Difficulty: Advanced   Keywords: Theory

Charles Although people (SL people and others) like definitions, I have reservations about the business of providing them. I wrote at some length about reification in go. That's closely related, really: one may imagine that there are solid rocks about which the stream of play swirls. Defining those correctly looks like the acquisition of some stepping stones into the harder parts of the game.

There are probably others, but some of the pages here related to definitions are

I'd regard these as of a different kind to the many discussions about Japanese terms and usage: what is elusive is on a formal level.[1]

I wonder what the motivation is, to pursue these points. If it is to teach a computer, I just back away (not myself much interested).

It could be that better definitions could help with the business of upgrading heuristics to versions with known exceptions. That can be clarifying, certainly - depending on your level. It is useful to know that certain rules are made to be broken, because they apparently rely on definitions that are simplistic. At a certain stage one may be interested in exactly how is it that 'rules' fail. I think my experience is that this is usually transient: given some idea of the limits of validity, one then goes back to treating a rule as heuristic.

On the whole, therefore, and without wishing to seem obscurantist, I put definitionism on my list of minor study techniques. The flow of the game matters more, most of the time.


RobertJasiek: Precise definitions are good not only for programming and rules application in tournaments but also for such research on strategy and shapes that contributes to solving the game completely in the long run. Even applications for normal teaching can be imagined. - I do not share your failed rules view.

Charles OK - normal teaching is one of my interests. Can it be based on rules, based on definitions? No: mostly that is a failure (can help dan-level players who are missing some concepts).

John F. Precise definitions have their place, but it's a limited one. I think I'm right in saying that no-one has been able to give a precise and easily usable definition of an elephant but even a pre-speech toddler can recognise one. I'm one of those that believes a game of go has something of the elephant about it. Not only do I find it quite unamazing that we post-toddlers manage to play the game with fuzzy rules, I find no difficulty in believing future go players will be able to make the same intellectual leap.

Like Charles, I would encourage contributors to let a subject breathe and not be too concerned with pinning things down.

In linguistics, we often make a distinction between prescriptive and descriptive models of a language. There is room for both even though, as with go, real life language just flows on regardless. My own experience with prescriptive models is that the intellectual effort needed to find rules and patterns can throw up valuable insights, but they are ultimately sterile. The river has moved on. Descriptive models have a wider range. They are split into diachronic and synchronic. In go I'm one of very few diachronics interested in the evolution of terms over time, but there are lots of people willing to talk about how terms are used here and now.

I would therefore suggest that discussions of terms and concepts on SL could usefully follow the synchronous model. That means not fastening on to a Japanese term and trying to define it to death, but trying to list all the equivalent terms in use, in any language, explaining where they are used and by whom, and why, how they differ and overlap. No one has to be a linguist or Orientalist to contribute then. To give a concrete example: I have heard the term pre-atari. I have never really understood it. I certainly have never understood why it's considered useful. Since I believe it may be a Matthew Macfadyen term, I'm worried it may be an advanced players' concept[2] - or is it used spuriously just to give an impression of class? Discussion of this may bring in little tidbits of history, personalities, attitudes to go and teaching, etc. That would be an awful lot more useful (and entertaining) than trying to define elephants.

BobMcGuigan: Definitionism might arise from attempts to codify things so that they can be taught on the basis of principles rather than by subconscious acquisition from repeated examples. As Charles says above, this is often related to attempts to "teach" computers. But humans almost always learn something by, at some point, "getting it", without being given an airtight algorithm. For example, I think all go players learn what a group is without having to use the recursive definitions employed by programs. And we all learn how to play without knowing a proper, complete rule set. The arcane minutia involved in rigorously defining some things detract from actually playing, I think, and for me, at least, playing well is the ultimate goal. I'm reminded of a story about a sports physiologist who decided to try consciously to follow the complete muscle sequence used when we start to walk. After thinking it through he found he couldn't take a step! Apocryphal no doubt but ...

RobertJasiek: If finding all seki shapes is the purpose, the one needs a precise definition of seki. If a go player's practical strategy is the purpose, then knowing the standard seki shapes is sufficient.

Bill: As far as learning go is concerned, learning the terminology can be a great help. But language can take us only so far in our understanding. That is the reason why I have been posting examples taken, not from my personal understanding, but from real go literature about real games or from problem commentary. But language can also confuse, and that is why it is also important to have clear and accurate definitions.

John F. Is this not an example of how language confuses? If you said it was "useful" to have clear and accurate definitions, I would agree entirely and would be happy to support such efforts because I'd see the work as not tied to the end goal - we can have happy diversions and we don't need to worry if we miss the final goal. But if you insist that it's "important" all the effort seems to go into, and be judged solely by, the final goal. I would not enjoy or support such work. My strong impression is that Bill, in practice, supports the "useful" rather than the "important" camp - and as a result I very much enjoy his contributions.

[1] Bill: I think that the Nakade / Discussion is very much in line with attempting to understand Japanese go terms. It arose from the questions of one person, Warp, who wants to understand nakade.

Charles Bill, on the other hand - you are quoting there the work of Landman on eye space. Sure, nakade is a Japanese term, used by English-speaking players. We could call it 'inside play', though, and still have just the same discussion about what inside means.

Bill: Actually, I did not quote Landman, and when he was quoted, I said that his definition was appropriate for math and logic, rather than for the language of ordinary go. But the definition on eye space. given by Dieter, it looks like, agrees with Landman's idea and is a good definition for regular people. There is no problem with the definition on SL of eye space.

One problem arose because nakade was defined in terms of eye space. That was a mistake, like defining a dog as a small domesticated quadrupedal mammal. While true, it is too broad to be a good definition, and could lead to confusion for somebody who did not already know what a dog was. Surely, Charles, correcting such a definition is not definitionism.

And we would not, or should not, get into a discussion about what inside means as far as nakade is concerned. As John points out, nakade occur inside a completely enclosed region.


Charles Pre-atari is a back-formation from some mathematical jargon (presheaf, pre-additive category ...). Just means two liberties left (whatever that means). Despicable stuff, in my view.

Bill: Hmmm. If you have more than two liberties, one might be an approach move, but if you have two liberties, both must be dame. There might be some significance there.

Definitionism last edited by Unkx80 on November 20, 2003 - 07:53
RecentChanges · StartingPoints · About
Edit page ·Search · Related · Page info · Latest diff
[Welcome to Sensei's Library!]
Search position
Page history
Latest page diff
Partner sites:
Go Teaching Ladder
Login / Prefs
Sensei's Library