Why are there grammars in languages in the first place?

Question

I recently took an interest in linguistics and is currently working through the various interesting phenomenon like x bar theory, wh-movement, binding theory, etc. It all sounds very fascinating to me. It amazes me that simple speeches are governed by such complex rules that most people are not aware of.

But why? Why is our speech subjected to grammar in the first place? Am I supposed to believe that when I'm trying to ask a question, my brain secretly draw up a x-bar tree in lightening speed and works out where to place the auxiliaries and wh-words? Is there any theory and hypothesis about the origin of grammar in human languages?

Answer 1

This is a very general question about acceptance of scientific theories, and you can analogously ask if you should really believe that your body is made up of little tiny cells, the cells are made up of really tiny atoms, atoms of made up of even tinier particles maybe even "strings", and so on. People didn't even believe in cells because you can't really see them without a gadget that wasn't invented until very recently (in human experience), and you still can't see quarks.

The primary question, IMO, is whether there is a "thing", grammar, that regulates language production and comprehension. I just don't see any alternative. I can't understand Chinese, Farsi, Hindi or Mongolian, so why not? I seem to lack something. You might object that it's because I don't know these languages, but I'm saying that "knowing the grammar" is a vital part of "knowing a language". If I don't know the grammar of Farsi and start randomly spewing out words of Farsi, Farsi speakers don't understand me and I won't understand them – why?

I personally do not believe that the mind draws syntactic trees at all. I am a phonologist so I don't have the depth of acquaintance with syntacticians and syntactic theories that the stars of syntax have, but I am pretty sure that syntacticians also do not believe that the mind draws, or that physical trees exist in the mind. Instead, a tree is a graphic representation that is easily understood by readers, which embodies more basic claims about human cognition. The challenge for any linguist is to figure out what the actual claim is – what thing in reality is claimed to exist?

For example, virtually any theory of syntax will say that "good chickens" is a (one) grammatical unit, and "chickens from" is not a unit. There are various reasons for believing those claims: let's just take it to be obvious that they are different. There are very many theories of what kind of thing "good chickens" is, for example N, NP, N", and even just "bare unit, no label". They all have in common that those two words are in a relationship, and that "chickens from" are not in a relationship. So you can focus on the reasons for thinking what the words are in a relationship (or not), and not be as concerned with the exact nature of the relationship (the labels and the theory governing those relationships).

Let's then settle on a specific theory of syntax and just accept it. The claim is that that system of rules corresponds to a thing in the mind: it's part of how we produce and understand speech. You believe that there is such a thing for a few simple reasons. You can't understand a stream of speech if you have no knowledge of this how the language is structured, nor can you produce it. You also can judge when grammatical garbage is "part of the language" versus "not part of the language", especially for sequences that you have never encountered before.

There is a current somewhat lively debate over how human language evolved. Personally, I think that debate is sterile and not sufficiently informed by facts – too many firm conclusions based on tenuous data. I think a more useful question that relates to evolution is simply, what is the role of language in human thought? People still believe that there is a secret "language of thought". I don't accept that claim, but that is because I have a different theory of what "thought" is (I don't equate "thought" and "cognition", but many people do). Deep explanations will probably elude us until we get a firmer picture of what "cognition", "reason", "perception" and "thought" are.

Answer 2

One of the most useful features of human language is that there's an infinite variety of statements that can be made with it. We constantly create new utterances, and the audience can understand them (for the most part -- they might not know some of the words we use).

But obviously the brain isn't infinite in size, it can't have neural structures for every possible sentence. This is especially clear since specific languages have to be learned, and you can't hear every possible sentence while learning the language. We know that infants relatively quickly learn the structure of their local language.

So the brain must have general rules that it follows when creating and understanding sentences. "Grammar" is the term we use to describe these rules. Brains in general (even in very primitive creatures) are extremely good at detecting and matching patterns, and this is what we're doing when we learn and apply grammars.

It's also natural for humans to arrange concepts into categories. This is an efficient way to organize knowledge, and it manifests itself in the way we group living things into species, arrange phone books, organize libraries and department stores, etc. I believe brain researchers have found that many of the same structures are involved when categorizing the natural world and processing language.

Much research has been done to try to learn how these processes actually work. Noam Chomsky popularized the theory of universal grammar: there are some general properties of all human languages that the brain is genetically hard-wired to recognize, and during language acquisition we learn how these properties apply to a specific language (a common example is whether the general rule is "adjective before noun" or "noun before adjective").

Answer 3

You know how to eat and how to sleep, without studying medicine. You can get job and earn a living without studying politics or economics. You can use a mobile phone without studying electronics, computing, or quantum mechanics.

In just the same way you can speak without studying linguistics.

Nonetheless, studying these things can help you understand what's going on, if you want to.

Answer 4

A grammar is a set of rules for describing how words in a sentence relate. This is important because some words have multiple 'slots' for relationships that could be filled by the same categories of word, and multiple of the same category of word can be present in a sentence. For example the set of words {The, The, cat, fish, ate}. Each of the 'The's has a slot for a noun to refer to, but which of the two nouns it is attached to is not something that can be inferred from set of words alone. Likewise, the word 'ate' has two relationships, but exactly which relationship (the actor and the acted upon) is which can not be inferred from just the words.

With grammar we can construct two sentences from this set of words with different meaning:

• The cat ate the fish.
• The fish ate the cat.

Both of these sentences are grammatically valid, and consist of the same words, yet by changing the relationships the grammar specifies we have changed the meaning of the words.

To move on to your second question, the brain probably does not construct x-trees, but x-trees are not the only way of expressing a grammar. Going back to the definition, it doesn't mention trees at all. Trees are just a way of expressing these relationships that humans find useful for abstract analysis. There are a number of alternative ways of expressing grammars, but one that is probably worth looking at is the concept of automata.

An automaton is a way of interpreting a sentence using a set of states, and rules for what a word means when the automaton is in a specific state, that steps through each of the words sequentially.

For example with the sentence the cat ate the fish:

1. The first word is 'the', we are now in the state of expecting a singular noun, or adjectives associated with it.
2. 'cat': We have just received the word 'cat' so at this point we know the actor in this sentence is a singular cat, and we can forget the word 'the'. We are now expecting a verb.
3. 'ate': the singular cat had eaten something in the past. We are now expecting a noun phrase to describe what was eaten.
4. 'the': The thing that was eaten was a singular item. So we expect a noun or adjective.
5. 'fish': The thing that was eaten is a fish.

The point of this is not that this is definitely how the brain processes language, but to show that it is possible to process a sentence in a sequential way that applies the grammar without having ever built a tree. Given that it is likely that the brain follows some form of sequential process.

Answer 5

Much discussion of grammar focusses on the words and the relationship between words - perhaps because people studying grammar are mostly linguistics. What often gets missed is that grammar is a feature of language, and language is merely intentional communication.

Communication can be extremely simple. If I throw a rock at random and it hits you, that is not communication. If I throw a rock at you as a warning not to go further, then rock-throwing is communication. If you continue coming forwards and I throw a larger rock as a more serious warning, then we have a grammar of communication using rock-throwing - the larger the rock, the more I want you to stay away.

When communication becomes more sophisticated, then getting someone to do something for you requires you to state what it is. That inherently gives us a verb of what needs doing. And the next clear step after that is a noun of what you're doing it to. At this point you still don't really need a grammar, because you only have two types of word.

Communication quickly gets more complex than that when you add relational concepts though, such as "put meat on rock". Both people then need to understand that the meat goes on the rock, and you're not lifting the rock to crush the meat with the rock. This is the point where your speech is subject to grammar. Grammar evolves from individuals working out a means to communicate, which is why grammatical rules are so damn awkward - they aren't planned, they're just a series of things added to each other over decades, centuries or millennia.

As apes with communication skills, we have literally evolved to do grammar. It's hardwired into our brains, babies babble to test their grammar skills, and the development of language skills in childhood is a process which is fairly well understood. We even have historical examples of deaf/mute children who developed their own personal sign languages to communicate with each other, as examples of how language can be developed spontaneously simply by both sides agreeing the method to communicate a given noun/verb/concept.

Yes, you are genuinely meant to believe that your brain is doing a ton of processing for you behind the scenes, because it is. Is it using the precise method described in your linguistics book? Almost certainly not, because the brain evolves its behaviour and does not run that way. (By analogy, consider the many studies looking at the difference between programmable logic designed by an engineer and comparing them to evolved programmable logic to do the same job.) Always remember that your textbook is descriptive of the end result, and is not proscriptive of the method to get that result.

Answer 6

Grammar is likely nothing more than an emergent property of language. Humans are structured, and so what we do forms structure. Eventually we may study whatever structures we've conjured up and catagorise them. I believe this is what happened with grammar - we eventually create complex structure by accident, through generations of improving upon previous generations' ability to communicate. At some point someone then catagorises this perceived structure as best they can. Once this perceived structure is officially accepted, it is then taught as the very thing that defines the language in the first place.

As a side note, I'm a non-native English speaker and I've found grammatical rules very convenient. Past a certain age, you simply cannot learn language via gut feel anymore, and that's where grammatical rules really start to shine.

Answer 7

Grammar provides additional information about how objects relate to each other. This can also be done with more words. You need less words when you use grammar, resulting often the shorter sentence that can be said in less time.

For instance, since grammatical cases were dropped from the English language long time ago, saying "I soup" gives no clue if you want a soup, are making a soup, want to use it for something, sitting in a soup or consider himself being a soup. Differently, in Lithuanian that still has 8 cases it is possible to say with two words and mostly be understood.

With additional cases Lithuanian language does not longer have it was possible to distinguish if you just move roughly towards the location, approach and stop at the border of the location or cross it into the location. This may be absolutely not the same and may require to be communicated. These forms are still used with some specific words but not with all, and helper words are used instead with the most of nouns. Hungarian language still have these specific cases.

Of, it obviously matter if something is happening now, planned for the future or has been in the past. How can you do without the grammatical tense? Again, it is possible but requires additional words to clarify the meaning.

Answer 8

The thing to remember is that language exists largely (entirely?) to convey "information" from a "source" to a "receiver", and that fidelity of transmission is typically of great importance.

The above is true regardless of whether we're talking about natural languages, constructed languages, or even computer-programming languages!

Grammar evolves (and/or is prescribed) as a set of constraints that promote the fidelity of that information transfer.

Answer 9

Grammar Exists to make Complex Speech Possible.

We use grammar as a way of creating relationships between the words in a sentence, and for expressing additional information without needing to use additional words. So lets pretend you tried to say "He has a daughter with a yellow hat. She went to school with it.".

Let's see what happens if you take away the gender rules:

It has a child with a yellow hat. It went to school with it.

Okay, so we are now less precise, but still readable. But pronouns are also a grammar rule; so, let's get ride of those too.

Johnathan has a child with a yellow hat. Elizabeth went to school with the yellow hat.

This actually made the sentence more clear, but longer; so, we can see grammar sometimes intentionally trades off clarity in favor of brevity, but we are still just getting started. Now let's take away those annoying prepositions and articles we use for establishing relationships between other words.

Johnathan has child yellow hat. Elizabeth went school yellow hat.

Congregation is also annoying, let's get ride of that too.

Johnathan has child yellow hat. Elizabeth go school yellow hat.

So, we've actually stripped away a lot of grammar rules, and still have a pretty readable sentence here. Sure, there is room for confusion, but we can work with this as long as we have a little context. The big reason we can still work with it is that if follows a familiar word order telling us a lot about how these words relate to one another. But when you remove familiar word order you get something more like:

Child Johnathan yellow has hat. Hat Elizabeth go yellow school.

Oh, and let's not forget, the very idea of a sentence is a grammar rule, so why not just say:

school has child yellow has hat hat Elizabeth go yellow Johnathan

So, as you can see, without the structure of grammar, any random arrangement of words would be equally valid meaning there is no way to compose words to be more complex than individual ideas.

Basically, the only way to really get away from grammar is to restrict a language to each word expressing a complete, uncomposed thought. Toddlers (and many animals) communicate this way. A toddler can say "milk" when they mean that they want milk. They can say "up" when they mean that they want you to pick them up. But you can not build a civilization on this kind of language. Part of using a composed language system means that we can explain words that other people do not already know. So, if we had a grammarless language with billions of words to express each idea we could possibly have, teaching it would be impossible because the sentence structure "A is like B but C" can not exist without grammar. You also can't short-hand learn anything with conjugation where you know "picked" is like "pick", but in the past because they have the same root with a known suffix.