Sometimes on this blog I do sentence diagrams, and they always have a tree-like structure to them like this:
I don’t just label all the parts of speech like this:
I thought it might be interesting to talk about why that’s done. Why draw upper and lower levels? Why can’t sentences be “flat”?
Intuitively, they are flat-looking things. We write out words in a pretty strictly linear order, and speech is necessarily linear. You have no choice but to articulate one sound after another. But this isn’t how language actually works. The way that we process language and the way that it’s mentally represented is as hierarchical structures. That means that there are some parts of the sentences which are “above” and “below” each other, and linguists even talk about certain parts of a sentence “dominating” and “commanding” others.
There are lots of patterns in English that demonstrate how this works, but the example I am going to use is question formation, specifically the pattern known as Subject Auxiliary Inversion. Let’s take a simple sentence:
The man in the corner is smiling.
How do you turn that into a question? You move the auxiliary (is) to the front:
Is the man in the corner __ smiling?
The underscores represent where “is” used to be.
That’s easy enough, but what happens when there is more than one auxiliary? For instance:
The man who is standing in the corner is smiling.
Now what do you move? Do you still move the first auxiliary? Nope, that makes for an ungrammatical sentence:
*Is the man who __ standing in the corner is smiling?
You have to move the other one:
Is the man who is standing in the corner __ smiling?
And you could, in principle, end up with any number of auxiliaries in a row:
The man who is standing in the corner which is shaded is smiling
*Is the man who __ standing in the corner which is shaded is smiling?
*Is the man who is standing in the corner which __ shaded is smiling?
Is the man who is standing in the corner which is shaded __ smiling?
So maybe the rule is that you always move the last auxiliary? That turns out not to work either:
The man is drinking a beer which is overflowing with foam.
*Is the man is drinking a beer which __ overflowing with foam?
Is the man __ drinking a beer which is overflowing with foam?
You can’t formulate a rule like “invert the first/second/last auxiliary”, because where the auxiliary falls in the linear word order has nothing to do with whether it should be moved to the front. What matters is the auxiliary’s structural relation to the subject. You have to move the auxiliary that is attached to the main clause, which is the “highest” clause. And you can only express these relationships with a tree-like structure. If you just label a string of words, then you are missing an important peice of English grammar, and you won’t be able to explain patterns like subject-auxiliary inversion.
Here is a syntax tree for “The man who is standing in the corner is smiling”:
Notice how there are two S nodes in the tree, representing the fact that there are two sentences here: “the man is smiling” and “who is in the corner”. (I’m simplifying by calling them sentences, feel free to argue in the comments.) The auxiliary that gets moved in question formation is the one that is directly attached to the top-most S. The other auxiliary, even though it comes first in the surface string of words, is attached to a different S, lower down, and should not move.
Syntactic structures are an example of what linguists refer to as your “mental grammar”. They don’t exist in the physical speech signal. You can’t look at a waveform and find relative clauses or noun phrases. Listeners and readers have to mentally create the structures that goes “on top of” the linear sentences that they perceive. And it’s a pretty cool trick we can all perform.