from a definitional understanding of what things are to a highly relevant view? My prediction (my motto: “Someday I’ll get it right!”) is that within the next few years, “dimension” will become an everyday word.
One way of looking at meaning is that a word is what its definition says, as if the definition were a long way of saying what the word says more compactly. But that’s not how we use or hear words.of empire strikes backwhen Princess Leia says “I love you” to Han Solo and he replies “I know”, the definition of those words completely misses what just happened.
Tagging has made it clear that things have very different meanings in different contexts and to different people. Tags make it clear that what (or how) things are depends on context and intent.
Machine learning is making us more familiar with this way of thinking than just words. For example, machine learning models for medical diagnosis are trained on health records, which include a variety of data such as a patient’s heart rate, blood pressure, weight, age, cholesterol levels, medications taken, and past health conditions. There may be history, places, food, etc. The more factors, the more dimensions.
