Categories or relationships

Stuart Henshall reports a Dave Snowden keynote. Very clever stuff though I did get lost in one or two places.

What sticks with me most is Dave’s question: What’s the odd one out from chicken cow and grass. Here in the West most of us would say grass but in much of the world they’d say chicken. That’s because we’re trained to filter by categories; elsewhere they filter for relationships. Dave then elaborates on how this profoundly affects they way organisations are run (favouring boxes and isolationism). But the world’s changing and it’s relationships we need to focus on now. That’s my shorthand version but see what you make of it.

8 thoughts on “Categories or relationships

  1. Daryl

    Hi Johnnie,

    We experimented with this in a workshop just the other day. We drew the cow, chicken and grass on a whiteboard and asked participants which one was the odd one out. The results? A quick straw poll indicated about 75% thought it was grass! Interesting little exercise to go through to demonstrate how classification-oriented we are – and to get people thinking about how this orientation effects their behavior and actions.

  2. Stuart Henshall


    Not surprised that this story appeals to you. It’s a good one and the shorthand is all most groups will need… as a prod! which only works on the cow and the chicken.Now is that a relationship or categorization?



  3. sig


    excellent one! Being a bit slow it took me some time before I made the same conclusion – and then had quite some code rewritten – bye, bye tags and other limited “knowledge enhancers”, hello “relationships”.

    Thing is that linking two objects (cow + grass) is like a sentence without a verb – add the predicator “eats” and it makes sense.

    Given that you already have the objects (for subject and object) you only need the predicator so it does not necessarily bloat the system – in exchange you get precise, and not culture-dependent relationships.

    BTW, the cow-chicken relationship would not be direct it would be more (cow, is a, domesticated animal) – (domesticated animal, has member, chicken).

    Hey, that’s the basics of Semantic Web technologies that so the standards exists… 😉

  4. thingamy

    categorising or relationships

    Thanks to Johnnie Moore and his links to Stuart Henshall referring to a keynote by Dave Snowdon I was yet again reminded about the importance of ditching the tags. Yep, the tags, a popular categorising tool still spreading, as an

  5. Nathan Snell

    A great example of our stinted preference towards categories versus organizing thought in terms of relationships.

    I wonder how different the results really would be given different countries. Since Aristotle was the basis of education in the early European universities, as well as Aristotle’s HUGE focus on categories (that’s basically what he did was categorize), it makes sense that’s what we naturally lean toward. It’s kind of what we got started on, and it makes sense (stereotypes are used for the same reason, etc).

    When it comes to the tests done, though, I wonder how much the chicken, grass, and cow example really tell (Granted, I’m no specialist in the field, just giving conjecture). I imagine people are more inclined to understand things on the basis of relationships when those things pertain most to relationships they’re familiar with. For example: If you give the cow and grass question to a farmer, 75% of them may tell you the chicken is out of place. That’s because those are 3 things of which a farmer is in close proximity to and has some form of relationship with.

    Just some thoughts 🙂

  6. Johnnie Moore

    Hi Nathan, thanks for the comment and reflections. And you’re absolutely right to point out that context is vital in understanding these things.

    I think the evidence suggests the chicken-cow-grass anedote is the anecdotal tip of an evidential iceberg. (Oh dear, sorry, getting carried away). The differences between relationship and category thinking are significant and appear quite deep rooted.

    If you’re interested, The Geography of Thought explores this in considerable depth. It suggests that there are significant cultural differences in how we understand the world at a pretty fundamental level. I blogged it briefly here.

  7. Nathan Snell

    Interesting, thanks for the tip on the book, Johnnie! I will have to check it out. It’s definitely an interesting subject.

    So, the question that most obviously follows this deep rooted concept to me is “so what?” Not in the conniving sense, but in the “How can this be applied sense?”

    For example, you mentioned tagging. Is it to say that the system of tagging is entirely western based due to its structural nature? Or is it to say that the system of tagging has been built in such a way that it’s open enough to allow both systems to co-exist in accordance to the user, similar to a piece of paper with lines (I imagine more of the second)? I guess what I am getting at more specifically is what are your thoughts on how it can be heightened?

    Thanks for the response 🙂

  8. Johnnie Moore

    Good questions. Of course this isn’t binary; the West isn’t entirely categorical etc etc. So tags are fine and dandy although I notice I’m getting quite lazy about using them myself lately!

    I think an example of category thinking might be leadership. We use a noun and so seem to think of it as existing separately from the things it relates too. Hence a lot of questionable literature on the subject. But a relational view might see leadership as a verb and something that happens between people, perhaps not led by one over another.

    I used to work in ad agencies which had Creative Directors. Now there’s a use of language with all sorts of assumptions about creativity, and most of them unhelpful. But many’s the time I was people in the Creative Department screaming at anyone in another department trying to “be creative”.


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.