I often say that learning its concepts is a great way in to understanding a subject that is new to you. It’s not the only way, of course, and it isn’t the usual way, at least in my experience. More often I think people approach or explain a subject by outlining what you do, or the reasons why, the structure of it or its history or key figures and events. All perfectly good, by the way – I’d just wave the flag for learning the concepts as well.
Well, if we go back to 1989, I guess, when I was first properly doing what we then called knowledge acquisition interviews with experts for the purposes of modelling their knowledge in AI, I had been taught this approach – discovering and understanding the concepts – by some of the best academics in the field (Breuker and Wielinga and their school at University of Amsterdam). So I tried it right from the start and I have experienced, over the years, that listening out hard for clues as to the concepts helped open up a subject area. The simple example I often give is working with ex-detectives, then employed by financial institutions to prosecute fraud cases, in order to understand how we could better apprehend fraud earlier on in the process. They talked of ‘full payers’ and ‘extended credit takers’. The first are people who pay their credit card off each month, the latter those who don’t clear their credit over longer periods. Now these aren’t hard concepts at all, but grasping them made it suddenly easy to see patterns in the data I might not have done without them. The point is that they, the detectives, had learned these concepts not so much from study but from the other end – from giving evidence in court, not from trying to spot fraud before it was identified, which is where we applied it. The power of the concepts was the ability to see something and reason about it, and, in this case, greatly reduce credit fraud and virtually eliminate payments fraud.
There’s a different kind of case of this that I’m really thinking about today which is a bit more like ‘tip of the tongue phenomenon’. That’s when you can’t think of a word that you’re certain that you do know. You writhe, puzzled by not being able access memory that you’re sure you have somewhere up there. When it comes to you, or is revealed by someone else, there’s a sense of relief and confirmation that, yes, I did know that word, but also a lingering disappointment with yourself that you couldn’t somehow find it. It’s like the catalogue to the library had a missing card (sooo old school!) which would have told you where your book was.
Similarly, I think a lot of education and learning is not about mastering something new, but bringing out and clarifying to you something you already knew. Bringing it out of the ‘unconscious competence or incompetence’ and into the ‘conscious competence’, perhaps. Yes, I always knew that, you think, but it was never clear before. Sometimes a concept has that effect of bringing out something you already knew, but couldn’t quite access. Now it’s in focus, the contrast level has been upped. The phantom steps out of the mist.
An example for me, topically, would be the knowing-doing gap, as revealed in Pfeffer and Sutton’s book of the same name. This is the concept that much of what is regarded as progress-making work is actually, in reality, intellectual by nature: ‘knowing’ rather than ‘doing’ (the vast amount of time, effort and attention lavished on ‘reporting’ in major corporations being a stand-out example). It’s one of those reads that had me nodding and smiling with that relief that (a) it’s not just me, if you know what I mean, and (b) that feeing when something you knew but hadn’t quite conceptualised suddenly takes shape. The tacit, experience-based knowledge, that is there and is informing your judgements and thoughts and feelings and actions, but that, nonetheless, you had trouble explaining, now comes into focus, becomes amenable to reasoning about with your slow brain. Maybe that example doesn’t work for you, but I expect you’ve had the same experience with something else – the “ah-ha” moment.
Sometimes it comes about like that: someone else has figured it out and you ‘get’ it when they explain it. Another way I feel that we can expose and clarify concepts is through conversation. I don’t mean through purposeful ‘meetings’, although any meeting between friendly humans is likely to include elements of conversation. What I mean is taking time to share stories, impressions and experiences with others. As we discuss the half-formed ideas that excite us or trouble us it’s likely that others will have similar, adjacent or even contradictory experiences to share, all of which are like pulling the focus on something that was vague and help to sharpen it up. It’s another reason to value conversation, and a rare case – I mean the case of concepts because of their exceptional enabling power in reasoning – when turning the tacit into explicit is a good move. Have more conversations with fellow practitioners at the limits of what you all understand, about the ideas that excite or trouble you and about which you’re still unclear.
AI4KM - Robert Taylor 