The Language Game
It seems to me that usually scientists such as physicists or chemists do not take philosophers very seriously who write about the philosophy of science or of physics. I was surprised to find that in linguistics this does not seem to be that case. I have just read a book by Morten H. Christiansen and Nick Chater titled „The Language Game”. The title is already a reference to the philosopher Ludwig Wittgenstein who introduced the term „language game”. Wittgenstein seems to be very popular among linguists and psychologists since for example Michael Tomasello in “Origins of Human Communication” also references him several times. I found it very interesting that many ideas I have read from philosophers such as Quine, Davidson and of course Wittgenstein appeared in this book. The main thesis of the book is that human language is much like the game charades. During communication people try to hint at certain things building on shared background knowledge so that others would catch these clues and guess what they want to communicate to them. They say language evolved gradually by playing charades and reusing moves in the game which worked well before. So they argue against the view that language and grammar are innate. I found this book very convincing on this topic, and I think this way of looking at language is very much compatible with a pragmatist philosophical outlook. Rorty always talked about language as being a set of tools which help us cope with reality, instead of representing it, which he also took from Wittgenstein. This book confirms this view from a linguistic aspect. However since this is a philosophy blog, I think the authors do not follow through the philosophical implication of this picture of language. This would not be a problem at all if they would not actually make some philosophical claims which are in fact contradictory to their view on language.
In one section of the book the authors make a distinction between “Natural world learning” (N-learning) and “Cultural world learning” (C-learning). For N-learning they say “the world provides the yard-stick with which learning is measured”. They give examples such as throwing a ball or a spear to hit a target. The outcome of this is said to be independent of what people think. About C-learning they write: “By contrast C-learning is about coordinating with other people – people who are just us. With C-learning, there is no predetermined universal ‘truth’ about what is right and wrong. The crucial thing is to do what others are doing.” They acknowledge in a note that there are cases where “C-learning is involved in N-learning”. But then they write: “ultimately what works and what doesn’t work in N-learning still depends on the external world and not on whether we do the same thing as other people do.”
I would like to say first that the authors simply don’t need this distinction. Their point is perfectly clear without it, which is that the way we speak has evolved to favour those grammatical practices which worked well and made communication easier. So it is not a miracle that children can learn language since the brain did not evolve to suit language, but language evolved in such a way to suit the brain. The children are participating in the game of charades and not trying to figure out abstract laws of grammar.
I find that the word “learning” makes this distinction very odd. Learning usually means gaining knowledge of some sort, so let’s say the outcome of N-learning is N-knowledge and for C-learning it’s C-knowledge. There is a well-known distinction between “knowing that” and “knowing how”. For example knowing that the first president of the United States was George Washington is a different kind of knowledge from knowing how to ride a bicycle. “Knowing that” means we know a sentence, “knowing how” means we acquire a skill.
Which type of knowing does N-learning and C-learning produce? For C-learning they give the example of the direction people shake their heads for saying “no”. This is a case of knowing that since we can express in a sentence that shaking the head in a given direction means “no”. For the example of throwing a spear as N-learning it seems we are going to know how to throw the spear, which is acquiring a skill. But it seems this is not all to N-learning, in the note they also give an example of learning which plants are edible. This is a case of knowing that a plant is edible or not.
I am going through this because my claim is that it is not really clear what is the role of language in N-learning and C-learning. It seems C-learning is always a case of knowing that involving langage. But N-learning can be either knowing that or knowing how. The problem is that if we have a case of knowing that for N-learning than we need to use language to state what we know, and then we must involve C-learning as well since the language we speak is of course transmitted through the culture we live in. If our culture is such that it has words for elementary particles then we can talk about elementary particles and if not then we can’t. So I think the N-learning, C-learning distinction is not as clear cut as the authors make it out to be. Even if we only consider N-learning as a case of knowing how, the problem is that language even affects perception and the book also talks about such cases when the authors describe how colour perception is shaped by language. So to sum up even the authors themselves provide examples which blur the N-learning/C-learning distinction.
The other problem is that they also make out this N-learning/C-learning distinction to be metaphysical when they say that for N-learning there is “predetermined universal ‘truth’”. Only sentences can be true or false, and if there is predetermined truth then that can only mean that language evolution actually has a goal, which is to reach the predetermined truth so in time we would reach stable meanings. This would be absurd if we think about language as charades, and the authors from the beginning argue that meanings are not stable.
I would like to end by saying that the book is wonderful and insightful; I just picked this topic to criticize because it seems to me the language-as-charades view supports very well a pragmatist anti-representationalism, and in this section the authors miss the philosophical moral of their own work.
