A different thread viewtopic.php?f=15&t=5477 has discussed the viability of "proving" certain facts or theorems about the game, and how useful those might be to improve.

A point of view has been expressed that Go is like a language. This I presume is based on the ideas that
(a) Go has a nearly infinite space of possible games, much like a language has a nearly infinite space of possible sentences and stories
(b) Much like children can swiftly pick up a language by being surrounded by people speaking that language (experts), despite not knowing any formal grammatical rules, the best evidence we have so for swiftly improving at Go and becoming an expert is to surround yourself by experts, learning from them in a largely intuitive way in the same way children learn languages.

The alternative idea is that rather than spending many years labouriously learning to become an expert by vague, unpredictable, osmotic fashion, we should rather divert our energy into distilling the game down to some fundamental principles and patterns, or "grammars", that we could learn, and learn to apply, in a dramatically shorter timeframe.

The language idea is quite appealing because it implies that we can use our clearly incredible innate abilities to learn languages, to the point that really skilled human go players will always (?) be superior to computers (at least until they learn to become experts at language). It's a little frustrating, because as someone who learnt the game as an adult, I know that, much like learning a new language, I can never ever be really fluent at that "language".

Any thoughts?

Also consider that most children do not learn languages by being surrounding by "experts". They learn it by being surrounded by average "players". Kyu-level language users if you will. That of course explains why the great majority of us are not experts at our native language.

Also consider that most children do not learn languages by being surrounding by "experts". They learn it by being surrounded by average "players". Kyu-level language users if you will. That of course explains why the great majority of us are not experts at our native language.

I think that, if children had to play and watch games of go every day, from an extremely young age, the way they have to use and listen to their native language every day, from an extremely young age, that most of them would reach either pro level, or at least very strong dan level.

Well, I'm not sure I agree with that. I think we are all staggeringly good at languages. Without even thinking, we can relay a complex tale, full of detail, nuance and often, humour, and no matter how deviant our structure from "strict" grammer rules, essentially all our listeners are able to immediately understand every part of it. We take this incredible ability very much for granted, and compare ourselves to Shakespeare and deem ourselves deficient. But if could play go as well as my seven year old son can communicate I would be a very strong player indeed.

Your other point about other things being akin to the same problem is valid enough. So are they like languages, where they can only be learned from immersion, or are there distillable rules and facts that can short circuit the many years it normally takes to achieve expertise?

I have to scratch my head over the claim that we are all especially good at languages. AFAIK, essentially all 7 billion people currently alive do or shortly will communicate smoothly in their native tongue. So what is so special about that? Doesn't the "special" tag come from comparisons to animals that do not have languages?

Is Go like a language?

If you use "language" in a broader sense that also includes low level languages. Like a Turing machine's "talk" when it writes a 0 or a 1 to the storage tape. When you exercise go, then it uses intersections instead of storage addresses (cells) and colours instead of digits.

Go is like a game tree.

I can put a stone on the board and...

...propose a trade.
...declare my intention to take influence instead of territory.
...ask my opponent to choose a direction.
...show, in a review, why a certain move won't work.

None of these need words. I understand, my opponent or student understands. We communicate.

None of these need words.

A game tree is a very inefficient representation of the game of go.

the difference between 10k and 9d is probably smaller than the difference between 25k and 20k. To those of us heavily involved in the game this doesn't seem true, and we're prone to say that there is more difference between the ranks as you move up (because it requires more time and effort to move), but to the outside world all moderately proficient go players probably seem equally fanatical.

*Some native speakers seem really proficient... so long as you let them talk about their favorite topics. As soon as you push them into a new topic, the stammering, the circular conversations, and the non sequiturs start.
*Some native speakers seem really proficient... so long as they're speaking, and can immediately correct any misunderstandings that they've caused. In writing, they're incomprehensible.
*Some native speakers seems really proficient... so long as you have no idea what they're talking about. If you understand the subject matter, it's easy to see that they're just randomly pulling at cliches that sound impressive.

And so on. All this without either invoking the glories of Shakespeare, or starting a fight about grammar. I think each of these has fairly clear parallels in Go... the first example is like people who memorize a few joseki but flounder outside of them, the second example is like people who are great at life and death if they can play out a few variations on the board but can never get it right in a game, and the third example is like people who make ambitious overplays without necessarily having the reading to back it up.